Research On Weak Signal Detection And Application Based On Composite Multi-stable Stochastic Resonance

Weak signal detection in strong noise background is a common problem in many engineering fields.As a nonlinear weak signal detection method using noise,stochastic resonance(SR)has been widely used in many fields in recent years,especially in the field of fault diagnosis.The research on the detection theory and method of SR has important theoretical significance and application value for the diagnosis of early weak faults.At present,the potential function models of SR mostly adopt bistable,monostable and tristable models.In order to further improve the weak signal processing ability of SR,this dissertation proposes an impro'ved novel composite multi-stable potential model,which is constructed by the joint of the tristable model and Gaussian potential(GP)model.The switching mechanism of the composite multi-stable model between different steady states is studied.And the output signal-to-noise ratio(SNR)is used as the index to study the effects of the system parameters of the composite multi-stable model and ? stable noise distribution parameters on the output of SR system under Gaussian noise and ? stable noise environments.According to the influence of the noise intensity D on the SR effect,the performance of the composite multi-stable SR system is analyzed.For different levels of weak signals,the output performances of SR systems based on composite multi-stable model,traditional tristable model,composite tristable model constructed by the bistable and GP models are compared and analyzed.The results prove that the proposed model has better performance.In addition,the cascaded system consisting of composite multi-stable SR is studied.The experimental results show that the cascaded SR system can obtain better detection effect,but the cascaded series should be selected according to the actual detection results.And the cascaded SR system has shaping and filtering effect on square wave signals.At the same time,the adaptive cooperative optimization method for multiple system parameters in parameter-induced composite multi-stable SR is studied.The differential brain storm optimization(DBSO)algorithm is used to realize the adaptive selection of the parameters in the composite multi-stable SR system.The multiple high-frequency periodic signals are detected effectively under Gaussian noise and a stable noise environments.And the results are compared with those of particle swarm optimization(PSO)algorithm.In addition,the non-periodic impact signal and UWB-IR signal are detected under Gaussian noise environment,and the optimal shape of potential well presented by the model and the traj ectories of particles in the potential well are analyzed.Finally,the weak signal detection theory and method of SR are applied to machinery fault detection.The detection of simulated fault signal,bearing inner and outer ring faults is realized,and compared with spectral kurtosis and empirical mode decomposition(EMD)methods.The research results in this dissertation provide a novel adaptive SR method,which enriches the research results of SR theory and broadens the application range of SR in practical engineering.