Managing the mechanical integrity of pressure equipment using a layers of protection framework and incorporating integrity operating windows

Abstract

A requirement of performance based regulatory regimes for managing process safety is to demonstrate the adequacy of control measures for reducing the risk of a major incident so far as reasonably practical (SFARP). Loss of mechanical integrity allowing loss of primary containment (LOPC) is recognized as a possible initiating step for a major incident. Often LOPC results from the action of damage mechanisms such as corrosion that are able to degrade or destroy the mechanical integrity of the containment envelope of pressure equipment. LOPC, allowing a major incident, generally arises from human error in the form of a lack of recognition of a damage mechanism that should have been identified in a HAZOP or risk assessment and captured in a standard operating practice supported by relevant training. In this paper, a layers of protection framework is proposed for assessing the performance standards and performance indicators required of control measures to allow demonstration that likelihood of LOPC has been reduced SFARP, where LOPC has resulted from a damage mechanism such as corrosion. The context is the requirement to demonstrate adequacy of the practices used to assure the integrity of the containment envelope of pressure equipment required to prevent a major incident. The assessment includes examining the adequacy of inspection as a control measure including the RBI practice of using industry failure frequency databases to underpin the justification of an extension of inspection intervals. An improvement is proposed that recognizes the role of integrity operating windows (following API RP 584) for managing the process environment including providing evidence of control of the service conditions and recording excursions from nominal service conditions. Underpinning the proposed improvement is consideration of the contribution of human error to LOPC events and the need to recognize performance shaping factors that minimize error.