Charge up when, where, how long? Insights from European Public EV Charging

We will explore a study conducted by researchers from CTU Prague and Stanford University, addressing EV charging insights from true datasets from Europe. Based on the insights, we introduce an innovative approach of using machine learning to simulate electric vehicle (EV) charging profiles in urban areas with limited data, a challenge of predicting EV charging behavior influenced by spatio-temporal factors. We will delve into the neural network architecture used to uncover latent charging profiles, focusing on peak power demand and daily load shapes. We highlight the significant impacts of Basic Administrative Units on predicted load curves, providing insights into optimizing EV charging infrastructure. We discuss how this model can help Distribution System Operators (DSOs) efficiently plan EV charging infrastructure expansion in urban settings, as well as balancing opportunities for the grid.

Bio

Marek Miltner is a Researcher of AI for Sustainable Energy with affiliations between CTU FEE in Prague and Stanford University at the S3L Lab, he is also the founder of LIMESS Lab for Sustainable Energy and Mobility and of AI Tinkerers Prague Chapter. His research focuses on the intertwined role of AI and Power Grids, and looks at specific ways of applying AI methods to accelerate the green transition of power grids, such as through better modelling, forecasting or control.