PSO Algorithm Based On Mutation Operator Optimizes RBF Network Application In Fault Diagnosis Of Gearbox

The gear box is an important part of the transmission machinery,and its operating state is related to the production of the entire machinery and equipment.The operating environment and working conditions of gearboxes are relatively complex,and gears are components that are prone to failure.According to statistics,the two fault types of gear cracks and tooth surface cracks account for nearly half of all gear failure types.Therefore,on-line monitoring of gear conditions and the identification and warning of cracks are necessary.Particle Swarm Optimization(PSO)algorithm is a swarm intelligence algorithm to find the optimal solution.This paper first introduces a mutation operator to PSO to improve the algorithm,and the improved PSO and traditional PSO.The superiority between the simulation analysis and analysis.Then based on the improvement of PSO and combined with the specific problems to be solved in RBF network optimization,a PSO-optimized RBF network model based on mutation operator is established.The basic idea is: adding a mutation operator to improve the traditional PSO,the improved PSO is used to optimize the parameters such as the weights and thresholds of the RBF network,and the optimized network model uses the gearbox fault identification.The purpose is to make timely and accurate identification of various faults in the gear box,improve the life and stability of the gear box,and minimize the economic losses caused by the failure of the gear box.In this paper,the crack of the driving wheel is selected as the research object,and the fault manifests in four different degrees of damage,and the vibration signal is designed and collected in the experiment.The original signal was denoised using wavelet packet decomposition and reconstruction,and eight kinds of time-domain characteristic parameters sensitive to the fault were extracted.Using the proposed neural network model to analyze the characteristic parameters,the results show that the PSO-optimized RBF network model based on the mutation operator proposed in this paper is feasible,and makes a precise judgment on the fault type of the gearbox.In addition,in order to explain whether this article makes sense for the traditional PSO improvement,the same sample is analyzed and diagnosed using the standard PSO optimized RBF network model.The results show that the proposed network model has advantages.Therefore,the research in this dissertation shows that the improved PSO proposes an effective optimization strategy for RBF network,and it also improves the efficiency and accuracy of the network model to some extent,and then optimizes and extends PSO and The application of neural networks combined.