Centrifuge rotor fault diagnosis method based on parameter self-adaptive stochastic resonance

Abstract

The invention relates to the centrifuge rotor fault diagnosis field, and especially to a centrifuge rotor fault diagnosis method based on parameter self-adaptive stochastic resonance, which includes the following steps of: using a particle swarm optimization to realize self-adaptive value optimization of bistable system model parameters a and b; sampling multiple groups of fault signals of a centrifuge rotor with faults and performing stochastic resonance treatment and fast Fourier transform in the bistable system; carrying out normalization processing on each group of fault character frequency data; intersecting and combining five Support Vector Machine (SVM) types and four kernel functions to set up 20 types of different SVMs, performing parameter optimization on the values of penalty factors c and g by using the particle swarm optimization, and finding out an optimal SVM module and the optimal values of the penalty factors c and g; and carrying out fault diagnosis on test set samples. The centrifuge rotor fault diagnosis method based on the parameter self-adaptive stochastic resonance can recognize a characteristic signal value in noise-containing signals in a low error rate, and improves the accuracy rate of the fault diagnosis.

Description

technical field [0001] The invention relates to the field of centrifuge rotor fault diagnosis, in particular to a centrifuge rotor fault diagnosis method based on parameter adaptive stochastic resonance. Background technique [0002] A centrifuge is a high-speed rotating machine with a complex structure. It is widely used in chemical industry, petroleum, food, pharmaceuticals, mineral processing, coal, water treatment and shipping. Cause parts aging, wear and other failures. The main fault types of the rotor include rotor unbalance, rotor misalignment, rotor friction, oil film oscillation, and shaft cracks. [0003] Since the centrifuge rotor vibration signal is a weak signal with low signal-to-noise ratio and heavy pollution, it is difficult to obtain obvious characteristic frequency and associated frequency value if the vibration signal directly obtained from the sensor is directly subjected to fast Fourier transform (FFT). LIQiang, WANGTaiyong, LENGYonggang, etal. In th...

AI Technical Summary

Problems solved by technology

[0008] The present invention provides a centrifuge rotor fault diagnosis method based on parameter adaptive stochastic resonance, which solves the problem of inaccurate extraction of characteristic frequency caused by stochastic resonance system with fixed parameter a=b=1 in the prior art. The specific technical scheme is as follows :

Images