Research On Gearbox Weak Fault Diagnosis Based On The Stochastic Resonance Theory

Taking the gearbox of urban rail train as research object, the paper focuses on the fault signal detection issue under strong noise with the method of stochastic resonance.Firstly, the article describes the background of urban rail train gearbox, combed major weak fault signal processing methods. Based on this, the paper proposes that treating the early weak fault of gearbox with the method of stochastic resonance, analyzes the failure modes and vibration mechanism, gives a simulation model.Secondly, stochastic resonance is limited by near-adiabatic conditions, can only handle low frequency signals. In order to break through near-adiabatic conditions, normalized parameter adjustment method is useed so that stochastic resonance can be applied to detect large frequency signals. Fault characteristic signal tends to be modulated with high mechanical natural vibration signal, in order to get a better fault information, the paper proposes that combining demodulated resonance with stochastic resonance to diagnose early gearbox crack information. And it achieves a good result.Thirdly, it needs a lot of work to adjust the parameters for there are two parameters need to adjust. In order to reduce the number of stochastic resonance tunable parameters, the paper uses simulated signals in matlab environment, studies the bistable stochastic resonance output characteristics from the shape feature of potential function, proposes two methods to control parameters, reduces the adjustable parameters of the system to one and the workload of parameter adjustment.Finally, the paper sets up the traditional bistable stochastic resonance circuit using Multisim12.0 software, simplifies the circuit design based on the research in Chapter 4, and tests the circuit improved. Paper designes voltage mode resonance demodulation circuit, and testes each part of the circuit. Data conversion module is writed using labview software, and using the experimental data after conversion tests the performance of the entire circuit.