By 2030, the number of electric vehicles (EVs) on the road is expected to increase to 11 million in the UK, meaning that there will be an increase in electricity demand. A potential solution to help manage this increase in demand is to use a technology called vehicle-to-grid (V2G) which is essentially a connection post that allows a bidirectional flow of energy, which means that EVs can charge and discharge when connected. Through this technology, the electrical grid can make use of the energy already stored in the battery of the EV.
This research aimed to understand the effects of EV availability on V2G technology within a microgrid and evaluated the feasibility of providing ancillary services. A predictive model, primarily trained on internal combustion engine vehicle (ICEV) trips, used the UK’s historical travel data to predict the location of EVs, achieving significant understanding of travel behaviour and EV availability. Split into two tasks—predicting start and end locations—this model utilised light gradient boosting machine (LightGBM) due to its superior performance. After fine-tuning, it yielded a weighted average F1 score of 0.900 and 0.902 for tasks 1 and 2, respectively. The model, when informed by new, real-world EV data, derived travel start and end locations, which was the fed into an optimisation model.
This optimisation model use a mixed integer linear programming (MILP) approach to schedule EV battery usage at the household level and study various case studies involving V2G technology. Simulations factored in different photovoltaic (PV) penetration rates, energy tariffs, and peer-to-peer (P2P) pricing mechanisms within a microgrid. First, the technical and economic benefits of home batteries, smart charging (V1G), and Vehicle-to-home (V2H) systems in EVs were evaluated, with an emphasis on performance and electricity bill reduction. The second case studied the potential of EVs to provide short term operation reserve (STOR) services. The third case explored a payment mechanism to optimise the state of charge (SOC) for EVs under V1G and V2H technologies for a week and estimate the energy available for restoration services.
The study reveals that both stationary home batteries and EVs, when integrated with solar power and dynamic tariffs, can effectively reduce electricity costs, despite the fluctuating availability of EVs. Notably, EVs, when combined with P2P energy sharing and V2H systems, offer comparable performance to stationary batteries, in addition to their transportation benefit. In terms of STOR provision, EVs meet the technical requirements, with their availability significantly influencing STOR provision. Factors like energy tariffs, solar power penetration rates, and P2P mechanisms have minimal effect on the STOR energy amount, but they do affect the overall microgrid performance. The study also highlights the need to maintain a 15% surplus of EVs within the microgrid for ensured resilience. Effective strategies to maintain a high SOC in EVs include higher payment rate systems, implementation of V1G and V2H strategies, and dynamic energy tariffs. The study, however, recommends limiting users to V1G to prioritise potential energy use for restoration services. Although EV availability affects the minimum SOC, it is not more significant than other factors such as EV penetration rates, energy tariffs, and P2P price mechanisms.
The findings imply that EV availability can reduce some of the benefits that stationary home battery have, such as surplus noon charging, while V2H might match home batteries in certain situations. EVs can offer STOR services as the fulfil most of the technical requirements, but the energy amount is dependant on available EVs during STOR events. EV availability had minimal effect on maintaining minimum SOC for a week that could potentially be used for restoration services, with energy tariffs and end-of-week incentives being more influential. Different PV penetration rates, energy tariffs, and P2P price mechanisms each have varied impacts on grid performance and V2G provision depending on the scenario.
