PREDICTION OF ELECTRIC VEHICLE ENERGY CONSUMPTION

Abstract

Electric vehicles(EVs) have attracted a lot of interest as a viable and sustainable answer to the problems associated with contemporary transportation. Accurate energy consumption forecasting is crucial for enhancing the usability of electric vehicles, as it allows for precise prediction of the State of Charge (SoC) at the end of a journey.. To lessen the range anxiety that prospective users frequently feel, these problems must be addressed. The thesis presents a comprehensive study on Energy consumption prediction of electric vehicles (EVs). Several power consumption prediction models are developed, including a linear regression model, a polynomial regression model, a decision tree model, and an artificial neural network (ANN) model, achieving R2 values of 0.694, 0.975, 0.9925, and 0.9975, respectively. Traction force equation is developed. Additionally, a speed prediction model is created which is optimized to take the past 10 minutes of speed profile data as input to forecast the speed profile data for next 5 minutes. An energy consumption model is also developed using gradient boosting, attaining an R2 value of 0.8425. Finally, a web application is built that leverages this energy consumption model to accurately predict the State of Charge (SoC) of an EV at the end of a journey, significantly enhancing EV usability through precise energy consumption forecasts.