Renewable and Sustainable Energy Reviews 196 (2024) 114344

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

European transport infrastructure as a solar photovoltaic energy hub

G. Kakoulaki a, *, S. Szabo a, **, F. Fahl F b, N. Taylor a, A. Gracia-Amillo c, R. Kenny a, G. Ulpiani a,
A. Chatzipanagi a, K. Gkoumas d, A. Jäger-Waldau a
a
European Commission, Joint Research Centre (JRC), Ispra, Italy
b
European Dynamics Belgium S.A, Belgium
c
National Renewable Energy Centre (CENER), Sarriguren, Spain
d
European Research Executive Agency (REA), Brussels, Belgium

A R T I C L E I N F O A B S T R A C T

Keywords: The urgency of meeting climate targets, increasing land use competition and falling solar photovoltaic (PV)
Transport infrastructure energy costs have created unprecedented opportunities for innovative deployment options. This study uses
Electricity generation geospatial data processing to quantify the potential for large-scale deployment of vertical solar panels along
Photovoltaic
Europe’s major roads and railways. Factors such as geography, environmental constraints, land use limitations,
Bifacial PV
TEN-T network
and techno-economic parameters, were carefully considered. An economic assessment is provided to take ac­
count of the costs and benefits and technical feasibility of the proposed PV system as compared to the fossil-fuel
based transport. The findings reveal a potential PV capacity of 403 GWp within the European Union (EU). This is
the equivalent to 55% of the EU’s total solar PV capacity target set for 2030. Using bifacial PV modules, these
systems could generate 391 TWh (terawatt hours) of clean electricity annually. Considering only railway lines,
the total annual PV electricity output could potentially reach 250% of the current annual electricity consumption
of the EU railway network. Additionally, the electricity generated from PV installations alongside roads would
not only be cost-effective in electricity markets but also serve as a viable alternative to fossil fuels in trans­
portation. Tapping solar PV energy along transport infrastructure can therefore significantly contribute to the
EU’s energy transition.

and restricted availability of agricultural land [3]. As a result, alterna­

1. Introduction tive PV deployment typologies are gaining ground, such as building
integrated PV, floating PV, and agrivoltaics to encourage the growth in
The United Nations (UN) Paris Agreement [1] has clearly recognized the sector even in land-limited contexts. The transport infrastructure
the urgent need for rapid decarbonisation to combat the adverse effects offers an additional avenue to accommodate PV systems in existing built
of climate change and the significant changes in the global energy sys­ areas. This study explores its potential at a pan-European scale.
tem and markets that need to take place. Its sustainable development The European Union (EU) Climate Law, in force since 2021, commits
Goal 7 focuses on affordable and clean energy as “key to the develop­ the EU to become climate-neutral by 2050. The 2022 REPowerEU
ment of agriculture, business, communications, education, healthcare communication calls for an acceleration of the 2030 goals for energy
and transportation” [2]. The long-term unsustainability of fossil fuels is transition [4] to achieve the EU’s targets. The accompanying EU Solar
compounded by security concerns and price instability related to Energy Strategy [5–7] aims to bring online more solar photovoltaic (PV)
geopolitical conflicts. installed capacity aiming to reach 384 GWp by 2025 and over 720 GWp
Solar photovoltaics (PV) currently stands as the cheapest source of by 2030 [5,6]. In 2023 the EU Council adopted the amended Renewable
electricity, suitable for deployment at different scales and indispensable Energy Directive (“RED III”) that sets a legally binding renewable energy
means of realizing global ambitions in climate action. At the same time, target of at least 42.5% by 2030 [8]. However, many scientific publi­
issues such as increasing land use competition, environmental and cations propose a more rapid electrification of the energy sector using
biodiversity impacts, water use, and indirect land use change emissions PV [9–12]. Market trends in the EU indicate that the total installed ca­
need to be addressed, in particular in regions of high population density pacity could reach 1 TWp by 2030, equivalent to 2246 Wp per capita [5,

* Corresponding author.
** Corresponding author.
E-mail addresses: georgia.kakoulaki@ec.europa.eu (G. Kakoulaki), sandor.szabo@ec.europa.eu (S. Szabo).

https://doi.org/10.1016/j.rser.2024.114344
Received 13 September 2023; Received in revised form 24 November 2023; Accepted 22 February 2024
Available online 8 March 2024
1364-0321/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Nomenclature TEN-T Trans – European Transport Network

STC Standard Test Conditions
Abbreviations WDPA World Database of Protected Areas
Agri-PV Agri-photovoltaics (agrivoltaics) WACC Weighted Average Cost of Capital
AYE Average Annual Energy Yield
AW Anticlockwise Notations/Symbols
API Application Programming Interface Et Electricity generation in year t
BEREC Body of European Regulator for Electronic Ft Fuel Expenditures in year t
Communications HT In-plane irradiation
BEV Battery-electric vehicle It Investment expenditure in year t
CW Clockwise Mt Operations and maintenance cost in year t
CS Case study n Investments period in years
EU European Union t Time in years
EV Electric vehicle Units
FF55 Fit for 55 package of proposals to revise EU legislation GWp Gigawatt peak power
GISCO Geographical Reference database of the European km Kilometre
Commission kWp Kilowatt peak power
JRC Joint Research Centre MW Megawatt
LCoE Levelized Cost of Energy MWp Megawatt peak power
MS Member State TW Terawatt
NUTS Nomenclature of territorial Units for Statics TWh Terawatt hour
OSM Open Street Map TWp Terawatt peak power
PV Photovoltaics WACC Weighted Average Cost of Capital
PVGIS Photovoltaic Geographical Information System Wp Watt peak power
RED III Renewable Energy Directive III Wp/capita Watt peak power per capita
SRTM-3 Shuttle Radar Topography Mission

13]. Even if the area requirement would be modest (1 TW of conven­ emphasizes the need to promote innovative forms of PV deployment. It
tional ground-mounted would require approximately 0.1% of total EU further supports initiatives at the sub-national level, where climate
land area), the availability of land for ground-mounted PV is limited by neutral ambitions are being translated into enhanced uptake of renew­
prioritisation of agricultural land for food production as well as nature ables. This is demonstrated well by the European 100 Climate-Neutral
and landscape preservation. In some cases, the distance from the elec­ and Smart Cities Mission aimed at delivering 100+ such cities by
tricity network infrastructure is the no-go factor. 2030. This unprecedented challenge would be greatly facilitated by
In the transport sector, the first flagship project in the EU Sustainable unveiling the solar photovoltaics potential of all surfaces cities can
and Smart Mobility Strategy [14] defines the enhancement of the leverage to increase their local renewable energy production [22],
penetration of zero-emission vehicles, renewable and low-carbon fuels, including transport infrastructure. A novel framework is proposed for
and related infrastructure as core targets to achieve the decarbonisation quantifying the energy that could be produced by vertical PV modules
of the sector. The EU’s 2021 Fit for 55 package (FF55) [15,16] proposed installed along EU roads and rails. This is applied at regional and na­
a ban on the sale of new CO2 emitting cars in Europe by 2035. This paves tional levels, taking into consideration environmental constraints, land
the way for full uptake of zero-emission vehicles, with electric vehicles use limitations and various techno-economic parameters. Advanced
(EV) expected to be the prominent technology [17]. In addition, it is geospatial techniques are used to identify areas where the maximum
foreseen to install passenger car recharging stations at least every 60 km benefit could be reached by using either bifacial or monofacial devices.
on TEN-T (Trans-European Network Transportation) main roads of at An economic assessment is proposed along considerations for technical
least 150 kW capacity by 2025 and for heavy-duty vehicles with a feasibility and optimisation, to promote a fully informed implementa­
minimum 350 kW capacity every 100 km by 2030 [18]. This implies tion and to unveil the cost competitiveness of vertical PV versus fossil
approximately 1040 stations, which would consume 680 GWh per year fuel-based transport. Overall, this study provides unique insights into
[15,19]. Currently the EU has over 307,000 publicly accessible charging how the transport and renewable energy generation sectors can be
points, concentrated primarily in the Netherlands (29.4%, 90,284 intertwined towards practical sustainability paradigms and carries
chargers), France (12.1%%, 37,128) and Germany (19.4 % 59,410), crucial policy implications to deliver on the European climate targets.
whereas 10 European countries do not have a single charging point per The paper is organized as follows: The background on PV for trans­
100 km of road (motorways, state, provincial and communal roads) port infrastructure and technology aspects are presented in Section 2.
[20]. The application of PV on or near the road infrastructure can help The data collection and methodology are described in Section 3, and
optimize the electrical grid for the charging of EVs. More advanced results presented in Section 4. Section 5 discusses the main findings of
options include inductive (wireless) charging infrastructure along the the analysis along with their interpretation and the conclusions are
carriageway itself which would also benefit from distributed power presented in Section 6.
generation using PV [21].
This study proposes the installation of PV systems along the EU 2. Background on PV for transport infrastructure
member states’ transport infrastructure, whose potential is largely un­
tapped, thereby aiding the decarbonisation of the transport sector, while Various schemes for integrating PV into transport networks have
simultaneously reducing the land competition issue. This is the first been proposed and put into practice, although only on a small scale.
systematic, pan-European investigation of the total PV potential asso­ Proposed PV mounting solutions include solar road surfaces [23,24],
ciated with roads and railways, considering different technologies and overhead PV (covering the road or railway line itself), or PV placed
configurations. It supports the EU’s Solar Energy Strategy, which between railway tracks (in the form of panels or PV sleepers) [25]. Solar

2
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Fig. 1. Case studies, workflow, and three exemplary road segments featuring different mean direction. A table with the properties of a road segment is presented as
obtained from publicly available data, for reference.

3
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

roads have multiple limitations [26]. Most of the proposed applications monofacial modules [43,44], yet they can absorb both direct light on the
involve placing solar panels above concrete slabs and covering them sun-facing side and ground reflected and diffuse solar radiation on the
with a transparent layer. Technical issues could include durability, rear side, leading to a higher overall power output [41,45]. In the
traction performance, and construction methods. More importantly, the literature, the actual gain achieved via bifacial modules varies among
high cost currently inhibits any large-scale uptake. Installing horizontal different case studies and depends on a range of factors. Depending on
PV between railway tracks has fewer downsides as the solar panels do the installation, bifacial modules can produce up to 20% more energy in
not interfere with the performance of the tracks. However, there can be side-by-side comparisons than equivalent monofacial modules as stated
issues regarding durability (panels are placed in a harsh environment, in Ref. [45]. In Refs. [46,47], bifacial PV modules showed a yield gain of
subjected to vibrations and air pressure loads from train transit) [27,28]. between 5% and 15% compared to monofacial devices due to ground
For these reasons, such highly integrated solutions are disregarded in albedo but, as the albedo depends on the surrounding area this should be
this study. Rather, the focus is on exploiting the areas along the sides of considered in the energy yield projections. For artificially treated
major roads and railways. Studies from around the world suggest that ground with an albedo of 50% (e.g., white concrete), the bifacial gains
these areas represent a huge untapped resource for PV installations, can be up to 30%. A recent literature review on roof-mounted PV further
including also existing and future noise barriers [29–32]. Noise reduc­ reveals that a 0.1 increment in albedo translates into +0.7% and +4.5%
tion measures along motorway and railway systems in built-up areas are energy production, for monofacial and bifacial modules respectively
required by law in EU countries, presenting an important opportunity [48]. In the case of vertical standalone configurations, bifacial modules
for PV systems. Existing studies focus primarily on specific stretches of yield more energy than monofacial panels anywhere on earth for
highways or regions and lack a Europe-wide perspective. albedos above 0.35 [49]. The present study assumes an albedo value of
In particular, PV on noise barriers (PVNB) was proposed in late 0.2, a value that is typical for aged concrete [50].
1980s and the first installation was completed in Switzerland in 1989
[33,34]. 3. Methodological approach
To the authors’ knowledge, there are now 38 PVNB installations in
Europe, ranging in size from 7.5 kW to 3 MW, with 20 in Germany and The analysis is based on open-access spatially explicit data. The
10 in Switzerland [35]. On a broader level, a study estimated the total workflow and datasets used are illustrated in Fig. 1 The PV energy
potential for PVNB to be 1145 MWp, in six EU countries [32,36]. In the output is obtained for three different case studies (CS). The first two CSs
Netherlands, the potential production from PVNB along national high­ (hereafter CS1, CS2) are idealized cases assuming either only monofacial
ways ranges between 5828–6938 MWh annually [37]. In Germany, a or bifacial technology respectively, while the third case study (CS3) is a
recent study estimates that noise protection structures along railway hybrid scenario combining both monofacial and bifacial PV depending
lines and trunk roads covering around 5800 km could generate on the direction of the road segment. This latter approach involves the
approximately 1412 GWh of solar power annually [38]. In 2023, use of monofacial modules for road segments with predominant east-
Switzerland awarded a tender for the development of solar energy along west (E-W) direction or axis (i.e., the module is south-facing) and bifa­
national roads to produce renewable energy from PV (overhead light cial modules for segments with north-south (N–S) direction or axis (i.e.,
weighted roof) to be primarily used for electric vehicles (EV) charging the module faces both east and west) as shown in Fig. 1.
infrastructure by 2027 [39]. Germany proposed to install PV arrays
along motorways and to create a vast smart grid by 2030, by using 3.1. Road and railway network datasets
steel-grid masts erected every 15 m along the sides of motorways [40].
PV modules would be installed on steel lattice girders hung between the The road network was derived from the OpenStreetMap project [51].
masts. Highway datasets for each Member State (MS) of the European Union
were acquired. A selection criterion was used to identify motorways,
2.1. PV technology aspects trunk roads, and primary roads, most suitable for PV installations. Sec­
ondary and tertiary roads were excluded due to their narrower width
This study focuses on vertically-mounted PV systems along major and limited space, as well as possible visibility issues for drivers. Sec­
road or rail lines, as it represents a technically straightforward approach tions with tunnels, bridges, and intersections were also discarded from
to PV integration in transport infrastructure. While the ideal PV instal­ the analysis. It is noted that there are no motorways in three states of the
lation typology along roads and rails is not yet clear, this enables a EU, namely Malta, Latvia, and Estonia.
conservative estimate of the potential using a standard and realizable The railway network was obtained from the Geographical Reference
technological option. Such systems can also serve as sound barriers and database of the European Commission – GISCO [52]. Underground lines,
in turn existing sounds barriers can be upgraded with PV modules, junctions, and stations were excluded from the analysis.
thereby providing additional functionality. Nevertheless, attaching The orientation of the modules is necessarily bound to follow the
panels to existing noise barriers may require new structural analysis and direction of the roads and railways, contrary to the situation for open-
certification. The sound reduction performance will also require field ground-mounted systems where, typically, PV modules can be
assessment in applications where sound energy is to be reflected, the optimally oriented. The analysis reflects this limitation, by considering
noise-absorbing layer can be omitted, and the acoustic properties are only orientation imposed by the road or rail direction at a given location.
related solely to the thickness of the transparent substrate (typically This was in turn determined by dividing the transport infrastructure into
made of glass), which forms the front surface of the module [31], and the 500m-long segments and estimating the mean direction of each
noise transmission characteristics of the frame used to secure the PV segment. The orientation of the corresponding PV structure was defined
modules in place. Where a noise-absorbing layer would be needed, the as perpendicular to the segment direction in both clockwise and anti­
PV module will have to be monofacial. clockwise senses. For monofacial devices, the front part of the PV
An important aspect of vertically mounted PV is that it lends itself module is assumed to be the side with highest AEY (facing the road or
naturally to the use of bifacial PV modules. In addition, vertically the exterior). For bifacial devices, the sides are oriented towards the
mounting is especially useful at high latitudes (i.e. above 45◦ ), where the road and to the exterior. The segment directions were then classified
solar altitude angle is typically low [41]. The choice between mono­ using eight-point cardinal directions (N, NE, E, SE, S, SW, W, NW).
focial or bifacial modules for such applications depends on several fac­ As a last step, a spatial overlay analysis with complementary datasets
tors such as the orientation and latitude of the system, ground albedo, was applied to provide supplementary information. In this process, the
and diffuse fraction, beyond the sound-barrier case mentioned earlier corresponding Nomenclature of Territorial Units for Statistics – NUTS
[42]. Bifacial PV modules are typically more costly than standard 2021 [53] classification at regional level was assigned to each discrete

4
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Table 1
Overall extent of the EU transport network considered in the study.
Network Suitability Area Type Network Type Length (km) % of Total (km)

Highways (motorways and primary roads) Inclusion Built-Up Regular 188,314 35.58 449,970 (85% of total roads)
TEN-T 26,576 5.02
Open Regular 199,305 37.66
TEN-T 35,775 6.76
Exclusion Protected Areas – 59,088 11.16 79,312
Tunnels, bridges junctions, other) 20,224 3.82
Railways Inclusion Residential Regular 31,112 17.14 152,082 (89% of total railways)
TEN-T 37,101 20.44
Non-residential Regular 43,218 23.8
TEN-T 40,721 22.43
Exclusion Protected Areas – 28,285 15.58 19,405
Others Underground lines 1120 0.61

segment. Additionally, the World Database on Protected Areas – WDPA

Table 2
[54] (supplementary material), built-up areas.
Parameters and values used in the LCOE model.
(supplementary material), and the Trans-European Transport
Network (TEN-T) [55] (supplementary material) were geographically Parameter Source/Value

matched. Table 1 provides a summary of the overall extent of the It PV system price, EUR/kWp 760 [67]
transportation network pertinent to the EU member states. r Discount rate (WACC for each country), Country specific, from BEREC’s
% report
3.2. PV electricity output n Investment period, years 30
Mt Annual O&M costs, % of initial capital 4.0% [68]
investment
The potential PV energy generation for each segment was made with Ft indicates the fuel expenditure in year (t) which is zero for photovoltaic
the free access Photovoltaic Geographical Information System (PVGIS) electricity
[56]. This online tool, developed at the European Commission Joint Et Annual electricity generation Location specific, From PVGIS 5.2
Research Centre (JRC), provides location-specific estimates of the PV
production of different PV technologies and system configurations,
have been simulated as two separate monofacial devices, with the same
using the PV electrical performance model developed by [57].
specifications as described, and applying a bifaciality factor of 90% (i.e.,
For this study, the coordinates of the midpoint of each 500 m
the rear side is assumed to be 90% as efficient as the front side). The
segment were used as an input o PVGIS via the API (Application Pro­
bifacial energy yield was obtained by summing that from the side or
gramming Interface) to estimate the average yearly total in-plane irra­
orientation with the highest average yearly irradiation values and that
diation (HT, kWh/m2) and the average yearly PV energy yield (AEY,
from the orientation with the lowest average yearly irradiation multi­
kWh) for the three PV system case study configurations. These estimates
plied by the bifaciality factor. In PVGIS, the “rear-side” irradiation is
are based on a complete series of hourly values the period 2005 to 2020,
influenced by the reflectance from the ground via the albedo factor. The
and take into consideration the surrounding topo graphical horizon
albedo is set to 0.2 in PVGIS, which is an appropriate nominal value for
which may block the direct irradiance at certain times of the day. The
this study [62,63].
horizon is defined by a Digital Elevation Model obtained from the
To validate this, a sensitivity analysis was performed applyin other
Shuttle Radar Topography Mission (SRTM-3) [58] with 90 m spatial
proposed methods for estimating bifacial energy output. The results
resolution, but does not include information on nearby obstacles such as
indicate that the three methods used in the sensitivity analysis have a
trees or buildings.
range of approximately 10% and the back-to-back approach used in this
For the monofacial PV technology, a crystalline silicon module was
study lie in between the other two (the details of this analysis are con­
assumed with dimensions 1 m × 2 m, efficiency of 22% and a nominal
tained in supplementary material). Furthermore, the more sophisticated
module peak power at Standard Test Conditions (STC) of 440 Wp. The
approach gives the highest estimated PV output values meaning that the
vertical height of the module mounting structure was assumed equal to
estimated AEY values obtained from the back-to-back simplified
4 m (as a typical noise barrier on highways has a height of 4.2 m [59])
approach are conservative, and that the selected method is sufficiently
with the first metre from the ground either used as a base of the PV
robust for the purposes of this study.
systems or left empty to reduce the impact of soiling, shading and stone
damage. The effective active area of PV is therefore 3 m2 per linear
metre. In addition to the 14% system losses considered in PVGIS, an 3.3. Economic assessment
extra soiling loss of 6% due to dirt from vehicular traffic was applied
based on previous studies [60,61]. The average AEY of the idealized The “Levelized Cost of Energy” (LCOE) [64] is the minimum price of
module was estimated for both clockwise (CW) and anticlockwise (AW) the electricity generated from a specific source to allow it to break even
orientation for each segment. The orientation with the highest value of with the capital and operating costs over the lifetime of the project
in-plane irradiation (HT) determined the favourable orientation (investment period). It is calculated as shown in Equation (1):
configuration for the monofacial PV system. ( )
For the bifacial PV technology, the nominal module peak power is It + Σnt=1 (1+r)
Mt

the same as for the monofacial module, since it refers only to front-side
t

LCOE = (1)
illumination. In operation a bifacial module should in principle also
Et
Σnt=1 (1+r)t

benefit from any rear-side illumination available, however there is a lack

of standardised methods for estimating the impact on annual energy where It is the investment expenditure in year (t), Mt represents the
yield. Also, there is no option for bifacial modules in the current version operations and maintenance (O&M) expenditure in year (t), Et is the
of PVGIS tool. Therefore, for the purpose of this study as PVGIS is the electricity generation in the year (t), r the discount rate (fixed for the
only detailed spatial tool available we had to use its monofacial simu­ lifetime period for each country), and the considered investment period
lation to estimate the bifacial output. To this end, he bifacial modules in years. The values used in this analysis are reported in Table 2.

5
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

The discount rate uses the Weighted Average Cost of Capital (WACC) kilometre of railways in the EU is approximately 0.45 GWh/km/yr for
reported by the Body of European Regulator for Electronic Communi­ CS1, 0.64 GWh/km/yr for CS2 and 0.61 GWh/km/yr for CS3.
cations (BEREC) in its 2021 report [65], and ranges between 2% and When the transport segments are grouped by overall orientation, the
13% among EU countries. The EU electricity prices and the transport largest for both roads and railways is east-west (E-W) with a total
vehicle km figures were derived by using data from Eurostat. In the installed capacities of 83 and 28 GWp respectively, followed by the
comparative analysis (LCoE for PV electricity and transport) between northwest-oriented roads with 73 GWp, see Table 4. CS2 (bifacial PV
the fossil fuel consumption and PV electricity-based consumption we only) shows the highest potential annual electricity generation per kil­
used the LCOE calculations from Ref. [66] which examines the energy ometre, up to 0.68 and 0.67 GWh/km for North-South (N–S) oriented
unit cost for PV-battery systems. roads and rails, followed by the Northeast-Southwest (NE-SW) and
Northwest-Southeast (NW-SE) oriented infrastructure. East-oriented
4. Results roads in CS2 have a higher generation output per kilometre than CS3,
whereas CS1 has lower values compared to CS2 and CS3. In the case of
4.1. Pan-European assessment railways, it is seen that for N–S orientated railways, CS2 and CS3 have
equal values of annual electricity generation per kilometre. Among
Looking first at the continental level, the estimated total energy different railway directions, E-W railways have the highest technical
production from PV systems installed along the EU’s transport infra­ potential of 28 GWp, but lower electricity generation per kilometre due
structure was examined for the three case studies (CS1, CS2 and CS3). to their higher total length compared to other directions.
The total potential installed capacity amounts to 403 GWp regardless The annual electricity output and technical installed capacity for the
of the PV module technology (monofacial or bifacial) since the nominal transport infrastructure outside built-up areas was estimated, with a
power capacity is that of the front side. The contribution of road in­ total of 183,000 km (30% of the total length, 116,000 km of roads and
stallations equals 300 GWp while the remaining 103 GWp comes from 67,000 km of railways), as shown in Table 5. In this scenario, the annual
installations along railways, as shown in Table 3. electricity generation for the road network is 53 TWh per year for
For the roadside systems, the total useable road length of motorways, monofacial modules, 75 TWh per year for bifacial modules, and 68 TWh
highways and primary roads in the EU is estimated to be 449,994 km. per year for the combination of both types. In terms of the railway
The annual electricity generation for CS1 (monofacial) is 211 TWh/yr, network, the utilization of bifacial devices could generate approximately
while for CS2 (bifacial) it increases to 294 TWh/yr. This latter value 51 TWh per year. This generated power is equivalent to 1.4 times the
corresponds to 12% of the net electricity consumption in 2022 (2495 current electricity consumption of the EU railway system. The potential
TWh [69]). The hybrid scenario, CS3, results in 242 TWh/yr (18% less installed capacity for the whole transport infrastructure is 122 GWp
than using only bifacial modules) [70]. The average annual energy po­ corresponding to a potential electricity generation of 119 TWh/yr, or
tential per kilometre in the EU is approximately 0.47 GWh/km/yr for 5% of the 2022 electricity consumption at EU level, see Table.
CS1, 0.65 GWh/km/yr for CS2 and 0.54 GWh/km/yr for CS3. In CS3, the Out of the 450,000 km of roads assessed, 115,500 km (26%) are part
length of roads considered suitable for monofacial PV installation is of the TEN-T network, see Table 5. The corresponding segments could
equal to 344,000 km, while 106,000 km would use bifacial PV. provide 42 GWp of PV installed capacity along roads. Moreover, the
For the railway network the total length studied is 152,000 km with a annual potential electricity generation along the TENT-T roads varies
potential installed capacity of approximately 100 GWp and an annual from 30 to 75 TWh depending on the chosen technology. Similarly, 51%
potential electricity generation varying from 69 TWh (CS1) to 97 TWh/ of the European railway network, amounting to 77,784 km, falls within
yr (CS2). CS3 achieves an intermediate value (93 TWh/yr). For both CS2 TEN-T. Along these railway tracks, 52 GWp of PV can be installed of­
and CS3, the potential PV output exceeds the current EU railway net­ fering the prospect of generating 51 TWh of electricity annually if
work’s consumption by a factor of 2.5 (36.5 TWh/yr in 2022, [71]) if bifacial modules are employed (approximately 71% than that with
using bifacial modules. The average annual energy potential per monofacial modules).

Table 3
Summary table of the annual potential energy output for the EU transport infrastructure using vertically mounted PV installation along roads and railways.
Technical potential for annual energy yield (TWh/yr)
Transport Infrastructure
Length (km*1000) Case study 1 monofacial Case study 2 bifacial Case study 3, hybrid Nominal Installed Capacity (GWp)

Road (motorways and major roads) 450 211 294 242 300
Mono Bif.
344 106
Railway 152 69 97 93 103
Mono Bif.
42 110
Total (Road þ Rail) 602 280 391 335 403

Table 4
Summary table of the PV output in TWh and potential installed capacity in GW as a function of the road or rail segment orientation. N–S: North –South, E-W: East- West,
NW-SE: Northwest-Southwest, NE-SW: Northeast-southwest.
Technical potential for annual energy yield

Segment Total Length of Segments Case study 1 (GWh/km), Case study 2 (GWh/km), Case study 3 (GWh/km), Installed
Orientation (km*1000) monofacial bifacial hybrid Capacity (GWp)

Road Rail Road Rail Road Rail Road Rail Road Rail

N–S 106 37 0.4 0.38 0.68 0.67 0.68 0.67 71 25

E-W 125 41 0.52 0.51 0.62 0.63 0.51 0.51 83 28
NW-SE 111 38 0.47 0.47 0.66 0.65 0.65 0.66 74 26
NE-SW 108 35 0.48 0.49 0.66 0.65 0.66 0.66 72 24

6
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Table 5
Summary table of potential PV energy yield on the TEN-T road and rail networks outside built-up areas.
Technical potential annual energy yield (TWh/yr)

Transport infrastructure Length (km*1000) Case study 1, monofacial Case study 2, bifacial Case study 3, Hybrid mono- Technical Installed
bifacial Installed Capacity
(GWp)

Roads outside built-up areas 116 (26% of total roads) 53 75 68 77

Railways outside built-up 67 (44% of total railways) 31 44 42 45
areas
Total outside built-up areas 183 84 119 110 122
Roads & TEN-T 62 (14% of total Roads) 30 42 40 42
Rails & TEN-T 78 (51% of total Rails) 36 51 49 52
Total 140 66 93 89 94
TEN-T

4.2. EU member state assessment observed for the railways as CS2 could produce 28% of PV output.
Fig. 3 compares the total potential electricity generation per country
Fig. 2 shows the annual potential electricity output and installed for roads and railways and the corresponding percentage it represents of
capacity along the transport infrastructure at the country level for each the current electricity consumption (2022). The CS2 scenario is used and
case study (panel (a) railways and (b) roads). Additionally, information the countries are ordered by the values for percentage of current elec­
about the total road or railway length and the percentage of the trans­ tricity consumption. It is noted that this metric is simply intended to
port infrastructure in the TEN-T network is included. The data is ordered illustrate the potential impact of this form of PV deployment, and that
based on the total length of transport infrastructure. future electricity consumption is likely to increase, not least due to the
France, Spain, Germany, and Italy are the top four countries with the electrification of road transport. Latvia stands out with a potential
longest road networks, and for the railways, Germany, France, Italy, and coverage exceeding 70%, and Lithuania, Portugal, Romania, Croatia,
Poland have the most extensive systems. As expected, these countries Hungary, Slovakia, Spain and Cyprus have the potential to cover more
have the highest potential for infrastructure electricity output and than 20% of their current electricity consumption. Spain, despite its high
installed capacity for all three case studies. Specifically, for the road electricity demand, can cover approximately 25% of its 2022 con­
network, CS2 yields the highest values of potential electricity output, sumption. France and Germany, with the highest annual demand, have
surpassing CS3 and CS1 by 19% (53 TWh) in the case of France, for lower percentages of electricity coverage, at 15% and 10% respectively.
example. For the railways, the difference between case studies is smaller Fig. 4 shows the potential coverage of 2022 per capita consumption,
but CS2 still has the highest electricity generation. again assuming that CS2 (bifacial PV) is implemented, along with the
Upon comparing the potential electricity production between CS2 corresponding average annual consumption in kWh per capita in 2022
and CS1, several member states could achieve an increase of over 30% [72]. Most of the MSs could significantly benefit from this form of PV
by employing bifacial modules instead of monofacial ones. The same is deployment, with coverage values of more than 10%. Remarkably,

Fig. 2. Annual potential electricity output and installed capacity along the transport infrastructure at the country level for each case study. The last column to the
right gives the percentage of the studied transport infrastructure in the TEN-T network.

7
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Fig. 3. Total annual potential electricity generation per country (blue bars) and potential coverage of current electricity consumption (2022) (yellow bars).

Cyprus and Latvia could achieve coverages exceeding 75%, while utilization of the generated electricity locally along road and rail sec­
Portugal and Croatia have the potential to cover current per capita tions. An analysis was conducted to estimate the feasibility of this
consumption by up to 30% and Hungary by up to 27%. approach and the self-sufficiency of transport on road sections through
the substitution of fossil fuels with PV-generated electricity. Since PV
4.3. EU regional assessment module arrays are installed near transport infrastructures, the cost of
electricity produced was compared not only to Member States’ retail
The potential for road and rail infrastrucure to generate PV elec­ electricity prices but also to that of fossil-fuel based transport options.
tricity at regional level (NUTS2) reveals significant variability, see Fig. 5 Additionally, the economic impact was evaluated in two ways, one
with CS2 as the basis for the analysis. The regions with highest potential based on electricity production capacity, and the other based on how
electricity generation are found in Portugal, France, and Spain, with much produced PV power can replace fossil fuels used in transport in the
values over 15 TWh/yr, while those exhibiting the lowest potential are identified road section.
located in Central, Northern, and Eastern Europe with values ranging Fig. 6 compares current retail electricity prices (which represent
from 0.02 to 5 TWh/yr. The average electricity generation per kilometre continuous supply and include transmission and distribution, but
of transport infrastructure ranges from 300 MWh/km (NL34, with AEY exclude taxes and levies) [52] in the EU MS with the average Levelised
300 kWh/yr) to 830 MWh/km (ES30, with AYE 710 kWh/yr). cost of electricity (LCoE) for roadside PV. On average, the cost of the
The Castilla-La Mancha region in Spain (NUTS2 code: ES41) shows latter cost is significantly lower (by 50%). However, this cost advantage
the highest energy output with an estimated annual generation of 33 is small in countries where retail prices do not reflect actual production
TWh/yr. This region has also a substantial technical potential installed costs due to direct or indirect subsidies (e.g., Hungary, Bulgaria,
capacity of 27.2 GW, with a combined length of 40,550 km for both Poland). Overall, the cost-effectiveness of the roadside PV electricity
railways and roads. The top five regions with the highest annual po­ will depend on a wide variety of factors, including the arrangements for
tential electricity generation are situated in Spain, Portugal, and France transmission to the points of use and the grid itself.
(NUTS2 code: ES41, ES42, ES61, PT16, FRK2). The cost assessment of the proposed roadside PV systems can also be
compared with the commercial PV system based electricity costs
4.4. Cost comparison of vertical PV and economic impacts throughout the European countries. The following map, Fig. 6, shows
the PV electricity costs that were calculated by using the PVGIS tool for
A potential benefit of roadside PV installations is the direct the electricity outputs and similar input parameters in the LCOE

8
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Fig. 4. Map of the potential percentage of annual kWh substitution per capita for EU27, if using bifacial PV modules along the transport infrastructure (roads &
rails). The numbers displayed over each country represent the current electricity consumption in kWh per capita (2022).

calculations as for the roadside PVs (CAPEX cost 768 EUR/kWp; 4% which has so far only produced modest results in its efforts to help
O&M yearly costs, 30 years as a lifetime, national discount rates from mitigate climate change.
BEREC report). It is evident that in the countries where the PV based Additionally, a comparison of the cost of fuel/100 km for the petrol
electricity is cheaper than the others, the roadside PV costs are also (e95) based transport and the PV electricity-based e-transport in each
cheaper. Beyond this similar trend from the maps it can also be derived EU Member States was performed. E95 fuel prices ranged from 1,28
in which part of the given country it could be more economic to locate EUR/l in Romania to 1,88 EUR/l in France and 1,95 EUR/l in Finland
standalone commercial PV sites (with good insolation properties) and (with the median of 1,58 EUR/l in Portugal) and its tax content was
where it is less favourable. excluded. Assuming fuel consumption values of 5 L E95/100 km and 20
Another aspect concerns the likely demand for electric vehicle kWh/100 km, the resulting PV-based fuel costs are on average two to
charging, which can be correlated with vehicle usage. Publicly available three times lower, and the differences can be seen in Fig. 9 for each MS.
statistics on the yearly average km/vehicle are only available for 14 However, since this calculation only includes the LCoE of the PV
countries (Fig. 7). Using the latest data for these countries, the per­ produced electricity, a range of the vehicle electricity charging costs
centage of vehicles kilometres that could be substituted with the elec­ based on a system cost calculation [66] that includes batteries (assuming
tricity produced from roadside vertical PV was estimated. For this, an EV LiFePO type batteries replaced every 9 years) was provided (optimal
consumption of 20 kWh/100 km [73]. Only the TEN-T road sections South facing PV systems were considered). From the system options, the
were considered since these will have a systematically distributed LCOE of the 2-system configuration that is consistent with the roadside
charging stations (Fig. 8). Losses due to storage and other trans­ PV of our analysis (South oriented fixed PV & LiFePO, East-West ori­
formations were not factored in. ented fixed PV & LiFePO) was used. Using these values, the PV based
On this basis, the percentage of overall transport km powered by option is still more competitive than the fossil fuel-based transport, and
traditional transport fuels that could be substituted ranges from 0.5% in as the battery costs are expected to decrease in the future through
the Netherlands to 15% in Spain. However, when the full range of road technological learning its economics will be further improved [74]. The
sections is considered, the substitution potential increases at least three- sizing of the battery array is highly dependent on the road section pa­
fold. This indicates the potential to help electrify the transport sector, rameters (size, time distribution of traffic, vehicle composition) that can

9
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Fig. 5. Annual technical potential electricity generation for roads and railway infrastructure at regional level (NUTS-2) across the EU27 for CS2 (bifacial PV).

be calculated on the local sections. further studies are needed to investigate in detail key technical, socio-
economic and environmental aspects of the vertical PV systems that
5. Discussion were simulated in this study. For instance, on certain road segments,
reflections from the modules may result in glare and safety issues for
A major finding in this study is that the potential installed capacity drivers. The structural design of the support structures for vertical PV
associated to the mass scale installation of PV along the transport also needs to be examined to ensure that they are a sustainable solution,
infrastructure amounts at 403 GWp without requiring additional land as well as cost-optimized and resistant to wind loads. Additionally, po­
usage, and corresponds to an estimated production of 391 TWh annu­ tential issues such as physical stress from stone chipping or increased
ally. Notably, these values correspond to approximately 55% (425 GWp) soiling because of suspended particles, often electrostatically charged,
of the 2030 target of total installed capacity of solar PV set by the EU should be evaluated. Appropriate solutions are needed for connecting
Solar Strategy, and 16% of the 2022 electricity consumption in the EU. PV arrays running over several kilometres and for leveraging and uti­
For the TEN-T railway network, the PV potential energy generation lizing connections to existing electricity grid lines and to EV charging
reaches 51 TWh annually, equivalent to 1.4 times the current EU railway points. Future analysis of the positive regional and local air quality
consumption. Comparatively, using bifacial devices along the entire impacts of fuel substitutions along the identified roads will also be
studied railway network could result in an estimated PV output that advantageous.
exceeds the current EU railway network’s consumption by 2.5 times. The detailed geospatial analysis used in the study serves several
For the case of bifacial modules, a factor that has not been taken into additional purposes. The datasets (publicly available) can help identify
consideration in the present study is that the modules must be mounted locations with high-quality solar resources and proximity to demand and
in a way that avoids shading from the support structure on the rear side. existing infrastructure. Additionally, the estimates of this data-driven
This would mean in practice that the module area per linear km might be approach can serve as a basis for planning future PV installation pro­
less for a given system height. This would not be the case for the mon­ jects and help stakeholders to develop national or regional energy
ofacial modules and could be a factor to consider in the choice of module infrastructure plans, considering also local regulations and conditions (i.
type where the bifacial gain is extremely limited. e. on roadside barrier heights, technology options and restrictions and
It is recognized that, before such systems could be widely realised, traffic-related data).

10
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Fig. 6. Levelised cost of electricity estimated at pixel level (100 m by 100 m) for the commercial use using optimal inclination.

Furthermore, national charging-point planning could be refined by while in this study it is 0.22 kWp/m2. More importantly [75], focused
incorporating the estimated PV electricity potential alongside the road only on noise barriers while our analysis was extended for longer road
infrastructures to supplement traffic related data. By identifying suitable segments. Another study that focused on PV installations on existing
locations for solar installations, charging stations for EVs can be stra­ sound barriers [34] estimated a technical potential capacity of 800 MWp
tegically placed to meet the growing demand for EV charging infra­ generating 680 GWh of electricity annually for Germany, France, Italy,
structure. Currently, most EV charging infrastructure is concentrated in Netherlands, Switzerland, and United Kingdom. Fig. 10a presents the
cities, at private residential or commercial buildings. Therefore, there is mapped annual potential electricity generation per MS and the corre­
an opportunity to promote sustainable travelling and to have a sym­ sponding technical installed capacity. Our analysis reveals that France,
metrical distribution and share of chargers across MS, while at same Germany and Italy have significantly larger potential, with a total ca­
time the EV can serve as energy storage. Additionally, EV charging is pacity of 162 GWp, and a potential electricity generation of 160 TWh,
considered a major instrument of demand flexibility, as shifting which is 160 times the estimated electricity generation from PV on
charging operations can reduce peak loads and thus reduce the need for existing noise barriers. This highlights the significant potential of ver­
grid expansion. In addition, as EVs become more common, incorpo­ tical PV installed along transport infrastructure when compared to other
rating solar power into charging stations can help reduce carbon emis­ PV applications like noise barriers. The lack of publicly available
sions and enhance the sustainability of transportation. geo-located data on existing noise barriers and adequate information on
On average, this study finds that the annual energy potential per their characteristics cannot confirm the potential annual electricity
kilometre of the EU’s transport infrastructure is approximately 0.65 generation of installing PV on them. This study can support studies for
GWh/km/yr for bifacial PV. This is higher than a previous study [75] new suitable locations for PV noise barriers or the development of EV
which combined PV and sound barriers and estimated an average energy charging stations, as the provided dataset is geo-located and provides
potential of 0.2 GWh/km. One reason for this difference is that this study information on the potential annual electricity output on specific
assumes a higher module efficiency in line with current products; the location.
power output per unit area used in Ref. [70] was equal to 0.14 kWp/m2, The results of this study can support the Sustainable and Smart

11
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Fig. 7. Comparison of electricity costs (excluding taxes and levies) and the Levelised cost of PV electricity (LCoE) in EUR/kWh for each MS.

Mobility Strategy and pave the way to drastic reductions in transport infrastructure if bifacial PV are used. It observed that the countries
emissions up to 90% by 2050, in compliance with the European Green where the cost of finance is lower, the LCoE is favourable for the PV
Deal. As seen in Fig. 10c, the potential total carbon savings in the EU was installation, such as Spin, Austria, Poland, Slovakia, Hungary and
estimated and it was found that, on average, a 16% annual reduction can Slovenia. Czech Republic, Belgium, France, Italy and Portugal have good
be reached by installing bifacial modules along the transport infra­ solar resources and favourable financing conditions; therefore, they are
structure. The Member States with highest carbon saving potential are at the lower end of LCoE range.
Germany, Italy, and Poland, which are the largest producers of carbon This estimation can be calculated in more precise way if local data
dioxide and could reduce their carbon intensity by 8–11%. Beyond like as connection to the grid, local regulation of safety are known. In the
Sweden, which exhibits already the lowest carbon intensity, the coun­ scenario where the fossil fuel substitution is considered, high resolution
tries with the highest CO2 emission reductions potential are Latvia temporal and spatial datasets on the electrified mobility, the EV char­
(36%), Lithuania (25%), Romania (24%) and Portugal (21%). The gers’ geolocation and road traffic data are necessary, to have a more
estimated potential installed capacity along the TEN-T network is equal precise calculation. Currently, 20% of the European railways transport is
to 42 GWp with a potential to generate 42 TWh/yr. This energy gen­ fuelled by diesel [76]. Thus, installing PV modules along the railway
eration can play a significant role in supporting the deployment of infrastructure may help to accelerate to power 100% electricity. This
electrical chargers. Assuming that one EV chargers should be install can meet air quality targets but also the decarbonisation of the sector.
every 60 km of TEN-T road network (62,000 km in this study), 1033 EV
chargers should be installed. The benchmark of 150 kW output sug­ 6. Conclusions
gested by the new regulation for the deployment of alternative fuels was
considered. If the average use of EV chargers is 12 h per day a total of 4 A methodology has been developed to estimate the technical po­
GWh/yr, is needed. Based on our estimation, the TEN-T network could tential of large-scale installation of PV along the EU’s transport infra­
support the installation of 280 thousand fast electrical chargers which structure at national and regional level. This provides the basis for
could supply an ample charging infrastructure along the transport routes quantitatively analysing the possible impact of such solar PV energy
and make electric vehicles a more viable option for long-distance travel. generation. Furthermore, it can provide support to planning an accel­
Fig. 10b shows the LCoE for PV application for the whole transport eration of decarbonisation efforts at national and regional levels, while

12
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Fig. 8. Percentage of current road transport that could be powered by PV installed on TEN-T road network (for countries where km/data is available).

• Cyprus and Latvia stand out as the biggest potential beneficiaries in

relative terms, with the potential to cover 75% of their current
electricity consumption per capita.
• Road-side PV could directly supply the network of e-charging sta­
tions foreseen on the TEN-T network and contribute to the substi­
tution of traditional fossil-based transport fuels in Europe. For some
countries this could amount to up to 15% of vehicle kilometres
travelled.
• the PV-based fuel costs are on average two to three times lower than
the traditional petrol-based transport.
• The average annual potential of electricity generation along the EU
railway network is 97 TWh, or 2.5 times the current consumption of
the railway system.

Lastly, the results of this study on PV energy supply demonstrate the

potential for creating synergies between PV energy generation and its
use, and further studies could, for instance, develop geospatial profiling
of the future demand patterns charging stations on the TEN-T network
Fig. 9. Comparison of the fuel costs of transport (EUR/100 Km) based on E95 for scenarios with a high proportion of EVs in road traffic. This, com­
petrol, electricity from PV only and PV battery combinations for each MS. bined with geospatial time-series data on roadside PV generation
developed here could lead to a more refined understanding of how de­
reducing land use competition and enabling a significant emissions mand and supply could be balanced in the charging network and the
reduction in the transport sector, which has historically proven chal­ consequences for investment policies and market related measures.
lenging to decarbonize.
The results of this research highlight that: Author contributions

• The potential power capacity of vertical PV systems along EU Conceptualization, S. S, N. T, A. JW., R.K, G.K. Methodology, G.K. F.
transport infrastructure is 403 GWp (300 and 103 GWp for roads and F, S. S, A.G. Data curation, G.K. and F.F.; Investigation, G.K, A.G., S. S.
rail respectively), with a corresponding annual electricity generation Visualization, G.K. and F.F.; Writing – review & editing, G.K, N.T, S.S, R.
varying from 280 to 391 TWh, depending on the PV technology K, A.G, A. JW, K.G, U.G, A.C.
employed.

13
 G. Kakoulaki et al.
14

Renewable and Sustainable Energy Reviews 196 (2024) 114344

Fig. 10. Maps for the entire studied infrastructure for the CS2 case (bifacial) per member state a) Colormap depicting the annual electricity generation (TWh/yr) labelled with the corresponding technical installed
capacity (GWp) b) Colormap displaying the average Levelized Cost of Electricity c) Colormap illustrating the decrease of the carbon intensity labelled with the corresponding average carbon intensity of the last 6 years
(Gt CO2 eq/kWh).
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

Disclaimer 28(6):439-453. Available from: https://onlinelibrary.wiley.com/doi/full/10.

1002/pip.3189.
[14] Pape M. Briefing eprs | European Parliamentary Research Service.
The views expressed are purely those of the authors and may not in [15] Fit for 55: transport MEPs want car-recharging stations every 60 km | news |
any circumstances be regarded as stating an official position of the Eu­ European parliament [internet]. https://www.europarl.europa.eu/news/en/press
ropean Commission. -room/20221003IPR42118/fit-for-55-transport-meps-want-car-recharging-stations
-every-60-km; 2023.
[16] MEPs adopt new rules for more charging stations and greener maritime fuels |
Declaration of competing interest News | European Parliament [Internet]. https://www.europarl.europa.eu/news/en
/press-room/20230707IPR02419/meps-adopt-new-rules-for-more-charging-stat
ions-and-greener-maritime-fuels; 2023.
The authors declare the following financial interests/personal re­ [17] Electric vehicles - IEA [Internet] 2023. https://www.iea.org/energy-system/tran
lationships which may be considered as potential competing interests: sport/electric-vehicles.
[18] Alternative fuels infrastructure: Council adopts new law for more recharging and
All the authors working for the JRC reports financial support was pro­ refuelling stations across Europe - consilium [Internet]. https://www.consilium.
vided by European Commission Joint Research Centre Ispra. europa.eu/en/press/press-releases/2023/07/25/alternative-fuels-infrastructure-
council-adopts-new-law-for-more-recharging-and-refuelling-stations-across-e
urope/; 2023.
Data availability
[19] New Regulation for the deployment of alternative fuels (AFIR) [Internet] Available
from: https://ec.europa.eu/commission/presscorner/detail/en/ip_23_1867.
The aggragation at national-level data can be downlaoded from [20] Enabling factors for alternatively-powered cars and vans in the European Union.
2022 [cited 2023 Nov 22], https://www.acea.auto/files/Research-Whitepaper-A-E
https://doi.org/10.6084/m9.figshare.24088590.v1. The geopackage
uropean-EV-Charging-Infrastructure-Masterplan.pdf.
containing the road and railway networks along with corresponding PV [21] In-road inductive charging tests demonstrate unlimited EV range [Internet]. htt
electricity generation data will be made available upon request. ps://newatlas.com/automotive/stellantis-road-charging-induction/; 2023.
[22] Ulpiani G, Vetters N, Shtjefni D, Kakoulaki G, Taylor N. Let’s hear it from the cities:
on the role of renewable energy in reaching climate neutrality in urban Europe.
Acknowledgments Renew Sustain Ener Rev 2023;183:113444.
[23] Ferri C, Ziar H, Tin Nguyen T, Van Lint H, Zeman M, Isabella O. Mapping the
The authors gratefully acknowledge the valuable comments and photovoltaic potential of the roads including the effect of traffic. 2021 [cited 2023
Sep 5]; Available from: http://creativecommons.org/licenses/by/4.0/.
suggestions from Christian Thiel and Theodoros Grigoratos, as well as [24] Dezfooli AS, Nejad FM, Zakeri H, Kazemifard S. Solar pavement: a new emerging
Teodora Lyubenova for the first discussions on bifacial PV modules. This technology. Sol Energy 2017 Jun 1;149:272–84.
study was performed under the European Commission Joint Research [25] Padmakala S, Alkawak OA, Reddy GVS, Narendranathan SK, Muthuraman MS,
Sendilvelan S. A train with automatic functioning based on IoT with solar energy
Centre Annual Work Programme 2022/2023. Fernando Fahl is a source. IDCIoT 2023 - int conf intell data commun technol internet things. SAVE
contractor working for the Joint Research Centre. Proc 2023:908–13.
[26] Zhou Bochao, Pei Jianzhong, Xue Bin, Guo Fucheng, Wen Yong, Zhang Jiupeng.
Rui Li,Solar/road from ‘forced coexistence’ to ‘harmonious symbiosis. Appl Energy
Appendix A. Supplementary data 2019;255:113808.
[27] Maghami MR, Hizam H, Gomes C, Radzi MA, Rezadad MI, Hajighorbani S. Power
Supplementary data to this article can be found online at https://doi. loss due to soiling on solar panel: a review. Renew Sustain Energy Rev 2016 Jun 1;
59:1307–16.
org/10.1016/j.rser.2024.114344. [28] Hassan S, Dhimish M. Review of current state-of-the-art research on photovoltaic
soiling, anti-reflective coating, and solar roads deployment supported by a pilot
References experiment on a PV road. Energies 2022;15:9620 [cited 2023 Sep 5];15(24):9620.
Available from: https://www.mdpi.com/1996-1073/15/24/9620/htm.
[29] Hasmaden F., Zorer Gedik G., Yüğrük Akdağ N. An approach to the design of
[1] The Paris Agreement - publication | UNFCCC [internet]. https://unfccc.in
photovoltaic noise barriers and a case study from Istanbul, Turkey. Environ Sci
t/documents/184656; 2023.
Pollut Res Int 2022. 2023;29:33609-36626. Available from: https://link.springer.
[2] Energy - united Nations sustainable development [internet]. https://www.un.org/s
com/article/10.1007/s11356-022-18625-0.
ustainabledevelopment/energy/; 2023.
[30] Zhong T, Zhang K, Chen M, Wang Y, Zhu R, Zhang Z, et al. Assessment of solar
[3] van de Ven DJ, Capellan-Peréz I, Arto I, Cazcarro I, de Castro C, Patel P, et al. The
photovoltaic potentials on urban noise barriers using street-view imagery. Renew
potential land requirements and related land use change emissions of solar energy.
Energy 2021 May 1;168:181–94.
Sci Rep 2021;111 [cited 2022 Feb 4];11(1):1–12, https://www.nature.co
[31] Gu M, Liu Y, Yang J, Peng L, Zhao C, Yang Z, et al. Estimation of environmental
m/articles/s41598-021-82042-5.
effect of PVNB installed along a metro line in China. Renew Energy 2012 Sep 1;45:
[4] REPowerEU [internet]. https://ec.europa.eu/commission/presscorner/detail/en/I
237–44.
P_22_3131; 2023.
[32] Nordmann T, Froelich A, Goetzberger A, Kleiss G, Hille G, Reise C, et al. The
[5] Chatzipanagi A, Jäger-Waldau A. The European solar communication—will it pave
potential of PV noise barrier technology in Europe. Sixt Eur Photovolt Sol Energy
the road to achieve 1 TW of photovoltaic system capacity in the European union by
Conf 2020 Oct 30:2912–6.
2030? Sustain Times 2023;15:6531 [cited 2023 Sep 5];15(8):6531, https://www.
[33] Clavadetscher L, Nordmann T. 100 kWp gridconnected PV plant A13 in
mdpi.com/2071-1050/15/8/6531/htm.
Switzerland - 10 Years and 1’000’000 kWh later. Sixt Eur Photovolt Sol Energy
[6] EUR-lex - 52022DC0221 - EN - EUR-lex [internet]. https://eur-lex.europa.eu/legal-
Conf [Internet] 2020 [cited 2023 Sep 5]; Pagesç 2249–52, eBook
content/EN/TXT/?uri=COM%3A2022%3A221%3AFIN&qid=1653034500503;
ISBN9781315074405. Available from: https://www.taylorfrancis.com/chapter
2023.
s/edit/10.4324/9781315074405-47/100-kwp-gridconnected-pv-plant-a13-swi
[7] Breyer C, Bogdanov D, Ram M, Khalili S, Vartiainen E, Moser D, et al. Reflecting
tzerland-10-years-1-000-000-kwh-later-luzi-clavadetscher-thomas-nordmann.
the energy transition from a European perspective and in the global
[34] Nordmann T, Clavadetscher L. PV on noise barriers. Prog Photovoltaics Res Appl.
context—relevance of solar photovoltaics benchmarking two ambitious scenarios.
12(6):485–495.
Prog Photovolt. Res. Appl. 2022;31(12):1369–95. https://onlinelibrary.wiley.com/
[35] Photovoltaic noise barriers [internet]. 2023. Available from: https://www.
doi/full/10.1002/pip.3659.
pvresources.com/en/pvpowerplants/noisebarriers.php#notes.
[8] Renewable energy targets [Internet]. 2023. Available from: https://energy.ec.
[36] (17) (PDF) Evaluation of the potential of PV noise barrier technology for electricity
europa.eu/topics/renewable-energy/renewable-energy-directive-targets-and-
production and market share, Final Report [Internet]. 2023. Available from: http
rules/renewable-energy-targets_en.
s://www.researchgate.net/publication/264011205_Evaluation_of_the_potential_of_
[9] Dutta R, Chanda K, Maity R. Future of solar energy potential in a changing climate
PV_noise_barrier_technology_for_electricity_production_and_market_share_Final_
across the world: a CMIP6 multi-model ensemble analysis. Renew Ener 2022;188:
Report.
819–29.
[37] Meppelink S. The potential of photovoltaics along the Dutch national high-and
[10] Haegel NM, Verlinden P, Victoria M, Altermatt P, Atwater H, Barnes T, et al.
expressways. Rijkswegen; 2015.
Photovoltaics at multi-terawatt scale: waiting is not an option. Science 2023;80.
[38] Cullmann J, Dilley M, Egerton P, Grasso VF, Honoré C, Lúcio F, et al. State of
[11] Kougias I, Taylor N, Kakoulaki G, Jäger-Waldau A. The role of photovoltaics for the
climate services. 2022. p. 11–21. 2022.
European Green Deal and the recovery plan. Renew Sustain Ener. Rev. 2021;144:
[39] 35 MW solar capacity approved along Swiss highways - TaiyangNews [internet].
111017.
2023. Available from: https://taiyangnews.info/markets/35-mw-solar-capacity-a
[12] Bogdanov D, Ram M, Aghahosseini A, Gulagi A, Oyewo AS, Child M, et al. Low-cost
pproved-along-swiss-highways/.
renewable electricity as the key driver of the global energy transition towards
[40] German highway PV could generate up to 200 TWh a year – pv magazine
sustainability. Energy 2021;227:120467.
International [Internet]. 2023. Available from: https://www.pv-magazine.com/20
[13] Vartiainen E., Masson G., Breyer C., Moser D., Román Medina E. Impact of
23/04/11/german-highway-pv-could-generate-up-to-200-twh-a-year/.
weighted average cost of capital, capital expenditure, and other parameters on
future utility-scale PV levelised cost of electricity. Prog. Photovolt. Res. Appl. 2020;

15
G. Kakoulaki et al. Renewable and Sustainable Energy Reviews 196 (2024) 114344

[41] Jouttijärvi S, Lobaccaro G, Kamppinen A, Miettunen K. Benefits of bifacial solar [58] Farr T.G., Rosen P.A., Caro E., Crippen R., Duren R., Hensley S., et al. The shuttle
cells combined with low voltage power grids at high latitudes. 2022 [cited 2023 radar topography mission. Rev Geophys 2007;45(2). Available from: https://doi.
Sep 5]; Available from: http://creativecommons.org/licenses/by/4.0/. org/10.1029/2005RG000183.
[42] Stein J, Reise C, Castro JB, Friesen G, Maugeri G, Urrejola E, et al. Bifacial [59] CEDR Technical Report 2017-02 State of the art in managing road traffic noise:
photovoltaic modules and systems: experience and results from international noise barriers ISBN: 979-10-93321-27-1. January 2017. Available from: https
research and pilot applications. 2021 Apr 1 [cited 2023 Nov 24]; Available from: ://www.cedr.eu/download/Publications/2017/CEDR-TR2017-02-noise-barriers.
https://www.osti.gov/servlets/purl/1779379/. pdf.
[43] Manikandan MV. Performance Evaluation of Bifacial and Monofacial modules in [60] Kimber A. The effect of soiling on large grid-connected photovoltaic systems in
vertical and latitude mounting at South India using PVsyst. IOP Conf Ser Mater Sci California and the southwest region of the United States.
Eng 2020;912(4):042066. [61] Wadhawan SR, Pearce JM. Power and energy potential of mass-scale photovoltaic
[44] Kopecek R, Libal J. Bifacial photovoltaics 2021: status, opportunities and noise barrier deployment: a case study for the U.S. Renew Sustain Energy Rev 2017
challenges. Energies 2021;14(8):2076. https://doi.org/10.3390/en14082076. Dec 1;80:125–32.
[45] Guo S, Walsh TM, Peters M. Vertically mounted bifacial photovoltaic modules: a [62] Albedo – PV performance modeling collaborative (PVPMC) [internet]. 2023.
global analysis. Energy 2013 Nov 1;61:447–54. Available from: https://pvpmc.sandia.gov/modeling-steps/1-weather-design-inp
[46] Vertical bifacial photovoltaic systems – innovative applications with enormous uts/plane-of-array-poa-irradiance/calculating-poa-irradiance/poa-ground-reflecte
potential for agriculture | Rödl & Partner [Internet]. 2023. Available from: https:// d/albedo/.
www.roedl.com/insights/renewable-energy/2020-08/vertical-bifacialpv-pote [63] Quantifying pavement albedo [internet]. 2023. https://eng.auburn.edu/research/
ntial-agriculture. centers/ncat/newsroom/2016-fall/pavement-albedo.html.
[47] Mouhib E, Rodrigo PM, Micheli L, Fernández EF, Almonacid F. Quantifying the [64] Badouard T, Moreira De Oliveira D, Yearwood J, Torres P. Rotterdam, 31 July
rear and front long-term spectral impact on bifacial photovoltaic modules. Sol 2020 Client: European Commission-DG Energy A4 Final Report-Cost of Energy
Energy [Internet] 2022;247:202–13. https://doi.org/10.1016/j. (LCOE) Study on energy costs, taxes and the impact of government interventions
solener.2022.10.035 [cited 2023 Sep 5]. on investments in the energy sector. 2020.
[48] Vasilakopoulou K., Ulpiani G., Khan A., Synnefa A., Santamouris M. Cool roofs [65] BEREC Report on WACC parameter calculations according to the European
boost the energy production of photovoltaics: investigating the impact of roof Commission’s WACC Notice of 6 November 2019. 2019.
albedo on the energy performance of monofacial and bifacial photovoltaic [66] Kosmadakis IE, Elmasides C, Koulinas G, Tsagarakis KP. Energy unit cost
modules. Sol Energy 2023;265:111948. Available from: https://linkinghub.else assessment of six photovoltaic-battery configurations. Renew Energy 2021 Aug 1;
vier.com/retrieve/pii/S0038092X23005819. 173:24–41.
[49] Khan MR, Sakr E, Sun X, Bermel P, Alam MA. Ground sculpting to enhance energy [67] Renewable IRNEA. Power generation costs in 2021. International Renewable
yield of vertical bifacial solar farms. Appl Energy 2019 May 1;241:592–8. Energy Agency; 2022.
[50] (17) (PDF) impact of ground albedo on the performance of PV systems and its [68] Steffen B, Beuse M, Tautorat P, Schmidt TS. Experience curves for operations and
economic analysis [internet]. 2023. Available from: https://www.researchgate.net maintenance costs of renewable energy technologies. Joule 2020;4(2):359–75.
/publication/277258834_Impact_of_Ground_Albedo_on_the_Performance_of_PV_ [69] Statistics | eurostat [internet]. https://ec.europa.eu/eurostat/databrowser/view/
Systems_and_its_economic_analysis. NRG_CB_E/default/table?lang=en&category=nrg.nrg_quant.nrg_quanta.nrg_cb;
[51] Highways - OpenStreetMap wiki [internet]. https://wiki.openstreetmap.org/wiki 2023.
/Highways; 2023. [70] Eurostat - data explorer [internet]. https://appsso.eurostat.ec.europa.eu/nui/sh
[52] Geographical information system of the commission (GISCO) - statistics explained ow.do?dataset=nrg_ind_peh&lang=en; 2022.
[internet]. https://ec.europa.eu/eurostat/statistics-explained/index.php? [71] Metrology for smart energy management in electric railway systems - EURAMET
title=Geographical_information_system_of_the_Commission_(GISCO); 2023. [Internet]. https://www.euramet.org/research-innovation/search-research
[53] Background - NUTS - nomenclature of territorial units for statistics - eurostat -projects/details/?tx_eurametctcp_project[project]=1484; 2023.
[internet]. https://ec.europa.eu/eurostat/web/nuts/background; 2023. [72] EUROSTAT. Simplified energy balances [Internet] Available from: https://ec.europ
[54] UNEP-WCMC and IUCN. Protected planet: the world database on protected areas a.eu/eurostat/databrowser/view/nrg_bal_s/default/table?lang=en.
(WDPA) [internet]. www.protectedplanet.net; 2023. [73] No Title [Internet] Available from: https://evbox.com/en/ev-home-charger-electr
[55] Trans-European transport network (TEN-T) [internet]. https://transport.ec.europa. icity-usage%0Ahttps://ev-database.org/cheatsheet/energy-consumption-electric-c
eu/transport-themes/infrastructure-and-investment/trans-european-transport-net ar%0Ahttps://www.sciencedirect.com/science/article/pii/S0301421521006121%
work-ten-t_en; 2023. 0A.
[56] Photovoltaic geographical information system (PVGIS) [internet]. https://joint-res [74] Utility-scale battery storage | electricity | 2022 | ATB | NREL [internet]. https://
earch-centre.ec.europa.eu/photovoltaic-geographical-information-system-pvgis atb.nrel.gov/electricity/2022/utility-scale_battery_storage; 2023.
_en; 2023. [75] Perez M, Perez R, Ferguson CR, Schlemmer J. Deploying effectively dispatchable
[57] Huld T, Gracia Amillo AM. Estimating PV module performance over large PV on reservoirs: comparing floating PV to other renewable technologies. 2018.
geographical regions: the role of irradiance, air temperature, wind speed and solar https://doi.org/10.1016/j.solener.2018.08.088 [cited 2022 Feb 4].
spectrum. Energies 2015;8(6):5159–81. [76] UNIFE, association TERSI. Diesel traction - UNIFE [internet]. https://www.unife.
org/activities/environment-and-sustainability/diesel-traction/; 2023.

16