GNGTS 2018 - 37° Convegno Nazionale

GNGTS 2018 S essione 2.3 541 NATECH QUANTITATIVE RISK ASSESSMENT: THE PATH FORWARD E. Salzano Dipartimento di Ingegneria Civile, Chimica, Ambientale e dei Materiali - Università di Bologna, Italy Introduction. Public perception of the disaster potential derived from the interaction of natural hazards with industrial installations has strongly increased in the last decades (Krausmann et al. , 2011; 2016; Salzano et al. , 2013). When process equipment containing large amount of hazardous material are structurally damaged, the loss of containment is likely, eventually triggering severe, additional accidental scenarios such as fires, explosions or toxic dispersions. These types of accidents are known as NaTech events (Natural Hazard Triggering Technological Accidents) (Krausmann et al. , 2016). For effective Natech risk reduction to take place, the risk first needs to be identified and assessed. Hence, there is a strong need for specific assessment methodologies (Fabbrocino et al. , 2005; Campedel et al. , 2008), as also recognised by the European Commission and more specifically by the last version of Seveso Directive, which specifically requires for the Natech risks to be evaluated. Despite these observations, reliable Natech risk assessment tools are still lacking unless academic applications, sometimes adopted by public authorities. That is mainly due to the strong complexity of the algorithm, which must implement and include different knowledge and expertise from geophysicist, structural engineers, industrial engineers and chemical engineers. The Aripar-GIS software was one of the first applications of Quantitative Area Risk Analysis techniques in the evaluation of risks for extended industrial area, developed at University of Bologna. It allows the calculation of individual and societal risk originating from multiple risk sources due to both fixed installations and transportation systems. The software is supported by a geographical information system (GIS) platform that allows the production of risk maps. As described in Campedel et al. (2008), the software now includes natural events and domino effects (a central issue in Natech), following the flow-chart defined in Fig.1. Other codes are the latest version of RiskCurves (TNO, 2018) as recently applied in the FP7 EU project STREST (Harmonized Approach to Stress Tests For Critical Infrastructures Against Low-Probability High-Impact Natural Hazards, 2013-2016) and the in-house code described in Fabbrocino et al. (2005). The Aripar-Gis code has been applied to flooding and lightnings. Besides, the algorithm by Fabbrocino et al. has been also applied to volcanic events, earthquake and tsunami, within the activity of the Italian Reluis research activity. So that said, Quantitative RiskAnalysis (QRA) is a complex and time-consuming task, often involving large number of experts and relevant costs. This is not affordable for large part of private or public stakeholders unless mandatory actions. Selecting relevant equipment, choosing which accident scenarios are applicable, determining their frequencies and modelling the physical effects are just a few of the hurdles a user finds on his way when performing a QRA. These issues clearly affect the overall figures as several arbitrary choices and simplifications are necessarily introduced. When Natural events are introduced, these difficulties are even more consistent. Hence, some clear and simple indications are needed, as proposed in the next section. Natural hazard. Probabilistic hazard analysis (PSHA) is the methodology to assess the exceedance probability of different thresholds of hazard intensity, at a specific site or region in a given time period. The analysis has been first applied by Cornell for earthquakes and may be extended to other natural events as flooding, storms, tsunami or other natural events. PSHA has been extensively debated over the years. However, criticisms are mainly related to the methodology rather than the numerical inputs used by practitioners (Mulargia et al. , 2012). The probability function given by PSHA can be combined with fragility function for the vulnerability of involved industrial equipment (see next section). Also, the cited fragility are rare and with large uncertainties.