Comparison of evolutionary algorithms for solving risk-based energy resource management considering conditional value-at-risk analysis

Energy management systems must evolve due to the widespread use of distributed energy resources in modern society. In fact, with the current high penetration of renewables and other resources like electric vehicles, the challenge of managing energy resources becomes more difficult. Uncertainty and unpredictability from distributed resources open the door for unique undesirable situations, often known as extreme events. Despite the low likelihood of occurrence, such severe events represent a significant risk to an aggregator's resource management, for example. In this paper, we propose a day-ahead energy resource management model for an aggregator in a 13-bus distribution network with high penetration of distributed energy resources. In the proposed model, we consider a risk-based mechanism through the conditional value-at-risk method for risk measurement of these extreme events. Due to the complexity of the model, we also propose the use of evolutionary algorithms, a set of stochastic search algorithms, to find near-optimal solutions to the problem. Results show that implementing risk-averse strategies reduces the cost of the worst scenario and scheduling. From the tested algorithms, ReSaDE provides the solutions with the lowest cost, which is an improvement from previous work, and a reduction of around 13% in the worst-scenario costs comparing a risk-neutral approach to a risk-averse approach. (c) 2023 The Authors. Published by Elsevier B.V. on behalf of International Association for Mathematics and Computers in Simulation (IMACS). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).