Switching Frequency Minimized Harmonic Mitigation: A Multiobjective Optimized Modulation Strategy for High- Power Converters

In order to explore the minimum switching frequency required to regulate certain amount of harmonic components, a novel switching frequencyminimized harmonic mitigation (SFMHM) model is proposed in this article, which can regulate N-1 harmonics with far less than N switching angles. By flexibly adjusting the threshold values, the proposed model can achieve both harmonic elimination and mitigation objectives. Based on pulsewidth modulation (PWM) discretization and supported by quadratic programming, the switching frequency in the proposed model is expressed as a quadratic objective function of the PWM waveform. The objectives as fundamental control and selected harmonic mitigations are realized by treating the numerical approximations of the Fourier coefficients as constraints, which are finally transformed into an optimization model with binary variables and can be easily solved by some optimization toolboxes, such as YALMIP. Some computing results show that, compared with the conventional selective harmonic elimination/mitigation (SHE/SHM) methods, the number of switching angles required to mitigate the same number of harmonics under the proposed method has been significantly reduced. The switching frequency can be reduced a lot compared with the conventional SHE/SHM methods, e.g., by 40% for some modulation indexes. Simulations and experiments verify the correctness of proposed SFMHM model.

Automated summary

In this article, a novel switching frequency-minimized harmonic mitigation (SFMHM) model is proposed to regulate N-1 harmonics with far less than N switching angles. By adjusting threshold values, the model achieves both harmonic elimination and mitigation objectives. The switching frequency is expressed as a quadratic objective function and the objectives are realized through numerical approximations of Fourier coefficients and an optimization model with binary variables, easily solved using optimization toolboxes such as YALMIP. Comparisons with conventional methods show significant reduction in the required number of switching angles and a notable decrease in switching frequency by up to 40% for certain modulation indexes. Simulations and experiments validate the proposed SFMHM model.