Prediction of EV Energy consumption Using Random Forest And XGBoost

Abstract

As climatic crisis increasing in the world due to the increasing pollution day by day, in which one of the important contributor is the increasing demand for energy,it has been found that 25 percent of the global energy consumption is only due to transportation sector, so in order to minimize the effect due to transportation sector we have to shift from Internal Combustion Engine (ICE) to battery based Electric Vehicle (EVs), there are several issues that need to be addressed to encourage the adoption of EVs on large scale, one of the severe effect due to large scale adoption of EVs is on the grid, large scale deployment of EVs causes overloading on the power grid due to the unscheduled charging of the EVs, in order to reduce this overloading of the grid proper scheduling algorithm are needed by which we can scheduled the charging of EVs and growing public charging demand for EVs. For this data-driven tools can be utilized which can predict the various parameters like energy consumption, time of charging, whether the EVs use charging stationtomorrow, use of DC fast charging etc, in this paper we are focusing on the prediction of energy consumption by using the historical charging data of the EVs by using various popular machine learning (ML) algorithm such as Random Forest, XGboost, linear Regression, ANN and DNN. The best predictive results are obtained by Random Forest and XGboost, and various result of past work is also discussed.