A Parameter correction method of motor system based on particle swarm and convolutional neural network

In order to generate accurate predictions in the vibration characteristics analysis of the motor structure, a motor system parameter correction method based on particle swarm optimization is proposed in this paper. Initially, focusing on the motor itself and utilizing modal analysis theory, finite element analysis software is employed for modal simulation analysis of the motor model. Subsequently, parameters such as shell stiffness, shell density, and spring stiffness between end caps and the shell are adjusted to better match the actual motor model. The particle swarm optimization algorithm is then used to optimize the structural parameters of the motor, obtaining a non-inferior optimal solution between the vibration performance and operational performance of the driving motor. At the same time, in order to reduce the manual workload, the image recognition model is established according to convolutional neural network to find out the target output image. Finally, the proposed method is subjected to simulation analysis. The results indicate that this strategy can to some extent achieve identification and optimization of motor system parameters. Additionally, the algorithm can reduce the complexity of modal calculations for motor systems and the image recognition model can find the target pattern accurately, demonstrating practical value.