Combining conventional and innovative risk modelling techniques to identify safety culture issues during the Three Mile Island nuclear accident

The partial nuclear core meltdown at the Three Mile Island (TMI)-2 reactor in 1979 was the first big nuclear accident of significant public interest. A combination of mechanical component failures, a lack of operator understanding of the developing accident progression, and major operator errors led to the first core meltdown in the history of commercial nuclear reactor operations. Many studies have analyzed severe accident progression and physical barrier failures, using conventional risk assessment techniques such as ETA and FTA, which are typically applied in risk modelling prior to the operation of, e.g., a nuclear power plant. To address human and organizational factors, a wide range of methods exists, and these methods are also applied prior to plant operation. In this paper, we adopt an alternative approach. First, we use the STEP method to analyze the sequence of events in the TMI-2 accident, based on existing investigation reports. Then we investigate how to apply two different methods for human and organizational factor assessment, namely the SPAR-H and Risk-OMT approach, to address more explicitly those human and organizational factor challenges pointed out in the TMI accident literature. The objective of this investigation is to determine the extent to which analytical techniques can identify key risk factors that can be mitigated prior to plant operation. For example, is it possible to address challenges related to, for example, safety culture, complexity, and so-called tight couplings in a more analytical manner.

Keywords: Nuclear accidents, Nuclear safety, Accident investigation, Human and organizational factors, Risk modeling, Human reliability analysis.