A hybrid AEGAN-PDO strategy for power quality enhancement in PV-based distributed generation with stacked multi-cell converter

Abstract

Conventional power plants pose a threat to the environment because of their substantial carbon emissions. Photovoltaic (PV) systems are becoming more and more popular as a sustainable alternative for clean electricity generation. However, because weather and environmental factors vary, partial shadowing affects PV output. The stacked multi-cell converter (SMC) provides a practical way to improve power extraction under these circumstances. This paper suggests a hybrid control approach for a photovoltaic (PV)-based distributed system (DS) using an SMC that is based on the attentive evolutionary generative adversarial network (AEGAN) and prairie dog optimization (PDO) algorithm. The AEGAN forecasts load requirements, while the PDO maximizes converter control to improve reliability, efficiency, and power quality (PQ). Under various load and irradiation circumstances, the system is modelled and verified in MATLAB/Simulink. Results from simulations show that the AEGAN-PDO approach performs better in both dynamic and steady-state situations. Transient disturbances on the load side are rapidly reduced with minimal overshoot. In contrast to traditional particle swarm optimization (PSO), ant lion optimizer (ALO), and archerfish hunting optimizer (AHO) controllers, AEGAN-PDO maintains the lowest THD (1.1%), least power loss (0.24 MW), and best efficiency (98.59%). These results validate the AEGAN-PDO approach as a reliable and effective way to operate renewable-integrated power systems in real-time, promoting improved PQ and grid dependability.