Ingénierie de la sécurité appliquée à des applications hydrogène-énergie

Abstract : Since a few years, hydrogen appears as a credible energy-vector. AREVA Stockage d’Energie develop in this framework energy-storage solutions allowing to transform electrical energy into chemical energy (here hydrogen) to store it. However, hydrogen applications are still considered dangerous, so that is feared a hazardous event like explosion which could occur if a hydrogen leakage happened. And it should be recognized that hydrogen leaks can produce extended explosive clouds because of the broad flammability range, and that hydrogen-air mixtures ignite extremely easily and burn/explode fast and violently. As compared to other fuels (ex: hydrocarbons), the flammability range is 5-10 times larger, the minimum ignition energy is 5-10 times smaller and the maximum burning velocity is also about 5-10 times larger. However, due to its physical properties, hydrogen may offer some appreciable advantages in terms of mitigation: because the flammable mixtures are much lighter than air, they disperse rapidly. Design engineers have to face this reality and the logical consequence is that the safety demonstration has to be very strong and clearly understandable. Safety is the essential issue for the introduction of hydrogen applications on the market. If the risk is unacceptable (too near to important issues, sensitive environment…), Design engineers should be able to change the process and them to reassess the risk. The objective of this study is to establish a tool that permit to do that. In a way, a safety engineering tool is aimed to be created. The main points of this study are : - The design of a risk matrix: she must include both the probability and the severity of the accident. The specific case of AREVA Stockage d’Energie is presented that takes into consideration the geometry of the application studied and probabilities that reflects the real risks (here: ATEX risk). But the matrix could be formed with different criteria. - Identification of the “Critical events” or EC (Laurent, 2011): for hydrogen applications, the critical events are principally leaks and secondarily bursting. The EC is the ultimate event of a succession of events that are linked together by cause-and-effect relations. Once the EC occurs the series of following events is judged inexorable, automatic and with cause-and effect relation between the events but those events follow a temporal sequence. The probability of the accident is the one of the EC ; - Risk assessment: For each EC, a fault tree and an event tree are associated. The fault tree permit to calculate the EC probability if the probabilities of the initial events are known. In the framework of the major risk where the precision required is average, generic database are used. Those database, rather old and related to hydrocarbon technologies, are not adapted to the sharpness of the tool developed and neither to hydrogen objects. Since the feedback is also limited, an approach based on the extension through the upstream of the fault tree have been developed; however, some keys probabilities are still missing. Otherwise the methods of consequences evaluation used for major accidents are also not adapted. To correctly assess the consequences of an accident and the sizing of the safety barriers (vents, flowrate limiter…), the characteristics of the flammable cloud must be known precisely. Experiments are realized and will allow the development of the specific tool box for the evaluation of the consequences.
Document type :
Conference papers
Complete list of metadatas

Cited literature [9 references]  Display  Hide  Download
Contributor : Gestionnaire Civs <>
Submitted on : Tuesday, August 28, 2018 - 11:14:10 AM
Last modification on : Wednesday, September 5, 2018 - 2:50:07 PM
Long-term archiving on : Thursday, November 29, 2018 - 2:50:00 PM


Files produced by the author(s)


  • HAL Id : ineris-01863200, version 1



Audrey Duclos, Christophe Proust, Jérôme Daubech, Franck Verbecke. Ingénierie de la sécurité appliquée à des applications hydrogène-énergie. 16. Congrès de la Société Française de Génie des Procédés (SFGP 2017), Jul 2017, Nancy, France. ⟨ineris-01863200⟩



Record views


Files downloads