Identifying macrocognitive function failures from accident reports: A case study

Abstract

Reliable macrocognitive functions are important for maintaining system safety. Few studies were conducted to investigate macrocognitive function failures in a complex system. NUREG-2114 proposes a cognitive framework connecting macrocognitive function failures, proximate causes, failure mechanisms, and performance influencing factor (PIFs). This model can serve as a model for analyzing human failure events in human reliability analysis (HRA). This study investigated macrocognitive function failures in a complex environment and also examined the usability of the cognitive framework in the HRA qualitative analysis. A total of 103 investigation reports of incidents and accidents from a petrochemical plant in China were involved. It was found that 35 % of the incidents and accidents could be attributed to human errors. Failures of action implementation and team coordination were the dominant failures. This study also gave the information of proximate causes, failure mechanisms, and PIFs for each macrocognitive function failure. The usability issue of the cognitive framework in NUREG-2114 was discussed. It seems that the current cognitive framework needs to be improved to inform HRA.