Smart Charging of Community Storage Units Using Markov Chains

Abstract

Community Energy Storage (CES) is emerging as an alternative solution to home local energy storage for increasing the utilization of Renewable Energy Sources (RESs) in households. In this paper, a stochastic smart charging framework for CES in residential microgrids is proposed. A linear optimization problem for scheduling the charging process of the community battery to times when electricity prices are low, while accommodating the aggregated surplus renewable energy of households, is formulated. The goal is to satisfy the aggregated residual load of households at every time slot. To do so, a Markov chain-based forecasting approach is used for generating synthetic aggregated surplus solar Photovoltaics (PV) power and residual load profiles day-ahead. Numerical results are obtained using a history of real load and solar PV generation profiles of 10 households in the city of Amersfoort, the Netherlands. The forecasting performance is evaluated and compared with a persistence model by means of Root Mean Squared Error (RMSE). Then, the technical and economic performance of the smart charging process is presented for different annual periods. Results show that based on a time-varying electricity tariff and depending on the annual period, the CES with a smart charging process can bring a cost saving up to 68% in comparison with the traditional scenario without a storage.