Risk Assessment: Theory, Methods, and Applications

CHAPTER 15

COMMON-CAUSE FAILURES

CCF analysis is an integral part of any PSA, from the unavailability of systems to the dominant accident scenarios.

—Apostolakis and Moieni (1987)

15.1 INTRODUCTION

Common-cause failures (CCFs) have been considered in probabilistic risk analyses of nuclear power plants since the early 1970s (e.g., see NUREG-75/014, 1975). This industry has since then had a continuous focus on CCFs, and has been a driving force for the development of CCF models and with respect to collection and analysis of data related to CCFs. The aviation industry has also given these failures close attention, and the Norwegian offshore oil and gas industry has since the mid-1980s focused on CCFs related to reliability assessment of safety instrumented systems (e.g., see Hauge et al., 2010).

More recently, the IEC 61508 (2010) standard focuses on the need to control CCFs in order to maintain the safety integrity level (SIL) of safety instrumented functions (see Section 12.4). The standard suggests a method of calculating the probability of failure on demand where the contribution of CCF is modeled by the well-known beta-factor model.

15.2 BASIC CONCEPTS

15.2.1 Dependent Failures

When components of a system fail, the failures cannot always be considered as independent events. Two main types of dependency can be distinguished: positive and negative. If the failure of one component leads to an increased tendency for another component to fail, the dependency is said to be positive. ...