RR789 - LNG source term models for hazard analysis

1 Health and Safety Executive LNG source term models for hazard analysis A review of the state-of-the-art and an approach to model assessment Prepared by the Health and Safety Laboratory for the Health and Safety Executive 2010 RR789 Research Report Health and Safety Executive LNG source term models for hazard analysis A review of the state-of-the-art and an approach to model assessment Dr DM Webber, Dr SE Gant, Dr MJ Ivings & SF Jagger Health and Safety Laboratory Harpur Hill Buxton Derbyshire SK17 9JN In 2006, the Health and Safety Laboratory (HSL) undertook a research project to develop tools for the National Fire Protection Agency (NFPA) Liquefied Natural Gas Technical Committee to evaluate liquefied natural gas (LNG) dispersion models . The work was commissioned by the Fire Protection Research Foundation (FPRF) and delivered a model Evaluation Protocol (Ivings et al.)

2 , 2007) which included a checklist of model evaluation criteria and a structure for complete model evaluation. A partial evaluation of some common current dispersion models was also carried out. Subsequently, FPRF let a subcontract to HSL to create a database of full scale experimental trials and wind tunnels tests which can be used to validate LNG dispersion models . This work has recently been completed. The current project is concerned with the assessment of source term models for LNG spills that are used to provide an input to the LNG dispersion models . This work has been jointly funded by the UK Health and Safety Executive (HSE) and FPRF and aims to provide a state-of-the-art review of source term models which predict the early development of a release of LNG and an approach for assessing the adequacy of such models .

3 This report and the work it describes were jointly funded by the Health and Safety Executive (HSE) and the Fire Protection Research Foundation (FPRF). Its contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE or FPRF policy. HSE Books Crown copyright 2010 First published 2010 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means (electronic, mechanical, photocopying, recording or otherwise) without the prior written permission of the copyright owner. Applications for reproduction should be made in writing to: Licensing Division, Her Majesty s Stationery Office, St Clements House, 2-16 Colegate, Norwich NR3 1BQ or by e-mail to ACKNOWLEDGEMENTS The authors would like to thank Ted Williams, American Gas Association and Yinqing Liu & Weiping Dai, Trinity Consultants for providing information on the SOURCE5 model and Graham Tickle, ESR Technology for providing information on the GASP model .

4 We would also like to thank Weiping Dai and Graham Tickle for reviewing our assessment of the SOURCE5 and GASP models respectively. The authors would like to thank the US Fire Protection Research Foundation and UK Health and Safety Executive for jointly sponsoring this work. ii CONTENTS 1 1 Background .. 1 source term model assessment .. 2 2 THE PHYSICS OF LNG SPILLS .. 4 LNG .. 4 LNG spills: processes leading to a flammable cloud .. 6 6 Pool formation .. 9 Vaporisation from within the 9 Rapid phase transitions .. 10 Pool spread .. 11 Pool vaporisation .. 11 LNG as a hydrocarbon 13 Transition between source and dispersion.

5 13 Conclusions .. 14 3 LNG source TERM 15 Raj and Kalenkar .. 15 Opschoor .. 15 SOURCE5 .. 15 16 16 16 SafeSite3G .. 17 17 17 17 ALOHA .. 18 LSM90/LPOOL .. 18 18 ABS Consulting 18 LNGMAP .. 19 Sandia .. 19 19 Brambilla and 20 Discussion .. 20 4 source TERM model 22 22 Factors affecting source 22 Previous validation data reviews .. 24 Data for LNG spills on 24 iii Data for LNG spills on 25 Example datasets .. 26 LNG source term model validation .. 31 5 LNG source TERM model 39 General 39 Application to gas dispersion 40 Application to source term 41 6 STATE OF THE ART 46 46 Early integral models .

6 47 Pool spread: Improvements in the state of the art .. 52 Pool vaporisation .. 60 source term model dispersion model 66 Summary .. 67 7 69 8 71 9 APPENDIX A model ASSESSMENT REPORTS FOR GASP AND 79 iv EXECUTIVE SUMMARY Background The dispersion of releases of hazardous fluids through from loss of containment to dilution below hazardous levels can be simply considered as comprising two stages: source term formation and atmospheric dispersion. The former occurs immediately after release when the behaviour of the fluid is dominated by conditions under which the fluid was stored and the particular conditions of release. Further downstream, as the influence of the source decays, the atmosphere becomes increasingly important and controls fluid behaviour.

7 In LNG hazard assessments these two stages are usually modelled separately by a source term model and a dispersion model . The output from the source term model , specifying the state of the fluid at that stage, is used as input to the dispersion model . Assessing the appropriateness of these models is an important but complex problem. While dispersion modelling has received much attention over the years, the assessment and development of source term models has received comparatively less attention, though its importance in the overall release process is widely recognized. This is probably due to the very complex and variable behaviour during this stage and the difficulty of obtaining definitive experimental data close to the source .

8 In 2006, the Health and Safety Laboratory (HSL) undertook a research project that led to the development of a model Evaluation Protocol (MEP) for LNG dispersion models (Ivings et al., 2007) which included qualitative and quantitative evaluation criteria and a structure for complete model evaluation. A partial evaluation of some common current dispersion models was also carried out. Subsequently, the US Fire Protection Research Foundation (FPRF) let a subcontract to HSL to create a database of full scale experimental trials and wind tunnels tests which can be used to validate LNG dispersion models . This work has recently been completed. The current work has been jointly funded by the UK Health and Safety Executive (HSE) and FPRF, and aims to provide a state-of-the-art review of source term models which predict the early development of a release of LNG and an approach for assessing the adequacy of such models .

9 The main focus is on models of pool spread and vaporisation. Objectives The objectives of this project were: To review the physics of accidental releases of LNG and in particular the spread and vaporisation of spills of LNG on land and water. To compile a list of available source term models in widespread use to predict the spread and vaporisation of LNG. To carry out a review of data for the validation of LNG source term models . To develop a systematic approach for assessing source term models based on scientific assessment, verification and validation, including recommendations on how the available data, and any future data, can be used to validate the models . v To carry out an assessment of the source term models SOURCE5 and GASP using the developed assessment methodology.

10 To undertake a critical assessment of the state of the art and make recommendations for further model development and/or validation. Main Findings This project has led to the development of a methodology for assessing the suitability of LNG source term models for providing an input into LNG dispersion models for use in hazard assessments. The approach is based on the EU SMEDIS project (Carissimo et al., 2001, Daish et al., 2000), and is therefore similar to the LNG dispersion MEP (Ivings et al., 2007) comprising the three key stages of: scientific assessment, verification and validation. The assessment methodology is mainly applicable to pool spread and vaporisation models . The key difference between this model assessment approach and the dispersion model MEP is that the former does not include a structured approach to validation including quantitative assessment criteria.