The application of multivariate state estimation failure warning in industrial process

A new condition monitoring methods based on the multivariate state estimation for the industrial process failure prognostic is presented in order to reduce the maintenance costs of industrial equipment. A new and improved matrix in memory is given. It can achieve a good coverage of the normal operation status of the industrial system. When a large difference between the estimated value and the actual measurement value, the existence of the fault is indicated. Sliding window residual statistical methods is used to analyze the residuals. When the residual trend beyond the set threshold value, the system generate the warning. The real-time data of a power plant is used to verify the method. The simulation results show that this method can achieve early failure warning effectively.