Neural Networks for Monitoring Mechanical Defects of Rotating Machines

The quality of maintenance is an important parameter in the performance of industrial production plants. Good monitoring of the deterioration in rotating machinery can result in reduced maintenance costs by minimizing the number of bad judgments and decreasing in the number of spare parts. An effective system of maintenance should be able to monitor parameters such as vibration, temperature, oil quality. In addition it can deduce the current operating state of a rotating machine, estimate the future state of a machine and prevent fatal breakdowns well in advance. Therefore, vibration measurements are regularly used as major indicators on the health of rotating machineries and on whether or not maintenance is to be proceeded. The information contained in these measurements is processed using new signal processing algorithms. In this paper, forecasting methods are developed to prevent deterioration of a rotating machine using autoregressive models and artificial neural networks. Index Terms – Maintenance – prediction – vibration – artificial neurons networks.