Optimal location of EV charging stations in a neighborhood considering a multi-objective approach
Introduction
The electric vehicles (EVs) sales have increased over the years. While this brings multiple environmental and economic benefits, their deployment requires comprehensive studies to optimally plan the location of charging stations (CS) that ensures the EV charging demand will be met. As EV charging duration is still high [1], planning the location of semi-fast CSs to serve a substantial amount of EV users has become a major issue [2]. Therefore, CSs must be installed in strategic locations, where people often spend sufficient time, such as shopping malls, supermarkets and office buildings [3].
The CS location problem has been widely addressed in the literature through the delimitation of CS service zones [4]. The zones can be arranged using different clustering methods, usually divided between hierarchical and non-hierarchical algorithms. Non-hierarchical methods, such as the k-means and fuzzy c-means method, divide the database into distinct groups, while hierarchical methods categorize data in a tree with a hierarchical structure called dendrogram [5]. On the other hand, the utility operator should consider the EV's flexibility to make optimal decisions. This flexibility is affected by the driving behavior and vehicle parking patterns that are inherently stochastic [6]. These uncertainties related to the EV nature could be properly modeled through stochastic simulation. However, a comprehensive simulation can lead to impractical computation time due to the high number of scenarios. Therefore, scenario reduction techniques such as backward probability distance algorithm [7] and clustering methods [8] can be applied to lower the computational burden while preserving the stochastic information from the original scenarios.
Public charging infrastructure can encourage EV users to drive longer distances, as the availability of public CSs reduces range anxiety [9]. Locating public CSs in a neighborhood can also reduce the EV total cost of ownership and increase the EV penetration, since some users do not need to purchase a home charger.
This paper deals with an up-to-date problem that will become even more relevant as EVs sales increases. In mid-term planning, EV penetration is expected to be low, as consumers are starting to buy the region's first EVs. Although, it is expected to increase in long-term planning, as the technology becomes more accessible and widespread among the population. According to the Costa Rica Decarbonization Plan, mid- and long-term planning refers to 2023–2035 and 2035–2050, respectively [10]. Developing countries are expected to experience a similar growth in EV sales, as the price is still a major barrier. So, it is expected that this planning horizon would be similar in other countries.
The main approaches in the literature in this topic consider aspects related to economics, impacts on distribution power systems and EV mobility [11]. However, the problem is often formulated to meet the objectives of the CS operator only. For the power utilities perspective, optimal solutions should consider locations that minimize grid reinforcements cost, while maintaining the grid operation at a standard level [[12], [13], [14]]. For a private agent operator, capturing more clients to maximize the CS profit may be the priority [15, 16].
Although most of the problems associated with CS planning may include different aspects, they are often approached in a single objective model. However, in practice problems in this field of knowledge are better evaluated considering the existing conflicting objectives. For example, authors in [17] minimize the total cost (construction and charging) and maximize coverage. Another key factor for planning CSs is the user's behavior. The uncertainties associated with the time of arrival, departure, and state of charge of the batteries directly affect the size of the CS [18, 19].
Other topics often evaluated in CS planning include multi type chargers [20], integration with the transport network [21] and Vehicle-to-grid (V2G) [22]. A relevant subject that is rarely discussed in this area is the influence of EV penetration in CS position, as addressed in [23]. However, the CS planning problem in [23] is formulated as a mono-objective (maximize power utility profit) and does not consider the uncertainties in EV user's behavior. Thus, to the best of our knowledge, none of the existing works considers CS siting optimization in long term-planning with a multi-objective approach and EV user's uncertainties.
The present work aims to evaluate the impacts caused by the location of semi-fast CSs in a real case example in Costa Rica. A multi-objective approach comprising a metaheuristic algorithm is performed as a possible decision-making tool for the location of semi-fast CSs. The CSs are located in the distribution transformers, which are selected using the Bat Algorithm. The optimization problem considers the constraints of CS service zones, which are delimited by a hierarchical clustering method. The proposed objectives encompass both technical and mobility aspects. The technical aspect includes power losses minimization to reduce costs. The mobility aspect includes consumer preferences represented in this work as the distance that EV users have to reach the closest CS (shorter distances are expected to better meet user's needs).
In specific, this work contributes to the literature in:
- •
Performing an analysis that considers both the distribution power system and the EV's charging demand in a multi-objective approach.
- •
Evaluating the difference in CS location for mid and long-term planning based on the CS occurrence rate, i.e., the number of times a transformer is selected to allocate a CS.
- •
Consider uncertainties related to EV user's behavior.
- •
Presenting a holistic approach where decision-makers can balance between EV users and power utilities conflicting objectives.
The paper is organized as follows: Section II presents a literature review, with typical methods used to create zones, as well as the most used objective functions in CS location problems. Section III presents the methodology as well as the zone construction using a hierarchical clustering method of the zone and the formulation of the problem. Section IV presents the case study, with information about the planning horizons and the distribution network. In section V the results are discussed and the conclusion is developed in section VI.
Section snippets
Methodology
The objective of this work is to optimize the location of CSs in a distribution network considering multiple objectives. The proposed model requires the following inputs: definition of EV penetration levels; distance among EV clients; CS load profiles generated considering user's behavior uncertainties; and detailed GIS grid data. Then, hierarchical clustering is used to create the CS's zones, while a scenario reduction technique is used to lower the computational time by decreasing the number
Case study
The proposed case study comprises four simulation planning horizons with different number of CS zones and EV penetration rates, as indicated in Table II. The EV penetration level is defined based on the number of houses that purchase an EV. The EV penetration rate is directly related to the distance among customers with EV, so that a rise in the penetration rate leads to a reduction in this distance and, consequently, an increase in the number of CSs. Planning horizons 1 and 2 fit into mid-term
Results
The planning horizons are evaluated according to the electrical losses and . For all planning horizons the following values of the Bat Algorithm parameters have been considered: sound amplitude decreasing rate , pulse emission increasing rate , and population . and varies in steps of 1/10, in a total of 10 iterations of Bat Algorithm. The stopping criteria used is the number of iterations, considered as 4, since the algorithm presents a good convergence [31]. For
Conclusion
This work has proposed a methodology to optimally locate CSs through a multi-objective approach, by minimizing the electrical losses and the charging zone center deviation. The CS service zones are defined using a hierarchical clustering method based on EV owners’ geographical position. Four planning horizons are proposed, such that EV penetration is associated with the number of CSs. The results are arranged through a set of Pareto Frontiers, which displays a range of solutions for the
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) - Finance Code 001, National Council for Scientific and Technological Development (CNPq), Fundação de Amparo à Pesquisa no estado de Minas Gerais (FAPEMIG) and INERGE. We thank Compañia Nacional de Fuerza y Luz, in Costa Rica, for providing the data for this research as part of its collaboration with EPERLab of the University of Costa Rica.
References (33)
- et al.
Probabilistic planning of electric vehicles charging stations in an integrated electricity-transport system
Electr. Power Syst. Res.
(2020) - et al.
Urban public charging station locating method for electric vehicles based on land use approach
J. Transp. Geogr.
(2019) - et al.
A review of spatial localization methodologies for the electric vehicle charging infrastructure
Int. J. Sustain. Transp.
(2019) - et al.
Optimization models for an EV aggregator selling secondary reserve in the electricity market
Electric Power Syst. Res.
(2014) - et al.
A risk-constrained decision support tool for EV aggregators participating in energy and frequency regulation markets
Electric Power Syst. Res.
(2020) - et al.
Classification of electric vehicle charging time series with selective clustering
Electric Power Syst. Res.
(2020) - et al.
Development of a Bayesian network model for optimal site selection of electric vehicle charging station
Int. J. Electr. Power Energy Syst.
(2019) Optimal planning of electric vehicle charging stations comprising multi-types of charging facilities
Appl. Energy
(2018)- et al.
Optimal location planning method of fast charging station for electric vehicles considering operators, drivers, vehicles, traffic flow and power grid
Energy
(2019) - et al.
GIS-based multi-objective particle swarm optimization of charging stations for electric vehicles
Energy
(2019)
Capacity planning and pricing design of charging station considering the uncertainty of user behavior
Int. J. Electr. Power Energy Syst.
User behaviour and electric vehicle charging infrastructure: an agent-based model assessment
Appl. Energy
Coordinated allocation of distributed generation resources and electric vehicle charging stations in distribution systems with vehicle-to-grid interaction
Energy
Optimal siting and sizing of distribution system operator owned EV parking lots
Appl. Energy
Decarbonising the transport and energy sectors: technical feasibility and socioeconomic impacts in Costa Rica
Energy Strat. Rev.
Placement and sizing of battery energy storage for primary frequency control in an isolated section of the Mexican power system
Electric Power Syst. Res.
Cited by (41)
A reliable optimal electric Vehicle charging stations allocation
2024, Ain Shams Engineering JournalElectric vehicle charging service operations: A review of machine learning applications for infrastructure planning, control, pricing and routing
2023, Renewable and Sustainable Energy ReviewsIdentifying optimal locations for community electric vehicle charging
2023, Sustainable Cities and Society