Wavelet modified Bayesian convolution energy-based underwater vehicle propeller fault diagnosis method

Abstract

The invention discloses a wavelet modified Bayesian convolution energy-based underwater vehicle propeller fault diagnosis method. The influences of random noise interference in signals are reduced based on wavelet decomposition, the amplitude of singular behaviors of dynamic signals is enhanced based on a modified Bayesian algorithm, the fluctuation of singular behaviors of the dynamic signals isreduced based on convolution calculation, the position of the minimum value in the dynamic signals is used as an energy region boundary, energy fault features are extracted, the fault features and a fault class are coordinated based on homomorphic transformation, and propeller fault degree classification is realized finally based on a support vector domain description algorithm. The fault degree classification precision in the method is little influenced by a modified Bayesian step size, and the classification accuracy is 100%.

Description

technical field [0001] The invention belongs to underwater robot fault diagnosis technology, in particular to an underwater robot propeller fault diagnosis method based on wavelet modified Bayesian convolution energy. Background technique [0002] The underwater robot moves under the thrust of the propeller. When the propeller fails, dynamic signals such as the speed signal of the underwater robot and the rate of change signal of the propeller control voltage usually produce singular behaviors, and the greater the degree of propeller failure, the greater the energy contained in the singular behavior. Based on this, energy fault features can be extracted for thruster fault diagnosis. [0003] The known signal amplitude square sum method is a commonly used energy feature extraction method. In this known method, the signal amplitudes in the singular behavior area are squared and then summed. In the process of extracting energy fault features using known methods, the following...

AI Technical Summary

Problems solved by technology

In the process of extracting energy fault features using known methods, the following problems are encountered: the amplitude of the singular behavior of the dynamic signal of the underwater robot is relatively weak and needs to be enhanced; the random noise interference in the signal affects the size of the fault energy feature; the dynamic signal singular behavior fluctuates strong, affecting the boundary identification of the fault energy area

Images