This study aims to develop a demand-responsive planning framework for EV charging infrastructure using empirical data from Politecnico di Milano campuses. First, a two-year dataset was analyzed to identify behavioral patterns and energy usage trends. The results revealed a strong annual growth in charging demand, with projections estimating over 34,000 charging sessions by 2030—a ten- fold increase from current levels. To address this, two infrastructure configurations were proposed. The basic configuration assumes each socket can serve two sessions per working day and provides a practical baseline for deployment. However, simu- lations show it may result in over 20% unsatisfied demand during peak hours. The advanced configuration, on the other hand, incorporates time-dependent demand and enforces a design constraint that limits unsatisfied demand below 20% at any time. Minimal socket additions were found sufficient to meet this requirement. Furthermore, user classification based on parking duration and energy consump- tion enabled a more targeted allocation of high- and low-power sockets, improving efficiency. These findings support a scalable methodology that integrates data- driven forecasting and behavioral analysis to inform EV infrastructure planning. The proposed approach can be adapted to various institutions and urban contexts expecting rapid growth in electrified transport.
Questo studio sviluppa un quadro di pianificazione dell'infrastruttura di ricarica per veicoli elettrici (EV) basato sulla domanda, utilizzando dati empirici raccolti nei campus del Politecnico di Milano. L'analisi di un dataset biennale ha evidenziato schemi comportamentali e tendenze di consumo energetico, con una crescita significativa della domanda di ricarica. Le proiezioni stimano oltre 34.000 sessioni di ricarica entro il 2030, pari a un incremento di dieci volte rispetto ai livelli attuali. Per far fronte a tale crescita, sono state proposte due configurazioni infrastrutturali. La configurazione di base, che presume due sessioni giornaliere per presa, offre una stima minima delle necessità infrastrutturali, ma può generare oltre il 20% di domanda insoddisfatta nelle ore di punta. Al contrario, la configurazione avanzata considera la variabilità temporale della domanda e mantiene la domanda insoddisfatta sempre al di sotto del 20%, con un'aggiunta minima di prese. La classificazione degli utenti in base alla durata della sosta e al consumo energetico consente una distribuzione ottimizzata tra prese ad alta e bassa potenza, migliorando l’efficienza complessiva. L'approccio proposto rappresenta una metodologia scalabile e adattabile a diversi contesti accademici e urbani, per supportare la pianificazione di infrastrutture di ricarica in scenari di crescente elettrificazione del trasporto.
Smart planning of EV charging stations via demand forecasting and user behavior analysis : a multi-campus University case study
KUMAZAWA, KOUSEI
2024/2025
Abstract
This study aims to develop a demand-responsive planning framework for EV charging infrastructure using empirical data from Politecnico di Milano campuses. First, a two-year dataset was analyzed to identify behavioral patterns and energy usage trends. The results revealed a strong annual growth in charging demand, with projections estimating over 34,000 charging sessions by 2030—a ten- fold increase from current levels. To address this, two infrastructure configurations were proposed. The basic configuration assumes each socket can serve two sessions per working day and provides a practical baseline for deployment. However, simu- lations show it may result in over 20% unsatisfied demand during peak hours. The advanced configuration, on the other hand, incorporates time-dependent demand and enforces a design constraint that limits unsatisfied demand below 20% at any time. Minimal socket additions were found sufficient to meet this requirement. Furthermore, user classification based on parking duration and energy consump- tion enabled a more targeted allocation of high- and low-power sockets, improving efficiency. These findings support a scalable methodology that integrates data- driven forecasting and behavioral analysis to inform EV infrastructure planning. The proposed approach can be adapted to various institutions and urban contexts expecting rapid growth in electrified transport.File | Dimensione | Formato | |
---|---|---|---|
2025_07_Kumazawa_Thesis.pdf
accessibile in internet per tutti
Descrizione: Article Format Thesis
Dimensione
3.01 MB
Formato
Adobe PDF
|
3.01 MB | Adobe PDF | Visualizza/Apri |
2025_07_Kumazawa_Executive Summary.pdf
accessibile in internet per tutti
Descrizione: Executive Summary
Dimensione
1.4 MB
Formato
Adobe PDF
|
1.4 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/239557