Application of compressed sensing in remote fault diagnosis of hydraulic pump

摘要

To avoid the problem of low transmission efficiency brought about by the large redundancy of the monitoring signal in hydraulic pump when remote fault diagnosis, the compression transmission of fault monitoring signal based on compressed sensing was proposed. According to compressed sensing, the original signal was sparsely represented by orthogonal basis. Then, the Gaussian matrix was used as the measurement matrix to compress the original signal, the remote transmission of the compressed data was realized by means of the Socket communication technology and TCP/IP network protocol. Moreover, the orthogonal matching pursuit algorithm was used as reconstruction algorithm to reconstruct the compression signal, the reconstruction value of original signal was obtained by the inverse sparse transformation. Finally, on the hydraulic pump fault simulation test bench, the remote fault diagnosis experiment of single slipper wear and single loose slipper are experimented. The original signal and the reconstructed signal were analyzed by Hilbert envelope demodulation. Based on power spectrum analysis, it can be found that they have the same ability for fault diagnosis. Results indicate the availability of the technology in remote fault diagnosis of hydraulic system fault.