Inspecting Nuclear Pressure Vessels: The Conundrum of Minimizing Risk

Abstract

The probability of a sudden, massive release of radioactivity from a light-water nuclear reactor through a breach of the containment is assessed on the basis of statistical data which partly consist of subjective estimates. This breach refers to the existence of crack-like defects remaining after a non-destructive examination of the main pressure vessel surrounding the reactor core. Two studies have recently been made of such sources of information about the effectiveness of non-destructive examination of pressure vessels with respect to defects. The results of these studies indicate that the data used as input in the probabilistic calculations do not possess the reliability that might be assumed from the assessments. This type of failure should therefore no longer be considered a de minimis case. In the present review the overconfidence in the efficiency of non-destructive examination is discussed from psychological, sociological and political science points of view. It is concluded that ingrained professional assumptions and values seem to be the main reason for the trust in the technology of inspection. However, there are also psychological constraints that can be understood only in their social and political contexts.