Research on Fault Diagnosis and Prediction of Power Plant Fans

Abstract

With the rapid development of artificial intelligence technology, intelligent recognition, diagnostic technology and trend prediction research on power production equipment failure are gradually being carried out. This paper does research into the induced draft fan based on the vibration signal data of the fan. It uses K-Means clustering and least squares support vector machine (LSSVM) to diagnose and trend the collected faults. Next, the trend prediction method for cracking failure of induced draft fan is also studied. Aiming at the large residual error caused by LSSVM regression prediction and actual value, a parameter optimization scheme based on PSO-LSSVM is proposed to improve the prediction accuracy.