WASH-1400: A Comparison of Experience and Prediction

In an Electric Power Research Institute report, it was concluded that the error bounds in WASH-1400, the Reactor Safety Study, although perhaps understated, are not necessarily “greatly understated,” as claimed by the Lewis Committee. The relationship between an experimental data base for operating light water reactors (LWRs) and the predictions of WASH-1400 have been examined. The data base consists of

1. U.S. commercial LWR reactors
2. world LWR reactors
3. world LWR plus U.S. Navy reactors.

Estimations of an expected increase in experience indicate that the WASH-1400 core-melt probability cannot be an underestimation. The maximum degree of unquantified probability in the WASH-1400 calculation compared to experience is that of the margin. The increase in WASH-1400 uncertainty cannot be greater than a factor of 4, and the WASH-1400 median cannot be low by more than a factor of 4.