Open Access
Add to BibSonomyAbstract
We have examined the performance of a see-through photovoltaics module that uses a low-concentration prism concentrator by undertaking ray-tracing analysis and an on-site experiment. The incident angle dependency of the prism concentrator makes it possible to concentrate direct solar radiation onto solar cells and transmit diffuse solar radiation. Fewer solar cells can then be used without sacrificing the conversion efficiency or lighting performance. The module generates approximately 1.15 more electricity than a conventional module while operating with 63% less solar cell area. We also introduce a design method for the concentrator geometry that adjusts the incident angle dependency for different latitude and tilt angles.
©2011 Optical Society of America
1. Introduction
Building-integrated photovoltaics (BIPV) are used in place of conventional exterior building materials to provide electric power directly to the building and any excess can be feed back into the grid. Photovoltaic (PV) modules can be easily attached to existing windows inside the building, or alternatively, part of the window can be replaced with PV modules in a similar manner to rooftop installation. Retrofitting old buildings and incorporating PV modules into new buildings is one of the most effective ways to expand the use of PVs.
PV modules integrated into window facets are usually see-through or semi-transparent and controlling the balance between electricity generation and interior lighting is an important consideration. Recently, some see-through PV modules have been developed and installed commercially. Most conventional see-through PV modules consist of Si solar cells embedded (sandwiched) at certain intervals in flat transparent sheets (see [1–3]). The amount of light energy transmitted through the interval is proportional to the ratio of the interval area to the total area of the module. In other types of modules, dye-sensitized solar cells or thin film solar cells can be used directly as see-through PV modules because they allow light transmission (see [4–6]). In this case, the amount of light transmitted is basically proportional to the transmittance of the cell. As a result of this transparency, installing see-through PV modules is likely to lead to energy savings in office buildings due to changes in the indoor thermal conditions [7,8].
Prism solar concentrators exploiting total internal reflection (TIR) within a material of high refractive index were described by Mills and Giutronich [9] as a family of nonimaging concentrators which includes compound parabolic concentrators [10]. Since then, several types of concentrating photovoltaics (CPV) modules with a low-concentration static prism concentrator (hereafter referred to as a prism CPV module) have also been reported [11,12]. Modules aimed reducing the use of solar cells and, thus cost, do so by replacing expensive solar cells with a cheap concentrator that is made of plastic and glasses. Prismatic optical structures have been used not only for CPV modules but also for functional control of lighting [13]. These developments have inspired us to investigate the concept of see-through prism CPV modules. To our knowledge, the feasibility of this concept has not been discussed in detail to date. The optical characteristics, such as incident angle dependency and practical performance, of see-through prism CPV modules should be quantitatively clarified.
In this paper, we design a see-through prism CPV module and test its performance via ray-tracing analysis and an on-site experiment. The incident angle dependency of the prism CPV makes it possible to concentrate direct solar radiation onto solar cells and allow diffuse solar radiation to be transmitted into the building interior. Owing to this characteristic, fewer solar cells can be used without any sacrifice in module conversion efficiency or lighting performance because direct solar radiation tends to be more intense than diffuse solar radiation. In addition, to examine the applicability of the proposed module, we report an example design method for the concentrator geometry to adjust the incident angle dependency for different latitude and tilt angles.
2. Ray-tracing analysis
Figure 1 (a) shows a schematic cross-section of a prism CPV module illustrating the working principle. This module differs from past studies [9,11,12] in that there is no reflective coating on facet AB. When the angle of the incident light θin satisfies the following equation,
the incident ray is concentrated onto the solar cell due to TIR. Here, θcr is the critical angle of TIR calculated via θcr = sin−1(1/n), where n is the refractive index of the material, and θin is defined for the case when φ = 0. The coordinate system for the analysis presented here is defined in Fig. 1 (b).
Fig. 1 Schematic of see-through prism CPV module. (a) Cross-section of the module and working principle. (b) Coordinate system for the ray-tracing analysis.
The angular optical efficiency of concentration ηPV (θ,φ) is defined as the ratio of the concentrated energy onto the solar cell to the total energy incident on the aperture of the module. For example, ηPV (0,0) represents the efficiency of a ray incident in a direction normal to the module aperture. To obtain ηPV (θ,φ), a number of rays incident on the entire area of the module aperture are required for each discrete incident angle. In the analysis, the geometric and optical parameters were given as: n = 1.49, θA = 28°, AC = 26 mm, and BC = 13 mm. This gives a geometric concentration ratio of Cg = AC/BC = 2. The Fresnel reflections at the air-material interface were rigorously simulated, however, volumetric absorption by the material was neglected.
Figure 2 shows the ηPV (θ,φ) distribution obtained from the ray-tracing analysis. The path of the sun at equinox and solstice times for the module installed vertically and facing south and east at Nagaoka city in Japan, where the experiment was carried out, are also shown. It was found that the module concentrates light over a wide incident angle range covering a large part of the yearly sun path. The highest efficiency (96.0%) appears at (θ,φ) = (20°, ± 22°). The efficiency is ηPV ≥ 70% from 07:22 to 16:23 at equinox times, from 10:25 to 13:07 for the summer solstice, and from 07:02 to 16:26 for the winter solstice. Furthermore, to a good approximation, the optical efficiency for daylight lighting can be estimated as ηDL (θ,φ) = 0.96 − ηPV (θ,φ), and this has been confirmed by the ray-tracing analysis. The 0.04 difference from unity accounts for a mean reflection loss. These results indicate that the present see-through prism CPV module allows more light to pass through with a larger incident angle and, therefore, more diffuse radiation penetrates into the room.
Fig. 2 Distribution of angular optical efficiency of concentration to solar cell. Black lines represent an example sun path when the module is affixed on a vertical surface in Nagaoka city, Japan (lat.: 37.44°N, long.: 138.85°E). Solid line: module facing south; dashed line: module facing east.
3. Experiment
Outdoor experiments were conducted using a test module that had the same geometric and optical properties as shown in the ray-tracing model. Objectives of the outdoor experiment are to test the concentration performance of the prism concentrator and to clarify agreement between experiment and simulation which used the result of ray-tracing analysis. Figure 3 shows photographs of the fabricated prism CPV module. The prism concentrator is made of acrylic resin (mainly PMMA) and was coupled with single-crystalline Si solar cells using optical glue.
Fig. 3 Photographs of the fabricated see-through prism CPV module. (a) Small incident angle view. (b) Large incident angle view. (c) Sample view.
Figure 4 shows a schematic diagram and photograph of the outdoor experimental system. For comparison, a conventional see-through PV module (labeled as “Flat PV” in the figures) made with the same single-crystalline Si solar cells was prepared. The ratio of solar cell area to aperture area is 0.5 for the prism CPV module (i.e., Cg = 2) and 0.87 for the conventional module. In other words, the prism CPV module solar cell area is equivalent to 57.5% of the solar cell area of the conventional module. The modules were placed vertically (i.e., 90° tilt) facing south at Nagaoka University of Technology (lat.: 37.44°N, long.: 138.85°E). A current–voltage (I-V) curve tracer EKO MP-160 measured the time variation of the I-V characteristics of the modules and a sun tracker EKO STR-21 equipped with a pyrheliometer (full opening view angle: 5° ± 0.2°) was used to measure the direct normal irradiance (DNI). The global (broadband solar) irradiance incident on the vertical and horizontal surfaces was also measured by two pyranometers. The diffuse component of the solar irradiation was estimated as the difference between the global irradiance and DNI. The incident angle (θ,φ) was calculated from the orientation data of the sun tracker.
Fig. 4 Apparatus for outdoor experiment. (a) Schematic diagram of the system. (b) Photograph of the system. The tilt angle of the PV modules is 90°, facing south. Location: Nagaoka city, Japan.
Results
Figure 5 shows the daily variations in the generated power of the modules at maximum power obtained from the I-V characteristics on June 6, 2010, a clear day, together with the irradiation data. The sun path on June 6 is close to that of the summer solstice shown in Fig. 2. The generated power per solar cell area is shown in Fig. 5(a) and Fig. 5(b) shows the power per module aperture area. Simulated variations calculated from ηPV (θ,φ), and the measured DNI and diffuse irradiation are also plotted. From the experiment, it was found that the see-through prism CPV module generated 1.4 times more power than the conventional module per solar cell area during 10:00 to 13:30 when the DNI was high. This indicates that actual optical efficiency ηPV was 70% during that period, which agreed with the result of ray-tracing analysis. On the other hand, the resultant power per module aperture area for the CPV module is only equivalent to 68% of the power for the conventional module. However, it must be noted that transparent area of the conventional module was much less than that of the prism CPV module in this experiment. Figure 5 also shows that the simulated variations fitted fairly well with the experimental variations within a 7% error range for both the prism CPV and the conventional module. We also carried out the same experiments and simulations on clear and cloudy days over half a year and obtained a similar level of agreement.
Fig. 5 Results of the outdoor experiment on June 6, 2010. (a) Daily variations in generated power per solar cell area. (b) Daily variations in generated power per module aperture area. Simulated variations are shown in red and blue. (c) Measured daily variation of global and direct radiation onto module aperture.
By using computer simulation, the yearly and daily performances of the present see-through prism CPV module were estimated based on irradiation data from Tokyo, one of the biggest potential markets for BIPV. Figure 6 shows the daily performance of the generated power and daylight lighting for vertically mounted modules facing south on a typical clear summer day. Figure 7 shows the yearly performance for the modules facing south, southeast and east. For comparison, the performance of the conventional see-through PV module, which achieves the same daily and yearly total amount of daylight lighting as the prism CPV module, was also simulated and is shown in Fig. 6 and Fig. 7, respectively. Here, the conventional module was assumed to use the same efficient Si solar cells.
Fig. 6 Simulated daily summer performance of the (a) see-through prism CPV module and (b) conventional see-through Flat PV module in Tokyo, Japan. The modules face south at 90° tilt. The total amount of transmission (lighting) of the prism CPV is equivalent to that of the Flat PV.
Fig. 7 Simulated yearly performance comparison between the see-through prism CPV module and conventional see-through Flat PV module facing (a) south, (b) southeast, and (c) east in Tokyo, Japan. The modules are mounted vertically. The total amount of transmission (lighting) of the prism CPV is equivalent to that of the Flat PV.
From Fig. 6, in order to ensure production of the same daily total lighting energy, the prism CPV module generates more electricity than the conventional one. This is emphasized at midday on clear summer days due to strong direct radiation. Furthermore, the prism CPV module utilizes more diffuse light for interior lighting than the conventional module. It should be noted that, only in early morning and early evening, the prism CPV tends to allow more direct radiation to penetrate the room than the conventional module. At noon, the electricity generation of the prism CPV module is 1.15 times higher than that of the conventional module with 63% of the solar cell area that is used in the conventional module. In addition, the ratio of diffuse radiation to total daylight lighting is 66% for the prism CPV module, while it is only 32% for the conventional module. This may be preferable for indoor thermal comfort because direct radiation is likely to be strong, especially in summer months. Figure 6 also shows a schematic projected front view of the prism CPV and the conventional modules defined by two geometric parameters: a see-through ratio γ and solar cell ratio χ. The see-through ratio γ is defined as the ratio of the projected transparent part area to that of the non-transparent part, and the solar cell ratio χ is defined as the ratio of the solar cell area to the module aperture area. A higher γ and lower χ are key features of the prism CPV module as compared to the conventional module. For the same see-through ratio, the prism CPV generates 2.23 times more power than the conventional module.
Examining the yearly performance comparison in Fig. 7, in order to achieve the same yearly total amount of lighting energy, the prism CPV still has better γ and χ ratios for different orientations. In addition, it is clear that the diffuse transmission accounts for a greater portion of the yearly lighting energy in the prism CPV module than in the conventional module. However, the advantage in electricity generation of the prism CPV module is not remarkable (approximately 1.07 times greater than the conventional one) in yearly operation compared to the case of daily operation on a clear summer day. However, if the comparison is made for the same see-through ratio, the prism CPV still generates a greater amount of electricity than the conventional module, i.e., by a factor of 2.10–2.34.
4. Prism geometry design method
The performance and advantages of the see-through prism CPV were described in the previous section. However, the module was assumed to be mounted vertically for all cases. Therefore, in order to ensure the applicability of the proposed module, it is important to introduce an amount of controllability over the incident angle dependency and be able to design the prism geometry so that the module can be applied in different places (i.e., different yearly solar elevation) and at different tilt angles.
Figure 8(a) shows how the incident angle depends on the prism geometry when keeping the simple prism shape. In Fig. 8(a), the angle θC was varied while keeping the geometric concentration ratio constant. The vertical axis represents the optical efficiency of concentration to the solar cell ηPV (θin,0°). The simple prism concentrator tends to perform well only in the range + 15° <θin < + 50°, which may limit the use of the present geometry. Figure 8(b) shows the incident angle dependency after adding sub-geometry to the TIR facet of the prism (facet AB). An evolutionary algorithm developed by one of authors [14] was used to generate the best possible sub-geometry for a set incident angle range so that the concentrator has the highest optical efficiency around θin = 0°. It is clear that the incident angle for peak optical efficiency was successfully shifted close to θin = 0°. This demonstrates that such sub-geometry affects the optical efficiency over a wider range than the simple prism geometry. This means that the module discussed here could also be used for small tilt angle facades after minor modifications.
Fig. 8 Incident angle dependency. (a) Simple prism geometry. (b) Example of an evolutionary algorithm generated prism geometry.
5. Conclusions
In this paper, the performance of a see-through photovoltaic module combined with a low-concentration prism concentrator was elucidated by carrying out ray-tracing analysis and experiment. It was found that the geometry of the prism concentrator makes it possible to concentrate direct solar radiation onto the solar cells and allows diffuse solar radiation to be transmitted into the interior of the building. Owing to these characteristics, the use of solar cells can be reduced without sacrificing the module conversion efficiency or lighting performance.
Results show that, at noon in summer in Japan, the electricity generation per module aperture of the prism CPV module is 1.15 times higher than that of the conventional module for the same daylight performance, even though the prism CPV module uses 63% less solar cell area than the conventional module. Emphasizing the total see-through area, i.e., the transparent rather than transmitted amount of lighting energy, the present module achieves approximately twice as much power generation than the conventional model. Furthermore, in order to ensure the applicability of the proposed design, we examined the incident angle dependency and presented a design method for the concentrator geometry to adjust for different latitude and tilt angles. An evolutionary algorithm generated sub-geometry in the prism that successfully widened the optical efficiency angular range of the prism concentrator. We expect that the present results will trigger a wide variety of window integrated photovoltaic module designs incorporating a solar concentrator.
References and links
1. D. H. W. Li, T. N. T. Lam, and K. L. Cheung, “Energy and cost studies of semi-transparent photovoltaic skylight,” Energy Convers. Manage. 50(8), 1981–1990 (2009). [CrossRef]
2. K. E. Park, G. H. Kang, H. I. Kim, G. J. Yu, and J. T. Kim, “Analysis of thermal and electrical performance of semi-transparent photovoltaic (PV) module,” Energy 35(6), 2681–2687 (2010). [CrossRef]
3. M. Biancardo, K. Taira, N. Kogo, H. Kikuchi, N. Kumagai, N. Kuratani, I. Inagawa, S. Imoto, and J. Nakata, “Characterization of microspherical semi-transparent solar cells and modules,” Sol. Energy 81(6), 711–716 (2007). [CrossRef]
4. M. Grätzel, “The advent of mesoscopic injection solar cells,” Prog. Photovolt. Res. Appl. 14(5), 429–442 (2006). [CrossRef]
5. A. Takeoka, S. Kouzuma, H. Tanaka, H. Inoue, K. Murata, M. Morizane, N. Nakamura, H. Nishiwaki, M. Ohnishi, S. Nakano, and Y. Kuwano, “Development and application of see-through a-Si solar cells,” Sol. Energy Mater. Sol. Cells 29(3), 243–252 (1993). [CrossRef]
6. Y. Hamakawa, Thin-Film Solar Cells: Next Generation Photovoltaics and Its Applications (Springer-Verlag, 2004)
7. T. T. Chow, K. F. Fong, W. He, Z. Lin, and A. L. S. Chan, “Performance evaluation of a PV ventilated window applying to office building of Hong Kong,” Energy Build. 39(6), 643–650 (2007). [CrossRef]
8. T. Miyazaki, A. Akisawa, and T. Kashiwagi, “Energy savings of office buildings by the use of semi-transparent solar cells for windows,” Renew. Energy 30(3), 281–304 (2005). [CrossRef]
9. D. R. Mills and J. E. Giutronich, “Ideal prism solar concentrators,” Sol. Energy 21(5), 423–430 (1978). [CrossRef]
10. R. Winston, J. Minano, and P. Benitez, Nonimaging Optics (Elsevier Academic Press, 2005).
11. T. Maruyama and S. Osako, “Wedge-shaped light concentrator using total internal reflection,” Sol. Energy Mater. Sol. Cells 57(1), 75–83 (1999). [CrossRef]
12. T. Uematsu, Y. Yazawa, Y. Miyamura, S. Muramatsu, H. Ohtsuka, K. Tsutsui, and T. Warabisako, “Static concentrator photovoltaic module with prism array,” Sol. Energy Mater. Sol. Cells 67(1-4), 415–423 (2001). [CrossRef]
13. G. Walze, P. Nitz, J. Ell, A. Georg, A. Gombert, and W. Hossfeld, “Combination of microstructures and optically functional coatings for solar control glazing,” Sol. Energy Mater. Sol. Cells 89(2-3), 233–248 (2005). [CrossRef]
14. N. Yamada and T. Nishikawa, “Evolutionary algorithm for optimization of nonimaging Fresnel lens geometry,” Opt. Express 18(Suppl 2), A126–A132 (2010). [CrossRef] [PubMed]
