Dependence assessment of human operators in advanced main control rooms considering uncertainty and dynamic situations

Abstract

In human reliability analysis (HRA), dependence assessment is of particular importance, especially in nuclear engineering. For operators in advanced main control rooms (MCRs), many dependence assessment methods are derived from the Technique for Human Error Rate Prediction (THERP), which may lack traceability and repeatability. Moreover, existing studies tend to overlook the dynamic aspects of operator dependence. We propose a combination of decision trees (DTs) and Dempster–Shafer evidence theory (DSET) for assessing operator dependence in advanced MCRs under uncertain, dynamic conditions. Firstly, experts identify influence factors, which analysts then assess against specific conditions. Secondly, analysts model these judgments with DSET and apply Dempster’s rule to fuse evidence and reduce uncertainty. Thirdly, a DT model classifies the operator dependence level, and a modified THERP method calculates the conditional human error probability (CHEP). Finally, a shift-change scenario illustrates dynamic dependence assessment. Case study on a residual-removal system breach demonstrates its practicality and flexibility.