Optimal Installation of Remote-Controlled Switches and Reclosers in Distribution Networks Using Q-Learning Algorithm

The installation of remote-controlled switches and reclosers is a common solution for reducing the duration of power outages and improving the reliability indices of distribution networks. However, their high cost and the optimal placement problem make their installation a challenging decision. In this paper, we propose a reinforcement learning-based approach using the Q-learning algorithm to determine the optimal number and location of remote-controlled switches and recloser’s with the aim of maximizing the profit resulting from reducing the energy not supplied to customers and the outage costs. Also, considering load uncertainty, error rate, and repair rate. The proposed algorithm takes into account uncertainties in the network, such as load variation, fault rate, and repair rate, and uses a Monte Carlo simulation to simulate these uncertainties. The proposed approach is evaluated on a 13 buses distribution network, and the results show the effectiveness of the proposed approach in improving the network’s reliability indices and reducing the energy not supplied to customers and the outage costs.