Q1 2017
ptq PETROLEUM TECHNOLOGY QUARTERLY
REFINING GAS PROCESSING PETROCHEMICALS
SPECIAL FEATURES OUTLOOK FOR 2017 AUTOMA AUT OMATION TION & CONTROL ENVIRONMENTAL
We create chemistry that lets individual needs love global innovation.
As the global leader in catalysis, BASF draws draws on the talent and expertise of more than 1,100 researchers working in close partnership with our customers. This collaboration results in innovations that drive new levels of performance and achievement, today and over the long term. When global catalyst innovations help our customers become more successful, it's because at BASF, we create chemistry. www.catalysts.basf.com
We create chemistry that lets individual needs love global innovation.
As the global leader in catalysis, BASF draws draws on the talent and expertise of more than 1,100 researchers working in close partnership with our customers. This collaboration results in innovations that drive new levels of performance and achievement, today and over the long term. When global catalyst innovations help our customers become more successful, it's because at BASF, we create chemistry. www.catalysts.basf.com
ptq PETROLEUM TECHNOLOG Y QUARTERL Y
Q1 (Jan, Feb, Mar) 2017 www.eptq.com
3 This internet thing Chris Cunningham 5 Outlook for 2017 11 ptq&a 33 Crude oil sourcing: price and opportunity Misha Gangadharan, S D Pohanekar and M D Pawde Hindustan Petroleum Corporation Limited 41 Cuba’s oil: due for development Amaury Pérez Sánchez University of Camagüey 49 Control system security Sinclair Koelemij Honeywell Process Solutions 55 Data operations transform fuels value Craig Harclerode OSIsoft 61 Profiting from plant data Douglas White Emerson Automation Solutions 71 Optimal processing scheme for producing pipeline quality gas Saeid Mokhatab Gas Processing Consultant Gerrit Bloemendal Bloemendal Jacob Jacobss Comprimo Comprimo Sulfur Sulfur Soluti Solutions ons 79 Correcting vacuum column design flaws Gary Martin Recon Management Services, Inc. 83 Analysing FCC hot spots Kenneth Fewel Technip Stone & Webster Process Techn Technology ology 91 Balanced distillation equipment design Soun Ho Lee GTC Technology 101 Firing high sulphur fuel Adil Rehman, C Steven Lancaster, Sandeepan Ghosh, Om Prakash Sahu and Pawan Kumar Sharma KBR Technology 107 Combating reactor pressure drop Ankit A Jain and Ajay Gupta Reliance Industries Ltd 113 The design temperature of flare systems Paul David Paul David Process Ltd 119 Laser scanning with dimensional control Peter Field Warner Surveys 124 Identifying contributors to flaring P Sridhar Indian Oil Corporation 126 Technology in Action
Cover Triple flare of the hydrocracking unit at Galp Energia’s Sines refinery, Portugal. Photo: Galp Energia
©2017. The entire content of this publication is protected by copyright full details of which are available from the publishers. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means – electronic, mechanical, photocopying, recording or otherwise – without the prior permission of the copyright owner. The opinions and views expressed by the authors in this publication are not necessarily those of the editor or publisher and while every care has been taken in the preparation of all material included in Petroleum Technology Quarterly and and its supplements the publisher cannot be held responsible for any statements, opinions or views or for any inaccuracies.
ACTION If raising the alky feed and distillate from your FCC unit is a top priority, you’ve come to the right place. That’s because with its continuing record of success, Albemarle’s ACTION is the only commercially proven FCC catalyst to maximize distillate, butylenes, and octane with minimal gasoline loss. Utilizing its unique zeolite and matrix technologies, ACTION has been successful cracking all types of feeds, from tight oil to heavy resid. Your success is too important to risk with unproven alternatives.
Achieve satisfaction and success with the undisputed leader...demand ACTION.
For more information on Albemarle ACTION catalyst or our exceptional portfolio of products and services, call (281) 480-4747 or visit www.albemarle.com
REFINING SOLUTIONS
ptq
This internet thing
PETROLEUM TECHNOLOG Y QUARTERL Y
Vol 22 No 1
Q1 (Jan, Feb, Mar) 2017
Editor Chris Cunningham
[email protected] Production Editor Rachel Storry
[email protected] Graphics Editor Rob Fris
[email protected] Editorial tel +44 844 5888 773 fax +44 844 5888 667 Business Development Director Paul Mason
[email protected] Advertising Sales Office tel +44 844 5888 771 fax +44 844 5888 662 Publisher Nic Allen
[email protected] Circulation Jacki Watts Louise Shaw
[email protected] Crambeth Allen Publishing Ltd Hopesay, Craven Arms SY7 8HD, UK tel +44 844 5888 776 fax +44 844 5888 667
PTQ (Petroleum Technology Quarterly ) (ISSN No: 1632-363X, USPS No: 014-781) is published quarterly plus annual Catalysis edition by Crambeth Allen Publishing Ltd and is distributed in the US by SP/Asendia, 17B South M iddlesex Avenue, Monroe NJ 08831. Periodicals postage paid at New Brunswick, NJ. Postmaster: send address changes to PTQ (Petroleum Technology Quarterly), 17B South Middlesex Avenue, Monroe NJ 08831. Back numbers available from the Publisher at $30 per copy inc postage.
I
n the nal quarter of 2016 there was a blizzard of announcements about systems and shared technology deals from the big hi tters of plant automation, all about shaping the future of the industrial internet of things (IIoT). It is a concept that seems to have arrived at a tipping point, just about where the plain old internet was in the 1990s, and is set to transform the value of plant data. PTQ was a guest of the Honeywell Users Group EMEA conference, an opportunity to hear about some real experience with IIoT, at the time of this perfect storm of automation news. In some quarters of the world of rening and petrochemicals the internet has yet to catch on in any form. The internet is too big a security risk, is it not? Much bigger issues surely are where to nd the next generation of plant engineers and how to avoid losing the experience gained over the years by retiring senior engineers. An opinion shared at the Honeywell meeting is that the IIoT, its cloud capture of data and the possibilities of analytics can go a long way to solving these issues of employment and expertise. Historically, that expertise is likely to have been acquired over the years in a single plant. Taking the case of a rener with a eet of producing sites, perhaps one involving several regions of the globe, what the IIoT offers (and to simplify somewhat) is a control system that captures and uploads key operating data from controllers, sensors and other ‘things’ across those multiple sites to a corporate cloud resource. This compares with each site normally having only its own plant historian as a private repository of performance data. Data held in the cloud is then available to engineers across the company to perform analytic routines that can interpret issues that arise – or, more importantly, could arise – during plant operations. This cloud historian combined with data analytics in effect represents expertise of a very different order, even compared with a senior engineer’s career-long experience. So, for instance, a plant engineer could explain, on the basis of company-wide operations over the years, why ooding occurs in the CDU. Better still, that engineer can interrogate years of multi-site records to determine which combination of key performance indicators could lead to ooding – a matter of envisioning the invisible. Regarding employment, the aim of any rener is to hire the best and keep them. The next-generation plant engineer, after all, wants to run applications on the internet and wants to know why the tools available at work in the plant might be inferior to the ones he or she runs at home. Then there is the more basic question of place of work. Does a young engineer want to move from ve years of education in a city centre with the prospect of spending some part of the career journey in a remote, neardystopian environment? The world is just not moving in that direction. The facility to control and analyse plant operations remotely, via a multiple site cloud resource, provides the opportunity to install workers far from their reneries, at sites much closer to their personal preferences. And with potential handover of operations across the globe, there is the opportunity for a control room in one continent running a renery’s operations on some distant land mass. Who, after all, wants to work night shifts? CHRIS CUNNINGHAM
PTQ Q1 2017
3
FLEXIBILITY FOR PROFITABILITY DIESEL
MID DISTILLATE
MID DISTILLATE/NAPHTHA NAPHTHA
Z-HD10
Y T I V I T C E L E S E T A L L I T S I D D I M
Z-503
Z-2513
Z-2723
Z-513 Z-MD10
Z-603
Z-2623 Z-MD20
Z-623
Z-FX10 Z-2723
Z-673
Z-FX20
Z-FX10 Z-FX20
Z-723 Z-NP10
Z-733 Z-803
Classic
Tailored
Next Generation
Z-NP20
Z-853 Z-863
ACTIVITY
Criterion and ZEOLYSTTM have developed another generation of specialized Advanced Trilobe (ATX) cracking catalysts that perform with improved saturation, higher activity and higher selectivity, which is already driving value across the globe. ZEOLYST has been producing market-leading catalysts for 25 years, and our scientists have been diligently working to keep raising the bar. NEXT GENERATION Z-2723
Z-FX10
Z-FX20
• High middle distillate yield • High middle distillate and • Paraffinic feedstock flexibility Naphtha yield • Improved hydrowax quality • Max middle distillate yield • Heavier feedstocks • High diesel quality • Improved diesel cold flow • Max liquid/low gas production • Max liquid/low gas production
This new generation is setting the standard in hydrocracking to improve refiners’ profitability. The ZEOLYST TM FX Series is the ultimate choice for converting even the most challenging feedstocks into high-quality products that adapt to seasonal refining demands in hydrocracking. With Criterion Catalysts, you gain our innovative science backed by leading technical service for the quality you need.
Leading minds. Advanced technologies.
www.CRITERIONCatalysts.com
Outlook for 2017
What are the key trends affecting the downstream processing industry in 2017? Senior executives from leading technology developers discuss the challenges and prospects facing refiners. To begin this review of the prospects for petroleum refining, we examine a longrange view of trends in the International Energy Agency’s latest World Energy Outlook.
I
f there is a key conclusion to be drawn from the eliminate net imports of oil by the end of the quarter International Energy Agency’s (IEA) latest predictions century in view. for the oil industry, it is that the demise of petroleum processing is some distance beyond the forecasters’ Prospects for growth horizon. The 2016 World Energy Outlook (WEO) points The rest of the producing world offers a less gung-ho out that the governmental promises on tackling future. Prospects for growth among other signicant greenhouse gas emissions made within the Paris producers will depend on investment decisions that Agreement will not lead to a peak in oil demand before pre-date the tight oil boom. The drop in oil prices hit 2040. In particular, the likelihood of nding sufcient the producing industry in Brazil and Canada particualternative feedstocks for major consuming sectors larly hard, but the IEA posits some reward for investincluding road freight, aviation and petrochemicals ment in those countries over the coming ve years. means that expected growth in just those three In the OPEC camp, production in Iran has risen by activities will far outweigh expected growth in oil around 700 000 b/d as a result of the lifting of internademand globally during the coming quarter century. tional sanctions at the start of 2016, although ofcial Of more immediate interest, the IEA draws up a estimates of further growth in production may prove to changing pattern of global demand for oil. For instance, be ambitious. Iraq has shown the biggest recent gain in India will become the leading source of future growth, output among OPEC members, but declining oil revewhile China will surpass the US as the largest national nues and continuing security issues may also affect the consumer after 2030. nation’s future prospects for production. Even Saudi Returning to the major competition for oil rening Arabia has its economic woes and is looking for opporin its key markets, the number of electric cars world- tunities to develop away from an oil-reliant structure. Towards the end of 2016, OPEC nally conceded that wide is set to grow rapidly, from 1.3 million in 2015 to over 30 million by 2025 and to more than 150 million by it could cut back production to hold up crude prices. 2040. That latter gure displaces 1.3 million b/d of oil But the organisation appeared to require that nonprocessing. Although battery costs will dip to less than OPEC members, in particular Russia, should share half today’s level, the time required to recoup the added the burden and march in step by cutting production. cost of investment in electric cars is sufciently higher However, a more ‘natural’ decline in production from than the equivalent for conventional cars that more existing conventional crude oil elds is likely to have a rapid growth in electric car ownership is held back. more lasting impact than any gathering of energy minThe worst case scenario, from rening’s point of view, isters. The IEA points out that the current decline in is that new support policy emerges globally, leading to output is equivalent to losing the current oil output of faster roll-out, with as many as 710 million electric cars Iraq from the global balance every two years. on the roads in 2040, a number that would displace The volume of conventional crude oil resources which received the go-ahead for development in 2015 fell to around 6 million b/d of petroleum processing. Predictions for the supply side of oil processing its lowest level since the 1950s and there is no current revolve largely around the world’s newest swing prospect for a rebound in output. If this decline were to resource, tight oil. The dip in oil prices of the last two continue deep into 2017, the IEA warns, there is a risk years led to predictions of mayhem in production in of a return to volatile oil prices: “There is no reason to the US tight oil industry, in particular through 2016. assume widespread stranding of upstream assets for But the fracking industry proved to be far more resil- oil, even in a drive to decarbonise the energy sector… ient than many expected. In turn, this means that a provided governments pursue unambiguous policies stable recovery in global oil prices will not have any to that end. Investment in oil and gas, albeit at reduced dramatic impact on tight oil’s development. According pace, remains an essential component of an orderly to the WEO, output from tight oilelds will reach transition to a low-carbon future. But abrupt changes a high point in the late 2020s, peaking at a little over to climate policymaking or misjudgements of future oil 6 million b/d. With greater fuel efciency and fuel demand by the oil industry could lead to more severe switching on the nation’s roads, the US will effectively nancial losses.”
www.eptq.com
PTQ Q1 2017 5
Refining capacity and runs by region (million b/d)
Capacity Net capacity Refinery runs 2015 change to 2040 2015 2025 2040 North America Europe Asia
21.3 16.5
-0.3 -1.2
19.3 13.7
18.2 11.8
16.0 10.0
OECD Asia Oceania China India Southeast Asia
7.6 12.8 4.4 5.0 6.2 8.8 3.3 2.1 4.9 94.8
-0.9 4.9 3.4 2.6 0.1 4.3 1.7 0.8 0.5 16.1
6.8 10.8 4.6 4.0 5.5 6.4 2.1 2.0 3.4 79.7
5.8 12.6 5.3 4.8 4.9 10.1 3.1 2.3 3.7 84.1
4.8 14.8 7.6 6.6 4.6 11.6 4.1 2.7 3.5 87.6
Russia Middle East Africa Brazil Other
World
Source: International Energy Agency
Table 1
Better for refiners Although upstream operations have felt the pain of collapsing oil prices, the news has been far from disastrous for the rening industry. Margins have reached their highest levels for some years as reners beneted from cheaper feedstocks while at the same time managing to prop up product prices. Even the European rening industry, where falling local demand and competition from exporting reners mean that the region has been on a steady decline for the past decade and more, has enjoyed a relatively good two years. For all that talk of European decline in rening, it is as well to note that about a third of the global increase in rening through put during 2015, amounting to about 2 million b/d, came from reneries in Europe. However, the IEA warns, higher margins and higher throughput are all very well if supply is in balance with demand. Even though recent growth in demand for oil products was at its highest for the past 10 years, exclud ing a post-recession bounce, it still was not enough to absorb all of the extra output of rened product in Europe. Added to this, some of the increased demand in recent years for petrochemical feedstocks has been met by natural gas liquids products including ethane and LPG. These are materials which are for the most part outside the scope of rening. Rened product stocks increased throughout 2014 and 2015 and started putting pressure on the margins at the beginning of 2016. As a result, reners in Europe have no option other than to cut back throughput to below their 2015 levels. This pull-back in throughput has not been limited to European reners. Such is the inated over-capacity in global rening, a slump in crude oil prices brought on by over-production upstream, coupled with a period of strong demand pushing up product prices, cannot deliver upward movement in rening margins for ever. The WEO also states that the headline growth in demand for oil distorts the picture for reners since nat ural gas liquids processing and other non-rened prod ucts claim increasingly higher market shares.
6
PTQ Q1 2017
Global growth For the long term outlook, global rening capacity will grow by over 16 million b/d by 2040, but global gures do not paint a picture of the winners and losers who will contribute to the net gure. Growth will be in large measure down to expanding renery throughput in the Middle East which is forecast to double as the region seeks to match rapidly growing domestic demand for rened product (set to grow by 1.3% compared with a global average of 0.4%) as well as the export markets it aims to win through lower-cost production and rening structures. Despite their rapid expansion of domestic rening, both India and China are expected to become net rened importers of oil product, even though the two nations will account for 3 million b/d and 4 million b/d respectively of the 16 million b/d global gure (see Table 1) Companies both from oil exporting regions such as the Middle East and Russia and from signicant cen tres of rening such as Japan and Korea are expected to start looking for downstream partnerships to expand their businesses. This search for markets leads to sub-Saharan Africa, where a hefty decit in transport fuels demands extra rening capacity. At the same time, rening capacity elsewhere will become surplus to needs in view of lower demand and renery runs. Almost 15 million b/d of capacity is at risk of closure during the next 25 years, with Europe accounting for a third of the total. Chris Cunningham is the Editor of PTQ
Eric Benazzi
Vice President Marketing & External Communication Axens In 2016, global growth should fall to 3.1% before rising back up to 3.4% in 2017 according to IMF’s latest estimations. Long-term per spectives in industrialised countries are moderate and a rming up of growth is forecasted at mid-term in emerging and developing countries. These forecasts rely on several important hypotheses including, in particular, the grad ual slowdown of the Chinese economy with growth rate close to 6% at mid-term. The Chinese govern ment will have to manage the transition from a growth driven by investments and exports to a model based on domestic consumption and development of services without leading to a major economic slowdown. India, another heavyweight in terms of global demography, should continue growing at a sustained rate of 7.6% in 2017. Growth in demand for crude oil and products associated to its extraction should be close to 1.2% to reach 97.5 million b/d in 2017. Uncertainty on crude oil over supply should remain and will equally depend on the
www.eptq.com
COMPLETE SOLUTIONS FOR YOUR REFINERY CHALLENGES Today’s Refinery Challenges
Processing tight oil Managing stringent sulfur limits Monetizing orphan streams Upgrading residuals
CB&I’s Comprehensive Solutions CB&I’s broad portfolio of both refining and petrochemical technologies, combined with our execution expertise, will help you maximize processing flexibility and achieve margin benefits in the widest range of scenarios. We are with you through every stage of the process plant life cycle, from feasibility studies to identifying plant optimization and upgrade solutions, through technology selection, full-scope EPFC, commissioning and start-up. PROCESS PLANNING AND DEVELOPMENT LICENSED TECHNOLOGIES AND C ATALYSTS FULL�SCOPE EPFC SERVICES PROJECT MANAGEMENT AND CONSULTING AFTERMARKET SERVICES
A World of Solutions Visit www.CBI.com
OPEC’s policy about its level of production and the tight oil level of extraction itself related to crude oil prices. As a nal result, crude oil prices should continue to uctuate between $40 and $60/bbl during 2017. This will contribute to maintain investments in exploration and production at low levels and could translate to a production shortfall as we approach 2018-2020, leading to a new oil crisis. For more than 20 years, gasoline and diesel have been gradually reformulated and their sulphur content reduced to comply with legislation mandating decreased emissions and lower levels of airborne pol-
Friday 4th November 2016 will remain a historic day for people and the planet with the Paris Agreement entering into force lutants. In 2017, this trend will continue. The US government is enforcing the Tier 3 regulation standards and large US reneries must comply by 2017. To help curb air pollution, which has reached worrying levels in Chinese cities, the government is aggressively implementing China V fuel quality standards. By 2017, on-road fuels should contain less than 10 ppm sulphur and non-automotive diesel will have to comply with China V standards by 2018. Air quality is also an issue in many Indian cities. The Ministry of Road Transport announced that Indian reneries will have to produce Bharat Stage 6 standard fuels, an equivalent of Euro 6, by 2020. Recently the IMO conrmed that the sulphur content of marine fuels will be capped at 0.5 wt% globally with effect from 1st January 2020. Those new regulations will require investment in additional hydroprocessing capacities. Dieselgate, which is witnessing new developments in late 2016, could lead to a partial balancing of the on-road diesel/gasoline demand ratio in Europe which experiences the highest rate of dieselisation worldwide at about 2.3 (v/v). In this ever-changing situation, improving the rening asset’s exibility enabling it to adapt to market trends virtually in real-time will become a must. Technologies, advanced catalysts and digital solutions will be the successful combination for a higher performance and protability. In addition and more than ever integration of rening and petrochemical sites will be required to mitigate risks related to raw material and product price variations, to overcome market trend changes, and to benet from the stronger dynamic of the petrochemical market and improve asset protability. Friday 4th November 2016 will remain a historic day for people and the planet with the Paris Agreement entering into force. With ever-tightening environmental regulations, new signicant challenges that will drive innovation are emerging from the energy transition and climate change.
8
PTQ Q1 2017
Energy consumption represents 34-69% of the renery’s total operating costs depending on its location and the local energy cost. Therefore, every fraction of energy savings in rening or petrochemical units is not only a way to reduce greenhouse gas emissions but also to create value. Our experience as a licensor has taught us that during projects, tight schedule and lack of time mean energy efciency is not sufciently optimised. To take full advantage of our strong knowledge in process and technology, we developed the CEED methodology, standing for Custom and Efcient Early Design. During the early stage of the project, in close collaboration with the rening and petrochemicals operators, alternate schemes are dened, carefully evaluated and the most appropriate option is selected taking into account: energy constraints, capex, opex and plant operability. The decisions made by the next US President and its administration will also be key factors that will have an inuence on the course of events in 2017 and after.
Chet Thompson President AFPM
The rst days of the Trump administration are about to get under way, and we are cautiously optimistic on the outlook for the fossil fuels industry. The contrasts between President Obama and President-Elect Trump are stark. While Obama has been hostile to fossil fuels, Trump has made positive comments about supporting and promoting energy growth through a sensible regulatory framework. For the rening and petrochemical industries, this can only be received as a sea change in our ability to work constructively with the White House and executive agencies. The 2016 elections also resulted in a Republican controlled Congress, creating a newly unied Republican government that could lead to pro-manufacturing and pro-energy policies. Already, the incoming administration and Republican controlled Congress are talking about an issue that AFPM has been promoting for years: the need to improve our nation’s infrastructure, which is aging and reaching a desperate state of disrepair. A safe and modern infrastructure allows our reneries and petrochemical manufacturers to get feedstock to our facilities and to distribute our products to consumers. Unfortunately, for the last several years, infrastructure projects have been politicised, and often in reckless and dangerous fashion. Under the new administration, we are optimistic that the ‘keep it in the ground’ fringe will be sidelined and the potential for a comprehensive bi-partisan infrastructure bill is real. Of course, optimism throughout the oil and natural gas industries must be tempered with the reality
www.eptq.com
that government is slow to move, bureaucracies can be intractable, and policy impacts take time to be felt. That is why representatives of our industry should not take anything for granted. We need to be as assertive and aggressive in our public affairs efforts, as we have been the last eight years, to achieve our policy goals. First and foremost, we at AFPM are eager for our nation to have a rational debate about a sustainable energy policy. One that includes the realisation that
In the coming years, we expect to have the need for up to two million skilled workers – from welders to PhD chemical engineers energy is vital for modern day life, and the recognition that the world is going to use petroleum for decades to come. We will win this debate. The facts are on our side. But our challenge will be to convert this to the development of sound policies that balance the need for affordable and reliable fuels, and a growing economy with sound environmental policies. In 2017, we will continue to ght f or repeal of the Renewable Fuel Standard (RFS) or for signicant reform. I am optimistic that we could see changes with this policy this year, due to growing support in the
US Congress, among a growing and diverse group of organisations that include f ood, motor vehicle and even environmental groups and, of course, the new administration. With the United States producing more of its own oil and gas than at any time in history, and producing it in a clean and efcient manner, the reasons for RFS policy are no longer valid. This increased production has made us keenly aware over the past few years of our industry’s great need to focus on attracting a new generation of skilled professionals. In the coming years, we expect to have the need for up to two million skilled workers – from welders to PhD chemical engineers – due to retirements and industry growth. We know that the rening and petrochemical industries offer great careers, with competitive salaries and benets, and are focused on programmes that reach out to a variety of groups and generations that help us spread the word. Lastly, the big issue on everyone’s mind is how will the regulatory culture change under a Trump administration. We believe he will end this government’s war on fossil fuels and take afrmative steps to reduce the regulatory headwinds that have impeded economic growth and job creation. We certainly do not believe we will get a free pass; nor would we want one. But, we do believe we will get reasonable regulations that are good for the economy and environment. One thing I will say is that 2017 is certainly stacking up to be an interesting year.
NEXT GENERATION
VAPOR PRESSURE TESTING MINIVAP VP VISION • • • • •
Premium Vapor Pressure Tester for Gasoline, Crude, LPG Widest Pressure Range: 0 - 2000 kPa Highest Precision and Accuracy Certified for Robustness and Durability Monitor Your Processes Through Cockpit™ Software - 24/7 Access to Multiple Analyzers - Enables Remote Diagnostics, Help and Troubleshooting - Enables Global Results Download, Collaboration & Audit
ACCESS. ANYWHERE. ANYTIME.
Phone +43-1-282 16 27-0 | Fax +43-1-280 73 34 |
[email protected] http://www.grabner-instruments.com/Products/VaporPressure/VpVision.aspx
www.eptq.com
PTQ Q1 2017 9
New 2nd generation HyBRIM™ NiMo catalyst
For significantly better performance, ™ The answer is HyBRIM The new 2nd generation TK-611 HyBRIM™ catalyst for high pressure ultra-low sulfur diesel and hydrocracker pretreatment service. • 25% higher HDS and HDN activity • Unmatched stability throughout the cycle • Superior volume swell for increased profitability Now that’s performance that pays.
Scan the code or go to info.topsoe.com/tk611
topsoe.com
ptq&a Q We are interested in buying new crudes on the spot market. How do we assess the potential side effects of processing them in refinery units?
A Andrea Fina, Technical Coordinator Development & Marketing Division, Chimec,
[email protected]
When a renery introduces new or unknown crude oils to its typical crude oil slate, possible side effects that cannot be predicted with conventional tools used by the planning department (crude assay, linear programming model and so on) may arise. Indeed, unlike distillate yields, nished products specications, fouling and emulsion issues induced by the feed blend incompatibility are very hard to foresee. If not predicted to be avoided or promptly solved, once these side effects arise it might be too late for cor rective actions: the consequences could be huge, result ing in economic losses that can eventually wear away the prots linked from the purchasing of opportunity crudes on the spot market. Emulsion stabilisation and fouling deposition in a crude distillation unit (CDU) can originate from many sources, however the most common is asphaltenes precipitation. Asphaltenes are by denition the fraction soluble in toluene and insoluble in heptane, characterised by a polynuclear aromatic core with various heteroatoms in the structure, and some aliphatic chains. Asphaltenes are dispersed in the oil thanks to the formation of a resins-asphaltenes complex; its solubility varies accord ing to the relative contents of aromatics and parafns. It is clear that such a delicate balance can be easily upset, leading to asphaltenes destabilisation and eventual precipitation. For example, mixing a parafnic crude oil with one rich in asphaltenes can lead to asphaltene precipitation. The temperature plays an important role as well: while the solubility of the asphaltenes increases with tempera ture, at the same time the consequent reduction of the oil viscosity allows a higher collision factor of Brownian occulation, hence resulting in possible precipitation. As already stated, asphaltenes precipitation promotes undesired side effects as strong emulsion stability in the desalting vessel and fouling build-up in the CDU pre heat train. Flocculated asphaltenes are able to stabilise the waterin-oil emulsion because they are adsorbed at the water droplet interface, avoiding, or at least inhibiting, coales cence of droplets. Indeed, asphaltenes are surface active molecules able to reduce the interfacial tension: this leads, together with the creation of interfacial gradients,
-
-
-
-
-
the droplets to behave elastically to tangential stresses. The formation of an elastic oil-water interface is the main factor of emulsion stabilisation. As a result, a sta ble water-in-oil emulsion leads to: High water content in the desalted crude oil with con sequent increase in energy consumption High salts content in the desalted crude oil, thus an increased risk of corrosion in the main fractionator overhead system, increased fouling deposition in the crude distillation unit and in downstream processes and increased catalyst poisoning of residue hydrocrack ing (if present) High hydrocarbon content in the brine water that can affect operations at the wastewater treatment plant. On the other hand, fouling deposition in the heat exchangers train reduces the overall heat transfer ef ciency, and consequently the heat recovered from the hot uids in the preheat train. This translates into a steeper furnace inlet temperature (FIT) decay: a higher fuel consumption is then required to reach the distil lation temperature. For instance, a CDU processing 10 Mt/y of crude oil, a nFIT reduction of 10°C/y has an energy cost of 2500 toe/y. Finding a reliable tool to predict in advance the asphaltenes’ potential instability, in any possible blend the rener can process, is a keystone to anticipate and handle the related issues. For this reason, Chimec built in new research on crude oils and their blends to develop the Crude Oil Stability Model (COSMO), a tool able to predict the stability of crude blends in terms of asphaltenes precipitation risk, starting from the stability analysis of each blend’s components. The COSMO program is based on the Chimec Stability Index (CSI), a value that directly indicates the intrinsic stability of the asphaltenes in the oil matrix. CSI is measured with ASTM D7157, by means of a Rofa Fuel Stability Analyser, so it can be applied to any hydrocarbon stream containing asphaltenes, crude oils, gasoils, residues and fuel oils. Thanks to the CSI it is possible to predict both the fouling tendency of each crude oil and/or the tendency in stabilising the emulsion. Long research efforts both in Chimec laboratories and directly on the eld led us to develop a huge database of CSI for the most common crudes processed. Reneries seldom process a pure crude oil; indeed the feed blend can usually consist of three to 10 crude oils. The stability of the feed blend can be calculated start ing from the stabilities of the individual blend’s con stituents thanks to the rigorous COSMO model. Mixing two or more crude oils can lead to a mixture that is less -
•
-
•
-
•
-
-
-
Additional Q&A can be found at www.eptq.com/QandA
www.eptq.com
PTQ Q1 2017 11
more recent assays in order to have more accurate results. CSI Real Prior to the utilisation of a LP 80 CSI Cosmo model, crude properties should be x e 70 reviewed using the selected crude d n i assays. Properties such as cut yields, 60 y t sulphur content, density and TAN i l i 50 b (acidity) will allow a comparison of a t a new crude oil with the usual crude s 40 c oils processed in the renery. Then e 30 m distillation cuts can be calculated i h 20 using specic software, critical prop C erties for process units feedstock such 10 as sulphur content, mercaptans in ker0 osene, cetane index in the gasoil cut, 0 10 20 30 40 50 60 70 80 90 100 metal content in vacuum gasoil can be Isthmus, vol% reviewed as they have an important impact on product properties, metal Figure 1 Asphaltenes stability of a crude mixture lurgy, catalyst life cycle and so on. From this rst phase of the crude stable than all its components. This is due the non- evaluation the conclusion may be that the concerned linear behaviour of asphaltene stability. For instance, crude cannot be processed individually but in a blend mixtures of Isthmus and Western Desert have a lower with other crudes available in the renery. In this case, stability than the pure crudes (see Figure 1). a LP model will be able to optimise the crude blend It is important to highlight that the major part of suitable for the renery while maximising renery crudes exhibit a behaviour on blending similar to the margin. one showed by Isthmus and Western Desert. Regarding process units, the LP model will take into Thanks to the predictions of COSMO, it is possible to account capacity constraints or other hydraulic limitacalculate the stability of a blend whatever its composi- tions as well as operating modes. Ranges of variation tion is. Once all the crude oils processed have been ana- on cut points for the specic crude units are also intelysed in terms of CSI, no other analysis is required until grated in the LP model. a new crude oil is processed. When this occurs, CSI The main LP model outputs are the optimal renery analysis of the new crude(s) is required only. margin, routings and unit operating modes when proThe CSI calculated by means of COSMO for a cessing the new crude oil (individually or in a blend). selected blend can be used to: • Continuously evaluate the asphaltenes deposition A Tim Olsen, Refining Consultant, Emerson Automation tendencies of the CDU feed Solutions,
[email protected], and Jake Davies, Global • Predict the behaviour of blends planned for the Marketing Director, Permasense,
[email protected] future, or even to dene a proper one to minimise the First, be sure to understand whether the opportunity risk of asphaltene precipitation crude you are interested in buying is from one source or • Optimise the demulsier and antifouling dosage a blend of two of more crude oils. More and more renaccording to the real emulsion and fouling tendency of ers are nding they have purchased a blend of crude oils, and many times this blend is considered incomthe crude oil blend. patible (crude unit preheat exchanger fouling accelerA Miren Carro-Usle, Studies and Support Group Manager, ated with asphaltene precipitation). This was a topic of Performance Programs Business Unit, Axens, Miren.CARRO- discussion at the 2016 Crude Oil Quality Association
[email protected] (COQA) meeting in Houston. On the spot market crude oils are traded on short terms. If the crude oil is from a single source (and not selfThe advantages are the attractiveness of the price as the incompatible), hopefully the crude oil properties match spot market is inuenced solely by supply and demand. the design and capabilities of the renery conguraAnother characteristic of such a system is the place- tion (sulphur content and major cut yields of naphtha, ment of oil for disposal around the world no mat- distillate, gasoil, and resid). More than likely, a single ter what country it comes from. Therefore, crude oils crude oil will not match the capabilities of the renery on the spot market can be unknown for a renery and and thus need to be blended with other crude oil(s). have to be evaluated in order to assess the renery Crude oil blending introduces the potential for insolumargin that can be obtained and how to process a new ble asphaltenes (incompatible crude blend). crude oil in the renery. With crude oil blending, a rener must learn the maxThe planning LP model of the renery or a less com- imum content of one crude oil that can be blended with plex LP model adapted to crude evaluation is helpful another crude oil without having a signicant impact for such an assessment. The crude assay data will come on the risk of asphaltene precipitation. This includes the from a crude assay library. It is important to select the order in which one crude is blended with the other if 90
12 PTQ Q1 2017
www.eptq.com
Better Technology and Increased Profits
ISOMIX®-e reactor internals design has been setting industry standards based off of more than three decades worth of experience, research, and hard work. It is today’s most innovative and environmentally-friendly design. It not only provides safer and cheaper operations and easier maintenance, it also maximizes catalyst utilization so that you can get the most out of the catalyst you’ve paid for. Furthermore, these state-of-the-art internals increase the run length and unit throughput so that you can be even more productive than ever before. Please contact us at:
[email protected]
done in a single tank prior to being processed. If crude these crude challenges will directly inuence the peroil A cannot be above 30% when blended with crude oil formance of the installed hydroprocessing catalysts B, then you do not want to add crude oil B into a tank and will, as a side effect, trigger that more additives are of crude oil A (rather, just the opposite… adding crude needed for corrosion and fouling control. From a hydrooil A up to 30% into a tank of crude oil B). Percentage processing catalyst point of view, the spot market crude matters, and this is valuable information to determine thus implies more exposure to Ni, V, Fe, and As from the maximum amount use of a discounted opportunity the crude itself and exposure to more Si, P, Ca, and so crude that will not increase the risk of fouling and/or on from the increased additive injection. corrosion. The increased level of contaminants that has to be Emerson has signicant experience with automat- handled by the hydroprocessing catalysts makes it even ing crude unit preheat exchangers to detect when the more important to install the optimal guard and gradrate of fouling has increased. At the inception of the US ing catalysts to protect the main high-active catalyst. tight oil boom when reners on the Gulf Coast started The key to handling high levels of heavy metal and using Eagle Ford, many reners asked Emerson to assist other catalyst poisons is best solved by using alumina with monitoring beyond the traditional monthly Excel based catalyst where the porosity (pore volume, pore spreadsheet manual checks. With WirelessHART (the size and surface area) is designed to allow diffusion and process industry’s rst international open wireless com- storage of the unwanted contaminants. munication standard (IEC 62591, EN 6259)), reners are Being the leading supplier of guard and grading catable to easily add instrumentation around each bun- alysts, Haldor Topsoe has the specialty products and dle to monitor the rate of fouling; fouling is not linear across all bundles, therefore it is important to monitor Buying spot market or each heat exchanger bundle individually. Tight oils tend to have more naphtha and less resid, opportunity crudes is a and the extra naphtha content can destabilise a crude with high asphaltene content that may on its own have great approach to increase a good asphaltene solvency (asphaltenes do not dissolve in crude oil but exist as a colloidal suspension). refinery margins Mixing stable crude blends and asphaltic and parafnic oil creates the potential for precipitating the unstable experience to solve contaminant problems. Dedicated asphaltenes by reducing the solvency capability of the Haldor Topsoe catalysts for trapping phosphorous, sil base crude. It is important to note that asphaltene con- icon, nickel, and vanadium are recognised to be top tier. tent is not directly correlated with asphaltene risk. High Inorganic iron, frequently seen to increase from running asphaltene content may be stable asphaltenes while low acidic crudes, is effectively stored in the inert TopTrap asphaltene content may be unstable asphaltenes; it is product. Faced with the challenges of spot market more of a factor of the solvency capability of the base crudes makes it very important for the rener to concrude. Emerson provides the automation tools to moni- sult with the hydroprocessing catalyst supplier because tor and enable a rener to determine the crude oil blend almost all catalyst solutions are unique and have to be designed to match the specic situation. percentages that produce accelerated fouling. Assessing the corrosion risks of opportunity crudes is very difcult. Most are more corrosive and some idea A Joe Ritchie, Integrated Project Solutions, Honeywell UOP, of the increased risk can be ascertained from the total
[email protected] acid number (TAN). However, there are many corro- New crudes are typically discounted when initially sion risks that are carried in crudes that are not cap- marketed. This presents an opportunity to improve tured in total acid number. Often these corrosion risks margin, but measures should be taken to anticipate are only activated under certain operating conditions, crude behaviour to mitigate risk. Potential side effects for example at elevated temperatures. Operators consid- of a new crude depend on renery conguration and ering opportunity crudes should carry out a risk evalu- prevailing constraints. The further such a crude departs ation for the components of the process equipment, and from the renery design basis, the greater the likelihood ensure that they have reliable corrosion monitoring in there will be adverse side effects. This is why many the areas of increased risk to ensure the impact of that reners run new crudes as part of a larger blend to mitiincreased risk (internal metal loss) to the asset integrity gate risk. Key crude properties are API gravity, sulphur, nitrogen, TAN, pour point, viscosity, metals, salts, and is adequately controlled in service. mercury. Lower gravities impair desalting, potentially A Michael Tinning Schmidt, Product Manager, Refinery increasing corrosion and heater fouling. Lower graviBusiness Unit, Haldor Topsoe,
[email protected] ties or higher nitrogen levels decrease FCC and hydroBuying spot market or opportunity crudes is a great cracker conversions. Higher sulphur levels increase approach to increase renery margins. However, pro- hydrotreater temperatures which shorten run length. cessing the cheaper crudes will impact the renery Higher TANs increase corrosion in gasoil circuits. Pour hydroprocessing units due to lower feedstock qual- point and viscosity will impact oil movements in the ity in terms of, for instance, contaminants, density, tank eld and to the crude unit. Higher metals levels and TAN number (acidity). For the contaminants part, can shorten cat feed hydrotreater or hydrocracker run
14 PTQ Q1 2017
www.eptq.com
At the end of the day, you want a technology supplier who works with you. You’re committed to progress and success. We’re committed to you. And we demonstrate our commitment through licensing world-class refining, gas, chemical technologies and specialty catalysts that drive exceptional performance. You can count on our proven technology and long-term collaboration to help you keep pace with the increasingly complex challenges of today’s evolving marketplace. From initial consultation through plant startup and beyond, our global team offers you practical guidance based on years of real-world operating experience. Our goal is your success. Learn more about how we can work together to grow your business. www.catalysts-licensing.com
. n o i t a r o p r o C l i b o M n o x x E f o s k r a m e d a r t e r a n i e r e h s e m a n t c u d o r p l l a d n a e c i v e d ” X “ g n i k c o l r e t n i e h t , o g o l l i b o M n o x x E e h t , l i b o M n o x x E . d e v r e s e r s t h g i r l l A . n o i t a r o p r o C l i b o M n o x x E 6 1 0 2 ©
Preliminary assay assessment and evaluation of crudes S.No Characteristics
1
Density and API, kg/m3 and APIº
2 3
Reid Vapour Pressure (RVP), kPa Pour point and cloud point, ºC
4
Kinematic viscosity, cSt
5 6 7 8 9 10 11 12 13
Asphaltene, wt% Carbon residue, wt% Salt content, PTB BS&W, vol% TAN, mg KOH/gm Sulphur, wt% Filterable solids, PTB ASTM distillation D-86, ºC Metals (Ni, V, Fe, Ca, Cu, Si), ppm
Significance
Weight to volume and vice versa calculation, checking consistency of crude oil, control of refinery operation and give a rough estimation of crude oil. API gravity of lighter crude oil may be of the order of 45 whereas in heavier API is 10-12. Volatility property of a liquid fuel. For estimating the relative amount of wax present in the crude oil. Cloud point gives a rough idea above which the oil can be safely handled. Viscosity indicates the relative mobility of various crude oils. Temperature has a marked effect on viscosity. Indicates the presence of heavier hydrocarbons in the crude. Measure of thermal coke forming property. Potential corrosion related issues. Potential tower operation issues and energy consumption. Potential naphthenic acid corrosion in high temperature. Impacts the finished product quality and environment. Potential emulsion and fouling issues. IBP, 5% vol recovered, 10% vol recovered…..FBP. Potential impact in downstream catalyst and furnace tube coking.
Table 1
lengths unless vacuum tower cut points are reduced. Marketers can supply references for reners who have processed the crude and those reners will often share their experience. Integrating the crude assay in the facility LP model will provide estimates of unit throughputs and performance. A Parag Shah, VP, Technical Services, Dorf Ketal Chemicals,
[email protected]
compatibility can overcome asphaltene precipitation, which otherwise can cause potential sludge formation resulting in poor tank preparation, rag layer in the desalter and fouling in preheat exchangers. Dorf Ketal has vast experience in treating opportunity and problematic crudes. Various reners around the world have beneted from the crude management programme which includes multifunctional emulsion breaker chemistry, reactive adjunct to address the tramp amine issues, and compatibility studies of over 200 crudes. Specic problems can be addressed separately with respect to a particular crude purchased on the spot market.
Crude assessment and evaluation is a crucial step which can impact renery operations and the quality of nished products. Typically, evaluation can be classied into different types: • Preliminary assay: reviewing key basic properties Reference and distillation data • Short evaluation: physicochemical properties, yield 1. Asian refiners struggle to process light Mexican oil, Hydrocarbon Processing , Article 3445593, 2015. and further characteristics of straight run products • Detailed evaluation: detailed studies of physicochemA Randy Rechtien, Field Tech Services Manager,Randy.Rechtien@ ical properties, yields and side cuts. For purchasing crudes on the spot market, detailed bakerhughes.com, and Gerry Hoffman, Desalter subject matter information/data may or may not be available and hence expert, Baker Hughes,
[email protected] it is wise to review the preliminary assay (see Table 1). The processing of a crude which has newly come to As there are chances of low quality crudes being market offers both opportunities and challenges. The pushed in spot markets at discounted prices, 1 one same can be said of any existing crude which has never should typically look for the following bad actors apart before been processed at a particular renery. In these from preliminary assay review: cases, while there is no perfect method for determining • Organic chlorides: originating from chemicals used future processing risks, there are several key parameupstream, these are known as non-water extractable ters which should be considered. chlorides distilling into the column overhead and side For existing crudes, the best method for identifying cuts, causing potential dew point and under-deposit potential problems is to review accessible documentacorrosion. tion and databases which provide details of industry • Tramp amines: nd their source from amine based experience. In addition, input from additive suppliers H2S scavengers used during transportation of crude; should be solicited. Baker Hughes maintains a benchcan impact the desalter performance and potential marking database of numerous crudes, which includes under-deposit corrosion in column. commonly faced processing challenges. • Hydrogen sulphide: one should typically look for For new crudes where little to no experiences are potential and existing H 2S which contributes to health available, there are four aspects of crude processing that hazards and potential high temperature sulphidic should be examined to help determine risks: corrosion. • Physical properties • Crude blend compatibility: understanding the blend • Production location
16 PTQ Q1 2017
www.eptq.com
Tower Technical Bulletin Random packing uniformity improves performance Background Random packing has been around for over 100 years. Over that time period, its geometry has evolved to improve performance significant ly. Looking at the various stages of random packing evolution, it is clear that structures with a more uniform surface distribution within the packing volume have better performance. Sulzer’s NeXRingTM random packing, designed with this in mind, gives you an industry leading combination of capacity, efficiency, and strength. Why Geometry Matters Column internals tend to flood from localized high velocity regions. As such, designers want to make the restriction across a bed as uniform as possible. So the apparent contradiction is getting uniform flow through a random bed. The solution is to make the individual packing structure as uniform as possible so, regardless of its orientation within the bed, the fluids flowing through the column experience approximately the same resistance.
Column Size Reduction Using NeXRingTM Packing Case Study: Amine Contactor With a typical amine absorber, high performance NeXRing packing can allow designers to substantially reduce costs in revamps and grass roots applications. Grass Roots: Using NeXRing instead of Pall Rings allows for a column diameter reduction of 18% while maintaining the same product quality. In high pressure applications, this can lower the vessel wall thickness by 18% and the vessel cost by 35%. Revamps: Changing from P-rings to NeXRing packing will increase capacity by over 40% while maintaining the same efficiency and product composition. Pressure drop will be reduced by over 50%. In either case, this advanced random packing will provide savings due to improved operating performance and /or lower CAPEX.
Sulzer NeXRingTM Random Packing
NeXRing packings are designed with reinforced ribs that are spaced evenly throughout the packing volume. This uniform mechanical structure ensures uniform fluid flow through the bed. There are other benefits to this design as well. The end flanges combined with strengthened ribs make NeXRing inherently strong. The open structure lowers the pressure drop by 50% versus conventional packings.
Europe, Middle East and India Sulzer Chemtech Ltd. P.O. Box 65 8404 Winterthur, Switzerland Phone: +41 52 262 50 28
[email protected]
www.sulzer.com Please check for your local contact
The Sulzer Applications Group Sulzer has over 150 years of in-house operating and design experience in process applications. We understand your process and your economic drivers. Sulzer has the know-how and the technology to design internals with reliable, high performance.
Asia Pacific Sulzer Chemtech Co., Ltd. 10 Benoi Sector Singapore 629845 Phone: +65 6515 5500
[email protected]
North and South America Sulzer Chemtech USA, Inc. 8505 E. North Belt Drive Humble, TX 77396, USA Phone: +1 281 604-4100
[email protected]
Legal Notice: The information contained in this publication is believed to be accurate and reliable, but is not to be construed as implying any warranty or guarantee of performance. Sulzer Chemtech waives any liability and indemnity for effects resulting from its application.
• Compatibility • Renery equipment/process conditions. A given crude’s physical properties and assay data can provide insight into the levels of contaminants that may result in more serious processing issues. Crudes with low gravity (API< 25°) and/or high solids content (> 50 ptb) place a greater burden on desalter operations and result in poorer salt removal. High solids loading along with high asphaltene levels can lead to preheat train and furnace fouling problems. Crudes with high sulphur content (>2 wt%) and/or high TAN levels (> 1.5 mg KOH/g) will generate more signicant corro sion problems in the crude distillation units and units downstream. When new crudes are produced from geographic areas that already have existing crude production, it is reasonable to assume that these new crudes may be comparable to those that have already been processed. As such, known industry data can be used to predict potential risks for the new crude. The compatibility of new crudes with crudes already being processed must be considered. In particular, there is always the potential for asphaltene precipitation when asphaltic crudes are blended with parafnic crudes. In these cases, the new crude itself may not represent processing concerns. However, the addition of the new crude to the overall blend can have a signicant impact on asphaltene precipitation. To this end, Baker Hughes has developed ASIT (Asphaltene Stability Index Test), which quanties the extent of asphaltene precipitation and allows reners to make proper blending decisions. Crudes can be processed reasonably well under the best operating conditions, but any crude can be problematic under less than optimal conditions. For example, some crudes should be avoided if desalter vessel temperatures are too low to provide acceptable performance. Naphthenic acid corrosion risk may be too severe if only lower grade metals are present in high temperature circuits at the crude unit. As such, a complete understanding of a given renery’s limitations is critical in evaluating whether a new crude will pose processing issues.
feed qualities will affect the unit and its performance. Fortunately, the resilient nature of the FCC means that it can process a wide range of different feedstocks; however it is vital that the unit engineer is able to make adjustments to the operation in order to counteract any negative impacts cause by changes in the feed. For example, increasing the fresh catalyst make-up rate to maintain constant E-cat metals during periods where the feed metals are elevated, or reducing the preheat temperature to maintain a constant catalyst to oil ratio when a higher feed MCRT content results in a higher delta coke. Many reners will have analysis of FCC feed density and metals on a daily basis and the distillation data 2-3 times per week, however less commonly other analysis such as refractive index or basic nitrogen are also very useful parameters to monitor. Basic nitrogen is an important feed quality when it comes to conversion and it is often overlooked. As it is a temporary poison, it can go undetected as it is not concentrated on the E-cat and therefore will not show up on the E-cat analysis like the other contaminants. If the feed qualities have not differed from typical, then there are a few other potential root causes that can be reviewed when trying to troubleshoot conversion loss. Linked with the feed quality is the E-cat qual-
Feed quality should be monitored on a daily basis by the unit engineer in order to preempt how the feed qualities will affect the unit and its performance
ity, however changes to the E-cat quality are not always a result of a change in the feed. The E-cat microactivity and total surface area (TSA) reported on the E-cat analysis should be maintained by continuous fresh catalyst make-up. If these are reducing with time then it is possible that the fresh catalyst make-up is too low, or there is a preferential loss of fresh catalyst from the ASIT is a mark of Baker Hughes Incorporated. unit. It is possible that there is a mechanical problem with the fresh catalyst autoloader, resulting in the target fresh catalyst make-up rate not being achieved, Q What are the chief causes of poor conversion in the FCC and this can be identied by reviewing the full cata unit? lyst balance, accounting for fresh catalyst and additive make- up, withdrawals, reactor side losses, regeneraA Kate Hovey, Senior FCC Technical Service Engineer, Johnson tor side losses and the regenerator bed density and true Matthey,
[email protected] level. A reduction in MAT activity or TSA as a result of The FCC is one of the most complex, albeit interesting, deactivation of the catalyst can be caused by the use of units in the renery which can make troubleshooting torch oil, direct air heater ring or excessive afterburn a conversion loss scenario highly difcult. However, in the regenerator. Similarly, poor stripper operation much like all renery operations, the FCC feed qual - can result in a high hydrogen in coke weight percent ity has the biggest impact on conversion on the FCC. and excessive temperature excursions. The enthalpy of The aromaticity of the feed, its hydrogen content, and combustion of hydrogen in coke to form water vapour the amount of contaminant metals in the feed are a few releases more than three times more energy than the of the things that can affect the cracking chemistry in combustion of carbon to form carbon dioxide. High the riser. Feed quality should be monitored on a daily regenerator losses can be caused by things such as high basis by the unit engineer in order to preempt how the nes content in the fresh catalyst, maldistribution in the
18 PTQ Q1 2017
www.eptq.com
Shell Global Solutions
STRIVE FOR NEW HEIGHTS Take advantage of our technology portfolio to help boost your refinery margins. Static or declining margins? Want to increase capacity or take advantage of opportunity crudes while meeting stringent specifications? By working together, we can help you to respond to today’s demanding environment profitably and responsibly while planning strategically for tomorrow. Ask us how our cutting-edge licensed technologies can help grounded margins take off and reach new heights. www.shell.com/globalsolutions
regenerator, or high bed levels to name a few examples. made a lot easier by having a clear understanding of Assuming the E-cat is looking healthy and the FCC the normal operation, and this is facilitated by carryfeed qualities are typical of normal operation, another ing out regular check runs and recording the unit operpossible cause of conversion loss is due to the heat bal- ating parameters, the feed quality, the catalyst balance ance which directly inuences the catalyst circulation. and E-cat quality, and the full product yields and qualConversion is impacted by catalyst to oil ratio which ities. Familiarity is key when trying to identify change is a function of the riser outlet temperature, the feed on any renery operation, however the complexity of temperature, the regenerator temperature, FCC feed the FCC makes it even more vital to have a clear underrate and riser steam. Insufcient riser outlet tempera- standing of the normal operating parameters. ture, high preheat temperatures and higher regenerator temperatures will slow the catalyst circulation. High A Alexis Shackleford, Technical Marketing Specialist, BASF regenerator temperatures may be caused by insufcient Refining Catalysts,
[email protected] Poor conversion can be the result of a number of things, catalyst cooler duties. On a similar note, the uidisation of the catalyst can but chief causes include poor feed quality, catalyst activplay a role in the conversion. Good distribution of cata- ity and hardware malfunction/failures. The number one lyst in the riser cross-section is important to ensure fast most important factor affecting conversion is feed qualand sufcient feed vaporisation and oil to catalyst con- ity. A rule of thumb is a change in feed density of +0.01 tact. This minimises the thermal cracking which forms S.G. (-1.5 API) causes a 2 wt% loss in conversion. Other mainly coke and dry gas. Additionally, if the catalyst feed quality parameters to look at include Concarbon, uidisation in the standpipes is insufcient, the cata- contaminants (nickel, vanadium, iron, sodium, calcium, lyst ow can become uneven and sluggish. The main nitrogen), hydrogen content (UOP K factor), PIONA cause of uneven catalyst circulation, assuming the ABD analysis (aromatics are harder to crack), and distillation and Fprop of the E-cat are typical, is insufcient stand- curve (especially a high boiling point tail). Next cause of pipe aeration. If the ABD and/or FProp have shifted poor conversion would be low catalyst activity. A rule and the unit is experiencing standpipe uidisation of thumb is one number decrease in FACT (uidised problems, then it is also likely that there will be cyclone activity measured on the E-cat) results in 0.5-1 wt% loss dipleg bridging issues, which will increase the catalyst in conversion. There are a number of reasons catalyst losses. Standpipe uidisation issues can be identied activity could be low from loader problems (not adding by low slide valve differential pressures, or excessive enough catalysts), high regenerator temperature, feed slide valve percent openings, and can be optimised by contaminants, poor stripping, and so on. Another cause adjustment of the standpipe aeration. of poor conversion is a hardware malfunction or failure Of the above described causes of conversion loss, a (though uncommon). A malfunction could include poor couple of the examples could be attributed to mechan- feed atomisation/vaporisation caused by heavier feed, ical damage. For example, damage to the spent cat- incorrect feed/steam ratio, too low feed temperature alyst distributor can cause maldistribution and causing high viscosity, uneven feed distribution through afterburn in the regenerator. Damage to the stripper nozzles, and so on. Other malfunctions/failures to look internals can cause reduced stripping efciency and for would be a refractory failure, steam ring failure, or high hydrogen in coke content. Poor atomisation of air distributor failure for example. If you suspect a hardthe feed and its dispersion in the riser cross-section ware problem, after standard troubleshooting including can be caused by erosion, coking or catastrophic fail- a pressure survey, consider more advance techniques ure of the feed distributors. Additionally, damage to including a gamma scan. the lift steam distributor can impact the catalyst distriA Jeff Knight, Senior Chemical Engineering Manager, bution and density at the feed injection zone. It is possible for the reactor bypass valve to pass, or the feed/ Honeywell UOP,
[email protected] slurry exchanger to be leaking, both of which will Commercial FCC units operate over a broad range of result in FCC feed being routed to the main column conversion values. At a high level, the conversion is a bottoms. And a nal example is coking in the riser or function of the intended operating mode, the feed type excessive lift steam rates which will cause reduced and quality, and the technology that is employed. Poor and potentially insufcient riser residence times. Lift conversion in the FCC unit can be viewed as a relative and feed dispersion steam should be periodically opti- term, as a comparison to its base or expected operamised by the unit engineer by carrying out step test- tion. Opportunities for conversion improvement and ing to identify the optimum rates for the current feed optimisation occur when the FCC unit departs from its intended operating objective or design envelope. quality and unit operation. The chief causes of this poor conversion can be genA few of the main causes of conversion loss have been summarised here, some of which may be more easily erally attributed to four primary categories: cataidentied by the use of one of Tracerco’s diagnostics lyst, hardware, feed quality, and operating conditions. studies, which can identify things like maldistribu- Because of the integrated character of the FCC process, tion in the riser, regenerator or stripper. Tracerco addi- there can be overlap in the categorical causes. The FCC tionally has the ability to check for exchanger leaks as operates at its best expected performance when it is well as identifying mechanical damage in the regen- harmoniously optimised and this requires attention to erator. Ultimately, however, FCC troubleshooting is all four categories.
20 PTQ Q1 2017
www.eptq.com
Keep calm and optimize. Get the most from your rening operation. Linde delivers cost eective solutions to improve eciencies and recoveries. We supply and service a full range of technologies, including PSA, steam methane reforming, amine treatment, furnaces/ heaters, sulfur recovery, o-gas hydrogen and liquids recovery, hydrotreating, emissions abatement and more.
Some of the services we provide:
Eciency improvements → Feasibility, debottlenecking and process studies → Furnace and heater re-lifes → Fitness for service inspections → Revamps, retrots, upgrades and expansions → Engineering, procurement and fabrication → Overall project execution and commissioning → Spare parts → Training → Outage planning and optimization services →
Linde Engineering North America Inc. 12140 Wickchester Lane, Suite 300, Houston, TX 77079 USA Phone +1 281.717.9090
[email protected] www.leamericas.com
When it comes to choosing a compressor, a one-size-fits-all solution isn’t always the best choice. That’s why every Ariel compressor we produce is manufactured to order based on your needs and requirements. ARIEL IS COMMITTED TO YOU.
Learn more about Ariel compressors www.arielcorp.com/For-You/
A Berthold Otzisk, Kurita Europe GmbH, berthold.otzisk@
Analysis of catalyst deposits
kurita.eu
The observation of a poor conversion is usually the result of feed quality changes, operating conditions, mechanical inuencing variables, catalyst properties or combinations of them. Less parafns and more aromatics or an increase of nitrogen, nickel, vanadium or sodium may lead to a lower conversion. Metals in the feed cause thermal deactivation in the regenerator, resulting in a decrease in microactivity (MAT) and a decrease in the surface area. More carbon on regenerated catalyst (CRC) temporarily blocks some parts of the catalytic sites. E-cat and fresh catalyst properties have a very large impact to lower conversion. Changed operating conditions like reduced reactor temperature or decrease in the addition rate of fresh catalyst signicantly lower conversion. Broken stripping steam distributors, plugged or damaged feed nozzles are known mechanical failures. When you observe poor conversion you should check if there are changes in feed properties, mechanical and operating conditions, or catalyst properties. Review the feed properties, check catalyst activity, fresh catalyst availability and source and try to nd out if catalyst to oil ratio, dispersion steam rate or reactor temperature conditions changed. Monitoring the feed nozzles pressure prole can help to observe mechanical failures. Hydrogen in coke is also an indicator for damages. Q We have found heavy deposits of catalyst fines in our FCC unit’s CO boiler stack. What could be causing this and what solutions are available?
A Jeff Knight, Senior Chemical Engineering Manager, Honeywell UOP,
[email protected]
Catalyst nes in the boiler stack can result from a loss in performance in either (or both) the regenerator secondary cyclones and tertiary collection equipment, with the result a loss of ne particle size material to the stack. This ne material will be the result of the natural attrition that takes place due to the circulation/deactivation/thermal cycling of the fresh catalyst. Atypical attrition can result and will produce additional ne or microne material and could be a more likely cause for this sort of stack deposition. A good loss troubleshooting protocol should be executed and this should include a particle size count of the catalyst nes leaving the regenerator, the distribution should be centred on a mean and median size. A bimodal distribution almost certainly indicates an atypical attrition process. A Alexis Shackleford, Technical Marketing Specialist, BASF Refining Catalysts,
[email protected]
Catalyst deposits form most commonly in expanders, cyclones, and third stage separators – which is where the majority of the research has been. Over the course of an FCC run length, there is always enough catalyst on the regenerator side to form deposits. But what makes it form or stick is the important part. So what causes catalyst nes to stick together? Often the ‘glues’ are alkali
www.eptq.com
Sample SiO2 Al2O3 Fe2O3 P2O5 TiO2 La2O3 Na2O CaO SO3 BaO NiO MgO V2O5 K2O PbO ZnO SrO Cr 2O3 ZrO2 MnO CuO
Expander 3ry cyclone Refractory 52.8 49.9 25.1 36.2 36.9 30.9 2.45 5.00 6.18 2.92 2.10 0.14 1.34 1.32 1.08 0.98 1.14 0.02 0.93 0.92 0.25 0.75 0.54 14.1 0.46 0.76 20.4 0.19 0.14 0.00 0.22 0.21 0.02 0.22 0.00 0.68 0.15 0.05 0.04 0.10 0.09 0.76 0.08 0.03 0.00 0.04 0.02 0.00 0.03 0.01 0.02 0.04 0.02 0.02 0.00
0.52 0.02 0.17 0.01
0.07 0.06 0.11 0.01
E-cat 51.6 41.8 0.87 2.53 1.46 0.94 0.27 0.08 0.00 0.00 0.07 0.00 0.09 0.04 0.04 0.02 0.00
E-cat fines 53.7 37.5 0.96 3.41 1.80 1.35 0.46 0.11 0.28 0.00 0.05 0.00 0.06 0.08 0.05 0.02 0.00
0.01 0.02 0.00 0.01
0.02 0.04 0.00 0.01
Table 1
metals such as sodium, calcium, and magnesium, often associated with high sulphur. Iron is another element seen enriched in these samples. For stack deposits, low temperature is a potential cause since the sulphuric acid present can form sulphates deposits that are very sticky. To help understand what is causing these deposits, you should have the material analysed by your catalyst supplier including elemental analysis by XRF. Compare the elemental analysis to samples of your E-cat, nes, and perhaps even refractory to understand what elements are enriched in the deposit and potential sources. In the example here, this expander deposit is enriched with iron, calcium, sulphur, barium and magnesium (a significant amount coming from the refractory). See Table 1. Q We are feeding coker naphtha to our hydrotreater and encountering heavy silicon build-up on the catalyst. Should we be looking for a guard catalyst or is some other solution available?
A Michael Tinning Schmidt, Product Manager, Refinery Business Unit, Haldor Topsoe,
[email protected]
Most coker naphtha hydrotreaters are limited in cycle length due to accumulating silicon deposits on the catalyst. Higher percentage of coker stock gives more Si present in the feed blend and makes the problem worse. It has been known for a long time that the silicone oils added as de-foaming agents upstream the delayed coker will thermally decompose and eventually end up in the coker naphtha fraction. These broken-down silicon species have a great afnity for reacting with the catalyst’s alumina surface, and this mechanism is what causes severe deactivation of the hydrotreating catalyst. To ease the problem, it is absolutely essential that coker naphtha hydrotreaters
PTQ Q1 2017 23
have catalyst systems installed with the highest possible surface area per reactor volume. This technical approach will boost the silicon trapping capacity. Such catalyst systems will include both low and high activity catalysts. Furthermore, they need to be tailor-made to balance the required amount of silicon capacity and deliver the needed activity for removing sulphur and nitrogen from the coker naphtha stream. Haldor Topsoe has conducted research in these types of catalysts for more than 25 years and has recently been able to bring a new type of catalyst to the market with a signicantly higher silicon capacity than seen up to now. The new product line brand is SiliconTrap; it will help the coker naphtha hydrotreater to reach longer and less troubled cycles than with conventional catalysts. A Vivek Srinivasan, Assistant Manager, Technical services, Dorf Ketal Chemicals,
[email protected]
Silicon is known to be the most notorious in terms of catalyst poisoning, as it restricts the catalyst pores, eventually blocking the active sides permanently. A subsequent consequence is reduced unit run length ranging from three to six months. Based on research papers and eld experience, it is seen that silicon can nd its source from crude due to
Though various mechanical solutions are in place to control the poison, it is better to reduce the contamination at source use of silicone based antifoams in oil production wells, which can land in the distillate/residual hydrotreater. However, this is restricted to few oil elds depending on the geographical location. The main source for silicon contamination in naphtha hydrotreaters is antifoams used in the coker unit to suppress coke drum foaming. Based on the quantity of antifoam added, one can expect degradation of silicone into low molecular weight silica fragments in coker gasoil and coker naphtha. Though various mechanical solutions are in place to control the poison, it is better to reduce the contamination at source. Careful selection of silicone chemistry, antifoam injection rate, injection philosophy and monitoring plays a major role. One can adopt the following measures to put a check on the silica content to less than 2 ppm in coker naphtha as a global benchmark. Silicone based antifoams are available in different viscosities. Based on lab scale testing and eld experi ence, it is seen that high molecular weight (high viscosity) products have a benet of lower dosages and high thermal stability to reduce the potential degradation in coker naphtha. Silicone antifoams are always recommended to be dosed in the coke drum overhead with carrier media like LCGO for better mixing and performance. Such
24 PTQ Q1 2017
high temperature antifoams are not advised to be added on continuous basis and one may have to review the injection standard operating procedure based on coke drum cycle and ullage. The antifoam should be typically added with a proper distribution system having appropriate quills installed in the top of the drum and one should ensure that there is minimum distance from the overhead vapour line to avoid as such residual antifoam carryover to the main fractionator. Better selection of level indicators can minimise/optimise the addition of antifoams. Neutron backscatter types have a cutting edge over gamma transmission/ nucleonic types in detecting true foam. A Gary Peterson, Senior Business Leader, Refining, Honeywell UOP,
[email protected]
As the availability of heavier and more sour crudes increases, many reners nd themselves feeding coker naphtha containing high levels of silicon to their reforming and isomerisation units. The traditional approach to removing silicon from these feeds is to stack load silicon trap catalyst having a nominal amount of desulphurisation/denitrication activity in front of the main hydrotreating catalyst. A better option is to use a dual purpose catalyst that exhibits both high silicon capacity and high activity. A Bruce Wright, Senior Technical Engineer, Baker Hughes,
[email protected]
There are several changes that can be implemented to reduce the amount of silicone that ends up in the cracked products from a delayed coker, including mechanical improvements, operational adjustments and implementation of state of the art foam control products: • From a mechanical perspective, the defoamer injection system needs to be 180 degrees away from the overhead vapour lines – the opposite side of the top of the drum. This orientation ensures that entrainment of antifoam injected into the drum is not immediately carried out with the high volume overhead gases. In addition, state of the art foam measuring systems allow for critical assessment of the foam height in the drum. These systems allow operators to see how well and how fast foam fronts are knocked down upon injection of antifoam. • Operationally, there are several items to consider. Increasing the drum temperature reduces viscosity slightly and foaming tendency decreases; however, the benets of this temperature increase can be offset by increased furnace fouling. A carrier stream for the antifoam is a critical aspect to ensure foaming is controlled with the minimum amount of product. The carrier stream helps to carry the antifoam agent to the foam front and should be injected at a rate of 100 parts carrier to one part antifoam. • All antifoam injected into the drum will eventually crack to fragments, distill in the coker main fractionator and add silicon to the coker products. Therefore, minimisation of the amount of antifoam agent being injected to the drum is the key to reducing silicon carryover into the products. Controlled usage of antifoam agent is a
www.eptq.com
Q Is there a chemical treatment for heat exchanger fouling
Silicon in cracked products – Plant 1
Si (ppm) using 60 000 cSt defoamer Si (ppm) using 600 000 cSt defoamer Si reduction
Naphtha
LCGO
HCGO
34.0 12.3 59%
7.9 3.2 59%
7.3 2.7 63%
Table 1 Silicon in cracked products – Plant 2 Naphtha
Si (ppm) using 600 000 cSt defoamer 33.8 Si (ppm) using Foamstop LCI defoamer 8.2 Si reduction 75%
LCGO
HCGO
28.7 3.2 88%
1.8 0.9 50%
Table 2
critical aspect for minimising the amount that goes to the cracked products; therefore, operator condence in the ability of the antifoam agent to rapidly knock down foam and control further foaming can lead to minimising overall usage rates. • There are several antifoam agent options. Polydimethyl siloxanes (PDMS) have been used for many years due to their ability to quickly control foaming. These PDMS additives are available in several molecular weight ranges, classied in terms of the polymer viscosity. PDMS viscosities being used in delayed coker operations range from 60 000 to 600 000 centistokes. If the renery objective is rapid, consistent foam control, with minimal concern over cracked product contamination, then the lower viscosity PDMS additives are the best choice. The industry has been switching to higher viscosity PDMS for several years because these larger molecules stay in the drum and control foam longer before they are completely cracked into volatile fragments, therefore less siloxane is required overall. As a result, silicon poisoning of downstream HDS catalysts is reduced. Table 1 presents Si loading at a renery where the defoamer was changed to the higher viscosity program. Baker Hughes has developed an antifoam additive containing a modied siloxane molecule, the Foamstop Low Catalyst Impact series of products. These products control foaming with less overall silicone, therefore contamination of coker cracked products is further reduced as shown in Table 2. FOAMSTOP is a mark of Baker Hughes Incorporated.
that will avoid taking our units apart at every maintenance break?
A Tim Olsen, Refining Consultant, Emerson Automation Solutions,
[email protected]
From Emerson’s automation perspective, having online monitoring and analysis capabilities around the crude unit preheat exchangers can identify when the rate of fouling increases, thus inferring the crude oil blends are incompatible. Percentages of crude blends matter; the rener can adjust the crude blend percentage to reduce the rate of fouling once detected. The rener can then learn the maximum percentage of a crude oil that does not result in accelerated fouling and share that information with production planners when purchasing quantities of crude oils. A Vivek Srinivasan, Assistant Manager, Technical Services, Dorf Ketal Chemicals,
[email protected]
Heat exchangers in any renery unit serve to preheat the feedstock or to cool down the product for effective energy optimisation. Fouling in such crucial exchangers can lead to operational limitations and unexpected downtime. Reners across the world adopt different cleaning frequencies and methodologies. Some have exchanger bypass provision for online cleaning and some clean the exchanger during annual turnaround or opportunity shutdown. The question is, is it necessary to clean the exchangers throughout the train during a maintenance break which can take signicant man hours and can the same be avoided using a chemical treatment programme? The answer can be explained depending on the type of fouling, feedstock processed, conguration and monitoring philosophy. Basically, fouling can be classied into organic and inorganic. Various factors like temperature, velocity and feedstock properties inuence the organic or inorganic fouling. Table 3 shows an overview of fouling in key exchanger circuits across a renery. Depending on the above mentioned location and type of fouling, one can choose the Dorf Ketal antifoulant chemistry to get key benets like enhanced run length, reduced cleaning and hence downtime. Adopting good monitoring practices to assess the health of exchanger like fouling factor calculation, heat duty and normalised pressure drop can help predict the intensity
Fouling in key exchanger circuits S.No
Unit
Fouling prone location
1
Crude unit
Preheat train and atmospheric bottom exchangers
2 3 4 5
Type of fouling
Predominantly asphaltene (organic) and in some cases influenced by inorganic (metals) Vacuum unit Vacuum column bottom exchangers Asphaltene (organic) Hydrotreater unit Preheat train Organic (polymer formation) and inorganic (corrosion products) Visbreaker unit Preheat train and residue exchangers Organic (asphaltene) + inorganic (metals) FCC unit Main column bottom train Organic (poly nuclear aromatic) + inorganic (catalyst fines)
Dorf Ketal solution
Asphaltene dispersant and metal passivator Asphaltene dispersant Range of chemistries including dispersant and antioxidant Asphaltene dispersant and metal passivator Coke retardant and metal passivator
Table 3
26 PTQ Q1 2017
www.eptq.com
Bringing Unity To Hydroprocessing
Honeywell UOP’s vast hydroprocessing catalysts lineup is called Unity ™ because we bring a unified approach, delivering complete catalyst, equipment, licensing and technical support solutions for hydrotreating, pretreat and hydrocracking.
© 2016 Honeywell International. All rights reserved.
Learn more at www.uop.com.
of fouling in a particular exchanger. Dorf Ketal’s heat exchanger assessment tool not only determines the exchanger performance but is also benecial in strategising cleaning sequence. This way one can avoid the humongous cleaning activity during a maintenance break if online bypass provision is not available. A Andrea Fina, Technical Coordinator Development & Marketing Division, Chimec,
[email protected]
The loss of heat exchange efciency is currently one of the main issues for reneries, as it often happens long before the date of the scheduled turnaround. After one or two years, in some cases even less, the heat exchange loss can be so important as to inuence the performance and the economics of the plant, forcing management to plan dedicated stops necessary to restore the heat exchange efciency. In recent years, there have been several attempts to develop a method to clean the heat exchangers without the extraction of the tube bundles. Many involved the use of chemicals at a high dosage or a chemical cleaner, to be applied through external circuits or using the system’s pumps. Anyhow, due to the poor results, the widespread view is that extraction and hydrodynamic cleaning is still necessary to recover proper efciency in the time frame that separates one turnaround from another.
The loss of heat exchange efficiency is currently one of the main issues for refineries
stop which is necessary to apply the cleaning. In the event that the preventive actions indicated have not been adopted and the fouling phenomena are unusual Chimec is able to propose innovative solutions of online (Gain Back Technology) or ofine (Deko Efciency Recovery) fouling removal. Gain Back Technology has been developed for the online removal of the fresh fouling deposit commonly observed in a crude distillation unit. Its strong dispersion feature can efciently recover the thermo-hydraulic capacity of critical exchangers, maximising the throughput of crude oil and avoiding by-bass procedures. Gain Back Technology is the solution when worsening of fouling trends is detected in real time, hence when the fouling is still soft. The technology is applied by shock on top of the usual antifouling and allows one to fully recover the capacity of the unit in some hours of treatment. Deko Efciency Recovery may be applied when the renery has already experienced the consequences of a fouling deposition. The target of Deko Efciency Recovery is to inject innovative chemicals on the bypassable exchangers, removing the fouling up to a sustainable level. If the exchangers cannot be excluded, a momentary stop of the plant must be taken into account. Our solution allows one to eliminate in the shortest possible time a signicant amount of fouling, with a rapid resumption of operation of the involved exchangers and the immediate recover of the heat exchange and ow rate of the process. The efciency of this solution is the result of the innovative research made to develop proper actives. The characteristics of fouling are not equal in each exchanger. The fouling of the heat exchangers before the desalter is different from those after the desalter. Several chemicals must be selected on the basis of available information considering whether one or more steps are needed. Chimec can apply products to solubilise organic deposits with a high level of unsaturation as well as dispersant actives on both organic and inorganic deposits, using oil based carriers or aqueous solutions. The comparison between applying Deko Efciency Recovery and a shutdown of the plant to extract and wash hydrodynamically the exchangers is very easy: the washing procedure can take 7-10 days, depending on the procedures and the means put in place. It involves removing of the content, blinding, extraction, cleaning, reintegration and restart. All these steps force the system to a stress and risks in each step. Deko Efciency Recovery only requires exclusion and displacement of the uids from the heat exchanger in order to allow the start of operations. A period of 16-32 hours, depending on the type of fouling and of applied phases, can enable recovery of heat exchange efciency between 50% and 70%.
The renery should act in advance to avoid the need for this kind of cleaning, preventing fouling formation through a continuous control of all the elements that may affect the efciency of the heat exchange train. A preventive approach is the only technical solution that can avoid or move over time the need for a treatment solution such as extraction and hydrodynamic cleaning. Chimec’s approach is based on an advanced antifouling programme tailor-made for the unit where fouling is observed. Chimec chemical and service programme is based on continuous injection of a customised antifoulant and includes advanced tools for the prediction and monitoring of the fouling trends (fouling factors, normalised duties and normalised furnace inlet temperature). Furthermore, the company has developed Cosmo, a technology that enables prediction of the asphaltenes stability of a planned crude slate and the related fouling potential. Based on the stability index calculated by Cosmo, it is possible to optimise the renery crude slate and/or the antifoulant dosage for preventing and reducing fouling issues in the unit. If properly employed and implemented, such a A Carl Weaver, Senior Scientist, Baker Hughes, CarlWeaver@ preventive approach allows one to minimise the loss bakerhughes.com of heat exchange and reduce both the extra energy A method of exchanger cleaning often referred to as costs and those related to the eventual extra plant ‘online cleaning’ is one means of removing unwanted
28 PTQ Q1 2017
www.eptq.com
YOUR BENEFIT: LOWEST LIFE CYCLE COSTS
API ��� HIGHEST AVAILABILITY COMBINED WITH BEST MAINTAINABILITY; FAST ACCESS TO ALL WEAR PARTS
Full range: Rod load up to 1’500 kN / 335’000 lbs Power up to 31’000 kW / 42’100 hp Lubricated up to 1’000 bara, non-lubricated up to 300 bara Advanced solutions for demanding and sour gas applications
Over 120 years of experience in valve design and manufacturing Outstanding durability of Redura® sealing elements for longest mean time between overhaul (MTBO) Worldwide Burckhardt Compression quality, engineered in Switzerland Your solution partner – From bare shaft compressors to turnkey solutions
Full range of services and top performing components through global organization and local service centers → www.recip.com/api618
fouling debris and restoring heat transfer without pulling bundles for mechanical removal and hydro-blasting to remove unwanted fouling debris. Given that there is redundant (two or more banks of heat exchangers) equipment, the term ‘online’ refers to keeping the unit in operation at reduced throughput, leaving a bank of exchangers in service while another bank of exchangers is chemically cleaned. The most common method of chemical cleaning the out-of-service exchangers is typically a two-step process. The initial step, using an elevated temperature aqueous circulation of a detergent or micro-emulsion cleaner, rst de-oils and removes organic binders liberating some debris. The next step, using a low pH aqueous solution of an inhibited acid or organic chelant solution, solubilises inorganic scales (for example, iron oxide, iron sulphide and silicates) for removal of unwanted inorganic debris while protecting the base metal of the process equipment. In completion of this step, a passivation rinse ushes the process equipment, leaving internal surfaces in a passive state and ready to be put back in service. The effectiveness of this approach is highly dependent on both the composition of the unwanted fouling debris (% organic material and how much of that might be coke and % inorganic material) and being able to circulate the appropriate chemical cleaning solution. If exchangers have fouled to the extent that there are plugged tubes, then the chemical cleaning solution will not have access to the area to be cleaned. Continuous antifoulant treatment appropriate to process design and the hydrocarbon feed being processed has shown signicant benet in reduced energy costs, extended run length as well as a means of inhibiting complete blockage of exchanger tube throughput which would impede online cleaning efforts. Q
What is predictive data analytics? What benefits can it bring to our unit control operations? And how do we implement it?
A Tim Olsen, Refining Consultant, Emerson Automation Solutions,
[email protected], and Jake Davies, Global Marketing Director, Permasense,
[email protected]
The traditional approach has been to collect and historicise process data, and then only use the data to look back and evaluate after an incident. The new approach is to utilise pervasive sensing and predictive data analytics software to automatically analyse data and turn it into information; this new approach looks forward and alerts before abnormal operation or imminent failure, thus the ability to take appropriate timely action to avoid asset failure. It is important that a rener should not forget the staff training and behaviour changes that are required to act properly on this new analysed information. Predictive analytics utilises both process and asset health measurements that involves extrapolation of historical trends (from previously acquired measurements) into the future. In asset integrity terms, this can be achieved with online wall thickness monitoring. A line of best t through the historical online wall thickness
30 PTQ Q1 2017
measurements can be extrapolated to estimate when, for example, the equipment will need to be replaced. This information helps to better plan timing and scope of turnarounds. See page 61. A
Christophe Romatier, Senior Manager, Strategy & Business Development, Honeywell UOP, Christophe.
[email protected]
Data analytics represents a family of techniques that can be applied to operating data to derive patterns. These patterns can in turn be used to infer or predict certain types of process unit behaviour. Honeywell UOP recently launched a new business, Connected Performance Services, which makes use of such techniques and is paired with our longstanding experience designing these units, their catalysts, and well as assisting our customers in troubleshooting their operations. With Connected Performance Services, customers have
Data analytics represents a family of techniques that can be applied to operating data to derive patterns access to a next generation process monitoring environment that can improve their unit reliability by detecting issues before they become costly. This is an example of data analytics in action and how it can lead to tangible operational and nancial benets. Q The feed flow rate to our sour water stripper is well below design level when flooding starts to occur. Is the problem likely to be foaming or tray damage and how do we investigate this easily?
A Nghia Pham Phu, Principal Engineer, Refinery Process Department, Haldor Topsoe,
[email protected]
The ooding could be caused by different reasons, for example: • Tray damage • Fouling • Trapping of hydrocarbons inside the column. Tray damage or fouling could prohibit the liquid ow downward and start ooding. The ooding will progress upward in the column. The sour water feed could contain a small amount of hydrocarbons due to entrainment. Hydrocarbons in a certain boiling range could be trapped inside the column and accumulate to a high level. These hydrocarbons will vaporise at high temperature close to the bottom and condense at low temperature close to the top, thus circulating inside the column. They accumulate gradually to a high level until foaming and ooding occurs. Flooding, foaming, and tray damage can be detected by gamma scan. A number of companies provide this service.
www.eptq.com
Meet our family... We’ve learned a lot in 70 years And we know precious metals recovery and refining better than anyone! Sabin Metal Corporation
E
Sabin Metal West Corporation
Sabin Metal (Canada), Ltd.
Sabin International Logistics Corporation
Sabin Metal Europe B.V.
Sabin Metal Corporation DMCC BRANCH
A L CO RP O E T R
M
N ach of these organizations is B A part of the global Sabin Metal Group of Companies. Our combined N U M resources enable us to return highest N possible value of recovered and refined precious metals quickly, safely, and with assured peace-of-mind. We recover precious metals from virtually all industries that use them in their processes or products. We’re proud to begin our eighth decade of leadership, and we’d like to share our success with you. Tell us what we can do for you today. I
A
T
I
O N
S
5
P
1
L
A T I
A
N I V E R S A R Y
0
2 5 4 9
• 1
Variability is predictable. What if you could manage variability across the refinery and safely add profit to your bottom line? You can. Our industry-leading integrated solutions leverage decades of expertise, unrivaled customer dedication, and innovative predictive technologies to transform variability into a competitive advantage.
Visit www.gewater.com/calculator to discover your increased revenue potential today.
GE Water & Process Technologies Predicting your success.
Crude oil sourcing: price and opportunity Sourcing crudes for the best refining margin needs to be supported by a detailed procurement strategy MISHA GANGADHARAN, S D POHANEKAR and M D PAWDE Hindustan Petroleum Corporation Limited
T
he world’s crude oil market includes not only spot markets featuring physical transactions but also highly developed paper markets, notably futures and forward delivery, thus forming a very composite market structure in which all these transactions are interrelated and reciprocally influential. The variation in prices between two grades can influence refinery margins significantly if corrective action is not taken at the appropriate time. Therefore identifying and procuring crude oil that results in maximum margins are the industry’s top priorities. In this article, various aspects relating to crude oil procurement for optimisation of refinery margins are outlined. How the oil market works The price of oil is set in the global marketplace. Oil is traded widely all around the world and can move from one market to another easily by ship, pipeline or barge. Therefore the market is worldwide and the supply/demand balance determines the price for crude oil all around the world. If there is a shortage of oil in one part of the world, prices will rise in that market to attract supplies from other markets until supply and demand are in balance. If there is a surplus in a region and the price drops, buyers will soon be drawn to that market. This explains why crude oil prices are similar all around the world. Prices vary only to reflect the cost of transporting crude oil to that market and the quality differences between the various types of oil. The global nature of the
www.eptq.com
market also explains why events people also trade in crude oil as a anywhere in the world affect oil commodity in financial markets. prices in every market. Several They purchase ‘futures’ –essentially key factors influence the oil price, bets on how much oil will cost at a however the four major factors that later date – and this in turn affects help in determining the price of oil how other people think oil should are supply, consumption, financial be priced. It also affects how much markets and government policies. oil the petroleum companies will As per the basic principles of release to the market. Oil trading in financial markets has been growing The way oil is traded bigger than ever in recent years. As a result, speculation has come to shape the price of crude oil more on the financial than ever before. Crude oil trading markets has a in financial markets has a surprising effect on crude oil prices – specumassive influence lators who buy large amounts of futures can swing the price one way on its price or another. Here is an example: a speculator who buys oil futures at economics, prices will be low if a price higher than the current marsupply exceeds demand, and the ket price can cause oil producers reverse applies: an increase in con- to hoard their oil supply so they sumption over supply will lead can sell it later at the new, higher to an increase in prices. However, ‘future’ price. This cuts the current crude oil pricing goes far beyond supply of oil in the market and just supply and demand. The way drives up both present and future oil is traded on the financial mar- prices. kets has a massive influence on its Government regulation also has a price. Similar to the stock market, big impact on oil prices. Rules and
Dubai Dubai & Brent Brent WTI WTI & Brent
Figure 1 Marker crudes by area
PTQ Q1 2017 33
margin point of view can be less attractive after a few months.
14 Brent / WTI differential l 12 b b / $ 10 , l a 8 i t n e 6 r e f f 4 i d e 2 c i r P 0
Brent / Dubai differential
−2
1 4 1 4 1 4 1 4 1 4 1 4 1 5 1 5 1 5 1 5 1 5 1 5 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 n n u p u p J a M a M a J S e N o J a M a M a J S e N o Figure 2 Price differential movements between marker crudes Characteristics of marker crude types
Key property
Brent
WTI
Dubai
Oman
API Sulphur, wt% Pour, °F Acid number, mg KOH/gm
37.5 0.45 30 0.01
39 0.3 -0.4 0.09
31.4 1.54 -11 0.05
30.4 1.45 -33 0.61
Table 1 Price determination methodologies for marker crude oil Supplier
Marker crude
Marker price determination
Month of announcing OSP
Saudi Aramco Oman Dubai avg. B/L month average M-1 Kuwait Petroleum Corporation (KPC) Oman Dubai avg. Nominated month average M-1 National Iranian Oil Company (NIOC) Oman Dubai avg. Nominated month average M-1 State Oil Marketing Organisation, Iraq (SOMO) Oman Dubai avg. Nominated month average M-1 Abu Dhabi National Oil Company (ADNOC) Outright Prices Nominated month M+1 Qatar General Petroleum Corporation (QGPC) Outright Price B/L month M+1 Yemen Oil (YOG) Brent 5 days after B/L date M-2 Egyptian General Petroleum Corporation (EGPC) Brent B/L month M Nigerian National Oil Company (NNPC) Brent Five days avg. after B/L Five days avg. before B/L M-1 Five days avg. after five days after B/L. Sonangol, Angola Brent 5 days around B/L M-1 Sonatrach, Algeria Brent 5 days after B/L date M-1 BPMigas, Indonesia Minas B/L month avg. M+1 Petronas, Malaysia Brent B/L month average M Brunei Shell Petroleum (BSP) Outright price Nominated month M+1 Oman’s Oil Ministry (Oman) DME Oman B/L month average M-2 National Oil Corporation, Libya Brent B/L month average M-1
Table 2
regulation on the sulphur content of fuel could raise demand for low sulphur crude oil. The incentive for fuel efcient cars, development of more efcient alternate modes of transport and so on will lead to
34 PTQ Q1 2017
demand for oil prices to go down, simply because the world will not need it as much. Thus selection of crude oil for processing at a renery is always dynamic and crude oil that was most desirable from a
Pricing and marker crude oil grades The price payable for crude oil is calculated based on the marker crude oil price plus or minus a price adjustment factor, which is set by the seller, or is mutually agreed between the buyer and the seller. While the marker crude oil price varies with events in international markets, including speculations, the adjustment factor is decided by the seller, taking various elements into account like quality difference with the marker crude, transportation cost difference with alternate grades, demand supply balance, competition from other suppliers/ grades, and so on. There are different crude oil markers, each one representing crude oil from a particular part of the globe. The marker crude oil is specic to the market and three major markers used in pricing of crude oil across the globe are WTI (Western Texas Intermediate) for the US market, Brent for the European/West African Market and Oman/Dubai for crude oil grades in the Persian Gulf for the Asian market. The applicability of marker crudes across the globe is shown in Figure 1. The details of each of the markers are: Dubai/Oman: this Middle Eastern
crude is a useful reference for oil of a slightly lower grade than WTI or Brent. It is heavier and has a higher sulphur content, putting it in the ‘sour’ category. Dubai/Oman is the main reference for Persian Gulf oil delivered to the Asian market. Brent Blend: about two-thirds of
all crude oil contracts around the world are marked to Brent Blend, making it the most widely used marker. ‘Brent’ actually refers to oil from four different elds in the North Sea: Brent, Forties, Oseberg and Ekosk. Crude from this region is light and sweet, making it ideal for the rening of diesel fuel, gasoline and other high-demand products. The supply is waterborne and thus it is easy to transport to distant locations.
www.eptq.com
West Texas Intermediate (WTI): WTI
refers to oil extracted from wells in the US and transported via pipelines. The crude is very light and very sweet, making it ideal for gasoline rening in particular. WTI continues to be the main benchmark for oil consumed in the US. The characteristics of these marker crude oil grades are shown in Table 1. Figure 2 indicates how price differentials between marker crude oil move for various reasons, swinging sourcing decisions for a renery.
Determination of marker crude oil price The determination of a marker crude oil price varies between suppliers, and the various ways of determining marker crude oil price are: bill of lading (B/L) month average, ve days around B/L, nominated month average, or as agreed between buyer and seller. Thus while nalising the sourcing of marker crude oil, determination of its price must also be kept in mind. Table 2 indicates marker crude oil price determination methodologies followed by various national oil companies for supply to Asia, as well as the time of announcing the price adjustment factor. Price adjustment factor, official selling price The price adjustment factor for a crude oil grade is adjustment over the marker crude oil price. The adjustment factor is declared before the beginning of the month by many national oil companies and is called the ofcial selling price (OSP), whereas if crude oil is purchased on the spot market the adjustment factor is as agreed between buyer and seller. The adjustment factor plays an important role in crude selection as the relative economics (grade to be procured) changes based on the adjustment factors declared by the national oil company for their various grades. To be competitive, the national oil companies generally maintain the differentials of their grades with other, similar grades. Some national oil companies declare outright price for their
www.eptq.com
l 6 b b / $ , 4 r o t c a f 2 t n e m 0 t s u j d a 2 − e c i r P −4
Iran Light, API 33.1, Sulphur 1.5% Arab Light, API 32.7, Sulphur 1.8%
1 4 1 4 1 4 1 4 1 4 1 4 1 5 1 5 1 5 1 5 1 5 1 5 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 n n u p u p J a M a M a J S e N o J a M a M a J S e N o
Figure 3 Price adjustment factor over Oman Dubai average l b b / $ , r o t c a f l a i t n e r e f f i d e c i r P
Arab Medium, API 30.6, Sulphur 2.6% 2 1
Kuwait, API 30.5, Sulphur 2.5% Iran Heavy, API 30.2, Sulphur 1.8%
0
−1 −2 −3 −4 −5
1 4 1 4 1 4 1 4 1 4 1 4 1 5 1 5 1 5 1 5 1 5 1 5 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 n n u p u p J a M a M a J S e N o J a M a M a J S e N o Figure 4 Price differential factor over Oman Dubai average
grades after the completion of the month basis price is discovered under spot market conditions. Figures 3 and 4 show how the price adjustment factors of grades
l 6 b b / $ , 4 r o t c a f 2 t n e m 0 t s u j d a −2 e c i r P −4
similar to each other but from different suppliers closely track each other when the price adjustment factor is declared before the beginning of the month. However, the
Arab Light, API 32.7, Sulphur 1.8% Upper Zakum, API 33.0, Sulphur 1.8%
1 4 0 1 4 0 1 4 0 1 4 0 1 4 0 1 4 0 1 5 0 1 5 0 1 5 0 1 5 0 1 5 0 1 5 0 2 2 2 l 2 2 v 2 2 r 2 y 2 l 2 2 v 2 n r y u n p u p J a M a M a J S e N o J a M a M a J S e N o
Figure 5 Price adjustment factor over Oman Dubai average
PTQ Q1 2017 35
l 6 b b / $ , 4 r o t c a f 2 t n e m 0 t s u j d a −2 e c i r P −4
Murban, API 39.6, Sulphur 0.7% Arab Extra Light, API 38.4, Sulphur 1.1%
1 4 1 4 1 4 1 4 1 4 1 4 1 5 1 5 1 5 1 5 1 5 1 5 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 n n u p u p J a M a M a J S e N o J a M a M a J S e N o
Figure 6 Price adjustment factors over Oman Dubai average
20 Gasoline Dubai l b b / $ , l a i t n e r e f f i d e c i r P
18
Gasoil Dubai
16 14 12 10 8 6 4
1 4 1 4 1 4 1 4 1 4 1 4 1 5 1 5 1 5 1 5 1 5 1 5 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 2 0 r 2 0 y 2 0 l 2 0 2 0 v 2 0 n n u p u p J a M a M a J S e N o J a M a M a J S e N o Figure 7 Crude and product price differentials
Crude blend properties Arab Extra Light A 39.1 1.1 71
Soroosh B 18.7 3.6 29
Blend A+B 32.8 1.7
Arab Light
Blend 1 API Sulphur % wt in blend
Arab Super Light A 50.6 0.04 10
Khafji B 28.5 2.8 90
Blend A+B 30.5 2.5
Kuwait
Blend 2 API Sulphur % wt in blend
Blend 3 API Sulphur % wt in blend
Arab Super Light A 50.6 0.04 30
Maya B 21.8 3.3 38
Table 3
grades whose prices are declared after the end of the month (outright price) do not see any relationship
36 PTQ Q1 2017
32.7 1.8
30.5 2.5
Thus, the economics of a grade whose price is declared after the end of the month vis à vis a similar, alternative grade whose price is declared before the beginning of the month keep changing from month to month as the differential price between two grades varies signicantly. Crude oil procurement strategy The relative attractiveness of crude oil for renery margin maximisation takes into account the differentials in crude price, freight, and product worth. If one scans the data over a period of years, there is no simple relationship on all three parameters and thus decisions on term contracting are required to be taken basis various scenarios on price variations and historical performance. In addition, if product demand and specications are seasonal, the crude suitable for one season may not be suitable for another season. Gas oil rich crudes will be most suitable in a high gas oil demand/price season whereas gasoline rich crudes will be suitable when gasoline demand/prices are high. Figure 7 shows monthly variation in gasoil and gasoline prices in the Arabian Gulf market and indicates the suitability of gasoline rich crudes since March 2015 over gas oil rich crude oil grades. The procurement strategy thus requires detailed insight into market variations, supply security considerations and attractiveness of grades traded on the spot market. The most important decision is how much to procure on term contracts, which are generally for a period of one year, to address concerns over supply security if the grades contracted are not optimum for the entire year.
Widening of the crude oil basket Arab Heavy Blend Arab Medium The quality of crude oil varies C A+B+C widely in terms of its properties: 27.8 30.65 30.6 API, sulphur, nitrogen, metals 2.7 2.27 2.6 and so on. The crude and vacuum 32 distillation unit in a renery is designed to operate within the limits of the hydraulics of streams sepwith similar grades whose prices arated from crude oil. However, are declared before the beginning secondary units may be limited of the month (see Figures 5 and 6). by the quality and quantity of
www.eptq.com
streams generated from crude oil. 3 The crude unit can handle crude oil Blend 3 with certain API variations. If the Blend 2 l b crude is very light (high API), the Blend 1 b / column overhead section as well $ , 2 as naphtha stabilisers will be lim n o i t iting. Alternatively, if the crude is c u heavy (low API), the bottom circuit d e and vacuum unit may become lim r 1 t iting. Certain crude oil properties s o like viscosity and metals content C can limit the preheat train/desalter 0 operation in a crude unit. Crude 4 4 4 4 4 4 5 5 5 5 5 5 oil with an acid number beyond an 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 l p v n r y u l p v acceptable limit cannot be accepted n r y u J a M a M a J S e N o J a M a M a J S e N o in the absence of crude oil blending facilities or a chemical programme Figure 8 Reduction in crude cost through blending to mitigate corrosion. The properties of product from processed crude can limit the sec4 ondary processing units as well as the final product specification. Sulphur recovery units may be lim l 3 ited for high sulphur crude, while b b / a hydrocracker may be limited due Blend 3 $ , to high nitrogen in vacuum gas oil, Blend 2 n 2 i and a diesel hydrotreater may be a Blend 1 g limited for processing high diesel t e yield crudes. Generally, linear pro N 1 gramming (LP) models are used for evaluation of crude oil grades. However, there are limitations in 0 a LP model as processing in the 1 4 0 1 4 0 1 4 0 1 4 0 1 4 0 1 4 0 1 5 0 1 5 0 1 5 0 1 5 0 1 5 0 1 5 0 2 2 2 l 2 2 2 2 2 2 l 2 2 2 model is for a group of crudes a n M a r a y J u S e p N o v J a n M a r a y J u S e p N o v J M M together and not for a single grade. Thus widening of the crude oil basket depends on the capability of Figure 9 Net gain through blended crude the refinery to process crude oil on a standalone basis in the absence of crude oil blending facilities. Recommendations Alternatively, if blending facilities Model representing of crude oil grades Input refinery units and are available, the widening of the under term and spot operation contracts basket is not a challenge as difficult to process crude oil grades can be Crude oil grades LP output for various Opportunity analysis blended with other crudes so as to for evaluation pricing periods Most desirable grades have a composite crude oil suitaProduct demand LP output for identifications ble for refinery configuration. The seasonal product Delivered-basis Availability checks demand variations challenge then shifts to schedulcrude oil prices under term and spot ing and blending. As the economic Product sales Quantity finalisation realisation under term and spot benefit of widening of the crude oil basket is expected to be high, it is essential for every refiner to Figure 10 Steps followed during crude oil evaluation explore options that can give flexThus widening of the crude oil crude oil basket with blending. The ibility to procure any crude oil grade that will maximise margin. basket and blending is essential to net realisation of processing blended Examples in Table 3 indicate how increase options and capture oppor- crude over neat crude basis fivegrades that cannot be processed tunities for maximisation of margins. cut (LPG, naphtha, kerosene, gasoil neat, if blended and processed can Figure 8 shows the price difference and fuel oil) is shown in Figure 9. reduce the crude cost over procur- between blended crude and neat The blended crude processing maring a single grade similar to the crude and indicates an opportunity gins are directionally higher than to reduce crude cost by widening the for neat crude of similar quality. blended crude oil grade.
www.eptq.com
PTQ Q1 2017 37
The salient features of procurement under term or spot Procurement under term
Procurement under spot
1 Available mainly from national oil companies or equity producers. Some of the traders can also supply subject to them having a contract with oil major/equity producers.
Available mainly from oil majors and traders.
2. The contract can be for a particular grade or for all the grades available from the national oil company
The contract is for a specific grade of crude oil.
3 The contract is finalised either through negotiation or through participation in a tender enquiry by the national oil company.
The contract is finalised either through negotiation or through participation by a supplier in a tender raised by the buyer.
4 Period of contract is generally for 12 months starting from April or from January.
The period of contract is specific to the cargo procured.
5 Price as per official selling price (OSP) as declared by a national oil company or at a differential to OSP as agreed between buyer and seller. Grades that do not have an OSP are priced as per the price declared by the seller according to actual trading during the month.
Price at premium/discount to marker crude or to OSP as agreed between buyer and seller.
6 Terms and conditions as per general terms and conditions of a national oil company or an oil major holding equity. Few specific conditions of the contract like payment terms and credit limit will be as agreed between buyer and seller.
Terms and conditions as per general terms and conditions of national oil company or oil major holding the equity. Few specific conditions of the contract like payment terms and credit limit will be as agreed between buyer and seller.
7 The grade and quantity is finalised on a barrels per day basis to be supplied during the year. Nominations for loading month and cargo size are to be indicated well in advance.
The grade, quantity and loading month/date range for loading is finalised at the time of contract.
8 The supply is mostly assured.
The most desired grade may not be available if all the cargoes are tied up.
9 The grade offered by the supplier may not be suitable for a particular pricing scenario and thus becomes sub-optimum.
The contract is finalised only if the grade appears suitable.
10 There are possibilities to surrender the cargo back to the supplier with adequate notice, in the event of any issues with refinery operations.
The cargo contracted is generally required to be lifted unless agreed by the seller. However, the option of resale is available.
Table 4
Grades available only from national oil companies and only through term contracts Country
Iran
Kuwait Neutral Zone Saudi Arabia
Crude oil grade
Forozan Iranian Heavy Iranian Light Lavan Nowruz Sirri Soroush Kuwait Hout Khafji Arab Extra Light Arab Heavy Arab Light Arab Medium Arab Super Light
Key properties
API 31.7 30.2 33.5 35.2 20.5 33.2 18.7 30.5 32.8 28.5 39.1 27.8 32.7 30.6 49.1
S% 2.1 1.8 1.5 1.7 3.5 2.0 3.6 2.5 2.0 2.8 1.1 2.7 1.8 2.6 0.1
Table 5
The crude oil grades considered for in detail by means of LP model runs blending need to be evaluated for before a decision on procurement is compatibility as well as evaluated taken.
38 PTQ Q1 2017
Formulation of crude oil purchase plans Sourcing of crude oil can be carried out under term or spot contracts with national oil companies, oil majors and traders. Sourcing on term conditions ensures security of supply, whereas sourcing on the spot market can capture opportunities for margin maximisation by sourcing the most suitable grade under the prevalent monthly market conditions. Most reners use a LP model (a true reection of the renery conguration) for evaluation of crude oil. The product sales plan, sales net realisation, renery processing units’ capacities/constraints, processing cost (fuel and power consumption, chemical and catalyst consumption), crude oil availability, and crude oil landed cost are input to the LP model for evaluating the various options. In developing a strategy for crude oil procurement, the crude oil basket to be considered for evaluation plays a very important role. Once evaluation under different pricing scenarios has yielded appropriate crude oil grades, a strategy for sourcing under term and spot conditions needs to be developed. Figure 10 shows the details of various steps followed during crude oil evaluation. With various options generated, a decision must be taken on how much crude to term up and how much to keep for procurement under spot. The sailent features of procurement under term or spot are shown in Table 4. If one reviews the features of securing crude oil under term or spot contract, the premium/discount applicable and date range for loading a cargo secured under term conditions is not known at the time of conrming the contract, whereas for crude oil procured under spot conditions there is an agreement on premium/discount to the marker crude and also on a date range for loading. Under term contract, the supply of agreed quantity is assured. As regards grade of crude oil to be procured, there is no issue if the contract is for a single grade; however, if the contract is for multiple grades the allotment of
www.eptq.com
desired grades is subject to availa bility. The date of loading is subject to acceptance by the seller. Some national oil companies market their crude oil grades only through term contracts and also with end user restriction. However, the equity crude oil of some national oil companies is available under spot as well as under term conditions without any end user restriction. Table 5 details major crude oil grades available only through term contracts, whereas Table 6 details availability of major crude oil grades actively traded on the spot market from countries of interest to India. As can be seen from Table 6 , there are many alternatives for grades that are supplied only under term contracts. If security of supply is not a concern for a rener, there is very little reason why one should seek procurement under term contract unless the grades available under term conditions are robust under all pricing scenarios. Conclusion The crude oil market is complex and procurement of the right crude can maximise renery margins. The procurement of a crude oil grade under term conditions depends on its attractiveness under various pricing scenarios and any need to address concerns over security of supply. The opportunity exists to source crude oil through spot market and maximise margins, as terming up of grades for a longer period may not be attractive under all pricing scenarios and may not be required if there are no issues on supply security. Widening of the crude oil basket through state of the art blending facilities can increase the options for crude sourcing for a renery and will also contribute to margin maximisation.
Misha Gangadharan is a Senior Engineer with
Hindustan Petroleum Corporation Limited India (HPCL) with experience in crude oil evaluations, production planning and optimisation. She holds a BTech degree in chemical engineering from Institute of Chemical Technology (ICT) Mumbai, India. Email:
[email protected]
www.eptq.com
Grades actively traded on the spot market 1 Country
Dubai Iraq Neutral Zone Oman Qatar UAE
Yemen Egypt Nigeria Angola Algeria Malaysia
Indonesia Other WAF countries: Cameroon, Chad, Congo, Gabon, E. Guinea, Ghana, Sudan Mexico Venezuela Brazil Colombia Equador Azerbaijan Libya Brunei Russia
Kazakhstan North Sea
Crude oil grade
Dubai Basrah Basrah Heavy Kirkuk Eocene Ratawi Oman Al Shaheen Qatar Land Qatar Marine Das Blend Lower Zakum Murban Upper Zakum Marib Light Masila Ras Gharib Belyam Suez Mix Various Grades Various Grades Saharan Blend Miri Tapis Labaun Kikhe Various grades Various grades like Kole, Doba, Djeno, Nkossa, Rabi Lt, Ceiba, Zafiro, Jubilee, Dar, Nile Blend Isthmus Maya Olmeca Various grades Various grades Cano Limon Castilla Cusiana Napo Oriente Azeri Light, Various grades SLEB Champion ESPO Sokol Urals, Vityaz CPC Blend Kumkol Various grades
Key properties API S%
31.4 28.8 23.7 34.2 18.3 24.2 30.4 30 40.7 32.7 38.3 39.9 39.6 33.0 45.1 34.6 20.9 23.6 30.8 29-47 22-39 45.7 29.7 42.1 30.2 36.7 20-48
1.5 3.1 4.1 2.2 4.6 4.1 1.4 2.4 1.2 1.8 1.1 1.0 0.8 1.8 0.1 0.5 3.8 2.7 1.5 0.05-0.3 0.1-0.8 0.1 0.1 0.04 0.1 0.07 0.02-0.2
21-40 33.2 21.8 38.8 11 -39 17-29 30.2 18.2 43.2 18.6 24.7 36 37-42 38.5 33.1 36 36.1 32.7 34.4 45.4 41.4 30-47
0.1–0.4 1.2 3.3 0.9 0.1-5.2 0.3-0.8 0.5 1.5 0.1 2.0 1.4 0.1 0.1-0.4 0.1 0.1 0.6 0.3 1.0 0.2 0.6 0.1 0.2-0.8
Note 1: There are many more grades over and above those mentioned above
Table 6 S D Pohanekar is a Senior Engineer with
M D Pawde is Head of Economic Planning
Hindustan Petroleum Corporation Limited (HPCL) India with experience in refinery production planning and optimisation. She holds a BTech degree in chemical engineering from Laxminarayan Institute of Technology, Nagpur, India.
and Optimisation with Hindustan Petroleum Corporation Limited (HPCL) India with over 30 years of experience covering refinery operations, crude oil evaluation and procurement, production planning and optimisation. He holds a BTech degree in chemical engineering from Nagpur University, India. Email:
[email protected].
Email:
[email protected]
PTQ Q1 2017 39
Cuba’s oil: due for development Cuba’s ambitions for energy self-sufficiency require major investment in the nation’s oil reserves and refining industry AMAURY PÉREZ SÁNCHEZ University of Camagüey
C
uba’s internal power demand is expected to rise in the coming years in response to growth in private business, expansion of foreign business investment in the country, and progress and expansion in important power consuming industries such as biotechnology, tourism and construction. Today, Cuba generates about 95% of its internal electricity from hydrocarbons and their byproducts. Unión Cuba Petróleo (CUPET) is Cuba’s largest oil company. It is owned and operated by the Cuban government and is involved in the extraction of petroleum deposits,
refining and distribution of petroleum products (see Figure 1). For the period to 2030, CUPET has conceived several ambitious plans and projects in order to enhance
its profitability and productivity, as well as to develop its oil refining industry. Among these can be mentioned: • Expand the refining capacity of
the oil refinery located in Cienfuegos province to 150 000 b/d • Modernise and update equipment and accessories operating in
the refineries, mostly combustion systems, boilers, vessels, tanks, heat exchangers, wastewater treatment structures, control valves and automation systems • Produce dielectric oils for 33kV electrical transformers at Sergio
Soto refinery • Install sweetening units for Jet A-1
(Merox) fuel in Ñico López refinery • Introduce liquefied petroleum gas (LPG) into the national power mix • Elevate the quality of national fuels to international standards • Increase storage capacity nation-
www.eptq.com
F Facilities F
G GM G F
F M F
M Maritime terminals R Refineries
W F F F M
R
R
W Oil wells
M
F W
W
W
G Gas plants F M F F
M
F W W W W W W W WW W W W W W W WW
F
F F
F
F M F
F RF
F
Figure 1 CUPET infrastructure and facilities
wide, both for crude oil and petroleum products • Reduce/optimise the costs of logistics operations • Intensify the exploitation of oil deposits located inland and offshore • Expand the production level of existing oil deposits by means of enhanced oil recovery (EOR) technologies.
operations, the supply of modern, oil related technologies, equipment and resources, as well as provision
of financial, engineering and management services.
The first business contract related to oil extraction operations between the Cuban government and a foreign
firm was signed in 1990, in order to
exploit the oil wells located at Block III south of Varadero beach. In the Private investment last three years, around 42 shared To achieve sustainability in energy production agreements (SPA) have in the near future, the Cuban gov- been agreed between CUPET and ernment has incorporated new per- foreign firms, in accordance with spectives and opportunities into the first Cuban Foreign Investment its oil industry offered by recently Law, approved in 1995, and 2014 approved foreign investment poli- legislation whose terms ease busicies, in order to increase efficiency, ness transactions and increase the reduce costs and boost production scope of the proposed investment capacity. Accordingly, the main objectives. objectives of joint ventures and In Cuba, the National Office of associations with foreign firms and Mineral Resources (NOMR) is the companies are threefold: search- government institution in charge of ing for unconventional oil; offshore certifying and approving potential exploration; and EOR. There are investment projects and investors other collateral activities requiring in the oil industry. Each business direct capital investment such as contract signed is protected under technical services for oil extraction Cuban government decree and
PTQ Q1 2017 41
Foreign investment opportunities in the Cuban oil industry Option 1 Title:
Oil exploration risk contracts and shared production agreements in production blocks located in Cuban shoal waters (eight blocks). Description: Dene the potential oil and gas reserves existing in determined areas of Cuban shoal waters. If positive results are obtained, exploit the energetic resources with protability. Location: The eight blocks available are located predominantly at the north of Pinar del Río, Matanzas, Villa Clara and Sancti Spiritus provinces, as well as in the south of the following prov-
inces: Pinar del Río, Artemisa, Mayabeque, Matanzas, Ciego de Ávila, Camagüey and Granma. Market: Cuban internal market in rst term, surplus production for exportation. Results: For an exploration block presenting a 30-year exploitation
Economic Zone (EEZ) in the Gulf of Mexico (52 available blocks). Description: Dene potential oil and gas reserves in the Cuban EEZ. If positive results are obtained, exploit the resources protably. Location: The Cuban EEZ comprises a 112 000 km2 zone in deep waters in the Gulf of Mexico, just to the north of the Pinar del Río, Artemisa, Mayabeque and Matanzas provinces, where there are 52 accessible blocks for oil exploration and further exploitation (see Figure 3). Market: Cuban internal market in the rst term, with surplus production for export. Results: For an exploration block with a 30-year exploitation contract, considering an average oil cost of
$128.2/bbl and applying 10% taxes: NPV: $1241.2 million; IRR: 18.5%; PP: 7.5 years. Option 3
Oil exploration risk contracts
interest to the Cuban government are not considered. Market: Cuba’s internal market in rst term, with surplus production for export. Results: For a block with a 25-year exploitation contract, considering an average oil cost of $122.6/bbl and applying 10% taxes: NPV: $191.7 million; IRR: 44%; PP: 3.4 years. Option 4 Title: Enhanced
recovery of exist-
ing oil deposits. Description: Increase the recov-
ery coefcient of the oil reserves contained in certain existing oil deposits, which cannot be recovered by conventional extraction methods. Location: The oil deposits opened
for negotiation under this option
are Santa Cruz del Norte, in Mayabeque province, and East
contract, considering an average
Title:
oil barrel cost of USD $122.8 and applying 12% taxes, a net present value (NPV) of $383.5 million, an internal rate of return (IRR) of 39.5% and a payback period (PP) of 2.8 years are estimated.
province. Other oil deposits will be in production blocks located in considered for future investments. Cuban national territory (25 blocks). Market: Cuban internal market in Description: Dene potential oil the rst term, with surplus proand gas reserves in the Cuban duction for export. national territory. If positive results Results: For a block with a 30-year are obtained, exploit the resources exploitation contract, using an protably. average oil cost of $95.1/bbl and Location: There are 25 available applying 12% taxes: blocks dispersed throughout Cuban NPV: $142.4 million; national territory. Urban zones, IRR: 46.3%; protected zones, or areas of special PP: 2.1 years.
Option 2 Title: Oil
exploration risk con-
tracts and shared production
agreements in production blocks located in the Cuban Exclusive
Varadero oil deposit in Matanzas
and shared production agreements
Table 1
crude oil exists are well identied from 25 to 35 years. Taxes are not and characterised. paid for the rst eight years, be they The economic value of Cuban municipal, regional, or those on the crude oil in the international market repatriation of earnings or prod- is about 60% of the reference crudes ucts. Bonuses are not paid upon WTI and Brent. It is estimated that signing, and taxes are levied by the the production cost of a barrel of oil National Tax Administration only from offshore platforms in Cuba is on net annual earnings. Nowadays, about $20-35, while the production could have a legal validity period
foreign companies interested in
investing in the Cuban oil industry prefer to participate mostly in EOR operations since the exploratory risk is smaller and the zones where the
42 PTQ Q1 2017
If oil is found, it is estimated that companies would have to invest in developing production capacity for
at least three to ve years before production could begin. However, production could be delayed due to, mostly, availability of offshore oil eld development services. Once oil production begins, it is expected to cost of oil produced from inland oil grow slowly. The main non-Cuban operator elds is $13-15/bbl. Table 1 outlines foreign invest- in Cuba is Sherritt International, a ment opportunities in the Cuban oil Toronto mining company which has industry. been active in Cuba for more than
www.eptq.com
20 years. This firm operates Puerto Escondido, Yumuri, and Varadero West oil fields under two production-sharing contracts (PSC). In May 2014 the company negotiated a 10-year extension to the Puerto Escondido-Yumuri PSC. It has drilled eight wells, one more than required by the extension terms, and has ended the extension drilling programme. Six of the wells produce oil, one is suspended, and one has been abandoned. Some studies have concluded that Cuba could produce enough oil in the future to become an oil exporter, but there are some uncertainties that still need to be considered first in order to support this conclusion. First, there are tangible reservations regarding when oil production will start and at what rate it could be obtained. Secondly, Cuba will need to offset the roughly 130 000 b/d of oil it currently imports to meet existing demand prior to becoming a net oil exporter. Thirdly, oil demand is expected to grow in the near future, taking into account growth in sectors such as construction, tourism and industry. In this case, Cuba is still likely to trade more oil – especially as refining capacity increases – but its net trade balance in oil may not necessarily shift to a significant oil export surplus. All will depend only on how much oil is found and developed, and what will happen with domestic Cuban demand. What is more certain to take place is that an increment in oil production may reduce Cuba’s dependence on oil imports from Venezuela.
o a a n i a r a a r a u n c M t ~ i n i N a S P a B a L L o l e r a d a n a a m r n j i u a e o c a b t C B a s H E
2
o a n j t e l e s L e a J r i r o e B F o N l e i 1 a r F 4
o c − r u a a c J n e l a d B a c i a o V B
i o u r o m u c n r u a j Y e b o t e L S e e s O
3
z r u C t a n a S
10km
r t o Y u mu r i e S eb or u u c o P o − id i d s a n n co a s C E
Block N37 3D seismic (Habana−Matanzas) Leads Leads too far from the coast (Block N37) Prospects ERW
r o d e r a a V
1 B J Norte (Breccia) 2 B J Norte (Veloz) 3 Santa Cruz Este 4 Santa Cruz Norte 6km horizontal displacement
Figure 2 Mapped leads and prospects in Cuba’s Northern Oil Belt
was applied the oil production rate increased markedly. In the last 14 years, more than 245 million of oil barrels have been extracted from the so-called Heavy Crude North Fringe (HCNF) or Northern Oil Belt (NOB), an oilrich area between Havana city and Matanzas provinces (see Figure 2). This area accounts for about 97% of Cuban oil production (an oil deposit located in that zone and qualified as ‘productive’ could generate about 2000 b/d), while there are other
small production sites located in Ciego de Ávila and Sancti Spiritus provinces, some of them with exploitation periods of more than 60 years. The most important oil deposit located in the HCNF, and in Cuba, is that of Varadero, which has 90 fields under exploitation and had produced about 185 million barrels of oil at the end of 2015. It is calculated that only about 6-7% of its potential oil reserves (estimated at 1.3 billion barrels) have been recov-
P A N G R E T GULF OF MEXICO S E A FLORIDA STRAIT
How oil is extracted in Cuba
Cuba has the second largest proven hydrocarbon reserves in the Caribbean area, surpassed only by those of Trinidad and Tobago. Since 1996, and taking into account that most oil deposits in Cuba are located offshore, CUPET changed its approach to extracting oil by introducing the horizontal perforation method (HPM), which has led to an increase in oil production levels since 2002. This change has had a substantial economic impact on the Cuban oil industry, considering that in most fields where HPM
www.eptq.com
S T R A I T O F Y U C A T A N
CARIBBEAN SEA
Figure 3 Oil exploration blocks located in the Cuban Exclusive Economic Zone, Gulf of Mexico
PTQ Q1 2017 43
Varadero Field is the largest in Cuba with over 11 billion barrels of oil in place (10.5-14.3º API)
Bolanos-1 (1991) recorded a recovery of 22º API oil from Shallow Sheet
Guadal-1 (1971) recovered >30 barrels light oil (24.5º API) on test from Shallow Sheet
Marti-5 (1984) flowed light oil (24.0º API) from Deep Sheet at unspecified rate Surface geology / structural trend Well − oil shows Well − strong oil shows Oil field Oil and gas field Lower sheet play − leads
Motembo Field (1881) has intermittent production of light oil (50-64.5º API) 0
km
50
Figure 4 Block 9 in Motembo region
ered to date because no secondary or enhanced extraction methods have been applied. At present, annual oil production capacity in Cuba is about 25 million barrels (about 50 000 b/d). This is used entirely for power generation and meets about half of national energy demand, while the rest (about 90 000 b/d) is imported from Venezuela under an exclusive payment agreement. Cuba accounts for only 0.05% of the world’s total production of crude petroleum. A new oil drilling campaign started at the end of 2016, continuing throughout 2017, in the Cuban Exclusive Economic Zone (EEZ) in the Gulf of Mexico (a 112 000 km2 zone divided into 59 blocks, 30 of them assigned to foreign firms under risk contract agreements, see Figure 3), taking into account tangible evidence that there is about 15 000 million barrels of extractable oil in this area. The first exploration
Ñico López (Havana) R
operations will be carried out by the Venezuelan company Petróleos de Venezuela (PdVSA) and the Angolan firm Sonangol. Up to 2020, there are concrete
At present, annual oil production capacity in Cuba is about 25 million barrels (about 50 000 b/d) plans to drill exploratory wells in several land sites located between La Habana and Santa Cruz del Norte (mostly in the towns of Boca de Jaruco, Tarará, Santa María and Santa Cruz). Since the end of 2015, oil deposits in Boca de Jaruco are being treated with steam injection technology in
Sergio Soto (Sancti Spiritus) R R
Camilo Cienfuegos (Cienfuegos) R Refinery Pipeline
Figure 5 Oil refineries in Cuba
44 PTQ Q1 2017
R Hermanos Diaz (Santiago de Cuba)
order to reduce the viscosity of the extra heavy crude, in a joint venture project signed between CUPET and the Russian state owned firm Zarubezhneft. Cuban oil is basically a sour, heavy crude (8-16 °API) and can be used only for power generation and to produce cement, lubricants and asphalts, although there are some oil deposits that supply substantial amounts of light and medium oils from time to time.
Does oil exist in Motembo? During July 2016 international media publicised a report published by the Australian firm MEO on the hypothetical discovery of high quality, light crude oil reserves of about 8200 million barrels in a zone known as Motembo, a 2380 km2 area located in the central province of Villa Clara, which is the first region in Cuba where oil was obtained. Logically, this notice generated great expectations both in Cuba and the rest of the world. However, some CUPET authorities declared in a press conference some weeks later that there had been erroneous misrepresentation of the note published by the Australian company, which did not use the terms “confirmed discovery” or “finding”, but only the “identification of potential, important amounts of oil which can be recovered under the application of future development projects”. According to announcements by MEO, confirmation and further validation that there are thousands of millions of barrels of oil at this site requires supplementary evaluation, exploration and analysis operations distributed in different phases: 2D seismic operations, profitability and prospective studies, deep drilling and oil quality testing. MEO signed a shared production agreement with CUPET in September 2015, and is carrying out oil exploration studies at Block 9 in Motembo region (see Figure 4), including review and reprocessing of the existing 2D seismic data; geochemical samples acquisition and evaluation; and completing new 2D seismic operations in an additional 200 km zone.
www.eptq.com
Fracking Fracking is one of the most contro versial oil extraction methods used today. This method has never been used for oil extraction in Cuba since there are more efcient and prot able techniques available to extract oil such as, for example, steam or chemicals injection. According to recent geological studies made by CUPET, the oil res ervoirs located in Cuba are in car bonate layers, so the best available option to extract oil on those sites is by means of acid or steam injection. In that sense, fracking is not consid ered today a viable method for oil extraction in Cuba. Figure 6 Camilo Cienfuegos refinery in Cienfuegos province
Oil refineries in Cuba There are four reneries currently under operation in Cuba, all owned and operated by CUPET (see Figure 5). They are: • Ñico López, in Havana province • Camilo Cienfuegos, in Cienfuegos province (see Figure 6) • Sergio Soto, in Sancti Spiritus province • Hermanos Díaz, in Santiago de Cuba province. Together, these reneries have a total nominal crude oil distilla tion capacity of about 300 000 b/d, although only 45% of this capacity is currently exploited (135 000 b/d). Table 2 shows the rening tech nologies applied by each Cuban renery as well as its rening capacity. The largest of the four reneries, Ñico López, accounts for nearly 40% of total rening capacity (36 400 b/d). It operates a 12 500 b/d catalytic cracker, and is the only facility with a catalytic conversion unit in Cuba. There is a ve-year plan to modernise and increase its storage capacity (mostly crude oil recep tion tanks and petroleum products storage vessels), as well as its auto mation and steam generation/dis tribution systems. Similarly, there are plans to install a Jet A-1 (Merox) fuel sweetening plant. The Camilo Cienfuegos renery was reopened in 2007 as the result of an agreement between CUPET and PdVSA, and is now operated by Cuvenpetrol SA, a Cuban/ Venezuelan joint venture. This
www.eptq.com
renery runs only on Venezuelan crude, and has a nominal ren ing capacity of about 65 000 b/d. More than 90% of the renery’s output, including gasoline, diesel fuel and fuel oils, is consumed by the Cuban domestic market. Several techno-economic design projects are under way such as expansion of its rening capacity to about 150 000 b/d; constructing a plant for olen and aromatics; increasing its storage capacity from 8000 bbl to 139 000 bbl; building a liqueed natural gas plant; and reactivating the pipeline between Matanzas and Cienfuegos, at a total preliminary cost of about $6 billion. Several tests will be car ried out to determine whether the renery can process up to 80 000 b/d. From July 2016, the renery reduced its actual processing capac ity to 50 000 b/d due mostly to reduced oil imports from Venezuela and the implementation of exhaus tive, complex maintenance opera tions in main equipment and related
accessories. In 2015, the facility
rened about 17.8 million bbl, while it is projected that it will rene only around 9.43 million bbl in 2016, 53% of its planned output. Recently, this renery has been processing and blending crude and petroleum products for sale to neighbouring countries. Hermanos Díaz renery has a nominal annual rening capacity of about 1 500 000 cu m of crude oil (30 000 b/d), although it is cur rently processing 22 000 b/d due to reduced oil imports. Several oil products are produced in this ren ery such as liqueed petroleum gas (LPG), naphtha, gasolines, vacuum gas oil, fuel oil, diesel and asphal tic cement. It supplies the market of Cuba’s eastern region and provides raw material for the catalytic crack ing unit in Ñico López renery. Several actions have been accom plished throughout this year in order to modernise some equipment and accessories, such as the total
Refining technologies and capacities applied by Cuban oil refineries Refining technology
Ñico López (Medium conversion)
Refining capacity (tobd) Atmospheric distillation Vacuum distillation Catalytic cracking Reformer Distillate hydro-refining
36.4 18.0 12.5 2.7 3.2
Refinery Camilo Cienfuegos Sergio Soto Hermanos Díaz (Hydro-skimming) (Hydro-skimming) (Hydro-skimming )
65.0 9.6 25.0
2.8 1.4 -
30.0 18.0 2.7 7.0
Table 2
PTQ Q1 2017 45
Cuban production, exports, and imports of refined petroleum products, 2010-13 (‘000 barrels) Product production
2010
2011
2012
2013
Naphtha LPG Gasoline Diesel fuel Jet kerosene Fuel oils Total
330 220 4163 8972 2412 17 856 33 953
601 183 3313 8943 2272 17 020 32 333
586 176 3225 8349 2214 16 720 21 270
513 147 2932 8063 2199 14 660 28 514
5336 528 5622 11 486
0 528 5417 5923
0 506 5087 5087
0 0 5116 5116
396 2,829 498 4193 4325 12 857
462 0 506 4200 3255 9031
469 0 147 4537 2404 8188
367 0 132 4552 2456 8114
Exports Gasoline Diesel fuel Jet kerosene Total
Imports LPG Gasoline Diesel fuel Jet kerosene Fuel oils Total
Table 3
insulation of ovens and kilns, implementation of automation, more efcient combustion systems, and incremental storage capacity. Finally, Sergio Soto renery operates atmospheric distillation and vacuum distillation columns, and produces primarily asphaltic liquids (AC-30 and RC-0 types), as well as minor amounts of naphtha, diesel, fuel oil, solvents, basic dielectric oils, and chemicals to produce pesticides. It is the only Cuban renery dedicated exclusively to processing and rening crude oil produced in Cuba (about 2800 b/d). In 2015, the amount of asphaltic liquid produced by this renery was 15 400 t, the highest output in its history. The vacuum distillation tower was recently revamped in order to increase production efciency from 42% to about 58%. The production of dielectric oils for 33-kV electric transformers, and the introduction of more efcient, high capacity boilers and cooling water systems is the most important investment task for the facility in the near future. Cuban reneries are able to meet current domestic demand for gasoline and diesel fuels, but the nation still needs to import other rened petroleum products such as kerosene jet fuels and fuel oils, mainly from Algeria (which accounts for 80-85% of total imports of these
46 PTQ Q1 2017
products), Venezuela (about 8-10%), the European Union (6%), Mexico (2%) and Russia (2%). Such imports account for about 25-30% of total domestic demand for rened petroleum products (see Table 3). From July 2016, PdVSA reduced its crude oil exports to Cuba by 19.5% (about 83 130 b/d), alleging
The present day represents a ‘now or never’ time for the Cuban oil industry regarding growth and economic feasibility declining crude prices in the international market. Cuba, in turn, has replaced the shortfall with similar volumes of crude and petroleum products from the Caribbean transshipment terminals of Bonaire, St. Eustatius, Aruba and Bullen Bay, and Willemstad, Curaçao. Cuban reneries process small amounts of Cuban oil. Most of the crude oil obtained along the country’s northern coast is diluted with naphtha and sent directly to power plants for combustion.
Conclusion Taking into account that about 95% of Cuba’s potential oil reserves are located either inland or in shallow waters, they remain unextracted because drilling operations require the use of advanced recovery technology that Cuba does not possess. The introduction of foreign capital technologies in this eld is a high priority for the country in order to achieve its intended self-reliance on energy. The proven existence of relevant oil reserves throughout the country, the application of attractive tax related policies to foreign rms interested in oil related operations in Cuba, the strong support received by the Cuban government and the socio-political stability of the country, the availability of skilled and experienced human resources, as well as the reputation of CUPET regarding business procedures and contracts are the most signicant advantages offered by the country to develop its oil industry by means of direct foreign investment. The Cuban energy sector needs to update and modernise its technological systems and infrastructure to modern standards in order to increase its protability and economic viability with respect to the environment, to meet future increases in internal power demand, and to compete with the renewa ble sector. Thus, several strategic programmes and projects are being designed or evaluated to cope with expectations and also to develop the country to higher standards. In this sense, the present day represents a ‘now or never’ time for the Cuban oil industry regarding growth and economic feasibility.
Amaury Pérez Sanchez is Adjunct Professor at the Department of Chemistry, University of Camagüey and a Research Engineer with CUPET company. With over 10 years’ experience in chemical/petrochemical processes, mostly in mass/heat transfer operations, he holds a master’s degree in chemical process analysis from the University of Camagüey, Cuba. Email:
[email protected]
www.eptq.com
Safety Measures Hard Hat Goggles High-Visibility Vest
Remote Mount Capability Keeps Workers Off Top of Vessel for Switch Modification Advanced Self Diagnostics Assures Reliable Performance Best-in-Class Safe Failure Fraction >91%
Insulated Gloves
Steel-Toed Boots
Dual-Point Option for Two-Alarm Safety Protocol
Safety Harness
Protect your plant with Echotel Ultrasonic Level Switches
®
ECHOTEL liquid level control technology measures up to the most rigorous safety standards, with intelligent design that ensures outstanding quality and reliability. echotel.magnetrol.com
magnetrol.com • 800-624-8765 •
[email protected]
© 2016 Magnetrol International, Incorporated
Control system security A major cyber attack on a refinery is a clear and present danger. What lessons can be drawn from other sectors? SINCLAIR KOELEMIJ Honeywell Process Solutions
S
ince the 2010 attack on Iranian bearings, the exact structure of the uranium enrichment facilities, physical installation, and the exact the cyber security of industrial code running in the control equipcontrol systems has attracted signif- ment and feedback from the icant attention and investment as sensors. companies have sought to improve All this meant the attack required their protection against cyber a substantial effort. For many, it threats. Rening companies and therefore seemed such an attack other industrial businesses have was unlikely to occur against revisited their approach to cyber targets where no signicant politisecurity, which was until then cal motivation existed, since only focused on protecting systems nation states would have the against malware and hackers. An resources, focus and motivation to attack on the physical installation undertake it. had been inconceivable. Damaging a physical plant instalEven so, the 2010 attack required lation or removing it from service considerable resources and skills to does not always require such a succeed. 1 Several factors contrib- complex attack, however. There is a uted to its complexity: variety of ways to create damage • It required stealth. Focused on through thermal, mechanical or damaging the centrifuge equip- hydraulic stress in a plant or to ment, it needed to be done quietly cause physical systems to wear to prevent early detection. To hide more quickly than anticipated. The what was happening from the operation window of the equipprocess operators, the feedback ment is in large part determined by from the control system also the data in the industrial control required manipulation. system. A determined attacker • It was targeted, aimed at one successfully accessing the control particular installation. It seems system will therefore not have unlikely it was intended to infect much difculty causing a plant other systems, since when it even- shutdown. tually did so resulted in the attack With the recent attacks on the being discovered and stopped. steel industry (Germany, 2014), the • It bridged an air gapped/isolated power industry (Ukraine, 2015), system – the control system for the and water treatment (US, 2016), an installation. A method had to be expanding range of business leaddeveloped to infect mobile ers are becoming aware of the risks. computer equipment that was ulti- And while the rening community mately connected to the control has thus far been spared a major system network for engineering/ attack, its leaders are looking across maintenance purposes. industry for best practice security • It required knowledge of the approaches. The net result of this physical system. The attacker heightened awareness is that chief needed to know the variations information security ofcers and in speed that would cause exces- chief security ofcers are slowly sive wear of the centrifuge expanding their focus beyond
www.eptq.com
corporate IT systems to ask critical questions of the automation systems, too. This article explores the differences between information security and control system security, and it uses learnings drawn from the manufacturing and processing industries to inform rening decision makers. While information security methodologies aim to secure the integrity and availability of the information in the control system, they do not analyse the consequences for the underlying physical system. This limits the ability of information security to predict and prevent attacks on the physical system. The main focus of control system security, by contrast, is on identifying and analysing the different types of failure scenarios and designing security to withstand attacks that could cause these. Control system security develops scenarios for such smart attacks and identies counter-measures against them. The article provides some examples and discusses ways to protect the system. Background Industrial control systems typically have three tiers: • The physical system (production system) • The production management system, including the distributed control system (DCS), supervisory control and data acquisition (SCADA) system and safety instrumented system (SIS) • The operations management system (optimisation, quality management,
PTQ Q1 2017 49
Business Management System (L4, L5) (information management) Information System Security
Operation Management System (L3) (information management) Industrial Control System
Production Management System (L2, L1) (automation)
Control System Security
Sensors and Actuators (L0) Crude
Refinery Installation (production processes)
Petrol
Figure 1 Industrial control system
environmental management, plan- tanks and pipelines), including the ning, and maintenance). sensors and actuators. Operations management systems Figure 1 shows the high level focus on the efficiency, effective- structure, including the levels ness and quality of the assigned by the ISA 95 standard.2 manufacturing process. Production The focus of this article is on levels management systems focus on the 0, 1 and 2, but some applications at manufacturing process itself – the level 3 can also impact the physical automation of the production system, so it is not always possible process. Finally, the physical to distinguish a clear boundary system (production system) is between control system security and composed of the various process information system security. units in the plant (such as the distil- Depending on the application, there lation column, boilers, furnaces, may be an overlap between the two.
e g d e l w o n K
Covert attack Disclosure
Zero dynamics attack
Eavesdropping DoS
Bias injection attack
n io t up r s Di
Figure 2 C yber physical attack space model
50 PTQ Q1 2017
Replay attack
The security of the systems at the business management layer differs considerably from the control system. A frequently mentioned example is the difficulty in updating the control system or installing security patches for it. Stopping production for this is difficult to justify, and changes in general are seen as a risk to the business continuity.3 This is why so many legacy systems are still active despite no longer having security patches available for them, with the result that they contain well-known vulnerabilities that can be exploited by attackers. Another difference is that an automation system such as a DCS requires a stricter operational environment than the average computer system. Real-time availability ensures every automation task in the system has sufficient time to complete its task. Real-time environments differ from a time sharing environment where tasks are allocated a specific time slot to complete their workload, over-runs are not allowed, and ultimately tasks may not be completed. In a real-time environment, overloading the system can result in resource starvation. The environment is therefore more vulnerable to denial of service attacks than business management systems, which can cope with a very broad range of delays. Finally, risk is different in an industrial control system. Consider the attack model developed by André Teixeira, Henrik Sandberg, Daniel Pérez, and Karl H. Johansson.4 Unlike cyber security for the business management systems where cyber security is focused on protecting data, they noted that cyber attacks on control systems have the potential to change the physical production process in different ways depending on the specific attack scenario. The attack space for a control system can be described by their model (see Figure 2). The attack space for cyber physical systems has three dimensions: • System knowledge: attacking the cyber physical part of the system
www.eptq.com
requires an understanding of the Failure occurs quickly production process, the various Critical limit high Failure occurs with parameters, measured values, and sustained operations actuators that can be used in the Standard level high attack. • Disclosure resources: an attack Target range high Operating can also require knowledge of the window Stable Safe to Target Optimal real-time state information of the Reliable operate system, process values, limits and Target range low control parameters. • Disruption resources: an attack Standard level low requires the capability to disturb Failure occurs with sustained operations the system, perhaps by modifying Critical limit low an output setting, a control paramFailure occurs quickly eter, or disabling a system action. Complex attacks, like the attack Figure 3 Operating limits and boundaries on the Iranian nuclear facility, use all three dimensions, but a simpler ment – for example, to increase The middle zone (green), between attack such as replaying Modbus wear or corrosion of the system, the standard limits (high and low), traffic to ‘freeze’ a process value rupture a pipeline, damage furnace is the zone designated for achieving only needs two dimensions. An hosing or damage a distillation operational targets. Outside those attack like the HAVEX attack 5 used column. limits, operator intervention is only one dimension – a Trojan • Disturb the production process – generally required to return the horse for eavesdropping. stop cooling systems, activate a process to this zone. Risk for a cyber physical system bypass or reduce product quality, Depending on the process variais a function of the likelihood and for example. ble, some limit ranges may not impact of a scenario: risk = • Compromise compliance – such have an upper and lower bound(scenario, likelihood, and impact). as with health, safety, and environ- ary. For example, tube skin Scenario describes the system mental regulations temperatures generally have only under attack and what steps the To accomplish any of these objec- upper limits. attacker needs to take to achieve tives, the attacker has three If the attacker breaches the integthe goals. potential targets: rity of the operating window, the Likelihood is a function of the • Plant operability, which includes physical process moves into a range effort an attacker needs to invest to the operating window, the safety where failure will eventually occur execute the scenario. The more resil- function, and monitoring and diag- if it remains there or, if too far ient a system is against a specific nostic functions. An attack on the outside the operating window, may scenario, the higher the investment operating window, for example, even occur immediately. in effort required from the attacker may increase it to cause damage. During plant design, the integrity to achieve his goals, and the lower • The measured or calculated operating window (IOW) is defined the risk of him doing so. values, with the attacker deceiving and all configuration parameters are Impact is the consequence of the operators or an automated function set in such a way as to protect the attack – in principle the attacker’s by modifying the apparent real- physical system from being goal. Risk calculations within infor- time state of the system. damaged. As an additional safemation security primarily use • The control actions – initiating guard for the production scenarios to determine the business an unwanted action, blocking management system, the safety impact, irrespective of the likeli- execution of a required action or a instrumented system (SIS) is used to hood. The scenario is a dimension combination of the two. force an emergency shutdown when not used in information security, such a condition would occur and but essential in control system Attacking the operating window safety would be at risk. Not every security because of the direct The operating window is defined operating window is guarded this by parameters in the production way, however, so in general, even connection to the physical system. management system – for example, when the ICS includes a safety funcControl system security range values for flow, pressure, tion, modifying the operating For control system security the temperature, rotation speed, level, window can cause damage. central question is what an attacker signal characterisation parameters, can do to harm the physical system rate limiting values, travel time, Attacking the safety function (production system) or disturb the output limits, and so on. Safety functions are generally diffiFigure 3 illustrates various types cult to attack. They run proprietary production process. There are three potential objec- of operating limits that create code and a physical key switch tives for an ICS attack: boundaries for any specific operat- usually controls downloads of the • To damage the process equip- ing window. code to the safety controllers. It is a
www.eptq.com
PTQ Q1 2017 51
well-protected environment and not easily breached. However, there are other tactics that can have an impact on the safety function, such as attacking the safety transmitters. Transmitters are frequently centrally managed. If the management system is attacked it can be used to modify the sensor conguration and indirectly inuence the safety system’s trip point. This will not always be possible, but where the management system has a direct connection to the IO multiplexer boards that connect to the eld instrumentation it is a risk. Attacking the actuator function Sometimes actuators are directly connected to the control network, are indirectly connected through Ethernet to serial converter boxes, or use Ethernet connected relays operating a circuit breaker. Even when actuators are connected with controllers or PLCs there are ways to (de-)activate them. They can be operated by the attacker through either spoofed message injections, modied messages or message replays. Situations where the attacker could initiate an action and obstruct the correction of this action or obstruct the operation of an actuator as a denial of service attack should also be considered. Finally, elements of the operating window, such as the travel time of a valve, are sometimes dened in the actuator. In these cases this information can be targeted, too. Attacking the sensor function Sensors can be attacked in several ways, depending on sensor type and the way they are connected and managed. Attack scenarios might involve replaying sensor reading messages to freeze the sensor reading, injecting sensor messages to modify the readings, or perhaps just blocking the sensor messages. An alternative could be to modify the sensor conguration, for example changing thermocouple characterisation settings to manipulate the sensor values. Sensors and actuators that allow over the network rmware upgrades are also vulnerable to denial of service attacks in which
52 PTQ Q1 2017
the attacker installs corrupted rmware to disable the sensor’s function. Analysing control system security The attack surface of a control system is much bigger than that of open systems such as Microsoft servers or network equipment. The internal hardness of a control system, achieved through the segmentation of the network into distinct security zones with restricted access, is vital. Making the attacker’s life difcult once he gains access to the control network is just as important as keeping him out to begin with. The more time the attacker requires to reach his objective, the more time you have to spot the attack and respond to it. Approaching control systems through the eyes of information security results in many vulnerabilities being missed. Control system security therefore uses a different approach, based upon cyber failure scenarios. A cyber failure scenario describes an action in the control system and how this can be executed by an attacker. A very simple scenario may describe how to open a circuit breaker in a power grid and prevent remote action to correct this. Cyber failure scenarios can be very process specic, but most are generic and can be used for a wide range of processes. The control system security analysis process will investigate which scenarios apply for the specic system, how the system is or can be protected against them, and how it detects the attack. This method uses the scenarios (documented as attack trees) and evaluates the likelihood that each branch of the attack tree is successfully executed. The critical branch is the branch with the highest likelihood (and therefore constituting the highest risk since the root of the tree is common to all branches). This can be used to rank the various cyber failure scenarios. Finally, the cyber attack space model (Figure 2) shows us that some (low effort) attacks require little system knowledge and can be relatively easily constructed in the absence of specic security func-
tions. Other attacks require more dimensions in the attack space model to accomplish and are therefore more difcult to undertake. Summary and conclusions Control system security looks at how the control system can be used to attack the physical system and damage it. This dimension is not addressed by the information security approach, which, if relied on, will result in gaps in the internal hardness of an industrial control system. Recent attacks such as that on a steel factory in Germany6 and on the power grid in the Ukraine7 have exploited these gaps. While there has not yet been a major attack on a rening company or infrastructure, there remains a high degree of risk and, as we have seen, learnings can be drawn from other sectors. It is a common misunderstanding that securing the open systems in a control system will secure the control system. During the design of the control system, many decisions are made that can result in potential vulnerabilities in the control system security. Ignoring these can lead to serious incidents. The industrial cyber security team from Honeywell is one of the few teams in the world to combine the disciplines of information security, control system security, and process automation. References 1 Zetter K, Countdown to Zero Day, Stuxnet and the launch of the world’s first digital weapon. 2 ANSI/ISA–95.00.04, Enterprise-Control System Integration, Part 4: Objects and attributes for manufacturing operations management integration. 3 Krebs B, Cyber Incident Blamed for Nuclear Power Plant Shutdown, Washington Post, Jun 2008. 4 Teixeira A, Sandberg H, Pérez D, Johansson K H, Attack Models and Scenarios for Networked Control Systems, Royal Institute of Technology. 5 Langill J T, Defending against the Dragonfly Cyber Security Attacks. 6 Bundesamt für Sicherheit in der Informationstechnik. Die Lage der IT-Sicherheit in Deutschland 2014. 7 E-ISAC. Analysis of the Cyber Attack on the Ukrainian Power Grid. Sinclair Koelemij is Technical Team Leader, Industrial Cyber Security EMEA with Honeywell Process Solutions.
www.eptq.com
Neles® NDX valve controller delivers performance perfected That’s how we make the big difference, the Metso Way.
The Neles NDX has been carefully designed and manufactured to make the big difference to our customers, regardless of end use industry or application. Savings are created through accurate, long-lasting and maintenance-free performance and through time saved as a result of extremely easy installation and use. The Neles NDX provides the reliability and robustness you’d expect from a valve controller by Metso, for all valve brands in standard applications. Find out more about the savings, safety and reliability that each Neles NDX valve controller offers at metso.com/ndx #TheMetsoWay
Get The Power of Big Data A PROGNOST® diagnostic system acquires and diagnoses 65 terabytes of sensor data every day. Impressed? Data itself means nothing The challenge is to convert data into meaningful information. Our intelligent machine and component diagnostics retrieve the knowledge operators need, to boost uptime while reducing costs. PROGNOST® keeps operators informed. Non-stop and in real-time. That´s impressive.
[email protected] www.prognost.com
Data operations transform fuels value A refiner applied advanced analytics to develop techniques for processing opportunity crudes which minimise negative effects on the plant CRAIG HARCLERODE OSIsoft
T
he petroleum industry is once again in the midst of titanic An integrated changes.1 Declining prices, downstream value chain expanding sources of supply, ris- Integrated fuels value chain ing regulatory requirements and, 4 refineries, POL 2 petrochem plants perhaps most importantly of all, a Logistics including dramatic shift in markets like trans2000 retail stations portation are forcing companies Bratislava SVK Refinery across the value chain to reconsider PI System overview CZR 4 HA collectives, long held assumptions about expan~400K tags sion, growth and customer demand. Duma AUT TYK Luckily, these new challenges are Refinery Elements coinciding with advances in big ~300 smart templates ROU ~21K elements SLO data, Internet of Things (IoT) and ITA Rijeka and growing predictive analytics and the abilRefinery Sisak ity to leverage to process opportuNotifications CRO Refinery SRB nity crudes and be more proactive ~150K templates BIH ~6K notifications and predictive in decision making. While the upstream oil industry has ~61K event frames been a somewhat enthusiastic adopincluding dynamic ter of digital technology, the midPI Coresight is the primary visualisation tool stream and downstream segments have been conservative and slow Figure 1 MOL’s downstream operations and operational technology infrastructure to adopt. That is changing with the Industrial Internet of Things (IIoT), generated by its distributed control (EBITDA) by $1 billion over a five advanced analytics and big data. systems (DCS) and other systems year period ending in 2016 through Collectively, we are inundated with as part of its operations. In 2012, more aggressive data modelling marketing messages that are add- MOL leadership, in response to and analytics. ing confusion and false promises, European competition resulting in Petroleum Economist named resulting in a good number of pro- low cracked spreads, embarked on MOL Downstream Company of the jects that go awry with limited or a business transformation enabled Year in 20162 while the FieldComm no business value and, worse, lost by digital technologies. Group gave the company its Plant opportunity costs. The results? MOL has developed of the Year Award for its Danube But we will also see implementa- techniques for processing oppor- facility. tions that will effectively serve as a tunity crudes while minimising blueprint because they will demon- the negative consequences such Background strate how digital technology as corrosion, operational issues MOL is one of Central Europe’s can reduce risks and costs while in areas such as the cokers, and largest downstream companies. It improving asset utilisation, yields, yields. Advanced corrosion ana- operates four refineries and two integrity and, most importantly, lytics such as high temperature petrochemical plants in eight counprofitability. hydrogen attach (HTHA) and other tries along with 2000 filling stations In fact, we already have such an forms of predictive corrosion have across 13 countries. To organise example. MOL, based in Hungary, been implemented across multi- data across its production facilities, has been on a journey to reinvent ple sites. In all, MOL estimates it MOL has been using the PI System its operations by better leverag- increased earnings before interest, from OSIsoft since 1998. The sysing operational data already being tax, depreciation and amortisation tem, which has expanded steadily,
www.eptq.com
PTQ Q1 2017 55
MOL downstream OT data model based applications Safety (PSM) and asset integrity
Interlock governance/DCS role tracking Operating envelopes Integrity operating windows (IOWs) Advanced alarm management Energy Energy monitoring management
Energy KPI breakdown (6 tiers) Column energy efficiency dashboards Hydrogen, utilities and energy balances Flaring CBM asset reliability
All critical rotating equipment Hydrogen pressure swing absorbers
Yields
Crude blending control Yield optimisation/reporting Product quality Analyser reliability Operational optimisation Plan vs actual analytics with future data
NG and fuel demand gas forecasting Peak electrical forecasting Normal mode of control loops APC monitoring PI AF and Sigmafine (PI AF) used for yield accounting and material movement
Figure 2 MOL’s downstream OT data model based applications
is divided into four high availa- together – the smart OT infrastruc bility collectives with a combined ture with PI Asset Framework, total of approximately 400 000 ‘tags’ and PI Coresight – MOL had built or data points. More importantly, a self-serve analytics and business MOL utilises PI Asset Framework intelligence environment where with smart asset objects to pro- operators and engineers who travide a configurable, dynamic smart ditionally used Microsoft Excel can operational technology (OT) infra- configure their own smart asset structure. Currently, MOL has over objects, combine them like Lego 300 smart asset object templates blocks and create their own dig300 templates, 21 000 elements, ital replica and experiment with and over 61 000 event frames for potential improvements, and then signalling the occurrence of key execute changes across the MOL parameters or events (see Figure enterprise with governance. 1). Tibor Komroczki, who leads the Information Integration and Challenge Automation team at MOL, refers to critical availability problems the PI System as the MOL common language as it enables the abstraction and nomination of a diverse tag and asset naming, units of measure, and time zones. MOL generates over 80 billion data points per year. The PI System served primarily as an operations system of record until 2010 when Komroczki led an effort for digital transformation. As a first step, MOL adopted PI Asset Hydrogen production plants are critical units in the refinery Framework to create a so-called Pressure-swing absorbers (PSAs) are ‘digital twin’ of different processes critical equipment in unit operation and equipment sets in a facility. Cyclic operation: heavy load on valves With PI Asset Framework, all of (9-10 open-close hourly) the relevant data streams, meta $1.2M loss in three years due to PSA data, calculations and analytics, valve failures and alerts and notifications from a process step are combined into a UPTIME program: 97% operational comprehensive, digital replica of availability the plant. Additionally at this time, it adopted PI Coresight, a visual- Figure 3 Advanced analytics can isation tool for displaying and/ predict the impact and ripple effects of or analysing AF models. Taken opportunity crudes
56 PTQ Q1 2017
With the smart OT infrastructure in place, MOL established a foundation for higher level efficiencies because it could connect its assets relatively easily and track its performance backwards and forwards. New applications can be added rapidly. Komroczki asserts that greater control over data has ena bled MOL to move from managing in a reactive sense to predictive management to management by exception as indicated by the existence of over 61 000 event frames. Some of the achievements include improved asset integrity and safety, asset health, improved energy efficiency, increased yield, reduced hydrocarbon loss, improved environmental reporting, and reduced maintenance costs (see Figure 2). Another plus: MOL reduced its IT costs and reliance on outside vendors because employees were able to quickly build their own functionality on top of their infrastructure and then replicate it across foundries and, in doing so, simplifying and standardising its application and solutions portfolio. Different data streams can also be analysed in tandem so that MOL could determine the full impact (financial, maintenance, energy consumption) on changes to output. MOL employed analytics to reduce the risk of high temperature hydrogen attacks (HTHA). By studying the relevant operational data, the company was able to pinpoint the temperature and pressure parameters that increased the risk of HTHA. They developed a smart asset HTHA application template that was deployed in six units in less than a week. Following the successful test, it was rolled out across MOL’s plants in 2015 to over 50 pipe nodes. Advanced analytics potentially can be applied in a wide variety of ways: energy modelling optimisation; the impact and ripple effects of opportunity crudes in areas of corrosion, fouling, and efficiencies; the economic gains to be achieved through opportunity crude processing; better understanding of advanced control; and preventative and prescriptive maintenance (see Figure 3).
www.eptq.com
Together facing a brighter tomorrow At Yokogawa, we believe the sky’s the limit. And to reach beyond today’s horizons, we work step-by-step with you to make the unimagined a reality. That’s how we move forward, through the synergy of co-innovation partnership. Join hands with us, and together we can sustain a brighter future. Yokogawa: Building a better tomorrow with you today.
Please visit www.yokogawa.com/eu
The integrated OT Smart infrastructure Yield accounting P&S Unit models
DCS
Financial data
Varies by site
ERP Financial data
EAM IIoT/Edge Laboratory data
LIMS PI System OT Object Model OT data model / infrastructure PI integrator for Azure
Opralog e-logbook
NICE Natural info centre
Microsoft Azure
Figure 4 MOL adopted Microsoft Azure machine learning for a production environment
Machine learning tion layered on top of what it had Once MOL had the smart OT already implemented. infrastructure across its value Following these successes, MOL chain with associated IIoT analyt- turned to improving the perforics, focus was turned to machine mance of its delayed coking units. learning and “big data analytics”. By using opportunity crudes, MOL MOL has become one of the rst, if estimated that it could gain $6 milnot the rst, large rener to adopt lion for each 1% gain in DCU yield. Microsoft Azure machine learning Gains in DCU yields with variain a production environment (see ble feed from opportunity crudes, Figure 4). Microsoft Azure works however, also increased the risk of in conjunction with the PI System: steam explosions during the hydrooperational data is uploaded to the cutting step. cloud and then analysed across Azure analytics combined with Microsoft’s cloud infrastructure. continual data feeds from the PI MOL has developed Azure System enabled MOL to thread the machine learning to predict the needle. DCU yields were increased impact of sulphur levels in var- by 2%, yielding an estimated gain iable feedstocks in their various of $12 for each unit per year. At the desulphurisation units. MOL had same time, steam explosions went been using ofine models for ana- down by 75%. Machine learning lysing sulphur. Not only did using enabled the ability to achieve two ofine models increase time, it seemingly contradictory goals at also increased the potential for the same time. The company has error. MOL estimated it was los- now positioned machine learning ing $600 000 per year across four for its four DCU units across its units because of its inability to enterprise to take full advantage of adjust unit parameters to optimise opportunity crudes. sulphur content in the products. MOL eliminated the losses thanks On-premise vs cloud? to better forecasting and continues No: on-premise plus cloud to roll out the technology across While data gets transferred and its infrastructure. As with its other stored in the cloud with machine improvements, MOL was able to learning analytics, cloud systems leverage its previous technology typically will not replace on-preminvestments: the new applica- ise storage systems. Real-time con-
58 PTQ Q1 2017
trol and insight are required for operational efciency as well as safety. Transferring data to the cloud invariably increases latency: data simply has to move far further before it can be used. It also increases risk because a disruption in the network can lead directly to disruptions in operations costing millions in downtime. (Anyone who has worked with offshore upstream companies is likely familiar with the risks of satellite links.) Instead, these systems complement each other. Companies are opting to maintain on-premise systems and transfer data, or summaries of data, on an as-needed basis, preferably during hours of low network trafc. A substantial part of the success revolved around the use of integrators that effectively automate the translation process of bringing OT data to IT-based analytics systems. Using a CAST (clean, augment, shape, transmit) methodology, MOL was able to avoid the data prep and ‘data janitor’ problem that can take up to 80% of the time of projects. Other companies in similar heavy industries have experienced similar results: Cemex, the large cement manufacturer, has reduced the amount of time required for preparing and gathering data across 70 plants for its reports from over 700 hours to less than one through CAST automation. The financial bottom line One of the more compelling features of MOL’s transformation, and likely a phenomenon others will experience, is that the changes are additive. Once the foundation for digital transformation is in place, additional applications can be added on top of the now existing digital infrastructure. As a result, incremental improvements can accelerate savings, rather than result in ever shrinking marginal gains. Over the four year period running from 2011 and 2014, for instance, MOL estimates that its digital transformation programme accounted for an additional $500 million in EBITDA. During the next two years, however, MOL added an additional $500 million to EBITDA,
www.eptq.com
bringing the total over five years to been mandated by regulations. $1 billion. An equivalent amount 3. Just do it. It is a journey of conof savings was achieved in roughly tinuous improvement. Search for half the time. improvements that can be impleWhile the curve may change over mented now and add others as time, one can expect that savings time goes on. Separating problems will compound. Each new improve- can allow plant managers to resolve ment potentially will cost less than individual problems more quickly the ones that went before it because as well as document progress for MOL can leverage all of the previ- upper management. ous advances. Improvements made 4. Determine where analytics are in the first year should also con- performed. Calculations such as tinue to grow as additional data is exchanger and pump efficiencies, continually fed back into the sys- energy utilisation, and yields or tem to achieve Kaizen-style gains. advanced CBM can and should be done in the OT infrastructure Next steps closer to the assets. Performing OT MOL continues to mine for ways analytics in the OT data infrastructo apply analytics to its business. ture will also enable the migration In 2017, it wants to increase white of analytics to the edge over time. product yield by 2.5% through Meanwhile, more extensive analytincreased conversion and more effi- ics that might require thousands of cient crude processing. To increase servers and multiple data streams its buffer against market swings, it are better suited for the cloud. One will additionally improve the diesel way to think of the difference is to mogas ratio from 2.4 to 2.8. Flare that analytics for individual plants gas recovery and hydrocarbon loss or processes are best conducted in management initiatives, tracked OT while enterprise-wide analytics through continuous improvements are most suited for the cloud. to monitoring and tracking sys- 5. Bridge OT and IT through tems, are under way. Automation. This can be accelerated by the use of an integration layer Lessons learned that CASTs operational data so that What did MOL and other leading it can be consumed in unstructured companies do differently in apply- IT systems. These data integrators ing IIoT? There are five lessons: effectively automate data prepara1. Do not forget that it is about tion and translation. delivering business value and not IIoT, advanced analytics and big applying IIoT and advanced analyt- data are here and growing, make ics for technology’s sake. MOL has no mistake. They will dramatically seen technology as a means to an transform our largest and oldest end, not the other way around. industries. If you approach their 2. Start the journey by creating a implementation and use strategifoundation for data. By creating cally with the approach presented a digital infrastructure MOL gave above, you will increase the probitself a scalable, coherent infra- ability of value sustainable attainstructure. It created both a virtual ment from your IIoT, advanced model of its plant through AF and analytics, and big data initiatives. a means to implement and measure those results in reality. The infrastructure approach also made it far References easier to develop new functional- 1 www.iea.org/publications/freepublications/ ities because the same basic foun- publication/medium-term-oil-marketdation could be used for multiple report-2015.html 2 www.petroleum-economist.com/ functionalities. articles/corporate/pe-award-winners/2016/ In the end, MOL created what downstream-company-of-the-year-2016one could call an ‘OT chart of mol-group accounts’ where all OT data gets aggregated across a portfolio similar to the financial or ‘IT chart Craig Harclerode is the Industry Principal, Oil of accounts’ which structure has & Gas, with OSIsoft.
www.eptq.com
Level Measurement LNG & Low Density applications
■
Meets the design standard of the LPG/LNG industry: – Includes NACE requirements
■
Reliable level indication without external power with three functions in one device: – Liquid level indication – Liquid level switch – Liquid level transmitter
■
Equipment with optional bi-stable change over or on/off switches and linear or non linear transmitters (also as 2-wire systems with 4…20 mA output, optional with HART ®)
■
CE-conform, Ex-approval (ATEX & IECEx)
■
For operating pressures from vacuum up to 50 bar, operating temperature from –196 °C up to +400 °C
■
Densities > 0.27 g/cm 3
■
Closed, non-pressurized floats
■
Construction in stainless steel grade 316 & 316L as standard.
Casing of indication rail in stainless steel, float failure indication on request i s i t u s P lea se v : we b s i te w a t o u r ne h c g. ka -a w w w .we WEKA AG · Switzerland Schürlistrasse 8 · CH-8344 Bäretswil Phone +41 43 833 43 43 · Fax +41 43 833 43 49
[email protected] · www.weka-ag.ch ■
ARCA Flow Group worldwide: Competence in valves, pumps & cryogenics
PTQ Q1 2017 59
5th
REFINING
INDIA
's international conference for India PTQ 's held in partnership with Industrial Development Services
Lalit Hotel, New Delhi 3-4 18-19 September 2017
ptq PETROLEUM TECHNOLOGY QUARTERLY
refiningindia.com
Profiting from plant data Predictive analytics is an evolving technology with many potential applications in – and gains for – the process industries DOUGLAS WHITE Emerson Automation Solutions
T
ypical plant problems include: 1. You are a process engineer Previous Today and suspect that the catalyst Status information in one of the site’s reactors is deactivating faster than it has historically. Is your suspicion correct, and if it is, how much faster and what is the cause? Will the catalyst last until the next shutdown? If not, what operatDiagnostic information ing changes does the plant need to Travel deviation Valve signature signature make to ensure it will last? Cycle counter Step response 2. You are a reliability engineer try Air supply pressure Dynamic error drive signal ...etc. ing to understand why some pumps Output signal need frequent overhauls and some do not. What are the characteristics of those that repeatedly require this service? Are there leading indicators that will allow you to identify beforehand when a pump problem is about to occur? 3. There have been several signif Valve Valv e position icant flare releases over the past few years with associated visible plumes. The source and composition of the releases are not obvious. You have been asked to find the cause and recommend changes to avoid or Figure 1 Valve feedback output data at least reduce the incidence of these In recent years, rapid progress has most recent three years of historical releases in the future. What is the common factor in all been made in innovative innovati ve algorithms one minute sampled data for these of these issues? Their resolution is and approaches. How are the pro- points. going to require deep analysis of cess industries taking advantage Even more data can be expected plant data including data beyond of these new techniques today and from process plants in the future. standard process measurements. how might they apply them in com- Much of the new equipment and Another commonality is that time ing years? devices purchased for process In a related trend, process plants plants is already equipped with is an explicit variable and solutions should generate predictions of are producing more and more data. multiple sensors to monitor interfuture plant and equipment behav- ‘Big data’ is a popular topic and the nal conditions and performance, iour that supports decisions. How process industries are no exception. embedded computing to analyse the can these questions be addressed One refining company indicated it data produced, and enhanced conmore efficiently – more quickly and produced 80 billion data items for nectivity to transmit the results. This storage in one year from four sites. continuing evolution in the numwith less manpower requirements? In this article, some of the new Another process company refer- ber of sensors and their computing methods that are available to assist enced a corporate historian with 10 and communication capabilities is with this data analysis and develop- million tags across 15 sites and plans sometimes referenced as the ment of solutions will be reviewed. to implement online access to the Industrial Internet of Things (IIoT).
www.eptq.com
PTQ Q1 2017 61
What is forecast to happen Predictive analytics
What is happening Real-time analytics
Future What has happened
Figure 2.
Descriptive analytics
Present
Past
Figure 2 Analytics time line
But how can plant staff extract value practice of capturing, organising, from all of this data? and analysing data to determine Figure 1 shows one example. In patterns, correlations, and concluyears past, there might be one (or sions. Engineers, of course, have even zero) feedback signal from a been doing analytics, in practice, for eld control valve. Today there can as long as they have been employed. be more than 100 including sophis- However, the continuing increase ticated diagnostic information and in computing capabilities has led to generated graphical performance developme development nt of practical algorithms that can address signicantly larger representations. Of course the increase in data is data sets and more heterogeneous not limited to the t he process industries. data types. Much larger increases are reported in other industries. Walmart is reported to collect and store 2.5 Refinery pentabytes of unstructured data every hour from the more than one million customers around the Action world currently purchasing items. 1 Algorith Algo rithms ms Amazon is estimated to have more than 1.5 million servers in their data Acquire centres.2 Dynamic event detection In this article the focus is on the and sensor data capture plant operational impact of these developments and their potential Analyse business impact. One characteristic Extract information of most of these operational issues is patterns that they have a temporal/dynamic character which leads to special data Prediction requirements. There has also been an equivalent and important impact Models on other corporate activities such as trading, customer and market analEnd users ysis and nancial services which Present: has been covered in many other HMI Distribution publications. Data analytics Data analytics is the theory and
62 PTQ Q1 2017
McKinsey 3 has ranked big data analytics as one of the top potential technologies that can increase productivity and GDP over the next few years. Among business sectors, manufacturing was the lead with an estimate of overall GDP yearly increase of $125-$250 billion. Data analytics can be classied in three categories based on the time scale of interest as illustrated in
Figure 3 Predictive analytics implementation
Descriptive/historical analytics is analysis of what has happened, when it happened and why it happened. This includes producing retrospective performance measures. An example is calculation of actual versus planned energy usage for plant equipment as well as the overall site for the past day, week or month. Real-time analytics is the use of the most recent and historical data to interpret current plant conditions. An example is comparison of plant operating conditions against the approved ‘operating window’ and issuing an alert if the operating window is being violated. Predictive analytics is prediction of what will happen and when it is likely to happen based on an analysis of current and past data trends and patterns. An example is estimating the likelihood that a pump will need servicing before the next shutdown. Predictive analytics is the focus of this article. Predictive analytics applications Figure 3 shows a typical predictive analytics application in the process industries. Relevant sensor and event data is captured in real time. They are automatically analysed for patterns of signicance and if one or more patterns is detected a prediction is generated from the model. If appropriate, an alert is generated or specic compensating action is initiated. An example is the pump condition question at the beginning of this article. The spectral properties of the vibration measurement from the pump and the statistical properties of the outlet pressure measurement might be the inputs of signicance. If both of these variables increase signicantly at the same time the
www.eptq.com
digitalrefining.com is the most extensive source of freely available information on all aspects of the refining, gas and petrochemical processing industries.
It provides a constantly growing database of technical articles, company literature, videos, industry news and events.
Common predictive model algorithms Problems types
Plant application examples
Types of algorithms
Types of models
General application examples
Comments
Regression
Predict energy use in plant based on planned operation and models Spectral data interpretation Video/photograph characterisation
Generalised linear regression
Linear – continuous variables output
Marketing campaigns
Still the most common algorithm used
Neural nets
Image recognition
Training requires lots of data
Convolutional neural nets
Non-linear – continuous variables Non-linear – continuous variables
Fingerprint identification; photo sorting
Principal component analysis (PCA)
Linear continuous variables
Human gene to disease correlation
Anomaly detection Fault detection; product quality tracking Classification
Fault detection Sorting documents
Logistics regression Decision trees – random forest Product quality prediction, Support vector data preclassification machine (SVM)
Binary outcome – Yes/No Fraud detection Discrete outcome – Recommend movies A,B,C,D to watch; music Binary outcome – Image recognition Yes/No
Clustering
Spectral data encoding; characterisation
Nearest neighbour; K-means,
Clusters similar data
Recommend movies to watch; music
Natural language processing
Text sorting
Text analytics
Textual analytics
Email spam identification
Random forest often most accurate but hard to interpret Handles high dimensionality and large data sets
Table 1
model might suggest a potential future pump problem and suggest action by operations and maintenance to address the problem. Predictive model development Central to the success of any predictive analytics project is the development of an appropriate model which accurately forecasts the occurrences under investigation. Many of the popular algorithms for model building have been known to statisticians for many years. What has changed is the ability to apply these algorithms to extremely large data sets and widespread availability of easy-to-use software for these applications. This has resulted in many recent applications in the consumer area including such familiar ones as fngerprint recognition on smart phones, recommendations for music and movies from previous selections, and driverless cars. Table 1 lists some of the major classes of algorithms and their existing and potential use in the process plant area and in more general consumer applications. There is much overlap in terms of potential applications and the approach used often depends on the preferences and experience of the user. For more background on these algorithms see references 4 and 5.
64 PTQ Q1 2017
Statistical methods for regression have been and remain the most popular for plant applications with a further subdivision into those employing linear models versus non-linear ones. Principal component analysis (PCA) is a subset of linear analysis that identifes a reduced set of derived variables from the data set that approximate the entire set. These variables can be used to identify the normal operating region for a piece of equipment. Deviations outside this region are often characteristic of faults. For non-linear models, neural net algorithms have been gaining importance because of their ability to model quite complex data such as photo recognition. However, in contrast to linear regression, the resulting models have no easy physical interpretation. Classifcation algorithms are often used when the results are discrete: binary such as yes/no, or multi-class such as A,B,C or D. Logistics regression is regularly used for binary cases with ‘if-then splitting’ (CART trees) often used for multi-output problems and large data sets. Support vector machine algorithms are popular for binary problems with a large number of input variables and large data sets and non-linear boundaries. Clustering is a technique to determine if there are
distinct groups within data sets with similar characteristics. It can be used to sub-divide large data sets into more manageable and meaningful sub-sets. Natural language processing is used to convert unstructured text into structured indices that can then be used for further analysis and to segregate the source text into relevant sub-groups. Most applications to date in the process industries have been in the equipment availability/reliability area with the system intended to warn of impending faults or of deteriorating performance. These applications will often use a combination of the algorithms listed above. For example, clustering can be used to identify groups of pumps with similar maintenance histories and these groups can then have individual predictive models with resulting higher accuracy than lumping all pumps together. The methods discussed above are primarily examples of ‘supervised’ model building. In this case there is a data set where the output is generally known and the desire is to determine if the other variables in the data set can predict the outcome. An area of current research is what is termed ‘unsupervised’ learning where the system processes a large amount of data to determine if there
www.eptq.com
break through the
NOISE
Better Signal-to-Noise Ratio Means Better Level Control Performance 4.40 Model 706 SNR
1.57 Competitor SNR
The ECLIPSE Model 706 transmitter has a signal-to-noise ratio nearly 3 times higher than competitors.
While transmit pulse amplitude (signal size) has helped to make guided wave radar technology the standard for accurate, reliable level measurement, the fact is signal-to-noise ratio (SNR) represents a far more critical indicator of level control performance. For superior SNR in all process conditions, no other GWR device beats the Eclipse ® Model 706 transmitter from Magnetrol®. To learn more about the breakthrough ECLIPSE Model 706 GWR transmitter visit radar.magnetrol.com or contact your MAGNETROL representative today.
magnetrol.com • 1-630-969-4000 •
[email protected]
© 2016 Magnetrol International, Incorporated
Equipment condition and performance Relational data Maintenance Inspection Work orders ...etc. Spectral data Rotating machinery Vibration Thermography Infrared/acoustic/ ultrasonic
Unstructured text Daily operating plans Targets Notes, etc. Email
Model building
Spatial
Operational real-time and historical data Process Offsites Utilities Terminal data Analysers (NIR) Lab data
Video Flares Process equipment ...etc.
GIS/location
Figure 4 Typical plant predictive analytics data requirement
is any structure to the data or relationships without prior specication of the desired answer. Process plants and analytics Much of the work on analytics has taken place outside the process plant environment. Why are plant problems more difcult? The rst area of difculty is the character of the data required. Figure 4 shows typical data inputs for plant predictive analytics problems. Note the diverse data types. For the catalyst example, operational real time and historical data is required but also types of feedstocks and their analyses. To predict whether the catalyst will last requires data on the planned feedstocks, rates and modes of operation. The pump example requires time synchronisation of operational data with maintenance records and use of equipment data sheets and history. As mentioned earlier, time is an explicit variable which means that all of these records have to be accurately time synchronised. For the are example, how can the historical video footage of the are be automatically searched to identify when there was a visible plume and coordinate the incidents with operational data and equipment work orders at preceding time periods? Much useful information may be in unstructured text such as operational logs that need to be analysed. To further complicate the issue each of these data sources typically has its own physical storage, data organisation, search programs and user interface. For operational real time and his-
66 PTQ Q1 2017
The other issue is that few of the analytic algorithms were developed to model dynamic systems, that is systems that are time varying. As emphasised here, that is exactly the type of models that are needed for most plant problems. Fortunately, there has been a recognition of this deciency and suitable extensions of the algorithms have been and are being developed.
torical data, process plants often have residence times that are signicantly longer than typical data sampling times. For the renery crude unit shown in Figure 5 , a uid element entering the unit can take one to two hours to exit. If product quality is measured at the current time and you want to correlate process conditions with the quality, you have to use the conditions at the time the uid element was transiting in that part of the plant. To accomplish this requires selecting the measurements from that part of the plant at past intervals corresponding to the residence time between that point and the sampling point. This is called ‘time shifting’ the data. If the product was stabilised naphtha and the time of sampling is t, you might have to use data from the stabiliser column at t-10 minutes, data from the crude column overhead system at t-20 minutes, data from the crude heater at t-55 minutes, and crude composition data at t-85 minutes to properly analyse the unit relationships. Identifying the actual delay times empirically from the data is commonly one of the rst steps that has to occur in analysis. In addition, process data is natively of poor quality. It is often corrupted by noise and has poor statistical qualities. It can be auto-correlated, cross-correlated, collinear, and non-stationary. ‘Data cleaning’ is the term used to describe the necessary and very important data preparation step in analysis. It includes handling missing data, removing outliers, scaling, compensating for non-stationarity, and so on.
What are the expected benefits? To evaluate the benets from the increased use of predictive analytics rst consider the operating objectives of a typical large continuous process plant. The goals can be summarised as ‘The Four Zeros’: • Safety: zero safety incidents • Sustainability: zero environmental incidents, excess energy use, and waste • Availability and reliability: zero unscheduled downtime • Financial optimisation: zero lost prot opportunities. How can predictive analytics contribute to meeting these objectives? Although there are many possibilities, the most common areas are: Improving safety performance • Avoiding incidents through early detection of potential hazardous situations – multivariable equipment operating window • Reducing personnel exposure to hazardous conditions through enhanced detection of likely hazards. Supporting sustainability • Comparing current utilisation of resources, such as energy and other utilities, to their expected usage under current conditions with a determination of possible causes of variation • Early identication of potential environmental emission events to facilitate mitigation. Increasing plant availability and reducing unplanned downtime • Anomaly detection – identifying precursor events to unscheduled equipment outages or problems, allowing required maintenance activities to occur before the event • Performance monitoring for condition based and preventative main-
www.eptq.com
Stabiliser data at time ( t− 10) mins Column overhead data at time ( t− 20) mins
Stabiliser outlet sample at time t mins Off gas
Reflux drum
Atmospheric tower
Naphtha stabiliser
Light naphtha
Pumparounds
Stripper columns T-6
T-5
Heavy naphtha
T-4
Jet fuel / kerosene
Crude oil feed quality at time ( t− 85) mins
Diesel
T-3
Preheat Crude
Intermediate gasoil
T-2 Desalter
Heavy gasoil
Desalter outlet data at time ( t− 70) mins
T-1
Heater
Steam
Heater outlet data at time ( t− 55) mins
Atmospheric residue
Figure 5 Time shifting data
tenance – detecting loss of process/ equipment performance before it impacts production capacity allow ing for scheduled rather than unscheduled maintenance Reducing the impact of corro sion and fouling on equipment with enhanced corrosion monitoring and forecasting Reducing turnaround duration and frequency through better knowl edge of actual equipment conditions.
Determining reasons for product quality/yield issues. •
-
•
-
•
-
Increasing margin and productivity Energy and utilities demand fore casting – predicting potential con tract penalties before occurrence Detecting and dissecting complex interacting constraints (sometimes multi-unit) on production to help overcome bottlenecks Increasing the yield of most val uable products by identifying opti mum operating conditions •
-
•
•
-
www.eptq.com
and the estimated nancial or other benets. While sometimes difcult, this is a prerequisite to gain ing approval and support from sen ior management which is important for eventual programme success.
-
Typical analytics project What are the typical steps in execut ing a predictive analytics project?
-
-
Problem definition Inventory and acquire available data The rst step is to develop a clear Identifying, locating and gaining idea of the objective of the project. access to available data is usually What do you want to understand and the next hurdle. While process data predict? How will results be used in a historian may be easily availa to improve business performance in ble, it is seldom all that is required. the plant? What are the specic per For the example, aring events formance measures or metrics to be may be documented in environmen used? For the are example listed tal reports, plant operating instruc at the start of the article this objec tions in daily instruction documents tive might be: identify the cause(s) and unusual operating situations in of recent aring events in the plant log books. Maintenance issues on and possible mitigating strategies to equipment may be in work order avoid them in the future. records.
-
-
-
-
Cost and benefits estimation Next, it is necessary to develop an expected cost for the programme
Data cleaning, consolidation and transformation It is often the case that gathering
PTQ Q1 2017 67
and preparing the data for analysis consumes from one-third to one-half the total project hours. What data should be collected? How should it be stored, particularly non-numeric data? How should search and retrieval of data of interest be implemented? Next is the data ‘cleaning’ phase, as discussed above. Data needs to be brought to a consistent format with a uniform timescale. Missing data and outliers need to be excluded. For numeric analysis, scaling is often required. The better the training data, the better the model. For the are example, data time synchronisation is likely to be important and consistent data formatting an issue.
Plant 1 Plant 2
Surge basin
Plant 3 Plant 4 Plant 5
Denitrification basin
Plant 6 Plant 7 Plant 8
Clarification basin
Plant 9 Plant 10 Plant 11
Filtration
racy and increase uncertainty in the model predictions. Accuracy should be audited regularly and model parameter updates performed when appropriate. Current data has to be cleaned and transformed prior to use by the prediction model just as the training data was. This procedure is automated in the production environment. Project review At the conclusion of the project the actual return should be evaluated against the initial objective. What was the actual nancial return versus the investment required? What are the lessons learned and best practices? How to get started If you are considering your rst application, how do you get started? Generally a phased approach that is targeted at a specic limited, but important problem that can be addressed in a few weeks or months is better than a ‘big bang’ type programme. A ‘pilot’ project that generates some benets can help build support for more extensive future programmes. Finding an executive sponsor who is interested in the programme and the results is much better than conducting an isolated study. Experienced consultants can help.
Model building and evaluation Plant 12 As discussed, there are many algorithms that can be used for model Settling pond building. Before embarking on complex models and analysis it is often Outfall worthwhile to start with visualisation and simple analysis with Figure 6 Wastewater treatment plant linear models, to see if any relationships seem to exist. Confusion there may be relatively few actual between correlation and causation is events from which to build a model. common and needs to be addressed. This typically requires special algoAt the end of the project, the model rithm approaches. has to make sense for it to be used with condence. Model deployment, monitoring and Previously, use of the various pre- updates dictive analytics algorithms required For the model to be useful it needs signicant experience and expertise. to be moved into the daily deciMore and more easy-to-use software sion making process. Who will take Case studies is becoming available which allows action based on the results of the pre- 1. Reducing effluent emission those interested to build analyti- diction? What are their needs for a violations cal models without rst becoming user interface? Standard questions A US chemical company installed an expert. These also often provide include: what infrastructure should biological denitrication in their an infrastructure for implementa- be used?; what tools should be used wastewater treatment plant to tion, which will be a requirement for for deployment?; should the data reduce ammonia in their water eventual use. be stored on premises or off-site (in discharge (see Figure 6). The proTesting and validation of the model the cloud)? As with model building, cess uses aerobic bacteria to conare the next step. Ideally there is there has been recent development of vert ammonia to nitrates. However, enough data to segregate testing and a number of new software offerings the company experienced occavalidation data sets from the training in this area. In contrast to the past, sional breakthroughs of ammonia data. After the model is developed it is possible to deploy a solution that caused the water discharge to with the training data, it is tested using these new infrastructures with- exceed specied efuent ammoagainst the test data set. If parameter out requirement for an experienced nia limits. To assist in identifying changes are then made to the model expert on the specic software. There issues, they decided to conduct a the improved model is checked is a standardised documentation for data analytics study to see if the against the validation data set and models – Predictive Model Markup breakthroughs were associated with the accuracy reported. This test- Language (PMML) – that can facili- specic plant conditions. One of ing and validation step is extremely tate movement from a development the issues in the identication was the large residence time in the sysimportant in developing an estimate environment to deployed status. of the accuracy of the model. There are often changes in the tem, from one to two weeks for the The are example is a case where plant over time that degrade accu- total system depending on individ-
68 PTQ Q1 2017
www.eptq.com
ual plant discharge rates, although centered maintenance” and predicthere were intermediate measure- tive diagnostics. The results were a ments that did identify likely break- 12% decrease in maintenance costs through. Another was that there while achieving a 2% increase in were multiple plants discharging to overall plant availability. the waste water system, each with different product production sched- Conclusion ules. More than 200 measured pro- In summary, predictive analytics is cess variables were considered as an evolving technology with many candidates and a thorough corre- potential applications in the process lation study over five years of data industries. However, implementawas performed using PCA. The con- tion in these industries has distincclusion of the study was that a drop tive issues due to the time series in the pH in the effluent from one character of the data and associated of the plants preceded many of the complex data requirements. Current breakthrough indications by one to applications are demonstrating two days. This drop in pH corre- benefits in process fault detection, sponded to a particular type of pro- increasing equipment availability, cess operation in the plant. improving safety performance and supporting financial optimisation. 2. Reducing maintenance, increasing There will be more of these applicaplant availability tions in the future. A large Middle East chemical comReferences pany commissioned a major new 1 McAfee A, Brynjolfsson E, Big Data: The site with over 6000 important Management Revolution, Harvard Business equipment assets requiring regular Review, Oct 2012. maintenance. After initial commis- 2 Clark J, 5 Numbers That Illustrate the Mindsioning of the plant, an integrated Bending Size of Amazon’s Cloud, Bloomberg, maintenance programme was 2014, www.bloomberg.com/news/2014instituted centred on “reliability 11-14/5-numbers-that-illustrate-the-mind-
bending-size-of-amazon-s-cloud.html 3 Bloom A, McKinsey on Big Data Analytics: The #1 Key to US Economic Growth, 2013, blog.pivotol.io/pivotal/p-o-v/mckinseyon-big-data-analytics-the-1-key-to-useconomic-growth 4 Hastie T, Tibshirani R, Friedman J, Elements of Statistical Learning – Data Mining, Inference and Prediction, 2nd Edition, Springer, 2008,
Free download at statweb.stanford.edu/~tibs/ ElemStatLearn/ 5 Rudin C, Prediction: Machine Learning and Statistics, Course (free) under MITOpenCourseware, ocw.mit.edu/courses/ sloan-school-of-management/15-097prediction-machine-learning-and-statisticsspring-2012/index.htm Doug White is a Principal Consultant with
the Process Systems and Solutions Group of Emerson Process Management. Previously he held senior management and technical positions with MDC Technology, Profitpoint Solutions, Aspen Technology, and Setpoint where he was responsible for developing and implementing advanced automation and optimisation systems in process plants around the world. He has published more than 50 articles on these subjects and holds a BChE from the University of Florida, an MS from California Institute of Technology, and an MA and PhD from Princeton University.
EXPOEUROPE 28 - 30 MARCH 2017 - AHOY, ROTTERDAM
Register now qq www.stocexpo.com
www.eptq.com
Organised by
PTQ Q1 2017 69
™
Adv anc ed F IBE R F IL M C ontact or T echnol ogy
Reduce Costs Improve Efficiencies Hydrocarbon Treating Made Better • Improve Treating Rates • Reduce Carryover • Reduce Plot Space
Introducing the next generation of FIBER FILM ® Contactor technology. FFC Plus™ is an advancement to FIBER FILM technology that makes mercaptan extraction more efficient than ever. The latest innovations to this commercially proven technology allow for more than double the treating capacity while reducing overall costs and plot space. FFC Plus
can also be easily retrofitted into existing FIBER FILM units as a drop-in solution that is simple to install during a unit turnaround, requiring less time than a standard cleaning. In addition, turndown can be maintained at current levels. Add capacity for tomorrow without losing operability today.
Optimal processing scheme for producing pipeline quality gas Exploring processing line-ups for optimal design of sour gas processing plants
SAEID MOKHATAB Gas Processing Consultant GERRIT BLOEMENDAL Jacobs Comprimo Sulfur Solutions
T
he gas processing plant must be a ‘t-for-purpose’ design, meeting project economics and environmental requirements. When determining the proper solution set for a gas processing pro ject, it is not only important to select the proper technology within each processing unit but also to adjust sequencing of the processing steps in order to optimise the whole operation, thereby increasing operating exibility within the overall gas processing plant. This article discusses the considerations required to determine the optimal processing line-up for sour gas processing plants producing pipeline quality gas. The choice of gas processing plant conguration and its complexity depend upon the feed gas compositions and the levels of treating and processing required in meeting product specications and emission limits. The values Sulphur recovery
the levels of natural gas liquid (NGL) components to be recovered Gas processing (see Figure 1). While contaminants plant Pipeline Raw and sulphur compounds must be quality Producing gas gas removed to meet emissions require pipeline quality gas ments, the extent of processing to recover NGL components is project Condensate Sulphur specic.1 A typical scheme for a gas proEthane cessing plant producing pipeline quality gas is shown in Figure 2. As Gas processing Lean gas plant can be seen, different gas processRaw to Producing gas sales/LNG ing steps are required to meet the natural gas plant product specications dened by liquids the gas processing plant’s owner. The number of combinations of e n e t e u r n a h a t a possible processing steps presents s l p o p B u e n S u r d P a challenge to determine the best o n C processing scheme to meet technological and economic targets Figure 1 Two configurations of gas while providing exible and reliaprocessing plants1 ble operability. In conguring the optimal processing line-up, the of the hydrocarbon liquid prod- plant designer must understand ucts can also be drivers for pro- the technology options available, cess complexity, which determines their integration opportunities and Tail gas treating
Sulphur degassing Acid gas removal
Raw gas
Inlet separation
Gas dehydration
Incineration
Flue gas
Liquid sulphur to export/solidification Mercaptans removal (if required)
Hydrocarbon dewpoint controlling
Gas compression (if required)
Pipeline quality gas
Off-gas
Water Condensate stabilisation
Condensate to storage and export
Figure 2 A typical gas processing plant producing pipeline quality gas
www.eptq.com
PTQ Q1 2017 71
Sulphur
Sulphur recovery
R-SH
Regen-gas treating (physical solvent) H2O
H2S, CO2
Feed gas
Gas dehydration and mercaptans removal (molecular sieve technology)
Acid gas removal (promoted MDEA)
Hydrocarbon dewpoint controlling (propane refrigeration system)
Regeneration gas
Pipeline quality gas
C4+
Figure 3 Processing line-up A
their limitations. Technical risk, licensor experience, degree of commercialisation, safety, health and environmental aspects all need to be weighed along with the process and economic performance of the technologies concerned. Proposed gas processing line-ups Natural gas processing is often considered a mature industry with little opportunity for improvements or innovations. However, changes in the product market continue to drive improvements in gas processing technology. Experience has proven that operating costs and investment constraints are becoming more important when selecting the proper gas processing solution for developing more unconventional and stranded gas reserves. In general, the following steps shall be followed for determining the proper technology line-up for a gas processing project:2 • Select the appropriate technology for each gas processing step
• Consider interactions between more often the technology selecdifferent gas processing units tion is driven by the requirement • Adjust the sequence of gas pro- to remove trace components such cessing units. as mercaptans (R-SH) and carThis section proposes three inte- bonyl sulphide (COS). Whereas gration schemes of the gas pro- deep removal of H2S and CO2 is cessing steps for a sour feed gas now mastered, mercaptan removal from a sour gas is still considered a challenge depending on the conChanges in the centrations involved. Different options for the combined removal product market of acid gas components and mercaptans have been used in various continue to drive gas processing plants. The optiimprovements in mum solution in many cases is the distribution of the mercaptans gas processing removal capabilities over the acid gas removal unit (AGRU), utilising technology a mixed chemical and physical solvent or a promoted methyl diethancontaining mercaptans. Typical olamine (MDEA) solvent, as well as schemes of each line-up have the molecular sieve unit (MSU).3 been illustrated in Figures 3 to 5. In line-up A, shown in Figure Note that for the selection of the 3 , the sour feed gas is sent to the acid gas removal technology to AGRU utilising promoted MDEA remove hydrogen sulphide (H2S) solvents in which all acid gas comand carbon dioxide (CO2), differ- ponents (H2S and CO2) and some ences are only marginal, where parts of the light mercaptans are
Off-gas to tail gas treating unit
Sulphur
Sulphur recovery
R-SH
Regen-gas treating (physical solvent)
H2S, CO2
Feed gas
Acid gas removal (promoted MDEA)
Gas deydration and hydrocarbon dewpoint controlling (propane refrigeration system)
Mercaptans removal (molecular sieve technology)
Pipeline quality gas
Regeneration gas
MEG injection
H2O
C4+
Figure 4 Processing line-up B
72 PTQ Q1 2017
www.eptq.com
Off-gas to tail gas treating unit
Sulphur
Sulphur recovery
R-SH
Regen-gas treating (physical solvent)
H2O H2S, CO 2
Feed gas
C4+ Gas dehydration and hydrocarbon dewpoint controlling (silica gel technology)
Acid gas removal (promoted MDEA)
Mercaptans removal (molecular sieve technology)
Pipeline quality gas
Regeneration gas
Regeneration gas
Figure 5 Processing line-up C
removed from the feed gas. The The Joule-Thomson (JT) expan- mercaptans are removed utilising sweet gas is then routed to the sion method would not be attrac- molecular sieve technology. gas dehydration and mercaptans tive unless the required outlet In line-up C, shown in Figure 5 , removal unit utilising molecular pressure is much lower than the silica gel technology allows the sinsieve technology, and nally to the inlet pressure or the inlet pressure gle step removal of both water and hydrocarbon dew point controlling is too high, having the operating heavy hydrocarbons from natural unit (DPCU) utilising a pro- range in the critical region. In the gas to meet the required pipeline pane refrigeration system, which case of using a turboexpander for quality gas specications. In this requires greater operator attention the hydrocarbon DPCU, the out- line-up, mercaptan removal would and maintenance than a silica gel let gas usually needs supplemen- be done with a molecular sieve system, but offers greater exibility. tal compression to fulll product system. In this case a side (slip) stream of gas pressure at the delivery point, In proposed processing linecold propane can be used to keep requiring high power demand, ups A to C, tail (off) gas from a the feed gas temperature of the though some of the power can Claus sulphur recovery unit (SRU), MSU at 40°C or below (as required be recovered from the expansion which invariably contains small by most molecular sieve vendors). process. quantities of sulphur compounds, Silica gel technology may be a In line-up B, shown in Figure 4 , shall be sent to a tail gas treating feasible and competitive alterna- the sweet gas from the AGRU is unit (TGTU) to remove the residual tive for some feed compositions rstly routed to the gas dehydra- sulphur species in order to meet (not rich in C3-C4 components) and tion and hydrocarbon DPCU using emissions regulations. In the past, operating conditions (temperatures a propane refrigeration system in the most common approach for a below 45°C and pressures above which monoethylene glycol (MEG) 99.8% plus sulphur recovery was 28.5 bara). Since total pressure drop is injected to prevent hydrate forma- a SRU followed by an amine-based across the silica gel bed is very low tion. Although methanol is a more TGTU.5 In this method, all sulphur (about 0.65 to 0.85 bar), this process effective hydrate inhibitor than compounds in the tail gas are condoes not require product gas com- MEG at low temperatures, MEG is verted to H2S by hydrogenation folpression. However, this process typically chosen since it is adequate lowed by H2S scrubbing by one of may require a pre-cooling system for dewpointing temperatures, safer the selective amine-type processes to achieve a maximum inlet gas to handle, and easier to regener- so that H2S-rich gas can be recytemperature of 45°C.4 ate than methanol. In this line-up, cled to the inlet of the Claus unit. To incinerator
CO2 /mercaptans
Lean acid gas from AGRU
AGEU
H2S-rich gas
Claus SRU
Sulphur
TGTU hydrogenating section
TGTU absorber
Rich amine solvent Lean amine solvent
Figure 6 Typical integration of tail gas treating and acid gas enrichment units
www.eptq.com
PTQ Q1 2017 73
SuperClaus process
Steam
Combustion chamber
Stack
Waste heat boiler
Incinerator
QC
S
S
S
S
Caustic scrubber
ABC FC
Caustic
Acid Air gas
Air
Figure 7 Jacobs’ Superclaus process followed by a caustic scrubber
Therefore, the only emission is Figure 7). A number of advantages from the CO2-rich vent gas. In this can be achieved by this combinacase, the TGTU can also be inte- tion: in the forefront are the lower grated with an upstream AGRU or installed cost and higher reliability. acid gas enrichment unit (AGEU) Investment costs are reduced, the to reduce emissions, simplify oper- process is less expensive to operate ations, and reduce capital cost.1 and to maintain, requires a smaller Figure 6 shows an integrated tail footprint and greatly simplies gas treating and acid gas enrich- overall operation. By combining ment scheme in which the AGEU the Superclaus and caustic scrubabsorber can be used to selectively ber technologies, the overall system absorb H2S from the lean acid gas can achieve greater than 99.9% sulfeed, producing an H2S-rich solvent and a CO2 overhead gas. Since the The best processing AGEU uses an aqueous solution of a selective amine, the CO2 reject line-up is dependent gas still contains some traces of H2S and most of the mercaptans as on the initial feed gas aqueous solvents have little afnity for these species. Upon passing conditions, the treated the reject gas over the hydrogenating section, most of the mercaptans gas specifications (more than 80%) will be converted to H2S, which could be subse- and environmental quently absorbed in an absorber requirements using the same solvent as in the AGEU.5 In view of cost, complexity and phur removal at compelling capreliability, it may be worthwhile ital and operating costs. At least to explore alternatives to the well- 99.0-99.4% of the H2S is captured known Claus sulphur recovery and recovered as elemental sulplus amine-based tail gas treat- phur by the Superclaus process and ing technology. One of the inter- the remaining sulphur is scrubbed esting options is the combination and converted to sodium sulphate of two well established processes; by the caustic scrubber. Because of Jacobs’ Superclaus process fol- the nature of the caustic scrubber lowed by a caustic scrubber (see process, the residual sulphur diox-
74 PTQ Q1 2017
ide content of less than 20 ppm in the ue gas is achievable without excessive cost. Line-up selection In selecting the best gas processing line-up that satises the business objectives, factors such as plant reliability, operational exibility, process guarantee, and life cycle costs must be considered. Design flexibilities and issues Determining the best gas processing line-up is very much dependent on the initial feed gas conditions, the treated gas specications and environmental requirements. As an example, if drying and mercaptan removal is necessary, drying options other than molecular sieve technology (such as glycol) likely require an additional processing unit for the removal of mercaptans. When the molecular sieve beds are regenerated, the water and mercaptans desorb into the regeneration gas, which can then be treated using a physical solvent process to remove the mercaptans and provide a concentrated mercaptan stream that can be sent to the SRU. Removing mercaptans in this way reduces the quantity of sulphur species sent to the hydrocarbon recovery unit and thus into the liquid products. This can result in a sig-
www.eptq.com
SEE US AT:
MIDDLE EAST SULPHUR
EGYPS
BOOTH 34 · ABU DHABI 12-16 FEBRUARY 2017
STAND 1E36, HALL1 · CAIRO 14-16 FEBRUARY 2017
A N S O S U S e 2 7 D L L m & i n 2 H I D P a r 8 I F H A d e A N I C U P t a R D i l s I A T R L L a t · A I N I O t i n T G N y u H E 2 r l . N co S 0 , 1 m G 7 / R s u E l p h E C u r E 2 0 1 7
THINK SULPHUR THINK SANDVIK We design, manufacture and commission equipment for every aspect of sulphur processing, from upstream handling and a range of granulation options to downstream conveying, storage and bulk loading. ■
Engineering & consulting
■
Efficient sulphur degassing
■
Premium Rotoform® pastillation
■
High capacity granulation
■
High performance remelting
■
Handling, storage and loading
■
Global service and spare parts
A WORLD LEADER IN SULPHUR PROCESSING AND HANDLING
Sandvik Process Systems, Division of Sandvik Materials Technology Deutschland GmbH, Salierstr. 35, 7073 6 Fellbach, Germany Tel: +49 711 5105-0 · Fax: +49 711 5105-152 ·
[email protected] · ww w.processsystems.sandvik.com
nicant reduction in hydrocarbon liquid treating equipment size, and Hydrocarbon Pipeline Acid gas Dehydration / Feed dewpoint quality can create new technology opporturemoval mercaptans removal gas controlling gas nities in this area (for instance, use molecular sieves instead of caustic to clean up the residual mercaptan). Figure 8 Feed gas pretreatment using silica gel technology7 Considering this, process line-up A, which utilises molecular sieve tech- • There is considerable loss of feed gas to remove heavy hydronology for simultaneous removal glycol caused by its solubility in carbons. The most commonly used of water and mercaptans from the hydrocarbon condensate. option for such a purpose is using feed gas, would result in an appro- • Rich MEG sent to the MEG silica gel technology, 7 where heavy priate design and would provide regeneration system will add to the hydrocarbons are usually removed to such an extent that the requireoperating exibility over the feed system cost. In line-up C, a slip stream of the ment for downstream hydrocarbon gas range. The volume of mercaptan-rich gas containing mercaptans (from dew point controlling is eliminated molecular sieve regeneration gas downstream of the silica gel unit) (see Figure 8). It should be stressed and therefore the physical solvent is used for regeneration of the silica that the treated gas hydrocarbon cost in line-up A is greater than gels that would reduce the lifetime dew point increase will become that in line-ups B and C. However, and performance of the system. more severe for sour feed gases line-up B has some challenges: 6 Also, the silica gel adsorption times with a high acid gas content, • The mixture of MEG and con- are usually in the range of doz- which also affects the silica gel densed water/hydrocarbon has ens of minutes up to 2-3 hours, performance. 7 high viscosity and requires heat- which on the other side reduces ing to ensure good separation. the expected life time of the adsor- Licensors’ preference Even then, uncertainty in composi- bent. The other disadvantage is the Leading technology licensors comtion and the extent of hydrocarbon high-pressure adsorbent vessels, monly take an integrated treatabsorption in the MEG may cause which can be expensive. ment scheme presented in line-up operational difculties. As mentioned above, one of the A for optimal design of gas pro• Glycol has some afnity for main parameters affecting selec- cessing plants with a sour feed hydrocarbons and co-adsorbs some tion of the processing line-up is gas containing mercaptans.8 For BTEX (benzene, toluene, ethylben- initial feed gas composition. For example, Figure 9 illustrates the zene, and xylene) compounds that example, if the feed gas contains overall arrangement of a typiwill end up in the regeneration a highly aromatic corrosion inhib- cal integrated treatment packvapour stream. itor or elemental sulphur solvents, age offered by Shell in which a • Operation is completely depend- these can cause heavy hydrocar- mixed physical/chemical process ent on the propane refrigeration. bon condensation and foaming (Sulnol) is used for the AGRU. When the propane system is una- problems in the AGRU that pre- This integrated treatment line-up vailable, there is no dehydration. vent the unit from performing can deliver enhanced exibility In fact, the purpose of the injected at its design capacity. This may and operability of the whole gas MEG is not to ‘dehydrate’ the impact the entire availability and processing plant. However, the gas but to prevent formation of capacity of the plant. In these best choice for hydrocarbon dew hydrates. cases, it is necessary to pretreat the point control needs to be selected
Treated gas
MSU
Sour feed gas
AGRU absorber
MSU regengas treating absorber
AGRU regenerator Mixed solvent
AGEU absorber
AGEU regenerator Aqueous amine solvent
Offgas
TGTU
Incinerator
SRU
Liquid sulphur
Flue gas
Figure 9 Overall arrangement of the Shell optimised solution with respect t o maximisation of treating train operability and flexibility9
76 PTQ Q1 2017
www.eptq.com
for each project in order to obtain the most cost-effective and t-forpurpose gas processing line-up. Conclusion
Comparison of the proposed alternatives for most sour gas processing requirements leads to the recommendation of the integrated treatment scheme presented in line-up A due to its practical and nancial benets. In fact, the treatment processes used in this scheme are based on leading technology licensors’ proprietary knowhow that take the feed stream and deliver the required end product in an optimal manner. However, there may be some cases where a techno-economic analysis of alternative schemes may prove to be superior after considering all technical and economic factors. References
Mokhatab S, Poe W A, Mak J Y, Handbook of Natural Gas Transmission & Processing, 3rd Edition, Gulf Professional Publishing, 1
Burlington, MA, USA, 2015. 2 Liebert R, Flexible Processing, LNG Industry , 39-43, May/Jun 2013. 3 Mokhatab S, Hawes P, Optimal Mercaptans Removal Solution in Gas Processing Plants, presented at the GPA Europe Spring Conference, Hamburg, Germany, 22-24 Apr 2015. M L, Di Campli Y, Adsorption 4 Zoccante Alternative for Lean Gases Dewpoint Control, presented at the 59th Laurance Reid Gas Conditioning Conference, Norman, OK, USA, 22-25 Feb 2009. 5 Bloemendal G, There Are More Roads That Lead to Rome, Which Is the Best Route for You?, presented at the 2013 Sour Oil & Gas Advanced Technology, Abu Dhabi, UAE, 24-28 Mar 2013. 6 Mokhatab S, An Optimum Line-up for Sour Gas Processing, PTQ, Gas, 19-24, 2010. 7 Zhang I, Graham C, Nielsen D, Schulte D, Debest M, Impact of Sulphur Solvent or Corrosion Inhibitor Solvent Injection on Treated Gas Hydrocarbon Dewpoints and Gas Treating Unit Performance, presented at the 87th Annual GPA Convention, Grapevine, TX, USA, 2-5 Mar 2008. 8 Mokhatab S, Meyer P, Selecting Best Technology Line-up for Designing Gas Processing Units, presented at the GPA Europe Sour Gas Processing Conference, Sitges, Barcelona, Spain, 13-15 May 2009.
9 Klinkenbijl
J, Grootjans H, Rajani J B, Best Practice for Deep Treating Sour Natural Gases (to LNG and GTL), presented at the GasTech 2005 Conference and Exhibition, Bilbao, Spain, 14-17 Mar 2005. Mokhatab is an internationally recognised gas processing consultant who has been actively involved in several large-scale gas field development projects, concentrating on design, pre-commissioning and start-up of processing plants. He has presented on gas processing technologies worldwide and has authored or co-authored nearly 250 technical publications. He has held technical advisory positions for leading professional journals, societies, and conferences in the field of gas processing, and has received a number of international awards in recognition of his work in the natural gas industry. Gerrit Bloemendal is a technology specialist with Jacobs Comprimo Sulfur Solutions in Leiden, the Netherlands, specialising in amine treating, sulphur recovery and tail gas treatment technology. He has published articles on low temperature TGTU catalyst and has given presentations on international seminars. He graduated from Utrecht University, the Netherlands, with a degree in chemistry, and fromTwente Technical University with a degree in chemical engineering. Saeid
s Valve Best 1867 since
www.eptq.com
PTQ Q1 2017 77
Support grid Scallops
Outer basket
Centerpipe
Outlet Collector
From design to manufacturing and field installation, Johnson Screens provides an extensive offering of products and services for the hydrocarbon processing industry. Johnson Screens offers a complete solution to your hydrocarbon processing industrial needs, including:
– Down Flow Reactor Internals – Radial Flow Reactor Internals – Installation of Internals
– Hydrocarbon Refining Reactor Internals – Petrochemical Processing Reactor Internals – Gas Processing Column Internals
Please visit our website to see the full line of Johnson Screens products, or nd a reactor internals specialist engineer at one of our global manufacturing plants to assist you in nding the proper solution for your application. We look forward to providing the quality engineered products you have come to expect when specifying Johnson Screens. North & South America Phone +1 651 636 3900 Fax +1 651 638 3281
[email protected]
Asia Pacific Phone +61 7 3867 5555 Fax +61 7 3265 2768 asiapacifi
[email protected]
www.tiny.cc/johnsonscreenshp
Europe, Middle East & Africa Phone +33 5 4902 1600 Fax +33 5 4902 1616
[email protected]
Correcting vacuum column design flaws Conventional design practices have been proven inadequate when maximising diesel recovery and gas oil cut point GARY MARTIN Recon Management Services, Inc.
T
o achieve a good return on investment (ROI) from a project requires developing minimum cost designs regardless of whether it is a grassroots or revamp project. However, low cost designs do not necessarily guarantee a good ROI. In recent years, a number of new fuels vacuum units have been designed to produce diesel off the top of the vacuum column (see Figure 1). The vacuum units are commonly designed for low pressure damp or wet operation to enable maximising HVGO cutpoint.1 These design conditions, in addition to higher LVGO pumparound duties associated with maximising diesel production, lead to large diameter LVGO pumparound sections. The operating conditions associated with the design make the possibility of excessive overhead slop oil production much more likely. Design practices of the past are inadequate and must be re-evaluated to obtain an acceptable design. This article is based on a revamp design to x the operation of a large scale Gulf Coast vacuum column with excessive liquid entrainment into the overhead system. The LVGO pumparound section of this vacuum column is 38ft (11.58m) in diameter and typically has a diesel product yield of 17 500 b/d (115.9 m3/hr). The vacuum column is designed for damp operation, meaning that it has a resid stripping section using stripping steam and the charge heaters utilise coil steam. Spray header distributors produce a range of droplet sizes and essentially all LVGO pumparounds designed with these distributors
www.eptq.com
Motive steam
Ejector sytem
Noncondensibles Sour water Slop oil
Vacuum column
LVGO pumparound
Diesel Reflux
Figure 1 Upper section of vacuum column
have some overhead entrainment. This unit typically operates at about 1700 b/d (11.26 m3/hr) of slop oil yield, with approximately 40% of this from equilibrium oil and the remainder from entrainment. While
Low cost designs do not necessarily guarantee a good return on investment this may seem high, in actuality it is worse. Due to problems with the overhead ejector system design, the column operating pressure is normally higher than its original design value. In addition, the operators limit the LVGO pumparound circulation rate to limit overhead entrainment. During operation with
increased LVGO pumparound circulation or during the winter when the conditions enable lower pressure operation, the vacuum column overhead slop oil make has been in the range of 2750 b/d (18.22 m3/hr) and at times higher. Due to economic yield incentives, the rener would prefer to operate at the lower column pressure and higher LVGO pumparound rate if these problems did not exist. At the lower column pressure and higher pumparound rate, the liquid entrainment into the overhead is in excess of 2000 b/d. Slop oil make is detrimental to the plant’s bottom line. At minimum, it adds to the operating costs to reprocess the oil and in the worst case it consumes plant capacity to process additional crude. To reduce the excessive slop oil make at lower pressure operation, the operators must adjust the heat balance on the column. It is modied to lower the LVGO pumparound heat duty. This
PTQ Q1 2017 79
ments or, as addressed in this article, that ignore the effects of liquid entrainment.
Figure 2 Spray header above a packed bed
reduces the required LVGO pumparound rate and consequently reduces the entrainment rate to a somewhat more acceptable overhead slop oil yield. This however signicantly reduces the diesel yield. LVGO pumparound The typical design of a vacuum column LVGO pumparound section includes a product/pumparound collector tray, a packed bed, and a spray header for pumparound return liquid distribution. A spray header has been the industry distributor of choice because it is relatively cheap and it provides for good heat transfer. All spray headers produce a range of liquid droplet sizes. Normally, the majority of droplets are large enough and the vapour velocity low enough that the entrainment is minimal. When using spray headers, a damp vacuum column design always makes the entrainment problem worse because of the corresponding higher top of the column vapour velocity. The problem becomes more of an issue with higher coil and stripping steam rates. Dry columns have very little vapour compared to damp columns when operating at the same column overhead temperature and pressure. Also, the more spray nozzles the higher the potential entrainment. This LVGO pumparound distributor has 61 nozzles while most vacuum units have 7 to 19.
80 PTQ Q1 2017
Spray header design Spray header distributors, as shown in Figure 2 , are good distributors for packed bed heat transfer service. This is due to the high surface area contact between liquid and vapour due to atomisation of the liquid. Heat transfer between the liquid and the vapour phase is occurring prior to entering the packed bed. However, in general, the liquid distribution to the bed is only roughly uniform and every nozzle has different distribution characteristics. While spray headers are relatively simple distributors, many engineered designs that have been produced are very poor. Common design mistakes include: layouts that produce excessive liquid overspray onto the vessel wall; too small an nozzle orice size that is not practical and leads to fouling; and designs that do not take into consideration turndown requireVacuum column overhead slop oil slop oil Vacuum column overhead Volume % 1 5 10 30 50 70 90 95 99
Table 1
D2887°F 93 219 268 378 483 558 617 638 687
Equilibrium and non-equilibrium oil in overheads High vacuum column overhead slop make can occur from operating problems in the vacuum column or from light material carried over from the atmospheric crude column. A case study of the latter is provided in the literature. 2 Table 1 lists the vacuum column overhead slop oil make distillation curve determined by the plant laboratory during one set of operating conditions. The column top pressure and temperature measured were 38 mmHg and 125°F (51.7°C) respectively. Proper evaluation and xing of the problem requires good test run data.3 Using a full set of test run data to model the vacuum unit yields a much lighter overhead slop oil composition based on equilibrium calculations at the measured operating conditions. The measured overhead slop oil make on this day of operation was 1726 b/d (11.43 m3/hr) with a 38.1 API gravity. To match the model results with actual operation, 978 b/d (6.48 m3/hr) of entrained LVGO pumparound liquid had to be added into the column overheads to match the material balance and slop oil composition. Equilibrium oil in the column overhead vapour is a function of the column overhead operating pressure and temperature. Non-equilibrium oil is heavier oil that should not be in the vapour phase at these operating conditions. In this case, the heavier oil is from the smaller atomised oil droplets formed from the pumparound distributor spray nozzles that are entrained by the rising vapours. The vacuum column also has temperature indicators measuring the vapour temperature in the top head of the vacuum column and in the column overhead vapour line. There was a temperature drop of 21°F (11.7°C), which is further indication of entrainment. As the rising vapours from the LVGO pumparound bed exit the column, they are mixed with the cold LVGO
www.eptq.com
pumparound liquid that is entrained into the overheads, resulting in a reduction of the measured vapour temperature in the overhead vapour line.
found throughout the literature.5 Terminal velocity can be calculated using the equation: Terminal velocity =
[ ] 2 − Mist eliminators In some applications, mist eliminators located at the top of the where column can be used to remove or g = gravitational acceleration minimise the entrained liquid. ρp = particle density However, years of bad experiences ρ = actual vapour density by the industry have proved that m = mass of particle mesh pads should not be used for Ap = projected area of particle this application. Experience has Cd = drag coefcient proved that they will foul, leading to high pressure drop and a correIt is understood that liquid distrisponding decrease in gas oil yields. bution within a spray cone varies for a given nozzle and further variPhase Doppler interferometry ation is found from one nozzle type To properly design a spray header to the next. It is also known that as requires knowing the estimated the nozzle pressure drop changes entrainment rate for the selected from one ow rate to the next the nozzles. The tendency for the atom- spray cone diameter varies, as well ised drops to be entrained is a as the liquid particle sizes. function of the drop size distribu- However, reliably determining tion, the vapour velocity of the drop size distribution was not up-owing vapours, and the available until testing with vapour and liquid physical proper- phase Doppler interferometry. ties. There are three forces that act Fractionation Research Inc. in on the liquid droplets from the Stillwater, Oklahoma has recently spray nozzles that determine if tested this method.6 Their test entrainment of liquid will occur. results also showed that an increase The forces are gravitation, the in liquid rate results in a decrease buoyant force which acts in the of the entrained droplet Sauter opposite direction from the gravita- mean diameter and a correspondtional force, and a drag force due to ing increase in total liquid the relative motion between the entrained. The tests also showed particles and the rising vapours. that increasing the gas rate The drag force acts to oppose the decreases the entrained droplet motion of the liquid droplets in Sauter mean diameter and has a the opposite direction of their corresponding increase in the total liquid entrained. movement. In a revamp case such as this, Using phase Doppler interferomoperating data to determine the etry to determine the drop size overhead vapour rate and composi- distribution along with determining tion is available from plant meters the vapour velocity of the up-owand lab data. However, if inade- ing vapours and the vapour and quate meters are available or if this liquid physical properties, the is a new design then air leakage, designer can evaluate the entraincracked gas, and so on must be ment rate expected for a selected estimated. Guidelines regarding nozzle. Different nozzles provide estimates for these values are avail- varied liquid distribution and drop able in the literature.4 size distribution. In fact some The quantity of entrainment can nozzles are designed to produce be determined by the particle very small droplets that if used in terminal velocity, obtained using this application would be even Stokes’ Law, and knowing the drop more detrimental to operation. This size distribution for the selected method can be used to predict the nozzle. Stokes’ Law can be used at entrainment rate and to effectively low Reynolds numbers and can be select optimum nozzles. !
www.eptq.com
÷
!
!
!
Conclusion Maximising production of diesel from the top of a vacuum column combined with a damp vacuum column operation to maximise gas oil recovery makes old design practices for the column internals inadequate. While properly designed trough distributors can x this problem, they do not have as good heat transfer efciency as spray headers. Vessel height limitations to increase the packed bed height may not enable achieving the required pumparound heat removal and thus not enable using a trough distributor. Using spray nozzle phase Doppler interferometry test data can be used to appropriately select the optimum LVGO pumparound spray nozzle for the specic operating conditions to minimise overhead oil entrainment. This approach can also be used for other column designs.
References 1 Martin G R, Vacuum Unit Design Effect on Operating Variables, PTQ, Summer 2002, 8591. 2 Martin G R, Lines J R, Golden S W, Vacuum System, Fundamentals, Encyclopedia of Chemical Processing and Design, Vol. 61, McKetta J J, Weismantel G E, editors, Marcel Dekker, Inc., New York, NY, 1997, 5-18. 3 Martin G R, Cheatham B E, Keeping Down the Cost of Revamp Investment, PTQ, Summer 1999, 99-107. 4 Martin G R, Nigg J M, Vacuum Pressure Control: Impact on Profitability, PTQ, Summer 2001, 73-81. 5 McCabe W L, Smith J C, Unit Operations of Chemical Engineering, McGraw-Hill Book Co., New York, 1976. 6 King B, et al, Measuring Entrainment from a Spray for Industrial Applications, AIChE Annual Meeting, Salt Lake City, 11 Nov 2015, Paper 155B.
Gary R Martin is a Senior Process Specialist with ReCon Management Services, Inc. He has been involved in revamps, including conceptual process design and process design packages for large capital revamps, optimisation, and troubleshooting services to the refining industry worldwide. He previously was a Principal of Process Consulting Services Inc. and also worked for Total Petroleum, El Paso Refining, and Glitsch Inc and holds a BS degree in chemical engineering from Oklahoma State University. Email:
[email protected]
PTQ Q1 2017 81
ITW Innovative Technologies Worldwide
CHANGE YOUR MIND ON FOULING REMOVAL
ARE THESE IMAGES FAMILIAR TO YOU ? FORGET ABOUT THEM ! Patented ITW Online Cleaning can remove any type of fouling from equipment, including the polymers, without the need to open or enter hazardous process equipment. An entire Process Unit can be cleaned by utilizing ITW Online Cleaning in as little as 24 hours on an oil-to-oil basis. Online Cleaning can be applied at any time during the run of the Unit, in order to solve the problems when they start appearing, rather than when they are no longer sustainable. This will in turn avoid throughput reduction, giveaway and energy loss, associated with fouling. The application of ITW Online Cleaning will be therefore driven by performance recovery and Opex improvement, rather than the economics for placing a turnaround.
ITW Online Cleaning can be applied to all Refinery/Petrochemical/Gas Field/Oil Field production Units, as well as storage tanks, and is the onl y technology which has an immediate Return On Investment (ROI). Regular application of ITW Online Cleaning will target an increased run length under
clean conditions,
with a far higher ROI.
For Turnaround applications, ITW Online Cleaning can eliminate/dramatically reduce the need for mechanical cleaning, thereby reducing downtime and improve operational HS&E. In a Turnaround application, an additional ROI increase can be targeted by applying ITW Improved Degassing/ Decontamination to achieve quick and effective safe entry conditions. Our patented chemistry does not create any emulsion and fluids can be easily handled by Waste Water Treatment Plant.
EVALUATE ITW ONLINE CLEANING AND ITW IMPROVED DEGASSING/DECONTAMINATION TODAY TO IMPROVE YOUR PLANT’S PROFITABILITY ! For more informations contact :
ITW S.r.l.- C.da S.Cusumano - 96011 Augusta - Italy Tel. +39 (0931) 766011 E-mail:
[email protected] www.itwtechnologies.com
Join ITW Team worldwide and send your Curriculum Vitae to :
[email protected]
Now hiring Professionals Worldwide
Analysing FCC hot spots Finite element analysis can be employed to improve the safety and quality of old or new designs and troubleshoot problems in the field KENNETH FEWEL Technip Stone & Webster Process Technology Technology
H
ot spots on FCC units are a common occurrence but relatively little is known about the implications for long term operation. Standard procedures in a FCC plant are to monitor on a scheduled basis the thermal scans of their equipment and use steam to cool if hot spots exceed a certain temperature threshold. Although this has served operators in a practical manner for years, there are larger safety issues that should not be ignored. ASME has developed standards for analysing thermal stresses that are suitable for design, maintenance and operation. Operators should be aware of these methods and take advantage of them for their hot spot maintenance and procedures. Finite element analysis (FEA) is a method that enables accurate stateof-the-art analysis of stress and strain in all types of solid materials for all types of loadings, including thermal. Thermal stresses are usually calculated for a design that is working perfectly to insulate the steel shell from high temperatures inside the FCC vessel. FEA has been used to validate traditional calculation and more accurately defne the state of temperature, stress and strain that will occur under normal operation in FCC units made from composite, refractory-lined steel, materials. Recently, as a result of hot spots appearing on FCC equipment, there was a need to analyse and determine the severity of stresses and strains under hot spot operation. This has provided an excellent opportunity to use FEA to explore these situations using quantitative analysis and make recommenda-
www.eptq.com
tions to operators about the safety of their particular situation. Before further explaining FEA, it is important to understand why hot spots occur and how stress and strain occurs. Hot spot causes Hot spots are caused by several mechanisms of operation that come together like a ‘perfect storm’. They do not appear randomly as there are reasons why hot spots form in FCC equipment.
Hot spots are caused by several mechanisms of operation that come together like a ‘perfect storm’ 1. Erosion by by catalyst Catalyst is, by nature, irregular in shape and very hard compared to carbon or stainless steels. The refractory used to line steel vessels or piping is designed to provide insulation and strength for stresses and thermal strains; it is not designed purely for erosion resistance. Therefore, the material is susceptible to erosion by catalyst. Fortunately, this insulating refractory is thick, usually 4-5in in piping and vessels. However, catalyst can quickly erode through this thickness with the help of the next mechanism. 2. Thermal stress and strain a. Wall stress: thermal wall stress can easily reach levels that will
crack the refractory. In many designs some cracking cannot be avoided. Longitudinal cracks in the refractory outer diameter of lining will form as soon as temperatures reach operational levels. This is due to thermal wall stresses in the refractory. These stresses have been determined from research experiments by Wygant and Crowley1 among others.2 The inner diameter of the lining is very hot, particularly in a regenerator, typically 1300°F (700°C). The inner refractory will go into compression when internal temperatures start to rise. The outer surface of the lining may not expand as much as the steel. Thus the lining will go into tension stress and the refractory may not bear upon the inner diameter of the steel piping. If so, thermal wall stress can cause the lining in the outer edges to crack. b. Thermal bending stress: thermal expansion forces in the piping and vessel can crack the refractory in bending, often transverse to the longitudinal direction. In addition, compression bending stress can open gaps between the steel and lining, leading to large gaps through which vapour and catalyst can easily circulate. These are the most damaging stresses to the refractory that can occur. The resulting hot spots will exacerbate the problem, making the thermal stress and gap worse. 3. Worm holes As a direct result of wall stress cracks in the refractory lining and gaps opened by thermal bending of the refractory on the outside of the lining, there is a strong possibility
PTQ Q1 2017 83
t n e l a v i u q e i t s o p p , s s t s o e r h t s n i e g n a h C
50000 y = 73.789x
45000
R2 = 0.6434
40000 35000 30000 25000 20000 15000 10000 5000 0 0
50
100
150
200
250
300
350
400
450
Change in temperature of the hot spot, ºF
Figure 1 Rise in equivalent stress as a function of rise in hot spot temperature
500
by the geometrical constraints constraints of the adjacent cooler pipe and, to a lesser extent, refractory anchors. But the gap between the steel and refractory is larger now and the circulation of catalyst is naturally increased. The hot spot temperature is increased and the thermal stress becomes higher. This instability will not stop until one of the following mechanisms occurs: 1. The catalyst streaming in builds up to ll the void left by the steel strain and the gas bypass ow is reduced or eliminated. 2. The strain hardening in the steel, already in the plastic region, reaches a point of equilibrium before the rupture point is reached. This is usually assisted by a geometric constraint and/or steaming. 3. The steel reaches rupture point. Catalyst and vapours begin to spew from the opening. Obviously every operator hopes for number 1 or 2 to occur. Despite denials, number 3 has been known to happen. Based on the Goodier3 equation for thin circular plates unconstrained at the edges, the following relationship can roughly estimate the thermal stress rise due to additional temperature rise in the elastic region:
that the hot vapours and catalyst reviewed either, but done only will nd alternate paths through the periodically. How can an operator piping. The wall stress cracks are know that his plant is safe for the probably not large enough for cat- next day or week, or until the next alyst to circulate through, and they scheduled shutdown? may ll with catalyst instead. A These questions should not be larger problem is the thermal bend- answered by rough estimates or ing strain in the refractory. Thermal guesses. Hot spots, once formed, bending stress and strain will open cause additional thermal stress. existing cracks or form new cracks When a hot spot appears, the situasitua large enough for catalyst to circu- tion will worsen and thermal stress late through. Worm holes can then and strain can quickly become form, leading to catalyst circulat- unstable. ing past the steel, overheating, and rapid erosion of the steel skin from When a hot spot the inside out. dσ dσ d τ 0.5⋅ E⋅ α ⋅ dT The most apparent result of a r t rt appears, the situation combination of these mechanisms is a hot spot. The steel will often reach where: will worsen and σ = radial stress in the hot spot temperatures in excess of piping r σ = tangential stress in the hot spot or vessel design values, typically thermal stress and t τ = shear at the edge of the hot spot around 650°F (340°C). When the rt dT = temperature change skin reaches temperatures higher strain can quickly E = elastic modulus of the material than design, the situation requires α = coefcient of thermal expansion cooling steam to keep the temper- become unstable atures under control. This is standOf course this must be added to ard maintenance procedure, but the normal operating weight, preswhat are the resulting safety and How hot spots cause thermal stress sure and thermal stress state which long term implications? When is a Take, for example, a thin plate held existed before the hot spot formed. shutdown necessary? If a shutdown rigidly at the edges and heated with The sample eld data employed in congis planned, what needs to be done? a torch at the centre of the plate. Figure 1 show that geometric congThe material in the centre will begin uration and pre-stress matter greatly Is the steel and/or lining matemate rial ruined or can it be reused after to expand and thermal stress will because there is much scatter in the shutdown? occur. The edges, if truly rigid, will equivalent stress data. Where preUsually, the answers are not so prevent expansion from occurring stress is largely compressive, hot clear to the operator as a potentially in the plane of the plate. Once the spot stress will simply add more dangerous situation arises. What if stresses exceed the proportional compressive stress to the negative the steam pressure is lost temporar- limit, the plate will begin to warp components of pre-stress. Where hap- pre-stress is mostly tensile, the hot ily? What if the temperature rises? out of plane. Imagine that this hapten Continuous monitoring of the steam pens on the surface of a large cata- spot will lower or reverse the tenlyst transfer pipe. The steel must sile components and raise the comcooling system is rarely done. The but will be resisted pressive components. In addition, thermal scans are not continuously bulge outwardly but
84 PTQ Q1 2017
www.eptq.com
www.cat-tech.com
MAKING YOUR PLANT WORK...
TOWER FIELD SERVICES Cat Tech Tower Field Services offers bespoke mechanical services to oil refinery and protochemical
SMARTER We thrive on exceeding our clients expectations and understand the need for upfront communication communication in relation to timescales and costs. No-one likes surprises in this industry and we ensure that everything is done, from pre-job planning, on site activities to post job reporting in a mannerwhich allows our customers to concentrate on their core business.
plants all around the world.
We work hard to maintain an industry reputation for safety, speed and dependabili dependability. ty.
Our team has vast experience in overseeing and
Our in house training allows us to staff a project with to produced customer satifaction. satifaction.
undertaking the installation of all types of Tower internals.
the relevant
skilled personnel
Our tower ower services services devision has deep roots, with
We would value the opportunity to opportunity to demonstrat demonstrate e the benefits benefits of utilising Cat Tech for all of your tower ower services services..
several of the senior team at Cat Tech being
The bene benefits of engaging Cat Tech are: are:
involved involve d in the founding of the Cana-Tex organisation. Since 1971, Cana-Tex had been the premier global Tower speciality contractor; performing Tower maintenance and plant turnaround projects around the world. These origins, with the addition of new team members, make Cat Tech Tower Field services a global operation which is both cost effective and technically experienced.
Our well established sales team have a wealth of experience. We take pride on ensuring all enquires are dealt with quickly and efficiently.
Experienced Tower specialists. Proven track record. Ability to supply spares and replacements. Spade to Spade service. Revamps or modifications to Towers owers / / Internals. Internals. Inspection of Tower internals. Supervision of external/client labour force. Detailed inspection reports. Cleaning of tower internals. Reactor tray field services. Turnkey reactor tray field services.
UK: Cat Tech (Europe) Ltd., 1 South Park Road, South Park Industrial Estate, Scunthorpe, North Lincolnshire, DN17 2BY UK Phone: +44 (0) 1724 871747 Fax: + 44 (0) 1724 861928 Bulgaria: Cat Tech Services (South Eastern Europe) Ltd., Devnia, Varna District, Povelyanovo, 20 Droujba Str, R Bulgaria, 9160 Phone: +(35) 9519 92052 Fax: +(35) 9527 46770
China: Cat Tech Services (Shanghai) Ltd., Room 1230 No.1, 58 Nong Huachi Road, Putuo District, Shanghai 20061 CN Phone: +(86) 2152 9175 08 Fax: +(86) 2152 9175 05 Singapore: Cat Tech Asia Pacific Pte. Ltd., 17 Fan Yoong Road, Singapore 629794 Phone: +(65) 6264 6261 Fax: +(65) 6264 6401 South America: Please contact the UK office on +44 1724 871 747
Thailand: Cat Tech Tech Services (Thailand) Ltd., 2/1 Soi Banbon Nuenpayom Road, Maptaphut Sub District, Muang Rayong District, Rayong Province 21150 TH Phone: +(66) 894 064 100 Fax: +(66) 638 692 382
Lining
Temperature, ºF 1423.10 (max) 1294.20 1165.40 1036.50 907.68 778.84 650.00 500.00 450.00 400.00 352.97 305.94 258.91 211.88 (min)
200ºC skin Steel pipe 0
2.5 5.0 7.5 10.0
1382.00
750ºC Typical hot spot
Figure 3 FEA sub-model of steel plate, refractory lining and stainless anchors
788.00 576.60 716.64 467.83 435.06
0 2.5 5.0 7.5 10.0 inches
Figure 2 Thermal gradients in typical lined FCC pipe modelled with heat transfer through both materials using FEA
geometric conditions such as crev- of thermal scans, the hot spot can ices, bending stress and other stress be replicated in a model. By doing risers will play an important part. so, the state of stress caused by the Finally, vibration will also increase hot spot can be determined with a the general state of stress with high degree of accuracy. These therreversing components that will com- mal stresses can be compared with plicate the situation. Vibration stress ASME allowable to determine if is almost always present in an FCC continued operation is safe. Figure 2 and is largely disregarded, despite illustrates a FCC lined pipe hot spot the fact that it can be significant. model and the associated thermal Combining all of these factors state. makes a traditional analysis very ASME Section VIII, Division 2, challenging. It can be done, but has developed an analysis method the results will be highly uncer- for determining the safety of steel tain, could be inaccurate, and likely thermal stress.4 The ASME piping code B31.35 can be used also, but misleading. Methods are available to deter- this will result in more conservatism mine the complete state of stress as the piping code is based on tradiand to mitigate the causes. This is tional calculations. Both have been where the application of FEA mod- used for most studies, as a check elling can help. against one another. The results were similar although stresses The FEA modelling approach to lined checked against B31.3 had a slightly pipe hot spot analysis lower margin of safety. An FEA model can analyse hot Refractory stress can be checked spots in detail. With the assistance also. Although allowable material
limits are not available in ASME, the manufacturer’s strength properties can be used with a safety factor applied. Compared to steel, refractory strength is very low. Although compressive strength has been stated as high as 5000 psi, tensile or rupture stress strength, determined from bending specimens, is generally less than 1500 psi, even though stainless needles are used in many formulations to improve tensile strength. Fortunately, the elastic modulus is generally less than steel and therefore flexibility is greater. In addition, the flexibility of the anchors is fairly high and this isolates some of the steel strain from the refractory strain and vice versa (see Figures 3 and 4). Despite the use of anchors and needles, refractory is very
662.00 482.00
755.60 Temperature, ºF 1423.1 (max) 1036.5 650.00
788.00 698.00
inches 0.51241×10−3 (max) 0.45547×10−3 0.39854×10−3 0.34160×10−3 0.28463×10−3 0.22774×10−3 0.17080×10−3 0.11387×10−3 5.6934e−5 0 (min) 0
2.5
5.0
377.12
Max 899.60
258.91 211.88 (min)
791.60 Min
7.5 10.0
inches
Figure 4 ANSYS FEA sub-model of steel and lining pulling apart. Colours show deformation with highest shown in red and lowest in blue. Refractory has been cut away to show the anchors clearly
86 PTQ Q1 2017
500.00 450.00 400.00 352.97 305.94
0
225
450
675
900
inches
Figure 5 Field case situation: hot spot temperatures found from thermal scans, steam on
www.eptq.com
susceptible to cracking under operating conditions due to thermal stress. If the design is not analysed carefully with the use of FEA, areas of critical bending stress in the refractory are usually missed. This is because most piping is analysed with a one-dimensional piping stress program. These programs do not provide for placing discrete refractory material in the piping model. Although the added strength of the refractory can be factored in, stress in the refractory material often cannot be determined. If the refractory reaches the bending rupture point, the contribution to overall strength is lost. Yet the one-dimensional approach would not take this into account. Therefore, this approach could be non-conservative if the refractory strength is factored in and the refractory is actually breaking. FEA is the most advanced means to analyse the refractory-lined pipe in FCC systems. No other method can supply the detailed stress state or the design quality for both steel and refractory lining in one model. Case history: FCC hot spot FEA A major European operator was experiencing an increasing number
0
10
20
30
40
inches
Temperature, ºF 1425.5 (max) 1036.3 650.00 500.00 450.00 400.00 352.97 305.94
662
258.90 211.87 (min)
Figure 6 FEA replicated hot spot on regenerated catalyst withdrawal well nozzle, top side. High stress existed here because of normal operating thermal tension load in the regenerated catalyst standpipe
of hot spots on transfer lines and nozzle connections. The plant had vessels bulging, pipes rupturing, and hot spots occurring across the piping system’s slide valve, cyclone dipleg nozzle and other transfer lines. The long standpipe and riser transfer piping had no expansion joint so thermal stress and strain were suspected. A transfer line model was set up to determine the complete state of stress in the affected piping. Figure 5 shows hot spot tagged by temperature with existing steam cooling. In addition to a model of normal
operation under weight, pressure and thermal loads, a second model was set up and the hot spots were added. Figures 6 to 8 show the temperature contours of model setup and resulting stresses and deformations for typical parts of a catalyst transfer system. The results of the study were very revealing. Figure 9 shows computed displacements later verified by the operator. Figures 10 to 11 illustrate the stresses in the reactor cyclone
47439
899.6
42300
Temperature, ºF 1423.10 (max)
Stress, psi 55852.00 49653.00 43454.00
791.6
1036.50 650.00
50
100
150
200
inches
Figure 7 Standpipe to riser wye piece; hot spots appear in high bending stress areas
www.eptq.com
27937
Max
6261.20 100.47 (min)
258.91 211.88 (min)
0
3.3257 2.7714 2.2172 1.6629 1.1086
53780
37255.00 31057.00 24858.00 18659.00 12460.00
500.00 450.00 400.00 352.97 305.94
0
Deformation, in 4.9886 (max) 4.4343 3.8800
25
0
225
450
675
0.55429 0 (min) 900
inches 50
75 100
inches
Figure 8 Stresses found with FEA in the wye, probably due to bending failure in the refractory
Figure 9 As a benefit of an accurate thermal strain state, FEA will yield very accurate thermal displacements. The operator measured a 4in lateral movement with hot spots, matching the FEA closely
PTQ Q1 2017 87
Maximum principal stress, psi 1.0363e5 (max) 7000.0 6222.2 5444.4 4666.7 3888.9 3111.1 2333.3 1555.6 777.78 0 −500 −1000 −50397 (min)
Stress, psi 1.2519e5 (max) 55000 48127 41255 24382 27509 20636 13764 6891 18.237 (min) 54654
0
5
10
15
20
inches 0
125
250
375
500
inches
Figure 11 Dipleg refractory lining stress contours at the hot spot indicating ultimate
failure of refractory throughout thickness
Figure 10 Skin stress on an external FCC
reactor with hot spot around the nozzle connection
dipleg skin and lining. Figure 12 summarises safety factors. All of the thermal hot spot stresses could pass ASME Section VIII, Div 2, part 5 rules. The operator decided to continue running the plant to the next scheduled shutdown with confidence. Pre-stress and hot spots: causal relationships In almost every case, high levels of operating stress and strain coincide with hot spot locations. The hot spot temperature can be correlated by the operating stress and strain at the hot spot. Figure 13 illustrates this with a graph of hot spot temperature vs steel operating stress. Although the data are scattered, the relationship between pre-stress and hot spots is clear. Faulty thermal design causes high operating stress and this leads to cracking of the refractory, perhaps during the first start-up. From there, it may be only a matter of time before the hot spots appear. Use of FEA to analyse FCC designs The detailed design process for FCC equipment should rely on an FEA stress model to determine the complete thermal stress state before an FCC is built. FEA will reveal where thermal stress and strain is high, allowing custom changes to the
88 PTQ Q1 2017
design and possibly preventing hot spots. This premise is based on experience with hot spots in the field. Most, if not all, cases are associated with thermal stress and strain, the key causal elements in hot spot formation. If these are eliminated, erosion is the likely reason for a hot spot to form and that could take many years to happen. Once a hot spot has appeared, an FEA analysis can determine if the stresses are high enough to cause alarm. Variables include temper-
ature, size, pre-stress, position, geometric factors and material properties. It is difficult to know with certainty if a hot spot is a safety issue without a detailed FEA analysis. The analysis can take all of these factors into consideration: anchors, refractory strengths, refractory expansion coefficients, piping material, weight, pressure and thermal stress. Even vibration and fatigue can create damaging refractory stress and strain to occur, leading
100 DPT normal FOS DPT ASME FOS DPT B3.13 FOS
y t e f 10 a S f o r o t c 1 a F
0.1
i p e p i p e c t i o n p º º 4 5 4 5 Y j u n D W D W D W W f o f W W o p T o t t o m B o
A 4
A 6
A 7 a l v e t i o n e v j u n c d i S l y e W
e r l e g R i s d i p e l o n c C y
Figure 12 Factor of safety (FOS) comparison for dead weight, pressure and thermal
(DPT) case at new condition vs hot spot operation. A factor of safety of 1 or greater is a passing grade, indicating that computed stress is lower than allowable stress. Normal FOS indicates as designed. For hot spot cases, ASME indicates Sec 8, Div.2 evaluation and B31.3 is a piping evaluation
www.eptq.com
to hot spots and the possibility of rupture. This can be analysed with a dynamic stress FEA model of the piping or vessel. Conclusion FEA is a maturing technology used more frequently as the tools become more widespread, user friendly and real world effective. With proper setup, verication of computations and experience to evaluate the results, this engineering tool can be a means to improve the quality of old or new designs and trou bleshoot problems in the eld with the vision like no other method. Do not discount FEA as unnecessary, too complicated or expensive as the investment returns can greatly advance engineering and product quality. References 1 Wygant J F, Crowley, American Ceramic Society Bulletin, Vol. 43, No. 3, 1964. 2 Buyukozturk, Oral, Tseng, Journal of the American Ceramic Society, Vol. 65, No. 6, Jun 1982.
1000
F º 900 , e r u t 800 a r e p m700 e t t o 600 p s t o 500 H
y = 0.0128x + 670.72 R2 = 0.7968 Hot spot temperature Linear hot spot temperature
400 0
5000
10000
15000
20000
Operating thermal stress as designed, psi
Figure 13 Correlation of design stress levels and eventual hot spot temperatures experienced in those stress locations in an FCC catalyst transfer line. This is one indication that higher operating stress leads to hot spots 3 Goodier J N, Thermal Stress, ASME J ofApplied Mechanics, Vol 4, No 1, 1937. 4 ASME Section VIII, Division 2, Part 5: Design by Analysis, 2007. 5 ASME Process Piping, B31.3, 2006.
Kenneth Fewel is a Senior Supervising Engineer with Technip Stone & Webster
Process Technology. With 40 years of industry engineering experience, he holds a bachelor’s degree in mechanical engineering from Southern Methodist University and a master’s from the University of Texas. PE licensed, he holds seven US patents, received numerous professional awards, and is a Fellow of the American Society of Mechanical Engineers. Email:
[email protected]
A New Way bolted bulk material silos for all kinds of bulk solids simple and quick installation short delivery times comes with complete finish resistant to explosion and pressure shock SILOBAU THORWESTEN GmbH, D-59269 Beckum (Germany), www.thorwesten.com
www.eptq.com
PTQ Q1 2017 89
Torch Oil Injector
SUPERIOR SPRAY.
SERIOUS RESU LTS.
Whether you need to cool gas, dissolve salts in an overhead line or inject chemicals to prevent corrosion, we can help optimize injector performance. Here's how:
• Assistance with nozzle selection, spray direction and injector placement. There are dozens of factors to consider before choosing a spray nozzle, determining whether to spray co- or counter-current and identifying the proper placement of an injector in a vessel. We can help you evaluate your process conditions and then design an injector to provide optimal performance • Design validation using Computational Fluid Dynamics (CFD) and Fluid Structure Interaction (FSI). We use powerful modeling tools to simulate your environment, confirm the injector will provide the expected spray performance and withstand process conditions such as thermal stresses, heat transfer, vortex shedding and more
• Proven track record. Companies like Technip, Mustang Engineering, Bechtel, Shell and many others rely on us to manufacture B31.1 and B31.3 code-compliant injectors and conduct radiographic, hydrostatic, ferrite tests and more
Learn More. Call 1.800.95.SPRAY or visit spray.com/injectors
CFD MODEL ILLUSTRATES SPRAY
NEW STEAMMAX™ NOZZLES
SLURRY RECYCLE INJECTOR.
CHARACTERISTICS BASED ON INJECTOR
USE PLANT STEAM TO
DOZENS OF OTHER TYPES
PLACEMENT IN PIPES & VESSELS
ATOMIZE LIQUID INSTEAD
ALSO AVAILABLE
OF COMPRESSED AIR
UNMATCHED GLOBAL ENGINEERING, MANUFACTURING & TECHNICAL SUPPORT NOZZLES
|
CONTROL SYSTEMS
|
HEADERS & INJECTORS
|
RESEARCH & TESTING
Balanced distillation equipment design Fouling resistance and efficiency requirements for distillation equipment are balanced and optimised for reliable unit performance SOUN HO LEE GTC Technology
F
ouling tendency is a critical issue in crude distillation units and should not be overlooked when designing crude distillation columns. Corrosion tendency can inuence fouling issues as well. Since fouling resistance has an inverse relationship to efciency in distillation equipment design, optimising equipment design between fouling resistance and efciency requirements must be precise. Poor
application know-how as well as poor equipment design often downgrade column performance and reduce unit run length. This article will discuss common fouling issues associated with crude distillation column design. Actual retrots for crude atmospheric columns are demonstrated through two case studies. These studies examine how fouling resistance and efciency requirements for distilla-
tion equipment are balanced and optimised through careful evaluation and design methodologies. Case study 1: crude distillation unit description and background The conguration of the crude distillation unit in this case is illustrated in Figure 1. Fractionated light and heavy kerosene streams through the crude atmospheric column and the side strippers are
Off gas
Crude atmospheric column
Unstabilised naphtha
Steam Light kerosene
Heavy kerosene pumparound
Kerosene
Steam Heavy kerosene
Light diesel pumparound Steam
Preflash drum Desalted crude
Light diesel
Heavy diesel pumparound Diesel Steam Steam
Heavy diesel
Atmospheric residue
Figure 1 Case study 1: crude distillation unit configuration
www.eptq.com
PTQ Q1 2017 91
fractionation sections. Kerosene or diesel intermediate product yield limitation was also experienced when kerosene or diesel boiling range material composition was increased in the charged crude slate. Charge crude compositions were frequently varied during operation.
Figure 2 Case study 1: valve/perforation hole wearing and corrosion
Figure 3 Case study 1: underside view of fouled trays
combined and rundown as a single kerosene intermediate product stream. A diesel intermediate product stream is also formed from a combination of light diesel and heavy diesel streams. These crude atmospheric columns and side strippers were originally designed with conventional movable valve trays, traditionally selected in the past. The exception was the wash section which was arranged with structured packing. Three pumparound circuits are arranged at the heavy kerosene, light diesel and heavy diesel range material locations. The naphtha/kerosene fractionation section is positioned as the crude atmospheric column top section. This column was designed without a top pumparound circuit in order to maximise fractionation between unstabilised naphtha and kerosene at a given column height.1
92 PTQ Q1 2017
This crude distillation unit faced two problems: fouling and corrosion of the distillation equipment in the unit were found during a turnaround inspection. Valve perforation hole wearing and corrosion were found in the trays for naphtha/
Fouling tendency is a critical issue in crude distillation units and should not be overlooked when designing crude distillation columns kerosene fractionation. Tray fouling was also identied in the trays for the light kerosene/heavy kerosene and heavy kerosene/light diesel
Case study 1: root cause identification Figure 2 shows that the naphtha/ kerosene fractionation trays suffered from valve/perforation hole wearing and corrosion. Some movable valve units dislodged from the tray deck. Perforation hole sizes on the tray deck were increased by wearing and corrosion actions.2 Low column top temperature required for target operation could accelerate hydrochloric acid corrosion and valve/perforation hole wearing. Signicant fractionation efciency loss between naphtha and kerosene was not recognised during the operation. The bulky fractionation nature of crude distillation service might result in fractionation efciency being insensitive to tray weeping. However, if this valve/perforation hole wearing progresses, signicant fractionation efciency loss will be noticed through substantial weeping. Fouled trays located for the light kerosene/heavy kerosene and heavy kerosene/light diesel fractionation sections are shown in Figure 3. A tar-like substance was discovered around the periphery of the valve legs. Phosphates used for crude oil production were suspected as the root cause. Boiled phosphates may react with kerosene boiling range material and make fouling deposits. A dedicated process evaluation for kerosene or diesel yield limitation was conducted. The original column and tray drawing revealed that intermediate side product and pumparound streams were withdrawn from fractionating trays directly. The originally designed side draw conguration is illustrated in Figure 4. Flow from the crude atmospheric column to the side stripper relies on gravity ow.
www.eptq.com
If the liquid head formed on the collector tray is not high enough to overcome total friction losses from the crude atmospheric column to the side stripper, the ow rate can be limited. Moreover, frothy liquid withdrawn from the fractionating tray’s active area can contain vapour. The presence of vapour can limit this gravity ow. Rigorous pipe line hydraulic evaluation revealed that the gravity line hydraulics could be limited at a maximum target draw rate.3 Case study 1: equipment modification Based on the aforementioned process evaluation and root cause analysis, the original movable valve trays were replaced by xed valve trays. This tray type conversion improved equipment resistance against fouling and valve/perforation hole wearing. The original fractionating trays at draw locations were converted to chimney trays to increase the liquid head for gravity ow. This chimney tray conversion also eliminates the chance of yield loss and start-up trouble through xed valve tray implementation and increases draw liquid residence time for vapour disengagement from liquid. However, this conversion resulted in losing one tray for each fractionation section: light kerosene/heavy kerosene, heavy kerosene/light diesel and light diesel/heavy diesel fractionation. GT-Optim high performance trays with various performance-enhancing features and xed valves were implemented for the rest of the fractionating trays. As described earlier, light and heavy kerosene streams are combined and rundown as a single kerosene intermediate product stream. Therefore, fractionation performance between light kerosene and heavy kerosene streams is not critical. The same rundown conguration of light and heavy diesel streams does not necessitate sharp fractionation between the two streams. However, fractionation performance between heavy kerosene and light diesel streams affects rundown kerosene and diesel intermediate product quali-
www.eptq.com
Crude atmospheric column
FT FT
LC
Side stripper
Pumparound
LT
Steam Rundown product
Figure 4 Case study 1: original design side draw configuration
0.7 F º ∆
0.6
,
n o i t c u d e r t n i o p g n i z e e r F
0.1ºF
0.5 0.4
Pre-retrofit test run
0.3
Retrofit design prediction
0.2 0.1 0
Number of theoretical stages gain ∆ for heavy kerosene − light diesel fractionation section
Figure 5 Case study 1: pre-retrofit sensitivity analysis
1.4 F º ∆
1.2
,
n o i t 1.0 c u d e 0.8 r t n i 0.6 o p
0.1ºF 3.0%
Pre-retrofit test run
g 0.4 n i z e e 0.2 r F
Retrofit design prediction
0
0
10
20
30
40
50
60
70
80
90
Top simulated reflux L/V ∆ for heavy kerosene − light diesel fractionation section, %
Figure 6 Case study 1: pre-retrofit sensitivity analysis
ties including the kerosene freezing point, one of the key specications for kerosene rundown. To predict kerosene freezing point change, dedicated sensitivity analysis was conducted.
The aforementioned high performance tray implementation could improve individual tray efciency. Nevertheless, extra individual tray efciency improvement was not counted to predict the retrot
PTQ Q1 2017 93
Case study 1: performance summary Case Parameter Yield balance
Pre-retrofit Test run
Post-retrofit Test run
Base Base Base Base
+∆12 + ∆6.0 +∆19.5 +∆12.1
Base Base Base Base Base
+∆1.0 -∆1.0 ∆0 + ∆0.2 - ∆15
72 4 7 16
62 6 10 21
Crude charge, BPD Unstabilised naphtha, LV% Kerosene, LV% Diesel, LV% Fractionation performance
Light kerosene 5% - naphtha 95%, 1 °F Light diesel 5% - heavy kerosene 95%, 1 °F Kerosene flash point, °F Kerosene freezing point, °F Heavy kerosene/light diesel internal reflux L/V, 2,3 weight basis % Heat balance
Heat removal - overhead condenser, 2 % of total BTU/hr Heat removal - heavy kerosene pumparound,2 % of total BTU/hr Heat removal - light diesel pumparound, 2 % of total BTU/hr Heat removal - heavy diesel pumparound, 2 % of total BTU/hr 1. ASTM D86 (LV%) 2. Simulated value 3. At the top tray of the section
Table 1
heavy kerosene/light diesel fractionation performance.
0.1°F by using a chimney tray conversion scenario. Figure 6 also shows another kerosene freezing point sensitivity curve per heavy kerosene/light diesel fractionation section internal reux L/V (liquid/ vapour) ratio. A freezing point increment of 0.1°F was predicted at a 3% lower heavy kerosene/light diesel fractionation section internal reux L/V ratio. Undetected kerosene freezing point changes were anticipated through the sensitivity analysis.
Case study 1: sensitivity analysis For reliable sensitivity analysis, simulation modelling was rst validated with pertinent unit test run conditions. Simulated kerosene freezing point value was reasona bly matched to actual value. The tray efciency and internal vapour/ liquid trafc prole for each fractionation section were quantied through model validation. A constructed kerosene freezing point sensitivity curve per varied theoretical stages is plotted in Figure 5. This curve predicted that the freezing point could be increased by
Case study 1: performance summary The pre- and post-retrot performances are summarised and compared in Table 1. Both retrot
1.4 F º ∆
1.2
, n o i t 1.0 c u d e 0.8 r t n i 0.6 o p
Post-retrofit test run
15%
0.2ºF 8%
g 0.4 n i z e e 0.2 r F
Retrofit design prediction
Pre-retrofit test run
0
0
10
20
30
40
50
60
70
80
Top simulated reflux L/V ∆ for heavy kerosene − light diesel fractionation section, %
Figure 7 Case study 1: pre- and post-retrofit sensitivity analysis
94 PTQ Q1 2017
90
test run conditions were obtained through the same operating mode, SOR (start of run) for fair comparison. Since the internal vapour/ liquid trafc and heat balances of the crude atmospheric column were not measurable operating parameters, these values were quantied through simulation modelling. Like the simulation model for the pre-retrot case, the simulation model for the post-retrot case was also validated with selected post-retrot test run conditions. The crude charge rate was increased during the post-retrot test run. The product yield balance reveals that the post-retrot test run crude slate contained more kerosene boiling range materials compared to pre-retrot crude slate. Kerosene and diesel yield limitation experienced in the past was eliminated. The simulated crude atmospheric column heat balance of the postretrot case was also shifted from that of the pre-retrot case due to a change in the crude slate composition. Laboratory test results showed that the post-retrot kerosene freezing point was relaxed by 0.2°F. Meanwhile, simulation modelling showed that the post-retrot internal reux L/V ratio for the heavy kerosene/light diesel fractionation was reduced compared to the pre-retrot value. The heavy kerosene/light diesel fractionation section performance through the post-retrot test run was evaluated and compared to the pre-retrot section performance. The post-retrot kerosene freezing point was plotted and compared to the pre-retrot sensitivity curve in Figure 7. Relaxing the kerosene freezing point by 0.2°F predicted an internal reux L/V ratio reduction of 8% on the pre-retrot sensitivity curve. A lower internal reux ratio of 15% was simulated with the same 0.2°F freezing point relaxation at the post-retrot test run conditions. Results indicated that the actual post-retrot heavy kerosene/light diesel fractionation section efciency was more satisfactory than the predicted retrot efciency value.
www.eptq.com
N c r P
Crude units can be designed for reliability.
Maximize Reliability in Grassroots Crude Units
DESALTING Desalter size is highly dependent on crude blend due to dramatic variation in required centerline velocity.
Crude unit operators are far too familiar with a long list
A unit must be designed with the flexibility to care-
of crude unit reliability problems including fouling in
fully control desalter temperature, which can range
heat exchangers and fired heaters, poor desalting, cor-
from 110°C to 150°C, by shifting heat from upstream to
rosion of piping and equipment, and coking in the vac-
downstream of the desalters. Vendors are often judged
uum column wash zone. Many millions of dollars have
on cost alone, which results in minimum sizing for the
been spent fighting these problems, yet they continue
design crudes and rates. Carefully consider whether
to force unplanned shutdowns with depressing regu-
long-term crude trends will soon render these desalt-
larity.
ers inadequate.
Revamps must address reliability issues, but project
CORROSION
scope is hindered by the limitations of existing equip-
In grassroots design, be realistic about metallurgy. Be-
ment. Grassroots design of crude and vacuum units
cause modern refineries do not run a steady diet of the
presents an opportunity to get everything right the first
same crude, consider the sulfur and TAN numbers of
time. Here are a few tips for designing a reliable and
potential crudes outside the unit’s design blend. Chron-
profitable crude/vacuum unit.
ic corrosion issues, or the inability to process high-margin opportunity crudes, will quickly overshadow the ini-
HEAT EXCHANGER AND HEATER FOULING
tial savings from choosing too low of a metallurgy.
High velocities in heat exchanger tubes produce high shear at the walls, preventing foulants from accu-
VACUUM COLUMN WASH ZONE
mulating. High shell-side velocities, coupled with ex-
Wash zones are not for fractionation, they are for
changer designs that minimize dead zones in the flow,
de-entrainment! Pursuing fractionation efficiency by
eliminate shell-side fouling. In fired heaters, high mass
specifying a deep bed with small crimp packing is a
fluxes maximize wall shear, shorten residence time,
recipe for rapid coking. e correct choice of packing
and lower wall film temperatures, all of which reduce
combined with the right wash rate and good distribu-
coking. Furthermore, reliable heaters must have cor-
tion will properly de-entrain while preventing coke for-
rectly sized burners with proper burner-to-burner and
mation.
COKING
burner-to-tube spacing.
3400 Bissonnet St. Suite 130 Houston, TX 77005, USA
+1 (713) 665-7046
[email protected] www.revamps.com
Case study 1: fractionation efficiency comparison Heavy kerosene/light diesel Fractionation section
Pre-retrofit Test run
Retrofit Prediction
Post-retrofit Test run
Base Base 55
+ ∆0.1 - ∆3 50
+∆0.2 - ∆15 65
Kerosene freezing point, °F Internal reflux L/V,1,2 weight basis % Section efficiency,1,3 %
1. Simulated value 2. At the top tray of the section 3. Overall efficiency (number of theoretical stages/number of fractionating trays)
Table 2
Table 2 shows simulated heavy
kerosene/light diesel fractionation section efciencies. As one fractionating tray was converted to a chimney tray, a lower number of theoretical stages were counted for the retrot performance prediction. Nevertheless, the simulated theoretical stage count through the post-retrot test run condition data was maintained at the same level as the pre-retrot theoretical stage count, resulting in improved fractionation efciency between heavy
kerosene and light diesel. The aforementioned GT-Optim high performance tray implementation contributed to the fractionation efciency improvement. Case study 2: crude distillation unit description and background The second case study also examines a crude distillation unit. Figure 8 illustrates the schematic of the unit. The crude atmospheric column in this case was designed as a fully structured packed frac-
tionator excluding the bottom stripping section. Unstabilised naphtha, kerosene and atmospheric gas oil intermediate products were distillated through the crude atmospheric column and the side strippers. The unstabilised naphtha stream was further separated into LPG (liqueed petroleum gas), light naphtha, and heavy naphtha through the naphtha stabiliser and naphtha splitter. The heavy naphtha stream was directed to the reforming unit for aromatic component production. The unit utilised two pumparound circuits as well as an overhead condenser for heat removal. One circuit was positioned as the top pumparound. The other circuit was located between kerosene and atmospheric gas oil boiling range materials. A second packed bed at the top of the crude atmospheric column had the function of fractionating naphtha and kerosene boiling range materials.
Off gas
Crude atmospheric column
LPG
Light naphtha
Top pumparound
Naphtha/kerosene fractionation section Heavy naphtha Steam
Bottom pumparound
Kerosene
Preflash drum Steam
Desalted crude
Atmospheric gas oil Steam Atmospheric residue
Figure 8 Case study 2: crude distillation unit configuration
96 PTQ Q1 2017
www.eptq.com
Register your attendance now at artc.wraconferences.com. Save 15% on your booking by using the code ‘MP15’
20th Annual ARTC Meeting 29 - 30 March 2017, Jakarta, Indonesia
Asia’s most respected event for core refining technology ARTC 2017 speakers include:
Asit Das Head of Refining R&D Centre Reliance Industries
Dempsy Robby Kambey GM Asset Management PT Kreasindo Resources Indonesia
Karambir Anand Partner EY
Matthew George Chief Manager, Petrochemicals Marketing, Indian Oil Corporation
Paul Kennedy Vice President, Asia Solomon Associates
Michael Costello, Technical Manager, JVs Chevron
Clive Gibson Partner Nexant
LTG Nugroho Widyotomo Secretariat General National Resilience Council Indonesia
Are you a refiner or petrochemical producer?
Contact:
The World Refining Association has reserved 50 complimentary passes for refiners and petrochemicals producers at attend the ARTC 20th Annual meeting as our guests. Places are reserved for Refiners and Petrochemical Producers such as Pertamina, Petronas, Petrovietnam, S-Oil, Thai Oil, Reliance Industries, Singapore Refining Company and will be given on a first come first serve basis. Apply for your free pass at artc.wraconferences.com
Kelly Tea
[email protected] +44 207 384 7807 artc.wraconferences.com
Case study 2: distributor modification summary Naphtha/kerosene fractionation Distributor drip point density, drip point/ft 2 Distributor operating range Drip hole elevation,1 inch Distributor drip hole diameter, inch
Original 16 2.5:1 1.5 3/16
Modification 4.6 1.7:1 3 7/16
target fractionation efficiency. However, the liquid distributor drip point density originally selected was excessive for the 1in crimp size. The basic equation used to size gravity liquid distributors is:
1. From the bottom of the trough
Table 3
H
=
Lv ⎛ ⎞ ⎜ ⎟ ⎝ k × N × HA ⎠
2
Findings showed that the H = Liquid height (‘Head’) above column’s inside wall cladding round shaped hole using Monel metallurgy was only Lv = Liquid volumetric flow applied to the portion where the N = Number of drip holes top pumparound section was posi- HA = Hole area tioned. The column inside wall K = Orifice coefficient portion of the naphtha/kerosene A minimum liquid head needs to fractionation section remained as be maintained to ensure uniform carbon steel. Rusted wall pieces liquid distribution. A certain from corrosion could accelerate number of drip holes, which indifouling. cates ‘drip point density’, is required for the desired distribution quality. However, unnecessarily high drip A minimum liquid point density reduces distributor drip hole size and increases a chance head needs to be of fouling. Distributor operating Case Study 2: root-cause maintained to ensure range affects hole size because the identification liquid head should be maintained at the minimum rate for uniform Inspection during unit turnaround uniform liquid showed that the trough-type liquid liquid distribution. The original distribution distributor for the naphtha/keroliquid distributor (as designed) was sene fractionation section was not properly optimised between A review of the original distilla- fouling resistance and liquid distrifouled. Several root causes of the fouling were identified through tion equipment drawing revealed bution quality. rigorous evaluation. that the gravity flow trough-type Amine-based corrosion inhibitor liquid distributor for the naphtha/ Case study 2: distributor was injected into the crude distilla- kerosene fractionation section was modification tion unit. Chloride present in the designed with high drip point The liquid distributor modifications column overhead may react with density and small drip hole size. for the naphtha/kerosene fractionathe inhibitor and form ammonium The naphtha/kerosene fractiona- tion section are summarised in Table salt, which can foul the distributor. tion packed bed was equipped with 3. In order to enlarge liquid distribFormed ammonium salt particles structured packing with a 1in crimp utor drip hole size, the drip point could reside in the top pumparound size and a 45° inclination angle. density was reduced in a new bed and also migrate to the naphThis packing size at the given bed design. The new density was careheight was suitable to achieve fully selected by considering the tha/kerosene fractionation bed. commercially proven drip point density in the given size packing 330 and application. Distributor operatPost-retrofit ing range was also adjusted further F Pre-retrofit º 320 , to increase distributor drip hole size. e r 310 The minimum end of the distributor u t a 300 operating range was increased. This r e p 290 adjusted distributor operating range m does not reduce the unit operating e T 280 range. The minimum rate of the liquid distributor does not have to 270 0 50 100 150 200 250 300 350 400 be matched to the minimum unit Unit run length, days charge rate.4 Heat balance shifting through pumparound adjustments Figure 9 Case study 2: performance trend – naphtha end point or increasing furnace coil outlet A particular unit limitation the refiner faced was that fractionation performance between naphtha and kerosene was substantially downgraded after a four-month operation. Substantial amounts of kerosene boiling range materials were downgraded to the naphtha stream. This downgrading not only limited the kerosene yield but also influenced the downstream reforming unit performance. The high rear end distillation point of the heavy naphtha stream adversely affected the reforming reactor catalyst activation.
98 PTQ Q1 2017
www.eptq.com
temperature can maintain the required minimum distributor rate during lower unit charge rate operation. This strategy can increase energy consumption during minimum charge rate operation. But it can assure efficient unit operation in the entire charge range and more efficient overall unit economics can be achieved. Distributo Distributorr drip hole elevation from the bottom of the trough was increased to slow down distributor fouling. The measured naphtha end point and kerosene flash point trends are plotted in Figure 9 and Figure 10 respectively. Plots in red indicate values gathered during pre-modification operating periods while plots in blue represent values achieved after the modification. Stable naphtha end points and kerosene flash points were maintained for more than eight months of operation. These case studies show how fouling resistance and efficiency requirements for distillation equipment are balanced and optimised
114 112
F º , 110 e r u 108 t a r 106 e p m104 e T 102
Post-retrofit Pre-retrofit
100 10 0 0
50
100
150
200
250
300
350
400
Unit run length, days
Figure 10 Case study 2: performance trend – kerosene flash point
for reliable crude distillation unit performance. This article is an updated version of a presentation given at AIChE’s Spring Meeting Distillation Symposium, 11-14 Apr 2016, in Houston, TX. GT-OPTIM is a mark of GTC Technology US LLC. References 1 Lee S H, et al, al, Optimising crude unit design, PTQ,, Q2 2009. PTQ
Middle East Sulphur 2017
2 Kister H Z, Distillation Operation, McGrawHill Company, 1990. 3 Libermann N P, Process Design for Reliable Operations, Gulf Publishing Company, 2nd Operations, Edition. 4 Bonilla J A, Don’t neglect liquid distributors, Chemical Engineering Progress, Progress , Mar 1993. Soun Ho Lee is Manager of Refining Application for GTC Technology US, LLC, in Euless, Texas, specialising in process design, simulation modelling, energy saving design and troubleshooting for refining and aromatic applications. Email: Email: Sounho Sounho@gtct @gtctech.co ech.com m
Platinum Sponsor:
12-16 February 2017 • Jumeirah Hotel At Etihad Towers, Abu Dhabi, UAE
Exploring the sulphur value chain
Organisers of the 33rd Sulphur Conference
CRU is delighted to announce a new conference devoted to exploring the entire sulphur value chain and the growing role of the Middle East as the world’s largest producer of Sulphur.
The commercial programme will be devoted to:
The technical sessions will focus on:
• A thorou thorough gh explo exploratio ration n of econ economic omics s underp underpinni inning ng the the oil and gas markets and the subsequent impact on the global sulphur market • The role of the Middle East as the world’s largest sulphur producing region • Demand updates from key consumers • Maintaining the quality of sulphur and identifying alternative uses for sulphur and new areas for sulphur demand
Promoting best practice on the technical and operational aspects of sulphur production with a particular focus on establishing operational best practice via the sharing of operational experience through operator-led papers incorporating: • The latest technologies and case studies for sulphur production • SRU performance and reliability • Energy efficiency • Emissions compliance • Sulphur forming & handling
Event Partners
Official Publications
Media Partners
CRU VIEW Global update on the sulphur market
Technical workshops & training
Exhibition & sponsorship Enquiries contact: Michelle Fisk +44(0)20 7903 2159
www.middleeastsulphur leeastsulphur.com .com and register To book your delegate place go to www.midd regi ster TODAY.
www.eptq.com
PTQ Q1 2017 99
WWW.ZWICK-ARMATUREN.DE
TRI-SHARK 100 % CONTROL VALVE 100 % TIGHT
Firing high sulphur fuel An evaluation of schemes to improve firing efficiency and ensure the reliability of combustion equipment ADIL REHMAN, C STEVEN LANCASTER, SANDEEPAN GHOSH, OM PRAKASH SAHU and PAWAN KUMAR SHARMA KBR Technology
P
ressure to reduce the carbon footprint of processing facilities means a reduction in fuel consumption by the energy consuming plant. At the same time, the ever increasing price of fuel means that reners are compelled to utilise cheaper fuels like fuel oil, renery gas, or sub-grade fuels with a high content of sulphur and other impurities. On the credit side, there is the possibility of increased prot margins as a result of increasing energy demand worldwide. Thus reners and petrochemical companies aim for process revamps in order to improve their prot margins as well as decrease their carbon footprint. Furnaces are major consumers of fuel in a typical renery or a petrochemical plant, and various process schemes are explored here in order to process dirty fuels without severe corrosion problems or compromising the life of the equipment and avoiding any unwanted shutdowns. Basis of study For the sake of the study, hypothetical base case and revamp case models have been developed, and simulations for a hydrotreater fractionator feed furnace have been carried out using typical industrial data and studied using commercial software, FRNC-5PC Version 4.18 Mod 7.6. Definition of cases The various process cases considered for the study are as follows: • Base Case A is the existing base case of a natural draft heater with an absorbed duty of 15 Gcal/h. • Revamp Case B1 [Option I] is
www.eptq.com
Exhaust to safe location
Fuel
Combustion system
Ambient air Air circuit Flue gas circuit
Figure 1 Existing process scheme (Base
Case A)
the revamp case of a balance draft heater with an absorbed duty of 15 Gcal/h with on-board heat recovery in the form of a conventional air preheater with multiple carbon steel tube bundles. • Revamp Case B2 [Option II] is the revamp case of a balance draft heater with an absorbed duty of 15 Gcal/h with heat recovery in the form of an on-board glass coated tube type air preheater. • Revamp Case B3 [Option III] is the revamp case of a balance draft heater with an absorbed duty of 15 Gcal/h with heat recovery in the form of an off-board ground mounted cast and glass type air preheater. • Revamp Case B4 [Option IV] is the revamp case of a balance draft heater with an absorbed duty of 15 Gcal/h with heat recovery in the form of an off-board ground mounted cast type air preheater with an upstream steam air preheater required to maintain the lowest tube metal temperature above acid dew point. • Revamp Case B5 [Option IV] is
the revamp case of a balance draft heater with an absorbed duty of 15 Gcal/h with heat recovery in the form of an off-board glass coated tube type air preheater. Comparing the base case scheme with revamp schemes Basic schematics of the existing base case along with various other process revamp schemes are shown in Figures 1, 3, 4, 5, 6 , and 7. Methodology for study Typical base case and revamp case process ow rate, duty of the furnace, process inlet and outlet temperature, process inlet and outlet pressure have been considered and a feed property grid has been generated using commercially available process simulator PRO II. Results of the various process schemes were compared to evaluate their relative advantages and disadvantages The study was carried out with the following assumptions and data for modelling: the base case is considered to be a process scheme with a 15 Gcal/h natural draft heater without any heat recovery, radiant heat loss of 1.5%, excess air of 20%, ambient air datum temperature of 15.6°C and relative humidity of 60%. The revamp case is considered to be a process scheme with a 15 Gcal/h balanced draft heater with heat recovery, radiant heat loss of 2.5%, excess air of 15%, ambient air datum temperature of 15.6°C and relative humidity of 60%. The typical renery fuel gas composition in mole% is as follows: hydrogen, 44; methane, 13; ethane, 15; propane, 22; isobutane, 5; hydrogen sulphide, 0.6; and isopentane: 0.4.
PTQ Q1 2017 101
Parameters and simulation results for base case and revamp cases Step 1
Building the base case simulation model, and running the simulation base case. Step 2
Building the revamp case simulation models, and running the simulation revamp cases. Step 3
Evaluating various process schemes available for efficiency improvement for the revamp cases. Step 4
Comparison of various revamp options with the base case. Step 5
Comparison of various revamp options with each other. Step 6
Evaluating merits/demerits of various revamp options.
Heater system under study
Base Case A
Revamp Case B1
Process flow, kg/h 259 416 Process in/out temp., °C 317/378 Absorbed duty, Gcal/h 15 Fired duty, Gcal / h 18.47 Fuel flow, kg/h 1551 % excess air 20 Efficiency,% 81.2 Acid dew point,°C 140 Lowest tube metal temp., °C 293 Max tube metal temp., °C 415 Avg. radiant heat flux, Kcal/h.m 2 18 869 Bridge wall temp., °C 787 APH duty, Gcal/h N/A APH air in/out temp., °C N/A Total CO2 emitted MTPD 105 NOx2 emission, ppm 46
259 416 317/378 15 16.72 1411 15 89.7 140 155 428 19 617 797 1 851/233 95 81
Revamp Case B2
Revamp Case B3
Revamp Case B4
Revamp Case B5
259 416 259 416 259 416 259 416 317/378 317/378 31/378 317/378 15 15 15 15 16.25 16.25 16.72 16.25 1372 1372 1411 1372 15 15 15 15 92.3 92.3 89.7 92.3 140 140 140 140 N/A 3 1555 155 N/A3 429 429 428 429 19 758 19 758 19 617 19 758 799 799 797 799 1.8 1.8 1 1.8 15.64/295 15.66/295 856/233 15.64/295 93 93 95 93 95 95 81 95
1 During low ambient conditions such as in winter, a steam air preheater or electric heater may be required to increase air inlet temperature for avoiding any acid corrosion issues. 2 Considering low NOx burner design. 3 Lowest tube metal temperature is not applicable here as the metal tubes are coated with glass. Thus more appropriate terminology may be the lowest glass temperature. Acid condensation would not cause any corrosion problem as metal is coated with glass in the cold zone. 4 During low ambient conditions, the air inlet temperature would be lower but would not have any acid corrosion issues due to glass tubes construction in the cold zone. 5 Lowest tube metal temperature shall be maintained by vendor in the cast section. 6 During low ambient conditions, the air inlet temperature would be required to be increased using a steam air preheater and/or air may be partly bypassed for avoiding any acid corrosion issues. *Note the figures mentioned in the tables are indicative in nature and meant for study purposes only.
Table 1 Step 7
Conclusion/inference.
Exhaust to safe location
Figure 2 Algorithm followed for the study
A typical algorithm followed for the purposes of the study is shown in Figure 2. Important parameters and simulation results for the base case and revamp cases are shown in Table 1.
On-board conventional APH
Fuel
Combustion system Preheated air
Results and discussions Air circuit Revamp Option I Flue gas circuit This scheme is identical to the base case scheme except that conventional on-board metallic tube bundles have Figure 3 Process scheme (Option I: been used for heat recovery, thereby Revamp Case B10 targeting improvement in fuel efciency. Overall structural weight gas side and air side. Thus, where would be increased due to the addi- there is a plot limitation, for examtion of an on-board air preheater, ple in a revamp, such a scheme is a thereby increasing load on the foun- feasible option for improving fuel dation as compared with the existing efciency, provided space is availa base case conguration. ble for fans and ducting. Although extra footprint is not In the event of acid condensaenvisaged for an on-board air pre- tion, the carbon steel tubes of an heater, ducting and footprint would on-board air preheater are expected be required for a ue gas fan and to corrode at an estimated rate of combustion air fan to take care of approximately 2 mm/y. Thus for extra pressure drop across the ue low ambient or winter conditions, a
102 PTQ Q1 2017
steam air preheater is envisaged for avoiding any acid corrosion issues, which will add to the hardware requirement of the Option I conguration. This situation is not envisaged with the base case as there is no heat recovery and no chance of acid corrosion. Efciency improvement for Option I is around 9% compared with the base case and thus would bring down fuel consumption substantially. CO2 emissions would fall from 105 to 95 t/d, thereby saving 3300 t/y of CO2 being emitted to the environment, considering 330 on-stream operation days. However, quantities of other emissions like NOx will go up for the revamp case compared with the base case conguration. Because this scheme introduces fans, electrical power consumption would be required compared with the base case, which essentially does not require any electric consumption. Soot deposits on the outside of the tubes, especially in the case of fuel oil ring, is a matter of concern as it reduces the efciency of the system. In the case of on-board air preheater
www.eptq.com
designs, water washing in general is not recommended for cleaning of soot since water may spill below and may cause unnecessary damage to coils, refractory and so on. Thus steam cleaning can be recommended by installation of soot blowers in the on-board air preheater region. This would require installation of access platforms for the operation and maintenance of these soot blowers. Carbon steel tube bundles are easy to manufacture and can be fabricated by most construction vendors compared with a cast and glass type air preheater or a glass coated tube air preheater, which are specialised designs available from only a few vendors. This option in general is mechanically more robust compared with a glass type air preheater or glass coated tube air preheater, which are thought of as fragile in nature. Since a carbon steel tube bundle is relatively robust compared with a glass type air preheater or glass coated tube air preheater, there are no transportability issues with this design. Modularisation is also possible with this type of design. If designed and operated properly – always maintaining the lowest tube metal temperature above acid dew point – this type of design would require minimal maintenance. However, if acid corrosion is encountered, then carbon steel tubes would corrode in a few years and unwanted maintenance shutdowns/turnarounds may be required. Delivery at site would be comparatively quicker than for a cast and glass type air preheater or glass coated tube air preheater, which are relatively fragile. Construction in terms of installation and erection would be relatively easy and quick, and can be done by most construction contractors compared with cast and glass type or glass coated tube designs, which may require supplier supervision and services during installation and erection.
substantially. CO2 emissions would fall from 105 to 93t/d, thereby saving 3960 t/y of CO 2 being emitted to the environment, considering 330 on-stream operation days. However, other emissions like NOx will go up for the revamp case compared with the base case conguration. As this scheme introduces fans, electrical power consumption would be required compared with the base case. Soot deposits on the outside of the tubes, especially in the case of fuel oil ring, is a matter of concern as it reduces the efciency of the system. In the case of on-board air preheater designs, water washing in general is not recommended for cleaning soot since water may spill. Thus steam cleaning can be recommended by the installation of soot blowers in the on-board air preheater region. This type of air preheater is a specialised design available from only a few vendors. In general, it would be mechanically less robust compared with an on-board metal tube air preheater design but may be slightly better than a glass type. Transportability may be a matter of concern compared with metal designs. Modularisation is possible with this type of design. The only concern is that this option may offer standard module dimensions governed by glass coated tube lengths and other fabrication aspects. However, tailor-made modularisation may be offered by suppliers upon request. This type of design will require minimal maintenance.
Revamp Option II
This scheme is identical to Option I except that conventional on-board metallic tube bundles have been replaced by glass coated metallic tube bundles in the cold zone region for heat recovery, thereby targeting improvement in fuel efciency. Overall structural weight would be increased due to the addition of an on-board air preheater, thereby increasing load on foundations compared with the base case conguration. Although extra footprint is not envisaged for an on-board air preheater, ducting and footprint would be required for a ue gas fan and combustion air fan to take care of extra pressure drop across the ue gas side and air side. Thus where there is a plot limitation, for example in a revamp, such a scheme is a feasible option for improving fuel efciency, provided that space is available for fans and ducting. Since metallic tubes are coated with glass, acid corrosion issues are not envisaged in this scheme. Maximum efciency improvement can be achieved with Option II of around 11% compared with the base case and thus would bring down fuel consumption
www.eptq.com
PTQ Q1 2017 103
Exhaust to safe location On-board glass coated APH
Fuel
Combustion system Preheated air Air circuit Flue gas circuit
Figure 4 Process scheme (Option II:
Revamp Case B2)
Delivery at site may be in general comparatively slower than for the other options since manufacture of glass coated components may be limited to few manufacturers. However, delivery time can be improved and is a matter of further optimisation during project execution. Construction and erection may be relatively slow as an on-board glass coated air preheater may not t exactly over the existing red heater convection section. This type of design may require supplier supervision and services during installation. Revamp Option III
This scheme is identical to Option II except that an on-board glass coated metallic tubular air preheater is replaced by a cast and glass type off board air preheater mounted on the ground for heat recovery, thereby targeting improvement in fuel efciency. Overall structural weight would be increased due to the addition of an off-board air preheater, but load on the heater foundation would be less compared with the existing base case conguration and on-board air preheater designs since the load would not be on the furnace but shifted to the ground. Extra plot area is envisaged for an off-board air preheater design scheme. Ducting and footprint would be required for a ue gas fan and combustion air fan to take care of extra pressure drop across the ue gas side and air side. Thus where there is a plot limitation
104 PTQ Q1 2017
such a scheme does not seem to be a suitable option for improving fuel efciency. Since this design comprises a glass module in the cold zone, acid corrosion issues are not envisaged. Maximum efciency improvement can be achieved with Option III and should be around 11% compared with the base case, which would bring down fuel consumption substantially. CO2 emissions would fall from 105 to 93t/d, or 3960 t/y, considering 330 on-stream operation days. Other emissions like NOx will rise for the revamp case compared with the base case. As this scheme introduces fans, electrical power consumption would be required. In the case of off-board air preheater designs, water washing is generally available for soot cleaning. This would require installation of access platforms for the operation and maintenance of soot blowers. This type of air preheater is a specialised design available from a few vendors. It would in general be mechanically less robust compared with an on-board metal tube design. Transportability may be a matter of concern compared with metal designs due to their more fragile construction. The cast components also contribute to weight. Modularisation is possible with this type of design. The only concern is that this option may offer standard module dimensions governed by glass tube lengths and other fabrication aspects. However, tailor-made modularisation may be offered by suppliers upon request.
Exhaust to safe location
Off-board ground mounted cast APH
Steam APH
Fuel
Combustion system Preheated air Air circuit Flue gas circuit
Figure 6 Process Scheme (Option IV:
Revamp Case B40
Exhaust to safe location Off-board ground mounted cast and glass APH
Fuel
Combustion system Preheated air Air circuit Flue gas circuit
Figure 5 Process Scheme (Option III:
Revamp Case B3)
This type of design may call for maintenance along with service by the supplier as and when required. Breakage of glass tubes may also occur during operation if excessive vibration is induced due to factors like fan motor and blade induced vibration, and air and ue gas induced turbulence. There are other maintenance issues like damage of polymeric sealing material in the tube sheets of the glass module at ue gas temperatures greater than 280°C. However, these types of issues may be addressed during the project execution stage. Delivery at site may be comparatively slow since the manufacture of glass components may be limited to a few manufacturers, whereas installation and erection are comparable with other options. This type of air preheater may require supplier supervision/services during installation/erection. Revamp Option IV
This scheme is identical to Option III except that an off-board cast and glass type air preheater has been replaced by a cast only type off board air preheater coupled with a steam air preheater upstream of the air circuit mounted on the ground for heat recovery. Overall structural weight would be increased due to the addition of an off-board air preheater and steam air preheater, but load on the heater foundation would be less compared with the base case conguration and on-board air preheater designs since the load of the air preheater would not be on the furnace.
www.eptq.com
Extra plot area is envisaged for an off-board air preheater and steam air preheater design. Ducting and footprint would be required for a ue gas fan and combustion air fan to take care of extra pressure drop across the ue gas side and air side. Thus where there is a plot limitation such a scheme does not seem to be a suitable option for improving fuel efciency. Since this design comprises a steam air preheater to maintain the lowest tube metal temperature above acid dew point, acid corrosion issues are not envisaged. However, there is an additional requirement for steam in the steam air preheater, so the utility requirement would be increased, which might be a limitation for some plants. Efciency improvement for Option IV would be around 9% or more compared with the base case whilst the impact on CO2 and other emissions would be comparable with the revamp options already described. The scheme introduces fans, thus electrical power consumption would be required compared with the base case. This type of air preheater is a specialised design compared with an on-board metal tube bundle type air preheater and may be limited to a few suppliers, but it would be mechanically more robust compared with an air preheater having glass components.
Fuel
Combustion system
Exhaust to safe location
Off-board ground mounted cast APH
Preheated air Air circuit Flue gas circuit
Figure 7 Process Scheme [Option V:
Revamp Case B5]
with Option V should be around 11% more than with the base case, whilst the impact on CO2 and other emissions would be comparable with the revamp options already described. This scheme introduces fans, so electrical power consumption would be required. Water washing is generally available for cleaning soot from off-board air preheater designs. This would require instal-
Various process schemes are available for efficiency improvement in high sulphur fuel scenarios
Revamp Option V lation of access platforms for the This scheme is identical to Option operation and maintenance of II except that an on-board glass soot blowers. coated type air preheater is replaced A specialised design, this type of by an off-board glass coated type air preheater may only be manuconguration for heat recovery. factured by a few vendors. It is less Overall structural weight would be robust compared with an on-board about the same as for an on-board metal tubes design but may be glass coated air preheater, with the slightly better than a cast and glass load shifted to the ground. Extra type air preheater and would need plot area is envisaged and foot- minimal maintenance and service print would be required for a ue by the supplier. gas fan and combustion air fan. Compared with the base case, Where there is a plot limitation, which is a natural draft system, all such a scheme does not seem to be revamp options would necessitate a suitable option for improving fuel replacement of natural draft burnefciency. ers with forced draft types because The design includes a glass of the additional air side resistance coated tubular construction in the of the air preheater. However, cold zone, so acid corrosion issues this is subject to review since the are not envisaged in this scheme. original equipment supplier may Maximum efciency improvement recommend some modications to
www.eptq.com
existing burners and total replacement might not be required. Conclusion Various process schemes are available and currently in operation in plants worldwide for efciency improvement in high sulphur fuel scenarios. Each of these schemes has its merits and demerits. The aim of this study is only to provide basic guidelines for comparison of the available options. Contractors and suppliers will typically provide tailor-made engineering solutions and provide project-specic optimisation. Selection of a particular scheme is frequently driven by project economics and client and operating companies’ previous experiences and preferences. Acknowledgement
Special thanks to Mr Sandeepan Ghosh for his contribution in carrying out detailed analysis of the study work. Adil Rehman is Technical Advisor – Process
with KBR Technology, Gurgaon, India. He has over 14 years of experience in process design in various companies and holds a master’s degree in petrochemical engineering from AMU Central University, India. Email:
[email protected] C Steven Lancaster was formerly with KBR
Technology. He has over 31 years of experience in furnace and fired heater design and specialised heat transfer applications and holds a bachelor’s degree in chemical engineering from Vanderbilt University, Nashville, Tennessee. Sandeepan Ghosh is Principal Technical Professional – Process with KBR Technology, Gurgaon, India. He has over 12 years of experience in process design in various companies and holds a bachelor’s degree in chemical engineering from Delhi College of Engineering, India. Email:
[email protected] Om Prakash Sahu is Technical Professional
Leader – Mechanical with KBR Technology, Gurgaon, India. He has over 16 years of experience in mechanical design with various companies and holds a bachelor’s degree in mechanical engineering from Barkatullah University, India. Email:
[email protected] Pawan Kumar Sharma is Chief Technical
Advisor – Mechanical with KBR Technology, Gurgaon, India. He has over 25 years of experience in mechanical design in various companies and holds a bachelor’s degree in mechanical engineering from Kurukshetra University, India. Email:
[email protected]
PTQ Q1 2017 105
What’s eluding your reactor?
Stability.
For all the engineering that goes into stabilizing your reactor’s performance, the most important factor is the most difficult to control: your feedstock. With Crystaphase reactor optimization, you can take control of the particles, poisons, and maldistribution that otherwise wreck your best-laid plans. The result: predictable performance, year after year, cycle after cycle. Put your reactor back on schedule. Email
[email protected] to find out how.
crystaphase.com +1 281-874-2110
[email protected]
Combating reactor pressure drop A refiner investigates the causes of pressure drop due to fouling in a fixed bed reactor and considers an appropriate mitigation strategy ANKIT A JAIN and AJAY GUPTA Reliance Industries Ltd
T
he dominant factor that limits this fouling problem have been the run length of fixed bed probed in detail. These scenarios units in a refinery is the are: deposition of fines in the voids Frictional losses build-up of pressure in the reactor. between the catalyst particles; and 1. Inlet diffusers 2. Distributors The possible causes of pressure irregular swelling of catalyst parti3. Trays, etc. drop in the reactor are shown in cles in the voids due to chemical Figure 1. Pressure drop due to reactions (see Figures 3b-c). The Particulates fouling requires scientific under- pressure drop across the catalyst 1. Iron sulphide standing and root cause analysis in bed of a single phase packed bed 2. Phosphate products order to develop an appropriate reactor is estimated by the Ergun 3. Carbon particles mitigation strategy. The typical equation. In this article, the correlaPressure 4. NaCl drop pressure profile seen in one of our tion is used to quantify with time (reactor) Organic species commercial fixed bed reactors is the pressure drop due to fouling 1. Olefins/diolefins shown in Figure 2. A rapid increase across the bed. Similarly, the 2. Metal naphthanetes in pressure drop across the catalyst approach discussed here may be 3. Asphaltenes bed was seen in both the single and extended to multiphase packed bed multiphase fixed bed reactors oper- reactors by making appropriate Coke formation ating in our refinery, leading to a changes in the governing pressure 1. Insufficient hydrogen premature shutdown of the reactor. drop correlations.1,2 2. Maldistribution On probing the different layers of 3. High bed temperature catalyst bed, it was found that the Case study: fouling of a packed layers of the bed were either bed reactor plugged with fines (iron sulphide, The two cases of fouling observed coke and other inorganic in the catalyst bed are shown sche- Figure 1 Possible causes of pressure compounds) or in some cases there matically in Figures 3a-c . Case 1 drop across a hydrotreater (black depicts was irregular change in the shape represents a situation in which pressure drop due to inherent sources; of the catalyst particles. In a fouling arises due to the accumula- blue depicts fouling sources) scenario wherein each manufactur- tion of fine particles in the voids er’s aim is to increase the intrinsic activity of the catalyst and, at the same time, refiners aim to pack the Fouling maximum amount of catalyst into a Normal operation given volume, it is imperative that the run length of the reactor should p o r be governed by catalyst activity d rather than by a rapid increase in e r pressure drop due to fouling or u s s other factors. e r This article describes an approach P to quantifying pressure drop due to fouling in a reactor. The unit in question is a single phase packed bed reactor, which was experiencTime ing a severe fouling problem and subsequent pressure drop problems. Two phenomena that led to Figure 2 Schematic representation of a pressure drop profile in a commercial reactor
www.eptq.com
PTQ Q1 2017 107
9 Case 1 8 o i t a r p o r d e r u s s e r P
Case 2
7 6 5 4 3 2 1 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Normalised time
HIRING EXPERIENCED PROFESSIONALS WORLDWIDE
Figure 4 Comparison of pressure drop for the two cases of fouling
A
10 L/Lt = 0.1 (Case 1)
o i t a r p o r d e r u s s e r P
9
L/Lt = 0.3 (Case 1)
8
L/Lt = 1.0 (Case 1) L/Lt = 0.1 (Case 2)
7
L/Lt = 0.3 (Case 2)
6
L/Lt = 1.0 (Case 2)
5 4 3 2 1 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Normalised time B
10 L/Lt = 0.1 m 9 o i t a r p o r d e r u s s e r P
L/Lt = 0.3 m
8 7 6
• refining/petrochemicals process specialty chemicals sales • refining/petrochemicals process technology, operations, maintenance, turnaround and should be willing to travel nationwide and worldwide, along with being prone to hard work and sales.
5 4 3
Having a technical degree and good market knowledge is also required.
2 1 0.10
0.15
0.20
0.25
0.30
0.35
0.40
Voidage
Figure 5a Comparison of pressure drop over the entire length of t he bed vs certain regions of the bed 5b Comparison of pressure drop over the entire length of the bed vs certain regions of the bed
its effect on the pressure drop prole. Comparison of pressure drop due to the two fouling phenomena The pressure drop due to fouling in
www.eptq.com
ITW is expanding and looking for experienced professionals worldwide. The candidates should have a minimum five years’ experience in at least one of these fields:
the two cases (Case 1: particle deposition and Case 2: particle swelling) has been compared in Figure 4. For the same rate of fouling (in terms of volume), the same extent of fouling (length of bed)
The available positions will cover: technical sales, implementation of ITW technologies in the field, sales and operations management. Interesting compensation plans will be given along with serious career possibilities. Please contact: ITW S.r.l., C.da S.Cusumano, 96011 Augusta ITALY Email:
[email protected] i www.itwtechnologies.com i
dp,f / dp,c = 50
Feed filter system
dp,f / dp,c = 10
10
Scale trap baskets
dp,f / dp,c = 5.0
9
dp,f / dp,c = 2.5
Graded material
o 8 i t a r p o r d e r u s s e r P
7
Sacrificial beds
6
High porosity graded material
5
Antifouling chemical
4 3 2
Figure 8 Pressure drop mitigation measures in hydrotreaters
1 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Normalised time
Figure 6 Comparison of pressure drop over the entire length of the bed vs certain regions of the bed
10
o i t a r p o r d e r u s s e r P
dp,f / dp,c = 0.10 (Case 1)
9
dp,f / dp,c = 0.17 (Case 1)
8
dp,f / dp,c = 0.25 (Case 1) dp,c = 2 mm (Case 2)
7
dp,c = 3 mm (Case 2)
6
dp,c = 5 mm (Case 2)
5 4 3 2 1 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Normalised time
Figure 7 Effect of diameter of catalyst on the pressure drop profile
and same diameter of the particle bed, the pressure drop is higher in the case wherein fines are concentrated in the voids between the catalyst particles compared to the deposition of fines on the surface of catalyst particles. At the end of the same cycle, the pressure drop in the first case is roughly twice that in the second case. Effect of fraction of bed fouled We varied the fraction of the region in which fouling takes place (usually the topmost layer of the bed). It can be noted from Figure 5a that the increase in pressure drop is most sensitive to the increase in the fraction of fouled bed; this is
110 PTQ Q1 2017
Pressure drop mitigation in hydrotreaters
evident from the slope of the increase in pressure drop for the L = 0.1 m to L = 0.3 m case. It can be seen from Figure 5b that at 0.27 of voidage an exponential increase in the pressure drop across the reactor begins. We define this as the critical voidage. Hence a reactor showing an exponential increase in pressure drop may have reached the critical voidage of 0.27 at a certain layer of the bed. It can also be inferred from Figure 5a that the distribution of fouling over the entire length of the bed, instead of being concentrated in a smaller region of the bed, is crucial to increasing the run length of the reactor. This indicates the impor-
tance of strategies that ensure the distribution of fines (fouling particles) over the entire length of the bed. Skimming of the top layer of the bed from time to time is applied in various units. The length of the bed to be skimmed can be estimated from these calculations. Effect of fines diameter Simulations to study the effect of fines diameter were done by keeping the volume of fines constant. These simulations kept the diameter of the catalyst particles constant. It can be seen from Figure 6 that as the particle size decreases, this leads to a higher rate of pressure drop. Hence strategies to trap the fine particles are important in order to ensure an increased run length for the bed. Effect of catalyst particle diameter Simulations wherein the diameter of the catalyst particles was increased (keeping the constant ratio of dp,c : dp,f in Case 1) showed that as the catalyst particle size increases there is a decrease in the rate of pressure drop, whereas in Case 2 the change in the pressure drop profile is negligible. Mitigation strategies Various technologies have been employed for mitigating pressure drop problems in a reactor. As discussed, localised fouling causes a rapid rise in pressure drop; normally this would occur in the topmost layer of the bed. One of the strategies could be to remove this fouled layer. For practical
www.eptq.com
reasons, this is possible if we have two separate beds: a main catalyst bed and a sacricial/guard bed. For low pressure systems, we may also have two guard beds in swing mode while for high pressure systems one needs to be careful in employing such a strategy. There are other commercial technologies such as cata lysts with macropores at the top layer of the catalyst bed to capture nes, injecting antifouling chemicals, and so on. A short overview of the various technolo gies is shown schematically in Figure 8. The calculations shown in this article prove the importance of having appropriate barriers to ensure that nes do not enter the catalyst bed. Summary and conclusions • Maximum availability of the catalyst bed, avoiding premature shutdown of the reactor due to factors such as fouling, plays a dening role in the protability of the renery. • Various factors leading to the rapid increase of pres sure drop in the reactor have been discussed in this article. • Two cases of fouling in a single phase packed bed reactor have been described. Case 1 assumes that foul ing is due to the deposition of spherical particles (nes) in the voids in the catalyst bed, while in Case 2 the assumption is that fouling occurs due to ellipsoidal swelling of particles in the voids of the catalyst bed due to deposition of nes or polymerisation reactions. It was observed that the pressure drop in Case 1 is higher than in Case 2 because of the higher surface area the uid ow encounters. • An approach to calculating the pressure drop due to fouling in the reactor is described. This approach can be extended to other xed bed reactors with an appropriate change in the pressure drop equations. • Sensitivity analysis with respect to various parame ters has been described. It is seen that the pressure drop is highly sensitive to the length of layer in which fouling takes place. The calculations prove the impor tance of strategies to ensure that nes are deposited in the entire length of the bed rather than accumulated in a concentrated region of the bed. The approach discussed in this article will be useful in quantifying the pressure drop due to fouling in a xed bed reactor and in designing an appropriate miti gation strategy. References 1 Al-Dahhan M H, Larachi F, Dudukovic M P, Laurent A, High-pressure trickle-bed reactors: a review, Industrial & Engineering Chemistry Research, 36, No. 8, 1997, 3292-3314. 2 Ranade V V, Chaudhari R, Gunjal P R, Trickle Bed Reactors: Reactor Engineering & Applications , Elsevier, 2011. Ankit A Jain is a Research Scientist in the Refining R&D department of Reliance Industries Limited. He holds a bachelor’s degree and PhD in chemical engineering from NIT Surat and IIT Bombay respectively. Ajay Gupta is Assistant Vice President and currently heads the fixed bed process development group of the Refining R&D department at Reliance Industries Ltd., Jamnagar, Gujarat, India. He holds bachelor’s, master’s and doctoral degrees in chemical engineering from IIT, Delhi, India.
www.eptq.com
L X X n i y t i l a u Q
ECOTROL® control valve ARCA Anti-Surge Control Valves Flow- and noise optimized valve housing and trim Stable control even under severe service conditions due to fully pressure balancing of valve plug Fast and accurate response with negligible overshoot allows efficient compressor operation close to the surge limit Rigid pneumatic actuators with proven pneumatic instrumentation Up to 28” (DN 700)
For further product information
ARCA Regler GmbH, D-47913 Tönisvorst Phone +49-2156-7709-0, Fax …-55,
[email protected]
www.arca-valve.com
ARCA Flow Group worldwide:
• Competence in valves, pumps & cryogenics • Subsidiaries and partners in Switzerland, the Netherlands, India, P.R. China, South Korea, Mexico and USA!
PTQ Q1 2017 111
THE SINGLE SOURCE FOR EMISSIONS CONTROL. We’ve been honing our expertise for more than 85 years. Investing in experts and assets. Innovating through research and development. And earning unrivaled combustion and vapor control experience at installations worldwide. Let us put our expertise to work for you.
johnzinkhamworthy.com | +1.918.234.1800
©2017 John Zink Company LLC. johnzinkhamworthy.com/legal-notices
The design temperature of flare systems A key objective when setting flare system design conditions is to maintain the integrity of the system during fire relief PAUL DAVID Paul David Process Ltd
A
great deal has been written about the minimum design temperature of are and blow-down systems because cold temperatures can cause brittle failure or over-stressing of are piping. Less has been written about setting the maximum design temperature of are system piping. American Petroleum Institute (API) Standard 521, 1 which is used throughout the industry by process engineers, gives only general guidance on how to approach the task of setting the (mechanical) design temperature. The standard advises that the extremes of temperature of the uids entering the header should be considered, heat transfer analysis may be performed, it is common to exclude the re case, and careful analysis is required. There is considerable scope for interpretation by individuals and companies of how this is to be implemented. The author’s experience is that company standards differ in their approach and in the level of detail they offer, sometimes giving rise to more questions than answers. The renery process engineer, faced with tables of relief case data, has to place a number in the box on the line list that says ‘Design Temperature’. It is a number for which he or she will be held accountable. The design temperature of process equipment is usually the higher of: the maximum normal operating temperature plus a margin (typically of the order of 25°C); or the highest temperature expected during start-up, shutdown or upset. Upset conditions include the operation of the pressure relief valve and for equipment containing a
www.eptq.com
saturated liquid the temperature is location remote from the original higher at relief pressure than at nor- re. The result would be a major mal pressure. The selection of relief escalation of the incident. It is therevalve body and ange rating are fore not illogical that parts of the normally based on the equipment are system should have a higher design conditions. design temperature than that of The re contingency is not nor- the equipment from which the re mally considered when setting the relief stream originates. However, design temperature of the equipment the header piping in on-plot are or its associated pressure relief valve systems is highly constrained (by (PRV). The re relief temperature for many piping branches) and likely a heavy hydrocarbon mixture, with to be large in diameter. Setting an a wide boiling range, can be very unreasonably high design temperamuch higher than the equipment ture will result in major mechanical design temperature. This can lead to design problems. the somewhat uncomfortable result There is no industry wide practhat the PRV re case operating tice for setting design temperature temperature is higher than its design based on re case data. This scenario temperature. requires careful consideration to If during the re the vessel PRV avoid either an inadequate speciopens to relieve vapour, the temper- cation or an infeasible specication. ature of the uid entering the are API Standard 521 states that it is system may be much higher than common practice to exclude the re the vessel design temperature. Such relief scenario when specifying the a possibility gives rise to the ques- design temperature of the are headtion: should the design temperature ers. It may be a common practice of the are system be higher than but, in the author’s experience, it is the design temperature of the vessel certainly not ubiquitous. To consider it serves? re case for the majority of renery A re on a hydrocarbon pro- PRVs but to ignore it when specicessing unit usually means that a fying the disposal system to which loss of containment has already they discharge might be considered occurred. When a vessel contain- questionable. After all, a re on an ing hydrocarbon is subject to heat oil renery is a reasonably foreseeainput from a major re, further ble contingency. failure (for instance at the vessel The following sections assume anges) should not be unexpected. that the re case will not be ignored Although this results in an escala- and consider some of the issues that tion of the re, the incident is still will therefore arise. contained in the same plant area. If hot vapour relief to the are What is the fire case relieving system causes sufcient stress at temperature? any point in the are header (by Wetted wall vessels thermal expansion) then signi- Many renery vessels containing cant damage can occur. This may wide boiling range hydrocarbons cause loss of containment at a will have very high calculated re
PTQ Q1 2017 113
relief temperatures. The heavier relatively short lived. Failure of the heat transfer through the pipe wall the hydrocarbon and the higher the vessel is likely unless the vessel is occurs, but the principle that the PRV set pressure, the higher will be effectively cooled by re-water. In velocity related temperature drop the relief temperature. If we take the either case the relief will cease. For is not fully experienced at the wall average boiling point as indicative these reasons, relief from a vessel remains. As vapour ows through pipe of of the temperature a vessel might containing vapour only is unlikely be expected to reach, then a distil- to heat up a signicant portion of constant diameter, its pressure will late stream with an average boiling the are piping and is unlikely to fall and its velocity will increase point of 300°C might be over 400°C determine the design temperature – this results in a reduction in temat a relieving pressure of 5 barg. The of the are system. perature. This effect may briey design of the are sub-header piping be reversed when a ow from a would be very challenging at this What is the temperature small pipe enters a larger header, downstream of the pressure sort of temperature. causing the velocity to reduce. Any fall in temperature due to velocity Actually, the relieving temper- relief valve? ature is unlikely to reach 400°C Overall, the ow through a relief increase in the are system will norand it is therefore unlikely that the valve is considered to be approx- mally be less than 10°C for heavy are piping would ever reach this imately isenthalpic – the nozzle hydrocarbons at typical back prestemperature. Hydrocarbons tend ow is isentropic, but this does sures. Again, care should be taken to start cracking if the temperature not continue throughout the valve. in accounting for this temperature exceeds a value of around 350°C; Therefore it is usual to expect a drop – the uid temperature close if the liquid in a vessel is boiling at temperature drop across the relief to the wall would be higher than the bulk temperature for adiaba350°C it is likely that some cracking tic ow. Note: API Standard 521 is occurring at the vessel walls. The If the calculated fire recommends the use of isothermal lower molecular weight materials ow calculations for estimating produced by cracking will tend to relief temperature reduce the effective vapour pressure PRV back pressures from the are is higher than 350°C of the liquid and make it unlikely piping. While this is reasonable that the temperature will continue in that it results in a conservative it should be treated to rise in the way predicted from back pressure estimate (for gases at above ambient temperature) isothe feed stream boiling range. with extreme caution thermal ow calculations are not Depending on hydrocarbon type generally suitable for estimating and molecular weight, it may well be that the hydrocarbon in the valve due to the pressure falling at the are temperature prole. vessel will become supercritical at constant enthalpy. This is typically Isothermal ow is equivalent to the relieving pressure and will no of the order of 1.5°C per bar drop adiabatic ow with the exponent in longer boil. Prediction of what hap- across the PRV for a hydrocarbon the equation PVγ = constant set to a value of 1. For heavy hydrocarpens inside the vessel is now even stream. The ow velocity at the relief bons, the ideal gas value of γ is not more difcult since cracking will still occur (or increase since the wall valve outlet is almost always much above 1, which is why the temperature is likely to increase). higher than at the relief valve inlet. calculated temperature drop (in the The formation of light hydrocarbons Rigorous ow simulators will indi- absence of heat transfer) is usually will tend to increase the critical cate a further temperature drop small. While all these effects are direcpressure of the mixture and also below the stagnation temperature cool the vessel contents since ther- which would be calculated by a tionally ‘helpful’ for the are process simulator at the valve out- system, their combined effect is relmal cracking is endothermic. If the calculated re relief temper- let. (Stagnation temperature is the atively small. When hot gases ow through a ature is higher than 350°C it should temperature that would occur if the ow was brought to rest isentropi- PRV into the are system, the heat be treated with extreme caution. cally.) This additional temperature transfer from the gas to the cold Vessels containing vapour drop will generally not amount to pipe will initially result in relatively When vessels containing only gas or more than a few degrees and should rapid cooling of the hot gas. The vapour are subject to re heat input, not be accounted for in simple heat heat transfer coefcient between very high initial relieving temper- transfer calculations to estimate the owing gas stream and metal atures can be calculated depending the wall temperature. Due to the is likely to be limited by the fouling on the ratio of normal operating to velocity prole, the velocity of the factor (renery are lines are typirelieving pressure. In some cases the uid next to the wall is zero and cally heavily scaled). Nevertheless calculated relief temperature will the uid temperature at the wall the inside coefcient will be well be infeasibly high and failure of the would be nearer the stagnation over an order of magnitude greater vessel would have occurred before temperature than the bulk owing than the outside coefcient and relief. In any case the mass relief temperature for adiabatic ow.2 The this results in the pipe wall heating rate is likely to be low and the relief ow is not actually adiabatic, since quickly. Since the outside coefcient
114 PTQ Q1 2017
www.eptq.com
Advancing the World’s Gas and LNG Value Chain
Full Conference Programme Now Available
Advancing the World’s Gas & LNG Value Chain Conference & Exhibition Programme
OfficialKnowledge Partner:
Organisedby:
For the latest programme and speaker line-up, visit www.gastechevent.com
INSIGHTFUL PRESENTATIONS FROM OUTSTANDING INDUSTRY SPEAKERS
Patrick Pouyanné Chairman & Chief Executive Officer
Ryan M. Lance Chairman & Chief Executive Officer
Total
ConocoPhillips
Alexey Miller Deputy Chairman of the Board of Directors, Chairman of the Management Committee
Sultan Ahmed Al Jaber Minister of State
Gazprom
Abu Dhabi National Oil Company
United Arab Emirates Chief
Maarten Wetselaar Integrated Gas & New Energies Director
Royal Dutch Shell
Executive Officer
25,000+
2,500+
600+
70+
130+
Exhibition Attendees
Conference Delegates
International Exhibitors
Countries Represented
Conference Papers
SUPPORTED BY THE JAPAN GASTECH CONSORTIUM
Download Gastech conference programme on www.gastechevent.com/ptq1
is very low, the steady-state pipe perature of the anges. Resultant perature drop as the vapour ows wall temperature is quite close to over-stresses, which might occur through the PRV and are piping the relief stream temperature. Near due to re relief, can be in the sub- • If the relief valve opens intermitthe PRV, the steady-state tempera- header or one of the branches and tently then are piping may not ture will be reached within minutes. can only be predicted by a full stress reach its equilibrium temperature Once the steady-state temperature analysis. This stress analysis will • The relief stream from the vessel is reached in the pipe segment just reect the fact that only part of the in question may mix with other downstream of the PRV, the gas sub-header may be at high temper- material as it enters the sub-header temperature into the next segment ature, depending on the location of • It takes time for the vapour is not much lower than at the PRV the re. Where the failure occurs is to heat the metal of the are outlet – heat transfer, limited by not necessarily intuitive. It may well sub-header. the outside co-efcient, is slow. The be that the failure is at a ange, but The heat-up time may also allow wall temperature in the rst seg - the ange that fails may not be one the re brigade to start cooling ment is also not much lower than that experiences the highest temper- sprays on the area surrounding the the PRV vapour outlet temperature. ature. The stress may be caused by re. Although the are header is As we move downstream through the growth of piping elsewhere. unlikely to be their priority, cooling the are system, this process keeps Failure in the are system during of the process equipment in the re repeating itself. The gas and wall a localised re is a highly undesira - area should reduce the relief rate. temperatures rise later (due to the ble event. The are is an important There are two methods by which larger upstream mass of metal to be safety system. As the re scenario the design temperature of the are heated) and the steady-state tem- unfolds, operators may vent gas sub-header is usually set: perature is lower (due to the greater 1. Large companies sometimes have upstream heat transfer area). The Failure in the flare a ‘not to be exceeded’ value based rate of both these effects is dependon experience. This value is derived ent on the relief case and the system system during a from many years of operating mulgeometry. However, on a typical tiple process units and is known to localised fire is a process unit the time to steady state be practical for design from the piping stress viewpoint. is still likely to be measured in min highly undesirable utes and the steady-state tempera2. The second method is based on ture is likely to be well within 50°C event rigorous uid ow with heat transof the PRV outlet temperature. fer calculations to give an expected Off-plot are headers are often of metal temperature for the are considerable length, are designed from numerous locations into the sub-header during the worst case with considerable exibility, and are. Failure will result in highly scenario. It is necessary to be careful have far fewer connections. They ammable or toxic gas venting with this approach – the uid ow are downstream of the sub-header, to atmosphere at an unexpected and heat transfer in the are system often downstream of a large knock- location. can be modelled with rigour but the out drum, and the header itself has process engineer should question considerable mass and relatively Setting the flare system design whether the model of the process large heat transfer area. All the temperature vessel, during the re, is equally considerations for the sub-header The are system does not usually rigorous. The calculated re case apply but, for the off-plot header as have a design temperature lower piping metal temperature is used a whole, the rate of temperature rise than the owing temperature for by the stress engineer to ensure the will be signicantly slower and the non-re contingencies. integrity of the are system and a Setting the design temperature suitable mechanical design tempersteady-state temperature will also of the PRV tail-pipe to a value of ature is back calculated. All short be lower. 300°C-350°C, based on the re case, term overstress allowances in the What is the hazard? might not be considered unrea- relevant piping design code should Process engineers are accustomed sonable for equipment containing be taken into account for the re to think of the anges as the weak heavy or wide boiling hydrocarbon case. points in piping systems and most streams. For units where the re relief of us are quite capable of assessing However, 350°C would be a very temperature is assessed at 350°C, a their pressure-temperature rating. high design temperature for the sub-header design temperature in For are sub-headers with multiple process unit are sub-header or the range 250°C to 300°C may not branches, the thermal expansion downstream piping and renery be unexpected. of the header can cause stresses for headers generally have design temThe off-plot are system design which the ange pressure-temper- peratures lower than this. There temperature is found by similar ature rating is not directly relevant. are several reasons for this, includ- considerations and methods to It is the temperature of the pipe wall ing those that have been discussed those applying to the sub-headers. (causing thermal growth) that is above: Generally, the design temperature important, not necessarily the tem- • There will be some heat loss/tem- is lower than for the on-plot piping
116 PTQ Q1 2017
www.eptq.com
since there is more opportunity for heat loss in the upstream piping. It is not normally considered that the PRV tail-pipe or are headers will experience re engulfment. On-plot are sub-headers will normally be at an elevation above the 7.6m (25ft) normally considered for re relief calculations and PRV tailpipes should be situated above their respective headers. Direct radiation is therefore not normally taken into account when determining piping design temperatures. Where there are particular concerns, it may be advisable to provide re-proof insulation although this is unusual on are piping. The provision of any insulation, of course, affects the heat loss calculations discussed above. 1
Conclusion
It is important to know the objective when setting are system design conditions. For large hydrocarbon processing plant, one key objective is to maintain the integrity of the are system during re relief. The consequences of the re may cause
upsets or shutdowns on other process units and the are system is a key utility under these conditions. A are system failure may result in a large release of ammable material at an unexpected location. The author is aware of an incident where a huge reball was caused by a process unit venting into a damaged are system. Although in this case the damage was not caused by thermal expansion it emphasised the importance of maintaining are system integrity, particularly during upset conditions. The use of relieving temperatures, and hence are design temperatures, of higher than around 350°C are unlikely to be warranted for typical renery hydrocarbon streams. For other materials, the behaviour of the material at elevated temperatures needs to be reviewed. Design temperature of downstream sub-headers and headers is based on the maximum expected metal temperature. This is often set by experience since the exact behaviour of the relieving material
when heated by re and the actual contingency is not known until it happens. In the absence of extensive, relevant experience, the large simulation effort required should be tempered by good judgement. The data generated during the re case study should be documented and reviewed with the engineer responsible for the are piping stress analysis. References
American Petroleum Institute Standard 521, Pressure relieving and Depressuring systems. Sixth Edition, Jan 2014. 2 Shackelford A, Temperature Effects for Highvelocity Gas Flow, Chemical Engineering, Jan 2015. 1
Paul David is
Director and Process Engineer with Paul David Process, providing process engineering and overpressure protection services. He holds a bachelor’s degree in chemical engineering from the University of Bath and has over 30 years’ experience in the industrial gases, chemical and oil refining industries. For more than 20 years he worked at a major UK oil refinery.
www.sogat.org SPONSORS
26–30 March 2017, Abu Dhabi
SOGAT Workshops
March 26-27
Workshops will focus on: Amine Treatment; Sour Gas Process Optimisation and Simulation; Novel Methodologies in Mercaptan Removal, and are separately bookable.
13th International SOGAT Conference
March 28-30
The Conference Programme will feature such technical topics as: Energy recovery in CO2 removal processes Case study of SRU and AGE facilities installed in a newly discovered gas field in Egypt Successful implementation of flare gas recovery systems Performance improvements in AGR from ultra sour wells Dealing with CO2 cycling due to CO2-EOR Operational process safety experiences at the Shah field
SOGAT Exhibition
March 28-30
Exhibitors include Energy Recovery, Huntsman, OHL Gutermuth Industrial Valves GmbH, Al Hosn Gas, DOW, Worley Parsons, Sulphur Experts, John Zink Hamworthy and many more. Please visit www.sogat.org/exhibition to review the shell scheme, floor plan and availability.
Media Partners
www.eptq.com
For further information on all aspects of SOGAT 2017 and to reserve your delegate places please refer to www.sogat.org or c ontact Nerie Mojica at: Dome Exhibitions, PO Box 52641, Abu Dhabi, UAE E:
[email protected] T: +971 2 674 4040
PTQ Q1 2017 117
What’s missing in this picture?
YOU!
JOIN THE FACES OF THE INDUSTRY. AFPM Annual Meeting March 19 - 21 Marriott Rivercenter San Antonio, TX
afpm.org/Conferences #AM17
Laser scanning with dimensional control Integrating traditional methodology with new scanning technologies to achieve higher order accuracies for critical interfaces and tie-in points PETER FIELD Warner Surveys
F
or oil and gas companies operating in a volatile climate, careful planning and tight control over construction and maintenance programmes have never been so vital. The sheer immensity of most projects and the scale of investment involved places huge pressures on all parties from stakeholders and engineers to engineering procurement, construction & design (EPCD) contractors, with projects often scheduled for completion within extremely short timescales, working around the clock. Errors or misinterpretation during design and fabrication can easily occur, leading to the need to rework not just minor parts but substantial plant items. A key contributor to project over-runs, reworking can account for up to 15% of total budget costs on new build developments and a significant proportion of the cost for scheduled maintenance and engineering works. These are costs that could be totally removed by anticipating issues ahead of time through closer collaboration and employing survey techniques up front. A huge leap forward for surveying, laser scanning enabled the remote capture and dissemination of large amounts of data, where previously each data point had to be measured individually and, frequently, physically. The technology is now used extensively for the accurate mapping of large scale developments and complex sites, including heavily congested or inaccessible areas with complicated assets in the hostile environments such as those to be found in
www.eptq.com
Figure 1 Laser scanning provides a reliable method for surveying congested and often inaccessible areas of plant. However, hostile site conditions may affect accuracy, and dimensional control techniques may be required for critical tie-in points
reneries and industrial plants (see Figure 1). Collaborative working Being able to call on a comprehensive database offers benets for collaborative working, where many different inuencers require input into the design process. The interoperability of modern 3D models created using a variety of software packages, central to the building information modelling (BIM) workow approach just now emerging in construction, has been employed extensively by the oil and gas sector over many years. Providing a visual representation of plant structural and operational aspects, the digitisation of data in this way creates a central information resource that saves time and greatly reduces
the risk of errors and need for reworking to ensure projects are delivered on time, on programme and on budget. Having a shared source of accurate data allows pro ject engineers and contractors to maintain control over all aspects of the engineering and fabrication phases of oil construction pro jects, delivering improved returns through an integrated approach. Early intervention pays dividends It is often the case that survey services are rst deployed on large projects at the point where fabricated elements built to original designs are brought to site to check. Reworking may be required where misalignment occurs, causing delays that could be avoided by surveying earlier in the process to
PTQ Q1 2017 119
could be up to 15 years old in some cases, increasing the chances of misalignment or poor t where plant elements are fabricated work ing to original designs. For this reason, it is advisable to consider commissioning additional laser scanning surveys to help plan the shutdown properly. Lack of detail could otherwise leave asset owners and contractors having to deal with considerable variations in scope during the execution of projects. Completion of design work in the point cloud or model environment in advance provides the necessary assurance that the process will be clash free.
Figure 2 Basic dimensional control techniques are used to ensure a single weld hook up
(SWHU) for fabrication and installation of new modules
Unplanned shutdowns Unplanned shutdowns can have a signicant effect on facilities and their operators. In addition to the production losses incurred during shutdown, unplanned events place huge pressure on resources when operations are restored, with capacity needing to be restored quickly. Outages of this kind can also affect the wider economy, with the lack of availability of fuel supplies creating a knock-on effect on industry and deliveries. Managing shutdowns and turnarounds successfully involves planning ahead to extend the period between shutdowns and eliminate unexpected downtime through preventative maintenance. Although survey companies including Warners will generally offer a rapid response service to cover emergencies, regular surveys to assess plant condition can avoid unexpected maintenance costs and business disruption, both of which may seriously impact on the bottom line.
approve the designs ahead of fab- replacement, with planned downrication, allowing build issues to be time periods offering the opportudetected and rectied before on-site nity for scheduling in revamps and construction commences. regeneration. Design errors or omissions can Whilst shutdowns are usually lead to inadequate project speci- planned when production is at its cation, with estimates of cost accu- lowest, and such maintenance is racy being underestimated in some important in maintaining produccases. The early involvement of sur- tive capacity, extensive planning veyors at ‘cold eyes review’ stage and control are required, often can help with assessing design con- several years beforehand as well cepts for constructability, providing as during execution. Turnaround access to site-experienced expert time is dependent on the extent of knowledge of the issues that can the project and any problems that occur and potential workarounds occur. Any derailment of timetato guide engineers and designers bles – for example where reworking with less experience to this level is required, with much of the work of project complexity. Early screen- difcult to scope in advance – could ing and availability to survey data result in the loss of millions of dolplus a fresh pair of eyes can offer lars for every day of lost production new solutions to deliver cost sav- and incur additional direct costs for ings and prevent project overruns, labour and heavy equipment usage. helping to secure an earlier return Key to the planning process is on investment. the availability of accurate legacy data for buildings and structures. Shutdowns and turnarounds – The availability of a comprehen- Laser scanning or dimensional minimising your downtime sive database generated through control? Shutdowns are an inevitable fea- laser scanning and the conversion The choice of survey type can ture of rening and can represent of point cloud data into computer sometimes be confusing for proa signicant proportion of a plant’s aided design (CAD) models pro- ject teams, with laser scan surveys yearly budget. Typically, these will vides a tool for asset management often requested when the required be scheduled every 3-5 years, tak - and a valuable guide for exten- tolerances in fact necessitate dimening plants or part-sections offstream sions and alterations, particularly sional control techniques with more to undertake inspections required in clash prevention, where pipes traditional instrumentation. to comply with regulations and and new elements of plant must be So how do you decide what you to carry out necessary repairs and integrated within already congested need? These denitions may help: Laser scanning is a rapid and reliable maintenance. A percentage of these sites or structures. events will involve major plant However, existing legacy data method for surveying often inacces-
120 PTQ Q1 2017
www.eptq.com
sible, complex or congested areas. Survey control is the essential, traditional survey activity providing the auditable accuracy to so many survey operations including laser scanning. Dimensional control is the name given to high accuracy survey techniques used to achieve a good t up between new, basic pieces of plant. Critical interface surveying raises the bar for dimensional control and relates to high accuracy techniques and instrumentation used to achieve rst time t-ups between new and old complex (often dimensionally corrupted) pieces of plant or structures. SWHU (single weld hook up) : although basically with the same objective as critical interface surveying, this has become the term generally applied to very large projects with multiple modules requiring a rst time t with no additional spool connections (see Figure 2).
Data accuracy Whilst ‘t for purpose’ for many design and construction applications, laser scanning is subject to inherent limits on accuracy, with terminal accuracy subject to and dependent on a number of variables. Good quality survey control, range, reectivity of the surface, type of scanner, temperature variations, stability of the survey subject and even the symmetry of a survey subject can all have an impact on results. In the real world, most survey subjects are exposed to a signicant temperature change, which alone can account for several millimetres’ difference in the data registered. Additionally, the ‘noise’ produced within a point cloud from a at surface can range from as little as 2mm to 10mm or more depending on the quality of scanning instrumentation and the surface being scanned. Interpolation of this ‘noise’ can lead to further distortion in the data. Where the objects being scanned are vibrating or subject to other inuences that can cause movement, this will also affect the results. For this reason, claims that
www.eptq.com
laser scanning can provide accu- teams can facilitate greater interracy to within +/-2mm should be operability, although this may be approached with caution. Using the hampered by the availability of the highest possible specication terres- necessary software in some cases. trial laser scanner in a stable, con- Often however there is no substitrolled environment at relatively tute for a full CAD model. short range (10-20m) and with high order survey control (+/-1mm), it Dimensional control – raising the bar may be possible to produce data For surveys of existing plant, where to that kind of auditable accuracy. the key aim is to get a record If the data is not auditable, accu- of what is present, accuracies of racy claims should be dramatically 10-15mm, laser scanning and 3D downgraded, with +/-5mm being modelling will be sufcient to more realistic. prove the design is clash free. Moreover, the concept that modA more inclusive and wide rangelled surfaces are more accurate ing approach, critical interface surthan the basic cloud data is only veying calls for a different order of true in certain circumstances. complexity and accuracy in dimenSurfaces are often irregular, the sional control. Essential in achievgeometry of old basic primitive ing a ‘rst time t’, this requires shapes becomes corrupted, and working to tolerances of 1mm or even the lack of dimensional integ- less (see Figure 3). rity of basic primitives can result Applications for this higher order in some software squaring off or accuracy include ensuring complex smoothing out such irregulari- new t-ups such as cone to cone ties. Although they result in a very connections between new and old good looking model, they do not plant are achieved on a rst time accurately represent what actually lift and t or high value reactor/ exists on site, so on this occasion regenerator heads and internal the accuracy of the model can actu- components are replaced in a sinally be dictated by the ‘quality’ of gle lift. Consistently achieving a the structure being surveyed. weldable 3mm root gap actually While a number of software com- requires surveying to a 1mm accupanies are actively involved in racy. Dimensional control is the key investigating how these limitations component of major plant replacecan be overcome, a commercially ment projects involving any major viable solution is still some way item of process plant whether cutoff. Another option available to ting heads off reactors and regenerdata users is to bypass the auto- ators or dimensionally controlling mated software corrections and the fabrication and the installation more complex ‘as-built’ geometries of new PRT lines. by interrogating the underlying Off-site fabrication databases directly. Frequently overlooked in the In the quest for greater consistency delivery of 3D models, the point and faster construction, prefabricacloud database captured by the tion where assemblies are manufaclaser scanning process offers a val- tured under factory conditions then uable resource for designers and transported to site for incorporation contractors. Containing high qual- into project civil engineering works ity, auditable data that can be inter- is increasingly common in revamp preted factually, these databases and new build projects. Offering greater programme offer the exibility to interrogate and view point clouds to produce certainty and lower labour costs, 2D plans and elevations directly. this approach also results in less This allows the freedom to request waste. However, the in-situ work prefabricated assemadditional deliverables from the abutting original survey team in the form of blies requires absolute accuracy sections or drill down into specic to avoid civils interface problems and achieve single weld hook-up areas at a minimal extra cost. In addition, opening up wider (SWHU) objectives. Warner Surveys applies high access to raw data within project
PTQ Q1 2017 121
Figure 3 Checking cone shapes for potential fit-up accuracy
order total station survey control, combining laser scanning with proven methodologies such as conventional survey control, dimensional control and critical interface t-ups and is active in the further development of traditional survey techniques for inspection and visualisation to include drone-based/ unmanned aerial vehicle (UAV) and vehicle-mounted scanners. It is the combination of laser scanning, conventional survey control, dimensional control and critical interface surveying that can provide the overall database for design, BIM and clash prevention applications along with 1mm accuracy dimensional control techniques for specic ‘corridors of interest’ or tie-in points.
plant or pipework connections will therefore be problematic, especially where designs are based on old data. Planning ahead by testing the designs against actual plant condition prior to manufacturing will avoid delays on site and potential reworking. Ambient temperature correction enables matching of connections between the new plant and existing structures subject to temperature variations in operation. The availability of improved intelligence removes the risk of clashes and poor t-ups.
Visualisations Using survey data to build 3D animations can help avoid problems where large built-in cranes are operating in conned spaces. Providing a visual walk-through, Other issues where surveys can help these animations can track the traTower installations jectory of crane movements to This task requires tried and tested ensure there will be sufcient clearstandard operating procedures ance with no part of the crane or including at least one visit to the the load likely to hit surrounding fabrication environment, on-site plant, something that is very useful surveys to the civils interface, during constructability studies. pre-shimming and checks against design elevations to ensure not Case study only a vertical installation but an Warner Surveys was awarded installation that successfully inter- an engineering project for Tengizchevroil Future Growth faces with adjacent plant. Project-Wellhead Pressure ManageReverse engineering ment Project with survey teams Continually vibrating and often based in Kazakhstan, Korea and hot existing plant can become Italy over a four-year period. The dimensionally corrupted over company is part of the team protime. Achieving a successful inter- viding front-end engineering design face between new designs and old (FEED) and engineering, procure-
122 PTQ Q1 2017
ment and construction management activities. The FEED stage is now complete and detailed engineering under way. Warner Surveys has been appointed to oversee all technical aspects throughout with the aim of ensuring rst time t on site (SWHU). While laser scanning models are being used to ensure clash prevention, process plant design and asset management, the technology will only be deployed during engineering and construction to provide nal clearance checks prior to shipping through dimensionally restricted inland waterway systems. The extent of the project and the off-site fabrication of modules require the application of high-end dimensional control techniques throughout to ensure nal correct positioning as per design, working within very tight tolerances of less than +/-2mm. Warner Surveys will be involved in the full life cycle of every module from fabrication and transportation to on-site positioning and the nal hook up. Working closely with the local fabrication teams from its local ofce in Kazakhstan, the company is responsible for the establishment of standard operating procedures and validating the accuracy of the fabrication process through to sign-off. Constant monitoring and reporting allows any occurrences outside of accepted tolerances to be checked back with designs and the necessary amendments made before shipping to site. As part of this brief, Warner Surveys is involved with marking the cut lines for pipe connections. On-site, the company will continue to manage the civils interface by setting out demarcation lines and match marks to allow cranes and other module transportation systems to manoeuvre and lower modules into place on the concrete foundations to achieve the delicate t required. The result is a perfect SWHU, minimising construction timescales and avoiding costly on-site reworking. Pushing the boundaries Continued investment in research and development focuses on the
www.eptq.com
assessment of different technologies Providing 3D movement monitor- and improved integration, with the and new instrumentation to nd ing and change reporting, the auto- reassurance of independent results. the best way to generate accurate mated system issues emails and/ data that is ‘t for purpose’ across a or text messages when pre-agreed Built-in protection range of diverse applications. Often key trigger values are exceeded, With the price of oil expected to the subjects to be surveyed can be in enabling timely action to be taken remain low, oil and gas companies aggressive or hostile environments, to prevent further deterioration, are increasingly concerned with subject to extremes of temperature which might lead to major failures. how they can reduce costs on their and may be hundreds of feet away: capital projects and operations. new solutions are emerging that can Structured support There will always be uncertainty at Increasingly, suppliers are work- this level of complexity, but taking overcome these problems. For example, many chimneys ing with project teams and organi- a few basic steps can eliminate the and industrial plants are in dis- sations on major long-term projects risk of additional costs and minitress. The use of UAVs for inspec- across a range of areas from con- mise delays. Land surveyors have a duty of tion removes any risks and can struction site engineering surveys generate highly accurate inspection to multiple site and large site top- care to ensure the services provided reports on movements or decay as ographic surveys as well as interna- are ‘t for purpose’ and delivered well as dimensional change where tional oil, gas and energy projects. to the required accuracy. Equally Working within a framework however, designers and construcdeployed by survey specialists. Warner Surveys is also currently agreement or on a call-off basis tion professionals need to be clear evaluating new products to meas- ensures the availability of support about what the project demands. ure movement quickly and ef- and expertise at every interface to Ultimately, good survey control ciently, including Leica GeoSystems’ deliver a seamless progression and techniques must lie behind and Kumonos system, which appears to offers an economic solution, xing support the laser scan, however provide a reliable method to map costs through the establishment of high its quality, providing auditaand report changes to cracks on an agreed schedule of rates. With ble proof that there is no degradastructures remotely, thereby remov- the necessary certication, stake- tion in the laser scan due to other ing access issues and improving holders also benet from the added outside inuences. A holistic approach to surveythe health and safety aspects of a assurance of compliance with safety standards. ing, which combines laser scanproject. In addition, Warner Surveys has ning and modelling for project Used largely by construction companies to date, 24/7 automated established a Quick Quote mech- stages where lower accuracies are monitoring and reporting also anism to simplify process man- sufcient, along with the higher offers potential benets for oil and agement and communications order accuracy of dimensional congas companies. Installed and set up by providing a fast turn-round trol for critical interfaces and tie-in in just a few days, the remote sys- for urgent and additional survey points, will always deliver the best tem provides real-time updates via requirements. It was also one of results. dedicated websites on how plant is the rst companies to set up an moving during operational cycles autonomous CAD department, a Peter Field is Managing Director of Warner and whether it is performing safely move designed to give designers Surveys Ltd and a fellow of the Chartered and as per design expectations. and contractors increased exibility Institute of Civil Engineering Surveyors.
dig it alr efining .co m: k eep up t o dat e w it h y our indust r y and mak e bet te r decisions
digitalrefining.com is the most extensive source of freely available information on all aspects of the refining, gas and petrochemical
www.eptq.com
processing industries. It provides a constantly growing database of technical articles, company literature, videos, industry news and events.
PTQ Q1 2017 123
Identifying contributors to flaring Applying an integrated flare and fuel gas monitoring system to identify and eliminate major contributors to refinery flaring P SRIDHAR Indian Oil Corporation
P
eople are often confused when it comes to are monitoring. In a renery, where hundreds of control valves, dump valves and pressure safety valves make their way into the list of causes, it can be difcult to monitor and minimise aring. It is a very complex task for a person to identify difcult and fre quent occurrences of aring, or to pinpoint contributors to aring. Operators can come into contact with a huge list of split ranges and dead bands when they look into this matter. In addition, those respon sible for are monitoring may not have an in-depth knowledge of every process involved, thus mak ing the issue more confusing.
A typical monitoring table Attribute\flare source
Control valve size in inches, D % Opening Upstream pressure in kg/cm 2 g, P Split range Dead band applied Actual control valve opening, %op P*D3/100
Point A
Point B
Point C
1 30 0.5 No No 30 0.3
10 7 2 No No 7 140
6 60 30 Yes (>50% flaring) Yes (+/-5%) 10 (including dead band ((60-55)*2) 648
Table 1
theless be identied or targeted as major contributors to aring depending on the percent opening of control valves. To avoid any confusion, it is proposed to sort aring points based on volumetric ow rate using the fol lowing formula: P*D3/100*% op
Common reasons for flaring A few common reasons for aring This can be used to set the order (apart from acid gas aring) in petro- and magnitude of aring. The for chemical plants and petroleum ren - mula cannot predict the exact eries are outlined below: amount of aring, but it can arrange A. Imbalance in the fuel gas system aring points in order of priority (see B. Process requirement where Table 1). column overheads are operated For all contributors to aring, per 2 at very low pressure (about 1 kg/cm ) cent opening can be normalised on C. Dumping from the hydrogen the basis of actual percent open header ing to are, considering dead band D. Process abnormalities leading to and split range. This method can be are adapted wherever a aring meas E. Process equipment abnormalities. urement is missing or not reliable. Point B above is a system require - For the above example, trivial Point ment to reject lighter gas; it cannot be A can be ignored and Points B and controlled. Point C can be focused upon to Point E can be solved during a reduce the level of aring. turnaround (if fouled) or by revampThe basis of the above equation to ing (if under-designed). Points A, C infer aring is as follows: and D can be addressed by effective At any point, the number of moles monitoring of the system. being ared through a control valve is α. Proposed methodology For the number of moles present in Trivial aring points might none- a control valve port, n:
124 PTQ Q1 2017
n = P * V/ (R * T) n = P * π D3 (6 * R * T)
where P is the upstream pressure, T is the temperature and V is the volume of gas being ared. D is the diameter of the control valve port. Because the temperature from the surge drum, from where most of the aring takes place, is almost constant: n
=K*P*D3 (K is constant) /K=P*D3 dividing by 100 on both sides n /K/100=P*D3/100 n /K/100 represents relative moles n
Integrated flare and fuel gas monitoring sheet Data for an online system can be imported directly from a real time database in Excel format. Flaring data should be retrieved on a ve minutes average basis to make the system dynamic. Flaring control valves can be arranged in order of maximum aring by using the expression:
Pressure in kg/cm2 g *(c/v diameter in inch) 3*% opening
www.eptq.com
INTEGRATED FLARE AND FUEL GAS MONITORING SHEET (IFMS) TIME
1/8/2015 9:21
FLARE SYSTEM Current Highest flaring points UNIT
TAG
RELATIVE MOLES
DHDS-HGU-2
GJA.1011PC1706.OP
160
HGU-3 PSA2
GJA.2041PIC3013.OP
94
JR current Flare factor FLARE STATUS
365
medium flaring
FUEL GAS SYSTEM STATUS
no acton required
JR FUEL GAS GENERATION
61854
JR FG CONSUMPTION
61722
Kg/hr
FGRS DISCHARGE TO GRE HEADER
2954
Kg/hr
GENERATION-CONSUMPTION
3086
Kg/hr
MAKE UP NG+LPG to FG SYSTEM
1190
Kg/hr
ERROR IN fuel gas system calcula ton
1322
Kg/hr
% ERROR
n o i t c u d e R
Kg/hr
0
Time, days
3
Crude processed
50800
MTPD
FG consumpton excluding inerts
60978
Kg/hr
% Fuel gas consump ton on Crude pro
4
Natural gas cost
21
LPG cost
19
USD/MMBTU
IFO cost
11
USD/MMBTU
HSD for Gt
20
USD/MMBTU
Advantage in using LPG in place of NG
3
USD/MMBTU
Advantage in using IFO in place of NG
10
USD/MMBTU
Advantage in using HSD in place of NG
1
USD/MMBTU
Figure 2 Reduction in fuel gas to flare control valve (202 days trend)
USD/MMBTU
n o i t c u d e R
Figure 1 Typical integrated flare and fuel gas
monitoring sheet consisting of total flaring from each unit and highest contributor
Fuel gas hydrogen generation and consumption data can be retrieved on a last one average basis to make it more reliable. These data can be sorted on the basis of their deviation from the last one week average. Based on this, deviation in the system can be identified to fix the actual problem. All of these data can be integrated onto one sheet, an integrated flare and fuel gas monitoring sheet (see Figure 1). Continual updating of these data facilitates easy, one-click understanding of a fuel gas and flare system sheet dashboard. Action to be taken can be incorporated into the sheet, based on the upper and lower limits of individual consumption points. Predetermined disturbances should also be considered in order to determine a strategy to mitigate flaring. For instance, delayed coker unit drum change-over and FCC unit catalyst make-up lead to extra gas generation which will go to flare if not absorbed into the system. Provision should also be made to highlight recovery of poor data as the system depends completely on the quality of data recovered.
www.eptq.com
202
0
200
Time, days
Figure 3 Reduction in hydrogen to fuel gas header (200 days trend)
n o i t c u d e R
n o i t c u d e R
Figure 4 Reduction in make-up gas to fuel gas system
Results The real time trends shown in Figures 2 to 4 demonstrate the effectiveness of a close monitoring system. Conclusion At present, there is no reliable and effective method for determining flaring from process units. The integrated system for flare and fuel gas monitoring described here will facilitate monitoring of a fuel gas and flare system at a glance. It will also enhance operational competitiveness among the various operating units for achieving max-
imum reduction in flaring and fuel gas consumption. Hidden problems will be identified and resolved in an appropriate manner. For instance, a persistent problem may be resolved by the addition of a condenser or by more timely cleaning of a condenser. Acknowledgment
The author would like to thank all colleagues and associates without whose efforts this article would not have been possible. P Sridhar is the Deputy Manager Production
with Gujarat refinery, Indian Oil Corporation. He has nine years’ operational experience in various refinery areas including FCC, atmospheric and vacuum distillation, and cryogenic, extractive and reactive distillation units.
PTQ Q1 2017 125
Technology in Action
Shaping up hydrotreating performance Trilobe (TL) shape
A fine balance of activity and pressure drop has long created a challenge when considering the maximisation of performance for hydroprocessing and hydrocracking units. It is especially a critical balance for high profile units in hydrocracking service that receive large margins for product upgrade and also have high incentive for incremental processing capacity. Recent margins have placed a great deal of pressure on refiners to maximise hydrocracking unit throughput up to hydraulic limitations, which in many cases is a limit set by reactor pressure drop. Limitations in reactor pressure drop can be mitigated by many means, but ultimately catalyst selection is the most critical factor in hydrocracker optimisation. Criterion developed the ATX (Advanced Trilobe eXtra) catalyst shape to allow hydrocracking units to reduce pressure drop and improve activity simultaneously. There are several significant advantages of the ATX shape (see Figure 1). Pressure drop in packed beds is commonly modelled by the Ergun equation which is applicable for single phase flow, but to model two phase flow through packed beds there are modified versions of the Ergun equation. Using the modified Reynolds number the Ergun equation can be simplified as:
150 =
=
+ 1 75
.
∗ ∗
∗
=
!
1
− !
Advanced Trilobe eXtra (ATX) shape
Proprietary technology offered for majority of HC catalyst Lower CBD → lower fill cost (~10% lower $) Better liquid yield: shorter diffusion path → reduced over-cracking Larger void fraction → higher particulate uptake → delayed onset of rapid ∆P build Lower SOR ∆P (15 to 20% lower commercial performance demonstrated vs conventional TL)
This equation illustrates that pressure drop is inversely proportional to both catalyst bed void fraction and the effective particle diameter. Void fraction itself is a function of loading method (dense vs sock) and particle shape. Differing diameter particles of the
Limitations in reactor pressure drop can be mitigated by many means, but ultimately catalyst selection is the most critical factor in hydrocracker optimisation same extruded shape load at the same void fraction and density since packing efficiency is determined by shape. However, larger particle diameter extrudates of the same shape will result in lower pressure drop.
126 PTQ Q1 2017
Conventional HC catalyst shape − used in all HC catal ysts offered after 1994
Figure 1 The Trilobe design allows hydrocracking units to reduce pressure drop and improve activity
This is due to the effect of liquid hold-up and relative velocity in the bed voids. In addition to particle size, particle shape has a significant impact on pressure drop. The following two commercial examples demonstrate the effect of the ATX shape applied in hydrocracking units to unlock additional capacity, creating increases in profitability. Case 1: Canadian refiner Figure 2 shows the normalised pressure drop from
a hydrocracking unit at a Canadian refinery where the previous cycle applied Criterion’s earlier generation Z-3723 TL catalyst and the current cycle is operating with Z-FX20 ATX. The current catalyst load was also dense loaded to maximise reactor performance and yields while still achieving reduced bed pressure drop. The reduction in pressure drop across the reactor has permitted an increase in unit capacity of 20%,
www.eptq.com
resulting in increased profits, better than $10 million annually. In addition to increases in unit profitability, the catalyst load itself saw a reduced fill cost as a result of lower compacted bulk density density.. Case 2: US Gulf Coast refiner
One of the highest capacity hydrocracking units in the world leverages the advantage of ATX shaped catalysts to maximise unit capacity. Over recent operating cycles the following unit has transitioned to Criterion’s ATX catalysts to reduce unit pressure drop. While only ~30% of the overall catalyst load is now ATX shaped cracking catalyst, the unit has been capable of increasing throughput by 35% (see Figure 3). Figure 4 shows the relative difference between the various shaped extrudates used in hydroprocessing applications in terms of pressure drop per bed height contribution ratio with the typical trilobe setting the scale at 100%. The chart shows that a quadlobe particle with the same outer dimensions as a trilobe particle will result in higher pressure drop across the same bed depth. This is well explained when considering the Ergun equation because the effective diameter of a quadlobe particle is smaller than that of a trilobe particle. Further, the ATX shape offers a significant reduction in pressure drop when compared to standard TL particles. While the pressure drop advantages of ATX are remarkable in comparison to conventional particle shapes, the greatest advantage afforded by the ATX shape is exceptional activity. Improved reactor pressure drop profiles with improved activity enables performance gains for these hydrocracking units.
15% lower ∆P a P k , P ∆
Previous cycle TL shape Current cycle ATX shape 0
100
200
300
400
500
600
700
800
900
Days on stream
Figure 2 Improv Improved ed pressure drop in a hydrocracking unit
i s p , P ∆
r o t c a e r d e s i l a m r o N
Previous cycle TL shape
Current cycle ATX A TX shape ~15% lower ∆P
0
0 0 0 5 0 0 0 5 0 0 0 5 0 0 0 5 1 3 4 6 7 2 9 1 1 1
0 2 0 7 0 0 2 1 2 4
Days on stream
Figure 3 Pressure drop performance at a Gulf Coast refinery
120 110 100
%90 , L / P 80 ∆
70 60
James Esteban Esteban is a Senior Technical Service Engineer with Criterion Catalysts & Technologies.
Criterion Catalysts & Technologies Technologies For more information:
[email protected]
www.eptq.com
50
b e l o d a Q u
b e l o i r T
X T A
r d e n i l C y
t l e l e P
r e h e p S
Figure 4 Standard particle shapes used in hydropr hydroprocessing ocessing of equal outer dimensions and loading method
PTQ Q1 2017 127
Alphabetical list of advertisers
AFPM Annual Meeting
118
Albemarle Catalysts Company
2
ARCA Valves
111
Jonell Filtration Group
40
Linde Engineering North America
21
Magnetrol International
48 & 65
Ariel Corporation
22
Merichem Company
70
ARTC Annual Meeting
97
Metso Flow Control
53
Axens
OBC
MERTC 2017
IBC
BASF Corporation, Catalysts Division
IFC
Middle East Sulphur 2017
99
Burckhardt Compression
29
OHL Gutermuth Industrial Valves
77
Cat Tech
85
Process Consulting Services
95
Prognost Systems
54
Refining India 2017
60
CB&I
7
Chevron Lummus Global
13
Criterion Catalyst & Technologies Crystaphase Products
4
103
Rezel Catalytic Technologies
47
63 & 123
Sabin Metal Corporation
31
DuPont Clean Technologies
25
Sandvik Process Systems
75
ExxonMobil
15
Shell Global Solutions
19
Silobau Thorwesten
89
DigitalRefining.com
106
Rembe Safety + Control
Gastech
115
GE Water & Process Technologies Grabner Instruments
32 9
SOGAT 2017
117
Spraying Systems
90
Haldor Topsøe
10
StocExpo 2017
69
Honeywell UOP
27
Sulzer Chemtech
17
WEKA
59
112
Yokogawa Europe
57
78
Zwick Armaturen
100
ITW Technologies
82 & 109
John Zink Hamworthy Combustion Johnson Screens
For more information on these advertisers, go to eptq.com/advertisers.aspx 128 PTQ Q1 2017
www.eptq.com
Inaugural
MERTC The New Flagship Refining and Petrochemicals Event for the Middle East 23-24 January 2017 | Manama, Bahrain Under the Patronage of
H.E. Shaikh Mohammed bin Khalifa bin Ahmed Al Khalifa Minister of Oil Kingdom of Bahrain
Esteemed advisors include:
Eng. Awadh M. Al-Maker Executive Vice President, Technology & Innovation SABIC
Mohammad Ghazi Al- Mutair i CEO Kuwait National Petroleum Company (KNPC)
Ahm ad A Al- Oha li Chief Executive Ocer
Saudi International Petrochemical Company (Sipchem)
To register go to mertc.wraconferences.com or contact Rosie on
[email protected] Save 10% when booking by quoting ‘PTQ10’ at the online checkout
Brought to you by the organisers of
and
17th annual