The Potential for Expert System Support in Solving the Pressurized Water Reactor Fuel Shuffling Problem

The fuel shuffling problem is posed by the need to reposition partially burned assemblies to achieve minimum X-Y pin power peaks in reload cycles of pressurized water reactors. This problem is a classic artificial intelligence (AI) problem and is highly suitable for AI expert system solution assistance, in contrast to the conventional solution, which ultimately depends solely on trial and error. Such a fuel shuffling assistant would significantly reduce engineering and computer execution time for conventional loading patterns and, much more importantly, even more significantly for lowleakage loading patterns. A successful hardware /software demonstrator has been introduced, paving the way for development of a broadly applicable expert system program. Such a program, upon incorporating the recently developed technique of reverse depletion, would provide a directed path for solving the low-leakage problem.