Comparison of recent metaheuristic optimization algorithms to solve the SHE optimization problem in MLI


Yiğit H., Ürgün S., Mirjalili S.

NEURAL COMPUTING & APPLICATIONS, vol.35, no.10, pp.7369-7388, 2023 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 35 Issue: 10
  • Publication Date: 2023
  • Doi Number: 10.1007/s00521-022-07980-1
  • Journal Name: NEURAL COMPUTING & APPLICATIONS
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Applied Science & Technology Source, Biotechnology Research Abstracts, Compendex, Computer & Applied Sciences, Index Islamicus, INSPEC, zbMATH
  • Page Numbers: pp.7369-7388
  • Keywords: Metaheuristic optimization algorithm, Multilevel inverter (MLI), Selective harmonic elimination (SHE), Total harmonic distortion (THD), SELECTIVE HARMONIC ELIMINATION, MULTILEVEL INVERTER TOPOLOGIES, PARTICLE SWARM OPTIMIZATION, MODULATION, EVOLUTION, EFFICIENT, SCHEME
  • Kocaeli University Affiliated: Yes

Abstract

Multilevel inverters (MLIs) are one of the most popular topics of power electronics. Selective harmonic elimination (SHE) method is used to eliminate low-order harmonics in the MLI output voltage by determining the optimum switching angles. It includes the solution of nonlinear sets of transcendental equations. The optimization becomes more difficult as the number of levels in MLIs increases. Therefore, various metaheuristic algorithms have emerged toward obtaining optimal solutions to find the switching angles in the SHE problem in the last decade. In this study, a number of recent metaheuristics, such as ant lion optimization (ALO), hummingbird algorithm (AHA), dragonfly algorithm (DA), harris hawk optimization, moth flame optimizer (MFO), sine cosine algorithm (SCA), flow direction algorithm (FDA), equilibrium optimizer (EO), atom search optimization, artificial electric field algorithm and arithmetic optimization algorithm (AOA), are employed as an attempt to find the best optimization framework to identify switching moments in 11-level MLI. Marine predator algorithm (MPA), whale optimization algorithm (WOA), grey wolf optimizer (GWO), particle swarm optimization (PSO), multiverse optimizer (MVO), teaching-learning-based optimization (TLBO), and genetic algorithm (GA), which are widely used in solving this problem, are selected for performance analysis. AHA, ALO, AOA, DA, EO, FDA, GA, GWO, MFO, MPA, MVO, PSO, SCA, SSA, TLBO and WOA methods meet maximum 8% THD requirement specified in IEEE 519 standard in the range of 0.4-0.9 modulation index. Simulation results show that MFO is superior other algorithms in terms of THD minimization, convergence rate, a single iteration time and robustness.