Rolling Bearing Fault Diagnosis Based On Adaptive LMS Algorithm

Rolling bearing is one of the key parts of the train running part.Its good condition is an important guarantee for the safe operation of the train.Therefore,the method of rolling bearing fault diagnosis is a hot research direction.How to choose an appropriate signal processing method to extract fault information effectively and realize fault diagnosis is very important.An adaptive processing method for rolling bearing fault diagnosis was proposed.The adaptive LMS filtering algorithm was used in the preprocessing of rolling bearing vibration signal.The main contents of the research are as follows:Firstly,the research background and significance of this subject were introduced.The research and development status of variable step length LMS algorithm and rolling bearing fault diagnosis were expounded in detail.After that,the theory and principle of adaptive LMS algorithm were introduced in detail,and the performance of the algorithm was analyzed from four aspects.Lastly,The application of the algorithm was introduced in signal noise reduction.Secondly,three methods were proposed for improving the LMS algorithm,and two commonly used improved algorithms were introduced.The first one was the normalized LMS algorithm,which solves the problem that the gradient noise would be increased when the input signal is larger.The second is variable-step LMS algorithm based on S-type function,which solves the problem that the traditional algorithm conflicted with the convergence speed and steady-state error requirements.Based on these two algorithms,this paper proposes an LMS algorithm of normalized hyperbolic tangent function,introduces the mechanism of the algorithm,and analyzes the influence of parameters on the performance of the algorithm and the possible optimal value of parameters.Thirdly,the convergence performance of the variable step size LMS algorithm mainly depends on the selection of the parameter values.The Traditional method rely on experience or multiple trial to determine the optimal value of the parameters,this method has certain limitations.In this paper,the genetic algorithm was used to optimize the parameters of the variable step size LMS algorithm step factor to obtain the optimal value of the parameters.At the same time,the performance of different algorithms selected optimal value were compared by computer simulation.The result shows that the proposed algorithm has better convergence performance.Then the noise reduction analysis of the adaptive LMS algorithm was carried out by using the rolling bearing fault simulation signal,and the noise reduction effect under different SNR and different noise conditions was analyzed,and the denoised signal is envelope demodulated.The comparison is different.The effect of feature extraction in the envelope spectrum under conditions.Finally,due to the low SNR of the rolling bearing fault signal caused by noise interference and the complex noise environment,in order to extract the fault information accurately and effectively,this paper proposed a fault diagnosis method for rolling bearing based on VMD and adaptive LMS algorithm.The signal was decomposed into several IMF components by VMD,and the two components with the largest kurtosis value were selected according to the kurtosis criterion for reconstruction.Then,the reconstructed signal was filtered by adaptive LMS algorithm,and then the envelope demodulation was performed to obtain the bearing fault characteristic frequency.The experimental analysis shows that the combination of VMD and adaptive LMS algorithm achieved good results in the fault feature extraction of rolling bearings,and can extract fault features and realize fault diagnosis effectively.