Solar Energy 77 (2004) 171–178 www.elsevier.com/locate/solener
Technical and economical system comparison of photovoltaic and concentrating solar thermal power systems depending on annual global irradiation Volker Quaschning
*
FHTW Berlin, Berlin University of Applied Sciences, Marktstr. 9, D-10317 Berlin, Germany Received 30 July 2003; received in revised form 30 March 2004; accepted 22 April 2004 Available online 31 May 2004 Communicated by: Associate Editor David Mills
Abstract Concentrating solar thermal power and photovoltaics are two major technologies for converting sunlight to electricity. Variations of the annual solar irradiation depending on the site influence their annual efficiency, specific output and electricity generation cost. Detailed technical and economical analyses performed with computer simulations point out differences of solar thermal parabolic trough power plants, non-tracked and two-axis-tracked PV systems. Therefore, 61 sites in Europe and North Africa covering a global annual irradiation range from 923 to 2438 kW h/m2 a have been examined. Simulation results are usable irradiation by the systems, specific annual system output and levelled electricity cost. Cost assumptions are made for today’s cost and expected cost in 10 years considering different progress ratios. This will lead to a cost reduction by 50% for PV systems and by 40% for solar thermal power plants. The simulation results show where are optimal regions for installing solar thermal trough and tracked PV systems in comparison to non-tracked PV. For low irradiation values the annual output of solar thermal systems is much lower than of PV systems. On the other hand, for high irradiations solar thermal systems provide the best-cost solution even when considering higher cost reduction factors for PV in the next decade. Electricity generation cost much below 10 Eurocents per kW h for solar thermal systems and about 12 Eurocents/kW h for PV can be expected in 10 years in North Africa. 2004 Published by Elsevier Ltd. Keywords: Concentrating solar power; Photovoltaics; Global irradiation; Electricity generation cost; Computer simulation; Economics; Photovoltaic systems; Solar thermal power systems
1. Introduction Concentrating solar thermal systems such as parabolic trough power plants or solar tower power plants and photovoltaic systems can convert sunlight to electricity. Recommended sites for the installation of
*
Corresponding author. Tel.: +49-30-55134-256; fax: +4930-55134-199. E-mail address:
[email protected] (V. Quaschning). URL: http://www.volker-quaschning.de. 0038-092X/$ - see front matter 2004 Published by Elsevier Ltd. doi:10.1016/j.solener.2004.04.011
concentrating solar systems are situated in the sunbelt of the earth, where the direct irradiation is high. On the other hand, many calculations have shown that concentrating solar thermal systems are more cost-effective than photovoltaic systems not only in the sunbelt (Trieb et al., 1997). However, until the year 2003 commercial solar thermal systems have been only built in the Mojave desert in California where the annual global solar irradiation is higher than 2100 kW h/m2 . It is not easy to estimate the minimum irradiation for concentrating solar thermal systems to be operated more economically than PV. Furthermore, conditions will change in the next decade. This paper shall deal with these open
172
V. Quaschning / Solar Energy 77 (2004) 171–178
Nomenclature G Gb Gd Gd30 Gdn Gr30 Grn H1ax
global irradiance in W m2 beam irradiance in W m2 diffuse irradiance in W m2 diffuse irradiance on a 30 sloped Southoriented surface in W m2 diffuse irradiance on normal plane facing to the sun in W m2 ground reflected irradiance on a 30 sloped South-oriented plane in W m2 ground reflected irradiance on a normal plane facing to the sun in W m2 direct irradiation on a one-axis tracked plane in W h m2
questions carrying out detailed technical and economical analyses of these two technologies. Therefore, the greenius simulation software has been used. This software was developed especially for carrying out such investigations and offers the possibility to compare different types of solar electricity generation systems at different sites. Two types of photovoltaic and solar thermal parabolic trough systems have been simulated at 61 sites. Altogether 183 simulation runs assuming today’s cost and expected cost in 10 years provide the basis for this paper. 1.1. Examined photovoltaics systems Since cost estimations for concentrating PV installations have a high uncertainty, only non-concentrating PV systems were considered. For the simulations in this paper two types of photovoltaic systems have been chosen: • non-tracking system with 30 slope angle, • two-axis tracking system. For the following calculations, the nominal efficiencies of both systems are the same, whereas the annual efficiencies are site dependent. During the simulations the hourly output of a reference PV system was calculated to obtain the annual efficiency. Part-load behaviour of PV modules and inverters were calculated with empirical models. The annual self-consumption of the two-axis tracking system was estimated with 10 kW h/m2 . 1.2. Examined solar thermal power plant 354 MW of solar electric generating systems (SEGS) parabolic trough power plants have been connected to the grid in Southern California since the mid-1980s.
H30 Hb Hgn Greeks a b c cax h
irradiation on a 30 sloped South-oriented plane in W h m2 beam irradiation on a normal plane facing to the sun in W h m2 global normal irradiation on a normal plane facing to the sun in W h m2 sun altitude tilt angle sun azimuth azimuth angle of tracking axis angle of incidence
Therefore, they represent the most mature solar thermal power technology (Quaschning and Blanco, 2001). To date, there are more than 100 plant-years of experience from the nine operating plants, which range in size from 14 to 80 MW. Most of the currently planned solar thermal power projects are also parabolic trough power plants. For this reason, this paper considers state of the art parabolic trough power plants as shown in Fig. 1. The output for these power plants was simulated using hourly meteorological data. The optical and thermal collector efficiency, the field losses and the power block efficiency were calculated for the whole year to estimate the annual system efficiency as described by Quaschning et al. (2002).
1.3. Use of simulation software Only a powerful simulation tool can provide the results of the system analyses presented in this paper. Many computer programs for simulating photovoltaic systems are available, but only a few software codes exist to simulate concentrating solar thermal power systems, e.g. the TRNSYS STEC library (Jones et al., 2001). However, since no available simulation software can perform detailed technical and economical calculations of both PV and solar thermal systems DLR has developed the new simulation environment greenius for the technical and economical analysis of renewable power projects such as photovoltaic, solar thermal or wind power plants. The inputs for the simulation are hourly solar irradiance, ambient temperature and other meteorological data. With this meteo data, site information, technical system data and economical specifications, this simulation environment calculates the system output, system efficiencies and other technical parameters and provides various economical key values (Quaschning et al., 2001).
V. Quaschning / Solar Energy 77 (2004) 171–178
173
Fig. 1. Scheme of a parabolic trough power plant.
The greenius software has already been used for the calculations at other studies (Quaschning et al., 2002). Thus, the greenius simulation environment was also used to perform the technical and economical calculations presented in this paper.
2. Site, system and cost specifications
Hb ¼
8760 X
ð1Þ
The chosen one-axis tracking concentrating solar thermal power system can use the direct irradiation H1ax on a single-axis tracking device (trough collector). Therefore, the beam irradiance Gb must be projected into the collector area: 8760 X H1ax ¼ Gb cos h 1 h: ð2Þ t¼1
2.1. Solar irradiance The further simulations have been carried out for 61 sites in Europe and North Africa. These sites cover a global annual irradiation range from 923 kW h/m2 a in Dublin (Ireland) until 2438 kW h/m2 a in Luxor (Egypt). The databases Satellight (Satel-light, 1998) and Meteonorm (Meteotest, 2003) provided mean hourly global G, diffuse Gd and direct normal or beam Gb irradiance values for the simulations. The free Satellight database contains irradiance data for whole Europe and Africa north of the 34th latitude using half-hourly images of weather satellites. The Meteonorm database was used for including irradiance data also of other North Africa sites, for instance Egypt. Meteonorm contains an extensive database of monthly irradiation values. It applies statistical models for estimating the hourly irradiance. Both databases show differences in the annual irradiation of more than 10% for single sites. Hence, there is always a relative high uncertainty for solar irradiation even when using sophisticated meteorological databases. A two-axis tracking concentrating solar system can use the annual beam irradiation Hb that is the sum of the hourly beam irradiance Gb :
Gb 1 h:
t¼1
Hereby the angle of incidence h (Stine and Harrigan, 1985) is calculated in dependence on sun elevation a, sun azimuth c, tilt angle b the azimuth angle cax of the tracking axis at the middle of each time interval h¼ arccos
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 1 ðcosða bÞ cos b cos a ð1 cosðc cax ÞÞÞ :
ð3Þ Methods to calculate the angles a and c for position of the sun are given by (Quaschning, 2003) or (Duffie and Beckman, 1991). The tilt angle b and azimuth angle cax of a horizontal trough collector oriented in N–S direction are zero. The fixed PV system with a tilt angle of 30 can use the global irradiation H30 on a fixed, 30 sloped surface oriented to the South: H30 ¼
8760 X ðGb cos h þ Gd30 þ Gr30 Þ 1 h:
ð4Þ
t¼1
Therefore, the hourly beam irradiance Gb the angle of incidence h on the sloped surface, the diffuse irradiance
174
V. Quaschning / Solar Energy 77 (2004) 171–178
2.2. System efficiencies
Gd30 on a 30 sloped surface and the ground reflection Gr30 must be calculated for every hour of the year. Methods to calculate the angle of incidence h and the diffuse and ground-reflected irradiation on a tilted plane are given by (Quaschning, 2003) or (Duffie and Beckman, 1991). Hourly diffuse irradiance and ground reflection are calculated using the hourly global irradiance that is provided by the above-described databases. The annual diffuse irradiation depends significantly on the diffuse irradiation model. Here, the model of Perez et al. (1990) was used. Finally, the two-axis tracking PV system can use the global normal irradiation Hgn This is calculated similar to the 30 sloped surface. However, the surface orientation must be changed hourly so that it is perpendicular to the direction of the solar irradiance 8760 X ðGb þ Gdn þ Grn Þ 1 h:
ð5Þ
t¼1
Fig. 2 shows the different irradiations depending on the global irradiation for the chosen 61 sites. For global horizontal irradiations below 2300 kW h/m2 a the direct irradiation on the one-axis-tracking system is lower than the global irradiation on a fixed system. This indicates clearly that the use of tracked and concentrated solar systems is difficult in Middle and North European regions. The incident irradiation on 2-axis-tracked systems is the highest for all sites. The absolute difference between the annual irradiations for all tracking variants is nearly the same over the full irradiation range. Hence, the advantage of two-axis-tracking systems compared to one-axis-tracking systems decreases with increasing annual irradiations. On the other hand, tracking systems have a much higher output than non-tracking systems especially at regions with high irradiation values.
Luxor
Cairo
Oujda
El Aricha
Palermo
Milan
Porto
Bern
Zurich
Toulouse
3500
Frankfurt
Dublin
kWh/m²a
Hamburg
4000
Almería
Hgn ¼
The used simulation software has calculated the hourly output of the investigated solar systems. The output of the PV systems depends on the site, position of sun, irradiance and ambient temperature. Humidity and air pressure influence additionally the output of the solar thermal system. Parameters of state of the art systems have been chosen for the simulation. The annual system efficiency of the monocrystalline silicon PV system of about 11% is almost constant with the site irradiation decreasing a little at higher irradiation values due to the negative influence of correlated higher ambient temperature (Fig. 3). This result is not astonishing, because PV module efficiency is almost constant over large irradiance ranges and decreases with higher temperatures. In contrast to PV systems, the annual system efficiency of parabolic trough systems increases significantly with the annual irradiation. The part load efficiency of a steam turbine cycle is much lower than the nominal efficiency. The efficiency is also reduced during days with fluctuating irradiance values due to the capacitive behaviour of the thermal system. Furthermore, sites with lower irradiation values are situated at higher latitudes where the solar altitude is also lower. This increases the optical losses of one-axis tracked trough collectors and therefore reduces the efficiency. Lower ambient temperatures at sites with lower irradiations increase the thermal losses of the collector and field and thus reduce the solar thermal collector efficiency. All these effects were taken into account for the annual performance analysis. The results show that the annual system efficiency of a today’s solar thermal trough power plant varies between 10% and 14% for the considered irradiation range.
3000
2500
2000
1500 global normal irradiation, 2 axis tracking system direct normal irradiation, 2 axis tracking system direct irradiation on 1 axis tracking system global irradiation on 30° tilt, non-tracking system
1000
500 900
1100
1300
1500
1700
1900
2100
2300
2500
global horizontal irradiation in kWh/(m² a)
Fig. 2. Direct normal irradiation, direct irradiation on a 1-axis tracked collector, global irradiation on a 30 tilted surface and global irradiation on a 2-axis tracked surface as function of the global horizontal irradiation.
V. Quaschning / Solar Energy 77 (2004) 171–178
175
14 %
annual system efficiency
12 % 10 % 8% 6% standard non-concentrated photovoltaic system 4%
standard parabolic trough system
2% 0% 900
1100
1300
1500
1700
1900
2100
2300
2500
global horizontal irradiation in kWh/(m² a)
Fig. 3. Annual system efficiency of photovoltaic systems and 1-axis tracked concentrating parabolic trough systems as function of the global horizontal irradiation.
2.3. Cost assumptions For comparability reasons all costs are related to square meters of effective system area. Assuming a PV module efficiency of 13.5% one square meter can hold PV panels with a capacity of 135Wp . Finally, overall system cost of 4500 a/kWp result in area related cost of 610 a/m2 . Operation results of existing PV systems have provided net present values of the cost for operation and maintenance of about 170 a/m2 . Installation and operation cost of tracked PV systems are higher than the cost of non-tracking systems. The cost assumptions of about 3000 a/kW for the parabolic trough power plant are valid for a system with a capacity of 30 MW without thermal storage. These costs are much lower than the cost of PV systems (see Table 1) but still in the same magnitude as in the 1990s, since the installation rates for solar thermal power plants are not very high today. For all systems a lifetime of 30 years and an overall discount rate of 7% were assumed. Looking at the PV learning curve, there is a cost reduction by 20% when doubling the market volume (Woditsch, 2000 or IEA, 2000). In the past, this doubling was achieved almost every 4 years. Assuming the same growth rates for the next decade, there will be a cost reduction by 50%. For solar thermal parabolic trough power plants the progress ratio is about 0.88 (Enermodal, 1999). In other words, a price reduction by 12% can be expected when doubling the market volume. On the other hand, possible growth rates of solar thermal power are higher. Their power plant installation numbers start from a lower annual production rate and they have not the same production limits as PV. Combining lower price
Table 1 Assumptions for today’s system costs Net present value in a/m2 for system
Installation
Operation
Total
Non-tracking PV system 2-axis-tracked PV system Parabolic trough power plant
610
170
780
800
230
1030
450
180
630
Table 2 Assumptions for system costs in 10 years Net present value in a/m2 for system
Installation
Operation
Total
Non-tracking PV system 2-axis-tracked PV system Parabolic trough power plant
305
85
390
400
115
515
270
108
378
reduction and higher growth rates leads to an overall cost reduction for parabolic trough power plants of about 40% within the next 10 years. Table 2 summarizes the assumptions for solar thermal power and PV. 3. Simulation results Fig. 4 shows the specific annual system output of the non-tracking and two-axis-tracking PV system as well of the solar thermal power plant. Up to annual global
176
V. Quaschning / Solar Energy 77 (2004) 171–178
irradiation values of about 1700 kW h/m2 a the output of the solar thermal system is the lowest because efficiency and usable irradiation are disproportional low. Since the system efficiency of solar thermal power plants in regions with very high irradiations is higher than the efficiency of PV systems, the specific annual output of the solar thermal system becomes nearly the same as of the two-axis-tracked PV system there. Fig. 5 shows the resulting levelled electricity cost combining the specific output of Fig. 4 and the specific
400
cost assumptions of Table 1. Since today’s cost of solar thermal power plants are lower than of PV systems, levelled electricity cost are also lower above global irradiations of 1300 kW h/m2 a although the specific output of the PV system is higher until 1700 kW h/m2 a. Nevertheless, there is a high uncertainty in the simulation results of solar thermal power plants at very low irradiations. Due to high investment cost for multimegawatt solar thermal power plants, sites with higher annual irradiations are recommended.
kWh/m²
350
specific annual output
300 250 200 150 100 2-axis tracking photovoltaic system parabolic trough system
50
non-tracked 30° tilted photovoltaic system 0 900
1100
1300
1500
1700
1900
2100
2300
2500
global horizontal irradiation in kWh/(m² a)
Fig. 4. Specific annual output of fixed and tracked photovoltaic systems and 1-axis tracked concentrated parabolic trough systems as function of the global horizontal irradiation.
1
c/kWh 2-axis tracking photovoltaic system
today’s levelled electricity generation cost
0.9
parabolic trough system 0.8
non-tracked 30° tilted photovoltaic system
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 900
1100
1300
1500
1700
1900
2100
2300
2500
global horizontal irradiation in kWh/(m² a)
Fig. 5. Today’s Levelled electricity generation cost for photovoltaic systems and 1-axis tracked concentrated parabolic trough systems as function of the global horizontal irradiation.
V. Quaschning / Solar Energy 77 (2004) 171–178
levelled electricity generation cost in 10 years
1
177
c/kWh
0.9
2-axis tracking photovoltaic system parabolic trough system
0.8
non-tracked 30° tilted photovoltaic system
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 900
1100
1300
1500
1700
1900
2100
2300
2500
global horizontal irradiation in kWh/(m² a)
Fig. 6. Levelled electricity generation cost in 10 years for photovoltaic systems and 1-axis tracked concentrated parabolic trough systems as function of the global horizontal irradiation.
In 10 years, the break-even irradiation for the generation cost of non-concentrating and solar thermal systems moves to higher irradiation values as shown in Fig. 6. In South Europe both technologies can operate with cost below 20 Eurocents/kW h. Solar thermal power plants remain the best-cost solution in South Europe and North Africa with possible generation cost below 10 Eurocents/kW h. Tracked PV systems have some cost advantages in North Africa.
4. Conclusions Simulation runs of non-tracking PV systems, twoaxis-tracking PV systems and solar thermal trough power plants have provided detailed results of technical and economical parameters for 61 sites in Europe and North Africa. These results demonstrate the high performance of simulation tools and indicate clearly the optimal operating ranges for solar thermal power and PV systems. Today, solar thermal power plants are more economically at sites with annual global irradiations of more than 1300 kW h/m2 . Below this value levelled electricity cost of PV systems are lower. The break-even irradiation will move to about 1600 kW h/m2 in 10 years due to higher achievable cost reductions for PV systems. Tracked PV systems show some slight cost advantages in regions with high global irradiations above 1800 kW h/m2 . The cost estimations have shown that levelled electricity cost can come down to a reasonable range significantly below 10 Eurocents/kW h for solar thermal power at about 12 Eurocents/kW h for PV systems in North Africa within the next 10 years. Thus, solar
electricity will get a more important role in the struggle against global warming.
References Duffie, J.A., Beckman, W.A., 1991. Solar Engineering of Thermal Processes, 2nd ed. Wiley Interscience, New York. Enermodal Engineering Ltd., 1999. Final Report Prepared for the World Bank: Cost Reduction Study for Solar Thermal Power Plants. World Bank, Washington, DC. International Energy Agency (IEA), 2000. Experience Curves for Energy Technology Policy. OECD/IEA Paris. Jones, S.A., Pitz-Paal, R., Schwarzboezl, P., Blair, N., Cable, R., 2001. TRNSYS modeling of the SEGS VI parabolic trough solar electric generation system. In: Proceedings of Solar Forum 2001, Washington, DC, April 21–25, 2001. Meteotest, 2003. Meteonorm Global Meteorological Database, Version 5 Edition 2003, Meteotest, Bern. Perez, R., Ineichen, P., Seals, R., Michalsky, J., Stewart, R., 1990. Modeling daylight availability and irradiance components from direct and global irradiance. Solar Energy 44, 271–289. Quaschning, V., 2003. Regenerative Energiesysteme, 3rd ed. Hanser, Munich. Quaschning, V., Kistner, R., Ortmanns, W., 2002. Influence of direct normal irradiance variation on the optimal parabolic trough field size: A problem solved with technical and economical simulations. ASME Journal of Solar Energy Engineering 124 (May), 160–164. Quaschning, V., Ortmanns, W., Kistner, R., Geyer, M., 2001. Greenius––A new Simulation Environment for Technical and Economical Analysis of Renewable Independent Power Projects. In: Proceedings of ASME International Solar Energy Conference Solar Forum 2001, Washington DC, 22– 25 April, pp. 413–417.
178
V. Quaschning / Solar Energy 77 (2004) 171–178
Quaschning, V., Blanco, M., 2001 Solar power-photovoltaics or solar thermal power? In: Proceedings of VGB Congress Power Plants 2001. Brussels, 10–12 October 2001. Satel-light, 1998. European Database of Daylight and Solar Radiation. Available from
. Stine, W.B., Harrigan, R.W., 1985. Solar Energy Fundamentals and Design. John Wiley & Sons, New York.
Trieb, F., Langniß, O., Klaiß, H., 1997. Solar electricity generation––A comparative view of technologies, costs and environmental impact. Solar Energy 59 (1–3), 89–99. Woditsch, P., 2000. Kostenreduktionspotenziale bei der Herstellung von PV-Modulen. In: Proceedings of FVS Themen 2000, ForschungsVerbund Sonnenenergie, Berlin, pp. 72– 86.