A design method for single-mode SH1 wave PPM-EMAT based on nondominated sorting genetic algorithms-II

Abstract The permanent periodic magnet electromagnetic acoustic transducer (PPM-EMAT) may generate multiple shear horizontal (SH) wave modes under identical excitation conditions, leading to interference in defect detection. Traditional time-space domain qualitative design methods can control modes but lack quantitative standards and a systematic parameter design process, often resulting in oversized EMATs and poor signal time resolution. To address this issue, we propose a design method for the single-mode SH1 wave PPM-EMAT, based on the non-dominated sorting genetic algorithm-II (NSGA-II). This method comprehensively considers the coupled effects of wavenumber, frequency, and bandwidth on multi-mode behavior. It optimizes the time-space parameters of the EMAT by balancing the purity of the target mode with the number of magnetic periods and excitation signal cycles. First, we derive the energy amplitude function and single-mode amplitude coefficients based on the analysis of the excitation force. Next, we analyze the effects of time-space parameters. Finally, multi-objective optimization is performed using the NSGA-II algorithm to obtain Pareto optimal solutions, with the optimal compromise solution chosen using the fan-shaped decision method. Experimental results demonstrate that the purity of the SH wave excited by the optimized transducer increased from approximately 80% to approximately 90%, compared to the reference group. Additionally, the number of excitation signal cycles and magnet periods was reduced by 55.6% and 40.0%, respectively, compared to traditional methods. The design method significantly reduced the magnetic periods and excitation signal cycles, thereby minimizing EMAT size and improving signal time resolution. Simultaneously, it maintains the generation of high-purity SH1 waves.