Efficient Compliance with Prescribed Bounds on Operations Parameters by Means of Hypothesis Testing Using Reactor Data

Abstract

A common problem in any reactor operations is to comply with a requirement that certain operational parameters are constrained to lie within some prescribed bounds. The fundamental issue which is to be addressed in any compliance description can be stated as follows: The compliance definition, compliance procedures and allowances for uncertainties in data and accompanying methodologies, should be well defined and justifiable. To this end, a mathematical framework for compliance, in which the computed or measured estimates of process parameters are considered random variables, is described in this paper. This allows a statistical formulation of the definition of compliance with licence or otherwise imposed limits. An important aspect of the proposed methodology is that the derived statistical tests are obtained by a Monte Carlo procedure using actual reactor operational data. The implementation of the methodology requires a routine surveillance of the reactor core in order to perform the underlying statistical tests. The additional work required for surveillance is balanced by the fact that the resulting actions on the reactor operations, implemented in station procedures, make the reactor "safer" by increasing the operating margins. Furthermore, increased margins are also achieved by efficient solution techniques which may allow an increase in reactor power. A rigorous analysis of a compliance problem using statistical hypothesis testing based on extreme value probability distributions and actual reactor operational data leads to effective solutions in the areas of licensing, nuclear safety, reliability and competitiveness of operating nuclear reactors.