Avoiding Low-Voltage Grid Congestion using Smart Charging of Electric Vehicles based on Day-Ahead Probabilistic Photovoltaic Forecasts

Abstract

Distribution System Operators (DSOs) can mitigate future grid congestion problems in low-voltage grids by applying smart charging algorithms to electric vehicles (EVs). However, application of real-time smart charging of EVs to mitigate local grid congestion could be problematic when aggregators cost-optimize EV charging by trading in electricity markets, as a deviation from the charging schedule for the provision of local grid services can lead to imbalance costs to the aggregator. Therefore, grid congestion problems should be forecasted, so aggregators can consider grid congestion in their electricity market bids and imbalance costs can be avoided. This study proposes a framework for mitigating grid congestion using EV smart charging, using probabilistic day-ahead forecasts of the grid load. The effectiveness of the proposed system in mitigating grid congestion is tested using day-ahead quantile regression forecasts for photovoltaic (PV) generation. Results indicate that transformer congestion problems reduce considerably when using probabilistic PV forecasts in EV scheduling. Considering a higher percentile in the PV generation forecast when scheduling EVs reduces grid congestion but marginally increases EV charging costs.