Application of artificial neural networks in industrial technology

The application of artificial neural networks to industrial technology is an area that has great potential for exploitation. Many of the applications developed at the University of Tennessee and Oak Ridge National Laboratory are related to equipment and facilities in nuclear power plants. However, the basic principles involved are the same whether the application is to a check valve, rotating machinery, inspection for wear in mechanical system or the evaluation of fatigue life. It makes little difference whether the component or system is part of a nuclear power plants, a manufacturing plant, or a chemical processing plant. Data are used to train a neural network to model the input-output relationships of the systems involved, and patterns associated with specific modes of behavior or characteristics are identified. The results from the artificial neural network must then be put in a form that is useful to the operators and/or engineers concerned with the system. This can involve a simple computer program, an expert system, a fuzzy system, or a human factors related methodology. >