Hydraulic pump fault diagnosis with compressed signals based on stagewise orthogonal matching pursuit

In recent years, many studies have been conducted in hydraulic pump fault diagnosis, which usually ask for sufficient monitoring data. However, in some engineering applications, such as remote monitoring, it is not always guaranteed due to the limitations of transmission bandwidth and computational resources. To solve this problem, this paper proposes, a fault diagnosis method based on compressive sensing theory, which could guarantee the monitoring data can be compressed, and then recovered in remote side. First, original vibration signal of hydraulic pump is used to structure compressed dictionary matrix. Second, Gaussian random matrix is applied to compress the vibration monitoring data of hydraulic pump. Finally, stagewise orthogonal matching pursuit is introduced to reconstruct test data and classify the compressed test data vectors for lessening the diagnosis error rate and reducing the fault diagnosis time. Experimental results demonstrate that the proposed diagnostic can effectively identify and classify typical hydraulic pump faults