Research On Fault Diagnosis Of Rolling Bearing Based On Optimized Elman Neural Network

Rolling bearings are relatively important components in mechanical equipment,and they are widely used,but rolling bearings are also prone to failure,which will affect the normal operation of mechanical equipment and even cause serious consequences.In order to avoid large losses due to the failure of rolling bearings,the research on the fault diagnosis of rolling bearings is particularly important.This article firstly analyzes the main failure forms of rolling bearings and their fault frequency characteristics,and then research and analyze from the two aspects of rolling bearing vibration signal feature extraction and the fault diagnosis of the rolling bearing.The main contents of this article are as follows:Firstly,aiming at the feature extraction of rolling bearing vibration signals,this paper proposes an improved wavelet packet transform algorithm.Compared with the traditional wavelet packet transform algorithm,this improved algorithm can effectively eliminate frequency aliasing on the extracted signal by using the Fourier transform algorithm and the inverse Fourier transform algorithm to zero the excess frequency.It is also proposed to use information entropy to select the optimal wavelet function.It has been verified by experimental simulation,which effectively illustrates the effectiveness of the improved wavelet packet transform algorithm proposed in this paper.Secondly,aiming at the problem of fault identification of rolling bearings,the fault diagnosis method adopted in this paper is to use Elman neural network to diagnose the fault.As the connection weights and thresholds of the Elman neural network cannot reach the optimal connection weights and thresholds as the number of network trainings increases,as a result,the Elman neural network is not easy to obtain the optimal value.This paper proposes an improved adaptive genetic algorithm with global search capabilities,the improved adaptive genetic algorithm is used to optimize the Elman neural network,the improved adaptive genetic algorithm can solve the shortcomings of genetic algorithm easy to fall into premature phenomenon.Finally,in order to prove the effectiveness of the proposed algorithm,this paper uses the bearing data of the Western Reserve University Experimental Center.Non-optimized Elman neural network and Elman neural network optimized by improved adaptive genetic algorithm were used to diagnose the processed bearing data.According to the simulation results,it is found that compared with the unoptimized Elman neural network,the fault diagnosis rate of the Elman neural network optimized by the improved adaptive genetic algorithm is obviously improved,which illustrates the effectiveness of the method in this paper.