Validation and application of agent-based electric vehicle charging model

Agent-based models are a class of simulation in which many autonomous agents interact such that the mix of stochastic and deterministic actions each agent undertakes results in some emergent behaviour across the entire population. Such models have been used to explore the impacts of electric vehicles (EVs) on electricity grids. However, there has been little data available against which to validate this approach. In this study, we evaluate an agent based EV model against real data observed during the “My Electric Avenue” project; an Ofgem funded 3 year trial aiming to identify the impacts of EVs on local grids. We find that, within the constraints of the available trial data, the agent model is able to replicate dominant charging pattern features. The behaviour of owners will inevitably play a role in the actual charging patterns observed and we further explore how consumer adoption of time-of-use tariffs and vehicle range (battery capacity) preference would impact on demands at the local substation. We show that simplistic adoption of time-of-use tariffs would have undesirable consequences for network peak demands and that the expected increase in EV range and thus battery size, will both increase peaks and total energy supply to domestic properties.