Stochastic Energy Management Strategy for Autonomous PV–Microgrid Under Unpredictable Load Consumption

Abstract

This paper introduces a novel energy management strategy incorporating stochastic elements designed for off-grid photovoltaic (PV) systems supplying multiple loads in environments marked by unpredictable power usage. In these PV microgrid applications, unpredictable power consumption can lead to discrepancies between energy supply and demand, compromising system reliability and efficiency. This issue is especially pertinent in providing reliable electricity to remote or rural areas where conventional grid infrastructure is not available or reliable. To overcome this challenge, this paper addresses the random variability in load consumption by modeling it as a Markov decision process (MDP). The MDP framework facilitates the development of an effective decision-making process, accounting for the probabilistic nature of energy consumption patterns. Furthermore, by integrating MDP-based load consumption prediction into the energy management system, real-time optimization of both PV power and battery charging and discharging within the microgrid is achieved. This integration balances energy production and consumption, enhancing overall system efficiency. Three scenarios were examined to evaluate the effectiveness of the suggested strategy in enhancing the real-time operation of off-grid PV systems: standard test conditions, time-varying climatic profiles, and real-time weather situations. The findings indicate that the proposed strategy can adapt to dynamic load profiles, ensuring efficient energy utilization while maintaining microgrid stability.

Keywords

• Photovoltaic system
• battery storage
• state of charge
• energy management
• stochastic control
• Markov decision process