Research on Selective Harmonic Elimination Pulse Width Modulation of Three-Level NPC Inverter Based on Adaptive Restart Genetic Algorithm

The harmonic elimination task of three-level neutral point clamped (3L-NPC) inverters is related to a complicated engineering optimization problem, which mainly involves constrained nonlinear transcendental equations with multiple solutions. In order to solve the problems of difficult initial value determination, slow convergence rates and susceptibility to local optima trapping in the process of using traditional numerical algorithms, This paper introduces a novel selective harmonic elimination pulse width modulation (SHEPWM) framework, enhanced by an adaptive restart genetic algorithm, to address the limitations of conventional optimization methods. By introducing the adaptive restart mechanism, the performance and robustness of the solution algorithm can be improved, so as to circumvent entrapment in local optima and secure the global optimum across the entire solution space. The simulation results show that there are divergent trends in the switching angles and the total harmonic distortion (THD) values of the output voltage obtained by the proposed method, but all of them can effectively eliminate the specified low-order harmonics. These findings verifies the practical applicability and solution diversity of the proposed method, as well as provides new ideas and methods for solving optimization problems in other engineering fields.