Future of LNG_Arup_April17.pdf

Gas and LNG Storage The Future of Modular LNG Tanks Tanks

Gas and LNG Storage | The Future of Modular LNG Tanks

1. Introduction 1.1 LNG and LNG to Power Market Overview The LNG supply market has doubled in the last decade to 301.5 MMPTA [1], and it is anticipated that the next decade will see further growth, particularly in the USA, Canada, East Africa and FLNG, increasing by 46% to 443 MMTPA by 2021 based on projects currently under construction. construction. This expansion is associated with very high and increasing LNG liquefaction costs. For those terminals coming on line by 2021 the estimated CAPEX is over $1,500/tonne. Efforts to lower the unit costs of liquefaction has seen a move away from very large scale, bespoke trains tr ains to a modular, multi-train approach, based on smaller, midscale 0.5 to 1.0 MMTPA trains, such as Energy World’s proposed plant in Sengkang, Indonesia. At the beginning of 2016 regasication capacity, or potential demand, was 757 MMTPA, including just over 10% FSRU capacity,, but from 2000 to 2015 utilization has remained capacity r emained between 30% to 40%. With With capacity only expected to expand to 810 MMTPA by 2021, utilization would need to increase to 44% to meet the estimated increase in demand. Clearly there has not been a lack of regasication capacity for the LNG supply, but some analysts have predicted an oversupply of LNG [3].

Global trade was 245 MMTPA in 2015. The average yearly growth of LNG demand since 2000 has been 6.6% pa. If this continues, demand would reach 358 MMTP MMTPA A by 2021, which would represent a utilization of 80% on the planned liquefaction capacity by 2021, without allowing for capacity taken ofine. Some recent reports [2] have suggested the so called glut in LNG has not materialized, and the numbers above could lend support to that view.

But the LNG market is changing, oil prices are lower, LNG prices are being driven down, even renegotiated, renegotiated, and buyers are seeking shorter term, more exible contracts. Despite these challenges there was 890 MMTPA of proposed new liquefaction capacity in January 2016, key regions being US, Canada, East Africa and FLNG. Clearly many of these projects will not proceed as they compete for supply contracts, but this should encourage demand side expansion. A number of factors will drive demand side expansion including conversion to cleaner cheaper fuels for power generation, to either reduce particulate pollution from coal red power stations or convert from fuel oil. In addition, those countries that seek to honor their COP21 commitments are likely to see natural gas and liqueed natural gas (LNG) as an essential transition fuel to a lower carbon future. For those countries with an established gas distribution network, large scale regasication terminals, in excess of 1.0 bcfd are appropriate, whereas archipelagoes such as the Caribbean [4], Indonesia and the Philippines need to consider a hub and spoke solution in which large scale LNG imports (7.0 MMTPA) can be distributed by smaller LNG carriers (30,000m3) directly to the power station. Some Some companies companies are now areconsidering now considering vertical integration vertical in which they provide both supply LNG asboth well as demand integration in which they of provide supply in terms of LNG to Power. According to Anatol Anatol Feygin of of LNG as well as demand in terms of LNG to Cheniere “that will be the major growth for LNG demand Power. According to Anatol Feygin Feyg in of Cheniere Chen iere going forward” and is a model it is looking to replicated “that will be theLNG major growth formaking LNGthe fuel globally [3]. Lower prices are also demand going forward” and countries is a model more attractive. However, for those thatitdoisnot have an established gas distribution the capital costs looking to replicated globallynetwork [3]. Lower LNG of receiving, storing and regasication at each power station prices are also making the t he fuel more attractive. can inhibit the development of LNG to power projects.

However, for those countries that do not have an established gas distribution network the capital costs of receiving, storing and regasication at each power station can inhibit the development of LNG to power projects.


Storage Tank Capacity (m3 )

Power Station Capacity(MW)

Offtaker

2018 (Conv)

2032 (New)

Total

By 2018

By 2023

Total

The Bahamas, BEC (NP)

393

320

713

70,000

90,000

160,000

The Bahamas, GBPC

240

0

240

20,000

20,000

40,000

Barbados, BL&P

60

245

305

35,000

45,000

80,000

Belize, BEL

62

40

102

5,000

15,000

20,000

1,025

1,800

2,825

370,000

460,000

830,000

Guyana, GPL

140

240

380

35,000

45,000

80,000

Haiti, EDH

238

560

798

55,000

225,000

280,000

Jamaica, JPS

621

1,320

1,941

140,000

155,000

295,000

Suriname, EBS

299

640

939

50,000

105,000

155,000

3,078

5,165

8,243

780,000

1,160,000

1,940,000

Dominican Republic, All

Total

Table 1 LNG Storage Tank Capacities for Caribbean hub and spoke scenario [4]

1.2 LNG to Power Storage Requirements For the Caribbean the IADB report [4] forecasts gas demand of 490 MMSCFD or almost 23,000m 3 of LNG per day. Assuming a hub and spoke scenario for distribution from the Dominican Republic, Table 1 summarizes the storage capacity by country required in 2018 and then expansion through to 2032 to meet the forecast demand.

For perspective a modern combined cycle gas turbine has a thermal efciency of between 45% to 55%. Therefore a 100MW power station consumes approximately 800m 3 of LNG per day or over 24,000m 3 per month. The report outlines conversion and new build forecasts for CCGT, single cycle gas turbines and reciprocating gas fueled power stations. The results highlight a key issue for development of LNG to power projects. The storage capacities at the end of each spoke are relatively small. If the capacity for the Dominican Republic is ignored the average storage capacity is approximately 50,000m3. If Belize is also ignored and a nominal tank size of 20,000m 3 is assumed, each phase of development could be based on multiples of a standard tank size. By 2018 the Caribbean market could require 20 x 20k m3 LNG storage tanks with perhaps another 40 x 20k m 3 or 20 x 40k m3 LNG tanks by 2023.

The Caribbean is only one example. Other countries, such as the Philippines and Indonesia have much greater demand for power station conversion and new build. Also, projects in Central America are being considered, but based on receiving 150,000m3 LNG carriers unloading to FSRUs or onshore LNG regasication terminals that are effectively oversized for the power station capacity.

1.3 Background to Modular LNG Tanks The authors have been involved with LNG tank design and development for almost 20 years. In that time the traditional solution for LNG storage in excess of 10,000m 3 has been a stick built 9% Ni steel single or full containment LNG storage tanks. Most LNG projects have targeted throughputs greater than 1000 MMSCFD or 7MMTPA. The storage volumes for this size of regasication or liquefaction plant have exceeded 160,000m 3. Indeed as the capacity of LNG carriers has increased up to 266,000m 3 (Q-Max) the onshore storage tank size has also increased to ensure lling or discharge can be achieved within 24 hours.


Figure 1 27,500 m3 ethane/ethylene/LNG carrier operated by Evergas ©

Relatively little work has been done to develop cost effective storage tank sizes for the LNG to Power market. Tank sizes greater than 160,000m3, required to receive a standard export LNG carrier, would provide 10 months of storage for a 100MW CCGT. Even for a larger power station it is clear that there is a mismatch between the storage tank and the exporting LNG carrier. Smaller carriers exist, using Type C or membrane technology, but there is a denite requirement for smaller ships to support cost effective LNG to power delivery. Ships in the range of 10,000m 3 to 30,000m3 would allow smaller marine facilities and be compatible with the required onshore storage.

Figure 2 Economies of scale- tank volume [8] or number of tanks [9]

Presentations at the Trinidad Oil and Gas Conference in 2014 [5] and Gastech 2015 [6] have highlighted the market opportunity for LNG to power and emphasized that the design and delivery of smaller LNG tanks is essential to reduce overall cost and schedule to ensure that the cost base is reasonable and the market sustainable. Another market that is expected to see signicant expansion is the LNG marine fuels business. Eagle LNG has recently completed its project in Maxville, FL, USA and Conrad Shipyard is building an LNG bunkering barge. The LNG volumes for each ship are suitable for Type C storage containers, but aggregated onshore LNG storage tank volumes in excess of 10,000m 3 are necessary.


1.4 The opportunities The small to midscale LNG market, supplying power stations or the marine fuels business, requires a smaller capacity LNG storage tank, in the range of 10,000m 3 to 100,000m3. The traditional solution based on 9% Ni steel technology is stick built on site. It is well known that the unit price of LNG stored reduces as the single tank size increases [8]. However economies of scale can also be achieved by production volume. The modular LNG tank seeks to reduce the unit cost for smaller LNG storage volumes by targeting offsite manufacturing productivity levels. The economies of scale are based not on the volume of a single tank but the number of units produced to achieve the required volume. A good reference case was the production of 25,000 m 3 tanks in South Carolina [9]. The estimated productivity improvements, interpolated from the stated productivity for the initial 10 tanks, are shown in Figure 2. It is noted that the rst sphere in that project experienced severe component t up and some welding issues. Since the basic tank unit can be in the range of 10k m 3 to 40k m3, larger total volumes can be achieved with multiple tanks, which can also align with project phasing goals. The following sections in this paper will provide an update on development of the modular LNG tank concept.

The key modular LNGLNG tank drivers are: The key modular tank drivers are: – –Standardize tanktank design by volume based on based site specic Standardize design by volume on seismic isolation

site specic seismic isolation

– Offsite tank pre-fabrication in parallel with foundation –construction Offsite tank pre-fabrication in parallel with

foundation construction

– Dedicated fabrication yard leading to improved productivities higher quality – Dedicated and fabrication yard leading to

improved productivities and higher quality – Offsite pre-commissioning of tank – –Reduced executed on siteof tank Offsitemanhours pre-commissioning These drivers target a “plugexecuted and play” capability – Reduced manhours on site while reducing costs and schedule compared to the stick built traditional solution.target a “plug and play” These drivers

capability while reducing costs and schedule compared to the stick built traditional solution.


2. Technical Development

Figure 3 Initial Modular Tank Concept [6]

2.1 Initial Concept

2.2 Current Concept

The initial concept [6] was based on either 9% Ni or membrane technology. To reduce the overall weight the modular tank provides single containment capability, thereby eliminating the concrete wall and roof. The tank was erected on a cellular concrete base which provided a robust susbstructure for subsequent transportation by water from the fabrication yard to the project site.

2.2.1 Design

At the project site the tank was supported on bearings, founded on shallow footings or piles. Trenches between the foundations allowed access for the self-propelled modular transporters (SPMTs).

After the presentations in 2014 [5] and 2015 [6], specic project opportunities focused further development.

The initial concept considered a maximum volume of 36,000m3, and this was considered to be close to the upper bound of what could, or should, be pre-fabricated and transported, before costs were negatively impacted. However, an opportunity to consider a 40,000m 3 single containment design on the US GoM coast provided the basis for the next phase of development. Technical assumptions are presented in Table 2. The updated design in shown in and Figure 5


Figure 4 40k m3 9%Ni Steel Single Containment Modular LNG Tank General Arrangement

Figure 5 40k m3 9%Ni Steel Single Containment Modular LNG Tank Details


Remark

Value

Remark

Value

Design Standards

NFPA59A, API625/620

Outer Tank

LNG Storage Tank Type

Single Containment

Material

Steel ASTM A36

Foundation Type

Piled supported, elevated

Outer tank diameter

40.000

m

Min width annular space

1.250

m

Inner Tank Material

9Ni ASTM A533 Type 1

Dome Roof

Net Capacity

40,000

m3

Material

Steel ASTM A36

Gross Capacity

42,696

m3

Spherical Radius

40.000

Inner Tank Diameter

37.500

m

Insulation Material

Height (ambient)

39.380

m

Bottom

Cellular Glass

Annular

Expanded Perlite

Suspended deck

Glass ber blanket

LNG Product

m

Seismic Design

Temperature

-170

oC

OBE (pga)

0.037

g

Density (BOG)

440

kg/m3

SSE (pga)

0.074

g

Density (max)

470

kg/m3

Wind

Latent heat of vaporization

511,000

J/kg

ASCE 7-05

63

m/s

Design boil o rate (vol)

0.05

%/day

Soils

US GoM typical

Maximum lling rate

850

m3 /hr

very soft to rm cohesive

0-30

ft

Max out pumping rate

2,250

m3 /hr

rm to sti cohesive

30-100

ft

slightly over consolidated

>100

ft

Pressures Maximum design pressure

190

mbar

Minimum design pressure

-5

mbar

Table 2 40k m3 Single Containment Tank Design Data

The key technical developments are summarized below. –

The tank is elevated above ground to provide both space for the SPMTs and also air ow to eliminate base slab heating.

–

The cellular concrete base slab is replaced with a steel grillage and concrete deck. This reduced weight which is a signicant issue for the larger tank volume.

–

9%Ni was chosen over membrane based on owner preference and concerns over permitting delays that might arise since membrane tanks have not yet been approved by FERC. This issue is discussed further in the next section 2.3.2.

–

Side wall discharge is proposed. This is consistent with NFPA 59 and if in-tank shut off valves are provided the design spill is signicantly reduced. The tank elevation also ensures that the pump does not need to be recessed below ground to achieve the minimum NPSH. Typical details were presented at LNG 12 [10] refer to Figure 6. The results of the techno-economic evaluation concluded that side wall pump discharge could reduce costs by up to $6MM for a 2 x 140k m 3 storage tanks (1998 prices). But the prize is even greater for the modular LNG tank. Not only is the pump platform signicantly reduced in size, refer to Figure 7, but the tanks can be manifolded reducing the total number of pumps. The pumps can also be located outside of the bunded areas with easy access for maintenance.

–

For larger total volumes, based on multiple units, the modular LNG tank will require individual bunded areas. This area can be optimized based on the work carried out by Coers (2005) [11].


Figure 6 Proposed side entry pump suction nozzle for a single integrity LNG tank [10]

Figure 7 Comparison of roof platforms with and withou t side wall discharge (courtesy Cheniere and Coers [11])

2.2.2 Execution

–

The hydrotest is not carried out at the fabrication yard. It was concluded that owners and or regulators may require proof that the 9%Ni inner tank was not damaged during transportation. Transferring the test to the project site signicantly reduces the foundation loads at the fabrication yard.

–

The inner and outer tanks are erected as complete prefabricated rings in a stepped sequence starting with the outer tank then the inner tank. A linear layout for multiple tank erection is shown in Figure 8. A heavy lift crane is used for ring installation.

–

The roof is prefabricated as one piece and lifted into position. No air lift is envisaged.

–

After roof erection the bottom insulation and inner tank bottom plate can be installed, providing weather protection to the insulation.

–

It is assumed that the fabrication yard has a bulkhead suitable for load out of 5,000 te, however temporary loading ramps can be used, founded on a piled ground beam where soil conditions are not strong enough.

Based on the design described above an execution plan was developed working with Great Basin Industries and Mammoet. The overall scope of work was divided into a number of work packages as summarized in Table 3. The following notes highlight some important issues regarding the execution plan. –

–

Fabrication yards do exist along the US GoM coast. The work to date has not undertaken a detailed evaluation of potential sites, but greeneld development is also an option. This approach will increase the initial startup costs and therefore it has been assumed an existing facility will be utilized. Fabrication facilities are not limited to the project country, indeed the modular LNG concept envisages regional fabrication yards that will support LNG storage tank in that area, thereby reducing the shipping times and costs.


Figure 8 Modular LNG Tank erection (courtesy of GBI and Mammoet)

Figure 9 Modular LNG Tank Transportation (courtesy of Mammoet)


Tank prefabrication

Tank Transportation

Project Site

Fabrication yard enabling works

Supply of all heavy lift equipment

Enabling works for receiving tank

Tank fabrication line foundations

Supply of all marine equipment

Construction of tank foundation

Material procurement

Load out at fabrication yard

Hydrotesting

Steel grillage fabrication

Tow to project site

Perlite insulation

Tank ring prefabrication

Ooad at project site

Tank hook up

Tank erection

Set down at project site on plinths

Bund construction

Tank roof prefabrication

Demobilization

Final pre-commissioning of tank

Roof Erection

Ready for cooldown

Pre-commissioning Preparation for transportation

Table 3 Execution Work Packages for 9% Ni steel single containment modular LNG tank

–

–

–

The tank will be moved on to the transportation vessel using SPMTs. Whether the SPMTs remain for the duration of the tow depends on distance. For short tows the SPMTs will travel with the tank, although the tank will be lowered on to temporary supports on the transportation vessel. For long tows (more than several days) two sets or SPMTs are require, one at the fabrication yard and one at the project site. Sea fastenings will depend on the specic tow route. For inland water way tows or sheltered water tows initial calculations indicate vessel motions will not r equire any seafastening for the inner tank. The outer tank will be fastened to the vessel deck. For longer tows or open water tows, temporary sea fastening of the inner tank will be required. Calculations have shown that the inner tank top ring stiffening and or shell thickness could be increased to cater for the inertial loading. Alternatively temporary restraints to the outer tank shell will provide resistance to the inertial loads. These restraints can be removed once the tank is installed at the project site and prior to hydrotesting.

Enabling works at the project site are relatively modest and cost effective. For load below 5,000 te temporary unloading ramps can be used. This will be founded on a piled ground beam. Temporary onshore mooring onshore points will be required for a traditional Mediterranean spread mooring pattern.

The key benets of the proposed execution plan are

1.

Tank erection is not waiting on construction of the project site tank foundation.

–

Regulatory processes normally prevent any construction on site before project permits have been secured.

–

Many LNG sites require signicant enabling works including, but not limited to, bulk earthworks before foundation construction can commence.

2.

Tank fabrication and erection can start once material is procured and delivered to the fabrication yard.

–

Many large LNG tanks have seen lead times for 9% Ni steel plate of 12 to 18 months. This is very market dependent but it has mitigated the schedule delay waiting for foundation construction and outer wall construction.

–

Material pre-ordering can reduce the lead times, and nancial commitments prior to nal regulatory approval can further reduce the schedule.

–

Tank erection commences with fabrication and erection of the steel grillage and outer tank carbon steel outer tank rings. This material is on much shorter lead times.

–

Based on an established fabrication yard, tank erection can commence well ahead of a stick built tank at the project site.

3.

Signicant, labor intensive activities are transferred from the project site to a dedicated fabrication yard.

–

Project site, stick built tanks are often remote from large resource centers, reducing productivity and or increasing labor costs.

–

Specialist welders are required for the inner 9% Ni tank which incurs a premium for remote sites. Further, in tight labor markets, the transient labor force may be difcult to secure, whereas an established fabrication yard can provide a more reliable resource.

4.

Improved productivities and quality

–

An established fabrication yard focused on tank fabrication can invest in training and equipment to increase productivities and reduce costs.

–

Prefabrication of tank parts can be done in covered areas, further increasing productivities and workmanship quality.


Figure 10 23m diameter tank under tow (courtesy of S mith Group)

Figure 11 Peru LNG Tank 130,000m3 with 256 Triple PendulumTM bearing (courtesy of EPS)

Figure 12 Incheon LNG Terminal founded in elastomeric bearings


2.3 Further Development 2.3.1 Standard Tank Design by Volume The work carried out on the 40k m 3 modular LNG tank conrmed technical feasibility and schedule advantages over a stick built solution. It also highlighted the importance of fabrication yard set up costs. When these are spread over many tanks they are not signicant, as for any pre-engineered, manufactured product. To ensure that competitive pricing is achieved from the start it was recognized that offsite pre-fabrication should not be delivered on a bespoke design basis for each project. The modular LNG tank concept would be enhanced if standard designs could be offered for any site, anywhere in the world. A standard tank design would permit the fabricator to further improve its fabrication and erection methods. Key site specic drivers for modular LNG tank design are: 1.

Soil conditions and foundation design

2.

Seismic conditions and inertial loads on tank and foundation

3.

Other environmental loading conditions (such as wind and snow loading)

4.

Temperatures and effect on insulation design

5.

Tow route, duration and storm conditions

The soil conditions will always be site specic and provided settlement criteria are satised then there is no direct impact on the modular tank design, except for seismic response. Other environmental loading conditions are not signicant drivers of tank shell and roof quantities and conservative assumptions could be made to eliminate this variation.

Preserving a standardize design is always a compromise. Perlite insulation could be maintain a constant thickness and heat leak variations addressed by changes in the roof and base insulation thicknesses. This would impact the overall height of the tank and is not necessarily the most efcient solution. Further work will be required to understand the sensitivity to this issue, but if insulation properties cannot be easily adjusted for a given thickness then conservative insulation thicknesses could be appropriate. Tank response during the tow has been investigated. It is clear that any extreme motions that would impact the basic tank design can be addressed with temporary sea fastenings and strengthening to the outside of the tank which can be ultimately removed and reused.

The key driver on tank shell design and quantities is seismic loading. This is the most signicant lateral load on the tank and in areas of moderate to high seismicity, will govern the tank geometry and shell weight. Some tank designs have adopted seismic isolation to reduce the inertial loading and shell quantities, refer to Figure 11 and Figure 12. According to Earthquake Protection Systems Inc. (EPS) [12] an 85% reduction in seismic loading was achieved, which reduced the overall cost of the tank construction. Despite the cost savings on the Peru LNG tanks, seismic isolation is not the default approach for dealing with moderate to high seismic loads. Lowering the tank aspect ratio (H:R), using inner tank straps to prevent uplift and advanced nonlinear dynamic soil structure interaction (DSSI) can be used to lower the inertial load effects on the tanks. Seismic isolation automatically elevates the tank and introduces a second foundation or base slab. This increases schedule and cost, to which the isolator cost is also added.

For the modular LNG tank these costs are already included and the elevated tank is part of the overall concept to allow for installation using SPMTs. In fact the modular LNG tank is very well suited to adopting seismic isolation because all components are included in the existing design for other reasons. Initial calculations conrm that tuning the elastomeric bearing will lower the seismic loads to those of the base design. The base design could be chosen utilizing the 33% over stress permitted under the Operating Basis Earthquake (OBE). For areas of high seismicity, friction pendulum bearings of the type provided by EPS may be required. The solution for any specic site requires a detailed analysis of the tank foundation system incorporating isolators. It is important that the foundation system (shallow or deep) is incorporated into the model, because signicant reduction in loads can arise due to non-linear response in the soil resulting in longer period response and higher levels of damping.


Figure 13 Effect of seismic isolation on acceleration and displacements [13]

Seismic isolation results in longer period response which is accompanied by an increase in tank transient displacements. This will impact the design of incoming pipework but experience has shown that differential movements can be accommodated in the piping design. If displacements are considered too high then viscous dampers can be added to the isolation system to reduce peak displacements. Isolation of vertical ground motions is not as common, and has not been proposed for LNG tanks to date. Vertical accelerations will increase the effective weight of the LNG and therefore the hoop stresses. In areas of high seismicity, such as the west coast of the US, peak spectral accelerations approaching 1g can occur, but careful DSSI can mitigate these effects. Long period ground motions cannot be isolated and these give rise to sloshing effects on the liquid surface. The codes are clear on the requirements for freeboard under both OBE and SSE conditions. As seismic intensity increases, the freeboard height for a given tank aspect ratio increases. To preserve a standard tank design, bafes could be installed on the underside of the roof to disrupt the sloshing wave, but this is a novel approach which might not be acceptable to owners or regulators. Alternatively, it is accepted that the tank height must be increased to address this issue. However it would require only a minor height adjustment to the standard tank design. Further work is required to understand the variations and impact that vertical and horizontal seismic accelerations have on the modular tank design, but initial results are encouraging and a standardized tank design is possible, which should translate into further reductions in cost and schedule.

2.3.2 Membrane Modular LNG Tank Membrane tanks are not new, indeed more than 100 onshore membrane tanks have been built since 1972, and over 85% of all LNG carriers utilize the membrane technology solution. Two membrane tanks are currently under construction for Energy World Corporation at Sengkang, Sulawesi, Indonesia and Pagbilao, Philippines. In addition there have been recent developments in international codes to recognize and incorporate design provisions for membrane tanks. Nevertheless, the dominant tank technology for LNG storage remains 9% Ni steel. A description of the membrane technology and comparison with above ground 9% Ni storage tanks is presented by Ezzarhouni etal (2016) [7]. Whilst this comparison was for a full integrity or full containment design there are many attributes of the system that are compatible with the objectives of the modular LNG tank and would enhance the overall concept, further lowering the costs and reducing the schedule.

Figure 14 Top view of the bottom oor showing membrane system (courtesy GTT)


These benets are summarized below and quantied in Section 3:

–

GTT has developed a highly modular membrane system based on pre-engineered, manufactured components. This is well aligned with the objectives of a standardized tank design.

–

There is only one structural tank and it is located on the outside. The inner 9% Ni and outer A36 shells are replaced with a 1.2mm stainless steel liner and A537 Class 2 outer shell. Total steel weight and costs reduce signicantly.

–

–

–

–

–

–

Additional benets of a membrane LNG tank are: The keydesign modular LNG tank drivers are:

–

Stainless steel and A537 Class 2 have much shorter procurement lead times and will continue to exhibit much lower price volatility. The total volume of wall insulation, based on PUF lled plywood boxes, is less. Hence, for the same overall external tank diameter and volume the corresponding tank height is reduced, further reducing the shell quantities. The tank transportation weight is lighter than the 9% Ni steel option, despite having all insulation installed prior to load out. Membrane tanks do not require hydrotesting. Leak tightness is demonstrated through the ammonia leak test. Foundation proof loading is of questionable value even for 9% Ni LNG tanks and is not required for membrane LNG tanks which use polyurethane foam (PUF) bottom insulation. No hydrotest means that the tank can leave the fabrication complete with all insulation installed and fully pre-commissioned. After installation at the project site the ammonia leak test could be rerun to satisfy the owner and regulator that no damage was sustained during the sea tow.

The design is fundamentally more robust with respect to transportation loadings. Recalling that 85% of all LNG carriers use the technology it is a well proven technology able to accommodate the strains associated with vessel motion. Further, all transportation loads can be designed into the outer tank which can easily accommodate seafastening and temporary strengthening. There is no thin walled inner shell to seafasten.

–

cycling of 9% Ni tanks is not recommended –Thermal Standardize tank design by volume based on because of the inner tank radial movements. However, site specic seismic isolation the membrane tank is not subject to the same constraints the linertank accommodates the thermal strains within –asOffsite pre-fabrication in parallel with the stainless steel corrugations. foundation construction The membrane insulation space is maintained under a

–nitrogen Dedicated yard leading to This purgefabrication which is continuously monitored. isimproved consideredproductivities a more effective and method of leak detection higher quality than temperature sensors which rely on a spill of LNG

–rather Offsite of tank thanpre-commissioning vapor.

– –The membrane liner permits the use on of sumps Reduced manhours executed site in the tank bottom thereby increasing the net useable tank volume. – These In summary, thetarget membrane modular tank takes drivers a “plug andLNG play” important steps towards the “plug andand play” objective. capability while reducing costs schedule

compared to the stick built traditional solution.


3. Comparison of 9% Ni Steel and Membrane Tanks Dimension

9% Ni Modular LNG Tank

Membrane Modular LNG Tank

Net LNG storage volume (m3)

40,000

Outer tank diameter (m)

40.000

Inner tank diameter (m)

37.500

38.800

Design Maximum Liquid Level (m)

38.802

36.280

Outer tank height to roof joint (m)

42.280

39.460

5.365

5.365

50.447

47.627

Roof rise (m) Overall tank height from ground (m)

Table 4 Comparison of principal dimensions for 9% Ni and membrane modular LNG tank

3.1 Quantities

3.2 Schedule and Cost

Table 4 and Figure 15 summarize the principal dimensions of the 9% Ni and membrane modular LNG tanks.

A comparison of construction schedules is shown in Table 7. The schedule is based on an EPC contract, with all design data, including soils information available at notice to proceed. The membrane tank is estimated to be ready for transportation at the same time as the 9% Ni but the overall schedule is 2 months quicker because there is no hydrotest and annular insulation to complete at the project site.

Table 5 compares the weights, and thereby the quantities, for 40k m3 9% Ni and membrane modular LNG tanks. The following notes explain the key differences. –

The outer tank shell weights are similar weight. The membrane tank is the same diameter, but is shorter because of lower wall and base insulation thicknesses. The membrane tank uses ASTM A537 Class 2 steel compared to A36 for the 9%Ni tank. This is a stronger steel and whilst more expensive per tonne, is more efcient in terms of weight and subsequent welding costs. Bottom shell thickness is 23mm compared to 16mm for the 9% Ni tank.

–

The inner tank compares the weight of ASTM A533 Type I 9% Ni steel with 1.2mm A304L stainless steel membrane. Since the membrane is not structural the weight is substantially less, saving 476te on the inner tank weight.

–

Roof insulation weights are similar, however the PUF insulation system shows a saving in weight of 336te over the perlite, resilient blanket and foam glass blocks used on the 9% Ni tank.

–

The elimination of the inner structural tank and use of PUF insulation has resulted in overall weight savings of 20%. Further the membrane transportation weight is less than the 9% Ni which excludes the perlite.

These results demonstrate that the membrane tank is a lighter design than the 9% Ni steel tank.

Costs are sensitive to local labor conditions and material costs. The costs have, therefore, been normalized and compared to a traditional stick built single containment LNG tank at 100%.


Figure 15 General arrangement for 9% Ni and membrane modular LNG tanks

Item



Total (te)

Transport (te)

Total (te)

Transport (te)

Shell

504

504

544

544

Base

74

74

74

74

Roof

107

107

107

107

Shell

467

467

45

45

Base

66

66

12

12

Bottom

502

502

234

234

Wall

368

300

300

Roof

50

50

54

54

Pump Platform

350

350

350

350

Concrete

990

990

990

990

Steel

212

212

212

212

3,690

3,323

2,922

2,922

554

498

438

438

4,244

3,821

3,360

3,360

Outer Tank

Inner Tank

Insulation

Grillage

Sub-total Contingency Total

Table 5 Comparison of tank weights for 9%Ni and Membrane Modular LNG Tanks


9% Ni Single Containment



100%

90%

80%

Table 6 Cost comparison of 9% Ni and Membrane Modular LNG Tank 40k m3 Activity

Months from notice to proceed 9% Ni Modular LNG Tank


0

0

Purchase and fabricate material

+5

+4

Grillage construction complete

+6

+5

Outer tank erection complete

+14

+10

Inner tank erection complete

+14

+16

Roof installation complete

+15

+11

Insulation complete at fab yard

+17

+17

Transport and set tank

+18

+18

Hydrotest

+19

n/a

Insulation complete at project site

+20

n/a

Final pre-commissioning

+22

+20

Ready for Cooldown

+22

+20

Notice to Proceed

Table 7 Comparison of schedules for 9% Ni and Membrane Modular LNG Tanks


4. Conclusions

The ongoing development work on the modular LNG tank concept has conrmed technical feasibility of both 9% Ni and membrane solutions. The membrane option will offer a more robust design for transportation and also lower costs and shorter schedules. More importantly, the concept of a cheaper and quicker prefabricated small to medium sized tank with “plug and play” capability, based on a standard design that can be installed for any site, anywhere in the world is achievable. Single containment is not appropriate for all projects and jurisdictions. Full containment options are too heavy to transport cost effectively, but initial work looking at precast wall panels and wire wound prestressing as used in the water tank industry, combined with the membrane technology should offer cost and schedule savings.

The small to mid-scale LNG and LNG to Power markets require smaller tanks. Cheaper and faster, smaller tanks will greatly assist this developing market.


References 1.

IGU (2016), “2016 World Energy Report”, International Gas Union

2.

Shell (2017) “Shell LNG Outlook 2017”, http://www.shell.com/ energy-and-innovation/natural-gas/liqueed-natural-gas-lng/lngoutlook.html

3.

4.

5.

Shiryaevskaya, A., Burkhardt, P., (2017), “Hottest thing in LNG is producing power as record glut looms”, Bloomberg news article 18 January 2017, https://www.bloomberg.com/news/ articles/2017-01-18/hottest-thing-in-lng-is-producing-power-asrecord-glut-looms Castalia (2015), “Natural Gas in the Caribbean – Feasibility Studies, Revised nal report (Vol I and II)”, Report to the InterAmerican Development Bank, 30 June 2015. Raine, B., (2014) “Onshore Mid-Scale LNG Terminal Storage and Modularization”, Trinidad Oil and Gas Conference, May 2014

6.

Raine, B., Powell, J., (2015), “Onshore Mid-Scale LNG Terminal Storage Modularization”, Gastech 2015, Singapore, 29 October 2015.

7.

Ezzarhouni, A., Powell, J., Elliott, S., (2016) “Why a Membrane Full Integrity Tank?” LNG 18, Perth, PO-8, 11-15 April 2016

8.

Long, B., (1998) “Bigger and Cheaper LNG Tanks? Overcoming the obstacles confronting freestanding 9% Nickel Steel Tanks up to and beyond 200,000m3”, LNG 12, Perth, 4-7 May 1998, Paper Session 5.6.

9.

Veliotis, P.T., (1977) “Solution to the Series Production of Aluminum LNG Spheres”, Society of Naval Architects and Marine Engineers Transactions, Volume 85, 1977, pp 481-504.

10. Antalffy, L. P., Aydogean, S., De la Vega, F. F., Malek, D. W., Martin, S., (1998) “Technical-economic evaluation of pumping systems for LNG storage tanks with side and top entry piping nozzles”, LNG12, Perth, 4-7 May, 1998, Poster Session B.8 11. Coers, D, (2005) “Transshipping LNG – Downscaling FieldErected Storage Tanks for Lower Prole”, 2005 (Presentation with photos provided by CB&I). 12. Peru LNG, Melchoriate, Peru, “Triple Pendulum bearings protect critical storage tanks”, Earthquake Protection Systems Inc, http:// www.earthquakeprotection.com/pdf/Peru_LNG_Dec08.pdf 13. Symans, M. D., “Seismic Protective Systems: Seismic Isolation”, FEMA, Instruction Material Complementing FEMA 451, Design Examples, Seismic Isolation 15-7-1, http://www.ce.memphis. edu/7119/PDFs/FEAM_Notes/Topic15-7-SeismicIsolationNotes. pdf

Contact us: [email protected] Find us on Twitter and LinkedIn

Future of LNG_Arup_April17.pdf

Recommend Documents