Meshing theory of point-contact conical-envelope cylindrical worm-face worm gear drive

Abstract

This paper proposes further developing mismatched modification technology in meshing theory via physics-informed neural networks (PINNs). Thus, a design approach that considers meshing performance and meshing theory is presented for the mismatched parameters. On this basis, an innovative point-contact face worm gear drive with symmetric benchmark points is developed and its meshing theory is systematically developed. This study addresses high-dimensional nonlinear equation systems in tooth contact analysis (TCA) via PINN technology to convert initial value problems in conventional iterative methods into physical boundary problems. A PINN technique driven by meshing theory and meshing performance is proposed, achieving a coordinated optimization design for multiple mismatched parameters. The accuracy and feasibility of the PINN technique are proven by comparing it with the traditional iterative method. Applying the established meshing theory, TCA and error sensitivity analysis are conducted for the drive. The numerical results demonstrate that the PINN model has excellent accuracy in solving high-dimensional nonlinear equation systems. The analysis of the meshing characteristics indicates outstanding performance and a reduced susceptibility to installation errors.