Deep Reinforcement Learning for the Optimal Angle Control of Tracking Bifacial Photovoltaic Systems

An optimal tilt-angle control based on artificial intelligence (AI control) for tracking bifacial photovoltaic (BPV) systems is developed in this study, and its effectiveness characteristics are examined by simulating a virtual system over five years. Using deep reinforcement learning (deep RL), the algorithm autonomously learns strategy real time from when starts to operate. Even with limited RL input variables, such as global horizontal irradiance, time, tilt angle, power, proposed AI successfully achieves 4.0–9.2% higher electrical-energy yield high-albedo cases (0.5 0.8) compared traditional sun-tracking control; however, energy of slightly lower low-albedo (0.2). also demonstrates superior performance there seasonal changes albedo. Moreover, robust against long-term degradation manipulating database used reward setting.