Intelligent Event Detection in Unbalance Active Distribution System Using Reinforcement Machine Learning Algorithms

With integration of Distributed Energy Resources (DERs) in the distribution system (DS), the power quality has decreased causing large deviation in voltage, frequency and, especially during disturbances, the security of the system may lose due to the maloperation of the relays. However, effective event detection can reduce the outage time and initiate faster restoration of the healthy section. Furthermore, the data collected from the DS is large in volume and variety and hence, to capture the accurate information from those data during any disturbance, is difficult and demands application machine learning techniques for event identification. In this paper, a Q-learning algorithm, is proposed for accurate event detection in distribution system following a disturbance. The previous and subsequent events data is utilized to identify the event. On the basis of events, the corresponding actions include isolating faulty sections, facilitating fast restoration, and load reduction as necessary. The efficacy of these algorithms is evaluated through simulations conducted on the IEEE 33-bus system with integration of photo-voltaic, utilizing MATLAB/SIMULINK 2023 and Jupyter Notebook and necessary actions are derived following the disturbance. The results obtained are satisfactory.