• A multinomial logit model is developed to analyze drivers’ gas station choices. • Model integrates travel survey data with gas station price and location information. • Average value of time (VOT) estimated at 15. 20/hour, varying across user groups. • Price sensitivity and detour aversion are influenced by income and trip purpose. • Refueling behavior is used to estimate monetary impacts of EV infrastructure. This study investigates drivers’ refueling behavior and the factors influencing their gas station choices. Using regional travel survey data combined with detailed gas station price and location information, we develop a multinomial logit model to estimate how socio-demographic characteristics, trip purposes, price sensitivity, and detour times affect refueling decisions. The model captures variation in the value of time (VOT) across different user groups, with an average estimate of 15. 20 per hour. Results show that gas stations farther from a driver’s route and those with higher prices are less likely to be selected, and that income level and trip purpose significantly impact sensitivity to price and detour time. While refueling and recharging behaviors are fundamentally different, we do not aim to model EV charging behavior. Instead, we apply the estimated VOT from our refueling model solely to approximate the monetary cost of transitioning to EVs under different infrastructure conditions. Our findings suggest that drivers with home charging can save approximately 143. 65 over a vehicle’s lifetime by avoiding gas station trips, whereas those relying on public charging infrastructure may incur costs up to 3, 463. 58 per vehicle due to detours and waiting times. These estimates are not predictions of EV behavior but serve as a benchmark to assess the cost-effectiveness of EV infrastructure planning based on observed refueling patterns.
Mehditabrizi et al. (Sat,) studied this question.