Using systems theory for additional risk detection in boiler explosions in Brazil

Abstract

Recent technological advances have allowed some human activities, including those related to safety, to be automated. However, these advances increased the complexity of sociotechnical systems, representing exponential growth of interactions between humans, computers, machines, and the environment. Moreover, growing productive and economic pressures demand safer and more reliable products and projects, at lower costs and time than those practised by competitors. To cope with this complexity and set of conflicting objectives, the STAMP (Systems-Theoretic Accident Model and Processes) emerged as a novel approach to analyse processes and accidents. In this study, an overview of boiler accidents in Brazil is presented and a causal analysis based on STAMP (CAST) is conducted to revisit one of the worst boiler accidents in the Brazilian scenario. Even without direct participation in the investigations, additional and more relevant risk factors are evidenced. Furthermore, it was found that government agencies generally refrain from reviewing their own control actions contributing to the hazard, limiting their potential improvements. This suggests a need for companies and government agencies to adopt new paradigms of risk and accident analysis and to work together for a systemic safety improvement approach.