Canada's Electric Highway Network

Electric vehicles (EV) are seen as a sustainable solution for reducing greenhouse gases from the transportation sector, however the lack of fast, Level 3 charging infrastructure is a major barrier to EV adoption. This project aimed to appropriately place coast-to-coast EV charging stations in Canada to fulfill the requirements of Natural Resources Canada’s (NRCan) Electric Vehicle and Alternative Fuel Infrastructure Deployment Initiative (EVAFIDI). Utilizing Closest Facilities and Origin-Destination Cost Matrix solvers available through ArcMap and ArcGIS Pro Network Analyst extension, 52 new locations for Level 3 EV charging stations are proposed with a focus on mid-west and northern Canada, as well as Newfoundland and Labrador. Stations were placed at existing gasoline stations using Origin-Destination Cost Matrix solver and are at most 200 km apart along the cross-Canada main corridor and are no more than 275 km apart in remote locations to accommodate currently available EV range capacities. The installation of fast charging stations will reduce range anxiety felt by consumers and further incentivize the wide-spread use of EVs, allowing Canada to better meet climate change commitments.

Range anxiety – the fear of not being able to recharge an EV due to a lack of infrastructure (Table 1) – is widely accepted as one of the major barriers to EV adoption [1,2]. The installation of fast charging stations is suggested to reduce range anxiety and further incentivize the wide-spread use of EVs [3]. As part of Canada’s commitment to reduce GHG emissions through the Pan-Canadian Framework on Clean Growth and Climate Change, the EVAFIDI program provides funding for the construction of new EV charging stations that are located in Canada, publicly available and not restricted to car make in an attempt to put more low-emission vehicles on the road [4]. While there is currently a network of EV charging stations across Canada, it has limited availability of fast, Level 3 DC charging infrastructure beyond central and western Canada. 

Objectives

Various mathematical model approaches are often used for infrastructure location problems [5], including variations on the use of origin-destination (O-D) models [6]. Here I used Closest Facilities and Origin-Destination Cost Matrix models to expand the current Level 3 DC charging infrastructure in Canada (Figure 1). Charging infrastructure deployment will focus on frequently travelled routes [7], be located at gasoline stations [6,8], and consider existing and planned EV infrastructure [4] and EV range capacity. Average range capacities of EVs on the market vary between 200 – 250 km, however newer models can drive up to 500 km on a full charge [9,10]. Therefore, distances of 200 km for the heavily populated southern main corridor and 275 km for remote areas are used to accommodate current range capacities on the market. 

See Table 1 for relevant data sources and their associated descriptions, purpose, and data quality characteristics.

Software Applications

Relevant software applications used includes ArcMap version 10.7 [11] and ArcGIS Pro version 2.4 [12].

Coordinate System

A custom projected coordinate system was created for Canada using North America Equidistant Conic to minimize distortion of distance (see metadata for further details).

Geocoding Available/Planned EV Station Locations

Available and Petro-Canada EV station data were cross-referenced to remove duplicates and subsequently merged, herein called available EV stations. Available (n=688) and planned (n=19) EV stations were geocoded to spatially display their address locations. 

Preparing Road Data Set for Network Analysis

Using Model Builder, selected roads were intersected with each province to be dissolved, merged, separated at their intersections, and placed into a new feature dataset for topology analysis (Figure 2). Unconnected, “dangling” roads were corrected using the topology and editor toolbars and verified with Esri World Topographic Base Map [13].

A network dataset was built using road length as the impedance value to determine distances between locations along the road network.

Closest Facility Analysis: Identifying Routes and Possible Locations for New EV Station Placement

To identify frequently use routes for intercity travel, major cities were chosen for use in Closest Facility solver using a search tolerance of 652 m (determined from spatial join) to create a route network.

Available/planned EV stations and gasoline stations along my routes network were "snapped" using the Near tool and extracted with a 100 m buffer along the routes layer. Based on previous driver behaviour studies [7,14], I assumed long-distance drivers were willing to detour up to 5 km to recharge. Using Closest Facility, stations ≤ 5 km from the routes network were selected.

Origin-Destination Cost Matrix Analysis: Choosing New EV Station Locations

Initially, new charging stations were placed at gasoline stations in cities without charging stations already present. Two Origin-Destination (O-D) Cost Matrices were created with a cut-off of 200 km: between charging stations, as well as between charging stations and gasoline stations. This identified all stations that connected to each other within 200 km to help identify new candidate charging station locations. Where there were no gasoline stations to satiate a 200 km gap, the cut-off was extended to 275 km. New charging stations were placed at gas stations where multiple charging stations could reach it (often at intersections) and closest to major highways within settlements when possible. To maintain spatial accuracy, charging stations were placed on the original gasoline station location and not the "snapped" location.

Along the routes network, 457 available and planned EV stations were extracted and used in O-D Cost Matrix analyses (Figure 3). This resulted in 52 proposed locations for Level 3 DC charging stations with a focus on mid-west and northern Canada, as well as Newfoundland and Labrador (Table 3). Charging stations are at most 200 km apart along the cross-Canada main corridor and are no more than 275 km apart in remote locations. The average distance between charging stations is 87 km with a standard deviation of 60 km. To visualize the entire network expansion, O-D Cost Matrix analyses with all 668 available and 19 planned Level 3 charging stations was done (Figure 5).

Issues with the geocoding of EV stations (Table 2) may have impacted the locations of proposed EV stations. Although this is not a significant issue since 99% of addresses were correctly matched, it does cause some uncertainty for spatial accuracy. For future studies, I would recommend using actual count traffic data to choose frequently travelled roads for analysis. This has been done previously and would better represent intercity driving behaviour [7,8]. 

This project aimed to expand the current fast charging station network in Canada by appropriately placing 52 proposed Level 3 DC stations across Canada. The locations chosen will help reduce range anxiety felt by consumers and further incentivize the wide-spread use of EVs. This expanded network was created to meet NRCan’s EVAFIDI program requirements to help assist Canada to better meet climate change commitments.

Click  here  for metadata.

[1] Skippon S, Kinnear N, Lloyd L, Stannard J. 2016. How experience of use influences mass-market drivers’ willingness to consider a battery electric vehicle: a randomised controlled trial. Transp. Res. A Policy Pract. 92: 26-42.

[2] Melliger M, Van Vliet O, Liimatainen H. 2018. Anxiety vs reality - sufficiency of battery electric vehicle range in Switzerland and Finland. Transp. Res. Part D Transp. Environ. 65: 101-115.

[3] Christensen L, Nørrelund A, Olsen A. 2010. Travel behavior of potential electric vehicle drivers. The Need for Charging. European Transport Conference, Glasgow Scotland.

[4] Natural Resources Canada. 2019. Joint audit and evaluation – The electric vehicle and alternative fuel infrastructure deployment initiative. Canada: The Government of Canada; [accessed 2020 March 10]. https://www.nrcan.gc.ca/nrcan/transparency/reporting-accountability/plans-performance-reports/audit-evaluation/reports-year/reports-2019/joint-audit-evaluation-electric-vehicle-alternative-fuel-infrastructure-deployment

[5] Dong J, Liu C, Lin Z. 2013. Charging infrastructure planning for promoting battery electric vehicles: an activity-based approach using multiday travel data. Transp. Res. Part C. 38: 44-45

[6] Chung S, Kwon C. 2014. Multi-period planning for electric car charging station locations: a case of Korean Expressways. Euro. Journ. Opera. Res. 242: 677-687.

[7] Kelley S, Kuby M. 2013. On the way or around the corner? Observed refueling choices of alternative-fuel drivers in Southern California. Journ. Transp. Geog. 33: 258-267.

[8] Sun X, Yamamoto T, Morikawa T. 2016. Fast-charging station choice behavior among battery electric vehicle users. Transp. Res. Part C. 46: 26-39.

[9] Nykvist B, Sprei F, Nilsson M. 2019. Assessing the progress toward lower priced long range battery electric vehicles. Energy Policy. 124: 144-155.

[10] Canadian Automotive Association. 2020. Canadian Automotive Association. Types of electric vehicles; [accessed 2020 March 8]. https://www.caa.ca/electric-vehicles/types-of-electric-vehicles/.

[11] [Esri] Environmental Systems Research Institute. 2018. ArcGIS Desktop: Release 10.7. Redlands, CA.

[12] [Esri] Environmental Systems Research Institute. 2018. ArcPRO: Release 2.4. Redlands, CA.

[13] [Esri] Environmental Systems Research Institute. 2012 Feb 19. "Topographic" [basemap]. Scale Not Given. "World Topographic Map". [accessed 2020 March 25]. http://www.arcgis.com/home/item.html?id=30e5fe3149c34df1ba922e6f5bbf808f

[14] Lines L, Kuby M, Schultz R, Clancy J, Xie Z. 2008. A rental car strategy for commercialization of hydrogen in Florida. Int. Journ. Hydrog. Energy. 33: 5312-5325.