Mixed‐integer stochastic evaluation of battery energy storage system integration strategies in distribution systems

Abstract

Abstract This paper presents a new approach to the problem of defining an investment policy in battery energy storage systems in active distribution networks, taking into account a diversity of uncertainties. The proposed methodology allows the selection of type, capacity, and location of battery energy storage systems in distribution networks with distributed generation and electric vehicle charging stations. A mixed‐integer stochastic programming problem is cunningly approached with a metaheuristic, where fitness calculation with stochastic scenarios is performed by introducing an approximation to the operation costs in the form of a polynomial neural network, generated according to the Group Method of Data Handling—GMDH method, with strong computing speeding‐up. The quality of this approximation for heavy Monte Carlo simulations is assessed in a first case study using a 33‐bus distribution test system. The optimization planning model is then validated in the same test system using real data collected from solar and wind sources, demand, prices, and charging stations. Four types of batteries are compared considering degradation impact. The results demonstrate the practicality and advantages of this process.