Prediction of Electric Vehicle Charge Profile using Battery Digital Twin

The transition towards vehicle electrification presents various challenges due to uncertainties in charging behavior and battery aging. This study proposes a method to generate an accurate charge profile with a battery Digital Twin (DT). The method is adaptive, has fast prediction and has low training cost. It is intended towards improving the charge scheduling for fleet operators by providing predictions on grid load and charging time. The method was tested with data from a real case study where, for a charging session, it showed an error of less than 2% for charge time and an improvement of 7% over a standard charge profile.