Optimal design of Halbach magnetized magnetic screw for wave energy converters based on KELM network optimized by weighted mean of vectors algorithm

This paper investigates the parameter optimization problem of Halbach magnetized magnetic screw (HMMS) for wave energy converter (WECs). A magnetic screw with improved Halbach magnetized PMs arrays is presented. To further enhance the thrust density of HMMS, a HMMS optimization model is proposed based on the kernel extreme learning machine (KELM) optimized by weIght meaN oF vectOrs (INFO) algorithm. The topology and working principle of HMMS are introduced. Based on comprehensive sensitivity analysis, the design space is stratified, and the nonlinear coupling relationships between parameters are addressed using the Kriging model to enhance the accuracy and efficiency of the optimization process. The INFO algorithm optimizes kernel parameters and regularization coefficients of KELM, which critically affect its output accuracy. By establishing the INFO-KELM optimization model, the final optimized structure is obtained. The FEA is utilized to assess the properties of the HMMS. Compared with the traditional radially magnetized screw, the thrust force of the optimized HMMS increased by 40.8%. Finally, a prototype is developed and platform tests are performed to validate the theoretical analysis results.