Weak Fault Diagnosis Method And Application Of Rotating Machinery Based On Time-delay Stochastic Resonance Potential Model

The fault diagnosis of mechanical equipment often needs to extract the sound and vibration signals of mechanical parts,and modern mechanical equipment is mostly operated under the background of strong noise,and these noises will cause interference to the extracted fault signal.Therefore,enhancing the filtering of noise and enhancing the output signal-to-noise ratio of useful signals are of great significance for the extraction of weak fault signals under the background of strong noise.Traditional signal processing methods are mainly based on the principle of eliminating noise and suppressing noise,that is,eliminating or suppressing noise components and retaining useful signal components.However,when the noise and the useful signal overlap,the elimination and suppression of the noise will also filter out the useful signal,causing distortion of the useful signal.In view of the above problems,this paper studies a technique that uses noise to enhance useful signal detection,which is called stochastic resonance signal detection technology,and also studies its application in fault diagnosis.Stochastic resonance uses noise to enhance useful signals through non-linear systems.This filtering mechanism and signal processing method are better than methods based on eliminating and suppressing noise.This article mainly studies the weak fault diagnosis methods of stochastic resonance for rotating machinery parts(bearings,gears,etc.)And carried out simulation and experimental verification.The main research contents include:(1)Research and summary of traditional stochastic resonance theory and basic theoretical research,and derivation of traditional stochastic resonance theory.The noise is classified,and the stochastic resonance of correlated noise is theoretically derived.Take measures and propose new methods to solve the problems of traditional stochastic resonance.(2)Aiming at the single stochastic resonance with a single potential model structure and the problem of output saturation,this paper studies several other potential model structures besides bistable states.The effects of potential parameters on the structure of each potential model are discussed separately.(3)Aiming at the fact that the traditional stochastic resonance does not consider the influence of historical information,this paper proposes a study on the weak fault diagnosis method of delayed stochastic resonance.At the same time,the tristable potential model and the periodic non-sinusoidal potential model are introduced into the delayed stochastic resonance.Time-delay Tristable stochastic resonance and Time-delay periodic non-sinusoidal stochastic resonance are theoretically derived.Signal-to-noise ratio,probability density function,and other indicators are analyzed.Two kinds of time-delay stochastic resonance are applied to weak fault diagnosis of bearings.Through simulation,experimental verification and engineering verification,the superiority of the proposed method is verified.(4)Considering the system inertia based on the time-delay stochastic resonance,a damping factor is introduced into the time-delay stochastic resonance to make the original first-order system into a second-order system.From a signal processing perspective,secondary filtering is performed.The theory of time-delay under-damped stochastic resonance is derived,and the influence of various parameters(time-delay term,damping factor,system parameters)on the signal-to-noise ratio is discussed.In the end,the simulation of the time-delay underdamped stochastic resonance and the fault diagnosis experiment of the bearing outer ring were carried out.Both the simulation results and experimental verification results show that the time-delay underdamped stochastic resonance can better improve the extraction effect of weak fault signals under the background of noise.In summary,this paper studies the weak fault diagnosis method of stochastic resonance in the background of strong noise,and improves on the basis of the traditional stochastic resonance method: A new potential structure,time-delay feedback term,and underdamped term are introduced to form a new stochastic resonance diagnosis method.Compared with the traditional stochastic resonance method,the advantages and practicality of the proposed method have been verified in fault signal processing.