AI-Driven Adaptive Energy Management of PV-Battery Integrated Owerri Urban Distribution Network Using MATLAB with Real-Time Uncertainty Handling

Abstract

Growing deployment of distributed renewable energy resources is essential to realizing sustainability and carbon neutrality in contemporary power networks. Nevertheless, the increasing penetration of these resources in power distribution networks can create operational uncertainties due to solar intermittency, stochastic load variations, and voltage instability. Thus, efficient coordination of these dispersed resources needs smart energy management schedules that have the capability to adapt dynamically under varying operational situations. This work proposes a PV–battery integrated 82-bus Owerri urban distribution network model using MATLAB/Simulink. To address the real-time uncertainties in solar generation and load demand, an AI– based energy management framework is proposed. This is done through a framework that integrates RL with PSO to dynamically cooperate with photovoltaic generation, battery energy storage, and grid power. BFS methodology is adopted for the computation of power flow applied to the distribution grid under stochastic operating conditions for the evaluation of voltage stability and power loss performance. Numerical results showed that the proposed AI-based energy management system is superior to traditional manual control strategies. It was equally observed that the RL-based controller enhanced the voltage magnitude from 0.86 to 0.96 p.u. while the value of real power loss declined by 42.9% (210kW-120kW). The findings validated that the proposed adaptive AI-based energy management can significantly increase the incorporation of PV systems, voltage stability, and operational resilience for forthcoming smart network distribution systems, in line with standards stipulated by IEEE, NERC, and NEMSA.