Resonance-enhanced mechanical transient feature extraction method induced by difference in transition rate of left and right potential well particles

Abstract

The invention discloses a resonance-enhanced mechanical transient feature extraction method induced by the difference in the transition rate of particles in the left and right potential wells, which overcomes the shortcomings of the existing stochastic resonance-based fault detection method that only enhances the frequency amplitude of a single fault feature and cannot extract repeated transient features. , which solves the problem that the asymmetry of the left and right potential wells drives the particle transition rate difference to induce the weak characteristic enhancement mechanism of the resonance. The present invention first establishes a single dynamic adjustable depth or width of the left potential well and an asymmetric double well potential with the right potential well, and constructs an asymmetrical double well potential of the left and right potential wells to induce an over-damped resonance system to generate unequal or asynchronous left and right potential wells Particle transition rate; then, build a left and right potential well asymmetric double-well potential-induced overdamped resonant array, input multiple sub-signals truncated by the rectangular sliding window of the mechanical original signal into the resonant array, and use the re-corrected kurtosis index as the genetic algorithm objective function to optimize the array parameter, which realizes the enhancement and extraction of weak and repetitive transient features of early mechanical faults, and improves the kurtosis index of the original signal by more than 80 times.

Description

technical field [0001] The invention belongs to the field of weak fault diagnosis of high-end equipment, and relates to a method for extracting resonance-enhanced mechanical transient features induced by differences in the transition rates of particles in left and right potential wells. Background technique [0002] High-end equipment has been exposed to extreme working conditions such as strong corrosion, strong radiation, high and low temperature, and strong impact for a long time, which can easily lead to fatigue failure of key components, resulting in huge economic losses and major catastrophic accidents, such as ships, aircraft, rockets, satellites, nuclear power plants, high-speed rails , wind power, etc. Facing the country's major strategic needs, the research on the detection of small damage on the surface of high-end equipment under extreme working conditions can not only guide the optimal design and manufacture of parts, but also guide the self-healing of minor dam...

AI Technical Summary

Problems solved by technology

[0005] The present invention aims to propose a resonance-enhanced mechanical transient feature extraction method induced by the difference in the transition rate of left and right potential well particles, aiming at the existing stochastic resonance fault detection technology which only enhances the single fault characteristic frequency amplitude and ignores the fault characteristic frequency multiplier or sideband Information, prone to misdiagnosis of faults, unable to enhance the transient characteristics of mechanical repetition, and does not consider the problem of the unequal transition rate of the left and right potential wells of Brownian particles, will establish a single dynamic adjustable depth or width of the left potential well and an asymmetric double well potential of the left and right potential wells , by adaptively adjusting the depth or width of the left potential well, the transition rate of Brownian particles that is not synchronized or equal to that of the right potential well is generated, so that the difference in the transition rate of Brownian particles in the left and right potential wells is used to induce resonance and capture noise energy to enhance mechanical repetitive transient feature extraction

Images