Photovoltaics on flat roofs: Energy considerations

Energy 36 (2011) 1996e2010 Contents lists available at ScienceDirect Energy journal homepage: www.elsevier.com/locate/energy Photovoltaics on flat roofs: Energy considerations Angel A. Bayod-Rújula*, Abel Ortego-Bielsa, Amaya Martínez-Gracia CIRCE/University of Zaragoza, Zaragoza, Spain a r t i c l e i n f o a b s t r a c t Article history: Received 24 November 2009 Received in revised form 7 April 2010 Accepted 10 April 2010 Available online 15 May 2010 Flat roofs present a large potential of suitable areas for installation of PV (photovoltaic) plants. Flat roof PV installations have the advantage of being able to be optimally positioned with support structures, and the inclination angle can be adjusted. Due to the important technological development existing in the PV sector, there are different PV technologies in the market, whose energy and economic features substantially differ. This paper describes some useful parameters to assess the technology and distribution of modules to be installed in flat roofs and terraces of buildings. The effect on the energy parameters of the modules tilt and disposition is analyzed in a case study, considering different technologies. Ó 2010 Elsevier Ltd. All rights reserved. Keywords: Photovoltaics Grid-connected photovoltaics Building integrated PV systems BIPV Flat roofs 1. Introduction Energy sector is constantly changing. With growing worries about the climate change and the high energy dependence of many developed countries, the research and new developments in renewable energies are become more and more needed. Photovoltaics are called to play a fundamental role in the embedded generation of electricity from renewable sources. The continuous development and advances in the PV sector facilitate new solutions for the solar radiation conversion, giving rise to less costly production rates. This reduction of prices together with the rising price of electricity is expected to contribute to the equivalence of prices between the electricity produced on PV installations and the grid electricity, i.e., the grid-parity. The building envelope provides a number of possibilities for the integration or superposition of PV [1e6]. The four main options are: - Slope roof Flat roof Façade applications Shading systems Abbreviations: a-Si, amorphous silicon; BIPV, building integrated photovoltaics; CdTe, cadmium telluride; CIS, copper indium selenide; De, density of energy; Dpp, density of peak power; EPBT, energy payback time; ERF, energy return factor; ES, energy surplus; GER, gross energy requirement; IRR, internal return rate of the investment; m-Si, monocrystalline silicon; NPV, net present value; PV, photovoltaic; p-Si, polycrystalline silicon; STC, standard test conditions; Yf, final system yield. * Corresponding author. Tel.: þ34 976 761920; fax: þ34 976 762226. E-mail address: [email protected] (A.A. Bayod-Rújula). 0360-5442/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.energy.2010.04.024 In sloped roofs, the characteristics of the roof (slope, azimuth, available surface) determine the performance of the PV installation. In flat roofs the modules can be optimally positioned with supporting structures, and the inclination angle can be adjusted. The optimum use of the available area has to be carefully observed. The technical requirements in façade installations are higher than for a flat of sloped roof installation, because of the wiring and the junction boxes, which have to be hidden, and the increased difficulty in fixing the array to the building. PV sunscreen devices are installed in shading systems; besides the generation of electricity they achieve the shadow of the building in summer. 2. Methodology and parameters of interest The sensitivity analysis of the PV production versus the inclination of the modules, the distance among module rows and space constrains and their consequences on the energy and economic parameters are of high interest for PV systems on roofs, terraces and covers [7e14]. The usual methodology starts from the analysis of the solar potential in the building location, by means of in-situ measurement devices or the use of solar radiation data bases. Next, the characteristics of the building are included, in particular the orientation and inclination of the areas free of shadows and other aspects related to the mechanical resistance and the visual impact. The optimum design of a PV system is determined by a complex combination of economic, technical, environmental and social considerations. The economic viability of a PV installation is determined by the profitability of an investment decision or the cash flow implications of the project. This involves taking into A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 1997 Fig. 1. Torres Quevedo Building, University de Zaragoza. account all costs and revenues of a project and allowing for the different timing of these flows and the potential interest charged on the funds involved. From an economic point of view, two parameters are of great importance: the IRR (internal return rate of the investment) and the NPV (net present value). The IRR determines the financial risk assumed in the project. It is especially important for the projects where external financing is required and for those whose cash flows are negative during the first years. NPV is defined as the total present value of a time series of cash flows. It is a standard method for using the time value of money to appraise long-term projects. It is an indicator of how much value an investment or project adds to the firm. In order to perform a complete analysis, the different PV technologies available in the market for the chosen configuration (flat roofs, sloped roofs or façades) can be included. Finally, those different technologies and their assembly possibilities have to be considered, attending to the suitable installation capacity, the annual production, the PV contribution to the electricity demand in the building, the costs, and the environmental impact, among others. The currently available simulation software is an important help for this analysis. The following parameters have been used in this paper to evaluate the energy behaviour of the different PV configurations on buildings: - Density of peak power, Dpp, (Wp/m2ground surface). This parameter is defined as the ratio between the peak power installed and the available ground. With this ratio it is possible to determine the solution that allows installing the highest power of the PV generation field in the building (in Standard Test Conditions, STC: Irradiance level of 1000 W/m2, solar spectral irradiance distribution AM 1.5 and cell temperature of 25  C). - Final system Yield, Yf, (kWh/kWp). This ratio expresses the energy produced per year by the unit of peak power installed, and it helps to determine the solutions with highest specific productivity. - Yearly energy per unit of collection area, (kWh/m2collection). This ratio is defined as the annually produced energy per unit of generation surface. It expresses the global efficiency referred to the area of the generation field. - Density of energy, De (kWh/m2ground surface). This ratio determines the energy produced (yearly) per unit of flat roof occupied with the PV installation. - Energy return factor, ERF, (dimensionless). The energy return factor is defined as the ratio of the total energy generation during the system operation lifetime and the total energy input during the system life cycle. ERF ¼ Energy ouput ðlife operationÞ Energy input Energy payback time, EPBT (years). The energy payback time is defined as the ratio of the total energy input during the system life cycle and the yearly energy generation during system operation. EPBT ¼ Energy input Energy output ð1 year operationÞ Energy surplus, ES, (kWh) is the gross amount of energy injected into the grid, minus the energy used directly and indirectly in the manufacture, assembly, operation and recycling of the whole PV system. This parameter has a strong relationship with the amount of CO2 emissions avoided. In this paper the evolution of these parameters with the distance among rows of modules or inclination angle for a specific location it is shown. The interest of the explained parameters is different depending on the specific requirement of the projects. 3. Case study As a case study, the possibilities of a PV installation on the Torres Quevedo Building, located in the Campus of the University of Zaragoza (Spain) are presented. A large flat roof exists in the central part of that building as it can be seen in Fig. 1. In order to analyze the energy performance of a PV plant it is necessary to establish the irradiation level and some climatic conditions. In absence of real measurements, the available solar Fig. 2. Monthly irradiation on horizontal surface in Zaragoza. 1998 A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 potential can be evaluated with data bases. The coordinates of the building are: Table 1 Characteristics of the modules. Technology Si mono Si poly CIS CdTe si-H Manufacturer Model Power (Wp) Efficiency (%) Voc (V) Isc (A) Vmpp (V) Impp (A) Length (mm) Width (mm) Weight (kg) Frame Wp/m2 Solarfun SF080M5-18 90 13.6 22.5 5.2 18.8 4.85 1210 554 8 Alum. 134.26 Photowatt PW 1000/105 105 11.86 43.2 3.2 35.9 2.97 1335 673 10.5 Alum. 116.87 Wurth WS11007/75 75 10.56 44.5 2.4 35 2.2 1205 605 13 Alum. 102.88 First Solar FS-275 75 10.41 92 1.2 69.5 1.08 1200 600 12 Laminate 104.17 Kaneka T-EC120 120 6.39 91.8 2.38 64.4 1.89 1919 990 28 Alum. 63.16 Latitude: 41 410 0100 Longitude: 0 530 1400 The detailed calculation has been performed with the software Meteonorm 6.0, where the solar potential is calculated by the Hay’s Model of radiation. The monthly values of irradiation on horizontal surface in Zaragoza according to Meteonorm are shown in the Fig. 2. PVSyst 4.34 [15] is the software package used in this work for the estimation of the yearly energy produced. It has been developed by the University of Geneva (Switzerland), for the study, sizing, simulation and data analysis of complete PV systems. It is suitable for a 1600 1550 1500 kWh/kWp 1450 1400 1350 1300 1250 1200 1150 1100 0 10 20 30 40 50 tilt Solarfun b Photow att Wurth Firstsolar Kaneka 1.15 1.1 kWh/kWp 1.05 1 0.95 0.9 0.85 -20 -10 0 10 20 30 40 50 tilt Solarfun Photowatt Wurth Firstsolar Kaneka Fig. 3. Evolution of the final system yield with the tilt angle, (a) in absolute values; (b) in per unit with respect to the value obtained with horizontal modules. A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 a 1999 200,00 180,00 kWh/m 2captation 160,00 140,00 120,00 100,00 80,00 60,00 40,00 0 10 20 30 40 50 tilt Solarfun b Photow att Wurth Firstsolar Kaneka 1.15 kWh/m 2captation 1.10 1.05 1.00 0.95 0.90 0.85 -20 -10 0 10 20 30 40 50 tilt Solarfun Photowatt Wurth Firstsolar Kaneka Fig. 4. Evolution of the yearly energy per unit of collection area with the tilt angle; (a) in absolute values; (b) in per unit with respect to the value obtained horizontally. grid-connected, stand-alone, pumping and DC-grid (public transport) systems, and offers an extensive meteorological and PVcomponents database. PVSyst has become reference software in the sector. Meteonorm generates a data file exportable to PVSyst. Commercial modules of five PV technologies (monocrystalline Silicon, m-Si; polycrystalline Silicon, p-Si; copper indium diselenide CIS; cadmium telluride, CdTe and amorphous Silicon, a-Si) have been considered. The studied modules and their physical and electrical characteristics are shown in Table 1. They are quite different in efficiency and price. The location chosen to install the PV system in the building must allow the integration or superposition of the modules avoiding shadows. The operation steps to achieve the best configuration imply the design of the building in a graphic design software able to carry out the shadows simulation. In this case, the commercial software SketchUp has been used. It allows the importation of typical draw files such as dwg or dxf. Fig. 1 shows the result of the building modelled by means of the mentioned software. 4. Sensitivity analysis of the different parameters with the inclination angle of the modules A first analysis evaluates the behaviour of the different modules with the change in the angle of inclination with which they are 2000 A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 Fig. 5. Modules mounted with a tilt of 30 and horizontally on the central flat roof of the Torres Quevedo Building. mounted. This analysis will determine the most favourable tilt and also it determines what modules are more sensitive to being installed by an angle of unfavourable inclination with regard to the ideal angle. The study has been performed with angles from 10 to 45 . The indicators to analyze will be the final system yield and the yearly energy per unit of collection area. The evolution of the final system yield with the angle of tilt of the modules is shown in Fig. 3. Maximum specific productivity is obtained in Zaragoza at 30 . According to the simulation with PVSYST 4.34, the best values are obtained for thin-film modules and the sensitivity is similar for the five modules (a bit lower in the case of the Kaneka amorphous Si). The efficiency in Standard Test Conditions of each module is different, so the production per unit of surface of module is also very different. The ratio yearly energy per unit of collection area is shown in Fig. 4. The best value is for Solarfun m-Si with optimum tilt (177.8 kWh/m2collection) Kaneka a-Si offers 91.9 kWh/m2collection (the efficiency is less than a half in the case of the a-Si modules than in the case of m-Si modules chosen in this study). In the case of the amorphous modules (Kaneka), the electrical generation is not very different for non-optimum angles. As a consequence, attending to its lower cost, it is reasonable to install these modules horizontally. Besides, there are no shadows between modules and the roof area is better used. 5. Sensitivity analysis of the different PV modules with the distance among different rows of modules Knowing the angle with which it is more suitable to realize the installation from the point of view of the specific production Yf (kWh/ kWp) it is necessary to determine in the facilities located in flat roofs what distance is best adapted to locate the rows of modules. Shading losses can be caused not only by obstacles but also by the arrays of modules themselves. The choice of the distance among rows influences the values of the parameters of study for PV facilities. Beam h0 α d Fig. 6. Distance d among structures, angle of tilt, a, and shading angle limit, h0. We have a trade-off in PV installation on flat roofs: - In Zaragoza, the best yield per square metre of active area is obtained at a tilt angle of about 30 . However, the PV modules cause shading and only a part of the flat roof area can be used. Hence, implying a lower ratio “module area/roof area” (see Fig. 5). - Most roof area can be used by horizontal integration as the PV modules have no shading effects, hence implying a ratio modules area/roof area of 1. With tilted structures to put an excessive distance will imply minor losses for shading but due to the limitations of space, it will provoke a lower production of energy per unit of surface of flat roof. On the other hand, if a too small distance is chosen, the yearly production of energy per unit of surface will increase (as a consequence of the highest installed power) but the final system yield (kWh/kWp) and the yearly energy per unit of collection area (kWh/m2collection) will drop. This fact will difficult the economic viability of the installation. For optimization of solar yield and the ratio module area/roof area a further element has to be considered: the shading angle limit (h0 in Fig. 6). The greater the angle of inclination of the PV modulearray and the closer the PV rows are, the higher the shading angle and the more the final system yield and the yearly energy per unit of collection are affected by shading. PVSyst can define and perform simulations with two kinds of near shadings: a) Linear shadings: The shading factor is the ratio of the effective shaded area to the total field area. This is shading at irradiance level, which leads to a minimum of shading loss. b) Module string shadings: In a string, the poorer cell drives the whole string current. That is, when one cell is shaded, the entire string is affected, and becomes almost inefficient within a set of several strings. Of course, this is not quite exact, but should represent an upper limit of the shading effect taking electrical behaviour into account. The real situation should lie between these two limits, so, to obtain a more real result, all simulations have been made with both simulation options, and the final result is the average of linear string shading and module string shading. To analyze this influence we can start by choosing a distance of reference among modules, calculated as the minimal distance that allows the absence of shading for the above mentioned modules at 12:00 of 21/12, and then analyze the effect of the variation of the distance in a range from a 15% of decrease up to an increase of 25%. In the following graphs, the results appear in absolute value and in relative value with regard to the existing value to that of reference. Afterwards, knowing the performance of the modules with an inclination of 30 (mounted at the distance of reference), we can compare the parameters with those obtained for an installation with modules horizontally installed. A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 2001 a 90,00 80,00 60,00 50,00 40,00 30,00 Wp/m 2groundsurface 70,00 20,00 10,00 Kanek a Firs ts olar Wurth Solarfun b Photowatt 0,00 20 15 10 5 -5 0 -10 -15 Dist (%) 25 20,00 15,00 Wp/m 2groundsurface (%) 10,00 5,00 0,00 -5,00 -10,00 -15,00 Kaneka Firstsolar Wurth Photow att Solarfun -20,00 -15 -10 -5 0 5 Dist (%) 10 15 20 25 Fig. 7. Evolution of the density of peak power with distance among rows. (a) absolute values; (b) percentage of variation. 5.1. Sensitivity of the density of peak power (Wp/m2ground with the distance among rows of modules surface) The distance among rows influences the nominal power that can be installed per unit of flat roof surface. The results of the simulation are shown in Fig. 7(a). From the point of view of a greater peak power installed in the surface of the building, the most efficient modules present the best numbers (70.2 Wp/ m2ground surface for Solarfun(m-Si)). When this ratio is of main interest, for instance for reduction of the peak power of the building, thin-film technologies are not advisable for their low efficiency. Fig. 7(b) shows the percentage of increase (þ) or reduction ( ) with respect to the reference distance. The behaviour is identical for the simulated modules. An increase of a 25% in the distance among rows implies a reduction of a 20% in the peak power installed. On the other hand, a 15% reduction in the distance among rows allows a 17.6% increase in density of peak power. 5.2. Sensitivity of the final system yield with the distance among rows of modules The ratio kWh/kWp (Yf) shows the energy that it is possible to generate with regard to the installed nominal power. This parameter 2002 A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 a 1600,00 1500,00 1300,00 kWh/kWp 1400,00 1200,00 1100,00 b Photowatt Solarfun Firs ts olar Wurth Kanek a 1000,00 20 15 10 5 -5 0 -10 -15 Dist (%) 25 20,00 15,00 kWh/kWp (%) 10,00 5,00 0,00 -5,00 -10,00 -15,00 Kaneka Firstsolar Wurth Photow att Solarfun -20,00 -15 -10 -5 0 5 Dist (%) 10 15 20 25 Fig. 8. Evolution of the final system yield with distance among rows. (a) absolute values; (b) percentage of variation. reflects also the profitability from the point of view of the energy production. This indicator enormously depends on the installed technology due to the different temperature dependences and use of the solar spectrum. The effect of shading is also included in this parameter. The results of this analysis are shown in absolute values (Fig. 8 (a)) and in percentage of increase or decrease as the distance among rows of modules varies (Fig. 8(b)). In those cases where the investment is fixed, there are enough room and the use of the area has not to be optimized, the best solutions are those that minimize the shading and maximize the production, with rows of modules distant enough to avoid shading. Thin-film technologies are a good option for these cases since, in spite of having less nominal power and low energy production they have the highest specific production (final system yield). An increase of a 25% in the distance among rows allows only a 2% more in the Yield factor. On the other hand, decreasing a 15% produces a reduction of nearly a 5% in the final system yield. A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 2003 Fig. 9. Evolution of the yearly energy per unit of collection area with distance for each module. (a) absolute values; (b) percentage of variation. Notice that the variation of the values of the parameter with the distance among rows is not so strong that in the case of the ratio Wp/m2ground surface. 5.3. Sensitivity of the yearly energy per unit of collection area with the distance among rows of modules The behaviour of the yearly energy per unit of collection area indicates the goodness of the installation from the point of view of the energy production. The modules that offer a major generation per surface unit are the most efficient. The values of this parameter are strongly affected by the shading (Fig. 9(a)),To change the distance affects the production per unit of collection area, similar to the study of the final system yield, as it is shown in Fig. 9(b). 5.4. Sensitivity of the density of energy and energy surplus with the distance among rows of modules The ratio kWh/m2ground surface (density of energy) reflects the use of the roof from the point of view of the yearly generation of energy. 2004 A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 a 80,00 60,00 40,00 kWh/m 2groundsurface 100,00 20,00 b Kanek a Firs ts olar Wurth Photowatt Solarfun 0,00 10 15 20 5 0 -5 -10 -15 Dist (%) 25 20,00 kWh/m 2groundsurface (%) 15,00 10,00 5,00 0,00 -5,00 -10,00 -15,00 Kaneka Firstsolar Wurth Photow att Solarfun -20,00 -15 -10 -5 0 5 Dist (%) 10 15 20 25 Fig. 10. Evolution of the density of energy with the distance for each module; (a) absolute values; (b) percentage of variation. It is a parameter of big interest when the aim of the PV installation is the generation of the electricity needed in the building. Generation at the point of use avoids transmission and distribution of electricity and the costs and losses associated. This argument is particularly strong for PV systems on commercial buildings where the demand for electricity (9 a.m.e5 p.m.) typically coincides with supply of electricity from PV systems. The results of this analysis in the central flat roof of the Torres Quevedo building are shown in Fig. 10(a) and (b). In Fig. 10(b) it is shown that an excessive increase in the distance among rows of modules produces a strong fall of the density of energy. An increase of 25% in the distance among rows produces a drop of 18.3% in this parameter. On the other hand, a reduction of 15% in the distance among rows allows a 12% increase in annual production. The most efficient modules are most adapted to optimize the energetic use of the available surface. Multiplying this value times the total surface of the flat roof (752.4 m2 in the case of the central flat roof of the Torres Quevedo A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 2005 a 4,00 3,50 3,00 2,00 EPBT 2,50 1,50 1,00 0,50 Firs ts olar Kanek a Photowatt Wurth Solarfun 0,00 20 15 10 0 5 -5 -10 -15 Dist (%) 25 b 20,00 15,00 10,00 EPBT (%) 5,00 0,00 -5,00 -10,00 -15,00 Kaneka Firstsolar Wurth Photowatt Solarfun -20,00 -15 -10 -5 0 5 Dist (%) 10 15 20 25 Fig. 11. Evolution of the EPBT with the distance among rows. (a) absolute values; (b) percentage of variation. Building) gives an indication of the total energy injected yearly into the grid. Considering depreciation in the performance of the modules along their useful life (according to the guaranties indicated by the manufacturers), the gross amount of energy injected into the grid along that time (usually more than 25 years) can be obtained. The gross amount of energy injected minus the energy used directly and indirectly in the manufacture, assembly, operation and recycling of the whole PV system provides an indication of the Energy Surplus of the PV installation. This parameter has also a strong relationship with the amount of CO2 emissions avoided. 5.5. Sensitivity of the energy payback time and energy return factor with the distance among rows of modules Until this section the study has shown the energy produced by the installation or the use of the area of the building, but it has not 2006 A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 a 18,00 16,00 14,00 10,00 8,00 ERF 12,00 6,00 4,00 2,00 Solarfun Wurth Photowatt Kanek a Firs ts olar 0,00 25 20 15 10 5 -5 0 -10 -15 Dist (%) b 20,00 15,00 10,00 ERF (%) 5,00 0,00 -5,00 -10,00 -15,00 Kaneka Firstsolar Wurth Photowatt Solarfun -20,00 -15 -10 -5 0 5 Dist (%) 10 15 20 25 Fig. 12. Evolution of the ERF with the distance among rows. (a) absolute values; (b) percentage of variation. been studied what happens with the parameters that reflect the behaviour of the energy payback of the plant. The parameters EPBT and ERF indicate the balance of the energy generated with regard to the energy that is consumed during its manufacture and assembly (and even recycling). The EPBT is expressed in years and indicates the time that passes until the energy generated by the modules equals to the energy needed for its manufacture and assembly. The ERF is dimensionless and indicates the number of times that the energy generated by the modules along the useful life overcomes the energy needed in their manufacture and assembly. Then, another consideration to bear in mind is the energy viability of the installation, since low levels of specific production would lengthen the period of energy amortization (EPBT) and would reduce the ratio between the energy produced by the installation and the energy required in its manufacture and assembly (ERF). To carry out the mentioned analysis a precise knowledge of the Gross Energy Requirement (GER) of the modules is required. The considered values in this paper taken from the literature, are: Monocrystalline Silicon: 5200 MJ/m2, Alsema et al. (2006) [16] Polycrystalline Silicon: 4000 MJ/m2, Alsema et al. (2006) [16] A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 2007 Fig. 13. Comparison of the parameters for 30 and 0 installation of modules (a) Density of peak power, (b) Yield factor, (c) Density of energy, (d) Yearly energy per unit of collection area, (e) Internal return factor, (f) Energy payback time, (g) Energy return factor, (h) Energy Surplus, (i) Net Present Value, (j) CO2 avoided emissions. CIS: 4053 MJ/m2, Raugei et al. (2007), [17] CdTe: 2281 MJ/m2, Raugei et al. (2007) [17] Amorphous Silicon: 2064 MJ/m2, Alsema (1998) [18] The energy demand for the structures has to be included as well [19]. Alsema et al. (2000) [20] quantified the GER for the structures in 500 MJ/m2. To be able to compare the energy delivered by the modules with the requirement of energy for their manufacture and installation it is necessary to express both quantities in the same form, as primary energy or as final energy. Both EPBT and ERF depend not only on the PV technology, on the irradiation in the considered location or on the performance ratio of the installation, but also on the efficiency of the systems of generation of electrical energy which it is replaced. Following the methodology used by the IEA in the Report IEA-PVPS-T10-01 (2006) [21] an average efficiency has been estimated in the mix of the systems of generation of in electrical energy of 31%. Besides, for the calculation of the ERF it is necessary to estimate the lifetime of the installation: a lifetime of 25 years has been considered for the modules. Nevertheless, reasonable efficiencies can be expected for more than 30 years, which would classified our values of ERF as conservative. The evolution of these parameters with the distance among rows is shown in Figs. 11(a), (b) and 12(a), (b) From this point of view, in the Torres Quevedo building the most suitable modules are those of CdTe (First solar) with energy payback time less than two years. Assuming a lifetime of 25 years, more than 15 times the energy used directly and indirectly in their manufacturing can be obtained. These modules have a low energy cost of manufacture and good efficiency. The other thin-film technologies present also good values of EPTB and ERF. Monocrystalline Silicon modules have the biggest production per area of module, as a consequence of its better efficiency, but they require a longer period of time to be amortized (energetically spoken) due to the important amount of energy used during their process of manufacture and assembly. The reduction in the distance among rows leads to higher shading losses and therefore a lower production of electricity and worst values of EPBT and ERF (nearly a 5% with a reduction of 15% in the distance; an increase of 25% in distance only modifies a 2% the values of these parameters). Nevertheless it is important to note 2008 A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 Fig. 13. (continued). A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 that a decrease of the distance implies a reduction of ERF but an increase in Energy Surplus and the amount of CO2 emissions avoided. 6. Analysis of the different parameters with inclination angles of 0 and 30 In this section a comparison between the results obtained with the modules in horizontal position and with an inclination of 30 and a distance among rows (following the criterion of absence of shades at 12:00 of the winter solstice) is presented. All the available area is used to install PV modules. In Fig. 13(a) it is shown that the installation of PV modules in a horizontal position results in a higher density of peak power. Specific dimensions (length and width) of both the flat roof and the modules affect the number of modules that fit and therefore the peak density of power. In particular, 134.26 Wp/m2 of the more efficient modules (Solarfun, m-Si) can be installed. At the other extreme are the modules of a-Si (Kaneka) with only 60.53 Wp/m2. With 30 tilt, the values are much lower (70.22 Wp/m2 for m-Si and 31.66 for a-Si, that is, only the 52% of the density of power in horizontal position) due to the required distance between rows. Thin-film technologies have better values of final system yield, as it is shown in Fig. 13(b). This is mainly due to their better behaviour with temperature (lower value of the coefficient of variation of power with temperature) and their superior spectral response. Although there are no shading losses in the horizontal arrangement, better values of Yf are obtained by placing the modules with 30 tilt due to increased irradiance (reduction of values of Yf range from 7.4 for a-Si and 8.8 for p-Si). However, the annual energy per unit area of roof area is higher for the more efficient modules (Fig. 13(c)). The highest value occurs for m-Si (162.31 kWh/m2). Polycrystalline p-Si (Photowatt), CIS (Wurth) and TeCd (First solar) modules present similar values (138.68, 133.40 and 133.53 kWh/m2, respectively), due to the combination of their values of efficiencies and yield factors. Again higher energy density values are obtained in the horizontal position (around a 75% higher than in horizontal position for all technologies). When the annual energy production is essential, it seems not advisable to use a-Si. The yearly energy per unit of collection area can be seen in Fig. 13(d). The values are slightly (8e9%) higher for 30 because of the higher collected irradiance. However this energy parameter has less interest (in the author’s opinion) than the density of energy obtained. From an economic perspective, the parameters IRR and NPV result of high interest. The profitability fluctuates depending on the conditions of the study: installation cost, financed quantity, bank interest rates, taxes, and tariffs for the kWh injected into the grid. These parameters are time and location dependent. Generalization is not possible. To calculate their values the current level of prices of the market for modules of crystalline silicon, CIS, CdTe and a-Si have been considered. A cost of assembly and Balance of System has been established for all the cases of 2500 V/kWp, so we obtain the following final prices: Solarfun: 5000 V/kWp Photowatt: 4800 V/kWp Wurth: 4600 V/kWp Firstsolar: 4000 V/kWp Kaneka: 4000 V/kWp For the economic analysis we have considered the operating and maintenance costs (including insurances) as 10% of the annual 2009 incomes, a 50% of financing from a banking institution at an interest rate of 5% (20 years), an inflation rate of 3.5% and taxes (25.00%). In Spain, the legal framework regarding the prices for kWh injected to the grid is given in the RD1578/2008 [22]. In this study a rate of 32 cents of euro per kWh has been used. The values of the parameters for the different modules are presented in Fig. 13(e) and (i). Because of their low initial cost and good values of final system yield, the best value of IRR is obtained for the aSi (Kaneka). CdTe modules (First solar) present good values of IRR as well. As expected, IRR values are higher with an angle of inclination of 30 , than with the horizontal option. It is due to the greater amount of electricity generated in every module with the same initial investment (Fig.13(b)). On the contrary, NPV values are higher for the modules placed horizontally, due to the higher installed capacity (Wp), and then, the increase in the annual produced energy and revenues by selling that electricity (Fig. 13(a) and (c)). By the same reasoning, the amorphous Si modules have a higher value of IRR, but the parameter NPV present lower values derived from to their lower efficiency and power density and energy. Less investment is needed but also the incomes are much lower. As for the values of ERF and EPBT, best values are obtained for the modules First Solar (CdTe) due to the combination of lower manufacturing costs and good energy density values (Fig. 13(f) and (g)). According to the simulations carried out in PVSYST 4.34, the energy used in its manufacture is recovered in less than 2 years. They are followed by Kaneka modules, a-Si. From the point of view of this energy parameter is more convenient to place the modules to maximize solar radiation. Reductions of 8e9% in EPBT are obtained in the 30 option, in comparison with the horizontal installation. For instance, the EPBT values for m-Si (Solarfun) modules are 3.02 years (horizontal) and 2.76 years (30 ). When the net energy generation over the lifetime and the avoided CO2 emissions to the atmosphere are considered, it is convenient to place the modules horizontally, because of their higher annual production (Fig. 13(j)). A conversion factor of 0.444 kg2CO/ kWh has been used [21]. The modules of Si-m horizontally placed are the best solution, despite their worst final system yield. In conclusion, due to the differences in energy required for manufacturing the modules of several technologies, efficiency, temperature, spectral performance and cost for each technology, the final decision could be different depending on the chosen parameter to be optimized. A PV installation with a tilt of the modules of 30 present better results for the parameters final system yield, energy payback time, energy return factor, IRR and yearly energy per unit of collection area. In contrast, the values of the parameters energy surplus (and CO2 avoided emissions), density of energy, density of peak power, and NPV, are much higher in the case of installing the modules horizontally. 7. Conclusions It is highly interesting to analyze the alternative options before carrying out PV installations on flat roofs. Parameters as density of peak power, final system yield, density of energy, yearly energy per unit of collection area, internal return factor, NPV, energy payback time, energy return factor, energy Surplus, CO2 avoided emissions could be analyzed. Depending on the chosen technology and its installation option, results will greatly differ. The careful analysis of the mentioned indicators is a key issue in order to perform a correct design of a PV system, especially where the room is a limiting factor. The owner, designer or promoter of the installation will give more importance to one parameter and the design of the system should be accordingly carried out. In general, the alternative of setting the modules at lower distance than the reference distance is not advisable for grid-connected plants. However, in those locations where room is an 2010 A.A. Bayod-Rújula et al. / Energy 36 (2011) 1996e2010 important boundary condition, this option can be interesting to study because of its higher production per square meter. It is also worth to highlight that the considered crystalline modules, in spite of presenting higher efficiency (and therefore a higher energy production) present worse values for EPBT and ERF. The reason is that they require more energy in their fabrication period than the thin-film technologies (a-si CIS and CdTe). In the latter, the lower net energy requirement leads to better values for the EPBT and ERF and parameters. Thin-film technologies are specially indicated for buildings without space restrictions, as it can be the case of an industrial warehouse. But the highest values of Energy surplus (and CO2 emissions avoided) are obtained with modules of higher efficiency, when all the available area can be used. In this paper, the study methodology has been shown, as well as its implementation in a case study, the central flat roof of the Torres Quevedo building of the University of Zaragoza (Spain). References [1] Prasad D, Show M. Designing with solar power, a source book for BIPV. Earthscan; 2005. Solarch Group, Centre for a Sustainable Built Environment, University of New South Wales, Australia. [2] Sick F, Erge T. Photovoltaics in buildings: a design handbook for architects and engineers. London UK: James & James Limited; 1996. [3] Kiss G, Kinkead J. Optimal building-integrated photovoltaic applications. Available at:. Colorado: National Renewable Energy Laboratory www.nrel. gov/docs/legosti/old/20339.pdf; November 1995. NREL/TP-472-20339. [4] Eiffert P, Kiss GJ. Building-integrated photovoltaics for commercial and institutional structures: a sourcebook for architects and engineers. Available at: www.nrel.gov/docs/fy00osti/25272.pdf; February 2000. NREL/BK-520e25272. [5] Yoo SH, Lee ET. Efficiency characteristic of building integrated photovoltaics as a shading device. Building and Environment 2002;37:615e23. [6] International Energy Agency. Potential for building integrated Photovoltaics. Available at: www.iea-pvps.org/products/download/rep7_04.pdf; 2002. Report PVPS T7-04. [7] Cheng CL, Sanchez Jimenez CS, Lee MC. Research of BIPV optimal tilted angle, use of latitude concept for south orientated plans. Renewable Energy 2009;34:1644e50. [8] Cheng CL, Chan CY, Chen CL. Empirical approach to BIPV evaluation of solar irradiation for building applications. Renewable Energy 2005;30:1055e74. [9] Mondol JD, Yohanis YG, Norton B. The impact of array inclination and orientation on the performance of a grid-connected photovoltaic system. Renewable Energy 2007;32:118e40. [10] Mondol JD, Yohanis YG, Smyth M, Norton B. Long-term validated simulation of a building integrated photovoltaic system. Solar Energy 2005;78:163e76. [11] Mondol JD, Yohanis YG, Norton B. Optimising the economic viability of gridconnected photovoltaic systems. Applied Energy 2009;86:985e99. [12] Oliver M, Jackson T. Energy and economic evaluation of building-integrated Photovoltaics. Energy 2001;26:431e9. [13] International Energy Agency. Guidelines for economic evaluation of building integrated PV. Available from: http://www.iea-pvps.org/products/download/ rep7_05.pdf; 2002. Report PVPS T7-05. [14] Hussein HMS, Ahmad GE, El-Ghetany HH. Performance evaluation of photovoltaic modules at different tilt angles and orientations. Energy Conversion and Management 2004;45:2441e52. [15] Mermoud A. PVSYST version 4.34. User’s manual. Available at: www.PVSyst.com. [16] Alsema EA, Wild-Scholten MJ. Environmental impacts of crystalline silicon photovoltaic module production. In: Proceedings of 13th CIRP Intern. Conf. on life cycle engineering. Leuven; 2006. [17] Raugei M, Bargigli S, Ulgiati S. Life cycle assessment and energy pay-back time of advanced photovoltaic modules: CdTe and CIS compared to poly-Si. Energy 2007;32(8):1310e8. [18] Alsema EA. Energy requirements of thin film solar cell modules. Renewable and Sustainable Energy Reviews 1998;2:387e415. [19] Crawford RH, Treloar GJ, Fuller RF, Bazilian M. Life-cycle energy analysis of building integrated photovoltaic systems (BiPVs) with heat recovery unit. Renewable and Sustainable Energy Reviews 2006;10:559e75. [20] Alsema EA, Nieuwlaar E. Energy viability of photovoltaic systems. Energy Policy 2000;28:999e1010. [21] IEA-PVPS T10-01. Compared assessment of selected environmental indicators of photovoltaic electricity in OECD cities. Available at: www.eupvplatform. org/fileadmin/Documents/IEA-PVPS_Report_T10-01-2006.pdf; 2006. [22] Ministerio de Industria, Turismo y Comercio, Spain, Royal Decree 1578/2008, de retribución de la actividad de producción de energía eléctrica mediante tecnología solar fotovoltaica para instalaciones posteriores a la fecha límite de mantenimiento de la retribución del Real Decreto 661/2007, de 25 de mayo, para dicha tecnología.