A fuzzy set analysis to estimate loss intensity following blast wave interaction with process equipment Ernesto Salzano a , Valerio Cozzani b, * a Istituto di Ricerche sulla Combustione, CNR, via Diocleziano 328, 80125 Napoli, Italy b Dipartimento di Ingegneria Chimica, Mineraria e delle Tecnologie Ambientali, Universita ` degli Studi di Bologna, viale Risorgimento n.2, 40136 Bologna, Italy Received 22 June 2005; received in revised form 13 August 2005; accepted 15 August 2005 Abstract Blast waves are able to produce structural damage to process equipment even at great distances from the source point of an explosion. A loss of containment may follow and, if hazardous substances are released, relevant secondary scenarios may be triggered, resulting in domino effects. The present study was focused on the assessment of the expected structural damage and of the associated intensity of loss of containment of process vessels loaded by blast waves. Hence, a knowledge-based fuzzy set analysis was used to assess the expected overall probability of occurrence of different damage states defined for several categories of process equipment items. The fuzzy approach was also used to obtain specific threshold values for the escalation sequences (domino effects), taking into account the hazard due to the expected secondary scenarios caused by the loss of containment following blast wave impact. q 2005 Elsevier Ltd. All rights reserved. Keywords: Domino effects; Explosion; Threshold values; Quantitative risk analysis; Probit analysis; Fuzzy set analysis 1. Introduction Blast waves are able to produce structural damages to process equipment even at great distance from the source point of the explosion (Lees, 1996). Loss of containment (LOC) may follow the structural damage, possibly triggering relevant secondary scenarios (domino effects). Although several studies were dedicated to the detailed analysis of blast wave damage to process equipment, only few simplified models are available for the assessment of equipment damage by blast waves in the framework of quantitative risk analysis and, in particular, of domino effect (Bagster & Pitblado, 1991; Eisenberg, Lynch, & Breeding, 1975; Khan & Abbasi, 2001). Previous studies carried out by the current authors were dedicated to the definition of simplified models and of threshold values for structural damage following blast wave interaction with process equipment (Cozzani & Salzano, 2004a; Cozzani & Salzano, 2004b; Cozzani & Salzano, 2004c). In particular, a probabilistic model for structural damage was developed for several categories of process equipment, relating the damage probability to incident (or side-on) static over- pressure. However, these studies were not specifically addressed to the estimation of damage extension and of the associated LOC intensity that may follow blast wave impact. Nevertheless, an assessment of the LOC intensity from the damaged item is of fundamental importance to evaluate the possibility and the relevance of escalation sequences that may lead to domino accidental scenarios. As a matter of fact, the severity of the secondary event is mainly influenced by the LOC intensity. Thus, the credibility of an escalation leading to a domino accident is highly dependent on the LOC intensity following structural damage. In spite of this, to the author’s knowledge, no engineering correlation is given to estimate the LOC intensity after the damage caused by a blast wave. The present study, was focused on the assessment of the expected LOC intensity caused by blast wave impact. Journal of Loss Prevention in the Process Industries 19 (2006) 343–352 www.elsevier.com/locate/jlp 0950-4230/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jlp.2005.08.002 * Corresponding author. Tel.: C39 051 2093141; fax: C39 051 581200. E-mail addresses: salzano@irc.cnr.it (E. Salzano), valerio.cozzani@ unibo.it (V. Cozzani).