To main content

New method for updating failure rates and proof test intervals of equipment groups within safety instrumented systems

Abstract

Safety instrumented systems (SIS) are to be followed up during operation at a facility. An important part of SIS follow-up is to ensure that all relevant SIS requirements from the design phase are fulfilled during the entire lifetime of the facility. The safety integrity level (SIL) and the corresponding probability of failure on demand (PFD) requirements should be verified regularly. The PFD of a component is a function of the dangerous undetected failure rate and the proof test interval of the component. For follow-up purposes it is relevant to group comparable components within the same equipment group, such as shutdown valves or level transmitters. For analysis purpose, the method focuses on equipment groups where all components within the same group are assumed to have a comparable (common) failure rate and PFD budget. This paper presents an approach for estimating updated failure rates and assessing the proof test intervals based on operational experience. The method for updating failure rates is based on a Bayesian failure rate estimation with additional periodisation. For proof test interval assessment two methods are given: One method based on comparison of the “experienced” failure rate and the failure rate assumed in design, and one method based on comparison of the “experienced” PFD and the PFD budget requirement.

Category

Academic chapter/article/Conference paper

Client

  • Research Council of Norway (RCN) / 295902

Language

English

Author(s)

Affiliation

  • SINTEF Digital / Software Engineering, Safety and Security
  • Norwegian University of Science and Technology

Year

2020

Publisher

Research Publishing Services

Book

e-proceedings of the 30th European Safety and Reliability Conference and 15th Probabilistic Safety Assessment and Management Conference (ESREL2020 PSAM15)

ISBN

9789811485930

Page(s)

797 - 804

View this publication at Cristin