A Fault Detection Method for a Practical Electro-Hydraulic Variable-Displacement Pump With Unknown Swashplate Moment

Abstract

Electro-hydraulic controlled variable-displacement pump (EHVDP) is a power source of the hydraulic system, which is extensively used in the electro-hydraulic system due to the high efficiency and controllability. Thus, for a good operation of the hydraulic system, effective fault diagnosis techniques are crucial to avoid unexpected breakdown and failure. However, the data-driven fault diagnosis methods, which are commonly adopted in axial piston pumps, suffer from a lack of considerably effective data. Modelling uncertainty due to the swashplate moment has not been adequately considered, though some attempts have been made for the model-based methods. In this paper, the pressure transition is assumed to simplify the swashplate moment. A Kalman filter with unknown input for the EHVDP is used to estimate the additional uncertainty due to the simplification of the swashplate moment. A cumulative sum (CUSUM) based residual evaluation is utilized to detect the fault of the EHVDP. The test bench has been established and three typical failure modes of the EHVDP, including the control valve, displacement regulation mechanism, and rotating group are investigated by the experiments. Experimental results under different working conditions validate the effectiveness of the proposed fault detection method. Comparative results show the superiority of the proposed method in sensitivity and rapidity over the general method not considering unknown swashplate moment.