The rapid growth of electric vehicles (EVs) challenges urban energy systems, particularly in cities with limited historical charging data. Accurate forecasting of EV charging demand is essential for sustainable energy planning, grid stability, and infrastructure design. This study proposes a simulation-based forecasting approach that integrates urban traffic modeling with data-driven demand prediction to estimate charging loads under varying fleet sizes and charging strategies. Using Ljubljana as a case study, synthetic traffic and charging datasets are generated through MATSim simulations to represent diverse operational conditions. A feedforward neural network combined with least-squares approximation is applied to predict daily charging profiles. The framework demonstrates stable predictive performance across scenarios and provides insights into peak formation and load distribution dynamics. The approach offers a practical tool for planning EV integration in data-scarce urban environments.
Víctor Fernández Pallarés (Fri,) studied this question.