PHAST-6-5-1-Tutorial

PHAST Applications the right tool for safety professional

Leading consequence analysis tool PHAST is the world's leading consequence analysis tool, used by governments and industry in over sixty countries across the globe.

PHAST is a comprehensive hazard analysis software tool which is applicable to all stages of design and operation across a wide range of process industries. Industries currently using PHAST include: • Oil and gas industries • Petrochemical companies • Chemical manufacturers • Government and regulatory authorities • Consultancy and design services • Insurance companies • Universities PHAST is one of a range of products developed by DNV Software, recognized leaders in the field of major industrial accident hazard assessment. As a result it has been adopted by many international companies and governments as a decision support tool in industrial risk and public safety matters. PHAST is used to identify situations consequences which have the for and causing unacceptable to potential population the environment. Such scenarios may be minimized by review of the design, process, plant layout or operational procedures. Those scenarios which remain can then be further scrutinized in a quantitative risk assessment (QRA).

Using PHAST as a Design Tool Key safety issues can be addressed in the conceptual stages which would otherwise give rise to costly alterations at final design. PHAST opens up the opportunity for an engineer to analyze a design for hazards at an early stage to gain some benefits, which include but are not limited to: • Reduction of costs related to post-design modifications • Increase in identification of hazards • Reduction of costs related to operational QRAs • Optimization of plant and pr ocess design With PHAST various scenarios – continuous, as well as time varying and instantaneous – can be modelled while in this stage in order to refine the design. Some of the available scenarios include leaks, line ruptures, tank roof collapse and long pipeline releases in pressurized and un-pressurized vessels/pipes. Some of the applications that have been modeled within various industries are: • Flare sizing • Environmental impact analysis • Equipment sizing

Using independently validated models PHAST is centred on DNV Software's proprietary Unified Dispersion Model (UDM) which has been independently validated and verified. The UDM is the only software in the world to completely model all stages of a dispersing plume.

Using PHAST as a Plant Siting Tool

Using PHAST as an Operational Tool

The location and layout the plant influences profitability and growth as anyofother aspect of the project. Therefore, this must be carefully examined. Various benefits can be achieved at this stage by using PHAST, including: • Reduction in construction and manufacturing costs • Improvement of engineer’s understanding of potential hazards • Increase in population safety • Flexibility for growth

Once a process is are in place, it isthat vitalrisk thatmanagement the proper safety precautions taken and remains a priority. Some of the benefits PHAST brings during this stage include: • Quicker response to hazardous incidents • Increased understanding of the impact of process and equipment changes • Reduction of costs related to losses and insurance

By using PHAST’s Unified Dispersion Model (UDM) distances to hazardous concentrations of interest can be obtained as well as the cloud behaviour as it transitions through the various stages – turbulent jets, buoyant gases, dense plumes and passive clouds. Additionally, the UDM considers droplet formation and pool re-evaporation. Multi-phase and multi-component modelling as well as indoor releases and toxic and flammable (fires, BLEVES, flash fires, vapor cloud explosions) effects can also be modelled within PHAST. Within the siting and layout stage PHASThas been used for many important decisions, including: • Building placement • Construction material • Roads and railroad placement • Space requirement • Tank farm locations

The variety of release scenarios available as well as the application of the UDM allow for numerous combinations of scenarios. PHAST also facilitates the direct modelling of flammable scenarios if the data is available. Some of these direct models include jet fires, pool fires, flash fires, explosions and BLEVEs. Once the scenarios are modelled, it is possible to consider possible mitigation measures. Some applications of PHAST as an operational tool include: • Emergency planning and response • Process/equipment/material changes • Forensic investigations • Risk mitigation plans

Offering a complete solution PHAST is available in three configurations, depending on your needs. It is also links to SAFETI for full risk analyses.

PHAST as a Compliance Tool

include the facility to overlay graphical output such as cloud footprints on maps, ability to connect to a GPS In addition to complying with legislative mandates and server to obtain maps and the inclusion of weather reporting to the corresponding agencies, general practice calls for companies to self-regulate. This means dependent parameters. Data that is common for a particular scenario set can they must conduct operations in a way that protects the health, safety and welfare of their employees and third be stored so it does not have to be entered for every party personnel, including visitors and the surround- scenario – i.e. weather data for a particular plant, mateing population. The benefits associated with using rials used in a plant. Only data specific to an individual case needs to be provided for each scenario, reducing PHAST for compliance and self-regulation drive PHAST’s wide use and recognition within the industry. data entry requirements and analysis time. PHAST has been used to comply with many regulations PHAST Materials Database world-wide, including Seveso II, RMP, and COMAH. DNV Software has acknowledged PHAST’s usage in PHAST uses the industry standard DIPPR database to this context and has facilitated it by including features calculate chemical physical properties. Additionally, the program is capable of creating mixtures and modwithin the program as necessary: eling them through all stages of discharge, dispersion • 10 minute continuous release and subsequent toxic or flammable effects. The DIPPR • Emergency Response Planning Guideline (ERPG) chemical database can also be used to obtain physical averaging times • Blast model • Average time flexibility

PHAST Interface and Data Handling The software is delivered with an on-line help system that includes theory of the models as well as a getting started guide and general help. The windows look and feel of the software and its user-friendly design help decrease the learning curve. Some of PHAST’s features

properties (vapor pressures, densities, etc.) for pure chemicals or multi-component mixtures.

For your engineering software needs PHAST performs complex calculations within a user friendly software package with a modern Windows look and feel.

Industry Leader PHAST is one of a range of products developed by DNV, a recognized world leader in the field of major industrial accident hazard assessment. As a result it has been adopted by many international companies and governments as a decision support tool in industrial and public safety matters. With this backup, PHAST subscribers can trust in DNV’s full commitment to the continued development of the product, where response will be made to operational experience and changing user requirements through regular product review and upgrades.

Validation and Verification PHAST has been extensively validated and verified. The theory and performance of the DNV Unified Dispersion Model (UDM) has also been independently reviewed as part of the EC funded project, SMEDIS, and it has excelled in both theory and performance. PHAST’s validation and verification documentation are provided with the software.

e-mail: [email protected] web: www.dnvsoftware.com

Visit us at: Main offices:

DNV Software branch offices:

Oslo DNV Software

Busan Det Norske Veritas

Kobe Det Norske Veritas

Shanghai Det Norske Veritas

Veritasveien 1 Norway NO-1322 Høvik, Tel: +47 67 57 76 50 Fax: +47 67 57 72 72

DNV Software Nambusan P.O. Box 120 Busan 613-011 Republic of Korea Tel: +82 51 610 7700 Fax: +82 51 611 7172

DNV Software Port P.O. Box 77 Kobe 651-0191 Japan Tel: +81 78 291 1305 Fax: +81 78 291 1330

DNV HouseSoftware No. 9, No. 1591 Hong Qiao Road Shanghai 200336 China Tel. +86 21 6278 8076 Fax. +86 21 6278 8090

Kuala Lumpur Det Norske Veritas DNV Software 24th Floor, The Weld Tower Jalan Raja Chulan 50200 Kuala Lumpur Malaysia Tel: +60 3 2050 2888 Fax: +60 3 2031 8080

Rio de Janeiro Det Norske Veritas DNV Software Rua Sete de Setembro 111/12 Floor 20050006 Rio de Janeiro Rio de Janeiro Brazil Tel: +55 21 2517 7232 Fax: +55 21 2221 8758

London DNV Software Palace House 3 Cathedral Street London SE19DE United Kingdom Tel: +44 (0) 20 7716 6525 Fax: +44 (0) 20 7716 6738 Houston DNV Software 16340 Park Ten Place Suite 100 Houston, Texas 77084-5132 USA Tel: +1 (281) 721 6700 Fax: +1 (281) 721 6880

2 2 0 1 0 5 0 n a G y b d te n ir P s e c i rv e S ic h p ra G E V N D : n ig s e D

DNV Software is the commercial software house of DNV serving more than 3,000 customers in the marine, offshore and process industries. DNV Software is a market leader in software development of design, strength assessment, risk and information management.

5 0 0 2 3 0 / 0 0 8

PHAST Tutorial Manual

DNV SOFTWARE Palace House, 3 Cathedral Street, London SE19DE, UK http://www.dnv.com/software © Copyright Det Norske Veritas. All Rights Reserved. No reproduction or broadcast of this material is permitted without the express written consent of DNV. Contact [email protected] for more information.

Contents Chapter 1

An Introd uct ion to PHAST

1

In the first chapter you open an example analysis provided with the program, explore its main features, and run the calculations and view the results – without having to enter or change any input data.

Chapter 2

Setting up your own Analysis

The second chapter guides you through the process of setting up a Study Folder for performing consequence calculations for a range of common types of hazardous event. The tutorial supplies all of the input values that you will need to complete the analysis.

18

Chapter 1: Introduction

Chapter 1 An Introduction to PHAST What to Expe ct o f thi s Tuto rial The aim of this tutorial is to make you familiar with the ideas and techniques involved in performing a consequence analysis with PHAST, and to give you practice in defining a range of common types of hazardous events. By the time you have finished the tutorial you should have a firm understanding of the issues involved, and be ready to start work on an analysis of your own. The tutorial is divided into two chapters. In this first chapter you will open an example analysis provided with the program, explore its main features, and run the calculations and view the results – without having to enter or change any input data. In the second chapter you will create a new analysis, defining a range of hazardous events and performing a consequence analysis for them. The tutorial should take 1-2 hours to complete. You do not have to complete it in a single sitting, and can take a break between chapters if you prefer.

Starti ng th e Prog ram R unn ing When you install the program, the installation process places a DNV Software folder under Programs in your Startmenu, and also adds a PHASTshortcut to your Desktop. You can use either method to start the program running.

The Main Wind ow When you start the program running, the main window will open as shown.

The Main Window on Startup

1


The first line in the Message Log should state that the “Licence is valid”. You must have a valid license for PHAST set up on your computer in order to be able to enter data and run the calculations. If the Message Log says that you do not have a valid license, you should contact product support using the details given under Product Supportin the Helpmenu. The window will normally open with no Study Folder loaded – where a “Study Folder” is a file that contains the definition of a consequence analysis – and you must open or create a Study Folder file before you can perform any modelling work with the program. If you wish, you can change the Installation Preferences under the Options menu so that the program starts by automatically opening a Study Folder (e.g. the Study Folder you worked on most recently).

Opening the PH AST Exampl e Stud y Fold er The program is supplied with an example Study Folder called “PHAST Example Study”, which is used in this chapter to give a quick introduction to the terminology and approach used in the program. To open the Study Folder, choose Open Example… from the Filemenu. A File Open dialog will appear as shown, displaying the contents of the Examples folder installed with the program files. There are several file-formats available for Study Folder files, but the default format is the *.psuformat, and the PHAST Example Study file is in this format. Select the file, and click on Open. The appearance of the main window changes when a Study Folder is open: there are many more toolbars, and there is a pane with five tab sections at the left side of the window, as shown. The pane is known as the “Study Tree” pane, and you work in its various tab sections to set up the input data for the analysis.

The Main Window with a Study Folder Open 2


The Stud y Tree Pane The Study Tree pane allows you to organise and edit the input data for your consequence analysis. The pane contains a number of tab sections, each of which covers a different type of input data, and these tab sections are described below.

The Models Tab Se cti on The term “Model” is used in two different ways in PHAST, though these different meanings are unlikely to cause you confusion. “Model”: a set of available calculations

The program has several different sets of calculations available, and each of these sets is known as a separate Model and has its own icon. For example, there is a Model known as the “Vessel/Pipe Source Model” that has a blue icon that represents a process vessel; this Model considers the release of material from its storage or process conditions in a vessel or pipe, through all the stages in its dispersion to a harmless concentration, and it also performs fire, explosion and toxic calculations to obtain representative effect zones for the dispersing cloud. There is another Model known as the “Fireball Model” that has a red and yellow icon that represents a fireball flame; this Model considers only thenot radiation effect a and dispersion modelling performed fireball, and does perform any zones of the from release by the Vessel/Pipe Source Model. There are eleven different types of Model in total. You define a given hazardous event that you want to analyse by selecting the most suitable Model from the list of the eleven Models. When you select the Model from the list, the program will insert an icon for that Model into the Models tab section. The icon represents an “instance” of that Model and will have its own set of values for the input data, and you can define any number of instances of a given Model in your Study Folder, each with its own set of input data to represent a particular hazardous even t. As shown in the illustration, the PHAST Example Study Study Folder contains ten instances of one Model (the Vessel/Pipe Source Model), and one instance of eac h of eight other kinds of Model. “Model”: one instance of a particular type of calculation Model

In practice, people rarely use the term “instance” to refer to a given use of a particular Model, and instead refer to the instance directly as a “Model”, so it would be more typical to say that the PHAST Example Study Study Folder contains eight Vessel/Pipe Source Models, one Pool Fire Model and one Fireball Model. The Model icons are organised in a tree structure. The top level represents the entire Study Folder, with the name PHAST Example Study , the next level is the Study (named example ), the third level contains several Folders, and the fourth level contains the Models themselves. You can create any number of Studies or Folders, depending on how you want to organise your analysis.

3


Inserting a Model

You cannot place a Model icon under the Study Folder itself, but only under a Study or Folder. To add a Model at a particular point in the structure, select the Study or Folder, and then select the appropriate Model from the Insertmenu as shown. You can also insert a Model by selecting the Model from the Insertcascade at the top of the right-click menu, or by selecting the icon for the Model from the toolbar.

The Weather Tab Section The Weather tab section contains a folder named Global Weathers with three definitions of weather conditions. The program performs a separate run of the consequence calculations for each separate weather conditions, giving a set of results that are specific to that Weather. The Weather tab section also contains a Study icon called Example Cases . In the Model tab section, all of the Models have been placed inside the Example Cases Study, but you create and use any number of Studies in an analysis. You can insert Weathers underneath a Study in the Weather tab section. Such Weathers are known as “Local Weathers”, whereas those in the Global Weathers folder are known as “Global Weathers”. When the program is processing the consequence calculations for a given Model, it will perform the calculations for every Global Weather and for any Local Weathers under the Study that contains the Model, i.e. the Local Weathers are specific to the Models in that particular Study. The Parameters Tab Section In PHAST, Parameters are background inputs that are applied to all calculations and are not specific to a particular Model. As with the Weathers, there is a set of Global Parameters, and you can also define Local Parameters that are specific to a given Study. If you define a local set of Explosion Parameters, for example, the values in this set will be used instead of the values for the global Explosion Parameters during the calculations for the Models in that Study. Green border to icon: shows use of default values

All of the icons in the Global Parameters folder have green borders. The program uses this border to show that all of the Parameters under that icon are using the default values that are supplied with the program. If you change the value of any of the Parameters then the green border around the icon will disappear. This allows you to see at a glance which aspects of an analysis are using alldefault values, and which are using changed values.

4


The Materials Tab Section The program is supplied with a set of System Materials that contains full property data for more than sixty materials. However, the Materials tab section does not show icons for all of these materials, but only for materials that have been selected in the input data for the various Models in the Study Folder, or for materials that you have added yourself while working in the Material tab section. PHAST currently only allows you to define Global Materials, and the same set of Materials data will be used in the calculations for all Model. You cannot currently define Local Materials to be used only for the Models in a given Study. There are three types of icon present in the Material tab section of the PHAST Example StudyStudy Folder: Green Icon: a Pure Material

The eight green icons are all pure Materials. Each icon has a green border, which shows that all of the input fields for the material have the values set for that material in the System Materials. You can change the values if you wish - e.g. to enter different probit values for a toxic material – and if you make changes the green border will disappear. All of the icons in the PHAST Example Study Study Folder are for pure materials that are supplied in the System Parameters, but the program also allows you to add your own materials. Yellow-and-Red Icon: a Mixture The yellow-and-red icon is a Mixture, and in the PHAST Example Study Folder it represents the plume of hydrogen chloride, nitrogen dioxide and sulphur dioxide produced by a fire in a pesticide warehouse – which is the situation modelled by the Warehouse Fire Model. This particular Mixture is generated automatically when you run the Warehouse Fire Model, but you can also define your own Mixtures, using any combination of the materials in PHAST, and select these Mixtures for use in the dispersion, fire and explosion calculations. Pink Icon: a Pesticide

The six pink icons are all Pesticides, and are used to describe the contents of the warehouse for the Warehouse Fire Model. Pesticides are only relevant to the Warehouse Fire Model and cannot be selected for any other type of modelling.

5


The Map Tab Section The Map tab section allows you to set up map image and geographic data so that you can view the regions and features affected by consequence results. The map image is defined by the powerstation raster image, and you view the image by selecting Map from the Viewmenu. The Map Window will open in the area to the right of the Study Tree pane, and you can use the options in the Map menu, the right-click menu and the Map toolbar to zoom in and out, to move around in the Map Window, and to control the display of the features of the window such as the scale bar and the legend.

The Map tab section and the Map Window

The Models are represented by dots on the Map. These dots can sometimes be difficult to see and to relate to the individual Models, but there are several options that can make this easier: Changing the Size and Colour of the Dots

Select Map from the Preferences cascade of the Options menu to open the Map Preferences dialog, and then move to the Model tab section. By default the colour is turquoise and the Point Size is 7 pixels, but if you change the colour to blue and the size to 10 pixels as shown, then the dots will be easier to see on the powerstation Map. Displaying the Model Names on the Map

If you move to the Models tab section, select any Model, and then select Labelsfrom the Viewmenu, the names of all of the Models will be displayed on the Map. To hide the names, deselect the Labelsoption. If there is more than one Model at a given location – as with the ChlorineModels and the Butadiene Models – then the names will be superimposed and may be difficult to read, although this will make it clear that there are multiple Models at the location.

6


Pinpointing an Individual Model

If you select a Model in the Study Tree and then select Pinpointfrom the Viewmenu (or press the F4 key), the dot for that Model will become centred in the Map window and will also be highlighted (i.e. displayed in a light turquoise colour). This allows you to locate a specific Model, which is useful if you cannot identify the name for that Model on the Map. You can close the Map Window by selecting Close All from the Windowmenu.

Viewin g Inp ut Data The section describes above introduced the main types of input and their organisation, and this section how to work on the details of thedata input data.

Opening the Input Dialog for t he Chlor ine Rupture Model Move to the Models tab section and double-click on the icon for the Model named Chlorine Rupture . The Vessel/Pipe input dialog will open as shown below. The dialog contains a large number of input fields organised over sixteen tab sections, but many of these fields are relevant only to advanced modelling options (e.g. for a sensitivity analysis), and you will typically only need to supply a small set of input data when defining a Model for use in an analysis, as you will see in the next chapter.

Input Dialog for the Chlorine Rupture Model

7


Getti ng Help on t he Input Data This tutorial does not attempt to describe every item on input data, but the program is supplied with comprehensive online Help. Every input dialog contains a Helpbutton at the bottom right. When you click on this button, the online Help will appear in a separate window, as shown.

The Help Window

The Help Window will be displaying a description of the current tab section, but you can use the links inside the topic and the Contents , Indexand Searchtabs to reach any topic in the Help system and gain a full understanding of the way that the input data will be used in the calculations and the appropriate values that you should set for the hazardous events that you want to model. Most dialogs also have a “What’s This Help” button in the form of a question mark at the right of the title bar. If you click on this button, the cursor will change to a question mark, showing that you are in “What’s This Help” mode, and if you then click on a field in the dialog, a popup window will appear over the field, describing the field and giving advice on setting values, as shown. There are some tab sections that appear in the input dialog for more than one Model. For example, the Material tab section is used for both the Vessel/Pipe Source Model, the User-Defined Source Model and the Bleve Blast Model. The Help is written in order to give full guidance for either Model, so there may be references in the Help to features that are not currently relevant to you. After you have finished exploring the input dialog, click on Cancelto close the input dialog without saving any changes you might have made. If you wish, you can move to the other tab sections and explore the input dialogs for other types of data.

8


Running the C alculations and V iewing t he Results In the Models tab section, select the Example Cases Study, and select the Modelsoption from the Run menu. The program will process the calculations for each of the eighteen Models in turn, performing the calculations for each of the three Global Weathers, and showing the progress through the calculations. When the calculations for a given Model have been completed for all three Weathers, the name of that Model will change from black to blue, which is the colour-coding that the program uses to show that a Model has run successfully and has a complete set of results. The calculations will take several minutes to complete, depending on the speed of your machine. You do not have to run the calculations for all Models and all Weathers. If you select a single Model or folder, then you can run the calculations just for that Model or for the Models in that folder, or you can select Batch/Weather Setup from the Run menu to select Models across different folders or to select only specific Weathers. The selection of Weathers in the Setup dialog will be used for all calculations, but the selection of Models will be used only when you select Batch Run from the Run menu.

Viewing t he Gra phs f or t he Chlorin e and Bu tadiene R eleases Select the Vessels or Pipe Sources folder and then select Graphfrom the Viewmenu, from the right-click menu or the toolbars. A dialog will appear as shown, prompting you to chose the weather conditions whose results you want to view. If youa had selected a single Models, Model rather than folder with multiple then the dialog would have checkboxes next to the Weathers instead of radio buttons, and you would be able to compare the results for several Weathers for that Model. If you choose a single Weather in this situation, then the graphs will have additional features that are not available when you are viewing the results for multiple Models or Weathers. For this example, select the F 1.5m/s Weather. This is the weather with the most stable conditions, and is likely to give the longest dispersion distances. When you click on OK there will be a pause of a few seconds, and then the Graph Window will open as shown in the space to the right of the Study Tree pane.

The Graph Window

9


The Graph Window will usually contain many tab sections, each with a different type of graph. The tab sections included for a particular combination of Model will depend on the type of the Models (e.g. Vessel/Pipe Source Model or Fireball Model), on the type of the materials (toxic or flammable), and on the details of the dispersion and effect behaviour (e.g. whether or not liquid rainout occurs). The Chlorine and Butadiene Models have graphs for cloud concentration, for pool vaporisation, for toxic effects, for jet fire, fireball and flash fire effects, and for explosion effects. The Concentration Graphs

The first graph is of centreline concentration. This will be showing the results at the time at which the cloud footprint covers the greatest area, which occurs at a different time for each weather. The graph will initially be showing results only for the four Chlorine Models. In the dispersion calculations, the program uses an averaging time that takes into account changes in wind direction over the course of the release, to give an average concentration at a given location, and it uses different averaging times for toxic and for flammable materials, reflecting the different time-scales that are relevant to each type of release. The concentration graphs always display results calculated with a specific averaging time, which is displayed in the legend for the graph. The default averaging time for this set of results is the Toxic averaging time, and the Butadiene Models were not modelled with that time so have no results to display. To view the concentration results for the Butadiene Models, you must change the selection of averaging time to display. To do this, select Properties… from the right-click menu or the Graphmenu to open the Plot Properties dialog, and then move to the Averaging Times tab section as shown. If you change to the Flammable Averaging Time, the graph will display the results for the four Butadiene Models only. The User Defined option will also be enabled, which shows that some of the Models have a user-defined averaging time defined in the Location tab section. In fact, all of them have such a time defined, and if you select User Defined as the averaging time for the graphs, the graph will display results for all eight Models. Results Displayed on the Map

After the six tab sections that show the results in terms of concentration, the next tab section is the Map graph, which allows you to view different types of effect zones superimposed on the map. When you first move to the Map tab section, the Map graph will be displaying Cloud Footprint results for a concentration of 10,000 ppm for the Toxic averaging time, and the only results displayed will be for the Chlorine Rupture and Chlorine Liquid Leak Models. The other Chlorine Models don’t produce this concentration level at the default height of ground level – as you can see from the Sideview graph – but if you open the Plot Properties dialog, move to the Distance tab and set the Height to 10 m, results for the Chlorine Vapour Leak Model will also appear in the plot.

10


The Footprint concentration results are the default form of results for the Map graph, but the Event field in the Display tab s ection of the Plot Properties dialog allows you to change to a different form, as shown. The list of types of effect will depend on the types of Models that are covered by the Graph, and will be similar to the range of tab sections in the Graph window. If you select Toxiceffects, then the Radiation/Toxic field will become enabled and you can choose between dose, probit and lethality results. If you view the Lethality footprint on the Map, you will see that the Chlorine Liquid Leak gives the greatest downwind effect distance for lethality. The RuptureModel produces higher peak concentrations at any given downwind location, but the short duration of the r upture means that the total dose received is lower than for the leak. The Map graph initially shows the effect zone with a northerly wind, but you can choose Wind Direction from the Graphmenu or the right-click menu to change the wind direction. The Pool Vaporisation graph does not show any hazardous effect distances, but the Toxic graph and the various Fire and Explosion graphs all include footprint-results of the form shown on the map, and most of them also include graphs that show the effect-level along the cloud centre-line as a function of distance downwind (e.g. radiation level for a jet fire, or lethality for a toxic release). If you look through the Fire and Explosion graphs, you will see that the greatest downwind effect distance is reached by the Late Explosion Worst Case for the Butadiene RuptureModel, which reaches a distance of about 880 m downwind. A late explosion is onedefault that occurs after the cloud has started dispersing away from the which releasemeans point, that and by the explosion is assumed to be centred at the cloud front, the explosion radius will reach beyond the flammable region of the cloud. The program calculates the results for such an explosion at regular intervals, and the Worst Case graph displays the results for the ignition-time that gives the greatest downwind effect distance.

11


Viewing the Re port s for t he Butadiene R upt ure Model The program also presents results in the form of reports. If you wish you can view a report that covers multiple Models – e.g. a report for all of the Chlorine and Butadiene Models – but if you want to compare the report-results for different Models it is easier to view separate reports for each Model and compare between two reports. To view the reports for the Butadiene Rupture Model, select the Model and then select Reportfrom the Viewmenu or from the right-click menu or the toolbars. After a pause of a few seconds, the Report Window will open to the right of the Study Tree pane as shown. The Report Window will probably hide the Graph Window, but you can use the options in the Windowmenu to move between the windows. You can have any number of Graph Windows and Report Windows open at the same time.

The Report Window

As with the Graph Window, the Report Window will normally contain several types of results, presented in different tab sections. A given tab section will present the results for all of the weather conditions that have been processed for the Model. For the Butadiene Rupture Model, the first tab section is the Input tab section, which lists the input data. The Audit tab section gives version details for the program, for parameters and materials, but all of the other tab sections give details of the consequence results that you saw summarised in the Graph window: The Summary Report

This report summarises the maximum downwind distance to different types of effects, and gives a direct comparison between the different weather conditions. For the Butadiene Rupture , D 5m/sis the weather that gives the greatest distances, although the difference between the three weathers is small.

12


The Discharge Report

This gives details of the discharge modelling, and the condition of the release immediately after expansion to atmospheric pressure – which is the condition used for the start of the dispersion calculations. This report and all the other results-reports give the results for each weather in turn. The Summary report is the only report which presents a direct comparison between the different weathers. The Dispersion Report

This report contains a table which describes the location and state of the cloud at a series of time-steps during aspect the dispersion. You might refer to this reportdepth. if you wanted to understand a particular of the dispersion behaviour in greater The Commentary Report

This report highlights the main events in the course of the dispersion, and allows you to see easily if and when differest types of behaviour occurred, e.g. touch-down on the ground, or the rainout of liquid droplets. The Averaging Times Report

The centreline concentrations given in the Dispersion and Commentary reports are all calculated using a “core” averaging time that is set in the Dispersion Parameters and that has a default value of 18.75 s. The Averaging Times report gives the centreline concentrations at a series of steps during the dispersion, calculated using alternative averaging times. For the Butadiene Rupture these alternative times are the Flammable Averaging Time (whose value is set in the Flammable Parameters) and the User-Defined Averaging Time (whose value is set in the Location tab section for the Model). In this analysis both of these times are also set to 18.75 s so for all the Butadiene Models the Averaging Times report gives the same concentrations as the other reports. However, if you viewed the report for one of the ChlorineModels, you would see results for the Toxic Averaging Time (whose value is set in the Toxic Parameters), and which has the default value of 600 s. The Fireball Report

The Fireball report gives radiation results for a fireball resulting from immediate ignition of the released material. The report first gives a description of the fireball flame (emissive power, liftoff height, etc.), then it gives the dimensions of the elliptical effect zones for up to five different radiation levels – where the levels are set in the Fireball tab section for the Model – and finally gives the radiation levels at a series of points downwind from the centreline of the release. The Jet Fire and Pool Fire reports have a similar form, giving the same three types of results.

13


The Early Explosion Report

For Butadiene Rupture , the tab for the Early Explosion report is named Early Expl.(TNT) , and this is because the explosion method selected for this Model is the TNT method. There are three methods available, and you select between them in the Flammable tab section for the Model. The TNT method is the simplest, requiring the smallest amount of input data, and it is the default method. The report is similar in form to the Fireball report, giving the dimensions of the circular effect zones for up to five explosion overpressures – where the overpressures are set in the Explosion Parameters – and also giving the overpressure levels at a series of points downwind from the centreline of the release. The Late Explosion Report

This report gives the overpressure effect distances for late explosions occuring at a range of times during the dispersion. For each ignition time, the report gives the location of the cloud-centre, the location of the centre of the explosion, the downwind distance to up to five overpressure levels, and the flammable mass in the cloud at the time of the explosion. By default the centre of the explosion is taken as the cloud front to 50% of the LFL, but you can change this setting in the Explosion Parameters.

Results for Two Time-Steps in the Late Explosion Report

The ignition-time that gives the greatest downwind effect distance is the one presented in the Worst Case Late Explosion graph, as described in the section above. The range of reports presented for a particular Model will depend on the type of Model and on the behaviour of a release, and there are additional reports that do not appear for the Butadiene Rupture Model. For example, if the material is toxic then there will be a Toxic report with a table of dose, probit and lethality results as a function of downwind distance, and if the liquid in the release rains out to form a pool, then there will be reports describing the spreading and evaporation of the pool and describing the series of “dispersion segments” used to represent the vapour produced from the pool. For most of your work with the program you will probably refer mainly to the graphs, since they present the results in the most direct form and allow easy comparison between different Models and Weathers. After you have finished examining the results, you can use Close All from the Window menu to close the windows.

14


Viewing th e Results for the Chimney Rele ase and L ong Pipeli ne Models The other two Vessel/Pipe Source Models in the PHAST Example Study Study Folder illustrate some of the special modelling features that are available. The Chimney Release Model

This models a release of methane from a chimney stack on top of a building, and takes into account the effects of t he building wake on the dispersion. If you view the graphs for the Model for all three Weathers and then move thewill Sideview graph,toyou see an outline of the building with the chimney on top, and with the three plumes emerging from the chimney. The building wake produces a zone of low pressure, and this pulls the plumes downwards. The model deals with this by adjusting the height at a specific downwind dista nce, which is 100 m in this case. In some situations the plume may be pulled down low enough that all or part of the plume is entrained in the building wake, but that has not occurred for any of the weathers for this Model. The Sideview graph show that plumes never approach closerFire to the ground about 58 m, but if you looks at thethe Explosion graphs and the Flash graph, youthan will see Worst Case Late Explosion distances of over 900 m, and Flash Fire distances of about 600 m to 50% of the LFL. When performing the modelling of late explosions a nd flash fires, the program can calculate the flammable footprint of the cloud either at the cloud centreline or at a specific height. The centreline method is selected by default in the Flammable Parameters since this will give the most conservative results, but you should check the Sideview graph and make a judgement about whether or not the effect zone would actually reach the areas of interest for your analysis. A flash fire in a plume 60 m in the air would not affect people on the ground, but an explosion in such a plume might well produce significant overpressures at ground level. The Long Pipeline Model

This models the rupture of a 250 m propane pipeline that has a pumped flowrate of 10 kg/s, where the rupture occurs 100 m downstream from the pump. The program performs discharge modelling for the complex, time-dependent flow regime inside the ruptured pipeline and then performs dispersion modelling for a representative averaged discharge rate. Select the Model, and view the graphs for the F 1.5m/s weather. For this analysis the discharge calculations are the same for all weather conditions, so you only need to view one weather if you are only interested in the discharge results. The first tab section in the Graph window will be the Long Pipeline tab. This contains a large number of sub-tabs, each of which shows the behaviour of a particular discharge variable against time. Move to the Flowrate sub-tab, since this shows the behaviour of the most important variable. 15


The Flowrate graph appears to show the flowrate dropping instantly from about 230 kg/s to about 10 kg/s, as shown. However, it is difficult to tell whether or not the drop is instant because the default scale on the time axis goes up to nearly a million seconds. To see the initial behaviour in more detail, you must set the scale yourself. Select Scale and Labels from the right-click menu or the Graph menu to open the Scale dialog, then uncheck the option for Automatic Scaling and set the Maximum Time to 60 s. With the changed scale, you can see that the rate takes about 45 s to drop to a steady rate of 10 kg/s, which is the pump rate. There are five lines plotted on the graph, and their meaning may not be immediately obvious. The two A lines describe the 100 m pipe-section upstream of the rupture, the two B lines describe the 150 m section downstream of the rupture, and the Totalline is the sum of the rate released from the two sections. The two Upstrea m lines show the pumped inflow into the section, which is 10 kg/s for Section A and ze ro for Section B, and the two Orificelines show the flow from that section at the point of rupture. If you want to hide any of these lines (e.g. th e Upstream lines), open the Plot Property dialog and deselect the lines in the Long Pipe tab section. For the first nine seconds, the orifice flowrates from both sides are almost identical, as the flash-front travels along each section at a similar speed, giving a similar flowregime. However, at 9 s the flash-front reaches the end of section A, and from this point onwards the pressure profile in that section is maintained at the profile produced by a pumped flowrate of 10 kg/s; the program stops the discharge calculations for Section A at this point which means that there are no results available to display on the graph after 9 s, but the 10 kg/s flow from Section A is added in to the Total, as you can see. If you move to the Distance sub-tab, you can see that the flash-front reaches the end of Section B after 14 s. However, the calculations do not stop for Section B at this point, and proceed to model the depressurisation of the section until it has emptied completely at 45 s.

16


If you view the Discharge report for the Model, you will see that the average rate used to represent the behaviour is 10.5 kg/s, taken over a time-scale of one hour. This may underestimate the hazard produced by the release, and there are options available for representing a time-varying release with more than one “release segment” so that you can investigate the significance of the type of short-term behaviour seen in this release. These options are described in more detail in the next chapter.

Viewing the Re sult s for t he Other Models The other eight Models in the Study Folder are not Source Models. Each models one specific type of behaviour and will produce a fixed set of graphs and reports. The Warehouse Fire Model

This models a fire in a pesticide warehouse and you can define multiple scenarios for each warehouse, where each fire scenario is defined by the surface area of pesticide involved and by the duration of the fire. There are special calculations that determine the release rate and composition for the toxic plume produced by the fire, and the dispersion and effects of this plume are then modelled in the same way as for the toxic cloud for the four ChlorineModels. The Three Flammable Models

The Pool Fire , Fireballand Jet FireModels perform the same type of radiation modelling as that associated with a Source Model, but they give you more control over the definition of the flame and they also allow you to specify in more detail the locations for which you want to calculation the radiation levels. The Four Explosion Models

The Baker-Strehlow , Multi-Energy and TNT Models perform the same type of vapour-cloud explosion modelling as thatofassociated with acloud Source Model, they give you more control over the definition the flammable and of the but results-locations. The BLEVE Blast Model calculates the overpressure levels produced by the rupture of a vessel under flame impingement, which is a type of explosion modelling that is not performed for a Source Model. The form of the results for all of these Models is similar to the corresponding dispersion, toxic, fire and explosion results for a Source Model, and you should find interpreting the graphs and reports very straightforward. You have now seen the main features of PHAST. When you are ready you should proceed to Chapter 2, which takes you through the stages in setting up your own analysis.

17

Chapter 2:Setting up your own Analysis

Chapter 2 Settin g up your o wn Analysis The For m of the Analysis This chapter will guide you through the process of setting up a Study Folder for performing consequence calculations. The tutorial supplies all of the input values that you will need to complete the analysis.

The Models De fin ed in the Analysi s The main aim of the analysis is to show you how you can define Models to represent the most common types of hazardous event, and how to take into account the main variables. The types of hazardous event that are considered in the analysis are as follows: •

A rupture of a vessel containing a toxic material

•

A pipework leak from the l iquid side of a vessel containing a toxic material

•

A pipework leak from the gas side of a vessel containing a toxic material

•

The equivalent three releases for a vessel containing a flammable material

•

The rupture of a propane tank wagon under normal operating conditions.

•

A fireball or BLEVE of the propane tank wagon as a result of fire impingement.

•

A liquid leak from the body of the propane tank wagon.

If you wish, you can omit events, d efine different events, or change the input values in order to define conditions that are more typical of your facility. However, if you do this you will obtain results that are different from those that will be shown in this manual.

Creatin g a new S tud y Folder To create a new Study Folder, select New from the Filemenu or the Toolbar. The program will close the PHAST Example StudyStudy Folder and a new Study Folder will open, with a name shown as “Untitled”.

Savin g the Study Folder You cannot save the Study Folder with the name “Untitled” and should save it wit h a real name immediately. Select SaveAs… from the Filemenu. The File Save dialog will appear and you should locate the DNVuser folder (the default location for saving Study Folder files), use the Create New Folder option to create a folder with your name, and then save the new file to this folder with the name Tutorialand the default file format of *.psu.

18


The Content s of a new S tudy Folder New Study

Folder files are not empty but will have some default data set up:

A Global Weather Folder containing three Weathers

The weathers are the same as those in the PHAST Example Study Study Folder. A Set of Default Parameters

As with the PHAST Example Study Study Folder, there is a set of Global Parameters, all of which are using the default values.

Setti ng up t he Map Dat a The tutorial uses a map of an area near two rivers, in a country which has a national grid system. The image for this map is supplied with the program the form of a *.tif file. If you have an image file for the area around your facility, you might prefer to use that instead.

Insertin g t he Raste r Image Image files that contain a description of each pixel in the image are known as raster images, and most common image files are in this form, e.g. *.tif, *.bmp, *.giffiles. The program can also display map data taken from a GIS Database, where an image is defined by describing the lines that form the image. The process of inserting a raster image into a Study Folder is very different from the process of inserting a connection to a GIS Database. This tutorial deals only with raster images, and you should refer to the online Help for details of working with GIS Databases. The process of inserting the raster images involves several stages. Ensure that there is a Raster Image Set in the Map tab section If the Map tab section does not already contain a Raster Image Set icon, select the Tutorial icon at the top of the tab section, and use the Insertmenu to insert a Set. The Set is a folder for raster images, and you have to insert raster images inside such a folder. Insert a Raster Image inside the Set

Select the Set, then select Raster Image from the Insertmenu. A dialog will appear as shown, and you must browse to locate the image file. The tutorial.tif file is located in the Examples folder for the installation of the program (which is typically under Program Files\DNVS\PHAST_6_5 ). When you first browse to this folder you will not see any files, since the list of File types is not set to *.tifby default. When you have selected a valid raster image file, the Placement Mode fields will become enabled; these are options for specifying the map co-ordinates covered by the image. Some files contain georeference data or header data that you can use to set the co-ordinate data for the image, but the tutorial.tif file does not and the only option available is the Interactive option, which is available for any raster image file.

19


Selecting a Co-ordinate System for the Map

When you click on OK in the Place dialog, a dialog called the “Co-ordinate system wizard” will open; this is the first step in selecting a co-ordinate system for the analysis. It is only essential to select a system if the Placement Mode is set to Georeferenced or to By Header, or if you want to use a GIS database in the analysis. When you are using the Interactive Placement Mode and will not be connecting to a GIS database – which is the situation in this tutorial - you can click on Cancelin the Wizard dialog and leave the co-ordinate system undefined. The Wizard dialog contains a Helpbutton, and this gives you a quick way of viewing an overview of the user and definition of co-ordinate systems in PHAST. Placing the Image in the Map Window

When you click on Cancelin the Wizard dialog, there will be brief pause and the Map Window will then open to the right of the Study Tree pane. The cursor will be in the form of crosshairs, and you must drag and drop to place the image in the window. This sets the initial values for the map co-ordinates for the images, which you will set to the correct values in the next step. Setting the Co-ordinates and Size of the Image

Double-click on the tutorialicon to open the input dialog for the image, move to the Geometry tab section, and set the values shown. The srcin for a map image is the top-left corner, and the values are in the national co-ordinate system for the country. When you click on OK the image will probably disappear from the Map Window because it has moved to a location beyond the scope of the window. Select Fit > Allfrom the Map menu, and the Map Window will change to display the image covered by the image; if the menu bar does not include a Map option, click on the Map Window to mak e sure it is selected, and the Map menu will appear in the menu bar. Setting a Large Number of Significant Figur es for Edit Dialogs

The co-ordinate values for the image will be in the national co-ordinate system for the country, and the values for the area covered by the map are six-digit numbers. By default, input dialogs display only four significant figures of any number that you are editing, and with this setting you will find it difficult to be sure that you have entered the co-ordinates. To change the setting for the number of sign ificant figures, select Preferences > General from the Optionsmenu and move to the Miscellaneous tab. The first field in the tab section is the Number of significant figures for edit windows, and you should make sure that this is set to six or more. Click on OK to close the General Preferences dialog and return to the Map tab section. If you open the dialog for the raster image again, you will be able to see that the values that you entered were stored in full.

20


The Location of the Site on the Map

For the tutorial, the facility occupies the long, narrow section of land to the north and west of The Village, between the east bank of the river and the road that runs parallel to the river, shown shaded yellow in the illustration.

The Location of the Facility on the Map

Definin g th e Firs t Model: for a ToxicRupture In the Tutorial.psu Study Folder, move to the Models tab section. The first Model you will define represents the rupture a vessel containing a toxic material, which is one of several Models dealing with aof toxic material. The vessel is a sphere with a radius of 3.37 m and volume of 120 m3 and a maximum fill-level of 85%, containing chlorine at saturation conditions and ambient temperature. The sphere is located near the centre of the site and is elevated 4 m above the ground. There is no bund surrounding the sphere.

Inser t a Folder to Gro up Toxic Releases Select the Studyicon, then select Folderfrom the Insertmenu or the toolbar to insert a folder. Use Renamefrom the Editmenu or the right-click menu (or press the F2 key), and give the folder the name “Toxic”. You will place all of the Models that represent toxic releases in this folder.

Turn on th e Opti on to Insert Models on t he Map In the Optionsmenu, select the option to Insert Models on Map . By default this option is turned off, and when you insert a Model the icon will appear immediately in the Study Tree. If you turn the option on, then the Model icon will not appear in the Study Tree until you have clicked on the Map to set the location for the Model. In this tutorial you will insert the Models on the Map in approximately the correct location, and then correct the location as necessary in the input dialog.

21


Insert a Vessel/P ipe Sour ce Model Select the Toxicfolder, then select Vessel or Pipe Sourcefrom the Insertmenu or the right-click menu. The Map window will open if it is not already open and the cursor will turn to crosshairs., and you should click at a point near the centre of the site as shown to place the Model. After you have clicked, an icon will be added to the Study Tree, and a dot will appear on the Map to show the location of the Model. Rename the icon to Cl2 Rupture . The icon will have a red border around it, showing that it does not have a full set of input data. You will not be able to run the consequence calculations for the Model until you have supplied values for all of the mandatory input fields, as will be described below. You use the Vessel/Pipe Source Model when you want to perform dispersion and effects calculations for a release from containment and you want to use the program’s in-built discharge calculations to determine the state of the material after expansion to atmospheric pressure, which is the state required for the start of the dispersion calculations. The program contains a second Source Model which is calle d the “User Defined Source Model”. This Model does not perform discharge calculations, but instead allows you to specify directly the state of the material after expansion to atmospheric pressure. You use it if you want greater control over the inputs to the dispersion and effect calculations, as will be described later in this chapter.

Setti ng t he Input Data Double-click on the icon for the Model to open the input dialog. All of the fields in the first tab section are blank, and those that are enabled have red borders . A field with a red border is a man datory field: you must supply a value for such a field, and you will not be able to ru n the calculations for a Model that has any mandatory fields unset. This section describes each tab section in turn, including those that are not relevant to this particular hazardous event. Click on the Helpbutton to open the online Help if you want further information at any point. The Material Tab Section

To set the Discharge Material, click on the button with three dots to the right of the Discharge Material field, and select CHLORINE from the list that appears. The list contains all of the materials that are defined in the System Materials. The vessel is a sphere with a volume of 120 m3. This Model will repr esent the vessel with the maximum degree of filling, which is 85%. Select Volume as the method of specifying the Inventory, and enter a value of 102 m3.

22


The chlorine is held under saturation conditions at atmospheric temperature. The temperature will vary depending on the season and time of day, but for this Model a value of 10oC will be used as representative. To set these Process Conditions, choose Saturated Li quidfrom the first dropdown list and Temperature from the second dropdown list, and set the Temperature to 10 degC, as shown. When you move the cursor away from the Temperature field the program will calculate the saturation pressure for this temperature and display it in the Pressure field. To define the process conditions for a material that is not held under saturation conditions (e.g. a gas or a padded liquid), you must select both Temperature and Pressure from the lists and give values for both. The Scenario Tab Section

You use this tab section to specify the type of hazardous event you want to model. The range of types available will depend on the process conditions you have specified. There is only one Scenario Type available for modelling the rupture of a pressurised vessel; this is Catastrophic Rupture, which is selected by default. The other scenarios are either longer-duration releases, or applicable only to insulated tanks. The vessel is out of doors, so you can leave the Outdoor / In-Building fields with the default selection of Outdoor. If you select In-Building Release, the program will model the build-up of concentration inside the building and th e dispersion calculations will start with the state of the plume as it is released from the ventilation system. The other fields in the tab section are not relevant to a rupture scenario. You can take the default settings for all of the fields in this tab section. The Pipe Tab Section

All of the fields in this tab section are disabled when the scenario is set to Rupture. They are relevant only to the Line Rupture, Disc Rupture, Relief Valve and Long Pipeline scenarios, as you will see later. The Vessel Tab Section

All of the fields in this tab section are disabled when the scenario is set to Rupture. For all of the other scenarios, some of the fields in the tab section will be enabled, with the combination depending on the scenario as you will see later. 23


The Location Tab Section

First, set the release coordinates. The Elevation has a default value of 1 m, taken from the System Parameters, but you should set this to 7.37 m, which is the elevation of the centre of the sphere above the ground. Set the East co-ordinate to 198492 m, and the North co-ordinate to 435063 m. The program requires a criterion for stopping the dispersion calculations: either a maximum distance, or a minimum concentration. For this tutorial, set the Concentration of interest to 100 ppm. When you set this concentration, the Uses averaging time field below the concentration will acquire a red border, showing that it is mandatory; you must specify the averaging time to be used in the calculations for stopping dispersion. Forassociated a toxic release, allows to choose the Toxic averagingthe time or the times with the the list ERPG , IDLHyou or STEL measures of toxicity, or to specify a User-defined time. For this release, select th e Toxic averaging time, which is set in the Toxic Parameters and has a default value of 600 s. The Location tab section allows you to select additional averaging times for which you want concentration values. If y ou make any selections in the final section of the tab, the results will be appear in the Averaging Times report, as you saw in the previous chapter. The Bund Data Tab Section

If there is a bund around the vessel and you want to take this into account in the modelling of pool-spreading and evaporation, you can check the Bund exists box and enter a description of the bund. For this sphere there is no bund, so you can leave the tab section with the default values. The Indoor/Outdoor Tab Section All of the fields in this tab secti on are disabled then the scenario is a catastrophic rupture outdoors. Some of the fields are enabled for the longer-duration scenarios as you will see later, while others are enabled for in-building releases. Flammable Tab Section

The fields in this tab section are disabled when the material is toxic only. For a flammable release, they allow you to choose between the three models for a vapour cloud explosion, and to choose between two models for jet fires. The Toxic Parameters Tab Section

The fields in this tab section are used in modelling the buildup of toxic concentration inside a building, and the exposure of a person inside the building. By default, these calculations are set to Unselected (i.e. they will not be performed), but for this tutorial you should change them to Selected . The calculations require information about the ventilation-rate for the building and about how long people remain in the building after the cloud has passed and the concentration is lower outdoors than indoors. By default these values will be taken from the Toxic parameters tab section for the Model, Ventilation Specification but if you choose WindSpeed Dependent for the , then the values will be taken from the data for the Weather, which means that the values may be different for each weather.

24


For this tutorial, leave the Ventilation Specification with the default value of Case Specified , and take the default values for the Building exchange rate and the Tail time. The TNT, Multi Energy and Baker Strehlow Tab Sections

The fields in these tab section are disabled when the material is toxic only. They are used in the modelling of a vapour cloud explosion. The Discharge Parameters Tab Section

The fields in this tab section are always enabled, and take their default values from the System Parameters. They are used in the discharge modelling for the Line Rupture, Disc Rupture and Relief Valve scenarios, so are not relevant to this Model. The Jet Fire, Pool Fire and Fireball Tab Sections For a flammable release, these tab sections allow you to choose between options for modelling each type of flame. A Summary of the Input Data

The input process involves examining a large number of input fields, but the number of values that you have to enter in order to complete the data for this Model is small, as shown in the table below: Tab Section Material

Location

Toxic parameters

Input Field Discharge Material Inventory Process Conditions Elevation East Co-ordinate North Co-ordinate Concentration of interest Uses averaging time Indoor Toxic Calculations

Value Chlorine 102 m3 Saturated Liquid at 10oC. 7.37 m 198492 m 435063 m 100 ppm

Toxic Selected

The default scenario for a Vessel/Pipe Source Model is a catastrophic rupture o ut of doors, so there is no need to change any settings in the Scenario tab section for this particular Model. If you have made all of these set tings, the input data for the Model are now complete, and you can click on OK to close the dialog. You should see that the icon no longer has a red border, showing that it has a full set of input data.

25


Run the Calcul ations and View the Re sult s Select the Model and select Run Model from either the Run menu or the toolbar. When the calculations are complete, view the graphs for all of the weathers. You will see that there is no Pool Vaporisation tab in the Graph Window, which means that the liquid in the release did not rain out; if you want more information about the behaviour of the liquid droplets in the cloud, you should view either the Commentary Report or the Dispersion Report. The concentration graphs only ever show the outdoor concentration, but if you move to the Toxic tab section you will see that the Probit, Lethality and Dose graphs display separate results for indoor and outdoor effects, and that there are separate Footprint graphs for outdoor and indoor effects. The L ethality graph shows that the greatest downwind effect distance is for the F 1.5 m/s weather outdoors, with a distance of about 2.5 km to a lethality level of 10%. The indoor effects for this weather reach about 2.25 km to 10% lethality. The shortes t downwind effect distances are for D 5 m/s indoors, which reaches about 1.3 km for a lethality level of 10%.

Definin g th e Second Rele ase: Toxic Liq uid fro m Pipework The second release is from the same chlorine sphere, but the hazardous event is the rupture of a one-inch liquid line attached to the bottom of the sphere, where the initial liquidh ead willbe4.6m . The line runs 4 m verticallydownwardsto10cmfromthe ground, then 5 m horizontally to an isolation valve; the rupture is assumed to occur just before the isolation valve.

Copy the First Model Much of the input data f or the vessel ruptu re is also applicable to the pipework failure, so you can use copy andpaste from the Edit menu or the right-click menu to create a copy of the Rupture M ode l, also in the Toxicfolder. Give thecopythename Cl2 Liquid Pipework .

Setti ng t he Input Data Openth e input dialogan d set the inputdataasfollows: Material Tab Section

Leave this tab section with the same values as for the rupture, since the material and process conditions are the same as for the rupture. Scenario Tab Section

Set the Scenario Type to Line Rupture, and the Phase to be Released to Liquid. The line rupture scenario models the full-bore rupture of pipework attached to a vessel, and the discharge calculations take into account the effect of friction in the flow from the vessel to the point of rupture. To model a release from the body of the vessel, with no frictional losses in the discharge, you would choose the Leak scenario. When the vessel contains saturated liquid, you will be offered a choice of release-phase for the line rupture scenario: a vapour release from the top of the vessel, or a liquid release from the bottom of the vessel. The list of phases includes “two-phase”, but this is only enabled for the disc rupture and relief valve scenarios, for modelling overfilling of the vessel.

26


Pipe Tab Section

The Pipe Length is the length of pipework between the vessel and the point of rupture, and you should set it to 9 m as shown. To set the Internal Diameter to one inch, click on “mm” to the right of the field, and then select “in” from the list of units that appears as shown. You can then enter the diameter directly in inches, rather than having to perform the conversion yourself into the default unit of mm. Leave the pipe roughness with the default value taken from the System Parameters. The number of valves is used in the modelling of frictional losses, and you can lea ve them as zero. The other fields in the tab section are relevant only to the long pipeline scenario, and are all disabled for the line rupture scenario. Vessel Tab Section

For the line rupture scenario and most of the other scenarios that involve a continuous release, the Time Varying Release option will be enabled in the Vessel tab section. If you do not check this option, then the release will be modelled with the initial release rate, and the duration will be the time required to drain the inventory at this initial rate. This will normally give conservative results in the consequence calculations. If you select the time-varying option, then you must supply information about the dimensions of the vessel. The discharge calculations will model the effect of the release on conditions in the vessel and the way that these conditions and the release rate change over time, and will represent these time-varying results either with a single rate (e.g. an average rate, or a rate at a particular time) or with a series of rates, depending on your selection for the Rates versus time. For this release, you will perform an initial run of the discharge calculations with t he time-varying modelling selected, then examine the results and decide on the most appropriate way to represent the behaviour for the rest of the consequence analysis.

27


Set the Liquid Head to 4.6 m, select the Time Varying Release option, set the Tank Type to Spherical, the Height of Discharge to zero, and the Diameter to 6.74 m. Leave the Rates versus time set to the default selection of Average rate with an averaging time of 3600 s; you can return to make a final selection after you have viewed the discharge results. Location Tab Section

Set the Elevation to 0.1 m. With this setting, the liquid droplets will probably not evaporate inside the cloud, and will probably rain out and form a vaporising pool. Leave the other fields with the same values as for the rupture. In reality, the releaselocation would be offset by a few metres from the centre of the sphere but this difference is insignificant compared with the effect distan ces for chlorine and can be ignored Bund Data Tab Section

Leave this unchanged, with no bund specified. Indoor/Outdoor Tab Section

For a continuous release scenario such as line rupture you must specify the Direction of the release. Cho ose Horizontal from the list, which is the correct setting for this type of unobstructed rupture of horizontal pipework. The list of directions includes a second horizontal option: Horizontal Impingement . You should select this option if the release is in a congested area and the release is likely to impinge on a wall or other equipment; the program will reduce the momentum of the release, which will reduce the amount of air mixed into the jet during the initial stages. Discharge Parameters

There is one bend in the 9 m of pipework, so you can set the Frequency of Bends to 0.11 per m. This completes the input data for this stage, and you can click on OK to close the input dialog.

Runn ing t he Disc harge Ca lcu lations Select the Model and then select Run Discharge from the Run menu, the right-click menu or the toolbar. This will run the discharge calculations alone, without peforming the dispersion and effects calculations. The calculations may take several minutes, depending on the speed of your machine. When the results are complete, view the reports and move to the TV Discharge Report. The rate drops by less than 3% in two hours of release, so the time-varying behaviour can be ignored for this release. There are two options for bypassing the time-varying discharge modelling in this sit uation: 1: Use the Averaged Discharge Re sults to Create a User-Defined Source Model When you performed the discharge calculations, the program calculated the average rate over the first 3600 s, and this is the representative rate given in the Discharge Report. If you decide that you want to use this average rate rather than the initial rate, you should select the Model, then select Create Source from the Editmenu or the rightclick menu.

28


The program will show a list of the weather conditions for which you performed the discharge calculations and for which it has results, and when you select one of these weathers the program will create a User-Defined Source Model with the name Calculated Discharge , as shown. The User-Defined Source Model has many of the same tab sections as the Vessel/Pipe Model, but instead of the Scenario and Vessel tab sections it has a Discharge tab section in which you specify the discharge rate and conditions directly, since the UserDefined Source Model does not perform any discharge modelling itself. The Calculate d Discharge Model will be created with Discharge data taken from the averaged results from the Liquid Pipework Model, but you can edit these values if you choose. 2: Edit the Model and Deselect Time-Varying Release

This is the simplest method for bypassing the time-varying discharge modelling if you decide that you want to use the initial rate to represent the entire release, and this is the method that will be used for this tutorial. The discharge calculations for this Model will run much more quickly with the time-varying option turned off. After this adjustment, the final set of input data for this Model can be summarised as follows, not including the values that are the same as those for the rupture model: Tab Section Scenario

Pipe

Vessel Location Discharge Parameters

Input Field Scenario Type Phase Released Pipe Length Internal Diameter

Value Line Rupture Liquid 9m 1 inch

Time-Varying Release? Tank Head Elevation Frequency of Bends

Not selected 4.6 m 0.1 m 0.11 per m

The default direction for a line rupture scenario is Horizontal , so there is no need to change any settings in the Indoor/Outdoor tab section for this particular Model.

Run the Consequence Ca lcu lation s and View the Re sult s Select the Model and select Run Model from either the Run menu or the toolbar. When the calculations are complete, view the graphs for all of the weathers. You will see that there is a Pool Vaporisation tab in the Graph Window, which means that the liquid in the release did rain out. If you view the reports and look at the Commentary Report, you will see that rainout fraction is only about 1%, so the formation and behaviour of the pool w ill have little effect on the dispersion or toxic effects. In the Toxic Lethality graph, the greatest effect distances are for the F 1.5 m/s weather outdoors, with a distance of 900 m to a lethality level of 10%, which is approximately a third of the distance reached by the catastrophic rupture. The least stable night-time condition, D 5 m/s, reaches only 350 m for 10% lethality outdoors.

29


Defining the Third Mode l: Toxic V apour from Pipework The vapour release is the rupture of a two-inch pipe attached to the top of the sphere. The line runs 3.4 m horizontally, then vertically downwards, and the rupture is assumed to occur 1 m from the ground. Create the Model as a copy of the Liquid Pipework Model, rename the copy to Cl2 Vapor Pipework , and change the input data as follows: Tab Section Scenario Pipe

Location Indoor/Outdoor

Input Field Phase Released Pipe Length Internal Diameter Elevation Direction

Discharge Parameters

Frequency of Bends

Value Vapour 13 m 2 inch 1m Down – Impinging on the Ground 0.08 per m

When the phase is set to Vapour in the Scenario tab section, the Building Wake Effect fields will become enabled. The sphere is in an open area so building-wake effects are not relevant to this release, and you can leave these options unchecked. The release rate from the two-inch vapour line is similar to that from the one-inch liquidli ne, and the twopipeworkr eleases give ve rysimilare ffect distances.

Defin in g Thr eF e lammabl e Rel eases There is a propane sphere at the far

north of the site. The prop

anehere sp has the same

dimensions as the chlorinesphere a nd samepheric design pip of ature. ework, and salsoi operating under saturation conditio nsatthe atmos temper

Setti ng the Input Da ta for the Mode ls Youcan define the rupture andthe two pipework failuresby copyingeth threetoxic Models and simply changing the selection of discharge material and the eastern coordinates. Copying the Models

Select the Toxicfolder, copy and paste it, and name the copy Flammable . In the name for each Model, change Cl2 to C3. Changing the Material Selection

Open the input dialog, click on the button with three dots to the right of the Discharge Material field, and change the selection from CHLORINE to PROPANE . The list of materials is arranged alphabetically, and you can move quickly to PROPANE by clicking in the list and then typing “P”, which will take you to the first material with this initial letter. When you return to the Materials tab section you will see that the program has recalculated the saturation pressure at 10oC and also the mass for the inventory. You must make this change for each of the three Models.

30


Changing the Location and Concentration of Interest

When you move to the Location tab section, you will see that the Toxicaveraging time is no longer set for Uses averaging time and that this field is now shown as unset and mandatory. The material is flammable only so the Toxicaveraging time is not included in the list, and the program is prompting you to make a different selection for the calculations of the stopping-concentration. For a flammable release you would not wantto calculate theconcentratio n to a value as low as 1 00 ppm, since the cloud willnot pose a hazardonceithas diluted below he low You er flammable limit2% of or 20,000 tppm. could set this concentration yourself, but for a flammable release you c analsole ave the Concentration of interest blank, as shown, and the program willauto matically stop the dis persion calculationsoncethe concentration has reached a given fraction of the LFL as calculated with the Flammable averaging time. By default the fraction is 50%, but you can change this in the Flammable Parameters if you prefer. For this tutorial, delete the value for the Concentration of interest, and set the East and North coordinates as shown above. You must make this change for each of the three Models.

Setti ng th e Input Data for th e Fire Modelling If you move to the Jet Fire, Pool Fire or Fireball tab sections, you will see that three levels radiation are specified, butcalculations that the calculations forcted radiation d ose, probit of and lethalityintensity are all unselected. These are not sele by default because they can be time-consuming, so you would normally only select them if you know that you need them for a particular analysis or a particular Model. For this tutorial you will set the lethality calculations to Selectedand specify five levels of lethality – but you will do this in the Parameters instead of in the Model data, since this is the most efficient method if you want to set values that will be used by all Models. Move to the Parameters tab in the Study Tree, open the input dialog for the Jet Fire Parameters, and set the Radiation Lethality data as shown. After you have clicked on OK to close the dialog you will see that the Jet Fire Parameters icon no longer has a green border, which shows that not it is not using the values set in the System Parameters for all of the fields. Next, open the input dialog for the Pool Fire Parameters, and set the same values for Radiation Lethality. Finally, open the input dialog for the Fireball and BLEVE Blast Parameters, and set the same values.

31


If you move to the Models tab of the Study Tree and look at the Jet Fire, Pool Fire or Fireball tab sections for any of the flammable Models, you will see that the lethality calculations are now selected, with the five levels set.

Runn ing the Consequence Calculations and viewing the Results Select the Flammable folder and use the Run Models option to run the calculations for all three Models. You can also view the results for all three Models at once. Select the Flammable folder and then select View Graphs . A Plot Setup dialog will appear, prompting you for the Weather for which you want to view resul ts. When you are viewing results for multiple Models you can only choose a single Weather, so the Weathers have radio butt ons beside them, whereas they have check boxes beside them when you are viewing results for a single Model. Select the 1.5/F Weather, which should give the greatest effect distances for dispersion. The Graph Window contains tab sections for Concentration graphs, as with the toxic Models, but it contains Jet Fire, Fireball and Flash Fire tab sections instead of the Toxic tab s ection. The propane releases do not produce any liquid rainout, so there are no Poo l Fire tab sections. The main features of the graphs are described below. Jet Fire Graphs

The Jet Fire tab section contains three graphs, which are presenting results for the two pipework failures. The first graph shows radiation level versus distance, the second shows Intensity Radii to the lowest of the three radiation levels set in the Parameters (4 kW/m2), and the third graph shows Lethality Radii to a lethality level of 1%, which is the lowest of the five lethality levels that you set in the parameters. The maximum downwind effect distance shown in these graphs is just less than 25 m, which is the distance for 4 kW/m2 for the liquid release. If a given Fire Radii graph is showing results for more than one Model or more than one Weather, then it will only plot a single level, which will be the lowest level set for that type of result (e.g. the lowest intensity level, or the lowest lethality level). If you want to see results for all of the levels, then you must view the graphs for a single Model and Weather. Fireball Graphs

The Bleve (or Fireball) tab section also contains three graphs. These are showing results only for the rupture, and this m eans that the two Radii graphs are able to show the results for more than one level. The maximum downwind effect distance is about 560 m, to a radiation level of 4 kW/m2, and the distance to a.lethality level of 1% is about 290 m. There is no ellipse for a lethality level of 100%, because the fireball does not produce the necessary radiation dose at the height of interest (set to ground level in the Flammable Parameters).

32


Explosion Graphs

The two Early Explosion graphs contain results only for the Rupture , since immediate explosions are assumed not to occur for continuous releases. However, the Late Explosion graphs contain results for all three Models. The Late Explosion Worst Case graph shows the effect radii for the explosion-time which gives the greatest downwind distance for the lowest overpressure set in the Explosion Parameters (0.02 bar), and the le gend for the Late Explosion Time graph gives the time at which the worst-case explosion occurs. The greatest downwind effect distances is 1,100 m, for the Rupture , and it occurs at 11.2 s. Flash Fire Graph

The Flash Fire Graph shows the zone for the c loud at the time that it covers the maximum area. For the rupture, this gives a maximum downwind effect distance of 350 m to 10,000 ppm, whereas for the two pipework releases this gives a distance of about 70 m to the same concentration. 10,000 ppm is 50% of the LFL, which is the fraction set by default in the Flammable Parameters as the boundary of the flash fire effect zone.

Al ter native Methods for Modelling Ea

rl y Expl os ion s

When you were setting the input data for the flammable Models you left the Flammable tab section with the default settings, which means that the early explosion for Rupturewas modelled with the default method, which is the TNT method. In this section you will create versions of the RuptureModel that use the other methods for modelling early explosions, and compare the results. Creating the Model Icons

Insert a folder inside the Flammable folder, and name it Rupture . Drag the Ruptur e Model inside this folder and then create two copies of the Model. Rename the srcinal Model TNT, name the first copy Multi-Energy , and the second copy Baker-Strehlow . Setting the Inputs for the TNT Explosion Method

For the TNT Model, move to the TNT tab section to check the input data for the modelling. You can leave the Explosion Efficiency with the default value, but for this Model you should set the location to Ground burst , which means that you are assuming that the explosion is sufficiently close to the ground that there will be reflection effects in the pressure waves. Click on OK to close the dialog for the TNT Model. Setting the Inputs for the Multi-Energy Explosion Method Open the input dialog for the Multi-Energy Model, move to the Flammable tab section, and choose TNO Multi-Energy as the Explosion Method.

Next, move to the Multi Energy tab section, where you can define up to seven regions of confinement within the cloud and also specify the strength of an explosion in the unconfined regions of the cloud.

33


By default there are no confined regions selected, which means that there are no mandatory fields in the tab section and that the Model will run even if you do not set any values in the tab section – but it also means that by default the Model will not produce any explosion results. For this tutorial you will define three regions of confinement, each occupying 30% of the volume of the cloud, and with a range of confinement strengths between 6 and 8, as shown. Values of 8 and 9 are typically used for process units, but the region around the propane sphere is relatively open. The strength of an explosion in the unconfined region of the cloud will be 2, as shown. Click on OK to close the dialog for the Multi-Energy Model. Setting the Inputs for the Baker-Strehlow Explosion Method

Open the input dialog for the Baker Strehlow Model, move to the Flammable tab section, and choose Baker Strehlow as the Explosion Method. Next, move to the Baker-Stehlow tab section. This tab section contains many mandatory fields, and you must complete this tab section before you can run the Model. For this tutorial, use the option to have the program calculate the speed of the flame (rather than supplying it yourself). For a propane release you should set the Material Reactivity to Medium , and for this release you should set the number of dimensions for the Flame Expansion to 2, and the Obstacle Density to Medium , as shown. The release is relatively close to the ground and there is likely to be some reflection of the pressure-waves off the ground, so you should set the Ground Reflection Factor to 1.6. Finally, the volume of the cloud assumed to be involved in the explosion is 500 m3. Click on OK to close the dialog for the Baker-Strehlow Model. Running the Calculations and Viewing the Results

Select the Rupturefolder, run the calculations, and then view the graphs for the 1.5/F Weather. In the Early Explosion Distance graph, the Baker-Strehlow Model has the highest peak overpressure, of about 1.02 barg, but the pressure declines rapidly with distance and there are no effects beyond about 300 m. The TNT Model produces a peak pressure of 1 barg and the pressure declines less rapidly with distance, so the pressure at 300 m is 0.2 barg, and there are effects out to 1,400 m. For the Multi-Energy Model, the graph shows results only for the unconfined region of the cloud, for which the peak overpressure is only about 0.02 barg.

34


However, in the Early Explosion Radii graph the results shown for the Multi-Energy Model are those for the worst case, and in this comparison the Multi-Energy Model gives the greatest effect distances of the three Models, with a distance of about 2 km to 0. 02 barg. If you view the graphs for the Multi-Ene rgy Model on its own and select only the 1.5/F Weather, you will be able to see separate Early Explosion Distance results for each of the regions in the cloud. These results show that the over-pressure levels close to the release are very strongly dependent on the value that you set for the strength of confinement. This analysis shows that, for this release, the default TNT method gives results that are close to the multi-energy results wit h a medium strength of confinement (i.e. with a strength of 7). It seems rea sonable – and simplest - to take the default method as representative for this analysis.

Flammable Releases from a Rail Tank Wagon The propane is delivered to the facility by tank wagon from a marshalling yard 10 km to the north. The deliveries take place once a week, involving two tank wagons, and are always during the day and never at night. The wagons are 10.6 m in length, 2.6 m in diameter with a volume of 54 m3, are raised 0.5 m above the ground, and are delivered with a fill-level of 80%. The propane is under the same conditions as in the sphere: under saturation conditions at atmospheric temperature (taken as 10oC). There are many hazardous events that could be modelled for the tank wagons, including leaks during the unloading process. This tutorial will consider only the rupture of a wagon under normal operating conditions, a leak from the liquid side of a wagon, and a fireball produced by catastrophic ru pture of a wagon under flame impingement. All events are assumed to occur whi le the wagons are at the unloading point 100 m south of the propane sphere.

Defin ing t he Ruptu re of the Wa gon First, create a folder and name it Tank Wa gon, and then copy the TNT Model from the Rupturefolder, which you will use as the starting point for defining the release. Name the Model Wagon Rupture . Open the input dialog and set the data as follows: Tab Section Material Location

Input Field Inventory Elevation North co-ordinate

Value 43.2 m3 1.8 m 435581 m

Defin ing t he Leak fro m the Liqui d Side of a Wagon Copy the RuptureModel and name the copy WagonLiquid Leak , and then open the input dialog and set the data as follows: Tab Section Scenario

Vessel Location Indoor/Outdoor

Input Field Scenario Type Hole Diameter Tank Head Elevation Direction

Value Leak 1 inch 1.95 m 0.5 m Down – Impinging on the Ground

35


For a release from the body of a vessel rather than from attached pipework, you should set the Scenario Type to Leak. This will give a larger discharge rate since there are no frictional losses during the flow to the leak-location. For the leak scenario, you specify the leak-size in the Scenario tab section. The leak is assumed to be at the bottom of the tank, which is the most conservative assumption for the tank head and the duration.

Defin ing t he Fireball Failure under Flame I mpin gement The program allows you to model immediate-ignition effects from fireballs and poolfires on their own, separated from any modelling of dispersion and delayedignition effects, and you do this by using the Fireball Model or Poolfire Model rather than the Source Models. Select the Tank Wagon folder, then select the option to insert a Fireball Model. Name the Model Wagon Fireball , then open the input dialog and set the data as follows: Tab Section Material

Fireball Shape Radiation Data

Input Field Material East Location North Location Burst Pressure Released Mass Mass Vapour Fraction Radiation vs Distance Maximum Distance Angle from Wind Height above Origin Radiation Ellipse Incident Radiation

Value PROPANE 197327 m 435581 m 8.57 barg 22.2e3 kg 0.25 Selected 500 m 0 degrees 0m Selected 4 kW/m2

The Burst Pressure is 60% greater than the normal operating pressure and is used in calculating the surface emissive power of the fireball. The Fireball Shape tab section gives you the choice between using a correlation to obtain the radius, duration and emissive power, or entering your own values. For this Model, you are using the correlation. The dialog also contains a Contour Data tab section that allows you to define a plane and up tothreer adiation levelsforwh ich you wantcontourresults.

Running the Calculations and Viewingthe Results Run the calculations for the Tank Wagon folder and then view th e graphs for the 1.5/F Weather, and then examine the Bleve or fireball results. The FireballModel gives slightly larger effect distances than the RuptureModel, with a distance of about 460 m to 4 kW/m2 compared with 440 m. This shows the effect of the higher vesse l pressure used in the Fireball Model t o model failure under ame fl impingement, whereas the RuptureModel con sidersa rupture under normal operating conditions which then has a probability of ig nitingimmediately and giving fireball effects.

36


Savin g the Study Folder with

the Re sults

You have now completed the tutorial, and you should save the Study Folder in order to save the changes you have made. By default, the program will only save the input data, which means that the next time you open the Study Folder, you will have to rerun the calculations in order to view the full results. However, if you select the Save As… option from the Filemenu, the Save As dialog will contain an option to Save results as well as you r input data. If you select this option, the program will save the full set of consequence results and you will be able to view the results immediately the next time you open the Study Folder – although you should be aware that the file may be large, e.g. 25 MB or more.

What Ne xt? This tutorial has not covered every feature of the program, but you should now have enough of an understanding of the approach and methods used in the program to be abletoexploretheremainingfeaturesyourself,with the assistance of the online Help . If you need further detai ls on any aspect of the program, or if you need guidance on how to model a particular situation for your facility, you should contact product support using the details given underduct Pro Suppo rt in the Helpmenu.

37

DNV SOFTWARE

PHAST Getting Started Manual Chapter 1: A Quick Tour of the Main Features

PHAST Getti ng Started Manual Chapter 1: A Quick Tour of the Main Features

Table of Contents Introduction......................................................................................................................... 1 Starting PHAST .................................................................................................................. 2 The PHAST Window.......................................................................................................... 3 Opening an Example Study Folder ..................................................................................... 4 The Study Tree Pane........................................................................................................... 5 Viewing Input Data............................................................................................................. 9 Running the Calculations.................................................................................................. 17 Viewing the Results .......................................................................................................... 18 Saving the Example Study Folder..................................................................................... 23

Introduction This manual will lead you through the main features of PHAST and PHAST Micro, by opening a pre-defined case so that you can view input data, run calculations and examine the results.

All Ex amp les are fro m PHAST Micro There are two versions of PHAST, and this manual covers both of them. PHAST Micro is the simpler version, containing DNV’s sophisticated dispersion modelling in full, but with some limitations to the options in other areas of the modelling. PHAST is the fullyfeatured version, offering control over most aspects of the modelling, and including stand-alone versions of the fire, explosion and pool vaporization models that are built into the integrated dispersion modelling. All of the examples in this chapter are based on PHAST Micro and are fully applicable to that version. If you are using PHAST, you will see some features in your program that do not appear in the illustrations and are not described in the text. Instruction on these features are given in the PHAST Introductory Training Course. For details on the PHAST training courses please contact your local Technical Support centre.

Page 1 of 25

Startin g PHAST When you install PHAST, the installation routine places a DNV folder under Programs in your Start menu, and you can start PHAST running by selecting the icon from the folder.

The start menu

The installation routine also places a PHAST icon on the desktop so you can also start

Page 2 of 25

The PHAST Windo w When you start PHAST running, the PHAST Window will open, as shown:

The PHAST Window on Startup

The window opens with no Study Folder loaded (a “Study Folder” is a file that contains the definition of a collection of consequence modelling calculations and you must open or create a Study Folder file before you can perform any modelling work with PHAST). In the Message Log, the program reports on the security checks, with either “Security OK” or “Security failed”. If the security has failed, you will not be able to save any changes to input data or run any of the calculations, although you will be able to view the features of the program that do not involve calculations. If you have problems with your security please contact your local Technical Support office. This manual assumes that the security has been set up correctly.

Page 3 of 25

Openin g an Example S tud y Folder In this chapter, you will open one of the Example Study Folders that are supplied with the program for a quick introduction to the terminology and approach used in PHAST. In the next chapter you will create a new Study Folder and perform a simple “worst case” analysis. To open an Example Study Folder, choose Open Example from the File menu. A dialog will appear as shown, listing the Example Study Folders supplied, each of which has the file extension PSU. Choose the Study Folder called Example, which is one of the simplest, supplied. When you click on Open, you will be returned to the PHAST window. Some messages will appear in the Message tab section in the “Log Window” pane along the bottom of Choosing the Example Study Folder the window, reporting on the process of opening and checking the Study Folder, and then the “Study Tree” pane will open along the left side of the window, showing the structure of the Example Study Folder, as in the illustration below.

The PHAST Window with a Study Folder Open

Page 4 of 25

The Study Tree P ane The Study Tree Pane allows you to organize and edit the values that are used in the calculations. It appears along the left side of the window whenever you have a Study Folder open, as shown in the illustration on the previous page. If you close the Study Folder, the pane will disappear. The inputpane data:contains a number of tab sections, each of which covers a different type of

Models Tab Section You use this tab section to add “Models” to the Study Folder, where each Model represents a different hazardous release for processing through the consequence modelling. The illustration on the previous page shows the eight models in the Example Study Folder; the first four represent different release scenarios for a chlorine vessel, and the last four represent the equivalent scenarios for a butadiene vessel.

This tab section contains a tree with several levels. The top level represents the entire Study Folder, with the name Example. If you click on the icon for the Study Folder, you will see that the red “Study” icon becomes enabled in the Toolbar:

You can use this icon to add a Study to the tree, and this can be useful if your Study Folder contains hundreds of models and you want to organize them in different groups. For a simple Study Folder such as the Example Study Folder, a single Study (also named Example) is sufficient.

The Study is the second level of the tree, and each new Study Folder is always created with one Study already defined, since each model must be assigned to a Study. If you click on the icon for a Study, you will see that the Folder icon and the blue Vessel and Pipe Source Model icon become enabled in the Toolbar:

You can use the Folder icon to organize models within a Study, and you can have multiple levels of Folders; the simple Example Study Folder does not use any folders. You use the Model icon to add a new Model to the Study Folder, placing it inside the current Study, or the current Folder. It is probably the most important tool in PHAST, and you will use it in Tutorial 1, in the next chapter.

Page 5 of 25

Weather Tab Sectio n Click on the Weather tab to move to the Weather tab section. This tab section shows the weather conditions that have been defined and can be used in the consequence modelling. You can define any number weather conditions and thenof select between them for a particular run of the consequence calculations. The tree in the Weather tab section shares the top levels of its structure with the tree in the Models tab section, so that if you add a Study to the Study Folder in Weather Tab Section either section then it will appear in the tree in the other section. However, the lower levels are not shared, and you can have different structures of Folders in each tab section. If you click on the icon for a Study, you will see that the Folder icon and the yellow and blue Weather icon become enabled in the Toolbar:

You use the Weather icon to add a new definition of weather conditions to the Study Folder, placing it inside the current Study, or the current Folder. However, each new Study Folder is created with three default weathers already defined, and for most work it may be sufficient to edit these, rather than creating any additional weathers.

Parameters Tab Section If you are running PHAST you will see a parameter tab in your Study Tree. The parameters are not available to view or edit in PHAST Micro.

Page 6 of 25

Materials Tab Section Click on the Materials tab to move to the Materials tab section. The Materials Property system in PHAST is based around a three level hierarchy, with only the bottom two levels of the hierarchy visible section.in the Materials tab

Materials Tab Section

The top level is the System level, which is the central store for all PHAST Property data supplied with the program, and which is not visible in the Materials tab section. If your copy of PHAST is installed on a network, then the System values will also be on the network, and will be shared between all people using PHAST network data. The System values can only be changed by an Administrator using the special administration options, which are described in the online Help. The next level is the Global level, which applies to an entire Study Folder. When you add a material to Global Materials folder in the Materials tab section, PHAST creates a copy of the material inside that Study Folder, using the values from the System level as defaults. PHAST will add a material to the Global list the first time you use it in a Study Folder, but you can also add materials yourself, using the two Materials icons that appear in the Toolbar when you have the Global Materials folder selected in the Materials tab section:

The Example Study Folder only uses two materials in its models—chlorine and butadiene—but you can see that there are many more in the Global list, and these were added using the icons in the Toolbar. You can edit the values for the Global version of the material, as described later in this chapter, and these edits will be used throughout the Study Folder. The lowest level in the hierarchy is the Local level. PHAST creates a Local Materials folder for each Study in the Study Folder, and you use these if you want to create a version of a particular material that will be used only by that Study, while all other Studies use the Global version. You can add a material to the Local Materials folder

Page 7 of 25

either by copying and pasting from the Global list using the Edit menu, or by using the Materials icons that appear in the Toolbar when you select the Local Materials folder. PHAST knows which fields for the Local material have not been edited and therefore still have the default values taken from the Global level. If you edit a field for a Global material, PHAST will update the field for any Local versions that are still using the Global default.

Map Tab Sectio n

Click on the Map tab to move to the Map tab section. This tab section shows the maps that have been defined and on which you can superimpose consequence results. You can define any number of maps and then select between them when viewing a particular set of consequence results. The tree in the Map tab section shares the top levels of its structure with the trees in the Models and Weather tab sections, Map Tab Section so that if you add a Study to the Study Folder in any section then it will appear in the tree in the other sections.of However, lower levels are not shared, and you can have different structures Folders in the each tab section. If you click on the icon for a Study, you will see that the Folder icon and the Map icon become enabled in the Toolbar:

You use the Map icon to add a new Map to the Study Folder, placing it inside the current Study, or the current Folder. Each new Study Folder is created without any Map defined, so you must create a new Map if you want to view any map-based results.

Page 8 of 25

Viewing Input Data The section above described the organization of the different types of input data, and this section describes how to open the dialogs for the input data and view the values that are set for the Example Study Folder. In the next chapter, you will set values when working on a tutorial analysis.

Setting fault Unitsthe input data, you should set the default units for PHAST to Beforethe youDe start viewing your preferred system of units. As you will see later, you can change the units for a given item of data from inside the input dialogs, but it is much easier to set a default system that will be used throughout PHAST, including any dialogs and results. To set the default system, choose Select Another System… from the Units cascade in the Options menu. A dialog will appear, as shown in the illustration on the next page. PHAST is supplied with four predefined systems of units, but you can also edit these to create your own. At this stage, simply choose the pre-defined system that is closest to your preferences, and click on Make selected system current to set this as the default system throughout PHAST. The

examples in this chapter use the USER system, which are mostly Metric units.

Setting Default Units

Page 9 of 25

Getting Help on the Input Data PHAST has a large set of input data. This gives it the flexibility to model a wide range of releases and situations, but can be confusing at first. If you are unsure of the purpose of a particular dialog or field, you can use the context-sensitive online Help to get a description.

Most dialogs have a “Help” button at the bottom right. When you click on this, the Help window will appear, with the Help topic for that dialog displayed in the right-hand pane, as shown:

Online Help on a Dialog

Most dialogs also have a “What’s This Help” button in the form of a question mark at the right of the title bar:

A What’s This Help button in a Title Bar

If you click on this button, the cursor will change to a question mark, showing that you are in “What’s This Help” mode, and if you then click on a field in the dialog, a popup window will appear over the field, describing the field and giving advice on values, as shown here. What’s This Help on an input field

Page 10 of 25

The popup window will disappear the next time you click with the mouse. You will see both of these features in the dialogs that are described below. You can also bring up the What’s This Help for a field by pressing the F1 key while the cursor is on that field. In addition, if you press the F1 key again while the What’s This Help is being displayed, the Help window will appear, displaying the Help Topic for the dialog, as described on the previous page. You may find the F1 key more convenient than the buttons for accessing the Help system.

Input Data for a Model In the Models tab section, double-click on the icon for the CL2 RUPTURE model. The Model Data dialog will open, as shown in the illustration on the next page. The full set of input data is large, and is divided over many tab sections. The illustration shows the tab section for Material data, where you set the material that is released, the amount released, and the process or storage conditions at the time of the accidental event which leads to the release.

Input Data for the Cl2 RUPTURE Model

Page 11 of 25

Although the full set of data is large, you do not have to decide on and enter a value for every item of data in order to model a release; PHAST is supplied with default values for many of the items, and if you accept these default values, then you can define a release easily and quickly.

Input Data for Weather In the Weather tab section, double-click on the icon for Weather 1, and the dialog will open as shown.

Input Data for the Weather 1 Weather Condition

The set of input data is much smaller than for a model, and the most important items are in the Weather Data tab section. All of the items in the Atmospheric Parameters tab section take their initial values from the defaults system, so you can either accept the default value, or enter your own.

Page 12 of 25

You can tell that the Atmospheric Parameters tab section takes all of its values from the default system without even moving to it, because PHAST uses italic lettering for the headings of all such tabs. When a tab has italic lettering, you know that there are no fields on that tab section that you have to complete before you can use the Weather (or Model, or Material) in a calculation; however, if the heading of a tab section uses bold lettering—such as the Weather Data tab section—then this tells you that there are fields in the tab section that are initially blank, and that you must complete. This system of lettering can be useful when you want to obtain preliminary results quickly.

Input Data for Materials In the Material tab section, double-click on the icon for CHLORINE . The set of input data for Chlorine is very large, and some of it is very specialized and technical. If you want to add a new material to the properties system in PHAST, you will have to gather and enter a lot of information before you can use the material in the calculations. However, since PHAST is supplied with full data for a large number of materials, it is unlikely that you will ever need to define a completely new material, and, indeed, you may use PHAST for years without ever making any changes to any materials data.

Input Data for Chlorine

Page 13 of 25

You are most likely to use the Materials tab section for defining a Mixture—made up of existing Pure materials—and for looking up property data. You can refer to the input dialog to obtain the values of constant properties (i.e. those that are not a function of conditions), and you can use the options in the Material cascade in the Run menu to calculate properties at a given pressure and temperature (e.g. vapour density, saturation conditions, etc.).

Input Data for Maps When you double-click on the icon for Map of region around plant in the Maps tab section, a separate window will open showing you the location and geometry of your map file.

The Map Window

To actually view the map on your screen, you need to click on the globe in the Map toolbar:

Page 14 of 25

The map will then appear, as shown below:

The Map Window

When entering a map in your study you can use a map with header or geographical data. The Geographical Information System used in PHAST will then place your map automatically.

If you are not using a Georeferenced map or one with Header data you must enter The insert raster image window

Page 15 of 25

the map in Interactive mode. PHAST will automatically enable the Georeference and Header data options if your map includes this information.

If you are using a flat map (without Georeferenced or Header data) the main information that you need is, the location of the srcin on the map, and the scale for the map. You define these after placing the map on your screen, by right-clicking on the selected map. You will define a new map in the next chapter, which gives details of these operations. Setting the scale and srcin

Page 16 of 25

Running the Ca lculations Direct Run of a Single Mode l, Folder or Study To run a single Model, or to run all Models in a single Folder or Study, select that Model (or Folder or Study) in the Models tab section and then choose Run Model(s) from the right-click menu or Model(s) from the Run menu, or press Ctrl+M. The Run Model(s) command processes all of the calculations, from discharge through dispersion to flammable and toxic effects. If you want to run the discharge calculations alone, without proceeding to the dispersion and effects calculations, select the Run Discharge(s) command instead, or press Ctrl+D. When you are running a single item in this way, the program performs the calculations for the Weather conditions that are currently selected for the Batch Run.

Page 17 of 25

Viewing the Re sult s If the Study Folder contains Models that have been processed successfully through the calculations—shown by the use of blue text for the names of the Models—then you can view the results. To view the results, select the Model in the Models tab section, and then select Report or

Graph from the View menu, or press Ctrl+R for the Report, or Ctrl+G for the Graph. A single Report or Graph can display the results for more than one Model, but the options for selecting the multiple Models are different for each, and described separately below.

Viewing the Re ports Reports are displayed in the Report Window, which appears in the free space inside the PHAST Window—i.e. in the space not occupied by the Study Tree and the Log Window—which is normally to the right of the Study Tree.

The Report Window

You can generate a Report that contains the results for more than one Model if the Models are in the same folder or Study. Select the folder or Study and then use the option to view the Report, and the program will generate a Report with the results for all of the relevant Models.

Page 18 of 25

The window will contain several Reports, depending on the Model and the type of results that are relevant to the Model. By default, the program will display all available reports, but you can use Preferences... in the Options menu to exclude Reports that are not of interest in the current analysis. The Reports and the options for displaying them are described in more detail in the next chapter. You use the tabs to move between the Reports. Some Reports are long, and cover many pages. You can move between the pages of a Report using the navigation buttons at the left of the Toolbar for the Report Window. You can also move to a particular part of the Report by using the Report Tree at the left of the window. When you expand the tree, it shows the structure of the Report, with the sections that cover the different Models (if the Report covers more than one Model), the sections that cover each Weather that was processed for the Model, and the sections that cover the different release segments for each Weather, shown as 1, 2, 3, etc. in the illustration.

The Expanded Report Tree

Most Models have a single release segment, but a Model may have more than one segment if you used time-varying discharge modelling (which is an option in the Vessel tab section of the input data), or if the release contains liquid that rains out to form a pool, and the pool then evaporates, since the evaporation is treated as a form of time-varying discharge. To move to a particular part of the Report, click on that part in the Tree (e.g. segment 4 for Weather 1 in the illustration), and the program will move to the page that contains the beginning of that part of the Report. The other main feature of the Report Window is the Export button in the Toolbar. Use the Export button to export the contents of the Report to an external file of a given format (e.g. Excel, HTML, text). The Export Button

Page 19 of 25

If the initial size of the window is small, you may find it difficult to view the Reports clearly, and in this case you should choose Full Screen from the Window menu or from the Toolbar, since this option expands the window to fill the entire screen. To return from Full Screen to Normal mode, press Ctrl+W, the escape key, or click on the CloseToggle Full Screen button that is always visible when you are in Full Screen mode. The Window Menu

The Restore Button

You can have more than one Report Window open at any time. Use the Window menu to switch between multiple Report Windows, or to arrange the windows so they are all visible at the same time.

Viewing the Graphs When you select Graph from the View menu, the Plot Setup dialog will appear, prompting you to choose between the Weather conditions that have been modelled, and to choose a Map on which to superimpose the footprint results.

Page 20 of 25

Choosing the Results to Plot

The list of Weathers will include all of the Weathers that have been defined for each Study, and not just the Weathers that have been processed for the current Model. If you select a Weather that has not been processed, an error message will appear when you click on OK. You can also choose the option to view a Graph from the Weather tab section of the Study Tree. In this case, the Plot Setup dialog will contain a Model tab section instead of a Weather tab section, and you can select multiple Models to plot for the Weather that is currently selected in the Study Tree. When you have chosen the items that you want to plot, the Graph Window will open in the area to the right of the Study Tree. The Graph Window contains many Graphs, and you move between them using the tabs. The Graphs and the options for displaying them are described in more detail in the next chapter.

Page 21 of 25

The Graph Window

As with the Report Window, you can have more than one Graph Window open at a time, and you use the Window menu to arrange the Graph Windows, and to switch to Full Screen mode. If you choose a single Weather and Model, the graphs will show the results for different concentrations, distances and overpressures, as appropriate for the type of graph. If you choose more than one Weather or Model, the graphs will show the results for a single concentration, distance or overpressure for each Weather or Model.

Page 22 of 25

Saving the Exa mpl e Study Folder Although you have not made any changes to the input data for the Example Study Folder, you have run the calculations. If you save the results with the rest of the Study Folder data, then the next time you open the Example Study Folder, you will be able to view the results immediately, without having to rerun the calculations. You should leave the Example.PSU file from the Examples folder unchanged, so that other users will be able to explore in itssince srcinal state. This means that you in should use the Save option from the File it menu, this would overwrite the file the not Examples folder. Instead, you should use the Save As... option from the File menu, so that you can save the Study Folder to a different location, creating your own copy of it. When you install PHAST, the installation program creates a folder to be the preferred location for Study Folder data. The default name and location for this folder are c:\DNVuser (if PHAST is installed on the c: drive), but you can set any name and location during the installation. If you have access to this folder, you should use it for your copy of the Example Study Folder. Before clicking on Save, you should ensure that the Save results check box is ticked, as shown in the illustration on the previous page. By default, PHAST does not save results for the Example Study Folder or for any new Study Folder, and you must use Save As... if you want to change this option. The results can make the Study Folder files very large. Since the calculations usually run very quickly, you may prefer to save your Study Folder files without the results, and then rerun the calculations every time you open the files.

Saving the Example Study Folder

Page 23 of 25

DNV SOFTWARE

PHAST Getting Started Manual Chapter 2: Performing a Worst-Case Analysis

PHAST Getti ng Started Manual Chapter 2: Performing a Worst-Case Analysis

Table of Contents Introduction..................................................................................................................................... 1 Starting PHAST .............................................................................................................................. 2 Introduction to the Analysis............................................................................................................ 3 Creating the Anysite Study Folder.................................................................................................. 6 Setting the Materials Input Data ..................................................................................................... 7 Setting the Weather Input Data..................................................................................................... 10 Setting the Map Data .................................................................................................................... 12 Defining the Ammonia Release .................................................................................................... 16 Defining the Hydrogen Cyanide Release...................................................................................... 21 Defining the Ethylene Release...................................................................................................... 22 Defining the Propylene Release.................................................................................................... 25 Viewing the List of Global Materials ........................................................................................... 26 Running the Calculations.............................................................................................................. 27 Viewing the Results ...................................................................................................................... 28 Summary of Worst Case Analysis ................................................................................................ 34

Introduction This manual will lead you through the mail features of PHAST and PHAST Micro, by opening a pre-defined case so that you can view both input data and results.

All Ex amp les are fro m PHAST Micro There two versions of PHAST and this manual covers both of them. PHAST Micro is the simplerare version, containing DNV’s, sophisticated dispersion modelling in full, but with some limitations to the options in other areas of the modelling. PHAST is the fully-featured version, offering control over most aspects of the modelling, and including stand-alone versions of the fire, explosion and pool vaporization models that are built into the integrated dispersion modelling. All of the examples in this chapter are based on PHAST Micro and are fully applicable to that version. If you are using PHAST, you will see some features in your program that do not appear in the illustrations and are not described in the text. Instruction on these features are given in the PHAST Introductory Training Course. For details on the PHAST training courses please contact your local Technical Support centre.

Page 1 of 36

Starti ng PHAST When you install PHAST, the installation routine places a DNV folder under Programs in your Start menu, and you can start PHAST running by selecting the icon from the folder.

The DNV folder in the Start Menu

The installation routine also places a PHAST icon on the desktop so you can also start PHAST running by clicking on the desktop icon.

Page 2 of 36

Intro ducti on to the Analysis In this chapter, you will perform a simple worst-case analysis for the Anysite chemical installation, to determine whether releases on the site have the potential to reach populated areas beyond the site boundary.

Hazardous Materials

There are four hazardous materials present on the site in significant quantities: Material

Type of Hazard

Anhydrous ammonia Hydrogen cyanide Ethylene Propylene

Toxic Toxic Flammable Flammable

Mass Present lb tonne 40,000 18.1 5,000 2.3 50,000 22.7 75,000 34.0

Hazardous Inventory for Anysite Facility

Storage Condition s The ethylene is stored under supercritical conditions, and the three other materials are stored under saturation conditions. For the worst-case analysis, the materials will be modelled at the maximum temperature experienced at the facility over the last five years, which is 90°F (32°C). At this temperature, the storage pressures for the materials are as follows: Material

Conditions

Anhydrous ammonia Hydrogen cyanide Ethylene

Saturation Saturation Supercritical

Storage Pressure Psig barg 180.1 12.4 18.7 1.3 700.0 48.3

Storage Conditions

Release Scenarios Different scenarios will be modelled for the toxic and the flammable materials, since different types of release cause the worst long-range effects. For the two toxic materials, the release scenario will be a release of the entire inventory over ten minutes, and for the two flammable materials, the scenario will be an instantaneous release of the entire inventory. For toxic releases, the duration and concentration profile at the populated areas are more important than the total mass in the cloud at any given time. A large continuous release will give

Page 3 of 36

a greater duration of exposure than the equivalent instantaneous release. It may also take longer to disperse to harmless concentrations, since air is mixed into the cloud from the sides only, whereas air is mixed into an instantaneous release across all exposed surfaces. For flammable releases, the greatest effect distances are usually produced by vapour cloud explosions, and the size of these explosions depends on the flammable mass in the cloud at the time of the explosion—which will be greater for an instantaneous release than for a continuous release.

Critical Effect Zones For the toxic materials, the calculations will obtain the dispersion distances to the Emergency Response and Planning Guidelines (ERPG) Level 2 concentration, which is the concentration that nearly all individuals can be exposed to for up to an hour without experiencing any irreversible adverse health effects or symptoms which could impair the ability to take protective action. For ammonia, this concentration is set at 200 ppm, and for hydrogen cyanide, it is set at 10 ppm. For the flammable materials, the calculations will obtain the explosion distances to an overpressure of 1 psig, which is an overpressure that may cause injuries as a result of minor structural damage (e.g. broken windows), but is unlikely to cause fatalities.

Weather Condition s The calculations will use a windspeed of 5 ft/s (1.5 m/s) and an atmospheric stability of F, which are common night-time conditions for the location. These conditions give low levels of atmospheric turbulence, and the release may travel long distances before being diluted to a harmless concentration. The average humidity for the location is 70%, which is typical for a temperate, maritime location. The calculations require a value for surface roughness, which is a measure of the turbulence induced in the air as it moves over the ground, and will be set conservatively to 0.06, a value for sea or for flat, treeless land. This assumes that the wind is blowing towards the town, and that the surface conditions upwind of the release determine the surface roughness.

Page 4 of 36

Location of the Anysite Facility As shown in the map, Anysite is a large, ocean-side facility, located in an industrial area, and nearly two miles from the nearest residential area.

The Anysite Chemical Facility

Page 5 of 36

Creating th e Any sit e Stud y Folder First, you must create a new Study Folder to store all of your work on the Anysite facility. Close any Study Folder that is currently open in PHAST, and then select New from the File menu. The program will create a new Study Folder called

Untitled with an empty Study called New Study.

Saving the Study Folder You cannot save a Study Folder with the name Untitled . Use either the Save or the Save As... options in the File menu to save the new Study Folder to the DNVuser directory with the name Anysite.PSU .

The New Study Folder in the study pane

Renaming the Study Click on the Study to select it, and then choose Rename from either the Edit menu or the rightclick menu. An insertion point will appear in the name of the Study, and you should edit this to change it to Worst Case.

Using Program Prefe rences to Open the Study Folder Autom

atically

All of the tutorials in this manual use the Anysite Study Folder. If you do not perform the tutorials in a single session, you will be returning to the Study Folder several times. The list of recently-used Study Folders at the bottom of the File menu makes it easy to re-open a Study Folder that you have been working on, but you can also use the Preferences for the program to make this even easier. Select Preferences... Installation… from the Options menu. The Preferences dialog will appear, and you should set the option in the Startup tab section to Try to open most recently used file, as shown in the illustration on the next page. If the file has been deleted or moved, the program will display a File Open dialog instead, so you can locate the file yourself. Setting the Preferences for Opening a Study Folder Automatically

Page 6 of 36

Setti ng the Mate ri als Inpu t Data In the database of System Materials supplied with the program, ammonia and hydrogen cyanide are defined as being both flammable and toxic. However, for the worst case analysis, you are only interested in the toxic effects, and you can simplify the input data and the results if you define them as toxic only for this analysis. You do this by creating local copies of the materials, and editing the property data. If you wish, you can omit this stage, since it is not essential. However, you may find it useful as a quick and straightforward introduction to the properties system.

Creating Loc al Versio ns of the Toxic Mate rials Move to the Materials tab section of the Study Tree, select the Local Materials folder under the Worst Case Study, and select Material... from the Insert menu. The Insert Material dialog will appear, as shown in the illustration below.

Inserting a Local Version of Ammonia

Page 7 of 36

The dialog offers three ways of inserting a material. The New option allows you to create a completely new material, with no pre-defined properties data. The Existing and Copy options both allow you to create a copy of a material that is already in the materials database at a higher level (i.e. at the System or Global level): the Existing option keeps a link to the srcinal material, and if the values for the srcinal material are changed, the program will automatically update the values for any fields that are still using the srcinal, default values; the Copy option does not keep a link, and the local version will not be affected by any changes to the srcinal material. Select the Existing option, locate and select AMMONIA in the list of materials, and then click on OK to add the material to the Local Materials folder. Next, repeat the process, selecting HYDROGEN CYANIDE as the material.

Editing the Mate rials Data for the Loc al Materials When you expand the Study Tree below the Local Materials folder, you will see the icons for the two materials. Double-click on the icon for AMMONIA to open the input dialog, and set the Flammable/Toxic field in the General tab section from the default value of Both to Toxic only, as shown in the illustration on the next page. Click on OK to save the changed data, and then repeat the process with the local version of HYDROGEN CYANIDE.

Page 8 of 36

The Changed Data for Ammonia

Page 9 of 36

Sett in g th e Weath er Input Data Before defining any of the worst-case releases, you will define the other aspects of the input data, which will be the same for all four releases: the Weather data, and the Map data. Each new Study Folder is created with a set of default Weather conditions defined for the default Study, as shown in the illustration. These default Weathers are representative of the range of common conditions, and they enable you to obtain results for a new Study Folder very quickly. For this Worst Case analysis, you are only going to model one condition—1.5 m/s with F stability —which is one of the default conditions. The Default Weather Conditions for a New Study Folder

Delete the Unwanted Conditions First, delete the other default conditions. You delete an icon from the tree by clicking on it to select it, and then using the Delete (Del) key or the Delete option in the Edit menu or the rightclick menu. You could leave the conditions in the tree, but it will make the design of the analysis clearer if you delete them.

Set the Detailed Weather Data Next, double-click on the Weather 1.5/F icon to open the dialog for input data, and set the following values in the Atmospheric Parameters tab section:

Page 10 of 36

The Atmospheric Parameters Weather options

All of the fields in the Atmospheric Parameters tab section take their initial values from the defaults system, which is shown by the green border around each field. When you change the values to those required for this analysis, you will see that the border disappears—the colourcoded borders mean that you can see at a glance which fields in a dialog are using the default values directly, and which have been changed.

Page 11 of 36

Sett in g t he Map Data In the Map tab section of the Study Tree, select the Raster Image Set option under the Anysite Plant, and then select the option to insert a new Raster Image using the icon in the Toolbar, the Insert menu, or the right-click menu.

Selecting the Bit map Image A bitmap image of the Anysite facility and its surroundings is supplied with PHAST. This can be selected by browsing to the C:\Program Files\DNV\PHAST_6_4_2\Examples\ folder and selecting the Anysite.bmp file.

The Insert Raster Image Screen

Once you have selected this raster image the Interactive Placement Mode option will become enabled. This means that you can place your raster image onto your Map Window interactively. The remaining Placement Mode options are not available because the Anysite.bmp file does not contain GIS data. If you use a GIS raster image, with Header or Georeference data, PHAST will place your map automatically for you. For further details on the GIS system in PHAST please refer to the Online Help system. When you press OK a Coordinate System Wizard will appear. Press Cancel to exit this wizard, as it is for use with true GIS raster images. Again, for further details on the Co-ordinate System in PHAST please refer to the Online Help system.

Page 12 of 36

The Coordinate System Wizard

A blank Map screen will now appear, for you to draw your raster image on to, using your mouse, which will appear as a cross. Draw your map onto the blank Map screen, ignoring the actual coordinates on the screen as you will set the srcin and scale of the map in your next step. NOTE: After pressing cancel you may have to wait a while before the blank map appears. Do not click elsewhere on the screen during this time or the program will switch to a new action and you will have to repeat the ‘Insert Raster Image’ step.

Page 13 of 36

Setting the Scale You set the scale by drawing a line between two points whose distance apart is known, then typing in the distance. In this case you will draw a line across the entire map, which is 5 km wide.

The Map Scale and Origin menu

Page 14 of 36

To set the scale, select the map by clicking on it, then choose Map Scale and Origin... Set Scale… from the Map menu or the rightclick menu. Then draw a line across the entire width of the map. A dialog box will appear asking for the length of your line, so you should type in 5 km. When you press OK to close the scale window you will see that the horizontal and vertical axes have been rescaled for the new value. Setting the scale

Setting the Origin The location of the map is defined when you first draw the raster image onto the blank map screen. The co-ordinates at that time are unlikely to be correct for the map, whether you are setting them with global co-ordinates or reference co-ordinates (in relation to a specific place). Therefore, you must set the raster image srcin manually after placing it on your screen and setting the scale. In this example you will set a reference srcin of (0, 0) in the middle of the Anysite Facility. You will later set the co-ordinates of the release to (0, 0), which will place them at this srcin.

Setting the Origin

To set the srcin, select the map by clicking on it, then choose Map Scale and Origin... Set Origin… from the Map menu or the right-click menu. The cursor will change to a cross-wire, and you simply click on a point on the map to set that point as the srcin. A dialog box will appear for you. For this worst case analysis, you do not have to place the srcin with great precision, and any location near the middle of the site will be suitable.

Page 15 of 36

Defin ing the Ammo nia Re lease The first worst-case release is the 40,000 lb (18.1 tonne) ammonia release.

Inserting th e Mode l Select the Study, and then insert a Vessel or Pipe Source Model, either by clicking on the icon in the toolbar, or by selecting Vessel or Pipe Source from the Insert menu. The new Model will be given the name Vessel/Pipe Source, and you should rename it immediately to Ammonia, as shown. When you insert the new Model, you will find that red boxes appear around all of the icons in the Models tab section of the Study Tree. The box appears around the new Model to show that it does not have a complete set of input data, and you will The new model in the study tree therefore not be able to process it through the calculations. When you have completed the data input for the Model, the box will disappear. The box appears around the Worst Case Study to show that a Model inside the Study has incomplete data, and similarly for the Study Folder; this effect on the higher levels of the tree can be useful in a large analysis with many Studies.

Setting the Material Data Double-click on the Ammonia icon to open the dialog for the input data, and set values in the Material tab section as shown in the illustration on the next page. To set the Discharge Material, click on the button with three dots at the far right of the dialog, and select AMMONIA from the list which appears, as shown. You will see that the list contains many materials, and not just the two materials that you inserted in the Local Materials folder. The Scope column shows these two materials as Local, whereas all

Page 16 of 36

Selecting a Material

of the other materials are shown as System. If the Scope is System, then there is currently no Global or Local version of the material. If you selected one of these System materials (e.g. BENZENE), the program would automatically create a version of the material in the Global Materials folder for the Study Folder, and the next time you opened the Select Material dialog, you would see that the Scope for the material had changed from System to Global; you will see this later, when you are defining Models for the two flammable materials. Now enter the remaining data in the Material tab, using the data shown in the screen below.

The Input Values for Material Data

When you select the material, the program automatically sets the Material to Track to AMMONIA . You only have to choose a material to track if the Discharge Material is a mixture. Note that you can use scientific notation when entering values, so you can enter the inventory as “40e3”.

Page 17 of 36

Setting the Scenario Data Move to the Scenario tab and set the following: Scenario Release Phase

10 Minute Release Liquid

For most other types of Scenario, you have to give additional data that will enable the discharge calculations to calculate the release rate. However, for the 10 Minute Release, the release rate is calculated (inventory/600 seconds) and not with any discharge calculations, so the Scenario input data areasvery simple.

Settin g t he Location Data Move to the Location tab, skipping the Pipe and Vessel tab, and set the values shown in the illustration:

Location Input Data

Page 18 of 36

The Elevation has a default value which is greater than zero, and you should leave it with this default value. If a release is located at ground level (i.e. the Elevation is zero), the program omits the detailed modelling of liquid droplets and their evaporation and possible rainout, and simply assumes that all of the liquid in the release rains out immediately; this is a reasonable assumption, since liquid droplets will have no opportunity to evaporate during the fall to the ground if they are released directly onto the ground. However, most releases will be at some elevation above ground level, and the program is supplied with a default Elevation that will give a treatment of the liquid droplets that is more typical of a real release. This worst case ammonia release is a vapour-only release, so the elevation is not as important as for a liquid or two-phase release, but it is still more realistic to place the release at some distance above the ground. Leave the North and East coordinates with the default coordinates of zero, which will place the release at the srcin for the Map, which is in the middle of the Anysite facility. Leave the three Distances blank. You can set a distance if you are interested in the effect levels at a particular location, but for this analysis you are interested in the maximum dispersion distance to a concentration of 200 ppm. Check the box for Concentration of interest, set a value of 200 ppm, and set Uses averaging time to Toxic. The significance of the Averaging time is described in detail below.

Av eragin g Times in PHAST: an Intro du ct io n The averaging time is important in PHAST, and is more prominent now than in previous versions. It is used to take into account the effects of changes in the wind direction over the course of the release, and the way that the changes cause the plume to meander from side to side. In order to interpret concentration results correctly, you must know the averaging time that was used in calculating the concentration, and the program allows you to specify different averaging times for different types of concentration results. The wind does not blow steadily in a straight line; its direction varies with time, which causes a cloud plume to meander from side to side. If you are standing downwind, at one moment you are in the centre of cloud, experiencing the peak concentration, and the next moment the peak has moved away to the side, and you are experiencing a much lower concentration—and in the moment after that, the peak comes back over you and off to the other side, and so on. The average concentration you receive over, say, 5 minutes will be much less than the peak concentration; if you stood at the location for 30 minutes, the average would be lower still. This factoring down of the peak concentration is carried out by the Averaging Time Adjustment— the longer the time window, or Averaging Time, the lower the calculated average concentration will be. For thetype Concentration of Interest , you material, can choose between several averaging times, depending on the of release. For a toxic-only there are five choices: a User-Defined time that you set in the User-defined field at the bottom left of the dialog group below; a Toxic time that is set in the Toxic Parameters; and the ERPG, IDLH and STEL times that are set as part of the definitions of these measures of toxicity, and cannot be changed. When you select a type of

Page 19 of 36

averaging time from the list, the value of the averaging time will be displayed in the field to the right of the list; the default toxic averaging time is 600 seconds, which is also the duration of this release.

Setting the Bund Data You can leave the bund data unset, since they are not relevant to this vapour release. For a liquid release, however, the presence and size of the bund can have a very large effect on the results: if there bund, then the pool from the anysurface liquid rainout can spread to if cover very wide then area,it givingisano high evaporation rate from of the pool; whereas therea is a bund, limits the area of the pool and the evaporation rate, as you will see in the next chapter.

Settin g t he Indoor/Outdoor Data Next, move to the Indoor/Outdoor tab and set the Release Direction to Horizontal. You can model a release as out of doors, where the only obstruction is the ground, or as inside a building, where the size and ventilation of the building affects the initial stages of dispersion. All continuous releases you must set the direction. Instantaneous releases do not require a direction as the inventory will be released in all directions.

Ignoring the Other Ta b Sections You skipped the Vessel tab section, and you can ignore all of the remaining tab sections and click OK to save the changes you have made. For a vapour release, the Vessel tab section is only relevant if you want to perform timedependent discharge modelling, in which case you must give information about the dimensions of the vessel and the liquid level. Such modelling is not applicable to the 10 Minute Release scenario, which requires only the simplest discharge modelling. The remaining tab sections allow you to change the default settings for explosion, fire and discharge modelling. For a 10 Minute Release of a toxic-only material, these tab sections are not relevant.

Page 20 of 36

Defin ing the Hydr ogen Cyanid e Release The ammonia and the hydrogen cyanide releases have the same data for the Scenario and Indoor/Outdoor tab sections, and differ only in the Material and Location data. To take advantage of this, you will create the Hydrogen cyanide model as a copy of the Ammonia model, and then edit the Material and Location data.

Copying the Ammo nia Mode l

Select the Ammonia icon, and then select Copy from the Edit menu or from the right-click menu. Then select the Worst Case Study, and select Paste from either menu. The program will give the copy the name Ammonia(1), and you must rename it to Hydrogen cyanide.

Setting the Material Data Double-click on the Hydrogen Cyanide icon to open the dialog, and change the values in the Material tab section to the following values: Discharge Material Inventory

HYDROGEN CYANIDE 2300 kg (2.3 tonnes)

Setting the Loc ation Data Move to the Location tab section and change the values to the following: 10 ppm Concentration of interest These changes complete the data for the release, and you can click on OK to save the edited data.

Page 21 of 36

Defin ing the Ethy lene Re lease The input data for the flammable releases are significantly different from those for the toxic releases, and there is nothing to be gained from copying one of the existing releases. Create the ethylene release by inserting a new Vessel or Pipe Source Model, and give it the name Ethylene .

Setting the Material Data

Double-click on the Ethylene icon to open the dialog, and set the values in the Material tab section as follows: ETHYLENE Discharge Material 22,700 kg (22.7 tonnes) Inventory Temperature Process Conditions Pressure 32 oC Temperature 43 barg Pressure The ethylene is stored under supercritical conditions, and you must specify both the temperature and the pressure. Before you complete the entry of the temperature and pressure, you will see that the program gives the Vessel Type and Phase as Unknown . At this point, the program has not checked the state of the material at these conditions, and does not know that it is supercritical. Once you enter the temperature and pressure then click or tab in another field the Phase will change to show the material is stored as a Vapour and the Vessel Type will change to show Pressurised Gas.

Setting the Scenario Data In this case you will accept all the Scenario tab default values, including the catastrophic release settings, for which there is a choice. You will see that in this tab the release phase has been set to Vapour. This happens because the program has checked the phase and determined that the material is supercritical (which the program models as vapour). Because you have set your release as a Catastrophic Vapour you can ignore both the Pipe and Vessel tabs.

Checking the Location Data Move to the Location tab section and view the data as shown in the illustration on the following page:

Page 22 of 36

The Location Input Data for the Ethylene Model

Unlike the toxic cases, you do not need to set a Concentration of interest or choose or set an associated Averaging Time. For flammable releases, the program automatically performs the dispersion to a fraction of the lower flammable limit (where the fraction is set in the Flammable Parameters), using the Flammable Averaging Time (also set in the Flammable Parameters). If you are interested in the details of the concentration results for a flammable material, you might set an additional concentration of interest and a user-defined averaging time, but for this analysis the effects from an immediate explosion are likely to be more significant than any later cloud dispersion.

Checking the Flammable Data Move to the Flammable tab section and check that the Explosion Method is set to TNT. PHAST now has three explosion models available.

Page 23 of 36

Leave the Early Explosion Mass Modification Factor with its default value of 3. This factor is used in calculating the mass involved in an early explosion. The program calculates the mass of vapour in the cloud after it has expanded to atmospheric pressure, and then multiplies this mass by the Modification Factor to obtain the explosion mass, with an upper limit set by the flammable mass released.

Setti ng the TNT Da tasection. The tab section is to the right of the Flammable tab section and Move to the TNT tab may not be immediately visible when the dialog first opens. If you cannot see the tab section, use the navigation button at the far right of the tabs to reveal the other tabs in the dialog. Leave the TNT Explosion Efficiency with its default value of 10%. This determines the fraction of the combustion energy in the explosion mass that is converted into explosion energy. Set Air / Ground Burst to Ground Burst, which means that the explosion occurs near the ground, i.e. at the same elevation as the release. For this type of explosion, the effects of reflection from the ground are assumed to double the amount of energy involved in an explosion, so this type will give the worst case results. These changes complete the data for the release, and you can click on OK to save the edited data.

Page 24 of 36

Defin ing the Prop yl ene Release The propylene release differs from the ethylene release only in the Material data, so you can create it as a copy of the Ethylene model, using the method described for creating the hydrogen cyanide release. Give the copied model the name Propylene .

Setting the Material Data

Double-click on the Propylene icon to open the dialog, and change the values in the Material tab section to the following values: Discharge Material Inventory Process Conditions Temperature

PROPYLENE 34,000 kg (34.0 tonnes) Temperature Saturated Liquid 32 oC

When you change the material, the program performs flash calculations to check the current process conditions and updates the reported Phase and Fluid Type if necessary. At 32 oC and 48 barg, ethylene is a supercritical vapour but propylene is a liquid. Therefore, you will see the Phase change from Liquid to Two-Phase and the Fluid Type change from Pressurised Liquid to Saturated Liquid.

Checking Scenario Move tothe the Scenario tabData section, and you will see that the only choice for Phase to be released is Liquid. The presence and design of a bund or dike can have a significant effect on liquid releases, but you should leave the bund data in the Bund Data tab section unset, as with the vapour releases, since this will allow the liquid pool to spread indefinitely, giving a larger evaporation rate than with a bund. This completes the release data, and you can click on OK to save the edited values.

Page 25 of 36

Viewin g the List o f Global Ma terials Move to the Materials tab section, and expand the tree under the Global Materials folder. You will see that icons for ETHYLENE and PROPYLENE have been added to the folder; each was added as a copy of the System version when you selected the material in the input dialog. The Ethylene and Propylene Models obtain their materials data from these Global versions. If you add Models for further ethylene or propylene releases, these Models will also use these versions. The global & local materials lists

Page 26 of 36

Running t he Calculations To run the calculations for all of the Models, move back to the Models tab in tab in the Study Tree and select the icon for the Worst Case Study, then start the run. There are three ways of starting a run: you can select Model(s) from the Run menu or Run Model(s) from the right-click menu, or you can press Ctrl+M. You can follow the progress of the run in the Progress Meter, and also in the Message Log tab section of the Log Window.

Running the calculations

Page 27 of 36

Viewin g th e Results The Graphs give the most direct way of viewing the results. To view the Graphs for a Model, select that Model, then click Ctrl+G, or choose the Graph option from the View menu or the right-click menu. The Plot Setup dialog will appear, prompting you for the weather to use, and when you click on OK, the program will generate the Graphs, and display them in the Graph Window. In this example we will view the Category 1.5/F weather, though you can view more than one weather if you wish.

The Results for A mmoni a There are seven Graphs which show concentration results. For this analysis, the most important Graph is the Map. When you first move to the Map tab section, there will be two concentration contours shown on the Map, for around 200 ppm (the concentration of interest) and around 400 ppm, as shown in the illustration. These contours are some distance from the facility, and show that the cloud has become detached from the release point because the time taken for the cloud to disperse to 200 ppm was much longer than the ten minute duration of the release.

The map graph view

This aspect of the release makes the results quite complex, and you may find them difficult to interpret at first, especially as the program gives much more detail in the results than in previous versions and provides many more options. The first thing to notice in Graphs of this type is the Time displayed in the legend. In the illustration above, the time is given as 1881 s, and this is the time after the start of the release at which the area covered by the 200 ppm contour (the contour for the concentration of interest) was largest.

Page 28 of 36

To see the contours for other times, select Dynamic from the Graph menu or the right-click menu, and the Cloud Dynamics Control will appear. You use this Control to view an animation of the cloud dispersion. Note: The Dynamic option will not be available if you have selected more than one weather from the Plot Setup screen. Click on the rewind button at the left of the control to set the animation time to the beginning of the release, and then click on the play button at the right of the control to start the animation. You will see the development of the cloud displayed on the Map, and when the time reaches about 600 seconds (as shown in the legend) you will see the cloud become detached from the release point. The cloud dynamics control screen

If you run the animation to the end of the release, you will see the 200 ppm contour go into the ocean, and disappear off the map. In this direction the cloud does not have reach any populated areas. However, the wind could also be moving from the south, and we should examine the effects from this direction. To model the release from the south select Wind Direction from the Graph menu or the right-click menu, and the Wind Direction control screen will appear. Move the bar to 180 degrees, as shown, to indicate the direction the wind is coming from.

The cloud dynamics control screen

After setting the Wind Direction you should again run the animation to see if the cloud reaches the town. You will then see that the worst case ammonia release does have the potential to reach populated areas offsite. The effect of the cloud will depend on the time that it takes to pass over the town, and you can see this in the Concentration vs Time graph. This shows the time-dependent concentration profile at a particular distance from the release source. When you first move to the tab section, the distance will be set as the mid-point of the cloud at the time that the contour covers the largest area (i.e. as in the first view of the Map Graph), but you can change this distance. The distance from the release to the middle of the town is approximately 4.3 km. To set this as the distance for the Concentration vs Time Graph, right-click on the centre of the graph image and select Properties from the right-click menu, and set the value as shown:

Page 29 of 36

Setting the downwind distance for the concentration vs time graph

When you click on OK and return to the Time Graph, the Graph will change, and you can see the concentration profile at the town. The Graph shows that a person at that point would only be exposed the cloud for about ten minutes, but theisconcentration during this time would be over 400 ppm.toThe 200 ppm concentration of interest based on an exposure of an hour, so the effects from this cloud should be small, but could still be unpleasant.

The concentration vs time graph

Page 30 of 36

This graph shows many vertical lines from 0 to 400 ppm. These represent the end points of different cloud segments which are produced because the Ammonia release forms a pool which vaporises at different rates. For further details on cloud segments please refer to the on-line Help system.

The Results for Hydrogen Cyanide Unlike the ammonia release, the hydrogen cyanide release does not become detached from the source. This is because the cloud rains out close to the source and the pool vaporises about an hour. The wider cloud positioned further from the source is produced by pool segments with a high vaporisation rate. The thinner cloud closer to the source is produced by pool segments with a lower vaporisation rate. If you use the Properties... option to set the distance of interest to 4.3 km and move to the Time Graph, you will see that the duration of exposure at the town is just under one hour. In addition, for some of this time, the concentration is almost four times the concentration of interest of 10 ppm. This indicates a much more significant hazard than the ammonia release, which reached twice the concentration of interest for only ten minutes. However, the difference in the values for concentration of interest makes it difficult to correctly compare the concentration results for ammonia and hydrogen cyanide using the Map and Time Graphs. However, the Lethality Graph (stored under the Toxic tab) allows you to compare the toxic effects directly, and you can also plot the results for the two Models on the same Graph. In order to plot the combined results, you must open a third Graph window from the Weather tab section of the Study Tree. To do this move to the weather tab and select the 1.5/F Weather, and then press Ctrl+G to generate the Graph. The Plot Setup dialog will open, with a Model tab section instead of a Weather tab section, and you can select the two toxic Models, as shown:

Plotting a graph for more than one model

Page 31 of 36

The Lethality Graph shows the results for both Models, and it shows that the toxic effects of the Hydrogen cyanide Model are worse than those for the Ammonia Model. However, neither gives a significant probability of death at the town.

The Probability of Fatality Graph

If you look at the Graphs for concentration, you will find that they are plotting the results for 200 ppm, for ammonia. The program cannot plot asocomparison resultsi.e. forthe 10concentration ppm, because of theinterest calculations for ammonia stopped at 200 ppm, it can only of the compare the results for 200 ppm. This comparison at 200 ppm may be misleading, because the inventory for the Hydrogen cyanide Model is much smaller than for the Ammonia Model, and the cloud is diluted to 200 ppm much more quickly. This emphasizes that some Graphs are useful for some purposes (e.g. getting the details of the results for a single Model, or for comparing Models that involve the same material) whereas other Graphs are useful for other purposes (e.g. comparing Models that involve different materials).

The Results for Ethylene a nd Propylene The two flammable Models have the same critical effect-level, i.e. 0.02068 bar, and their results can be compared directly. Move to the Weather tab section, select the 1.5/F Weather, press Ctrl+G, and then select the two flammable Models in the Models tab section of the Plot Setup dialog. When you first move to the Map Graph, it will be displaying the concentration contours for the two Models. This is the default option for the results displayed on the Map, but you can use the Properties... option to change this.

Page 32 of 36

In the Display tab section, change the Map Event from Dispersion to Late Explosion , as shown.

Setting the graph properties for viewing the late explosion overpressures on a map

When you have clicked on OK and the program has redrawn the Map Graph, you will see that the overpressure radii to 0.02068 bar do not extend outside the boundary of the site, and pose no threat to the town, as shown in the illustration.

The explosion results for the flammable models

Page 33 of 36

Summary of Worst Case Ana lysi s The Worst Case analysis shows that the hydrogen cyanide inventory poses the greatest offsite risk, although no scenarios are capable of causing fatalities at the town. If you require any further information on any of these cases please contact your local Technical Support desk or sign up for one of our training courses.

Page 34 of 36

PHAST-6-5-1-Tutorial

Recommend Documents