Human reliability analysis applied to rig operations using a cognitive framework and defined performance influencing factors

Abstract

This work presents a qualitative and quantitative approach to introduce human factors aspects in the reliability and risk assessment of offshore rig operations. The qualitative part covers the task analysis and searching for potential causes of failures based on a cognitive framework. In the quantitative stage, the combined effect of factors on Human Error Probability is computed from normalized relative weights and evaluation of the called degree of compliance. The model is initially drawn by means of binary trees and subsequently converted into a Bayesian network in order to express conditional relationships between human factors and cognitive failures. The approach was applied to the scenario of rig emergency disconnection and the degree of compliance scored considering a literature review on the operational context of the Macondo well blowout. The main contributors identified for this accident were associated to organizational aspects, risk communication, attention and team collaboration that mostly affected situation awareness of the crew. To verify the model, sensitivity analyzes were performed through variations in the human factors under two extreme operating conditions and in the probability of the cognitive function failures. The results obtained allowed to validate the developed proposal.