Daily EV Load Prediction Using Fuzzy Inference: A Microgrid Planning Perspective

The rapid rise in electric vehicle (EV) adoption highlights the critical need for a reliable charging infrastructure to ensure the stability of power distribution networks. This research introduces a fuzzy inference system (FIS) designed to forecast daily EV loads essential for developing microgrids to meet the increasing demand for EVs. The present work considers four factors for FIS designing: travel distance, parking duration, battery state of charge (SoC), and expected arrival times at charging stations. By developing fuzzy logic rules for these variables, a probabilistic charging is generated, improving both the precision and adaptability of load forecasts. This study also explores the impact of future EV adoption on microgrid load demand, analyzing adoption rates of 53%, 68%, and 84%, providing crucial insights for planning microgrids. The discrepancy between estimated and actual EV loads is found to be 0.078, demonstrating a reduction in prediction error. This effectively mitigates uncertainties related to EV user behavior and supports the design of resilient and flexible microgrid systems.