Antonio C. Caputo Mem. ASME Department of Engineering, Roma Tre University, Via Vito Volterra 62, Rome 00146, Italy e-mail: antonio.caputo@uniroma3.it Fabrizio Paolacci Mem. ASME Department of Engineering, Roma Tre University, Via Vito Volterra 62, Rome 00146, Italy e-mail: fabrizio.paolacci@uniroma3.it Oreste S. Bursi Mem. ASME Department of Civil, Environment and Mechanical Engineering, University of Trento, Via Mesiano 77, Trento 38123, Italy e-mail: oreste.bursi@unitn.it Renato Giannini Department of Architecture, Roma Tre University, Via Aldo Manuzio 68 L, Roma 00153, Italy e-mail: renato.giannini@uniroma3.it Problems and Perspectives in Seismic Quantitative Risk Analysis of Chemical Process Plants Earthquakes represent a class of natural-technical (NaTech) hazards which in the past have been responsible of major accidents and significant losses in many industrial sites. However, while codes and standards are issued to design specific structures and equip- ment in both the civil and industrial domain, established procedures for seismic quantita- tive risk assessment (QRA) of process plants are not yet available. In this paper, a critical review of seismic QRA methods applicable to process plants is carried out. Their limitations are highlighted and areas where further research is needed are identified. This will allow to refine modeling tools in order to increase the capabilities of risk analy- sis in process plants subjected to earthquakes. [DOI: 10.1115/1.4040804] Keywords: process plants, quantitative risk analysis, Na-Tech events, seismic vulnerability 1 Introduction 1.1 Background and Motivation. Besides process-related hazards, chemical process plants (CPPs) are vulnerable to natural hazards, such as earthquakes and hurricanes, which may trigger technological accidents usually referred to as natural- technological (NaTech) events, leading to equipment damage, release of dangerous substances, disruption of services and infra- structural, life, and economic losses [1]. The tremendous impact of NaTech events was demonstrated by the recent Tohoku earth- quake and the following Fukushima disaster in 2011. The problem is quite relevant as up to 5% of industrial accidents, involving the release of dangerous substances, which are triggered by natural hazards [2]. A major natural hazard for process plants is undoubt- edly seismic events which, therefore, have to be systematically included in current risk assessment procedures. The problem is especially relevant for plants classified as major-risk facilities according to the Seveso-III Directive [3], which regulates the con- trol of major accident hazards involving dangerous substances. The Seveso-III Directive does not explicitly consider natural haz- ards, even though thousands of major-risk European facilities are located in areas of medium to high seismicity and often near pop- ulation centers. As a result, there are several open research ques- tions related to quantitative risk assessment (QRA) for CCP. Recalling that QRA represents an analytic evaluation of risk by rigorous, replicable methods evaluated under agreed protocols of an expert community and peer-reviewed to verify the underlying assumptions, some of these questions are listed herein:  The existing procedures for QRA of process plants, including nuclear ones, cannot adequately account for the impacts of natural hazards. In fact, Na-Tech events often represent a common cause leading to the simultaneous occurrence of several interacting faults, which can often interfere with res- cue operations. Conversely, the traditional QRA of process plants starts with the analysis of a single failure or a loss of containment (LOC) in a single equipment; in this situation, it is much easier to define a few logic models (event trees and fault trees) representing accident dynamics.  To facilitate probabilistic Na-Tech risk analysis, some meth- odologies have been recently proposed. These are based on the estimation of on-site natural hazard parameters, the deter- mination of damage probabilities of plant units, and the assessment of probability and severity of potential Na-Tech events [4]. Nonetheless, these analyses usually rely on approximate fragility curves and equipment damage models, which were largely evaluated in the sixties [5], and, there- fore, are not suitable for more recent equipment.  Critical components like tanks, pipe elbows, and bolted flanges exhibit unusual and very complex phenomena. For instance, local wall thinning in pipe elbows is caused by mechanisms such as flow-accelerated corrosion and liquid droplet impingement erosion; for these reasons, wall thick- ness inspection is required and difficult to perform. FE mod- eling to detect wall thinning or local flaws and on-site system identification based on dynamic testing is still a challenge, and requires long inspections or large computing time [6].  Chemical process plants can be thought as part of the net- work of infrastructures, characterized by high consequence risk of complex systems with limited knowledge behavior. As such, resilience has emerged as a fundamental concern Manuscript received December 3, 2017; final manuscript received July 5, 2018; published online December 14, 2018. Assoc. Editor: Tomoyo Taniguchi. Journal of Pressure Vessel Technology FEBRUARY 2019, Vol. 141 / 010901-1 Copyright V C 2019 by ASME