Quantitative assessment of safety barrier performance in the prevention of domino scenarios triggered by fire

Highlights

• We developed a methodology for the quantitative assessment of safety barriers.

• We focused on safety barriers aimed at preventing domino effect triggered by fire.

• We obtained data on effectiveness and availability of the safety barriers.

• The methodology was exemplified with a case study of industrial interest.

• The results showed the role of safety barriers in preventing fired domino escalation.

Abstract

The evolution of domino scenarios triggered by fire critically depends on the presence and the performance of safety barriers that may have the potential to prevent escalation, delaying or avoiding the heat-up of secondary targets. The aim of the present study is the quantitative assessment of safety barrier performance in preventing the escalation of fired domino scenarios. A LOPA (layer of protection analysis) based methodology, aimed at the definition and quantification of safety barrier performance in the prevention of escalation was developed. Data on the more common types of safety barriers were obtained in order to characterize the effectiveness and probability of failure on demand of relevant safety barriers. The methodology was exemplified with a case study. The results obtained define a procedure for the estimation of safety barrier performance in the prevention of fire escalation in domino scenarios.

Introduction

Domino accident scenarios triggered by the escalation of fires were responsible of severe accidents that affected the chemical and process industry [1], [2], [3], [4], [5]. Past accident data analysis confirmed that in more than half of the industrial accidents involving a domino effect occurred in the past fifty years escalation was triggered by a primary fire. Secondary targets more frequently affected by escalation were pressurized tanks, atmospheric tanks, process vessels and pipelines [4], [6], [7].

The awareness of the hazards posed by domino effect led to important efforts aimed at the prevention of such scenarios. In the European Union, the legislation on the control of major accident hazard (the so-called “Seveso-III” Directive, 2012/18/EU [8]) includes measures to assess, control and prevent domino effect [3], [9], [10]. Moreover, several technical standards introduce the use of protective systems or barriers to reduce the likelihood or possibility of domino events. In industrial facilities where such hazard is present, protections from escalation is usually obtained adopting multiple safety layers [11] that can include: the basic process control system, safety instrumented systems, passive and active devices, safety shutdown systems, protective systems (post-release actions) and emergency response plans.

The specific feature of escalation due to fires is the time lapse present between the start of secondary events with respect to the start of the primary fire [4], [12], [13], [14]. In other escalation scenarios, as those triggered by overpressure or fragments, the secondary scenarios start almost simultaneously to the primary event. The delay in the start of secondary events in escalation triggered by fire is due to the damage mechanism of secondary vessels when exposed to fire. Actually, time is needed before the temperatures of the shell and of the internal fluid are able to jeopardize the structural integrity of the target vessels [15]. This time lapse, occurring between the start of the primary fire and the failure of the secondary equipment is generally termed “time to failure” (ttf) [15], [16], [17]. The ttf represents a key parameter to describe the resistance of equipment to external fires. The ttf depends on both the characteristics of the primary fire scenarios and the features of the secondary equipment involved in the fire [3], [9], [18], [19], [20]. A key point in the assessment of escalation probability in fire scenarios is that in most cases both factors may be modified by the installation of mitigation barriers and by appropriate emergency measures.

Therefore, an accurate assessment of escalation probability needs to include the analysis of the available fire protection systems and safety barriers. However, an exhaustive approach to the quantitative assessment of protection layers relevant to the prevention or mitigation of fired domino effect is still lacking. Besides, a comprehensive approach to the quantitative evaluation of the performance of all categories of protection layers (passive, active, procedural) in reducing the probability of escalation still represents an open issue.

The present study aims at the integration of a systematic quantitative analysis of safety barrier performance with probabilistic models for the assessment of escalation developed in previous studies [12], [15]. A methodology to assess the performance of safety barriers in the prevention of escalation was developed. The performance of active, passive and procedural safety barriers for escalation prevention was assessed considering both availability and effectiveness, introducing a LOPA (Layer of Protection Analysis) approach. A database of expected performance reference data was obtained for standard safety barriers adopted in escalation prevention in different types of facilities. Equipment vulnerability models based on probit functions were integrated with the LOPA results. Modified escalation probabilities, including the influence of safety barriers, were thus obtained. The approach allowed assessing the reduction in escalation probability provided by each protection layer as well as by the overall system of safeguards implemented. The application to a case-study allowed the exploration of the features and potentialities of the methodology.