Multi-period planning for electric car charging station locations: A case of Korean Expressways

Highlights

• We propose three methods for multi-period planning to locate electric charging stations.

• We compare the three methods with the Korean Expressway data.

• Excluding short/low-demand paths make problems tractable without losing coverage.

• Multi-period location decisions can be significantly different from single-period decisions.

• We consider five types of demand profile for additional insights.

Abstract

One of the most critical barriers to widespread adoption of electric cars is the lack of charging station infrastructure. Although it is expected that a sufficient number of charging stations will be constructed eventually, due to various practical reasons they may have to be introduced gradually over time. In this paper, we formulate a multi-period optimization model based on a flow-refueling location model for strategic charging station location planning. We also propose two myopic methods and develop a case study based on the real traffic flow data of the Korean Expressway network in 2011. We discuss the performance of the three proposed methods.

Introduction

The use of electric passenger cars has gained considerable attention over the last decade as an environmentally friendly alternative to conventional cars that consume fossil fuels and emit greenhouse gases. There are, however, several barriers for electric cars to become more popular. The first and foremost barrier is the limited range that electric cars can be driven without recharging (Capar, Kuby, 2012, Lim, Kuby, 2010, Romm, 2006, Wang, Wang, 2010). Most electric cars available in 2014 have ranges, from a fully charged battery, of about 60 kilometers to 160 kilometers depending on various factors such as weather conditions, traffic congestion, and road types. Such driving ranges may be insufficient for electric cars to be used as a primary transportation mode. Another critical barrier is the lack of charging station infrastructure (Kuby, Lim, 2007, Melaina, Bremson, 2008, Ogden, 1999, Shukla, Pekny, Venkatasubramanian, 2011, Wang, 2011). Since it would be difficult to increase the driving range of electric cars dramatically within the next few years, it is particularly important to have a well-planned charging station infrastructure. The goal of this paper is to help establish a multi-period strategic plan to build charging stations to maximize the total traffic flows covered.

There are a sizable number of papers in the literature that study and address the limited infrastructure issue and optimal locations for refueling or charging stations. Kuby and Lim (2005) suggest a flow-refueling location model (FRLM) to help find optimal refueling station locations for alternative-fuel vehicles (AFVs) that are powered by hydrogen, biofuels, or natural gas. The FRLM is based on the flow-intercepting model, proposed by Hodgson (1990) and Berman et al. (1992), that maximizes the total traffic flow passing a given number of facilities such as service stations. The FRLM extends the flow-intercepting model and incorporates the requirement that AFVs need multiple refueling stations, rather than a single refueling station, for long trips. Lim and Kuby (2010) propose several heuristic methods to solve the FRLM. Wang and Lin (2009),Capar and Kuby (2012), and Capar et al. (2013) provide alternative formulations that are numerically more efficient. MirHassani and Ebrazi (2013) propose a network expansion method to improve the computability of the FRLM, which is the base of our proposed model.

While the FRLM is suggested for refueling station location problems for AFVs that usually require short refueling time, it can also be applied to electric vehicle charging station location problems under mild assumptions (Capar et al., 2013). Since the FRLM assumes drivers will stop at charging stations on the way to the final destination to gain additional driving range, it is apparent that Level 1 or 2 charging technologies, for which drivers need to wait 2–8 hours to fully charge their vehicles, are inappropriate. Therefore, we assume in our model that Level 3 fast charging or battery swapping technologies are used with about 20-minute long waiting time and that charging stations are uncapacitated, under which the FRLM and its variants would provide meaningful results.

Unlike other studies, our paper focuses on a multi-period optimal construction plan, since it may not be practical to build a sufficient number of stations within a short period of time due to, for example, the limited budget. Indeed, the authority responsible for building such infrastructure will not invest until there are enough electric cars to use the infrastructure. On the other hand, the potential consumers will be less inclined to buy electric cars unless there is sufficient charging station infrastructure (Bento, 2008); a so-called chicken-and-egg problem arises (Kuby, Lim, 2005, Leiby, Rubin, 2004, Melaina, 2007, Wang, Wang, 2010). We note that a market-driven approach may not resolve such an issue; thus, a strategic infrastructure plan controlled by a central authority is needed. In that vein, strategic multi-period planning is required to find a first stage construction plan, followed by next stage construction plans, and thereby to provide an overall plan over the planning horizon. In this paper, we propose three methods: a multi-period optimization method, a forward-myopic method, and a backward-myopic method. In our case study employing the real data of the Korean Expressway network, we show and discuss the results of the three proposed methods.

In traditional facility location problems that consider optimal initial, intertemporal, or terminal (re)locations of facilities, a multi-period scheme has been studied extensively since the seminal work of Wesolowsky (1973). Among others, it is worth mentioning Drezner (1995) for a dynamic p-median problem, Contreras et al. (2011) for a multi-period uncapacitated hub problem, and Albareda-Sambola et al. (2009) for a multi-period service facility location problem. However, to the best of our knowledge, our paper is the first to consider a multi-period refueling/recharging station location problem for alternative-fuel vehicles including electric vehicles. Miralinaghi (2012) considers multi-period travel demands, but not multi-period locational decisions.

In our case study, we apply the proposed methods to the Korean Expressway, which is mainly operated by Korea Expressway Corporation (KEC). KEC is a government-owned company, which determines locations of rest areas, facility types and sizes. When KEC would plan for charging stations, it would cooperate with Korea Electric Power Corporation (KEPCO) that is also government-owned. This makes a market-based approach for the Korean Expressway network much more unrealistic in addition to the fact that it is unsuitable for a charging station infrastructure problem in general due to the chicken-and-egg problem mentioned above. This paper considers methods for central planning.

Our contributions are summarized as follows: (1) we propose three methods to help construct a multi-period plan for charging station infrastructure; (2) we perform an extensive numerical case study with the real Korean Expressway data to compare the three proposed methods; (3) to further investigate the differences among the three proposed methods, we perform another numerical study using five different demand profiles; (4) we show that multi-period location decisions from the three methods can be significantly different; and (5) we show in our case study that excluding short-distance and low-demand paths makes the problem solvable with a standard optimization solver within a reasonable time without losing coverage.

The remainder of this paper is organized as follows. In Section 2, we discuss the formulation of the single-period FRLM. In Section 3, we introduce the three methods for multi-period planning. We describe the Korean Expressway network in Section 4, and explain how we collected, organized, summarized, and manipulated the network topology traffic volume raw data. We also provide descriptive statistics that are helpful to understand the Korean Expressway traffic pattern. In Section 5, we report extensive computational results under a variety of scenarios and describe insights gained. We conclude this paper in Section 6 with some remarks on future research directions.