Repozytorium publikacji - Politechnika Gdańska

Electric vehicles (EVs) have been widely adopted to prevent global warming in recent years. The higher installation of Level-1 and Level-2 chargers in residential areas soon poses challenges to the distributed network. However, such challenges can be mitigated through the adoption of smart charging or controlled charging schemes. To facilitate the implementation of smart charging, accurate forecasting of EV charging demand in residential sectors is essential. This study focuses on utilizing machine learning (ML) techniques to predict EV charging demand based on charging data from Trondheim, Norway. A key contribution of this research is its systematic approach, providing a step-by-step implementation process of EV load forecasting using ML algorithms. By addressing the pressing issues surrounding the increasing demand for EV charging in residential areas, this paper offers valuable insights into sustainable transportation energy management. The results, obtained through Linear SVM, Wide Neural Network, Naive Bayes, and K-Nearest Neighbors algorithms implemented in MATLAB software, emphasize the effectiveness of ML techniques. This paper serves as an intelligent-based residential EV load forecast approach for researchers, policymakers, and industry professionals seeking effective strategies to mitigate the impact of EV charging on distributed networks.

Autorzy

• Sheetal Deshmukh,
• dr hab. inż. Atif Iqbal,
• Mousa Marzband,
• Mohammad Amir,
• prof. dr hab. inż. Jarosław Guziński link otwiera się w nowej karcie