LSTM-based forecasting on electric vehicles battery swapping demand: Addressing infrastructure challenge in Indonesia

Abstract

This article aims to design a model for forecasting the number of vehicles arriving at the battery swap station (BSS). In our case, we study the relevance of the proposed approach given the rapid increase in electric vehicle users in Indonesia. Due to the vehicle electrification program from the government of Indonesia and the lack of supporting infrastructure, forecasting battery swap demands is very important for charging schedules. Forecasting the number of vehicles is done using machine learning with the long short-term memory (LSTM) method. The method is used to predict sequential data because of its ability to review previous data in addition to the current input. The result of the forecasting using the LSTM method yields a prediction score using the root-mean-square error (RMSE) of 2.3079 x 10-6 . The forecasted data can be combined with the battery charging model to acquire predicted hourly battery availability that can be processed further for optimization and scheduling.

Keywords

• battery swap station (bss)
• demand forecasting
• long short-term memory (lstm).