Design and Implementation of Hybrid GA-PSO-Based Harmonic Mitigation Technique for Modified Packed U-Cell Inverters

• Published in: Energies (Basel) Vol. 18; no. 1; p. 124
• Main Authors: Iqbal, Hasan, Sarwat, Arif
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.01.2025
• Subjects: Algorithms; Alternative energy sources; Efficiency; Electric inverters; Fault tolerance; GA-PSO; Mathematical optimization; Methods; MLI; modified packed U-cell; Optimization techniques; Performance evaluation; power quality; renewable energy; Renewable resources; THD
• ISSN: 1996-1073; 1996-1073
• DOI: 10.3390/en18010124

Multilevel inverters have gained importance in modern power systems during the last few years because of their high power quality with lower THD. Various topologies developed include the packed U-cell inverter and its different modified versions that have emerged as a compact and efficient solution to distributed energy systems. Most of the available harmonic mitigation techniques, that is, passive filtering and individual optimization techniques, which include GA and PSO, are susceptible to a variety of shortcomings regarding their inherent complexity and inefficiency; hence, finding an appropriate convergence may be quite hard. This paper proposes a hybrid version of the GA-PSO algorithm that exploits the exploratory strengths of GA and the convergence efficiencies of PSO in determining the optimized switching angles for SHM techniques applied to modified five-level and seven-level PUC inverters. By utilizing the multi-objective optimization method, the approach minimizes THD while keeping voltage and efficiency constraints. Simulated in MATLAB/Simulink, the results were experimentally verified using hardware-in-the-loop testing on OP5700. A large THD reduction in both MPUC7 (11.68%) and MPUC5 (17.61%) was obtained. The proposed hybrid algorithm outperformed the standalone approaches of GA and PSO with respect to robustness and with precise harmonic suppression. Other appealing features are reduced computational complexity and improved waveform quality; hence, the method is highly suitable for both grid-tied and standalone renewable energy applications. This work lays a basis for efficient inverter designs that can adapt well under dynamic load conditions.