Modelling electric vehicle charging events using a random parameter duration approach

To promote the uptake of Electric vehicles (EVs), public charge points (or infrastructures) are crucial. A critical question for the planning and operational management of EV infrastructures is understanding different dimensions of temporal utilisation, which assists in improving the existing infrastructure and demonstrating the need for new infrastructures. Therefore, this paper analyses charge event durations of EVs using data from Transport for Greater Manchester. A hazard-based duration modelling approach is adopted to model the duration of charge events as a function of temporal, geographic, and prevailing weather conditions. The model reveals the charge events during the spring season can be longer and shorter, which can be attributed to temperature fluctuations requiring more higher/smaller energy consumption, and similarly, with an increase in windspeed, charge duration may increase/decrease. These findings can be useful for infrastructure operators in determining the factors that affect user behaviour on their EV charging networks and could feed into the development of commercial strategies to control EV charging networks, such as how to spread charge demand over different periods of the day and how to discourage the occurrence of overstay. For instance, charge events with longer connection durations could be identified and targeted by infrastructure providers to delay their occurrence to periods when power demand is not constrained through such means as providing preferential prices during off-peak periods.<p></p>