Coal mill fault diagnosis based on Gaussian process regression

Abstract

A typical operating set of equipment can be obtained through cluster analysis of historical data. Two state monitoring models for HP medium speed coal mill are established based on Gaussian process regression and the similarity index calculated by this model can be used for measuring the operating status of HP mills. Finally a method for fault diagnosis of HP mill based on Gaussian regression modelling is proposed combined with fault diagnosis knowledge base of this HP mill. Taking the HP medium speed mill of a 660MW thermal power unit as an example, the real operating data is collected and used for modelling and analysis. Results shows that the equipment parameter estimation calculated by Gaussian process regression is accurate. It can be used for early-warning and diagnosed of equipment fault and also for practical engineering application.