Comparing identification of human failure events with two HRA methods in an excontrol room action

We compare two human reliability assessment methods, THERP and Phoenix, in human failure event (HFE) identification in an ex-control room setting. First, we consider briefly what HFE identification should encompass. Then we introduce the example scenario used, a part of a post-earthquake scenario where two operators attempt to even out water levels in steam generators utilizing an emergency water source. We describe briefly how HFE identification is conducted in the two methods, and conduct HFE identification in the scenario with each method. We compare the methods by conducting HFE identification with them on the example scenario. A major difference between the two is that THERP is procedure-based whereas Phoenix is model-based. Another difference is that THERP does not guide HFE identification much, which gives analysts freedom to apply their judgment, but may lead to considerable differences in analysis results by different analysts. In contrast, the HFE identification process of Phoenix is structured, model-based and with considerable guidance, and therefore likely promotes reduction of differences in analysis results between different analysts.

Keywords: human reliability analysis, human failure event, ex-control room action, THERP, Phoenix, seismic.