Cost Optimization of Shared Energy Storage Using Machine Learning /

Abstract

Distributed Energy Resources (DERs) are becoming pivotal in the evolution of smart grids. They include generating and storing energy. The stored energy is either utilized independently by a household or it can be shared among members within a community. Private energy storage systems have historically received focused attention, now community energy storage (CES) is also becoming prominent in countries worldwide. The current models are inclined towards optimizing private energy storage and fall short in addressing the collective sharing of storage capacitates. In this study, we explore setting up and optimizing the cost of a CES for a community in Islamabad, Pakistan. It incorporates diverse appliance load profiles, grid, PV generation, private energy storage, and community energy storage configurations. This research aims to reduce a household’s capital cost of setting up energy storage, minimize electricity operational costs, and provide an environment-friendly alternative to PES. For optimization modelling, mixed integer linear programming (MILP) is implemented within the PyCharm environment using Pyomo and Gurobi. The computational outcomes highlight the benefits of CES versus private storage, demonstrating significant cost reductions and savings and supporting the future environmentally responsible implementation of community energy storage.