AI-Based GEVs Mobility Estimation and Battery Aging Quantification Method

Abstract

AbstractThe bi-directional link between the electrical system and electric vehicles allows vehicle batteries to provide balancing services for the system in a flexible, low-cost, and quick-response manner. However, two critical issues should be solved in realising the benefits of vehicles to grid (V2G) services. Firstly, grid-connected electric vehicle (GEV) mobility may cause uncertainties in the grid’s energy storage capacity, which may further impact the power quality and stability of the power grid. Thus, in V2G scheduling, it is necessary to access electric vehicle (EV) mobility and estimate its schedulable capacity and charging requirements (SC&CR) information in advance. Furthermore, the key factor that keeps GEVs owners from becoming the prosumers of the grid is the battery life loss caused by additional operating cycles in V2G service, as well as the concern about expensive battery deterioration costs. To promote the adoption of V2G services, battery life loss should be evaluated and mitigated through a behaviour management algorithm. This chapter investigates and compares the performance of existing V2G capacity prediction methods, including statistical model, learning-based model, and rolling prediction model. Thereafter, it introduces a life loss quantification model to analyse battery aging characteristics when providing V2G services. With the predicted GEVs mobility information and battery aging cost analysis model, V2G resources can be better utilized by producing more efficient strategies.KeywordsElectric vehicleVehicle-to-gridDeep learningV2G schedulable capacityBattery aging model