An Efficient Robust Design Optimization Approach of Electromagnetic Relay Based on Surrogate Model and Evolutionary Algorithm

Abstract

The reliability and robustness of electromagnetic relay's performance characterize directly related to the performance, reliability and robustness of the aviation, aerospace, intelligent equipment industry, and its design and optimization have initiated boundless concern. The efficiency of the calculation of performance characteristics and the accuracy of the optimization process are two critical aspects affecting the design optimization of electromagnetic relay reliability and robustness. In response to the above issues, this study proposed an efficient computational burden reduction strategy for the multi-objective design and optimization aimed at the electromagnetic relay with substantial non-linearity and insufficient convergence characteristics. The dual response surface approach extracted the surrogate model of the mean value and variance for the critical static attractive force point. On the basis of the surrogate model, a particle swarm evolutionary algorithm for multi-objective design and optimization of electromagnetic relay has been modified. Then, the modified particle swarm evolutionary algorithm was used to optimize the dual response surface model, verifying the algorithm's feasibility. At last, the effectiveness of the proposed approach in this study was verified by a case study of the differential polarized magnetically maintained electromagnetic relay double permanent magnets.