Estimation of Photovoltaic Potential of Solar-Powered Electric Vehicle: Case Study of Commuters on Donghae Expressway, Korea

Abstract

Studies on solar electric vehicles (EVs) have focused on calculating the power generation in a specific environment without discussing its practical utility. To expand the awareness of the utility of solar EVs, their potential should be evaluated by considering the operation methods of users. This study investigated the photovoltaic (PV) potential of an EV integrated with PV modules while driving on an expressway. Tunnel and shadow areas were identified to determine unpowered areas on the expressway. The PVWatts model was used to evaluate the PV potential by the time of the year. For a single vehicle traveling at 60 km/h on the Donghae expressway section during both the summer and winter solstices, the amount of power generation is within 0.208–0.317 kWh, corresponding to 0.94–1.43% of the electricity consumed for driving. Furthermore, this study assumed that office workers commute on the Donghae expressway. Under the scenario considering the time of operation (traveling to and from work and parking at work) and the shading ratio, the rechargeable amount was more than 10% of the electricity consumption. The results showed that solar roofs are potential charging supplements for EV batteries. This study can provide the efficacy and optimal operation method of solar EVs for commuters.

1. Introduction

With the evolution from climate change to an era of climate crisis, efforts to transition to cleaner renewable energy are becoming increasingly important [1]. The transportation sector, which relies considerably on fossil fuels, is one of the leading sources of carbon emissions and is at the center of global efforts to reduce greenhouse gas emissions [2,3]. The European Union is on track to achieve carbon neutrality in mobility, with targets to reduce carbon emissions from new cars by 55% by 2030 and 100% by 2035 compared with 2021 [4]. Global automotive companies are experiencing a paradigm shift from internal combustion engines to electric vehicles (EVs) [5,6]. Although EVs offer clean and sustainable alternatives to internal combustion engines, they have several limitations. Issues such as limited range, long charging times, and insufficient charging infrastructure are major barriers to their widespread adoption [7]. Furthermore, the environmental benefits of EVs are partially negated if the electricity used for charging comes from nonrenewable sources. Addressing these issues is essential for maximizing the environmental and economic benefits of EVs and improving their functions in sustainable transportation ecosystems.

Solar EVs (also referred to as solar cars or solar-powered EVs) or vehicle-integrated photovoltaics (VIPVs) integrate photovoltaic (PV) systems into EVs and are considered promising alternatives to minimize battery charging issues in EVs [8]. VIPVs integrate PV modules into various parts of the vehicle, such as the roof, bonnet, rear window, and side doors, to generate electricity from PV power while driving or parked [9]. Because these PV modules are mounted on the vehicle shell and connected to the EV drive battery, the generated power can be used for charging [10]. Thus, these systems contribute to increasing the range of EVs, reducing their dependence on external charging infrastructure, and reducing carbon dioxide emissions. Owing to these advantages, leading automotive companies worldwide are developing and expanding the production of various solar EV models (Sono Motors’ Sion; Lightyear’s Lightyear One; Hanegy’s Solar O, L, R, A; Aptera Motors’ Luna; Hyundai Motor Group’s Sonata Hybrid; and other models) [11]. However, despite the declining manufacturing costs of solar EVs, high unit costs owing to technical limitations in manufacturing complexity and low public awareness of their utility remain issues that need to be addressed [12]. Therefore, an accurate assessment of the amount of electricity that can be generated and the economic and environmental impacts of operating a solar EV is crucial for optimizing the design and functionality of the EV and evaluating its utility as a battery-charging supplement.

Various studies have been conducted to advance this technology and increase the utility of solar EVs. Sierra and Reinders [13] presented a design study on conceptual PV applications for solar mobility systems, highlighting multidisciplinary approaches and evaluating design challenges in integrating PV technology for various transport modes. Commault et al. [14] reviewed and summarized the rapid advancements in VIPV (e.g., car, ship, drone, plane, bus, train, truck) technologies, highlighting the increasing deployment of various PV cell and module technologies for EVs. In solar EVs, PVs are primarily utilized for propulsion or powering accessories. Cells are mainly silicon mono-crystalline (20–24% efficiency), with triple junction cells (34%) and GaAs III-V flexible (29%), also reported in vehicles. Advancements in PV technology or more data on different solar cells and efficiencies for solar EVs can be found in Green et al. [15]. In Japan, researchers have examined the applicability of different types of solar cells to develop high-efficiency or low-cost solar cells and modules that can be applied to solar EVs and have presented the expected cost savings or benefits [16,17,18]. Schuss and Fabritius [19] examined the impact of various environmental conditions, including parking and driving conditions, on the feasibility of solar EVs. The International Energy Agency (IEA) report [20] identified several factors that reduce the efficiency of solar EV systems, including shading losses (−30%), maximum power point tracking losses (−5%), DC/DC conversion losses (−5%), and DC charging/discharging losses (−2%). One of the most active areas of research is EV charging stations, with many studies analyzing the considerations, optimal locations, and technical and economic implications of their introduction [21,22,23,24]. Several studies have investigated parking space for solar EVs, including optimal installation and operation planning [25,26] and the prediction of the charging rate of solar EVs during parking time by considering the insolation [27,28]. To identify shaded areas that reduce solar EV power generation, a street view image was utilized to derive the hourly shaded area on a road [29], and a shading matrix was created by comparing modeling and field measurements to calculate the hourly percentage of shaded area for a given location [30]. In addition to the technical aspects of solar EVs, research has also been conducted on the policy side. Popiolek and Thais [31] observed that France is evaluating various innovation policies to promote solar mobility by 2030, aiming to achieve environmental, economic, and social goals. They found that a policy combining research and development subsidies and a carbon tax would be most effective, and public acceptance is crucial for the widespread adoption of solar mobility technology.

Recently, several studies have evaluated the PV potential of solar vehicles including buses, trains, and cars. A model to predict the power production of a solar bus with a solar roof has been developed and validated via in situ experiments [32,33]. An algorithm for predicting the PV potential of a solar train has been proposed, assuming that the modules are attached to the roof of the train and reflect spatially and temporally varying insolation and shadow effects during operation [34]. Park et al. [11] evaluated the potential of an EV with PV modules on all four sides of the vehicle (roof, rear window, and doors on both sides), depending on the driving route and travel time. By considering the tilt and orientation angles of each PV module attached to the vehicle and the tilt and orientation angles of the road along the driving route, the sunlight collection efficiency and potential of each EV module were evaluated. However, the above-mentioned studies are limited in that they focused on developing a model to calculate the power generation of solar mobility in a specific environment without discussing its practical utility. To expand the awareness of the utility of solar EVs, their potential should be evaluated by considering the operation methods of users. To date, no studies have been conducted to estimate the PV potential of commuters using solar EVs based on their vehicle operation methods.

The objective of this study is to estimate the PV potential of a solar-roofed EV while a commuter is driving on an expressway. To fulfill the addressed goals, the study built a database of spatiotemporal weather characteristics in a geographic information system (GIS) environment and presented the potential for each driving section and driving time of the EV. Under the assumption of a commuting scenario for office workers in the study area, the study predicted battery charging and consumption during driving and parking. This study differs from previous studies in that it considers the operation of solar EVs by commuters and discusses the utility and feasibility of solar EVs as a battery-charging supplement.

2. Study Area

The Donghae expressway, located east of Gangwon State, Republic of Korea (hereafter Korea), was selected as the study area. The name of the expressway, ‘Donghae’ indicates the eastern sea of Korea, which is surrounded by a sea on three sides. An expressway was selected as the study area because, unlike roads in urban areas, vehicles can travel at a constant speed without stopping, rendering it suitable for evaluating the potential of solar EVs. However, on general roads, vehicles travel at irregular speeds, rendering it challenging to estimate the time at which they stop. Consequently, the potential of solar EVs cannot easily be predicted.

The Donghae expressway connects Sokcho (north) in Gangwon State to Samcheok (south) via Yangyang, Gangneung, and Donghae in a northwest–southeast direction, with four lanes for round trips (Figure 1a). The total length of the section from Sokcho (latitude 38°21′, longitude 128°51′) to Samcheok (latitude 37°38′, longitude 129°21′) is 123.98 km on the Donghae expressway. The expressway is mainly flat with little slope variation, and most of the other sections have a gradient of less than 5°. Because the Donghae expressway is built on a site carved out of the eastern part of a mountain range, the western side of the expressway is mostly mountainous, whereas the eastern side is mostly open, with views of the sea or only partially surrounded by mountains (Figure 1b). Therefore, mountainous terrain on the western side of the expressway shades some sections of the expressway only during the afternoon hours (when the sun is to the west). In addition, few buildings are constructed in the mountains around the Donghae expressway. Therefore, the shaded areas on the expressway are mainly created by the mountainous terrain.

Among the five cities (Sokcho, Yangyang, Gangneung, Donghae, and Samcheok) that the Donghae expressway passes through, Gangneung-si has the highest concentration of commercial facilities and the largest number of residents (approximately 210,000). The other areas are home to 90,000 people or fewer. Regarding intercity traffic, Gangneung has the largest number of commuters traveling to work from other areas. Moreover, according to the Korean National Transportation Database [35], the average commute time in Gangwon state is approximately 30 min.

3. Methods

In this study, a research procedure was designed, as shown in Figure 2, to evaluate the potential of solar EVs when traveling on expressways and parked in workplaces. First, a database of expressway, geospatial, and weather data and the specifications of solar EVs in the study area was established. Thus, the tunnel area where power generation was not possible in the expressway section was extracted and the shaded area (by the position of the sun and surrounding terrain) was modeled. Then, for the areas where PV potential was possible, the PVWatts model [36] was used to predict the PV potential by hour of the year, using weather characteristics depending on the state and specifications and design values of the PV module of the EV as inputs. To further evaluate the utility of the EVs, this study assumed a commuter vehicle operation scenario and evaluated the electricity consumption of the EVs during traveling (commuting) and electricity charging during traveling and parking (working) hours. Finally, the sensitivity of electricity consumption and charging to changes in the driving speed of the EV were analyzed and are presented in the discussion section.

3.1. Identification of Unpowered Areas

To assess PV power potential, areas where power generation is not possible should be identified. In this context, non-solarization means that solar radiation does not reach the module because of the external environment (i.e., obstructions) or because the normal direct irradiation is less than 120 W/m² [37]. Because the area of interest in this study is a roadway, tunnel sections and shaded areas are considered non-solarization areas. Tunnels cannot be generated regardless of the time of day, whereas shaded areas are defined by the surrounding environment and the sun’s position (altitude and azimuth), which varies depending on the time of day. When a solar EV travels through such areas, battery electricity is consumed but cannot be recharged. The ratio of unpowered area can be calculated using Equation (1). When calculating the ratio of unpowered area, overlapping tunnels and shadows are not double counted.

$$Ratioofunpoweredarea(\%)={\displaystyle \frac{Lengthoftunnels\cup shadedareas\left(\mathrm{m}\right)}{Lengthofexpresswaysection\left(\mathrm{m}\right)}}$$

(1)

3.1.1. Tunnels

Because tunnels are fixed sunlight obstructions regardless of time changes, sunlight cannot be developed within tunnels. To extract the location and length information of tunnels, the latest national road network GIS data of Korea provided by the National Transportation Database (https://viewt.ktdb.go.kr/) (accessed on 15 July 2024) [35] were obtained, and through data preprocessing, data in the form of a polyline for the tunnel section of the Donghae expressway were extracted.

3.1.2. Shaded Areas

The position of the sun and year-round shading created by the terrain surrounding the expressway over time were modeled. The shaded area applied in this study corresponded to the height of the solar roof of an EV traveling on an expressway. Therefore, expressway data were generated by adding the height of the expressway (which varied by section) and that of the EV solar roof (1.7 m) and modeling the shading for the height-corrected expressway sections. To this end, 15 topographic maps (scale 1:5000) covering the Donghae expressway section from the Geospatial Information Platform (https://www.vworld.kr/ (accessed on 15 July 2024)) [38] were acquired, and the maps were spatially integrated using the union analysis method in ArcGIS Pro 3.2 software. The topographic maps included contour data within a radius of 1–10 km from the centerline of the expressway. Therefore, they can reflect the surrounding environment, which forms the shaded areas. To understand the terrain undulations around the expressway, contour lines were extracted from the integrated topographic maps. Then, a triangulated irregular network was created and converted into digital elevation model (DEM) data with a spatial resolution of 30 m × 30 m (Figure 3a). To accurately model the original shaded area, a digital surface model (DSM) that considers the surrounding terrain and buildings should be employed. However, as aforementioned in the study area section, the expressway in this study area was mostly constructed by cutting through the middle of a mountainous area. Therefore, a DEM, rather than a DSM, was adopted as the input because few buildings are constructed in the surrounding mountains or the altitude is low, which has little effect on expressway shading.

The shaded areas were modeled using the Hillshade analysis method [39] in the ArcGIS software (Figure 3b) based on the DEM data around the expressway and the hourly position of the sun throughout the year. This algorithm displays the degree of shading by considering the position (altitude and azimuth) and shadow of the sun (light source) from gridded raster data showing terrain undulations. A database was built for the modeling results of the hourly shading area for all sections of the Donghae expressway. Shaded areas appear in the form of black grid cells, as shown in Figure 3c, and if the location of the shaded area and expressway overlap, the solar EV cannot generate power because of sunlight obstruction.

3.2. Estimation of the PV Potential from Solar EVs on Donghae Expressway

Solar EVs with attached or integrated PV modules can generate electricity while driving, provided the site has PV potential. Generally, the PV power potential can be assessed based on the weather characteristics of the area of interest, specifications, and design values of the PV system components (modules and inverters). However, the difference is that, unlike fixed PV systems, PV mobility generates electricity while traveling; therefore, weather conditions and shading vary depending on the location of the main activity, including outside of hours.

3.2.1. Weather Characteristics

This study predicted the expected annual potential in the near future rather than targeting a specific year. For meteorological data, typical meteorological year (TMY) data were adopted, which are long-term observations that provide a good hourly representation of the climate characteristics of the region. TMY data have been used in various studies to evaluate PV power potential and are therefore considered valid input. However, as shown in Figure 1, the Donghae expressway, which is approximately 120 km long, is located in the north–south direction; therefore, weather characteristics may differ depending on the location. Herein, TMY data containing hourly insolation values for three cities, Sokcho, Gangneung, and Donghae, which are the main cities through which the Donghae expressway passes, were obtained from the SolarGIS website (https://www.solargis.com/ (accessed on 15 July 2024)) [40] (TMY data for Sokcho, Gangneung, and Donghae, which have existing meteorological observatories, were obtained from the SolarGIS website and are referred to as meteorological observation points in this study). To construct meteorological data for the Gangwon state, data from the Korea Meteorological Administration’s Weather Data Open Portal (https://data.kma.go.kr/ (accessed on 15 July 2024)) [41] also were reviewed but were not adopted because of missing data and errors.

To reflect the weather characteristics corresponding to the driving location and time of solar EVs, the nearest neighbor classification method was applied between the Donghae expressway and three weather stations. This classification methodology is one of the most basic classification rules in machine learning and is a methodology for finding the closest point or data to a specific point based on the Euclidean distance or a specific objective function. As shown in Figure 4, this study used the TMY data from the closest of the three observation points (Sokcho, Gangneung, and Donghae) to the point where the solar EV was located at the time of driving. The Donghae expressway was categorized into three sections considering the location of the latest weather observation point from the vehicle. The red section was termed Section A (reflecting the weather characteristics of Sokcho), the green section was termed Section B (reflecting the weather characteristics of Gangneung), and the blue section was termed Section C (reflecting the weather characteristics of Donghae). For example, if a solar EV is driven or parked in the red section, it can be considered to reflect the Sokcho weather characteristics.

When traveling between two or more sections in a drive exceeding one hour, the weather characteristics change because of time and location variations. In this case, the average value of the weather characteristics of the two cities (e.g., solar irradiation at the nearest hour to the departure city and solar irradiation at the nearest hour to the arrival city) is applied, as shown in Equation (2). For example, if a solar EV departs from Sokcho at 8:00 a.m. and arrives in Gangneung one hour later, the average of the weather characteristics of Sokcho at 8:00 a.m. and those of Gangneung at 9:00 a.m. is used. If the departure or arrival time is not on time, the traveling or stopping time is calculated in minutes, and the weather characteristic value is applied using the ratio of minutes to hours.

Hourly weather characteristics for traveling between cities (place, time)
={weather characteristics_{(City A, t)} + weather characteristics_{(City B, t or t+1)}} ÷ 2

(2)

3.2.2. Specifications of PV Module on EVs

Because the main sources of PV power in this study were the PV modules attached to the ceiling and bonnet of the vehicle, specification information is crucial to evaluate the potential. The solar EV selected for this study is a SION solar car from Sono Motors, Germany, which has more than half of its body area covered by PV modules (Figure 5). The vehicle has PV modules made of high-efficiency silicon monocrystalline cells attached to the front of the body (roof, bonnet, and both doors). Detailed information about the PV modules of the vehicle is presented in

Table 1. Specification of solar EV (Sion) used for PV potential calculation [42].

Items	Design	Value
Module type		mono-Si
Module capacity	Designed originally	1208 W
	Considered in this study	1095 W
Module area	Designed originally	7.5 m2
	Considered in this study	6.8 m2
Module efficiency		24%
Electric efficiency		5.6 km/kWh

.

In this study, only the PV module capacities of the bonnet and roof areas, where the PV modules are attached almost horizontally, were considered to estimate the potential during driving, excluding the side modules, where light collection efficiency is significantly affected by the driving path and direction. The PV modules considered for the PV potential estimation had a capacity of 1095 W and a module area of 6.8 m². The electric efficiency of the vehicle was 5.6 km/kWh, which was used to calculate the ratio of electricity consumption to potential when the EV was driven.

3.2.3. PV Potential Estimation Model and Parameter Setting

The PVWatts model developed by the United States National Renewable Energy Laboratory was employed to estimate the potential of solar EVs. This model, which can easily estimate PV power potential from eight inputs, has been used in numerous studies; therefore, it is deemed reliable. The specific PV power potential of the PV Watts model can be expressed using Equation (3), and the solar cell temperature estimation can be expressed using Equation (4).

$$PowerGeneration\left(\mathrm{k}\mathrm{W}\mathrm{h}\right)={\displaystyle \frac{I}{1000}}A{\displaystyle \frac{n_p}{100}}{\displaystyle \frac{n_i}{100}}{R}_{da}{\displaystyle \frac{100-(25-T_{cell})\gamma }{100}}{\displaystyle \frac{L}{100}}$$

(3)

$$T_{cell}=T_{air}+{\displaystyle \frac{NOCT-20}{800}}I$$

(4)

I indicates the hourly global horizontal irradiance (GHI) values from the closest meteorological station to the driving location of the EV, where the GHI (W/m²). This reflects the difference in insolation due to location, as the area of interest is more than 120 km long. A is the area (m²) of the PV modules installed horizontally on the bonnet and roof of a solar EV. n_p denotes the efficiency of the PV module (%), n_i is the efficiency of the inverter (%), R_da is the DC-to-AC conversion efficiency (%), T_cell is the temperature of the PV module (°C), γ is the temperature coefficient (%/°C), and L is other losses (%). Here, T_cell (°C), which represents the temperature of the PV module, can be derived from the outside air temperature (T_air) (°C) and nominal operating cell temperature (NOCT).

Except for the weather characteristics, module specifications, and design values of the solar EV, the remaining factors to assess the PV potential of the solar EV employ the values presented in

Table 2. Parameter values of the PVWatts model used for the PV potential calculation.

Parameter	Value
Inverter efficiency	96%
DC to AC ratio	1.1
Temperature coefficient	−0.5%/°C
Other losses	20%
NOCT (nominal operating cell temperature)	50 °C

. Other loss rates that can significantly affect the PV potential results were assumed to be conservative to account for different environments and changing conditions. In addition, this study assumed that the EVs travel at a constant speed of 60 km/h to estimate the PV power potential. This is in accordance with the minimum speed limit applied to some sections of the Donghae expressway, and the maximum speed limit is 100 km/h.

3.3. Estimation of the PV Potential from a Commuter’s Solar EV According to the Operation Scenario

Solar EVs can generate power (recharge batteries) while driving on the road or parked at rest stops or other places of interest. The most common example is when a commuter traveling to and from work utilizes a solar EV. Therefore, in this study, the potential of commuters to use solar EVs on sections of the Donghae expressway was estimated. The difference is that commuters can use solar charging when traveling between expressway sections or parked at a specific place during work.

To evaluate the potential for commuting with solar EVs, four different scenarios were assumed for the commuter’s residence and work location, commute time, driving section, working hours, and degree of shading (percentage) of the parking area, as presented in

Table 3. Commuter’s operational scenarios and assumptions for solar EV.

Item		Scenario 1	Scenario 2	Scenario 3	Scenario 4
House		Sokcho		Yangyang
Workplace		Gangneung		Gangneung
Driving section		Section A + Section B1		Section B1
Driving distance (round trip)		120 km (one way of 60 km)		60 km (one way of 30 km)
Driving hours	Go to work	1 h (8 a.m. to 9 a.m.)		0.5 h (8:30 a.m. to 9 a.m.)
	Back home	1 h (5 p.m. to 6 p.m.)		0.5 h (5 p.m. to 5:30 p.m.)
Shadowed region at expressway		Identical condition in Figure 3
Parking hours		8 h (9 a.m. to 5 p.m.)
Shadow ratio at parking lot	09:00	0	30%	0	30%
	10:00		20%		20%
	11:00		10%		10%
	12:00		0%		0%
	13:00		10%		10%
	14:00		20%		20%
	15:00		30%		30%
	16:00		40%		40%

. In all scenarios, the commuter was assumed to arrive at work at 9 a.m., leave at 5 p.m., and drive at a speed of 60 km/h on a one-way section of the Donghae expressway to and from work. Scenarios 1 and 2 denote when a commuter goes to work from home (Sokcho) at 8 a.m., drives for 1 h on Sections A and B1 of the Donghae expressway, a distance of 60 km each way, parks the solar EV outside during the 8 h they work at their workplace (Gangneung), leaves work at 5 p.m., and drives for 1 h on Sections B1 and A back home. However, because the proportion of shaded areas may vary depending on the time of day when parking at work, this study assumes that no shadow is formed from 9 a.m. to 5 p.m. (Scenario 1) and that the proportion of shadows on the vehicle varies with time (Scenario 2). The reason for assuming that the percentage of shadows varies with time is that when driving on an expressway, the shadowed area can be determined by location through modeling; however, when parking near work, the driver can choose where to park. Conversely, for Scenarios 3 and 4, the distance between home (Yangyang) and work (Gangneung) is 30 km, and as the commuter drives on Section B1 of the Donghae expressway for 0.5 h, the time of arrival at work is 8:30 a.m. and the time of departure is 5:30 p.m. Other conditions, such as parking time and shading ratio at the workplace, were the same.

4. Results

4.1. Results of Unpowered Area

Some results of the analysis of the tunnels and shaded areas where power generation is not possible on the Donghae expressway are shown in Figure 5. The grey line running from northwest to southeast is the expressway, the rectangular area with rounded corners is the tunnel section (only the approximate size is shown), and the black grid cells represent the shaded areas. The length of each section of the Donghae expressway, the length of the tunnel, and the percentage of road length that cannot be developed are presented in

Table 4. Length and ratio of tunnels on the Donghae expressway.

Item	Section A	Section B	Section C	Total
Expressway length (km) a	35.98	53.61	34.38	123.98
Tunnel length (km) b	4.56	11.28	8.41	24.25
Ratio (%) b/a	12.67	21.04	24.46	19.56

. As shown in Figure 4, the tunnel length of Section A is 4.56 km, the tunnel length of Section B is 11.28 km, and the tunnel length of Section C is 8.41 km. The area that could not be developed because of the tunnel is approximately 20% of the total length of the Donghae expressway.

Generally, tunnel segments and shaded areas are separated, but as shown in Figure 5, they overlap in some areas. This is because the study extracted them separately during the analysis process, and calculating and summing their lengths can cause an overlap. Therefore, in cases where the tunnel and shaded areas overlapped, this study considered all shaded areas as tunnel sections and excluded the results of the overlapping shaded areas.

In the case of shaded areas, unlike tunnels, the position of the sun changes daily; therefore, areas where power generation is not possible vary. During the summer solstice, when the sun’s altitude is the highest, the shaded area is relatively large, and during the winter solstice, when the sun’s altitude is the lowest, the shaded area is small. Therefore, this study presents the results of the shadow analysis on the summer and winter solstices, when the difference in the shaded area based on the hourly position of the sun is the largest, as presented in

Table 5. Shaded regions on the Donghae expressway on the summer solstice and winter solstice calculated by shadow analysis.

Date	Time (hour)	Section A		Section B		Section C
		Pixel (ea)	Length (km)	Pixel (ea)	Length (km)	Pixel (ea)	Length (km)
Summer Solstice	07:00	195	5.85	183	5.49	100	3.00
	08:00	38	1.14	129	3.87	13	0.39
	09:00	15	0.45	31	0.93	8	0.24
	10:00	0	0.00	2	0.06	3	0.09
	11:00	0	0.00	0	0.00	1	0.03
	12:00	0	0.00	0	0.00	0	0.00
	13:00	0	0.00	0	0.00	0	0.00
	14:00	0	0.00	0	0.00	0	0.00
	15:00	0	0.00	1	0.03	12	0.36
	16:00	22	0.66	78	2.34	17	0.51
	17:00	142	4.26	98	2.94	110	3.30
	18:00	227	6.81	231	6.93	166	4.98
	Average	53	1.60	63	1.88	36	1.08
Winter Solstice	09:00	245	7.35	238	7.14	169	5.07
	10:00	210	6.30	194	5.82	144	4.32
	11:00	168	5.04	158	4.74	107	3.21
	12:00	150	4.50	146	4.38	98	2.94
	13:00	166	4.98	164	4.92	114	3.42
	14:00	195	5.85	202	6.06	138	4.14
	15:00	228	6.84	252	7.56	163	4.89
	16:00	238	7.14	262	7.86	182	5.46
	17:00	245	7.35	271	8.13	184	5.52
	Average	205	6.15	210	6.29	144	4.33

. At this time, only the results from 7 a.m. to 6 p.m. on the summer solstice and from 9 a.m. to 5 p.m. on the winter solstice are shown, excluding the time when power generation is not possible because of insolation conditions. Therefore, the average shadow area per day on the summer solstice of the Donghae expressway was 4.56 km (Section A: 1.60 km, Section B: 1.88 km, and Section C: 1.08 km), and the average shadow length per day on the winter solstice was 16.77 km (Section A: 6.15 km, Section B: 6.29 km, Section C: 4.33 km). The shaded area on the road was longer on solstice days than on equinox days because of the low altitude of the sun. Therefore, the amount of shading is significantly influenced by season.

The length of the section that could not be powered due to the tunnel was 24.25 km along the entire section of the Donghae expressway (123.98 km long), and the average value of the road length that could not be developed due to shadow was 4.56 km at summer solstice and 16.77 km at winter solstice. Notably, tunnels in the study area cause approximately 1.4 to 5.3 times more sunlight obstruction than shadows caused by terrain. In addition, the average ratio of the area that could not be developed because of tunnels and shadows to the total length of the Donghae expressway was approximately 23% during the summer solstice and 33% during the winter solstice. This is calculated by entering the aforementioned length of the tunnel, the average length of shadows on the summer and winter solstice, and the length of the Donghae expressway into Equation (1).

4.2. Results of the Estimated PV Potential from Solar EVs on Donghae Expressway

Among the three cities (Sokcho, Gangneung, and Donghae) through which the Donghae expressway passes, insolation is the highest in Sokcho in the morning, highest in Gangneung in the afternoon, and generally lowest in Donghae during the summer solstice (Figure 6a). In contrast, Donghae had the highest insolation, and Sokcho and Gangneung had similar insolation (Figure 6b). In other words, the insolations of the three cities were different, and the ranking of the insolation of each city by season also differed. Therefore, the weather characteristics of the three cities should be reflected, rather than the weather characteristics of a specific city in the study area.

Table 6. PV potential of one solar car on the summer solstice and winter solstice.

Date	Time	Ratio of Unpowered Area (%)			PV Potential (kWh)
		Section A	Section B	Section C	Section A	Section B	Section C
Summer Solstice	07:00	28.9	31.3	33.2	0.026	0.018	0.008
	08:00	15.8	28.3	25.6	0.059	0.024	0.015
	09:00	13.9	22.8	25.2	0.086	0.048	0.017
	10:00	12.7	21.2	24.7	0.106	0.068	0.021
	11:00	12.7	21.0	24.6	0.103	0.120	0.045
	12:00	12.7	21.0	24.5	0.076	0.171	0.075
	13:00	12.7	21.0	24.5	0.090	0.189	0.065
	14:00	12.7	21.0	24.5	0.109	0.180	0.051
	15:00	12.7	21.1	25.5	0.120	0.127	0.036
	16:00	14.5	25.4	26.0	0.053	0.120	0.022
	17:00	24.5	26.5	34.1	0.031	0.096	0.012
	18:00	31.6	34.0	39.0	0.029	0.045	0.006
	Average	17.1	24.6	27.6	0.074	0.101	0.031
Winter Solstice	09:00	33.1	34.4	39.2	0.012	0.015	0.013
	10:00	30.2	31.9	37.0	0.027	0.040	0.034
	11:00	26.7	29.9	33.8	0.035	0.063	0.053
	12:00	25.2	29.2	33.0	0.060	0.078	0.065
	13:00	26.5	30.2	34.4	0.074	0.088	0.065
	14:00	28.9	32.4	36.5	0.065	0.082	0.058
	15:00	31.7	35.1	38.7	0.039	0.063	0.041
	16:00	32.5	35.7	40.4	0.019	0.039	0.018
	17:00	33.1	36.2	40.5	0.005	0.010	0.008
	Average	29.8	32.8	37.1	0.037	0.053	0.039

presents the percentage of areas (roads) where power generation is not possible because of tunnels and shadows on the Donghae expressway at different times of the day on both the summer and winter solstices and the results of the evaluation of the PV potential that can be expected from driving a single solar EV. On both solstices, the proportion of roads with no power generation to the length of each road section was the highest in Section C, followed by B and A. The proportion of non-generating areas by road section ranged from 17.1% to 27.6% during the summer solstice and 29.8% to 37.1% during the winter solstice. This value can be interpreted as a loss (amount or percentage) in terms of PV potential. The average hourly power generation of the solar EVs on each road section was within the range of 0.031–0.074 kWh for the summer solstice and 0.037–0.053 kWh for the winter solstice. In Sections A and B, the average value of the generation potential on the summer solstice was twice that of the winter solstice, whereas, in Section C, the converse was observed. This is because of the different hours of power generation on the two days (11 h on the summer solstice and 8 h on the winter solstice) and the insignificant amount of power generation during the morning and evening hours of the summer solstice. Therefore, comparing the average values of PV potential on the two days was not meaningful.

To evaluate the potential of a solar EV driven in all sections of the Donghae expressway (Sections A, B, and C) for approximately 2 h, the following method was used. The PV power potential values presented in Table 6 reflect the length and driving time of each road section; therefore, these values can be considered the power generated at the time of entering the road section. For example, assume that a solar EV leaves its starting point at 1 p.m. on a day and travels through Sections A, B, and C at a constant speed of 60 km/h and arrives at the endpoint at 3:00 p.m. As it takes off in Section A at 1 p.m., it has a generation potential of 0.090 kWh. As it passes through Section B at 1:36 p.m., it generates 0.1836 kWh (the value at the 60th percentile between the values of 0.189 and 0.180 at 1:00 and 2:00 p.m., respectively). As it enters Section C at 2:30 p.m., it generates 0.0435 kWh (the midpoint between the values of 0.051 and 0.036 at 2:00 and 3:00 p.m. respectively), for a total of 0.3171 kWh generated (charged). If a solar EV is driven simultaneously on the same day at sunset, a total of 0.2079 kWh is generated, including 0.074 kWh in Section A, 0.0844 kWh in Section B, and 0.0495 kWh in Section C. This is automatically calculated using the departure time from the electricity generation database by the time of year for each section of the Donghae expressway.

Solar EVs consume electricity while driving and charge simultaneously. The electricity consumption estimated from the electricity cost of a solar EV and the mileage of the Donghae expressway is equivalent to 22.138 kWh. Therefore, the values of electricity consumption compared with electricity charge during the driving time of the solar EV are approximately 1.43% for the summer solstice and 0.94% for the winter solstice. The ratio of electricity consumption to charging during the driving time is somewhat lower than these values. However, these values do not consider parking (or stopping) time.

4.3. Results of Estimated PV Potential from a Commuter’s Solar EV According to the Operation Scenario

The results of the daily generation potential of solar EVs for the four scenarios are presented in

Table 7. Estimated hourly PV potential of the solar car of a commuter on the summer solstice according to four scenarios.

Time (hour)	PV Potential (kWh)
	Scenario 1		Scenario 2		Scenario 3		Scenario 4
	Section A	Section B1	Section A	Section B1	Section A	Section B1	Section A	Section B1
08:00	0.059	0.011	0.059	0.011		0.013 a		0.013 a
09:00		0.062		0.043		0.062		0.043
10:00		0.087		0.069		0.087		0.069
11:00		0.152		0.137		0.152		0.137
12:00		0.217		0.217		0.217		0.217
13:00		0.239		0.215		0.239		0.215
14:00		0.228		0.183		0.228		0.183
15:00		0.161		0.113		0.161		0.113
16:00		0.160		0.096		0.160		0.096
17:00	0.031	0.043	0.031	0.043		0.054 b		0.054 b

^a: 08:30~09:00 (0.5 h). ^b: 17:00–17:30 (0.5 h).

. Comparing the results of Scenarios 1 and 2, the potential during commuting hours (8:00 a.m. and 5:00 p.m.) is the same (owing to the same driving conditions and different shading conditions when the EV is parked), whereas the potential during working hours (9:00 a.m. to 5:00 p.m.) is different. Comparing the results of Scenarios 1 and 3, the potential during commuting hours is different (owing to the different conditions of EV transportation and the same shading environment when parking), whereas the potential during working hours is the same.

Figure 7 shows the sum and ratio of the generation potential for each scenario for commuting (traveling) and working (parking) hours and the daily ratio of electricity production (charging) to electricity consumption for each scenario. In all scenarios, the parking time (8 h) exceeded the travel time (1 or 2 h) and the time for the prevailing insolation conditions; thus, the ratio of the generation potential during parking was approximately 7–9 times higher than that during travel. The generation potential of commuters depending on their travel times (Scenarios 1 and 2 versus Scenarios 3 and 4) was compared. The PV potential for a 2 h round trip is approximately 1.06 times that of a 1 h round trip, indicating a slight difference in terms of potential because this value corresponds to the difference in power generation in a 1 h commute time when insolation is low. However, a significant difference is observed in terms of the electricity consumption-to-charge ratio because the electricity consumption is approximately twice as much when the driving time is doubled. In the case of a 2 h round trip, the maximum charge-to-electricity consumption ratio is 6.8%, and the maximum charge-to-electricity consumption ratio is 12.8% for a 1 h round trip. When comparing the potential of different shading conditions during parking, the potential of Scenarios 2 and 4 (assuming realistic shading conditions) was approximately 84% that of Scenarios 1 and 3 (assuming no shading conditions). In other words, the difference in potential is significant depending on the degree of shading during parking; therefore, the shading conditions should be precisely and practically evaluated.

4.4. Green House Gas Reduction Effects from Solar EVs

In the previous section, the amount of electricity generated was presented when a solar EV is driven one-way on the Donghae expressway during the summer and winter solstices. To calculate the amount of carbon dioxide reduction per trip based on the average of the power generation on the summer solstice and winter solstice, the power generation was multiplied by the greenhouse gas emission standard value (0.4585 kg·CO₂/kWh) provided by the Korean Ministry of Environment [43]. This was calculated by summing the average PV potential of each expressway section (A, B, C) for the summer solstice and winter solstice presented in Table 6. The summer solstice has a PV potential of 0.206 (=0.074 + 0.101 + 0.031) kWh, while the winter solstice has a PV potential of 0.129 (=0.037 + 0.053 + 0.039) kWh.

Results showed that approximately 0.09 kg and 0.06 kg of carbon dioxide could be reduced per one-way trip of a solar EV for the summer and winter solstice, respectively, and approximately 34.47 kg and 21.59 kg of carbon dioxide could be reduced per year (365 days) for one daily trip. This is based on the assumption that a solar EV drives a length of 123.98 km in approximately 2 h; therefore, the carbon dioxide reduction cannot easily be considered high.

5. Discussion

5.1. Comparison of the PV Potential of the Solar EV with Other Renewable Energy Sources

The PV potential of solar EVs was compared with those of ground-mounted PVs and wind turbine systems to help in understanding the relative advantages and limitations of solar EVs. To make this comparison, it is necessary to calculate the amount of electricity that an entire fleet of vehicles traveling on the Donghae expressway could generate in a year, rather than a single vehicle’s one-way trip. As such, the potential was also calculated when considering the traffic on the Donghae expressway instead of a single solar EV in this study. Hourly sectional traffic data for the Donghae expressway for 2021 were collected and compiled from the public data portal of the Korean Expressway Corporation [35] to reflect the average number of vehicles per hour for each section (

Table 8. PV potential on the summer solstice and winter solstice considering the mean number of vehicles per hour.

Date	Time	Number of Vehicles (ea)			PV Potential (kWh)
		Section A	Section B	Section C	Section A	Section B	Section C
Summer Solstice	07:00	225	327	347	5.9	5.9	2.8
	08:00	341	447	479	20.1	10.7	7.2
	09:00	501	516	444	43.1	24.8	7.5
	10:00	645	618	517	68.4	42.0	10.9
	11:00	790	753	553	81.4	90.4	24.9
	12:00	744	747	574	56.5	127.7	43.1
	13:00	756	753	577	68.0	142.3	37.5
	14:00	737	772	593	80.3	139.0	30.2
	15:00	713	773	606	85.6	98.2	21.8
	16:00	659	756	597	34.9	90.7	13.1
	17:00	584	696	557	18.1	66.8	6.7
	18:00	421	520	421	12.2	23.4	2.5
	Sum	7116	7678	6265	574.5	861.9	208.2
Winter Solstice	09:00	501	516	444	6.0	7.7	5.8
	10:00	645	618	517	17.4	24.7	17.6
	11:00	790	753	553	27.7	47.4	29.3
	12:00	744	747	574	44.6	58.3	37.3
	13:00	756	753	577	55.9	66.3	37.5
	14:00	737	772	593	47.9	63.3	34.4
	15:00	713	773	606	27.8	48.7	24.8
	16:00	659	756	597	12.5	29.5	10.7
	17:00	584	696	557	2.9	7.0	4.5
	Sum	6129	6384	5018	242.8	352.9	201.9

). This is an estimate of the potential based on the assumption that the total number of vehicles traveling on the Donghae expressway, which is used by office workers, is the same as that of the solar-powered EVs used in this study. The average number of vehicles per day for each section ranged from 6265 to 7116 ea, and the potential was within the range of 208.2–861.9 kWh on the sunset day and 201.9–352.9 kWh on the moonrise day. The estimated potential of solar EVs was 1644.6 kWh on the summer solstice and 797.5 kWh on the winter solstice. The total potential of the summer solstice was approximately twice that of the winter solstice, and the potentials by section were B, A, and C in decreasing order.

If it is assumed that the average potential on the summer and winter solstices is the potential for a single day, the total power generation (charging) of a solar EV traveling on the Donghae expressway for a year (365 days) is approximately 445.7 MWh. This is compared with the power generation of a 341 kW land-based fixed PV power facility, considering that the average peak sun hours in Korea are approximately 3.58 h per day based on the recent literature survey [44]. Moreover, this is almost equivalent to the amount of power that a 217 kW wind turbine facility can generate taking into account the 23.5% yearly capacity factor of wind power facilities in Gangwon state [45].

While generating less power than other renewable sources, solar EVs offer distinct advantages. VIPVs can produce power while driving and parking, extending their range and reducing their dependence on charging stations, and they enhance the environmental image among consumers. Environmentally, EVs help mitigate climate change, reduce air pollution, lower carbon footprints, and promote sustainable energy use.

5.2. Sensitivity Analysis of PV Potential According to Driving Speed

In this study, the PV potential of an EV was evaluated under the assumption of a driving speed of 60 km/h on the Donghae expressway. However, on actual expressways, the driving speed may vary depending on various conditions and environments, such as driver preferences, traffic congestion, driving time, and weather. Therefore, in a sensitivity analysis, the electricity cost, electricity generation, and electricity reduction (the difference between consumption and generation) as the driving speed of an EV increased were compared.

Three driving speeds were considered: 60, 80, and 100 km/h. Official information regarding the change in electricity cost with increasing driving speed of solar EVs was not obtained; therefore, the electricity cost values of EVs with similar specifications were applied. Figure 8a shows the evolution of the electricity cost with the increasing driving speed of the solar EV in this study. The electricity cost decreased by approximately 37.5% as the driving speed increased. Figure 8b shows the average power generation per section with a change in driving speed, indicating that the power generation decreases by approximately 40% when the driving speed changes from 60 to 100 km/h. Figure 8c shows electricity reduction (difference between electricity consumption and generation during driving) as a function of driving speed, indicating that the average electricity consumption increases by approximately 64% based on electric efficiency, as shown in Figure 8a, when the driving speed changes from 60 to 100 km/h. As the driving speed of a solar EV increases, the electricity consumption of the battery increases because the electricity cost decreases with high-speed driving, and the power generation decreases because the power generation time decreases as shown in Figure 8b. Therefore, the potential of the solar EV presented in this study may increase or decrease depending on the driving speed.

5.3. From Simulation to Real-World Application

While this study used data and a couple of assumptions to assess the PV potential of solar EVs, there are factors that should be considered when applying the data in the real world. Therefore, the possible differences between simulation-based PV potential and real-world application were addressed along with their causes.

Firstly, clouds and trees are two environmental factors away from the solar EV that can reduce the accuracy of the estimated PV potential of solar EV. In real-world scenarios, clouds and trees act as light obstructions that can shade the solar roof of the EV, thereby reducing power generation. However, accounting for these factors in simulations is often challenging. The location and size of clouds can be estimated from weather data and satellite imagery, but their real-time variability makes them difficult to incorporate accurately beyond short-term forecasts. Trees around the expressway, depending on their size and density, can be considered in input data. For instance, the distribution and height of forests can be integrated using vegetation data, but individual trees with low density pose challenges for precise shade modeling. Therefore, trees can be considered as an input factor for shading modeling depending on their density. However, in this study, the shading effects of clouds and trees were not included in the power generation prediction. Consequently, actual power generation may be lower than the simulation predictions when field verification is conducted.

Secondly, soiling and bird droppings are the two environmental factors on the car reducing power generation efficiency in real-world applications. If these factors continue to occur, they can also cause damage (wear and tear) to the PV modules, resulting in further degradation. These are the same factors that reduce the efficiency of ground-mounted PV modules. If certain parts of a PV module are covered by soiling or bird droppings, those cells will not generate power, which can also affect the performance of the whole module. With cells connected in series, a problem with one cell can affect the power production of the entire circuit. Although the impact of soiling and dirt is acknowledged, quantifying their effects on EV solar roofs in terms of probability or magnitude remains challenging. However, if the driver checks, cleans, or maintains these events between trips, the reduction in power generation can be minimized.

The third is the variability of power generation due to uncertainty in site weather. In this study, TMY data, which are representative of long-term weather characteristics, was used to evaluate the PV potential of solar EVs throughout the year. However, the uncertainty or variability of weather conditions is the biggest reason for prediction errors. Therefore, for short-term forecasts such as tomorrow or the day after tomorrow, this may be a way to reduce uncertainty by evaluating the power generation potential using short-term forecast data rather than TMY data.

The fourth is the driving style of the commuter that can influence the PV potential of solar EVs in real-world conditions. Although autonomous vehicles or cruise control functions can maintain a constant speed, this study assumes that commuters are manually driving the vehicles, resulting in potential variations in driving speed depending on the route. Consequently, not all commuters exhibit the same driving style. The subjective driving habits of drivers are significant as they affect both power generation and fuel efficiency, as demonstrated in Figure 8. Further research on these factors is necessary.

5.4. Challenges and Practical Issues in Solar EVs

Technological challenges exist for solar EVs, but efforts are being made to address them. Adopting solar EVs faces challenges such as low efficiency, limited energy storage, and infrastructure needs. As mentioned earlier, current PV cells have 20–25% efficiency, with ongoing research on multi-junction cells and perovskite materials to improve this. Perovskite cells are regarded as the future of VIPVs because they hold significant potential for achieving high efficiency, low cost, ultralow weight, and flexibility compared with those of silicon crystalline-based cells [14]. That is, perovskite cells can be attached to multiple sides of a vehicle. However, this technology still faces many technical challenges for the mass production and commercialization of solar EVs due to limitations in longevity and reliability compared to large-scale silicon crystalline cell-based modules. Thus, if these issues are solved in the future, the PV potential of solar EVs could be significantly increased. Moreover, efforts should be made to include advanced battery technologies to increase storage capacity and designs that use lightweight materials to improve efficiency. Vehicles should be designed to be lightweight and aerodynamic, utilizing innovative design techniques to reduce weight and drag.

There is more to discuss on the factors concerned with an estimation of the PV potential of a solar EV from a practical point of view. First, PV modules integrated into vehicles should exhibit exceptional durability to endure diverse operational conditions, including adverse weather phenomena and potential physical impacts. Any damage sustained by these modules could significantly impair the vehicle’s capacity for power generation, thereby compromising its overall performance [14]. Second, PV modules experience a degradation rate of approximately 0.5% per year according to the published literature [46]. This study has not accounted for this degradation in potential calculations, as the estimates focused on the PV potential for a single trip or year. However, this degradation should be considered in analyses spanning multiple years. The third is the effect of vehicle design on the PV potential of a solar EV. VIPV systems vary in design, including curved PV modules with flat glass structures, lightweight flexible modules, and rigid lightweight modules. These differing geometries and weights lead to varied power generation efficiencies. Due to higher partial shading frequency compared to ground-mounted PV modules, more resilient electrical architectures, such as increasing the number of bypass diodes in series-connected cell modules, are recommended for EV applications [14]. The vehicle design factor will need to be considered when estimating the PV potential of solar EVs in the future.

6. Conclusions

This study evaluated the potential generated when a commuter drives a solar EV on an expressway to work and parks it at work. The results showed that the potential of a solar EV was approximately 1.14 kWh for a 1 h drive (average commute time in Gangwon state: 30 min) and 8 h of parking (work and lunch breaks, general shading condition) per day, representing approximately 10.6% of the electricity consumption during driving. Moreover, the solar roof on an EV can be useful as a battery-charging supplement. Furthermore, the generation potential increased by approximately 20% to 1.373 kWh when the parking spaces around the workplace were unshaded. Because insolation is generally higher during working hours (between morning and evening) than during commuting hours (morning or evening) and exposure to the sun is longer, the potential of a solar EV is significantly influenced by the parking time and surrounding environment of the parking lot. Therefore, commuters should consider the shade environment when parking near their workplaces to maximize the charge of their solar EVs.

The following are considerations to improve the limitations of this study and derive more reliable findings and implications. Uncertainty in meteorological data is one of the most influential factors in predicting PV potential. In this study, TMY data, which represent weather characteristics on an hourly basis, were used to evaluate the potential of solar EVs. However, because EVs are moving, weather characteristics are constantly changing, and in certain cases, the driving time is as short as 1 h. Therefore, weather characteristics should be applied at the subdivision level. On urban roads, the shading characteristics cannot easily be evaluated because of light obstacles, such as streetlights and trees; thus, this study considers the shading environment of parking spaces. If a high-resolution 3D model of urban areas is built in the future, shaded areas when driving on urban roads and expressways would be accurately modeled. In addition, although this study simulated various scenarios, field experiments should be conducted and verified using solar EVs sold in the future.

From the perspective of a commuter operating a solar EV, the results of this study are expected to provide useful data for evaluating the effectiveness of VIPVs as battery-charging supplements in EVs. Furthermore, if the reliability of the results is improved, it can be used as a basis for developing a navigation service that provides PV potential depending on the driving route and time or the optimal solar charging route as shown in Figure 9. Whereas traditional navigation offers drivers the option of the fastest, shortest, or free route, in the future it will be possible to offer options for routes that are suitable for solar charging if they have similar travel times. In addition, high-efficiency perovskite [47] and transparent solar cells [48] that can be attached to vehicle glass have been developed recently. Therefore, it would be interesting to conduct a study to predict the potential of solar EVs using these technologies in the future.