Advances in Power and Energy Systems

Visual impact assessment for small and medium size PV plants E. GARCIA-GARRIDO, P. LARA-SANTILLÁN, E. ZORZANO-ALBA, M. MENDOZAVILLENA, P. ZORZANO-SANTAMARÍA, L.A. FERNÁNDEZ-JIMÉNEZ, A. FALCES Department of Electrical Engineering University of La Rioja Luis de Ulloa 20, 26004 Logroño SPAIN [email protected] Abstract: - This paper presents a methodology, based on a geographical information system, for the assessment of the visual impact of new solar photovoltaic plants. The methodology uses geographical and population data in order to compute a relative visual impact index, which is a function of the observers that can see the new plant, and of the distance from their homes or work places to the site selected for the facility. All the feasible locations in a geographical zone are evaluated jointly, which allows to identify the places with the lowest relative visual impact index.

Key-Words: Solar power, photovoltaic plant, power generation, visual impact, geographical information systems place is taken into account and consequently weighting factors are applied. Additionally, in [4] is presented a function for the calculation of an indicator of objective aesthetic impact of wind farms. The factors used include visibility, fractality, colour, climatology and continuity of the wind turbines. The function is applied to two wind farms in Spain and Wales, and the results are compared with those obtained by means of surveys with photographs among inhabitants living nearby. A Geographical Information System (GIS) is used in [5] to the visual impact assessment of a wind farm. In this work four parameters are taken into account: orography, land-cover height, the facility height and width, and the observer height. The population in the vicinity and the users of roads and railways are also considered, and the distance to the facility is included as a divisor of a visual impact index. The authors calculate the visual impact indexes of six wind farms. The results show the relationship between the number of wind turbines and the visual impact index values. In the VIA for photovoltaic (PV) plants, the distance between the facility and the observer has also a relevant importance. This parameter is considered as the most important among others that include the effect of the scale, topography, vegetation and weather [6]. This work uses fuzzy techniques, based on Hassell visual assessment matrix [7], which considers that an object starts to be 50% less visible from 5 km away.

1 Introduction The significant increase worldwide in power plants based on renewable energy sources is mainly due to their important environmental advantages, such as CO2 reductions with respect to conventional power plants. But despite the advantages, there are some restrictions to the social acceptance of renewable energy technologies. One of these restrictions is the visual impact that large power plants may generate on people, causing undesired changes of the landscape [1]. The opposition of the local population can slow down and even block the construction of new power plants, so the selection of the places with lowest visual impact in a zone, can contribute to accelerate their construction. Most of large power plants based on renewable energies are wind farms and solar plants, what justifies the works about visual impact assessment (VIA) for these plants. Most of these works have been published in recent years [2-7]. In general, visual impact assessment takes into account the distance to the power plant. So, in [2] and [3] are assigned different coefficients according to the distance to the power plant: places with a distance (to a reference point or location) lower than 500 m correspond to the highest coefficient (value 1), places with a distance from 500 m to 6 km correspond to coefficient values that decrease linearly with the distance, and places further than 6 km correspond to an 0.1 coefficient value. The people than can see the power plant from each

ISBN: 978-1-61804-128-9

57

Advances in Power and Energy Systems

characteristics (panels height, extension of occupied terrain, etc.). For a single observation point (village, hamlet, farm, lookouts, etc.) we define the “visual pollution” as a magnitude, that depends on the number of affected persons, on the distance between them and the observed object (because the capacity of the human visual perception diminishes with the distance), and also depends on the orography of the terrain between the observation point and the observed object. The human visual perception can be modelled with a mathematical function of the distance that presents the maximum value for small distances and the minimum value for large ones. The statistical distribution that represents the human visual capacity responds to a Gaussian curve, because not all the humans are able to view the landscape with the same definition degree, and a limit of vision in the horizon exists (that depends on the height and size of the objects). Thus, in order to model the visual perception for a single observer, we propose a sigmoid function whose values decrease with the distance, and its inflexion point corresponds to the mean of the Gaussian curve that represents the human visual capacity. This function acquires a negligible value for the distance corresponding to the horizon limit. Beyond this limit is not possible to establish a vision line between the observer and the objects. Fig. 1 plots this sigmoid function.

The risk of glares from PV plants is focused in [8], with a study of the visual impact produced by a PV plant with 3085 modules sited in Italy. The authors conclude that although glares can increase the visual impact, they take place only in very short time periods, so they can be neglected. So, although an important effort has been done in order to assess the visual impact of new power plants based on renewable energies, it has done for specific sites, not over a whole area in order to select the locations where the visual impacts are the lowest ones. This paper presents a methodology, based on the use of a GIS, for the selection of the places with the lowest visual impact for the installation of a new medium-size PV plant. The methodology takes into account the population in the vicinity, the orography, and the height of the facility, calculating the “visual pollution value”. This value is calculated for all the places in a geographical zone. Afterwards, the values are normalized obtaining a “relative visual impact index” for each place. The results, in form of map indexes, allow identify graphically the best places (under the visual impact standpoint) to build the new PV plant. Although the methodology has been applied to PV plants, can be used for other kind of power plants, such as wind farms and even conventional plants. The paper is structured as follows: section 2 presents the proposed methodology, in section 3 is presented two case studies with the visual impact indexes for new PV plants in a zone of La Rioja (Spain), finally section 4 presents the conclusion.

Inflexion point

2 Proposed methodology

ISBN: 978-1-61804-128-9

Sigmoid function

The goal of this work is the achievement of a relative visual impact index map. The data of this map are stored in raster format in a GIS. The raster format keeps the geographical information in cells that represent a small square area in the geographical zone under study. In the GIS, a numerical value is associated to each cell of the raster format. In our case, the value that is contained in each cell is the relative visual impact index (for all the potential observers) due to a landscape modification as a result of the PV plant installation in the geographical area represented by that cell. This relative index depends on the terrain characteristics in the studied zone (orography, represented into the GIS by the Digital Terrain Model, DTM), also depends on the existing communities in the area under study (observation points and observers), and the PV plant

Visual horizon limit Sf  1 -

1 1   e -  (x - c)

Distance

Visual capacity distribution

Fig. 1. Sigmoid function for visual perception.

58

Advances in Power and Energy Systems

operation, the obtained final result is a GIS raster in which each cell (position) presents a value between 0 and 1. This last value is a relative visual impact index, that is, the cells with highest value correspond to the places with the greatest visual impact in the zone.

Fig. 2 represents the methodology used to calculate the visual pollution caused by the PV plant, taking into account only one community (village, hamlet, farm, etc.). The process stages are as follows: 1. The starting point is the DTM with the selection of the area for the PV plant (position or cell i) and the area corresponding to the community (position or cell j). 2. Data corresponding to observer height above the ground (h0) and number of potential observers (N0) from the community are available and stored in the GIS. 3. Data corresponding to the height above the ground (hi) and the occupied extension (Si) of the PV plant are available and stored in the GIS. 4. We apply the “observability filter”. This filter corresponds to a function which returns a value 0 if the PV plant in position i (with the characteristics stored in step 3) is not observable from the community in position j (with the height stores in step 2). The filter returns a value 1 when the PV plant is observable. 5. The GIS calculates the Euclidean distance between the areas represented by cells i and j. 6. The GIS calculated the visual perception for a single observer applying the sigmoid function. 7. The GIS multiplies the value obtained with the observability filter, the visual perception for a single observer, and the number of potential observers in j. The result corresponds to the visual pollution.

Start

Digital Terrain Model (DTM)

Observability filter

Position j

Sigmoid function

Observers in j

Product

Visual pollution

End

The steps 2 to 7 are repeated for all the communities in the zone under study. Afterwards, the values obtained in step 7 for each community (each of the possible position j) are aggregated to obtain the global visual pollution value, that is, the visual pollution for all observers or inhabitants in the area. With this action, we obtain a numeric value that is stored in the cell corresponding to the position i. A map with the global visual pollution can be obtained if the previously presented process is applied to all of the cells that correspond to the zone under study (possible positions i). Each cell of this map contains the global visual pollution value caused by the PV plant in the area represented by that cell. To obtain a normalized index, we divide the global visual pollution value of each cell by the maximum value in the studied area (the maximum value of global visual pollution). With this

ISBN: 978-1-61804-128-9

PV Plant location Position i

Fig. 2. Visual pollution calculation for only one community.

3 Case study The proposed methodology has been applied to a zone of approximately 400 km2 in La Rioja, Spain (Fig. 3). In that zone there are three communities: Arnedo (14457 inhabitants, 531 metres above sea level), Quel (2096 inhabitants, 479 metres above sea level) and Autol (4458 inhabitants, 458 metres above sea level). So, the population, potentially affected by the construction of a new PV plant in the zone, is 21011 persons (6.51 per cent of inhabitants in La Rioja in the year 2011). The

59

Advances in Power and Energy Systems

greater value of visual pollution due to the height of solar trackers. The effect of the distance to the observed object can be seen in the yellow area (with the shape of an arc centred in the greatest community), which correspond to intermediate index values, decreasing with distance. Evidently, PV plant equipment with higher height (as the plant with solar trackers of the Fig. 5), show bigger relative visual impact index that those with lower height (for example, fixed groundmounted panels which relative visual impact index map is shown in Fig. 4). There are more grey cells in the proximities of the considered communities in Fig. 4 than those in Fig. 5.

values of the sigmoid function parameters were  = 10,  = 0.0011 m-1, and c = 1600 m. The study has been realized to evaluate the relative visual impact of two different PV plants: a PV plant with fixed ground-mounted panels, with two metres average height, and a capacity of 0.5 MW; and a PV plant with solar trackers, with seven metres average height, and a capacity of 1 MW. The area needed for both types of facilities is one hectare (10000 m2). Therefore a resolution of 10000 m2 (the size of the raster cell is 100x100 metres) has been chosen, to assign to every cell of the map the possible location of the PV plant.

Visual Impact index Value

≤1 >0

ARNEDO

QUEL

ARN EDO QU EL AUT OL

AUTOL

Fig. 3. Map of La Rioja where is highlighted with a box the studied area. Applying the methodology described in section 2, the relative visual impact index maps are obtained for the zone under study. These maps are shown in Fig. 4 and Fig. 5. In these figures the dark purple colour corresponds to the places where the relative visual impact index is greater. The dark green colour corresponds to the areas where the relative visual impact index for an observer placed in any of three points of observation considered, Arnedo, Quel or Autol, has a low value. The cells in grey colour correspond to the places where the relative visual impact index is null (hidden from all the observation points by the effect of the local orography) and therefore, they are the suitable sites to locate PV plants visual impact standpoint. The light purple and light green colours represent intermediate values. Comparing the relative visual impact index maps shown in Fig. 4 (fixed ground-mounted panels) and Fig. 5 (solar trackers), we can observe a main difference: a greater surface with high visual impact index around the communities in Fig. 5 with respect to Fig. 4. This difference is caused by the

ISBN: 978-1-61804-128-9

Fig. 4. Relative visual impact index map for fixed ground-mounted PV plant. Visual Impact index Value

≤1 >0

ARNEDO

QUEL AUTOL

Fig. 5. Relative visual impact index map for solar trackers PV plant. In the presented case study only three communities have been considered. For a more

60

Advances in Power and Energy Systems

detail study, also other human common places as farms, factories, lookouts, etc, can be considered, using the number of inhabitants, workers or visitors as the number of observers. Any PV plant promoter can select easily, in the corresponding relative visual impact index map, the places with the lowest values. These places correspond to those, in the studied region, where the construction of the new PV plant presents lower visual impact and, therefore, the time needed to obtain the permissions from the local authorities can be shorter.

References: [1] J.Zoellner, P. Schweizer-Ries, C. Wemheuer, Public acceptance of renewable energies: Results from case studies in Germany, Energy Policy, Vol. 36, No. 11, 2008, pp. 4136-4141. [2] J.P. Hurtado, J. Fernández, J.L. Parrondo, E. Blanco, Spanish method of visual impact evaluation in wind farms, Renewable and Sustainable Energy Reviews, Vol. 8, No. 5, 2004, pp. 483-491. [3] T. Tsoutsos , A. Tsouchlaraki, M. Tsiropoulos, M. Serpetsidakis, Visual impact evaluation of a wind park in a Greek island, Applied Energy, Vol. 86, No. 4, 2009, pp. 546-553. [4] A.C. Torres-Sibille, V.A. Cloquell-Ballester, V.A. Cloquell-Ballester, R. Dartonb, Development and validation of a multicriteria indicator for the assessment of objective aesthetic impact of wind farms, Renewable and Sustainable Energy Reviews, Vol. 13, No. 1, 2009, pp. 40-66. [5] M. Rodrigues, C. Montañés, N. Fueyo, A method for the assessment of the visual impact caused by the large-scale deployment of renewable-energy facilities, Environmental Impact Assessment Review, Vol. 30, No. 4, 2010, pp. 240-246. [6] SRK Consulting, Draft Visual Impact Assessment for the proposed SATO holdings Photovoltaic project, near Aggeneys, Northern Cape, Sato Energy Holdings Report, No. 435209_VIA, 2012. [7] P. G. Haack, Taralga Wind Farm Development Landscape Visual Assessment, 2005, On-line: http://www.planning.nsw.gov.au/asp/pdf/taralg a_app_d_hassell_report04.pdf [8] R. Chiabrando, E. Fabrizio, G. Garnero, The territorial and landscape impacts of photovoltaic systems: Definition of impacts and assessment of the glare risk, Renewable and Sustainable Energy Reviews, Vol. 13, No. 9, 2009, pp. 2441-2451.

4 Conclusion A new methodology of the VIA for small and medium size PV plants has been presented. It is based on GIS, what allows the achievement of visual maps. The relative visual impact index map has been defined. This map represents, in a raster format, a normalized index of the visual pollution. This value is proportional to the number of visually affected people in the zone, as a consequence of the installation of a new PV plant. The final results, relative visual impact index maps, take into account the characteristics of the PV plant, the orography of the zone, and the number of inhabitants or visitors. These maps allow the selection of the places with lower visual impact in the zone, what can contribute to speed up the construction of new PV plants. The proposed methodology can be adapted to any kind of power plant, if its characteristics (surface size and height), the DTM of the zone, as well as the number and location of inhabitants, workers or visitors in each possible observation point are known. Acknowledgements The authors would like to thank the Spanish Ministry of Science and Innovation for its support under the Project ENE2009-14582-C02-02.

ISBN: 978-1-61804-128-9

61