Track: Masters Thesis Competition

Abstract

With the rise of Electric Vehicles (EVs) and Electric Vehicle Batteries (EVBs), effective management of these components has become crucial. This study focuses on forecasting sales of Battery Electric Vehicles (BEVs) and predicting the volume and return dates of their EVBs using advanced statistical methods. By analyzing BEV sales data for trends, seasonality, and stationarity, the research employs a combination of AutoRegressive Integrated Moving Average (ARIMA), Seasonal ARIMA (SARIMA), and linear regression models. The SARIMA model emerged as the most effective, exhibiting a Mean Absolute Percentage Error (MAPE) of 6.769%, thus providing a reliable forecasting framework. The models were also benchmarked against traditional forecasting methods such as naive and seasonal naive approaches. Furthermore, the study utilizes Monte Carlo simulations to predict EVB return dates, incorporating factors like battery lifespan and delay distributions. The effect of technological advancements was also considered in the battery lifespan. The simulations predict a substantial surge in battery returns around the mid-2030s, highlighting a crucial phase for resource allocation in battery recovery initiatives, including recycling and refurbishment. These forecasts provide valuable insights to support stakeholders in the EVs industry by enhancing strategic planning and promoting sustainable practices in the production and end-of-life management of EVBs.