VISUAL STUDIES OF TRANSPARENT PV - ELEMENTS Anne Gunnarshaug Lien SINTEF Civil and Environmental Engineering, N-74XX Trondheim, Norway, Tel. No. +47 73 59 26 21, Fax No. +47 73 59 82 85, E-mail Anne.G.Lien@civil.sintef.no Anne Grete Hestnes Faculty of Architecture, Planning and Fine Arts, Norwegian University of Science and Technology (NTNU), N-7491 Trondheim Norway, Tel. No. +47 73 59 50 40, Fax No. +47 73 59 50 45 E-mail: annegrete.hestnes@ark.ntnu.no Abstract - The paper presents results from the EU-JOULE project Amorphous Si Photovoltaics for Commercial Buildings (ASICOM). The objective has been to describe the visual qualities of elements with PV cells in laminated clear glass and partly transparent amorphous Si photovoltaic elements (a-si PVelements). The PV cells in laminated glass are usually of the crystalline type (c-si PV-elements), and further studies of this type are also presented. Materials that are transparent, translucent or partly transparent have different appearances in different lighting situations. Seen from the inside of a room a clear glass window is almost invisible during the day but looks like a black mirror at night. A diffusing window material can be glary in direct sunlight but transmits comfortable daylight from an overcast sky. For new materials that are partly transparent to light, such as PV cells in laminated clear glass or partly transparent a-si PV-elements, it is necessary to describe the appearance of the material for different lighting conditions to know how they can be used. The issues that are described concern visual amenity and include contrast, glare, and light modelling of a room. What can be seen through the transparent material - the view or the visual contact, is also described. The methods that are used for the visual studies are studies of prototypes (or mock-ups), studies of scale models, and studies with three-dimensional computer models. The variables described concern lighting situations and transparency of the elements, and the results are meant to be of interest for architectural integration of transparent PV-elements. 1. INTRODUCTION The paper presents results from the EU-JOULE project Amorphous Si Photovoltaics for Commercial Buildings (ASICOM). The objective of this part of the project has been to describe the visual qualities of different types of amorphous Si photovoltaic elements (a-si PV-elements) and "mosaic a -Si PV elements". The mosaic elements consist of pieces of a-si PV in laminated glass with transparent spaces between them. a-si PV are thin film cells and can be made with different shapes and sizes. The colour of the cells is dark brown, and the efficiency is 4-8%. The amorphous silicon film is a transparent reddish brown film but the back electrode is usually an opaque aluminium layer. The aluminium layer may consist of a grid if a more transparent module is wanted. The objective of the ASICOM project has been to study different types of more or less transparent a-si PV for use in commercial buildings. The paper also present some further studies of visual qualities of crystalline PV cells in laminated glass. The objective of the visual studies has been to find out how to maintain the transparent appearance and the aesthetic quality of glass facades when a significant part of the facade is covered with PV cells. The influence on the visual amenity inside the rooms has been an important issue to investigate. Materials that are transparent, translucent or partly transparent (all three are often called transparent) have different appearances in different lighting situations. Seen from the inside of a room a clear glass window is almost invisible during the day but looks like a black mirror at night. A diffusing window material can be glary in direct sunlight but transmit comfortable daylight from an overcast sky. For new building elements that are partly transparent to light, such as PV cells in laminated clear glass or partly transparent a-si PV-elements, it is necessary to describe the appearance of the building elements for different lighting conditions to know how
they can be used. The visual amenity in an office is one of the most important issues. Visual amenity includes contrast, glare, and light modelling of a room. What can be seen through the transparent material - the view or the visual contact, is also important. Three methods that can be used for visual studies are studies of prototypes (or mock-ups), studies of scale models, and studies with three-dimensional (3D) computer models. None of these methods result in exact answers. Studying prototypes gives a good impression of the material, but it does not show how it will look like when used in a particular room or on a particular building. A scale model study can be used for studying the material used in a particular situation. Distribution of the daylight in a room can be studied in a model in scales between 1:10 and 1:15. With the use of 3D computer models, a large number of material types and building situations can be studied, and nice representations can be made. The disadvantage of this is that the representation of the light might not be correct and that the focus in the picture will not be as for the human eye. Good results of modelling the distribution of daylight in a room can be obtained with the computer program RADIANCE, however, but this type of modelling is still time consuming and should only be done for the most promising results of the study. The 3D program FormZ can be used for a large number of studies of both the interior and the exterior of a building. The pictures are good, but the representation of the daylight distribution is rather simple. The methods used for the ASICOM project are: - studies of 5 different prototypes of a-si PV elements - 3D computer modelling of PV-elements used on single facades (FormZ) The methods used for further studies of PV in clear glass are: - studies of mock-ups of PV-elements - 3D computer modelling of PV-elements used on both single and double facades (FormZ) - studies of a scale model of a room with a double facade (scale 1 : 10) 2. ANALYSES OF VISUAL AMENITY The objective of studying visual amenity is to find out how potential comfort problems, such as glare from the transparent PV-element, can be reduced. Both the distances between the cells, the type of surface on the back of the cells, and the location of the PV-elements will have an impact on the visual comfort achieved in the room behind. The distance between the cells is also critical for the visual contact, i.e. for what can be seen through the element. The study of visual amenity involves studying glare, contrast, and light modelling of a room. All these aspects describe how the occupant experiences the room. These analyses are therefore mostly done from the inside of a room. The study of visual contact is also done from the inside, looking out. Glare is discomfort caused by bright light. When the sun shines on a transparent PV-element, the solar cells will throw shadows into the room and glaring light will come through the spaces between them. A wall with such PVelements might be uncomfortable to look at when the eye has both the opaque back-side of the cells and the transparent glazing in the visual field at the same time. See Figure 1. Figure 1 The sun shining on a mock-up of a PV element. Contrast between surfaces might also cause discomfort. The contrast between the transparent part and the opaque part of the transparent PV-element will always be large when the light shines through it. During the day, the back of the cells looks dark and the sky looks bright. The contrast problem during the day is very different for different lighting conditions. While direct solar radiation causes discomfort, the overcast sky causes few problems. At night, when the light is turned on inside, the back of the cells looks light and the sky looks dark. This situation causes no problems. However, the colour of the backing is still important for the aesthetic appearance. Light modelling: The direction of solar radiation creates light and shadow in a room. Transparent PV-elements will cause shadow patterns on the walls, and this has an impact on the visual appearance of the room. Coloured transparent materials, like transparent a-si PV elements, make the room appear coloured.
Visual contact: What can be seen through the transparent element is first of all dependent on the size of the transparent parts. The size of each solar cell, the distance between them, and the distance from the observer to the element does also have an impact. When seen from a long distance, an element with small cells will look more transparent than an element with large cells. 3. STUDIES OF PROTOTYPE a-si ELEMENTS Five different 300mm x 300mm modules of amorphous silicon PV were studied in order to describe the visual and aesthetic quality of the different types. The modules were placed on the inside of a window and two photos are taken from the inside of the room with different lighting situations, see Figure 2. The left photo is taken during the day with diffuse light from an overcast sky. The right photo is taken at night with electric light in the room. At night all the modules have a reflective appearance similar to conventional glass. The elements have lighter colors, though, due to the metallic backing. increase the transparency of the module. The difference can hardly be seen, though. For this sample the grey backing is matt because it had a chance to oxidise before the module was encapsulated. The matt backing is much more comfortable than the specular one. Module C is a mosaic sample of the normal opaque type. The size of the cells are 100mm x 100mm. The spaces between the cells are 30mm. Module D is a transparent type with green perforated aluminium backing. The aluminium layer has holes with 0.6mm diameter and with 1.0mm space between them. The transparency is limited, and little can be seen through it. The element looks red and green in daylight and green in electric light at night. These colors have a strong influence on the appearance of a room. The element will most likely not be useful for the majority of buildings, as the red color of the material gives red light to the room. Module E is the most transparent module, as it is has no aluminum back contact. The view can be seen quite clearly through it. The element is red in daylight and grey in electric light. The stripes are laser scribes in the TCO and a-si layer. Both the mosaic type and the transparent type will usually disturb the view. From a short distance the eye will focus either on the material or on the view. To illustrate this problem, photos can provide a good representation of what the human eye is seeing. As long as the eye can decide what to focus on there will be no problem. If the eye is not determined what to focus on, and shifts between the material and the view, it will feel disturbed and irritated. 4. COMPUTER STUDIES OF MOSAIC a-si PV ELEMENTS Figure 2 A-Si PV- modules with varying transparency. Module A - middle left. Module B - lower left. Module C - lower right. Module D - middle right. Module E - upper right. Module A is the normal opaque type. Only a small amount of light is transmitted through the 0.5mm stripes in the aluminium layer. The stripes appear red in daylight and grey in electric light. The grey aluminium backing has a strong mirror reflection both in daylight and in electric light. For use in offices or other rooms were the PV-elements are close to the occupants, the mirror effect of this specular surface will be quite disturbing. Module B is a modification of the normal opaque type. The stripes in the aluminium are made 1.0 mm wide to The use of elements with transparent spacing between the modules opens for a large variety both in design of new PV-elements and in new facade designs. The computer pictures in Figure 3 show some different types of mosaic PV-elements where both the size of the cells/modules and the size of the space between them vary. The mosaic elements also cover different parts of an exterior office wall. Although many new and exciting designs can be made with this kind of facade elements, some new comfort problems may occur. These problems will mostly be related to visual amenity and include glare problems and focusing on the view. To avoid this problem the mosaic PV-elements should not be located in the best view area. Glare may occur in direct sunlight because the contrast between the opaque and the transparent parts of the mosaic elements will be very large. This is also the reason why the back of the PV-cells looks dark in
daylight although the backing has a light color. Usually the glare problems for this kind of walls is not worse than for a conventional glazed wall and can be solved by the use of blinds or curtains. Figure 3 A variety of mosaic elements and designs of an exterior office wall. Computer pictures: P.O.Pettersen
5. COMPUTER MODEL STUDIES OF CRYSTALLINE PV ELEMENTS Further studies of visual qualities of crystalline PV cells in laminated glass were also carried out. The parameters analysed are listed below. All the three methods are used to describe these parameters. - the spacing of the cells location of PV-elements: on most of the facade and under/over the windows the colour of the back side of the cells sectioning of the cavity in a double facade situation different inner facades: totally glazed inner facade and two cases with an opaque part and a window The first parameter analysed is the spacing between the cells and the influence on the view and the light level in an office with single facade. Computer model studies have been done with the 3D program FormZ. Figure 4 shows how the daylight level will increase with increasing distance between the cells, and that the view is increasingly easy to distinguish with increasing distance between the cells. The PV-wall in the first picture has a very small spacing between the cells. The appearance is almost opaque, and very little light is transmitted through the PV-part of the wall. Very little of the view is seen. In the second picture, with a wider spacing between the cells, light is distributed to the whole room, and the view is more distinguishable. The contrast is smaller between the PV-part and the window part. For the last two pictures, with even wider spacings, the light distribution in the room is good and the view is even better distinguished. The aesthetic appearance of these walls is not the best, though. The PV-cells look like spots on a window and the PV-part and the window part sort of merge. The two first pictures look better, because the window part and the PV-part appear as two different elements. The distribution of the daylight in the rooms of these pictures looks very realistic. The view through the PVelements also looks realistic. To check the computer pictures against a more realistic situation, with real sunlight, a scale model was made. Figure 4 Computer pictures of an office with PV elements on the south facing facade. The PV elements have different spacing between the cells. Computer pictures: P.O.Pettersen
6. COMPUTER MODELLING COMPARED TO USING A SCALE MODEL The scale model (scale 1 : 10) represents a standard office that is 2.8 m high, 2.4 m wide and 4.2 m long and has a double facade. The outer facade is made of Plexiglas with PV-cells of grey plastic film. The inner facade is made of one layer of Plexiglas or of a combination of Plexiglas and wood. The spacing between the two facades, the width of the cavity, was regulated with long screws. Light from electric lamps sent through 40mm thick solid Plexiglas tubes with a mat surface was used as electric light The model was only used for studies of the interior, and the colour of the cells is light grey to look like the back side of the real PV-cells. The plastic used as cells is not totally opaque. Compared to real PV-cells with the same grey colour on the back side, our cells therefore look a little lighter. The model is studied when exposed to direct sun, to a diffuse sky, and with electric light inside the room at night. Natural daylight was used. The problem of using natural daylight is that the light and the sky vary all the time. For direct sun the angle of radiation will vary over the day. Photos taken in direct sun will therefore have different angles of incidence. For diffuse light from an overcast sky the light level varies quite a lot. The variation could be from 10 000 lux to 20 000 lux from minute to minute. A partly cloudy day can vary from 10 000 lux to 60 000 lux. The light level of the photos taken in diffuse light can therefore not be compared for the purpose of obtaining exact differences but only as an indication of differences. All pictures in one series are taken with the same aperture and shutter, but the exposure time in the copy process was not controlled. The second computer model picture in Figure 4 can be compared to the first scale model picture in Figure 5. The angle of incidence in the scale model picture is lower and the light is not so bright as in the computer model. The contrasts in the PV-element are larger in the scale model, though, and this element looks more glary. The distribution of the light in the room looks quite different for the two cases. For both cases the cells on the upper part of the wall are darker than the cells on the lower part. On the lower part light is reflected from the floor back to the cells. For studying glare, contrast, and light distribution, the scale model gives the more realistic results. Figure 5 Scale model photos of an office with PV elements on the south facing facade exposed to direct sunlight, diffuse light and electric light. The three scale model pictures above show a south facing office exposed to direct sunlight, diffuse light from an overcast sky, and electric light from inside of the room. The wall has one surface with two very different appearances, one part that is transparent, and one part that is partly transparent. The transparent part is a conventional window, a familiar building element. The other part is so far new and unfamiliar and quite difficult to look at. To make the wall as uncomplicated as possible for people to look at, the two different surfaces should appear as different elements with different locations. The PV-elements should be located at the upper part and/or at the lower part of the wall and the windows should be located in the area of view.
7. THE SCALE MODEL COMPARED TO A MOCK-UP A mock-up of a PV-element was made to study an even more realistic situation. The view seen from different distances, the contrast and glare problem, and the colour of the back of the cells were studied. The mock-up was also used to compare the effects observed in the scale model with almost real PV-cells. The cells are made of paper prints of monocrystalline cells in full size glued on a plastic sheet. The cells are totally opaque and the back is light grey. The element covers a window that is 1,5m high and 0.8m wide. The cell distance is 30mm. The view is increasingly easy to distinguish with increasing distance between the observer and the PVelement. It is more uncomfortable to look through the element from close up than from a distance. The problem is that the eye will focus either on the cells or on the view, and with both parts in the visual field, this is confusing. This confirms the conclusion from the scale model study saying that the PV-elements should not be placed in the area of view. At least not in an office or in situations were people sit close to the facade during the work hours. For corridors and halls these comfort problems will not be as noticeable. Figure 6 shows one photo of the mock-up and one photo of the scale model. In both cases the distance between the cells is the same. Even though the cells in the scale model are not totally opaque, the appearance is the same as for the mock-up. In both cases the cells look very dark toward the bright sky. On the picture of the mock-up, reflected light from the desk hits the back of the cells and makes them lighter. On the scale model, light is reflected from the floor to the lower part of the PV-element and makes it lighter. The scale model also has an inner Plexiglas sheet that reflects light onto the back of the outer facade. Based on the comparison with the mock-up, it was concluded that the scale model looks realistic and that it is well suited for further studies. 8. PV ELEMENTS ON THE OUTER FACADE AND DIFFERENT INNER FACADES The scale model is therefore used for further investigations. Three series of studies of an office with double facade with PV-elements on the outer facade and different types of inner facades are made. For all cases the PV-elements are placed on the upper and lower part of the wall and the spacing between the cells are in the middle range. The three series show the office room exposed to direct sun, exposed to diffuse light from overcast sky, and illuminated with electric light. The upper three photos of Figure 7 show the office room exposed to direct sun before and after noon in September in Trondheim (63,5 N). The three photos in the middle show the office room exposed to diffuse light from an overcast sky. The three lower photos show the office room illuminated with electric light. The dark spot in the middle of the picture is the reflection of the back wall with a hole for the camera lens. Figure 6 Upper photo: the mock-up with PV cells in full size. Lower photo: Scale model with comparable spacing between the cells. The three photos to the left show the office with a totally glazed inner facade. Compared to Figure 4 the location of the cells on the upper and lower part of the wall is easier to relate to. The need for shading for the window part, will be the same as for a single south-facing wall. For the PV-part a simple shading system, such as a light curtain, will be sufficient to eliminate the glare problem. The three photos in the middle show the office with an opaque parapet on the inner facade. The appearances of these rooms are very different from the ones with totally
glazed inner facade. The cell shadows are only seen on the wall and not on the floor, and only the upper PVelement is seen from the office. The room looks more like a conventional office with some kind of shading device hanging outside the window. The three photos to the right show the office with an opaque wall both over an under the window of the inner facade. For this case the outer facade has very little impact on the appearance of the office room. All the cases above have open cavities; i.e. it is not divided with floors. The light distribution in offices is different for cases with an open cavity and for cases with a closed cavity, i.e. with a cavity divided with floors. With an open cavity, the daylight reaches deeper into the room because there is no obstructing floor over the windows. For the cases with a closed cavity, the area around the window will have more light, because light is reflected from the cavity floor to the ceiling close to the window. Figure 7 Scale model photos of an office with PV elements on the outer facade and different inner facades.
9. CONCLUSIONS These visual studies were carried out to investigate aesthetic qualities of new transparent PV elements. Similar studies has been done for new transparent insulation materials (Lien, 1995). The purpose was both to understand problems and limitations, to investigate options for new products, and to develop new ideas for facade integration of PV cells. As we had not been able to find any relevant literature, indicating that not much work had been done in this area, our work in part consisted of finding the right types of experiments/developing an appropriate methodology. The types of experiments we chose to use, mock-ups, scale models, and computer studies with the program FormZ, turned out to be quite complementary. In particular, the FormZ studies made it possible to investigate a large number of situations, while the scale model studies helped confirm the results and was useful for better understanding of the lighting situation. The scale model was also used to "calibrate" the lighting options of the computer model. For new products such as PV in laminated glass the prototypes or mock-ups will always have to bee used at some stage in the development of new building elements. With the use of computer models and scale models quite a few issues can be sorted out first, however. REFERENCES Aschehoug et al. BP Amoco Solar Skin - a double facade with PV. Eurosun2000. Lien, Anne Gunnarshaug. "Transparent insulation materials for low energy dwellings in cold climate" Thesis, NTNU, Norway, 1995.