Energy Efficient Framework to Find Optimized Route for EVs Movement

Abstract

There has been tremendous increase in the use of renewable resources in 21st century which is essential to save the environment from the hazardous impact of non-renewable resources. It is projected that India emits approximately three gigatonnes of greenhouse gases (GHGs) per year, accounting for around 7% of global emissions [1]. It is also pertinent to note that road vehicles are responsible for 76% of the total CO2 emitted by the transportation sector [2]. Therefore, the governments are taking new initiatives all over the world to reduce GHGs production and cut down their reliance on fossil fuels. The evolution of non-fossil fuel-based vehicles or alternative fuel vehicles (AFVs) became a prominent choice for ensuring environment-friendly and long-term transportation sustainability. Thus, AFVs, especially electric vehicles (EVs), are now widely recognized as one of the most effective ways to alleviate GHG emissions, technological developments, and world-wide government incentives. Despite the profound significance of employing EVs, the factors such as limited range of these vehicles, underdeveloped charging infrastructure, and ease of charging services pose a barrier to the mass adoption of EVs. Indeed, the EVs’ charging times are substantially large due to CSs’ equipment and the composition of the batteries, resulting in increased wait times in CSs. Therefore, we need to have a system that can efficiently manage all the available resources of EV, transportation network, EV and CS. Another challenge lies in the fact that EVs are likely to meet a large number of transportation demands in the near future, complicating the decision-making process due to the coupling of routing and charging simultaneously, which makes it harder to solve such problems. Recent scientific contributions in joint routing and charging have been mainly classified into four domains: i) heuristic approaches; ii) commercial solvers like CPLEX, Gurobi, etc.; iii) machine learning-based approaches; and iv) problem-tailored based on particular structural information. However, each of them has its own set of constraints in terms of computation resources and computation time. None of them can guarantee the solution’s optimality. The problem based on structured information of combined routing and charging of EVs has not been thoroughly explored and needs further exploration with some real-time metrics such as battery SoC, traffic condition, state of charging station, ToU energy pricing, etc. In this research work, a modest attempt has been made to find the much-needed solution to the problem of energy-efficient EV routing. Firstly, the introduced problem has been addressed by inculcating the effect of various road surface conditions (icy, dry, wet, and snowy) and providing the solution by incorporating the principles of Artificial Bee Colony (ABC). Secondly, this problem has been extended by identifying the suitable charging station (CS), keeping in mind the objectives of minimal waiting time and charging cost at the iii CS. The K-shortest path (KSP) algorithm has been used to make the route and charging planning more effective. Lastly, this dissertation extends the mathematical model of vehicle’s energy consumption estimation with the features of vehicle’s start/stop energy expenditure and recapturing energy effect and it also introduces an Amplified-ACO (𝐴2𝐶𝑂)(Amplified-Ant Colony Optimization), routing algorithm based on ACO principles that makes use of the probabilistic selection model, to efficiently solve the underlined problem. Keywords: Energy-efficient EV routing, CS identification, Regenerative braking, Heuristic approaches, Non-identical road surfaces, Improvised distributed system