Spatial Markov chain model for electric vehicle charging in cities using geographical information system (GIS) data. (1st December 2018)

• Record Type: Journal Article
• Authors: Shepero, Mahmoud; Munkhammar, Joakim
• Is Part Of: Applied energy. Volume 231(2018)
• Page Start: 1089
• Page End: 1099
• Publication Date: 2018-12-01
• Subjects: EV electric vehicle -- GIS geographical information system -- GPS global positioning system -- KNN K-nearest neighbor -- LiDAR light detection and ranging -- OSM OpenStreetMap -- PV photovoltaics -- RES renewable energy sources -- SOC state of charge; Electric vehicles -- Spatial electric vehicles charging model -- Public charging infrastructure -- Spatial load of electric vehicles; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2018.09.175
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: Geospatial maps are used to estimate the charging load of electric vehicles in cities. Three distinct charging profiles are assumed in the city: Home, Work, and Other. Charging stations belong to a mixture of profiles depending on their nearby buildings. Using 22 kW chargers resulted in a load with a peak of 1.47 kW/electric vehicle. Fast charging causes high variability in the load when many cars start/stop charging. Abstract: In the recent years, the number of electric vehicles (EVs) on the road have been rapidly increasing. Charging this increasing number of EVs is expected to have an impact on the electricity grid especially if high charging powers and opportunistic charging are used. Several models have been proposed to quantify this impact. Multiple papers have observed that the charging stations are used by multiple users during the day. However, this observation was not assumed in any previous model. Moreover, none of the previous models relied on geospatial maps to extract information about the parking lots—where charging stations are installed—and the charging profiles of the potential users of these charging stations. In this paper, a spatial Markov chain model is developed to model the charging load of EVs in cities. The model assumes three distinct charging profiles: Work, Home, and Other. Geospatial maps were used to estimate the charging profile, or mixture of profiles, of the charging stations based on the nearby building types. A case study was made on the city of Uppsala, Sweden—a city with approximately 44, 000 cars. The results of the case study indicated that the aggregate load of the EVs in the city reduced the charging impact. For example when using 22 kW chargers, the peak load in the city per EV was estimated to be 1.29 kW/car in case of spatio-temporally opportunistic charging, and 1.47 kW/car in case of residential only opportunistic charging. This is to say that the Swedish grid operators can expect that every EV in the city will increase the peak load by at most 1.47 kW due to aggregation; this is assuming that 22 kW chargers were used. In addition, we showed that the minute-minute variability of the charging load in cities might cause some future challenges. In our case, up to 3% of the EVs in the city simultaneously started charging. This caused a one-minute-ramp in the charging load of 1.1 MW—if charging using 3.7 kW. Charging with higher powers will exacerbate these ramps, e.g., charging with 22 kW will cause sudden one-minute-increases as high as 6.7 MW in the charging load. Such a finding indicates that using high charging powers might cause high variability in the charging load of EVs in cities. This high variability might limit the synergy potentials between EVs and variable renewable energy sources (RES). The proposed model can potentially be used along with RES models to estimate the spatio-temporal synergy potentials between the two technologies. Evaluating the synergy potentials might be of value to grid operators, policy makers, market traders, etc.