Optimized hybrid multilevel inverter with vertical extension topology for energy management and power quality improvement in renewable energy systems

The increasing combination of renewable energy sources (RESs) into modern power systems requires effective energy management (EM) and improved power quality (PQ) under varying operating conditions. But, conventional inverter based methods exhibit high harmonic distortion, high switching losses and poor dynamic responses. To address these issues, this paper suggests a hybrid BTGO-SiGNN technique integrated with a hybrid 13-level multilevel inverter (MLI) for enhanced EM and PQ improvement in renewable energy (RE) systems. The suggested method involves using Banyan Tree Growth Optimization (BTGO) to optimally tune inverter switching parameters for harmonic reduction and switching loss minimization, and using the Spike-Induced Graph Neural Network (SiGNN) to control the energy flow and system operation under various dynamic renewables and load conditions. The suggested framework is implemented in MATLAB and tested under various operating conditions. Performance comparisons with existing methods including Genetic Algorithm and Particle Swarm Optimization (GA-PSO), Snow Ablation Optimizer and Matrix Diffractive Deep Neural Network (SAO-MDDNN), Anti-Predatory PSO (APSO), Equilibrium Optimizer (EO), Artificial Rabbits Optimized Neural Network (ARONN), Greylag Goose Optimization (GGO) demonstrate the superiority of the suggested approach. The suggested BTGO-SiGNN framework helps to achieve minimum total harmonic distortion (THD) of 3.32%, voltage regulation and reduce switching losses, transient response and stable operation under different load conditions. The obtained results confirm that the suggested hybrid 13-level MLI provides improved harmonic suppression, efficient energy coordination, and enhanced overall PQ performance for RE applications.