Efficient (2N+1) selective harmonic elimination in modular multilevel converters using an evolutionary many-tasking approach with prior knowledge

(2N+1) selective harmonic elimination pulse with modulation (SHE-PWM) is an important technique for a modular multi-level converter (MMC) to enhance the control ability and improve the fundamental wave amplitude accuracy of the output waveform. While finding optimal switch angles by solving the complex non-linear equation system is one of its main challenges. A current popular solution is to run a metaheuristic or evolutionary algorithm repeatedly on various modulation indexes. These methods ignore the fact that some of the optimization tasks are similar, and the optimization experiences could be shared among these similar tasks. To fill this gap, this study proposes an evolutionary manytasking approach with prior knowledge (EMT-PK). EMT-PK facilitates positive knowledge transfer between similar tasks and is capable to solve all the sub SHE-PWM problems simultaneously through a single run. Numerical results on 7-, 9- and 11-level MMCs show that EMT-PK provides superior results for the (2N+1) SHE-PWM in comparison with 8 other widely used metaheuristics/evolutionary algorithms in this field including differential evolution (DE), ant colony optimization (ACO), teaching and learning-based optimization (TLBO), bee algorithm (BA), genetic algorithm (GA), particle swarm optimization (PSO), generalized pattern search (GPS) and asynchronous particle swarm optimizationgenetic algorithm(APSO-GA). The superior performance of EMT-PK is further validated by the MATLAB Simulink simulation and the laboratory experiment on a 7-level MMC.& COPY; 2023 Published by Elsevier B.V.