Research On Feature Extraction Method Of Weak Signal Based On Cascade Second-order Tristable Stochastic Resonance Model

This paper mainly studies the weak signal feature extraction based on stochastic resonance in the background of strong noise.Although many scholars have also proposed methods to extract weak signal features using stochastic resonance,these methods have shortcomings and can be improved.Aiming at the shortcomings of traditional stochastic resonance methods,this paper proposes corresponding solutions.Finally,the proposed method is applied to the extraction of weak signal features of mechanical equipment.The research content of this paper mainly includes the following aspects:(1)Aiming at stochastic resonance using most intelligent algorithms for parameter optimization,it is easy to fall into the local optimization problem.This paper proposes an Adaptive Second-order Tristable Stochastic Resonance(ASTSR)method.This method solves the limitations of the stochastic resonance theory through the ordinary variable scale method,transforms the large frequency signal into a small frequency signal far less than 1Hz;uses the ASTSR output SNR as the objective function,and relies on the chaotic ant colony algorithm to calculate the stochastic resonance Output the best parameter combination when the SNR is maximum,and input the stochastic resonance system to get the best result,and finally realize the feature extraction of weak signals.Through simulation and bearing fault data analysis,the effectiveness of the weak signal feature extraction of the ASTSR method is verified.(2)Aiming at the general effect of single-stage stochastic resonance,there may be problems with interference frequencies and poor extraction of eigenfrequency,a Cascaded Adaptive Second-order Tristable Stochastic Resonance(CASTSR)method is proposed,that is,multiple ASTSR systems are connected in series.Using the principle of stochastic resonance,the noise energy in the signal under test is continuously transferred from the high frequency area to the low frequency area,thereby enhancing the weak signal characteristics of the low frequency.Simulation and bearing failure data analysis verify the effectiveness of CASTSR method for further extraction of weak signal features.(3)Aiming at the problem that the empirical mode decomposition(EMD)method has poor decomposition quality under strong noise conditions and poor extraction of weak signal features,a Cascaded Adaptive Second-order Tristable Stochastic Resonance(CASTSR)reduction is proposed.Noisy EMD method.This method first uses the CASTSR method to preprocess the signal to be denoised,then uses EMD to decompose,and finally extracts the characteristic frequency from IMF1.Simulation and bearing fault data analysis results show that this method can improve the quality of EMD decomposition and realize weak signal feature extraction.