The Multiple-Objective Design and Optimization of a Linear Vernier Motor with Spoke Structure Based on an Extreme Learning Machine

Nowadays, direct-drive systems are widely used in the actuators of computer numerical control machine tools. Linear motors are widely used in high-end computer numerical control machine tools due to their high positioning accuracy, good dynamic response, and simple transmission structure. First, a high-thrust-density concentrated magnetic linear permanent magnet vernier motor is proposed in this paper, which is designed by machine learning and optimized through an artificial intelligence optimization algorithm, to improve the air-gap magnetic density of the motor and improve the thrust density of the motor in principle; compared with traditional linear permanent magnet synchronous motors, the thrust density is increased by 40%. Second, using finite element calculations, a regression machine learning algorithm is proposed, which involves introducing a regression machine learning algorithm (called extreme learning machine (ELM)) to solve the computational modeling problem; compared with traditional ELM networks, it has faster training speed and higher stability. By mapping the nonlinear complex relationship between input structural factors and output motor performance, the superiority of the intelligent optimization algorithm is confirmed by comparative verification.