Core Damage Severity Evaluation for Pressurized Water Reactors by Use of Fuzzy Logic Algorithms

Abstract

The problem of core damage severity evaluation during a core-threatening accident of a pressurized water reactor is addressed. An expert system, Core Damage Evaluator (CoDE), is developed that makes an adept utilization of the inferring capabilities of fuzzy logic to classify the core in the damage severity category: “intact,” “clad failure,” or “core melt” or a combination of the last two. If it is determined that some form of core damage exists, the logic model enters a quantification stage to provide a numerical assessment of the damage.The model is provided with two row vector inputs at a rate of 100 to 150 vector pairs per minute. The qualitative vector consists of 69 elements, whereas the quantitative one contains 83. These elements constitute instantaneous physical parameter values provided by the plant instrumentation. The inferencing procedure employed in this problem is the generalized modus ponens (GMP), which has its origin in the field of approximate reasoning.