The safe operation of a nuclear reactor in a power plant is of utmost importance to the nuclear engineering community and quite vital to creating a positive attitude towards nuclear energy among the rest of the society. This thesis is an attempt to demonstrate how Artificial Neural Networks (ANNs) can increase the operational safety of nuclear reactors by being the basis of a fault diagnostic system in a power plant. It is hoped that neurocomputing, as the science of neural networks is sometimes called, will provide a better approach to recognizing and classifying operational transients at a nuclear power plant.