ANALYSES OF POSSIBLE FAILURE MECHANISMS AND ROOT FAILURE CAUSES IN POWER-PLANT COMPONENTS USING NEURAL NETWORKS AND STRUCTURAL FAILURE DATABASE

This paper describes analyses of case studies on failure of structural components in power plants using hierarchical (multilayer) neural net works. Using selected test data about case studies stored in the struc tural failure database of a knowledge-based system, the network is tra ined: either to predict possible failure mechanisms like creep, overhe ating (OH), or overstressing (OS)-induced failure (network of Type A), or to classify a root failure cause of each case study into either a primary or secondary cause (network of Type B). In the present study, the primary root cause is defined as ''manufacturing material or desig n-induced causes,'' while the secondary one as ''not manufacturing mat erial or design-induced causes, e.g., failures due to operation or mal -operation.'' An ordinary three-layer neural network employing the bac k propagation algorithm with the momentum method is utilized in this s tudy. The results clearly show that the neural network is a powerful t ool for analyzing case studies of failure in structural components. Fo r example, the trained network of Type A predicts creep-induced failur e in unknown case studies with art accuracy of 86 percent, while the n etwork of Type B classifies root failure causes of unknown case studie s with an accuracy of 88 percent. It should be noted that, due to a sh ot-cage of available case studies, an appropriate selection of case st udies and input parameters to be used for network training was necessa ry in order to attain high accuracy. A collection of more case studies should, however, resolve this problem, and improve the accuracy of th e analyses. An analysis module for case studies using the neural netwo rk has also been developed and successfully implemented in a knowledge -based system.