AC Transmission System Expansion Planning using a Q-learning Algorithm

Abstract

The Transmission Network Expansion Planning (TNEP) is a complex optimization problem. Metaheuristic techniques and mathematical methods are currently used to obtain reliable results for linearized or simplified TNEP formulations. However, solving the TNEP problem using the full AC equations remains an issue. The use of machine learning in the context of power systems optimization has increased in recent years. Therefore, this research explores the application of reinforcement learning to solve the AC TNEP problem, framing it as an episodic decision-making process where each episode represents a distinct planning scenario. The results obtained for the Garver 6-bus system show that it is a viable method.