Determinação Do Pfd De Prd, Usando o API 581 Como Referência

A company's risk analysis process usually begins with a qualitative analysis of the incident scenarios and, according to the classified risk level, is completed with semi-quantitative analysis, usually using the LOPA methodology (Layers of Protection Analysis). The objective is to ensure the maximum possible accuracy in risk assessment, as well as better define independent protection layers (IPL) for mitigation. Among traditional IPLs for risk mitigation are pressure relief devices (PRD), mostly represented by PSVs. For these devices, their mitigation capacity is directly dependent on their probability of failure on demand (PFD), for which is normally adopted the value of 1E-02, as some of the area's literature suggests. However, every organization is faced with PSVs installed in systems with quite different characteristics in terms of operating conditions (pressure class and temperature) or types of fluids (clean or obstructive). Faced with this, the question arises. How can we use a generic value for probability failure on demand for relief devices that have different failure modes?

Braskem then started technical studies, in order to identify values more appropriate for the characteristics of each system. The standard API 581, in Chapter 7, describes a methodology for calculating PFD for PRD where it considers some factors, such as: inspection interval, categories of service severity, type of relief device, alignment of relief, environmental factor, effectiveness of inspection and tests programs, test result, pressure set point and maximum system feasible pressure.. With this, the Process Safety team understood that this methodology applied to the company's systems and should be the methodology used to define the PFD of each PRD. After defining the use of the methodology, an Excel spreadsheet was developed to allow the multidisciplinary team participating in the LOPA to quickly calculate a PFD value of each PRD to be used in the studies without affecting the time of analysis.

Using the concepts of management of change, Braskem Process Safety specialists understood the need to train participants in LOPA studies. In order to implement this strategy and new PSVs approaches to LOPAs, some presentations were performed to professionals from process engineering, production and maintenance. Those presentations aimed to clarify technical approach and possible misunderstands. With the implementation of this methodology, the PFD of each PRD is better suited to the operating conditions, guaranteeing a more accurate risk value for each scenario. With this, the scenarios are being reviewed within the five-year cycle of risk analysis for the correct characterization of risk. Some important aspects have already been identified. For example: Reduction in the interval between PSVs tests, where possible, can be performed for risk mitigation, which previously was only possible with the inclusion of a new IPL, another PRD or a new one Safety Instrumented Function (SIF); evaluation of campaign increases for systems or units so that the risk level of the scenarios stays within acceptable criteria, etc.

The Learning from this new approach has been continuous and the perception of a more precise risk evaluation is solid. However, some gaps and possible uncertainties regarding to PSVs adoption are still facing in industry. It is important to keep alert to be precise as much as possible deeps on their roles as IPLs, in order to grant robust risk analysis and preserve Process Safety Management in a high level of priority and technical consistency.