Analyzing the Travel and Charging Behavior of Electric Vehicles -- A Data-driven Approach

TL;DR

This paper presents a data-driven approach using machine learning to predict electric vehicle travel and charging behavior, aiding in better electricity demand forecasting for power systems.

Contribution

It introduces a novel method combining survey data and machine learning to accurately model EV travel and charging patterns, addressing uncertainties in demand prediction.

Findings

Effective estimation of daily charging demand patterns

Acceptable accuracy in travel parameter forecasting

Machine learning models outperform baseline methods

Abstract

The increasing market penetration of electric vehicles (EVs) may pose significant electricity demand on power systems. This electricity demand is affected by the inherent uncertainties of EVs' travel behavior that makes forecasting the daily charging demand (CD) very challenging. In this project, we use the National House Hold Survey (NHTS) data to form sequences of trips, and develop machine learning models to predict the parameters of the next trip of the drivers, including trip start time, end time, and distance. These parameters are later used to model the temporal charging behavior of EVs. The simulation results show that the proposed modeling can effectively estimate the daily CD pattern based on travel behavior of EVs, and simple machine learning techniques can forecast the travel parameters with acceptable accuracy.