Research On Gearbox Fault Diagnosis Based On VMD And Blind Source Separation

Gearbox is the basic components in machinery and its operational status is critical to the safe operation of the entire plant.Due to the complicated structure of the gear box and the harsh working environment,it is prone to failure during operation,thus affecting the operation of the entire equipment.Therefore,it is necessary to use the appropriate signal processing technology to accurately extract the useful fault signal from the vibration signal of the gear box and diagnose the fault of the gear box.In this paper,the composite fault of gearbox is combined with the maximum correlated kurtosis deconvolution(MCKD)noise reduction,variational mode decomposition(VMD)method and fast independent component analysis(FastICA)method.The fault signal of the gear separation wheel and the rolling bearing is successfully obtained from the gearbox vibration signal,and the fault location is found.The main works of this thesis are as follow:Firstly,the vibration models of gears and rolling bearings and their common failure modes were introduced.The evaluation indexes of time domain signals were put forward,and the common signal processing methods for gearbox fault diagnosis were summarized.At the same time,the drivetrain diagnostics simulator(DDS)test bench was used to simulate the composite fault of the gear and rolling bearing in the gearbox,and the single-channel gearbox vibration signal was collected for analysis,which proved that the conventional method directly processed the composite fault signal is not ideal.Secondly,the blind source separation algorithm and the blind source separation algorithm in single channel were discussed.It was verified that the VMD algorithm is superior to the empirical mode decomposition(EMD)and the ensemble empirical mode decomposition(EEMD)algorithm.The single channel signal was converted into multi-channel signal by VMD decomposition,and the effective Intrinsic Mode Function(IMF)component reconstruction signal was selected according to the three criteria of correlation coefficient,kurtosis and variance.The FastICA algorithm was used to separate the reconstructed signals,and the feasibility of the blind source separation algorithm was proved by simulation signals.Finally,the problem of blind source separation in strong noise environment is studied,and the MCKD-VMD-FastICA algorithm was proposed.The superiority of the proposed algorithm was verified by simulation signal and measured vibration signal,which can effectively combine from single channel.Gear faults and bearing faults were separated from the fault signal and their position was determined.