Charging Demand Prediction for Public Electric City buses of Helsinki, Finland by Support Vector Machine

Abstract

Global warming, crisis of energy, and degraded air quality index have compelled electrification of the transport sector. Public Electric Buses (e bus) are the first candidates for electrification as majority of public transport is dependent on them. Electrification of the public e buses will increase the load demand of the power grid thereby creating technological and commercial challenges. The stability and resilience of the power grid may be affected if the charging activities are performed in an uncoordinated manner. Thus, charging load prediction of the e buses is a crucial issue for maintaining smooth and hassle-free operation of the power system. Motivated by the aforementioned factor, this work aims to delve into charging demand prediction for e buses. A novel Support Vector Machine (SVM) based model is proposed for charging demand prediction. The model is validated for predicting the charging demand of e buses for the city of Helsinki, Finland. Simulation results establish the efficacy of the proposed approach. Further, a sensitivity analysis is also performed to validate the robustness and efficiency of the proposed approach and for obtaining the optimal values of the parameters of SVM.