Abstract:Multi-scale permutation entropy(MPE) is a nonlinear dynamic method,which has been widely used in fault diagnosis of rotating machinery.However,in permutation entropy,it does not considered that time series with the same permutation pattern may have different amplitudes,and the coarse-grained method has defects.In order to solve the above problems,a time-shift multi-scale amplitude aware permutation entropy(TSMAAPE) was proposed.Time-shift time series was used to improve the shortcomings of coarse-grained time series in MPE and amplitude-aware permutation entropy was introduced.The robustness of TSMAAPE was verified by comparing with time-shift multi-scale permutation entropy and multi-scale amplitude aware permutation entropy.Taking into account the advantages of TSMAAPE in feature extraction,combined with the kernel extreme learning machine optimized by the whale optimization algorithm,an intelligent fault diagnosis method for hydraulic pumps was proposed.The results show that this method has good classification accuracy for different faults of hydraulic pumps,and has broad application prospects in the field of fault diagnosis.