Electric Vehicles to Support Grid Needs: Evidence from a Medium-Sized City

Comi, Antonio; Atumo, Eskindir Ayele; Elnour, Elsiddig · 2026 · Crossref

DOI: 10.3390/vehicles8020030

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the challenge of integrating electric vehicles (EVs) into sustainable energy systems through Vehicle-to-Grid (V2G) technology. While V2G offers significant benefits for grid stability and renewable energy integration, practical implementation is hindered by spatial and temporal uncertainties regarding EV availability and charging demand. Existing research often relies on simulated scenarios or land-use assumptions, lacking empirical methodologies to identify optimal V2G hubs and forecast available energy using real-world vehicle movement data. To bridge this gap, the authors introduce a data-driven approach to identify optimal V2G regions, estimate surplus battery capacity, and forecast grid-transferable energy. The methodology utilizes Floating Car Data (FCD) collected from GPS-enabled vehicles over 58 days in Viterbo, Italy, a medium-sized city. The process begins with spatial k-means clustering to segment the study area and identify candidate zones based on parking proximity. These clusters are joined with administrative census boundaries to correlate parking behavior with land-use variables. The optimal V2G hub is selected based on aggregated parking duration and vehicle density. Surplus energy is calculated by deducting daily mobility needs and a mandatory 20% battery reserve from the total battery capacity, accounting for energy consumed before parking and required for subsequent trips. Finally, the study compares two forecasting models—Autoregressive Integrated Moving Average (ARIMA) and Long Short-Term Memory (LSTM)—to predict next-day energy availability. The empirical findings demonstrate the effectiveness of the proposed framework in quantifying V2G potential. In terms of forecasting performance, the ARIMA model significantly outperformed the LSTM model under the current data conditions. ARIMA achieved a Root Mean Square Error (RMSE) of 52.424, a Mean Absolute Error (MAE) of 36.05, and a Mean Absolute Percentage Error (MAPE) of 12.98%. In contrast, the LSTM model yielded higher error rates, with an RMSE of 99.09, MAE of 80.351, and MAPE of 53.20%. This indicates that, for the specific dataset and temporal resolution used, classical statistical methods provided more accurate predictions than deep learning architectures. The study concludes that V2G technology plays a critical role in supporting the grid needs of medium-sized cities, particularly as EV penetration increases. The results highlight the importance of using real-world FCD and predictive approaches to make informed decisions regarding V2G infrastructure placement and operation. By providing empirical evidence on spatio-temporal uncertainties and validating forecasting techniques, the research offers a robust framework for leveraging EVs as distributed energy resources, thereby enhancing grid efficiency and supporting the transition to sustainable energy systems.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• v2x connected vehicle