Identifying safety indicators for safety performance measurement using a system engineering approach

Abstract

The paper presents a method for the development of safety indicators for a process industry application based on a system engineering perspective. Traditional approaches use probabilistic risk assessment or linear accident models which assume that accidents are linear chains of events and do not consider complex systemic factors and interactions. After BP's Texas City refinery incident, the investigation committee reported that BP had a false sense of safety performance due to providing more focus on managing personal safety rather than process safety. System engineering concepts may help the process industry to operate their activities without any severe accidents by establishing a better safety management system. This paper adopts the STAMP (System-Theoretic Accident Model and Processes) accident causation model to identify system specific indicators and also describes the proposed method with the help of a simple process industry application which is an LNG ship to ship transfer process. It compares the developed method with other methods for practical case applications. The first step of the present method is to establish the safety control structure, then safety performance indicators are identified. Further work is necessary to investigate to what degree these STAMP based indicators are complementary to indicators developed by other methods.