Reinforcement Learning Optimization Algorithm to Minimize Power Losses in Electric Distribution Networks through Allocation of Resistively Distributed Generators

This paper proposes a new solution to reduce power losses through the allocation of Distributed Generators (DGs) in a resilient manner using the combined approach of three-dimensional Kalman Filter (۳d-KF) and Reinforcement Learning (RL) in radial distribution networks. The objective of this task is to minimize the real power losses of the network which satisfies the operational constraints of the DGs and the distribution network. The proposed approach uses the IEEE ۳۳ bus standard to reduce the percentage of power losses and errors. The simulation results have shown that the proposed method can provide a better solution quality than many other methods for the considered scenarios and has been able to help reducing power by ۱۱%.

Keywords:

Distribution Networks (DNs) , Distributed Generators (DGs) , Power Loss , Fault , Kalman Filter , Reinforcement Learning