Research On Weak Signal Detection Based On Coupled,Unsaturated And Asymmetric Stochastic Resonance Systems

Stochastic resonance(SR),an important scientific phenomenon,is widely used in the weak signal detection(WSD).It is involved in mechanical fault diagnosis,biomedicine,image processing and other fields.Unlike the traditional detection methods which are designed to attenuate or eliminate noise,SR takes advantage of the noise to form a cooperative relationship with the nonlinear system and the signal and then part of the noise energy could be transferred into weak signals.The weak signal could be enhanced,what the system output signal-to-noise ratio(SNR)is improved.Because of its the unique characteristic that utilizing noise,SR has been a research hotspot in theoretical research and engineering applications.The thesis first elaborates the theories about the revelant research.Then,based on the theoretical analysis,this thesis studies different kinds of SR systems and their applications in WSD.The followings are the main works and contributions of the thesis:(1)The WSD method based on a linear coupled bistable system is studied and the SR phenomenon under Lévy noise is also analyzed.Firstly,the generation method and the probability density function of Lévy noise are introduced.Then,the mean of signal-to-noise ratio increase(MSNRI)is selected as the measurement index and the SR phenomenon is further researched.Finally,combined with the fourth Runge-Kutta algorithm and the adaptive algorithm,this system is applied to detect different kinds of signals,which verifies the effectiveness and engineering practicability of this system in WSD.(2)An unsaturated bistable system with the piecewise nonlinear structure is constructed.The SR phenomenon of the system driven by trichotomous noise is discussed and it is used for bearing fault diagnosis.The thesis first analyzes the output saturation phenomenon of the traditional bistable system,but the unsaturated system can break the limitation of output saturation to some extent.Based on the adiabatic approximation theory,the theoretical expression of the system output SNR is derived.The unimodal and bimodal SR phenomenon are analyzed under the MSNRI index.Finally,in practical engineering applications,the unsaturated bistable system is applied to the bearing fault diagnosis.The experiments demonstrate that the detection effect and performance of the system are superior to the traditional bistable system.(3)The time-delayed asymmetric bistable system influenced by the multiplicative and additive Gaussian white noise is studied and the theoretical expression of the output SNR is also derived by using a small delay approximation method.The relationship between parameters and SNR is analyzed.Then,the results show that the asymmetric factor and time delay can further improve system performance.Finally,the effectiveness and superiority of the system are verified by bearing fault diagnosis.