A multivariable model for estimation of vapor cloud explosionoccurrence possibility based on a Fuzzy logic approach for flammablematerialsM. Javidia, B. Abdolhamidzadeha, G. Reniersb,c,d,*, D. RashtchianaaCenter for Process Design, Safety and Loss Prevention (CPSL), Chemical and Petroleum Engineering Department, Sharif University of Technology,Tehran, IranbSafety and Security Science Group, Faculty TBM, TU Delft, Jaffalaan 5, 2628 BX Delft, The NetherlandscARGoSS, Faculty TEW, University of Antwerp, 2000 Antwerp, BelgiumdCentre for Economics and Corporate Sustainability, KU Leuven, Campus Brussels, 1000 Brussels, Belgiuma r t i c l ei n f oArticle history:Received 17 May 2014Received in revised form6 November 2014Accepted 7 November 2014Available online 26 November 2014Keywords:Vapor cloud explosion (VCE)ProbabilityPossibility estimationFuzzy algorithmEvent treea b s t r a c tIn this paper, a new method based on Fuzzy theory is presented to estimate the occurrence possibility ofvapor cloud explosion (VCE) of flammable materials. This new method helps the analyst to overcomesome uncertainties associated with estimating VCE possibility with the Event Tree (ET) technique. In thismulti-variable model, the physical properties of the released material and the characteristics of thesurrounding environment are used as the parameters specifying the occurrence possibility of interme-diate events leading to a VCE. Factors such as area classification, degree of congestion of a plant andrelease rate are notably affecting the output results. Moreover, the proposed method benefits from ex-perts opinions in the estimation of the VCE possibility. A refrigeration cycle is used as the case study andthe probability of VCE occurrence is determined for different scenarios. In this study, sensitivity analysisis performed on the model parameters to assess their effect on the final values of the VCE possibility.Furthermore, the results are compared with the results obtained using other existing models. 2014 Elsevier Ltd. All rights reserved.1. IntroductionEvents caused by the catastrophic release of a flammable sub-stance in the surrounding environment, such as a storage tankrupture or a hole in a pipe, can lead to a variety of outcomes. Theoccurrence of these outcomes depends on factors such as operatingconditions of the source, type of released material and surroundingconditions. Events such as pool fire, fire ball, Vapor Cloud Explosion(VCE), flash fire, and BLEVE are some of the possible outcomes.Vapor cloud explosions (VCEs) are a major hazard in industrialplants where large amounts of flammable materials are stored orprocessed. In the last few decades, several VCEs have occurred inprocess plants resulted in almost total destruction of the thoseplants (Sharma et al., 2013). Several studies about past accidentsindicate the importance of VCEs (Abdolhamidzadeh et al., 2011;Sharma et al., 2013; Salzano and Cozzani, 2005). Today, Quantita-tive Risk Assessment (QRA) has become an efficient tool in decisionmaking for safety and process experts. To assess the risk of anyaccident scenario, it is necessary that the probability of that eventbe determined in addition to estimating the probable conse-quences. Event tree analysis is known to be a standard tool forcalculating the frequency of incident outcomes. However, the ex-istence of many effective and sometimes neglected parameterscauses uncertainties in the calculations of event frequencies. Abetter understanding of the influencing factors for each accidentscenario will lead to a more precise estimation of eventfrequencies.Several studies on uncertainties in event trees have been per-formed which indicate the importance of this tool for incidentfrequencies estimation (Neri et al., 2008; Catalyurek et al., 2010;Julwan Hendry Purba, 2014).For instance, while developing conventional event trees, thetype of released material does not play a role in determining theintermediate probabilities that build the overall structure on theevent tree. Nonetheless, it is proven in several studies such as* Corresponding author. Safety and Security Science Group, Faculty TBM, TUDelft, Jaffalaan 5, 2628 BX Delft, The Netherlands.E-mail address: genserik.reniersua.ac.be (G. Reniers).Contents lists available at ScienceDirectJournal of Loss Prevention in the Process Industriesjournal homepage: 2014 Elsevier Ltd. All rights reserved.Journal of Loss Prevention in the Process Industries 33 (2015) 140e150Abdolhamidzadeh et al. (2011) that the type of material has a sig-nificant effect on the outcome of an event tree.Badri et al. (2013) proposed a model based on eventtreeanalysisto calculate the vapor cloud explosion frequency, in which theyhave tried to consider the effect of various parameters influencingthe frequency of events. Their idea was to convert the unit intosubsets such that the material properties and operating conditionsin each specific subset are constant and for each subset, the VCEfrequency is ca