Bottom-up Analysis of the Solar Photovoltaic Potential

for a City in the Netherlands

A Working Model for Calculating the Potential

using High Resolution LiDAR Data

B. B. Kausika

, O. Dolla

, W. Folkerts

, B. Siebenga

, P. Hermans

and W. G. J. H. M. van Sark

Copernicus Institute of Sustainable Development, Utrecht University,

Heidelberglaan 2, 3584 CS Utrecht, The Netherlands

Solar Energy Application Centre (SEAC), High Tech Campus 21, Eindhoven, The Netherlands

I-Real, Stationsweg 30, Terborg, The Netherlands

Aurum Europe, Zandsteen 6, Hoofddorp, The Netherlands

Keywords: Solar Photovoltaic Potential, LiDAR Data, Bottom-up Approach.

Abstract: This paper presents a working model to estimate the solar photovoltaic potential using high- resolution

LiDAR data and Geographic Information Systems. This bottom-up approach method has been selected to

arrive at the potential as this gives a better estimate than a top-down approach. The novelty of the study lies

in estimating the potential at high resolution and classifying the rooftop as suitable or not for solar

photovoltaic installations based on factors like irradiation, slope and orientation. The city of Apeldoorn in

the Netherlands has been selected as the study area. The model was able to successfully locate suitable sites

for photovoltaic installations at rooftop level. In addition, the area feasible for the installations and the

potential power output has also been calculated. We conclude that the city has a potential of 319 MWp

capacity, which would yield 283.9 GWh/yr in relation to the 304 GWh/yr consumption from residential

buildings in the area.

1 INTRODUCTION

Photovoltaic (PV) solar energy in Europe has been

increasing rapidly in the past few years.

Technological developments and research efforts

have brought PV in the renewable energy sector to a

new level. Estimation of the actual potential of PV

in the residential sector creates various business

opportunities and would assist in policy making. In

addition, consumers are also increasingly aware of

how PV could benefit them, as in many countries

retail grid parity is present (Olson et al. 2014).

Several top-down studies have been performed on

solar PV potential in the Netherlands, sometimes in

conjunction with potential studies for Europe

(Alsema and Brummelen 1992; Bergsma 1997; De

Noord et al. 2004; Krekel et al. 1987). Many studies

also mention different capacities based on different

top-down approaches. (Krekel et al. 1987), (Alsema

and Brummelen 1992), (Corten and (Bergsma 1997).

De Noord et al. re-assessed these potential estimates

and presented the realistic potential of solar PV in

the Netherlands to be 400 km

(80-120 GWp) for

building integrated PV (BIPV) and 200 km

(40- 60

GWp) for ground-based PV (GBPV). The latest

figure for PV potential is presented by Lemmens et

al., at 150 GWp and is based on the present

electricity consumption in the Netherlands of 120

TWh (Lemmens et al. 2014).

To summarize present top-down estimates for

BIPV potential, in the Netherlands it ranges from

200- 400 km

, or 40-80 GWp. Land based PV

installations would perhaps add another 200 km

Total country potential thus ranges from

80-120 GWp. At the end of 2013, the total amount

of installed PV was estimated at 722 MWp (CBS

2014). It is predicted that in the year 2020 an amount

of 4 GWp will be installed in the Netherlands

(KEMA 2012). With present annual growth rates,

this may be a conservative estimate.

Since the top-down assessment values are

difficult to rely upon these should be validated using

129

B. Kausika B., Dolla O., Folkerts W., Siebenga B., Hermans P. and G. J. H. M. van Sark W..

Bottom-up Analysis of the Solar Photovoltaic Potential for a City in the Netherlands - A Working Model for Calculating the Potential using High Resolution

LiDAR Data.

DOI: 10.5220/0005431401290135

In Proceedings of the 4th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS-2015), pages 129-135

ISBN: 978-989-758-105-2

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

bottom-up assessments that now are possible, using

tools such as “Solar Atlas”, in Dutch Zonatlas

(Zonatlas), which are based on aerial photographs

and solar irradiation. But determining the actual

solar potential of BIPV using high-resolution data

can be very challenging due to the complexity of the

urban areas.

High resolution rooftop potential studies are

relatively new and not much has been done in this

area at rooftop level for estimating the technical and

geographical potential for PV deployment.

(Izquierdo et al. 2008) estimated the technical

potential of roof integrated PV systems using easily

available data and stratified-samples of

Geographical Information Systems (GIS) maps at a

regional level. Based on this work, PV solar energy

potential estimations at municipal to regional level

was conducted in Italy with the help of global solar

radiation maps taken from the Joint Research Centre

of the European Commission (Bergamasco and

Asinari 2011). (Hofierka and Kaňuk 2009) proposed

a methodology for PV potential estimation in urban

areas based on the open-source solar radiation tool

r.sun (developed by (Šúri and Hofierka 2004)) and

3-D city model in GIS. Furthermore, models to

estimate solar potential on building rooftops using

GIS and statistical approaches to create roof-top

solar radiation maps were explored by (Karteris et

al. 2013; Kodysh et al. 2013). (Redweik et al.

2013)developed a model to calculate the solar

energy potential of the buildings taking into account

both the roofs and the facades using high resolution

LiDAR (Laser Imaging Detection And Ranging)

data and applied the model to the campus of

university of Lisbon. However, all the mentioned

studies fall short in estimating the potential at

individual rooftop level.

In the present study, we estimate the rooftop PV

potential in Apeldoorn, a city in the Netherlands

using high resolution LiDAR data and GIS

techniques. Only roof integrated PV is addressed

here. With the use of Solar Analyst (Fu and Paul M

Rich 1999) of ArcGIS solar irradiation over large

geographic areas is computed accounting for

atmospheric effects, sun angle, elevation and effects

of shadows by buildings, elevation and orientation.

Classification of the solar irradiation map was done

to differentiate between optimum and less optimum

suitable sites. These were the basis of potential

estimation, where further energy potential

calculations are made taking into account the slope

and orientation information. These estimations

would help in looking at the trend of PV diffusion,

create business opportunities and additionally

provide an insight for policy implementations.

2 METHODOLOGY

The area chosen for the study was the city of

Apeldoorn (52° 13′ N, 5° 57′ E), in the Gelderland

province of the Netherlands. For locating the

potential PV sites and for calculating the PV

potential a digital elevation model (DEM) derived

from LiDAR data was used. This was obtained from

Actueel Hoogtebestand Nederlands (AHN)

(Nederland 2013). This key input has a resolution of

50 cm (point spacing of 9 points per m

, which is

well suited for estimation of solar radiation at

roof-tops. The study area chosen in shown in Figure

1. The city itself is at low elevation, while in the

West one recognizes a hilly region called De

Veluwe. Another important dataset was a vector file

of the footprints of residential buildings in the study

area. In this paper we focus on the residential sector.

The recent building footprint layer was obtained

from Basisregistratie Adressen en Gebouwen

(BAG), which is a part of the government cadaster

system.

Figure 1: City of Apeldoorn which is taken as the study

area in this research.

The estimation of solar potential in this study

was calculated in two steps. First, suitable locations

for roof-top PV were singled out, and then potential

estimation calculations were performed based on

GIS data analysis. We specified some requirements

in order to characterize suitable locations; and

performed all the calculations using ArcGIS.

The criteria chosen for locating suitable PV sites

were solar irradiation, slope and orientation. This

has been adopted from the work of (Chaves and

Bahill 2010). The Area Solar Radiation Tool of the

ArcGIS Spatial Analyst automatically performs the

solar irradiation calculation based on the model by

SMARTGREENS2015-4thInternationalConferenceonSmartCitiesandGreenICTSystems

130

(Fu and Paul M. Rich 1999). This model takes DEM

as the main input and other parameters relating to

slope, shade and transmissivity of the atmosphere

and calculates the solar irradiance during the time

specified and produces an output image having pixel

values in units of Wh/m

The other inputs for the model were slope and

orientation, which were also created by the Spatial

Analyst tool in ArcGIS.

All the three images were masked to show only

residential buildings and were converted into binary

raster images taking the following criteria:

• Feasible Slope: less than or equal to 38

degrees

• Feasible Solar Irradiation: greater than 70%

of the annual maximum received in the area

which has been taken at 600kWh/m

according to the modelled irradiance

• Feasible Orientation: (a) South facing and

(b) other orientations.

South facing slopes have been considered as

optimum while the other slopes have been taken as

less optimum in this study. The binary rasters were

then combined together to create a final binary

image, which was then filtered to create a smooth

and continuous image.

A raster to polygon tool was used to convert the

suitable areas into a vector polygon layer. Attributes

like area, potential capacity and power were then

attached to these polygons. A value of 150Wp/m

has been taken as the PV power density that can be

installed. Therefore, the final output has been

classified as follows

• 0 : for unsuitable areas shown in red

• 1: partially suitable areas (with high solar

irradiance and orientations other than

south) shown in yellow and

• 2: optimally suited areas( high irradiation

and south facing slopes) in green.

In addition, the production from the estimated

capacity was determined using values determined by

(van Sark et al. 2014). This study states that the

annual production of a PV system in the Netherlands

can be estimated at 875 kWh/kWp. Therefore, for

optimum (south facing) oriented areas 950

kWh/kWp has been chosen and for other, less

optimal orientations 750 kWh/kWp has been taken.

3 RESULTS

The results are explained in the following

subsections. The first subsection shows the model

inputs and in the second subsection binary outputs

after the application of criteria are shown. The third

subsection shows the final output, which is the result

of a binary (AND) operation followed by a raster to

polygon transformation and addition of attributes

and finally the potential estimations.

3.1 Model Inputs

In this subsection the inputs taken in the model are

displayed. Figure 2 shows the height map of the

buildings. Figure 3 shows the annual solar radiation

image in kWh/m

. The area receives an annual

maximum irradiation of 960 kWh/m

in a year

according to the model-based calculations. Figure 4

shows the orientation image or the direction of the

slope. Followed by a slope image where a distinction

between flat and sloping roofs is vivid.

Figure 2: AHN height information derived for residential

buildings. Height information is in meters.

Figure 3: Solar Irradiation image derived for building by

running the Solar Radiation tool. South facing slopes are

seen to receive greater irradiation.

Bottom-upAnalysisoftheSolarPhotovoltaicPotentialforaCityintheNetherlands-AWorkingModelforCalculatingthe

PotentialusingHighResolutionLiDARData

131

3.2 Binary Outputs

Binary outputs after applying the mentioned criteria

for slope, solar irradiation and orientation have been

determined and are presented here.

Figure 5 shows in green the optimum irradiation

map of areas receiving greater than 600 kWh/m

per

year. We see that most of the building rooftops are

selected along with a few roads or empty areas. The

right image in Figure 5 shows the feasible slope

areas in green, which are 38° and below. The white

areas show unfeasible areas, which we can identify

mostly as facades or vegetation.

Images in Figure 6 are the optimum orientation

map, which shows south facing slopes in green (left

image) and other orientations image (right image).

Figure 4: Left: Orientation image showing the direction of slope of the rooftops. Right: Slope image classified in classes to

distinguish between flat and sloping roofs.

Figure 5: Optimum irradiation image (left) and feasible slope image (right).

Figure 6: Optimum orientation image (left) showing south facing slopes and other orientations image (right).

SMARTGREENS2015-4thInternationalConferenceonSmartCitiesandGreenICTSystems

132

Figure 7: Final output showing the geographic potential. Gridcode 0 shows unfeasible areas, 1 represents partially suitable

area and 2 shows best suits areas for the deployment of PV.

3.3 Final Output

The final output is a polygon layer that shows 3

classes (Figure 7). Areas in green are optimally

best-suited locations for PV. These areas receive

maximum amount of solar irradiation and have an

optimum slope and south orientation. South facing

slopes in the Northern hemisphere receive maximum

amount of solar irradiation.

The areas in yellow are partially optimum or the

other orientations, which still receive about more

than 70% of the average solar irradiation in the

region. These areas are suitable for PV but may not

show high energy yields, as they do not receive

maximum solar irradiation throughout the year.

The red areas are categorised as totally

unsuitable. These regions receive either minimum

amount of solar radiation or have unfeasible slopes

(facades or steep slopes) or are either shaded from

trees or nearby buildings.

The final output presented below is the result of

a smoothing filter on a raster, which was then

converted into a polygon shapefile. These polygons

were then intersected with the building information

from BAG so that the final output has address

information along with the building properties as

shown in Figure 8.

Figure 9 shows the attribute table associated with

the final map. Each record corresponds to an address

and each address is further categorised based on the

grid code, which is 0, 1 or 2.

Figure 8: Final map with information on address, potential

capacity and power.

Bottom-upAnalysisoftheSolarPhotovoltaicPotentialforaCityintheNetherlands-AWorkingModelforCalculatingthe

PotentialusingHighResolutionLiDARData

133

Figure 9: Attribute table for the final output.

3.4 Potential Estimation

Potential estimations for the city of Apeldoorn have

been calculated using the field calculator of ArcGIS

for each of the final polygons as can been seen from

the table in Figure 9. Figure 10 shows the rooftop

area in relation to grid code. Total area is about 3.9

. A constant power density of 150 Wp has been

used to estimate the potential capacity per square

meter. This value has been multiplied with the total

area available. Potential estimation has not been

performed for grid code values of 0.

The potential PV capacity for the city of

Apeldoorn thus was estimated at 319.9 MWp for the

residential buildings (Figure 11). This would mean a

power production of 283.94 GWh. Note that the

present PV capacity installed in the region is 3.4

MWp and the annual demand is around 230 GWh at

the rate of 3500 kWh/yr per household.

Using an annual average household

consumption, PV would be able to provide the

annual energy demand of 65,730 households, which

is more than 100% of the total households in

Apeldoorn.

Figure 10: Graph showing rooftop area covered under

each class after analysis. The total of these classes

corresponds to the total roof area of residential buildings

in Apeldoorn.

Figure 11: Potential capacity in MWp and expected yield

in GWh of the optimally suitable areas (Grid code 2) and

partially suited areas (Grid code 1).

4 CONCLUSIONS

In this paper a working model for the estimation of

solar PV potential using high-resolution LiDAR data

and GIS techniques has been presented. Detailed PV

potential estimation studies require high-resolution

height information models. The model presented in

this paper shows great potential and is easy to

implement. The calculations showed that the city of

Apeldoorn has great PV potential in its residential

sector. Based on an average electricity consumption

of residential houses in Apeldoorn of 3500 kWh/yr,

the potential electricity that could be generated

would be able to cover the electricity demand of the

city completely and even produce more.

The application of this methodology to a city has

shown that this method could be deployed in the

whole country for accurate bottom-up determination

of PV potential. This method could also be applied

to the whole of the Netherlands but proper

extrapolation techniques have to be developed. This

is currently under investigation.

REFERENCES

Bergamasco, L. and Asinari, P., 2011. Scalable

methodology for the photovoltaic solar energy

potential assessment based on available roof surface

area: Application to Piedmont Region (Italy). Solar

Energy, 85(5), pp.1041–1055. Available at:

http://dx.doi.org/10.1016/j.solener.2011.02.022.

CBS, 2014. CBS StatLine - Energy consumption private

dwellings; type of dwelling and regions. Available at:

http://statline.cbs.nl/Statweb/publication.

Chaves, A. and Bahill, T.A., 2010. Locating Sites for

Photovoltaic Solar. Available at: http://www.esri.com/

news/arcuser/1010/solarsiting.html.

Fu, P. and Rich, P.M., 1999. Design and implementation

of the Solar Analyst: an ArcView extension for

modeling solar radiation at landscape scales. In

Proceedings of the 19th annual ESRI user conference,

SMARTGREENS2015-4thInternationalConferenceonSmartCitiesandGreenICTSystems

134

San Diego, USA. Available at: http://professorpaul.

com/publications/fu_rich_1999_esri.pdf.

Fu, P. and Rich, P.M., 1999. Design and Implementation

of the Solar Analyst: an ArcView Extension for

Modeling Solar Radiation at Landscape Scales. 19th

Annual ESRI User Conference, pp.1–24.

Hofierka, J. and Kaňuk, J., 2009. Assessment of

photovoltaic potential in urban areas using open-

source solar radiation tools. Renewable Energy,

34(10), pp.2206–2214. Available at:

http://www.sciencedirect.com/science/article/pii/S096

0148109000949.

Izquierdo, S., Rodrigues, M. and Fueyo, N., 2008. A

method for estimating the geographical distribution of

the available roof surface area for large-scale

photovoltaic energy-potential evaluations. Solar

Energy, 82, pp.929–939.

Karteris, M., Slini, T. and Papadopoulos, a. M., 2013.

Urban solar energy potential in Greece: A statistical

calculation model of suitable built roof areas for

photovoltaics. Energy and Buildings, 62, pp.459–468.

Available at: http://dx.doi.org/10.1016/j.enbuild

.2013.03.033.

KEMA, 2012. Nationaal Actieplan Zonnestroom,

Kodysh, J.B. et al., 2013. Methodology for estimating

solar potential on multiple building rooftops for

photovoltaic systems. Sustainable Cities and Society,

8, pp.31–41. Available at: http://dx.doi.org/10.1016/

j.scs.2013.01.002.

Nederland, A.H., 2013. AHN - Actueel Hoogtebestand

Nederland - homepage.

Available at: http://www.ahn.nl/index.html.

Olson, C.L. et al., 2014. Is grid parity an indicator for PV

market expansion in the Netherlands? Solar Energy,

2013, p.2012. Available at: ftp://130.112.2.101/

pub/www/library/report/2013/m13041.pdf.

Redweik, P., Catita, C. and Brito, M., 2013. Solar energy

potential on roofs and facades in an urban landscape.

Solar Energy, 97, pp.332–341. Available at:

http://dx.doi.org/10.1016/j.solener.2013.08.036.

Van Sark, W. et al., 2014. Opbrengst van

zonnestroomsystemen in Nederland, Available at:

http://pers.uu.nl/nederlandse-zonnepanelen-opbrengst-

kengetal/.

Šúri, M. and Hofierka, J., 2004. A New GIS-based Solar

Radiation Model and Its Application to Photovoltaic

Assessments. Transactions in GIS, 8(2), pp.175–190.

Available at: http://onlinelibrary.wiley.com/

doi/10.1111/j.1467-9671.2004.00174.x/abstract.

Zonatlas, GA NAAR DE ZONATLAS | Zonatlas

Apeldoorn. Available at: http://www.zonatlas.nl/

apeldoorn/ontdek-de-zonatlas/.

Bottom-upAnalysisoftheSolarPhotovoltaicPotentialforaCityintheNetherlands-AWorkingModelforCalculatingthe

PotentialusingHighResolutionLiDARData

135