Introduction
flacs.png

About this publication

FLACS v10.9 User Manual

Copyright © 2019 Gexcon AS
All rights reserved
Updated: May 06 2019

Printed in Norway
Intellectual property notice
No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, or otherwise, without written permission from Gexcon AS.

Gexcon AS hereby grants permission to use, copy, and print this publication to organisations or individuals holding a valid license for one or several of the software packages described herein.

For further information about Gexcon AS, please visit the Gexcon web site.

Exclusion of liability
Gexcon AS has distributed this publication in the hope that it will be useful, but without any warranty, without even the implied warranty of merchantability or fitness for a particular purpose.

Although great care has been taken in the production of this publication to ensure accuracy, Gexcon AS cannot under any circumstances accept responsibility for errors, omissions, or advice given herein.

Registered trademarks

  • FLACS, DESC, CASD, and Flowvis are registered trademarks of Gexcon AS.
  • Linux is a registered trademark of Linus Torvalds.
  • Windows is a registered trademark of Microsoft Corporation.

Other product names mentioned herein are used for identification purposes only and may be trademarks of their respective companies.

Preface

Computational fluid dynamics (CFD) is a branch of fluid mechanics that uses numerical methods and algorithms to solve and analyse problems that involve fluid flow, with or without chemical reactions. Current use of CFD covers a broad range of applications, from fundamental theoretical studies involving models primarily derived from first principles, to practical engineering calculations utilising phenomenological or empirical correlations.

Many of the hazards encountered in society, and especially in the process industries, involve accident scenarios where fluid flow in complex, large-scale, three-dimensional (3D) geometries play a key role. FLACS is a specialised CFD toolbox developed especially to address process safety applications such as:

  • Dispersion of flammable or toxic gas
  • Gas and dust explosions
  • Propagation of blast and shock waves
  • Pool and jet fires.

The development of FLACS started in 1980 at the Department of Science and Technology at Christian Michelsen Institute (CMI). CMI established Gexcon (Global Explosion Consultants) as a consultancy activity under the Process Safety Group in 1987. In 1992, the Science and Technology department at CMI became Christian Michelsen Research (CMR), and CMR established Gexcon as a private limited company in 1998. Gexcon AS is a wholly owned subsidiary of CMR, and holds the full proprietary rights to the CFD code FLACS.

The purpose of this manual is primarily to assist FLACS users in their practical work with the software. In addition, the manual aims at documenting both the physical and chemical models, and the numerical schemes and solvers, implemented in the CFD code. Ample references to published literature describe the capabilities and inherent limitations of the software.

Application areas

FLACS is a specialised computational fluid dynamics (CFD) tool for safety applications. Major disasters continue to cause severe losses in industry and society in general, and some of the most severe accident scenarios entail fluid flow, with or without chemical reactions, as well as blast wave propagation, in complex geometries. Examples of such events include (Mannan, 2012):

  • Loss-of-containment (release) and dispersion of flammable, asphyxiating, malodorous, toxic and/or radioactive material in gaseous, liquid and/or solid form;
  • Gas explosions, vapour cloud explosions (VCEs), mist explosions, dust explosions, colliery explosions, hybrid explosions and vapour explosions (physical explosions);
  • Detonation of condensed explosives and propagation of blast waves (accidents or malicious attacks); and
  • Jet fires and pool fires.

A majority of the 100 largest property losses in the hydrocarbon industries from 1972 to 2011 involved fires and explosions (Marsh, 2012).

In their daily work, safety engineers use a variety of tools for assessing the consequences of accident scenarios, and for optimising the design of process facilities and safety measures. Examples of such models include:

  • Simple analytical expressions and empirical correlations, typically prescribed in safety standards and implemented in software packages,
  • Phenomenological tools of varying complexity,
  • Sophisticated CFD tools that account for the actual initial and boundary conditions, and solve the governing equations for conservation of mass, momentum and energy.

Computational fluid dynamics (CFD) represents the state-of-the-art in consequence assessment for flow-related accident scenarios. In some situations, the simpler models may under- or overpredict the consequences by a factor ten or more (Zalosh, 2008), even within their prescribed range of application. Regardless of the complexity of the model used, it is essential for the quality of a quantitative risk assessment (QRA), and hence for the safety and security in the facility, that safety engineers understand the underlying assumptions and inherent limitations of the tools they use, as well as the level of accuracy they can expect in the results. New users of FLACS must attend an introductory training course that conveys adequate knowledge of fluid dynamics and teaches the proper interpretation of simulation results. Beyond that, this User Manual provides comprehensive information on how to set up simulations, the modelling and theory underlying FLACS, and its limitations.

The following paragraphs summarise the main application areas of the CFD tool FLACS, and highlights key aspects of the model system.

Computational fluid dynamics

Although the governing equations for turbulent fluid flow are well established (Bradshaw, 1994), analytical solutions are primarily of academic interest, and discrete solutions by direct numerical simulation (DNS) can only be realized for idealised systems. In recent years, models based on large eddy simulations (LES) have gained increasing popularity at universities. However, within the context of simulating industrial accident scenarios, most commercial CFD tools, including FLACS, still rely on turbulence models based on Reynolds-averaged Navier-Stokes (RANS) equations, such as the $k-\epsilon$ model (Launder & Spalding, 1974), complemented with sub-grid models to account for the influence of objects that cannot be resolved on the computational grid. For turbulent reactive flows, it is necessary to add models for chemical reactions, and to couple the resulting model system (Hjertager, 1982). When it comes to describing real industrial systems, it is important for users of advanced CFD tools to keep in mind that most simulations are inherently ‘under-resolved’, and that a significant degree of sub-grid modelling is required. This implies that solutions may not converge as the spatial or temporal resolution increases, and it is important to follow the guidelines provided by the software vendor.

The model system

FLACS is an advanced tool for performing engineering calculations related to safety and security in industry and society in general. The software includes a 3-dimensional (3D) CFD code that solves Favre-averaged transport equations for mass, momentum, enthalpy ( $h$), turbulent kinetic energy ( $k$), rate of dissipation of turbulent kinetic energy ( $\epsilon$), mass-fraction of fuel ( $Y_F$) and mixture-fraction ( $\xi$) on a structured Cartesian grid using a finite volume method. The RANS equations are closed by invoking the ideal gas equation of state and the standard $k-\epsilon$ model for turbulence (Launder & Spalding, 1974). FLACS solves for the velocity components on a staggered grid, and for scalar variables, such as density, pressure and temperature, on a cell-centred grid. The accuracy of the Flacs solver is second order in space and first/second order in time. FLACS uses the SIMPLE pressure correction scheme (Patankar, 1980) extended with source terms for the compression work in the enthalpy equation for compressible flows, and the SIMPLEC scheme for non-compressible flows.

One of the key features that distinguishes FLACS from most commercial CFD codes is the use of the distributed porosity concept for representing complex geometries on relatively coarse computational meshes. With this approach, large objects and walls are represented on-grid, whereas smaller objects are represented sub-grid. The pre-processor Porcalc reads the grid and geometry files and assigns volume and area porosities to each rectangular grid cell. In the simulations, the porosity field represents the local congestion and confinement, and this allows sub-grid objects to contribute with flow resistance (drag), turbulence generation and flame folding in the simulations.

The use of CFD for consequence assessments is not limited to any particular market or industry. The basic conservation laws apply equally well to safe design and optimisation of new technology, including the emerging field of renewable energy: the use of hydrogen as an energy carrier, pipeline transportation of carbon dioxide as part of the carbon capture and storage (CCS) chain, production and use of various types of biofuel, etc. There is significant potential for increased use of CFD with respect to simulating accident scenarios in industry and society in general.

Release and dispersion

FLACS models flow in the atmospheric boundary layer (wind) by forcing profiles for velocity (wind speed and wind direction), temperature and turbulence parameters ( $k$ and $\epsilon$) on the inlet boundaries. The choice of profile should reflect the surface roughness and atmospheric stability class (Pasquill class) or the Monin-Obukhov length. The model accounts for buoyancy effects through additional terms in the momentum and turbulence model equations.

Gas explosions

While premixed combustion under constant volume or constant pressure conditions is straightforward to describe, gas explosions in complex industrial environments are complex phenomena. The key physical phenomenon to model is the positive feedback loop between expansion-generated flow and increased rate of turbulent combustion, which leads to flame acceleration, pressure build-up and generation of blast waves. A suitable starting point for the novice in the field of gas explosions is the Gas Explosion Handbook (Bjerketvedt et al., 1997).

The development of dedicated CFD codes for simulating gas explosions in realistic industrial geometries started around 1980. The combustion models implemented in explosion codes for industrial applications usually assume premixed combustion and fast chemistry. The main source of validation data are extensive series of large-scale experiments with natural gas, and the main area of application is offshore installations. The significant spread in the results from repeated explosion experiments at large scales represents a challenge for model validation (Evans et al., 1999; Skjold et al., 2013).

The current trend is to apply the same CFD tools to risk assessments for onshore processing facilities. This creates several challenges with respect to model validation. There is increasing awareness of the potential for realising deflagration-to-detonation transition (DDT) during flame propagation in large congested fuel-air clouds (HSE, 2009; Johnson, 2010; Tomlin & Johnson, 2013). Flammable mixtures containing gaseous fuels such as propane or ethylene react far more violently than methane-air mixtures. The research community has limited understanding of the effects of flow-turbulence interactions, anisotropic effects, and the influence of Landau-Darrieus, Kelvin-Helmholtz, Rayleigh-Taylor and/or Richtmyer-Meshkov instabilities on flame speed and pressure build-up, especially in large-scale industrial geometries. Hence, Gexcon invests significant resources in developing improved combustion and turbulence models that can be applied on a coarse computational mesh, and still capture the dominating mechanisms for flame propagation.

The technical reference chapter describes the combustion model used in FLACS-GasEx.

Hydrogen safety

FLACS-Hydrogen is a special variant of FLACS developed for hydrogen safety. The functionality is similar to FLACS-GasEx and FLACS-Dispersion, but limited to hydrogen as fuel.

Dust explosions

FLACS-DustEx, previously known as DESC (Dust Explosion Simulation Code), is a special variant of FLACS developed for simulating industrial dust explosions. Users define a dedicated combustion model for a specific dust from pressure-time data experimentally determined in a standard 20-litre explosion vessel (Skjold, 2007). FLACS-DustEx treats the dust cloud as a dense gas, assuming that the dispersed phase (particles) is in thermal and kinetic equilibrium with the continuous phase (typically air).

The specifics of FLACS-DustEx are not covered in this manual.

Blast wave propagation

FLACS-Blast (previously known as FLACS-Explo) is a special variant of FLACS that simulates the propagation of blast waves arising from the detonation of condensed explosives. The Blast simulator does not model the detonation process itself but a specified amount of explosive material is transformed into a high-pressure high-temperature region that is used as initial condition for the simulation. FLACS-Blast solves the Euler equations with a conservative shock-capturing scheme, the so-called flux-corrected transport (FCT) scheme (Boris & Book, 1973), together with the SOLA-ICE algorithm and a second order flux correction, instead of the SIMPLE algorithm used by the standard FLACS gas explosion simulator. Since the Euler equations are solved, it is not required to include sub-grid contributions such as turbulence. The porosity file is binarised (grid cells with less than 50% porosity are assumed to be fully blocked, the rest are fully open). FLACS-Blast treats the explosive like a bursting “balloon”, which starts with the initial condition of a sphere of high temperature and pressure gases at zero velocity. This balloon analogue method has been described previously (Brode, 1959; Ritzel and Matthews, 1997; Donahue, et al., 2004). The equation of state for the ideal gas law (ratio of specific heats equal to 1.4) is used, which corresponds to air (or any other diatomic gas). The scaling relation for the source volume is obtained from the standard TNT blast curves, so that the diameter of the “balloon” is proportional to the mass of explosive material. The initial condition is calculated based on the heat of reaction and the temperature of the combustion products (for TNT detonations), which are then isentropically compressed to a pressure of 808 bar and a temperature of around 10,000 K (consistent with the TNT blast curve). FLACS-Blast also has relationships for RDX, where the isentropically compressed state of the “balloon” is 936 bar and $13,288.6^{\circ}\mathrm{C}$. Since there is no case-specific calibration involved, it is possible to perform reasonably accurate predictions within the inherent limitation of the achievable grid resolution (Nolde & Skjold, 2010; Skjold et al., 2012; Davis & Hinze, 2014).

For limitations of FLACS-Blast see the relevant section in the Best practice chapter.

Future developments

Gexcon has plans to further develop FLACS so that it can simulate mist and spray combustion (explosions), as well as hybrid explosions. The steady increase in computational resources (speed, memory, etc.), accompanied by parallelisation of the numerical solvers, allows for faster calculations on larger computational domains. However, there is still a need for further speed-up, improved accuracy, and reduced sensitivity of the results with respect to the spatial resolution used in the simulations; technology based on adaptive mesh refinement (AMR) and hybrid parallelisation represents a promising solution for achieving this goal.

Validation and documentation

The validation and documentation process represents a fundamental challenge for developers of any model system that aspire to describe a wider range of physical phenomena, or other initial and boundary conditions than the ones that can be mapped out by a finite number of experiments. Both government bodies and industry show increasing awareness of the need to qualify models for particular applications, for instance by requiring modellers to demonstrate the capabilities of their models to reproduce results from specific sets of experiments (Ivings et al., 2007). The current trend in software development for application-specific CFD tools, such as FLACS, entails an integrated framework for model validation, implemented as a natural extension of the continuous integration and life-cycle management system for the software (Skjold et al., 2013).

Acknowledgements

The development of FLACS would not have been possible without the generous contributions received from supporting companies and government institutions throughout the years. The activity started at Christian Michelsen Institute (CMI) in 1980 with the Gas Explosion Programs (GEPs), and FLACS-86 was the first version distributed to the supporting companies.

flacs86_M24.png
The M24 compressor module represented in FLACS-86

The development of FLACS continued with the Gas Safety Programs (GSPs) and related projects up to around 2000:

  • BP, Elf, Esso (Exxon), Mobil, Norsk Hydro, and Statoil supported the development of FLACS-86 during the First GEP (1980-1986).
  • BP, Mobil, and Statoil supported the development of FLACS-89 during the Second GEP (1986-1989).
  • BP, Elf, Esso, Mobil, Norsk Hydro, Statoil, Conoco, Philips Petroleum, Gaz de France, NV Nederlandse Gasunie, Bundes Ministerium für Forschung und Technologie (BMFT), Health and Safety Executive (HSE), and the Norwegian Petroleum Directorate supported the development of FLACS-93 during the First GSP (1990-1992).
  • BP, Elf, Esso, Mobil, Statoil, Philips Petroleum, Gaz de France, HSE, and the Norwegian Petroleum Directorate supported the development of FLACS-94, FLACS-95, and FLACS-96 during the Second GSP (1993-1996).
  • BP, Elf, Exxon, Mobil, Norsk Hydro, Statoil, Philips Petroleum, Gaz de France, HSE, Agip, MEPTEC, and the Norwegian Petroleum Directorate supported the development of FLACS-97, FLACS-98, and FLACS-99 during the Third GSP (1997-1999).
  • BP, TotalElfFina (TEF), Norsk Hydro, Statoil, Gaz de France, Philips Petroleum, Mobil and supported the LICOREFLA project (2000-2001).

Since 2000, the more recent FLACS releases have been supported by various Joint Industry Projects (JIPs), funding from the European Commission (EU), the Norwegian Research Council (NFR), and support and maintenance fees from an increasing number of commercial costumers. In addition, several specialised versions of FLACS, such as DESC (Dust Explosion Simulation Code), FLACS-Dispersion, and FLACS-Hydrogen have been developed. The chronological development has been:

  • FLACS-Dispersion and FLACS-Hydrogen became available in 2001 (hydrogen had been available as a gas in FLACS since 1989, but can be purchased as a dedicated tool since 2001. FLACS-Hydrogen was strongly improved with FLACS 8.1 in 2005.).
  • FLACS v8.0 was released in 2003, including a test release of FLACS-Explo.
  • FLACS v8.1 was released in 2005.
  • DESC 1.0 was released in 2006.
  • FLACS v9.0 was released in 2008, including a test release of FLACS-Fire.
  • FLACS v9.1 was released in November 2009.
  • FLACS v10.0 was released in December 2012.
  • FLACS v10.1 was released in May 2013.
  • FLACS v10.2 was released in December 2013.
  • FLACS v10.3 was released in July 2014.
  • FLACS v10.4 was released in June 2015.
  • FLACS v10.5 was released in May 2016.
  • FLACS v10.6 was released in April 2017.
  • FLACS v10.7 was released in November 2017.
  • FLACS v10.8 was released in summer 2018.

Gexcon has also developed several in-house R&D tools, including FLACS-Aerosol and FLACS-Energy.

Gexcon is grateful to all companies, government institutions, and individuals that have participated in the development of FLACS. We intend to honour these contributions by continuing to develop the software, and thereby contribute to improved safety in the process industries.

About this manual

This User Manual describes a family of computational fluid dynamics (CFD) software products from Gexcon AS, generally referred to as FLACS:

These programs constitute a specialised CFD tool, FLACS, or 'standard FLACS', designed to study releases of flammable gas and gas explosions in complex congested geometries, both onshore and offshore.

Note
A newer version of this User Manual might be available on the FLACS User Portal.


Note
In some cases, the cached help file may be out of sync with the installed FLACS manual. If an older version of the manual opens when a newer version is expected, then the cache must be deleted. This can be achieved by removing the following directory with all of its contents:
Windows: C:\Documents and Settings\USER_NAME\Local Settings\ ...
... Application Data\assistant\gexcon\flacs_v10_9
or: C:\Users\USER_NAME\AppData\Local\assistant\gexcon\flacs_v10_9
Linux: ~/.local/share/data/assistant/gexcon/flacs_v10_9

The standard FLACS package includes the full functionality of FLACS-Hydrogen and FLACS-Dispersion, whereas DESC and FLACS-Explo are separate products. FLACS-Energy, FLACS-Fire and FLACS-Aerosol are still in-house R&D tools. The acronym FLACS (FLame ACceleration Simulator) refers to the complete package of products, whereas the term Flacs refers specifically to the numerical solver in the CFD code.

FLACS version 9.0 (FLACS v9.0) represented a major upgrade to the graphical user interfaces (GUIs), and was the first version that ran under both the Linux and Windows operating systems.

Getting started presents a detailed example for new users of FLACS, and Best practice examples contains further examples that highlight various applications of FLACS, including some of the specialised variants.
Technical reference contains technical reference material.

Printing conventions in this manual

  • The symbol '>' followed by text in typewriter font indicates command line input, e.g.:

    > command -options arguments  (general syntax for commands)
    > find -name flacs            (command line input in Linux)
    

  • The symbol '*' followed by text in typewriter font field input commands, e.g.:

    * exit yes yes

  • The symbol $\rightarrow$indicates a path through nested menu items or dialog box options, e.g.:

    File $\rightarrow$Save

    Scenario $\rightarrow$Ignition $\rightarrow$Time of ignition

  • Certain features of the software may only be accessible through text file input, and the content of a text file is also printed in typewriter font:

    THE FIRST LINE OF THE FILE ...
    THE SECOND LINE OF THE FILE ...
    ...                          ...
    
  • The format for describing keyboard and mouse input follows the pattern:

    CTRL+C

    CTRL+left_click

  • The use of bold or italic font emphasises specific words or phrases in the text.
  • The Nomenclature chapter lists the symbols and abbreviations adopted in this manual.

Special messages

Warning
Look out for the potential pitfalls pointed out by this heading!
Note
Be aware of practical information pointed out by this heading.
Remarks
Take notice of the points summarised under this heading.
See also
Follow up the additional sources of information suggested by this heading if required.

Job numbers

The typical application of FLACS is to quantify potential consequences of industrial accident scenarios involving compressible fluid flow, with or without chemical reactions. Proper characterisation of a particular problem may involve several simulations, and it is usually convenient to organise the files from related scenarios in a dedicated directory. The individual FLACS simulations are assigned job numbers, or simulation numbers, or simply jobs. In Linux, to start job number 010100 type on the command line:

> run flacs 010100

The job numbers are constructed from a six-digit string ijklmn, where traditionally:

  • ij is the project number.
  • kl is the geometry number.
  • mn is the sequence number.

The default job number used in many of the examples in this manual is 010100, i.e. project 01, geometry 01, simulation 00. However, each of the six digits in the job number may in principle take on any integer value from zero to nine, and the references to project, geometry, and sequence numbers only apply when the job numbers are derived from the file database in CASD.

What is new in this FLACS release?

FLACS version 10.9r1 contains the following improvements and bug-fixes:

CASD

The pre-processor

  • Import/export CO-files from database.
  • Support for Autocad 2018 format.
  • Improved CAD import dialogue.
  • Various improvements to graphics engine.
  • Various improvements to CAD import, especially with regards to mesh objects.
  • Added functionality to open FLACS User's Manual PDF version directly from help menu.

Flowvis 5

The post-processor

Simulators

the different simulators

FLACS Runmanager:

Running simulations:

FLACS-Risk

Gexcon's new unique software solution for 3D CFD risk modelling and visualisation:

  • Functionality to specify custom job numbers.
  • Added functionality to disable version control to improve responsiveness.
  • Ability to open exported scenarios in CASD.
  • Improvements to 3D exceedance GUI.
  • Support for diffuse leaks.
  • Functionality to modify exported scenarios in a safe and persistant manner.
  • Various improvements for FLACS-Cloud.
  • Various other improvements.

Utilities

Handling FLACS data:

FLACS-Cloud
Run your simulations on the FLACS high performance computing service - straight from the user interfaces you are used to:
  • Support for all FLACS files (setup, custom gasdata, dump, etc), except a few DustEx specific files.
  • Automatic download of results.
  • Partial download of results (incl. support in Flowvis 5).
  • Stability and responsiveness improvements.

Compatibility

Compatibility between FLACS v10.8 and v10.9

The Flacs simulators included in v10.9, include a new gas mixing rule/algorithm for Hydrogen-Inert mixtures. When enabled this rule will calculate more accurate flammability limits and laminar burning velocities of Hydrogen-Inert mixtures. This modification fixes an issue in all previous versions for these mixtures. See Significant overprediction of lower flammability limits for hydrogen+inert mixtures and Hydrogen-Nitrogen specific laminar burning velocity model for more information.


Apart from this change the Flacs simulators included in v10.9 contain only very minor changes to improve robustness and should not change results compared to v10.8.


CASD and Flowvis in v10.9 can read all scenario and presentation files, respectively, of previous versions. The new features of v10.9 (see above) are not supported in previous versions, so that input involving these will not be understood by, for example, CASD, Porcalc and Flowvis of v10.8 or earlier.


The licensing system have been updated in FLACS v10.9. If you are using a network license the licensing software installed on the license server must be updated to the latest version. See section License manager runtime software for information about how this can be done. The newest version of the licensing software is compatible with all earlier versions of FLACS (i.e. updating the licensing software on the license server should not affect earlier versions of FLACS).


For specific compatibility information related to LNG studies in the US, see section on
Compatibility between FLACS v10 and v9.1.

Compatibility between FLACS v10.7 and v10.8

The Flacs simulators included in v10.8r1 contain only very minor changes to improve robustness and should not change results compared to v10.7r2.

CASD and Flowvis in v10.8 can read all scenario and presentation files, respectively, of previous versions. The new features of v10.8 (see above) are not supported in previous versions, so that input involving these will not be understood by, for example, CASD, Porcalc and Flowvis of v10.7 or earlier.

Compatibility between FLACS v10.6 and v10.7

CASD and Flowvis in v10.7 can read all scenario and presentation files, respectively, of previous versions. The new features of v10.7 (see above) are not supported in previous versions, so that input involving these will not be understood by, for example, CASD, Porcalc and Flowvis of v10.6 or earlier.

FLACS-Risk
The algorithm for setting simulation job numbers has been made more robust. Because of the changes, when a pre-v1.1r3 Risk project is opened and re-exported with FLACS-Risk v1.1r3, the simulation job numbers may change and may no longer be consistent with previously calculated result files. FLACS-Risk with indicate this, but it may require rerunning most simulations. If you have old projects with many results, please use one of the following workarounds:

  1. Open the project in FLACS-Risk v1.1.r2.
  2. Open the project read-only. Note that this is only an issue when a re-export is done. If the existing project is only opened, it will keep the existing job numbers.

Compatibility between FLACS v10.5 and v10.6

CASD and Flowvis in v10.6 can read all scenario and presentation files, respectively, of previous versions. The new features of v10.6 (see above) are not supported in previous versions, so that input involving these will not be understood by, for example, CASD, Porcalc and Flowvis of v10.5 or earlier.

The Flacs simulators are included in v10.6 without changes compared to v10.5, except the Flacs-Fire solver: Due to the introduction of the (default) automatic DTM domain and far-field models, simulations rerun in FLACS v10.6 will likely be somewhat different, especially in the far field. In most cases the differences are expected to be modest. It is still possible to enable the full DTM domain (using the DTM domain constraint setting) in v10.6, which should give very similiar results to those obtained in v10.5. For the variable QWALL, significantly lower values may be obtained; previous results were excessively conservative.

Compatibility between FLACS v10.4 and v10.5

CASD and Flowvis in v10.5 can read all scenario and presentation files, respectively, of previous versions. The new features of v10.5 (see above) are not supported in previous versions, so that input involving these will not be understood by, for example, CASD, Porcalc and Flowvis of v10.4.

The Flacs simulators are included in v10.5 without changes compared to v10.4r2.

Compatibility between FLACS v10.4r1 and v10.4r2

Concerning file formats etc. v10.4r2 and the predecessor v10.4r1 are fully compatible. Flacs results should only change when monitor point output from Flacs-Fire is considered; here, the differences can be significant due to the improved model implemented in v10.4r2.

Compatibility between FLACS v10.3 and v10.4

The simulators included in the FLACS v10.4 package have undergone minor improvements and bug fixes. Changed simulation results have to be expected mainly due to several corrections in Porcalc, which is included with a new default version 2.8. This version is recommended, but for backwards compatibility in ongoing projects, the previous versions are included in the package. The magnitude of differences that can occur in simulation results due to the improvements in Porcalc is exemplified below.

The Flowvis versions 4.4 and 5.3, which are both included in FLACS v10.4, use different file formats for presentations. A presentation created with Flowvis version 4.4 cannot be handled in version 5.3 and vice versa.

Changes in results due to bug-fixes in Porcalc

As mentioned above, several recently discovered bugs have been fixed in Porcalc for FLACS v10.4r1. While the modelling principles are the same as before, the elimination of those mistakes in the implementation does lead to changes in the results of Porcalc and consequently also Flacs simulations.

To quantify the changes in the results, Gexcon has re-run the test suite of approximately 950 simulations that has been used for this purpose before. The simulations concern 15 different geometries, ranging from medium-scale experiments to full scale on- and offshore installations. The key output parameters of this test are the maximum overpressures in a number of monitor points for each scenario. The differences in the results arise indirectly due to changed turbulence generation and flame folding parameters based on the geometry and are shown in the plots below.

On closer examination, some of the bigger deviations can be attributed to a geometry that has a number of cylinders that are aligned with the grid in exactly the way that was affected by the implementation bug, and the ignition location is situated close to these obstructions. In this constellation the effect in turbulence generation and flame acceleration is particularly emphasised.

The effect of nested primitives should be minor for custom-built geometries and those filtered properly in the CAD export. The change is relevant (and has been discovered) when importing complex/as-built geometries into FLACS. For such geometries with many nested primitives the effect can be considerable, with the uncorrected versions of Porcalc tending to overestimate the explosion effects.

Objects inside empty space created by a left difference operation created incorrect porosity patterns in the Porcalc version prior to 2.8. This should be a relatively rare situation and not play any significant role for most practical cases.

In summary, the software changes between Porcalc 2.7.1 and 2.8 must be expected to entail noticeable changes in the results of FLACS simulations. These changes are due to the corrections of mistakes in the modelling and should therefore be seen as improvements.

Compatibility between FLACS v10.2 and v10.3

  • The simulators included in the FLACS v10.3r2 package are fully compatible with the previous release, v10.2r2, and results should not change when simulated with these two versions on the same hardware and using identical input files, except for the following types of scenarios:
    • when a setup file with the name cs<JOBNO>.SETUP is present that changes the default behaviour of the simulator but was not used when running with v10.2r2,
    • wrongly specified (too big) area leaks will trigger errors rather than warnings.
    Except for scenarios that are affected by the above changes, no differences in results are to be expected. Gexcon's test suite has confirmed that only very minor numerical variations occur. These may be caused by compiler or optimisation changes, operating system or hardware differences etc.
  • The Flowvis versions 4.4 and 5.1, which are both included in FLACS v10.3r2, use different file formats for presentations. A presentation created with Flowvis version 4.4 cannot be handled in version 5.1 and vice versa.
  • The interpolation algorithm for visualisation of surface values in 3D plots has been significantly improved in Flowvis 5.1 included in FLACS v10.3r2. It now accounts for porosities and alignment of large non grid aligned objects. Previously, the interpolation algorithm used only linear interpolation without accounting for porosities and solid walls. So in many cases this resulted in lower values (by up to a factor of 2x) being shown. The new algorithm is enabled by default. For legacy reasons, the previous interpolation algorithm from FLACS v10.2r2 is also still available by selecting 'Fast interpolation' in the menu. However, it is strongly recommended to no longer use the previous FLACS v10.2r2 version or the legacy 'Fast interpolation' option.

    Note: This change only applies to 3D surface plots. Therefore 3D volume plots and 2D cut plane plots are not affected and should not show differences between FLACS v10.2r2 and FLACS v10.3r2.

Compatibility between FLACS v10.1 and v10.2

  • The simulators included in the FLACS v10.2r2 package are fully compatible with the previous releases since v10.1r1, and results should not change when simulated with these versions on the same hardware and using identical input files, except for the following types of scenarios:
    • Flacs-Explo will give slightly different (more correct) results due to a bug-fix which improves symmetry for symmetric scenarios; the fix was first delivered with v10.2r1 and also applies to other Flacs simulator variants when using blast blocks.
  • The Flowvis versions 4.4 and 5.0, which are both included in FLACS v10.2, use different file formats for presentations. A presentation created with Flowvis version 4.4 cannot be handled in version 5.0 and vice versa.

Compatibility between FLACS v10.0 and v10.1

The FLACS v10.1r1 package is fully compatible with the previous release, v10.0r1. Results should not change when simulated with these two versions on the same hardware and using identical input files, with the following two exceptions:

  • Dispersion scenarios with a non-constant vertical wind and temperature profile may yield different results close to the outflow boundaries; v10.1 has been re-aligned with v9.1r4 regarding the vertical profiles at the inflow. If the v10.0 parameterisations are desired then the compatibility key "ABL=f240" can be used.
  • In simulations using the incompressible solver together with quiescent or low-wind initial conditions, the length of the first time step is more rigorously limited than in previous versions of FLACS.

Compatibility between FLACS v10 and v9.1

Scenarios that have been created in FLACS v9.1 can be run in FLACS v10. However, the porosities for the v9.1 scenarios should be recalculated with the Porcalc included in the latest FLACS v10 package before running them with the v10 simulator. This is due to an improvement in the porosity calculation by Porcalc, with a related change in the production of sub-grid turbulence. You can expect to see slight changes in the results compared to FLACS v9.1. In validation runs for hydrogen, and other highly reactive fuels, more accurate and reliable over-pressures were obtained. In simulations of the BFETS and HSE modules (with different congestion levels and ignition points) the results were nearly unchanged.

It is possible to use the sub-grid turbulence production model of FLACS v9.1r4 in FLACS v10 by using the key STF=f228. When this key is used, the log file (rt.dat3, or the tt-file on a Linux computer) will contain the line # SUBGRID TURBULENCE FACTOR STF="f228:0.600", which indicates that the sub-grid turbulence production model of FLACS v9.1r4 (simulator version flacs2.2.8) is used, even though the log file contains the current simulator version, e.g. FLACS, Version 2.5.2, July 2014, Gexcon AS.

Attention
When using this key you must use porosity files produced with the corresponding Porcalc version (v2.6), not the latest one included in v10!

Due to other changes and improvements, the results may still be slightly different from those obtained with FLACS v9.1r4.

Attention
Currently only FLACS v9.1r2 has been approved by PHMSA for LNG vapor dispersion modelling scenarios according to US federal regulations (49 CFR 193.2059). A formal request to also have FLACS v10.x approved has been submitted, but this is potentially a lengthy process and we don’t know when the approval will be granted. Until v10.x has been approved, FLACS v9.1r2 is the latest approved version and is the version that should be used for all US LNG studies subject to 49 CFR 193.2059.

FLACS version history

The following table gives an overview of the versions of the major FLACS programs which were part of the various FLACS releases.

FLACS version history
FLACS Flacs CASD Flowvis Porcalc geo2flacs RunManager Risk
10.9r12.6.07.12.05.10.02.8.22.8.01.71.1.5
10.8r22.5.77.11.05.92.8.22.7.01.61.1.4
10.8r12.5.77.11.05.92.8.22.7.01.61.1.4
10.7r22.5.67.10.15.82.8.12.6.11.51.1.3
10.72.5.67.105.82.8.12.61.51.1.2
10.6r32.5.57.95.72.8.12.51.41.1.1
10.6r22.5.57.85.62.8.12.51.41.1
10.62.5.57.75.52.8.12.51.41.0
10.52.5.47.64.4, 5.42.8.10.9.11, 2.41.4
10.4r22.5.47.54.4, 5.32.80.9.11, 2.31.4
10.42.5.37.44.4, 5.22.80.9.11, 2.21.4
10.3r22.5.27.34.4, 5.1.12.7.10.9.11, 2.1.11.3
10.32.5.27.34.4, 5.1 2.7.10.9.11, 2.11.3
10.22.5.17.24.4, 5.0 2.7.10.9.11, 2.01.2
10.12.5.07.14.4, 5.0b2.70.9.11, 2b1.2
10.02.4.07.04.32.70.9.111.2
9.1r42.2.84.10, 6.34.22.60.9.111.1
9.1r32.2.84.9, 6.24.12.60.9.101.1
9.1r22.2.84.9, 6.24.12.50.9.101.1
9.12.2.84.8, 6.14.12.50.9.91.1
9.02.2.74.7, 6.04.02.50.9.71.0

Feedback from users

Feedback on the content in this manual is most welcome, and FLACS users may submit their comments or suggestions by e-mail to: flacs.nosp@m.@gex.nosp@m.con.c.nosp@m.om
When submitting comments or suggestion to the content of the manual, or when pointing out misprints in the text, please indicate the relevant page numbers or sections, and the corresponding version of the manual (date issued).

flacs-small.png

FLACS v10.9 - Mon May 6 2019 Copyright © 2019 Gexcon AS FLACS