Research On Weak Signal Detection Method Based On Stochastic Resonance

There are two definitions of weak signals,one is that the absolute value of useful signal amplitude is very small,the other is that the amplitude of useful signal is very weak compared with noise.This paper mainly discusses the weak signal in the second case,that is,the periodic signal submerged in the background of Gaussian white noise.Traditional weak signal detection methods improve the signal-to-noise ratio of the system by suppressing or filtering noise.But when the signal and noise are in the same frequency band,the harmful noise is filtered,while some useful signals will be destroyed,which makes the accuracy of weak signal detection greatly reduced,resulting in the poor results.It is these problems that the nonlinear weak signal detection method is born in compliance with the problem.Random resonance is one of them.Stochastic resonance can make use of noise to be turned into treasure,that is,under the action of nonlinear system,noise energy can be transferred to signal energy,and the signal to be tested can be detected.Because there are many systems that produce stochastic resonance,Duffing oscillator with rich nonlinear dynamic characteristics is chosen as the research object to study its application in weak signal detection.Firstly,the traditional Duffing oscillator stochastic resonance under the excitation of Gaussian white noise is studied.The output signal-to-noise ratio is used as the measurement index of the effect of the stochastic resonance.The influence of system parameters,noise intensity and damping ratio on the output signal-to-noise ratio is studied.The factors affecting the performance of traditional Duffing vibrator are summarized.The effectiveness of traditional Duffing oscillator stochastic resonance system to detect weak signal is verified by the analysis of the fault diagnosis example of bearing outer ring.Secondly,a new coupled Duffing oscillator stochastic resonance model is constructed.Two traditional Duffing oscillators are nonlinear coupled into a multi-state Duffing oscillator stochastic resonance system by a new coupled form.It is proved that the model also has the characteristics of random resonance and analyzes the coupled parameters,system parameters,periodic signal amplitude and noise intensity to the coupled Duffing oscillator stochastic resonance system The influence of the output signal-to-noise ratio is summarized,and the parameters of the maximum output SNR of the coupled Duffing oscillator are obtained.At the same time,the fault signal of bearing outer ring is input into the newly constructed coupled Duffing oscillator stochastic resonance system,which verifies the effectiveness of the new coupled Duffing oscillator stochastic resonance system to detect weak signal.Finally,the coupled Duffing oscillator stochastic resonance system is connected in a cascade way,and a new cascade coupled Duffing oscillator stochastic resonance model is constructed.The model is applied to weak periodic signal detection,and the detection effect is analyzed and the parameters of the model have better output response effect than the coupled Duffing system,The influence of cascade number on the signal-to-noise ratio of the system output is analyzed.