A Multi-Model Diagnosis Method for Slowly Varying Faults of Plunger Pump

Abstract

As the energy supply component of hydraulic transmission systems, the plunger pump is widely used in the field of ship and ocean engineering. Thus, its fault diagnosis is of great importance. The multi-model fault diagnosis method based on the Kalman filter is slow in detection and isolation in the process of slowly varying fault diagnosis, and it may be diagnosed as a false failure. In this article, to improve the performance of the multi-model fault diagnosis method, we combine the method and support vector machine and propose a new method by fusing the conditional probability of the multi-model with the posterior probability of the support vector machine. The experimental results on a marine plunger pump illustrate the effectiveness of the proposed method. With the appropriate weight coefficient, the detection speed and isolation speed of the joint multi-model method are improved after the combination of the support vector machine, and the new method has better robustness.