No abstract

The presence of bypass diodes in photovoltaic (PV) arrays can mitigate the negative effects of partial shading conditions (PSCs), which can cause multiple peak characteristics at the output. However, conventional maximum power point tracking (MPPT) methods can develop errors and detect the local maximum power point (LMPP) instead of the global maximum power point (GMPP) under certain circumstances. To address this issue, several artificial intelligence (AI)‐based methods have been proposed, but they result in complicated and unreliable methodologies. This study introduces the driving training‐based optimization (DTBO) method, which aims to address the partial shading (PS) problem quickly and reliably in maximum power point (MPP) detection for PV systems. DTBO improves tracking speed and reduces fluctuations in the power output during the tracking period. The proposed method is extensively verified using the Typhoon hardware‐in‐the‐loop (HIL) 402 emulator and compared to conventional methods such as particle swarm optimization (PSO), and JAYA, as well as the recently proposed adaptive JAYA (AJAYA) method for MPPT in a PV system under similar conditions.

The application of iterative techniques for solving the Selective Harmonic Elimination Pulse Width Modulation (SHE‐PWM) problem, such as the Newton–Raphson (NR) method, can tend to get stuck at local optima in the solution space. Additionally, these methods may be sensitive to the initial value estimation of the solution. In contrast, metaheuristic approaches demonstrate resilience in seeking out the optimal solution. As such, this study utilizes the Runge–Kutta (RUN) metaheuristic optimization algorithm to demonstrate the SHE‐PWM technique in multilevel inverters (MLIs), including 5‐ and 7‐level modified H‐bridge (MHB) topology and a 9‐level asymmetric cascaded H‐bridge (CHB) inverter topology. The switching angles are obtained by varying the modulation index from 0.01 to 1.0 in steps of 0.01 for 5‐, 7‐, and 9‐level MLIs. The performance of the RUN algorithm in minimizing the total harmonic distortion (THD) value is verified through simulations in MATLAB/Simulink software. The superiority of RUN is established by comparing the results with state‐of‐the‐art metaheuristic algorithms, such as the Differential Evolution (DE), Genetic Algorithm (GA), and Grey Wolf Optimizer (GWO). Additionally, the switching angles obtained through RUN are validated by hardware experiments. According to the simulation and experimental results, the proposed RUN method exhibits better performance in terms of objective function values, algorithm robustness, fundamental harmonic magnitude, and THD values. The findings confirm the elimination of fifth harmonic, fifth, and seventh harmonics, and fifth, seventh, and ninth harmonics in 5‐, 7‐, and 9‐level MLIs, respectively.

No abstract