Research On Rolling Bearing Fault Diagnosis Method Based On QPSO-BONN Under Spark-GPU Platforms

Rolling bearing is an important part of rotating machinery and equipment.In industrial production,the running state of rolling bearing is related to the safe operation of the equipment and the normal operation of the entire production process.Therefore,the research on rolling bearing fault diagnosis has important theoretical value and practical significance.Traditional rolling bearing fault diagnosis methods based on signal processing technology are difficult to effectively extract the characteristics of different types of rolling bearing faults in the face of increasingly complex vibration signals.At the same time,with the rapid growth of rolling bearing monitoring data,how to achieve efficient fault diagnosis is also an important challenge that needs to be faced.To this end,this paper proposes a rolling bearing fault diagnosis method based on the quantum particle swarm optimization-back propagation neural network(QPSO-BPNN)under the Spark-GPU platform,which overcomes the shortcomings of traditional fault diagnosis methods that are difficult to identify complex vibration signals in different working environments.At the same time,it achieves efficient training and diagnosis of rolling bearing fault diagnosis models in the face of big data.In-depth research is carried out on the design of BP neural network structure,the optimization of neural network algorithm,and the distributed parallelization of the model.The specific research content mainly includes the following two aspects.In order to realize effective diagnosis of complex rolling bearing faults under different working conditions,a method of rolling bearing fault diagnosis based on QPSO-BP neural network and DS evidence theory is proposed.Firstly,the original vibration signal is decomposed by three-layer wavelet packet,and the eigenvectors of different states of rolling bearing are constructed as the input of BP neural network.Then,the optimal hidden layer nodes of BP neural network are automatically searched by dichotomy method to determine the network structure,and the initial weights and thresholds of BP neural network are optimized by quantum particle swarm optimization algorithm to improve the convergence speed.Finally,the fault classification results of multiple QPSO-BPNN are fused by DS evidence theory,and the fault diagnosis model of rolling bearing is established.The experimental results show that the proposed method can effectively predict the rolling bearing faults under different working conditions.In order to effectively improve the training efficiency,diagnosis efficiency and diagnosis accuracy of the rolling bearing fault diagnosis model in the face of massive rolling bearing vibration data,in view of the Spark-GPU platform provides powerful distributed parallel computing capabilities and back propagation neural network optimized by quantum particle swarm optimization algorithm has the characteristics of low computational complexity and high diagnosis accuracy,a rolling bearing fault diagnosis method based on parallel QPSO-BPNN under Spark-GPU platform is proposed.First,the distributed parallelization of QPSO-BPNN model based on Spark-GPU platform is realized,which can improve the training efficiency and diagnosis efficiency of rolling bearing fault diagnosis model in the big data environment.Second,in order to improve the convergence speed of fault diagnosis model,a parameter update strategy suitable for the distributed parallel training of QPSO-BPNN model is designed.At each iteration during training,the local parameters of each worker node are collected to the master node,and the global parameters are updated according to the weights and synchronized to each worker node.Third,a combination strategy of multiple QPSO-BPNN models based on ensemble learning is proposed.The weighted voting method is adopted to combine the output results of different QPSO-BPNN models to obtain the best fault diagnosis result of a sample,which can improve the fault diagnosis accuracy to a certain extent.Experimental results show that the proposed method can quickly perform model training and fault diagnosis for large-scale rolling bearing vibration data,and the fault diagnosis accuracy reaches 98.73%.