Kolloquium: Joint Optimization of Electric Vehicle Charging Station Placement and Charge Scheduling in a Future Energy System

As renewable energy sources like solar and wind hold the potential to reduce CO₂ emissions from electricity generation, their volatility presents challenges for the charging infrastructure of battery electric vehicles (BEVs). Our study introduces a mixed-integer linear programming model that simultaneously optimizes the placement of charging stations and the scheduling of vehicle charging, taking into account mobility data, electricity prices, and emissions.

While the placement of charging stations and the scheduling of EV charging have often been treated as separate decision problems, the Charging Station Placement and Electric Vehicle Charge Scheduling Problem (CSP-EVCSP) addresses these two challenges within a unified model. This approach sheds light on trade-offs between the timing of charging processes and the strategic placement of charging infrastructure. Due to computational limitations in large-scale scenarios, we propose a hybrid approach that combines the Set Covering Problem (SCP) with a Charge-Level Heuristic and the Electric Vehicle Charge Scheduling Problem (EVCSP).

Our focus is specifically on a greenfield expansion scenario, where the variability in renewable energy input leads to fluctuations in electricity prices. To address this, charging processes are evaluated over a representative period, ideally covering a full year, with a sufficiently large sample of vehicles, each with individual mobility profiles.

Given the lack of detailed public data, we generate synthetic mobility patterns based on empirical data, distinguishing 231 different activities of individuals throughout the day with a time resolution of 10 minutes. These daily vehicle movement profiles are mapped to geographic locations, which can be obtained from OpenStreetMap (OSM) for many cities and regions, creating geographically and temporally resolved driving patterns.

Electricity prices fluctuate significantly even within a single day, driven by demand levels and available supply. To capture these dynamics, we use the fundamental electricity price model ParFuM (Parsimonious Fundamental Model), which accurately reflects key features of future energy systems with high shares of renewable energy. We fit this model to publicly available time series data, demonstrating that actual price developments can be replicated with reasonable accuracy.

We apply our model to a case study in Essen, Germany, and investigate three scenarios for the year 2030. The results demonstrate that controlled charging strategies can significantly reduce costs compared to uncontrolled charging. Our findings offer valuable insights for municipalities, charging infrastructure operators, and policymakers, supporting the cost-effective and sustainable integration of BEVs into the electricity grid.