A novel approach for efficient power conversion in solar photovoltaic-based cascaded H-bridge multilevel inverters with DC–DC converters for grid integration

With the depletion of conventional fossil fuels (FS)and the rising demand for energy in the world, solar energy has emerged as one of the most promising renewable energy sources (RES). However, the power conversion efficiency and its integration to the grid performance are affected by nonlinear characteristics and harmonic distortions of PV systems. The main objective of this study is to design an intelligent and optimized control strategy to enhance the power quality (PQ), maximum power extraction and overall efficiency of the PV-based cascaded H-bridge multilevel inverter (CHB-MLI). The methodology proposed integrates Loss-Attentional Physics-Informed Neural Network (LA-PINN) to generate the switching signals in the MLI in an adaptive manner, while maintaining the physical consistency of the control under dynamic operating conditions. The House Swallow Optimizer (HSO) algorithm is utilized to improve maximum power point tracking (MPPT) and duty cycle control of DC–DC boost converter. The proposed LA-PINN model integrates the inverter physical constraints to ensure the strong performance. The HSO algorithm also optimizes PV power extraction and inverter modulation parameters by using a multi-objective fitness function of minimization of THD and minimization of switching losses and keeping the DC-link stable. The proposed system is simulated in MATLAB and its performance is evaluated. The total harmonic distortion of the proposed system is minimized significantly to 1.4% and system efficiency is enhanced up to 98.5%. The proposed LA-PINN-HSO approach offers superior performance compared to conventional approaches like ant colony optimization (ACO), particle swarm optimization (PSO), and Black Widow Optimization (BWO) algorithm, including better power quality and quicker convergence.