Simulation and artificial intelligence holistic approach for nationwide charging station deployment to solve electric vehicle range anxiety
The climate crisis caused by greenhouse gas emissions is the most important problem that humanity is currently facing on a global scale. One of the largest contributors to global greenhouse gas emissions is transport, which in Spain is the sector with the highest contribution: 27.7% of total emissions in terms of equivalent CO2 in 2020. Therefore, transport must be decarbonised, making a transition to electric vehicles. However, range anxiety, a fear of running out of electricity before reaching another available charging station, is one of the biggest barriers (probably the most important one) in the widespread adoption of electric vehicles, in combination with a very limited availability of charging stations (CSs). In order to solve the range anxiety problem, it is essential to build an adequate network of highway CSs at a national level. A key challenge is to determine their optimal location, as this is dependent on the interconnection between different factors.
In the SAINEVRA project we will build a framework that integrates simulation (digital twin) and artificial intelligence (AI) techniques to optimize the deployment and management of a nationwide highway network of charging stations, to solve the range anxiety problem and foster long range electric vehicle intercity trips.
It will take into account traffic behavior from the main national highways, and the layout of the national high voltage electric grid, using data from Spain, to find an optimal deployment of CSs that simultaneously optimize the following set of relevant objectives: the electric vehicles long-range travel times between cities considering dynamical traffic congestion, the distance deviation from the optimal trip to recharge, the waiting time in the CSs, the building cost of new CSs by considering the distance to current electric grid substations, and the global highway traffic carbon footprint.
The project will employ a combination of disruptive digital technologies, and will explore and compare different methodologies as building blocks: simulation techniques to form the basis of a digital tool with a holistic approach using different models: hybrid cellular automata – agent based, mesoscopic, and discrete event system specification models; different AI techniques to optimize the location of CSs: genetic algorithm, simulated annealing, deep reinforcement learning; and deterministic optimization algorithms. In addition, we will use high performance computing techniques to speed-up the execution.
PI: José Luis Guisado Lizar / Fernando Díaz del Río
Type: Plan Estatal 2021-2023 – Proyectos Investigación Orientada
Reference: PID2023-151065OB-I00
Funding by: Ministerio de Ciencia, Innovación y Universidades
Start date: 2024
End date: 2027
Web Page : https://grupo.us.es/sainevra/index_eng.html