Electrical Vehicle Fleet Management for Industrial Environment with Battery SoH prediction through Neural Networks

The design of the grid architecture for electric vehicle fleets in industrial environments (e.g., battery-powered forklifts) requires considering different parameters. The increased use of renewable energy sources and the need for efficient energy use, leads the designers to make considerations about the best size of the renewable energy system, energy storage system capacity, number of vehicles and their autonomy. Artificial intelligence algorithms could be used and make crucial changes to the design approach. In this work, we present a novel approach to define the most suitable grid architecture. Behavioral Matlab models, validated through tests carried out on a reduced-scale system, together with artificial intelligence algorithms for the battery state-of-health are used to determine the number of vehicles, initial investment cost, power grid consumption costs, CO2 footprint, and vehicle working time specifications. The optimum sizing of the system has been defined considering economic, technological, or environmental aspects.