Latest revision as of 11:32, 1 September 2026

Abstract

This paper introduces a numerical optimization model (mixed-integer) of reliability-based design of energy storage systems (ESS) in the highrenewable-based power grid. This is done by developing a mixed-integer nonlinear programming (MINLP) model that incorporates probabilistic reliability indices (Loss of Load Probability and Expected Energy Not Supplied) directly into the objective function and also optimizes ESS sizing, placement, and operational scheduling. An approach that is a hybrid solution strategy that incorporates outer approximation and metaheuristic enhancement is used to deal with non-convexities in reliability calculations and modelling of storage degradation. Convergence analysis demonstrates that the algorithm attains optimality gaps of less than 3.5% in 150 iterations, 18–22 times more efficient than standard solvers, computationally. It has been shown with a 320 MW peak load benchmark system based on the IEEE 118-bus topology under five operating conditions that the optimized ESS configuration saves Loss of Load Probability from 0.0214 to 0.0062 (71% improvement), as well as Expected Energy Not Supplied from 1245 to 358 MWh/year, and unnecessary ESS cycling is reduced by 29%. The framework provides system planners with a computationally manageable tool to optimally deploy ESS in a renewable-dominated grid.OPEN ACCESS Received: 12/03/2026 Accepted: 29/06/2026

Document