Fault diagnosis of planetary gear drive system based on sparse decomposition optimised by generalised matching pursuit algorithm

摘要

Due to its complex structure and poor working environment, the vibration signal of the planetary gear transmission system experiences high levels of noise. In mechanical fault diagnosis, the feature extraction of weak signals with high levels of noise is always a difficult problem.The sparse decomposition algorithm proposed in recent years can solve this problem to some extent. However, in the process of sparse decomposition, a large number of inner product calculations are involved, which has the disadvantages of large computational demand and low efficiency. In thispaper, the generalised orthogonal matching pursuit (gOMP) algorithm is applied to optimise the computation process of the matching pursuit algorithm in order to reduce the amount of computation required in the process of the operation and improve the real-time performance in the process offault diagnosis. It is proven that the generalised orthogonal matching tracking method can solve the shortcomings of the large computational demand to a certain extent and will not affect the precision of the fault signal reconstruction.