Applied Thermal Engineering 31 (2011) 3588e3594

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Simulation analysis of thermal storage for concentrating solar power

Roman Adinberg*
Weizmann Institute of Science, Solar Energy Research Facilities, Rehovot, Israel

a r t i c l e i n f o
Article history: Received 19 January 2011 Accepted 16 July 2011 Available online 22 July 2011 Keywords: Solar power plants Hybrid systems Thermal storage Solar fraction System modeling

a b s t r a c t
The aim of this study was to evaluate the capacity and analyze the performance of thermal storage required for solar thermal electric power plants in order to increase their capability to supply base load power with less need for back up from fossil fuels. For this purpose, a mathematical-statistical model of hybrid solar-fossil power cycles was developed, which is based on energy balance equations and historical hourly data of direct normal irradiance and load proles available in the literature. As follows from the computations performed for base load operations, an extremely large storage capacity equivalent to near a thousand full load operating hours should be available to a power plant to achieve continuous electricity production entirely on solar energy (solar fraction equal 1.0) during an annual operating cycle. For state of the art thermal storage technologies having a potential capacity of 10e14 full load operating hours for large-scale parabolic through solar power plants, the assessed solar fraction was 0.4e0.5 respectively, with relation to the specic conditions of calculations. The performance characteristics of thermal storage presented in the paper cover the whole extent of solar fractions from 0.2 (no storage applied) to 1.0 (pure solar operation of a power plant). 2011 Elsevier Ltd. All rights reserved.

1. Introduction In the last two decades, the concept of Concentrating Solar Power (CSP) has successfully demonstrated its capability of producing high-temperature steam to power the conventional Rankine cycle for electricity generation. Today, the CSP technology is under a wide deployment of large-scale solar power plants for 50 MW and more electrical power capacity in various sun rich regions around the globe [1e4]. The greatest technical challenge of producing electricity from the sun is the high intermittency of solar power supply that makes it incompatible with common types of electrical load proles, such as domestic, commercial or industrial [5,6]. In order to stabilize power delivery and prolong daily operating hours, solar thermal power plants have the options of using either or both solar thermal storage and fossil fuel combustion. Depending on the installed backup power capacity, the solar plant can be run continuously at full load during the day and several hours in the nighttime [7e9]. The share of solar energy in the annual electricity production capacity of hybrid solar-fossil power plants is called the solar fraction or annual solar capacity factor. In fact, it is a primary indicator of the sustainability of solar thermal electricity generation technology. The capability of hybrid solar-fossil power systems
* Tel.: 972 8 934 3779; fax: 972 8 934 4117. E-mail address: roman.adinberg@weizmann.ac.il. 1359-4311/$ e see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2011.07.025

without solar energy storage to match the typical grid load demands is limited to a solar fraction a 0.13e0.25 [4,6,8]. Consequently, a major part of the thermal power consumption, 75% or more, must come from conventional fossil energy sources, such as coal, natural gas, etc. In terms of thermal storage capacity, it is customary in addition to units of energy to use units of time, e.g. hours, as an operating duration of a power plant when the full load demand is provided solely from the energy storage. Simulations performed for a parabolic trough power plant including a thermal storage capacity for the typically considered 6 full load operating hours yielded a 0.4 versus 0.25 without storage [10]. To reduce signicantly the fossil fuel dependency of hybrid power plants by making the most of solar energy, CSP systems should have the ability to accumulate a large amount of solar energy during sunlight hours in order to retrieve the storage on a seasonal basis. For the purpose of storage, the CSP system must be increased so that part of the available solar power can be used to charge the thermal storage simultaneously with operating the power block, whenever the solar ux is sufciently high. In principle, adequately sized seasonal thermal storage should permit uninterrupted electric power generation during on- and off-sun hours, 24 h a day, all year round, to a so-called pure solar power plant, for which the parameter a 1.0. Although the subject of thermal storage has received considerable attention in the literature during the past few decades, basic

R. Adinberg / Applied Thermal Engineering 31 (2011) 3588e3594

3589

Nomenclature1 CSP E GS LV MR Q t yr concentrating solar power storage energy [kWh/m2; also, h, as full load operating hours or day 24 h] Granada, Spain Las Vegas, USA Mitzpe Ramon, Israel thermal power [W/m2] time [h] annual operating cycle [8760 h]
Rejected Power Qr j

QR

Solar Power Q1

QF

Fossil Fuel Backup Electric Power

CSP System Q0 Q2

Power Block QL

Q3 Energy Storage ES

Power Control System

Fig. 1. The schematic model of a solar thermal electric power system.

Greek letters solar fraction solar collector factor nominal storage capacity [kWh/m2; also, h, as full load operating hours or day 24 h] q annual average daily discharge duration of thermal storage [h] c coefcient of variation

a b 4

2. Energy balance modeling The studied model of solar thermal electric power plants is shown in Fig. 1. It consists of three major components: a CSP system (considering in particular a parabolic trough solar collector), a thermal energy storage facility (in general, a well-insulated tank containing the thermal storage material), and a power block (e.g. the Rankine cycle including fossil fuel backup). The core of the model is a power control system, as depicted in Table 1 and explained below. Primarily, the CSP system converts the incident solar power QR (direct normal irradiance) into the thermal power Q0 of the high-temperature heat transfer uid (HTF: e.g. 300e400  C, thermal oil). Then, the available power is distributed by means of the HTF ow between the power block, Q1, and the thermal storage, Q2. Simultaneously, some amount of power, Q3, can be extracted from the storage and directed to the power block. In addition, fossil-fueled power generation, QF, might be required in order to operate the power block at a specied load level, QL. The control concept of the power ow through a solar power plant is based on the following energy balance being evaluated at every instant of operating time:

Subscripts F fossil fuel power in the hybrid mode L thermal power equivalent to electric load R direct normal irradiance rj rejected solar power S storage; bend point in Figs. 4 and 5 0, 1, 2, 3 thermal power variables shown in Fig. 1

researches and developments concerning CSP applications have been mainly focused on short-term storage systems capable to provide a few full load operating hours [1,7]. Until now, there are only a few credible studies concerning the impact of large-scale storage systems on the operating efciency of solar power plants. As follows from simulations carried out in [11] for solar thermal electric power plants, solar fractions as large as a 0.75e0.9 are achievable with modest storage capacities ranging from 4 10 to 50 full load operating hours respectively, for the site of Albuquerque, USA. According to the estimates of storage capacity presented in [12], a solar fraction value resulting from 4 12 h is a 0.53, and due to [13], a larger storage capacity 4 15 h leads to a 0.65e0.71. As a whole, the capability of thermal storage to replace fossil fuels in backing solar operating cycles has not yet been thoroughly explored and deserves to be studied in depth. The objective of this work was to analyze the general case of hybrid solar thermal power systems combining thermal storage and fossil fuel backup facilities to produce electric power required by grid and to explore the ways of reducing the fossil fuel consumption by increasing the capacity and operating efciency of energy storage. For this purpose, an energy balance model of yearlong solar operating cycles comprising historical hourly data of direct normal irradiance and load demand was developed and applied to base load power plants located at different geographic sites. The computations performed allowed quantifying the amount of stored energy as function of the solar fraction in the whole range of storage capacities, from a few to more than a thousand full load operating hours.

Q1 Q2 Q0 Q1 Q3 QF QL

(1) (2)

The basic simplication of the considered approach is that the storage heat losses and auxiliary energy consumptions were not taken into account for these factors are rather related to the system design and in any case should constitute a relatively small fraction of the energy balance.2 Then, the contribution of solar energy in the continuous generation of power over an annual operating cycle is expressed with the aid of the solar fraction a in the following integral form:

Zyr
0

Q0 tdt a

Zyr QL tdt;
0

0<a

(3)

The amount of solar energy directly delivered to the power block is part of the total amount of energy provided by the CSP system:

Zyr
0

Q0 tdt b

Zyr Q1 tdt
0

(4)

Energy and power variables (e.g. E, Q and 4) pertain to one square meter of the solar collector aperture area.
1

2 A few sophisticated software packages are available for a comprehensive techno economic evaluation of particular thermal storage designs integrated into CSP systems [14,15].

3590

R. Adinberg / Applied Thermal Engineering 31 (2011) 3588e3594

Table 1 Logic operations of the power control system being performed at regular time intervals. CSP Output Direct Power Input Storage Charge (Solar Surplus) Storage Discharge Fossil Fuel Input Power Load Q0 Q1 Q2 Q3 QF QL !QL QL Q0 e Q1 0 0 <QL Q0 0 Q L e Q0 QL e Q1 e Q3 Q1 Q3 QF

The parameter b can be considered as some analogue of the solar multiple [16], a factor by which the solar eld size must be multiplied in order to supply energy for storage in addition to the amount of energy going directly to the power block at nominal conditions. The variable energy content of the thermal storage has the following integral form expression:

Fig. 3. Thermal energy variations ES(t) over a year at the MR site for two beginning dates of the pure solar power cycle, a 1.0: 1 - March 26, 2005, 2 - August 1, 2005 (lled area).

Zt ES t
0

Q2 s Q3 sds;

yr

(5)

Correspondingly, the direct solar power Q1 is determined as one of the two choices below:

Q 1 t Q L t or Q 1 t Q 0 t

(10)

The thermal storage is supposed to be empty at the beginning and fully utilized at the end of the annual operating cycle, so that

ES 0 ES yr 0

(6)

The highest quantity of thermal energy measured during the year represents a nominal storage capacity, denoted here (in the energy units, kWh/m2) as

4 Maximum ES t
Or, if expressed in full load operating hours (h), it is

Following the control strategy of Table 1, the solar power surplus, Q2 > 0, when available (8), is transferred to charge the storage unit, and under the opposite condition (9), when solar power is relatively low or entirely unavailable (at night time), the storage is discharged, Q3 > 0, to secure the basic assumption (2). While the current storage energy ES (t) is insufcient to meet load demand, the system makes use of the supplement (fossil fuel) source of power, QF > 0. 3. Basic results and discussion

4 Maximum ES t*yr=

Zyr Q L tdt
0

(7)

Numerical computations based on the above algorithm were carried out with the aid of the Mathematica program (Version 7.0, Wolfram Research). The input data included hourly values of direct

Table 1 is used to resolve the power variables Q1, Q2, Q3 and QF by estimating the conditional relationship between the solar power Q0 and load QL at every moment of time, so that either

Q 0 t ! Q L t or Q 0 t < Q L t

(8) (9)

Fig. 2. Depiction of two consecutive days (the LV site) with quite different solar conditions (1-h resolution data), where 1 - direct normal irradiance QR and 2 e base load level QL (for a 1.0).

Fig. 4. Nominal storage capacity 4 as a function of solar fraction a for hybrid power parabolic trough plants. The same line notations as in this graph are used in the next gures, where results for the different sites, GS, LV and MR, are compared.

R. Adinberg / Applied Thermal Engineering 31 (2011) 3588e3594

3591

Fig. 6. Typical diurnal-scale thermal energy variations ES(t) as obtained for a storage capacity 4 10 h (the GS power plant site). Fig. 5. The ratio of mean daily discharge time q [h] to storage capacity 4 [h] versus solar fraction a for the three power plant sites studied.

normal irradiance (DNI) for the years 2004e2006 related to different geographical locations, namely Granada, Spain [17], Mitzpe Ramon, Israel [17] and Las Vegas, Nevada, USA [18]. The selected sites presented total annual DNI values ranged between 2.0 and 2.5 MWh/m2. The conversion from a DNI dataset (denoted here as QR) to the CSP system output Q0 was calculated using the efciency factor for parabolic trough solar collectors operated on thermal oil as the heat transfer uid [10,14,16]. The electric load was set constant 24 h a day considering the case of base load power plants, though some other load proles were estimated as well. All the calculated power and energy variables pertain to one square meter of the solar collector aperture area and have a 1-h time resolution. Fig. 2 demonstrates a characteristic sequence of two operating days for a virtual base load solar power plant (the LV site): one day with an extremely high (left) and the other day with a very low (cloudy skies) direct normal irradiance QR. It is seen that a lot of solar energy as compared to the load (QR QL) can be saved during daylight hours of the rst day for use in the night, and also at a later date, provided the storage capacity is large enough. In contrast with this condition, the next operating day requires extensive support from either energy storage or fossil fuel backup almost all hours of the day. 3.1. Pure solar power cycle The variation of storage thermal energy ES (t) in a base load power system working solely on solar energy (a 1.0) during a year is shown in Fig. 3. With the beginning of the operating cycle 1, the storage energy grows strongly with time from zero to a top value of about 190 kWh/m2 attained after nearly half a year of the run. This result leads to an extremely large nominal storage capacity 4 53 full load operating days (24 h a day) and a solar collector factor b 4.8. The power operating cycle 1 was set off on 26th March in order to fulll initial conditions (6). With a different start date, the charging of the storage could not be completed, as is demonstrated with the cycle 2 initiated on 1st August. It is seen by comparing cycles 1 and 2 that the last one is insufcient until the end of March, and only afterwards the storage is getting charged properly. For the other two plant sites (GS and LV), the pure solar storage cycles were determined to commence in early April. In general, the start date may vary depending on the plant location and the specic solar power data set QR.

Quite similar pure solar ES(t)-proles have been obtained for the three plant sites simulated, so that each of them has a value of 4 ranged between 53 and 62 full load operating days. Such a large storage capacity is more than two orders of magnitude above the capability of available thermal storage technologies3 and therefore is entirely hypothetical. The signicance of the pure solar cycle is that it allows estimating the upper bound of storage capacity in the range of solar-fossil hybrid power cycles to be considered below. 3.2. Hybrid power cycle Fig. 4 presents the dependence of storage capacity 4 on solar fraction a varied from near 0.2 (no storage applied, 4 0) to 1.0 (the pure solar power cycle). The graph explicitly shows that the results of simulations for the different plant locations are rather similar and the function 4(a) has a marked change at the point S specied by the coordinates aS 0.5 and 4S 14 h. To the left of the point S the rate of increasing nominal storage capacity with solar fraction is comparatively low, in regard to the region of a > aS, where 4(a) grows rapidly to very large storage capacity values 4 w103 h. Solar fraction values a < aS are relevant to the state of the art thermal storage systems utilizing either sensible or latent heat of molten salts or various other storage materials [7,19e22]. The relatively low energy density of available thermal storage materials limits designing of storage systems for large-scale power plants (50 MW electrical power capacity and more) to below 4 10e14 full load operating hours due to an enormous amount of storage materials needed to ll the storage tanks. Another important attribute of thermal storage is an annual average number of full load operating (discharging) hours during a day that in the case of a base load power plant (QL const) can be written as

q 24

Zyr Q 3 tdt=Q L *yr

0

(11)

The discharge parameter q is sometimes referred to as equivalent to the thermal storage capacity 4 expressed in full load operating hours. However, such an interpretation might be misleading due to the different denitions given to these parameters, as can be seen from a comparison of Eqs. (7) and (11).

3 Particularly, the 2-tank molten salt system containing 28500 tons of the storage material provides 4 7.5 h for the 50-MW Andasol-1 parabolic trough power plant in Spain [2].

3592

R. Adinberg / Applied Thermal Engineering 31 (2011) 3588e3594

Fig. 7. Detailed depiction of the hourly changes of thermal energy from Fig. 6 during a few summer (left) and winter days.

Fig. 5 depicts the relationship of quantity between q and 4 as a function of solar fraction a. In fact, the ratio q/4 indicates the annual average storage capacity utilization rate. It is seen that the initially growing value of q/4 reaches saturation near the characteristic coordinate aS 0.5, at which q equals about 0.40e0.47 fraction of 4, depending on the power plant site. Hence, in the practically useful range of solar fraction values a 0.3e0.5, the mean daily discharge duration q is at least two times less than the nominal storage capacity 4. Regarding, for example, the GS plant site with a solar fraction a 0.42 and using Figs. 4 and 5, one can obtain 4 10 h and q 4.6 h. Furthermore, immediately following aS, the ratio q/4 decreases dramatically, reaching as low as about q 0.014 4 at a 1.0, with the approximate values of q 18 h and 4 1290 h. The last results provide clear evidence on the size of long-term or seasonal storage that should have a nominal storage capacity 4 within an order of magnitude between 1 and 2 in relation to the daily used amount of the storage energy represented by the discharge parameter q. Mechanisms behind the complex behavior of thermal storage variables such as 4 and q in the midrange of a, around the bend point S (Figs. 4 and 5), are discussed in the next section below. 4. Diurnal vs. seasonal energy proles The storage thermal energy ES as dened in Eq. (5) is a historydependent variable inuenced by the preceding processes of solar power supply and load demand. Fig. 6 shows the thermal energy variation with time in a storage system with a nominal capacity 4 10 h, which provides a solar fraction a 0.42 to the base load power plant. A higher resolution level of ES (t) during some selected

days of the annual cycle is presented in Fig. 7. For this relatively short-term storage, diurnal-scale variability of thermal energy is obvious, in contrast with the yearlong period of ES (t) observed with the pure solar system in Fig. 3. The diurnal storage has not only a short operating time after sunset in the summer days, as shown in the left graph of Fig. 7, but also a quite low energy level and irregular performance in the wintertime (the right graph), when solar power conditions are often inadequate to fully charge the storage during sunlight hours of a day (regarding the Northern Hemisphere). The signicant difference in the performance of diurnal storage in summer and winter seasons is clearly seen in Fig. 8, which demonstrates the variation of month-mean values of solar fraction during a year using the data from Fig. 6. It should be mentioned that due to the total daily discharge of short-term storage, the start date of its annual operating cycle is not as important as for the long-term storage shown in Fig. 3, and may be chosen arbitrarily. The variability of thermal energy during an annual operating cycle can be effectively characterized by a statistical coefcient of variation c calculated as the ratio of standard deviation to its mean. Fig. 9 depicts the relationship between the variable c and the solar fraction a based on the simulation data for the three solar power plant locations. In general, it could be expected that a higher solar fraction leads to a lower value of c because large storage systems should be less inuenced by differences in the daily solar energy balance. The data of Fig. 9, however, exhibit a deviation from the tendency of monotonic decreasing of c with solar fraction that occurs in the a midrange. It is seen that the coefcient of variation c has an internal minimum near the characteristic coordinate aS 0.5 followed by rising c between a 0.5 and 0.6, in some case up to 0.7. For the higher solar fractions, a > 0.6, there is a denite reduction in

Fig. 8. Month-mean values of solar fraction (white bars) based on the data of Fig. 6; annual average a 0.42.

Fig. 9. Thermal energy coefcient of variation c as a function of storage capacity 4 for the three power plant sites studied.

R. Adinberg / Applied Thermal Engineering 31 (2011) 3588e3594 Table 2 Basic thermal storage capacity scales (related to the GS site). Type Diurnal Intermediate Seasonal Annual Storage capacity, 4 0e14 h 14e100 h 5e59 days 59 days Solar fraction, a 0.2e0.5 0.5e0.60 0.60e1.0 1.0

3593

Solar collector factor, b 1.0e2.2 2.2e2.8 2.8e4.4 4.4

Fig. 10. Thermal energy variations ES(t) calculated for a storage capacity 4 100h (the GS power plant site).

energy variations from about c 2.0 to a numerical value of 0.6, which is specic for energy storage in the pure solar cycles. The irregular behavior of energy variations observed in the solar fraction range a 0.5e0.6, or, respectively, within the storage capacity values 4 14e100 h (Fig. 4), appears to be associated with the transformation of storage energy variations from the diurnal timescale, like that seen in Fig. 6, to a much longer period related to seasonal storage, up to the yearlong pure solar cycle shown in Fig. 3. Figs. 10 and 11 demonstrate the development of the seasonal energy prole in response to increased storage capacity 4. For 4 100 h in Fig. 10, the massive energy surge occurring in the middle of summer can be attributed to the origination of seasonal storage, though in the remaining time of the year this energy prole has the typical diurnal appearance. With further increasing of storage capacity towards tens of full load operating days, as shown in Fig. 11, the seasonal component of the thermal energy prole expands largely, nally reaching the outline of a pure solar cycle. The above results of the energy statistical analysis imply a denition of different thermal storage capacity scales such as is given in Table 2. Certainly, the presented 4-range division is quite conditional, as it may somewhat depend on a number of system-related factors, however, the resulting impact is not expected to be substantial. 5. Solar collector factor As shown in Fig. 12 for solar power systems including thermal storage, the solar collector factor b evaluated from the energy balance Eq. (4) is linearly related to the solar fraction a within the

whole range of its variation, 0.2 < a 1.0. The following values of the parameter b were obtained for the different pure solar plants sites (a 1.0): LV e 4.0, GS e 4.4, and MR e 4.8. By taking a value of b above that resulting from the energy balance, it is possible to obtain a higher solar fraction a in comparison to what is shown in Fig. 12, without changing the respective storage capacity. The larger solar eld has the advantage of providing additional amount of solar energy in the low-sun season (winter) when the utilization of energy storage is rather low (e.g. Fig. 6) and the corresponding month-mean values of solar fraction are signicantly below its annual average (e.g. Fig. 8). However, this method of increasing the solar fraction of a power plant, if applied, would inevitably cause some amount of solar energy to be rejected by the system (as shown in Fig. 1) that might occur when the storage is fully charged, most likely in the summer months. Concerning estimation of the time dependent values of the rejected power Qrj, the following control statement was used in addition to Table 1:

If Q 0 ! Q L and ES

4; then Q rj Q 0 Q 1 Q 2 ; else Q rj 0

(12)

Fig. 13 demonstrates the effect of the solar collector factor b on the performance parameters of a base load power plant within the diurnal and partly the intermediate storage capacity ranges. Here, the solid lines are contours of b and the dashed curves are contours of the storage capacity 4. It is seen, for example, choosing a storage capacity 4 10 h, that changing from b 1.8 to 2.2 causes an increase in solar fraction from 0.42 to 0.48, respectively. In this case, about 4% of the total annual amount of solar energy is rejected (some of this energy could be utilized as an auxiliary heat source for covering the plants needs). As follows from Fig. 13, the most favorable conditions for altering the parameter b are found below approximately 5% of the rejected energy, where the ascent of the 4-curves is rather low as compared with that in the upper part of the diagram. Moreover, the

Fig. 11. Seasonal storage cycles for the GS power plant site. The attached numbers indicate the values of storage capacity 4 in full load operating days.

Fig. 12. Solar collector factor b as a linear function of solar fraction a for the three power plant sites studied.

3594

R. Adinberg / Applied Thermal Engineering 31 (2011) 3588e3594

advanced thermal storage systems. The state of the art solar thermal storage technologies utilizing either sensible or latent heat of molten salts or various other materials are limited to the diurnal range of capacities as they suffer primarily from a relatively low energy density of the exploiting storage materials and also from signicant heat losses from the system during lengthy operating cycles. Hence, the further development of thermal storage technology should be particularly aimed at vastly more powerful energy storage concepts extending the storage capability into the intermediate and seasonal scale ranges, in order to achieve a higher level of sustainability for solar thermal electric systems.

Acknowledgements This research was sponsored in part by the FP7 European Research Infrastructures Project e SFERA: Solar Facilities for the European Research Area (http://sfera.sollab.eu/).
Fig. 13. The effect of the enlarged solar collector factor b upon both the solar fraction a and the amount of rejected energy, as the annual integral of Qrj(t), calculated for storage capacities 4 7, 10, 14 and 20 h (the GS power plant site). The rejected energy is given as percentage of the total annual amount of solar energy, the left part of Eq. (3).

References
[1] Program on technology innovation: Evaluation of concentrating thermal energy storage systems. EPRI 1018464, Palo Alto, CA, 2009. [2] Solar power and chemical energy systems, SolarPACES annual Report 2008. in: M. Geyer (Ed.). Germany, DLR, Koln, 2009. [3] A. Fernandez-Garcia, E. Zarza, L. Valenzuela, M. Perez, Parabolic-trough solar collectors and their applications, Renew. Sust. Energ. Rev. 14 (2010) 1695e1721. [4] F. Trieb, C. Schillings, M. OSullivan, T. Pregger, C. Hoyer-Klick, Global potential of concentrating solar power. SolarPACES Conference Berlin (September 2009). http://www.trec-uk.org.uk/reports/Solar_Paces_Paper_Trieb_Final_Colour_ corrected.pdf (accessed June 2011). [5] Variability of wind power and other renewables: Management options and strategies. IEA, 2005. http://www.uwig.org/IEA_Report_on_variability.pdf. [6] M.Z. Jacobson, Review of solutions to global warming, air pollution, and energy security, Energ. Environ. Sci. 2 (2009) 148e173. [7] U. Herrmann, D.W. Kearney, Survey of thermal energy storage for parabolic trough power plants, J. Sol. Energ. Eng. 124 (2002) 145e152. [8] H. Mller-Steinhagen, F. Trieb, Concentrating solar power: a review of the technology, Ingenia, Royal Acad. Eng. 18 (2004) 43e50. [9] W.-D. Steinmann, M. Eck, D. Laing, Solar thermal parabolic trough power plants with integrated storage capacity, Int. J. Energ. Technol. Policy 3 (1/2) (2005) 123e136. [10] H. Price, A parabolic trough solar power plant simulation model, NREL/CP550e33209, National Renewable Energy Laboratory. Cole Boulevard, Golden, Co, 2003. [11] J.J. Iannucci, The impacts of storage upon solar plants: general principles and seasonal applications, SAND808242. Sandia National Labs, Livermore, CA, 1981. [12] H. Price, E. Lpfert, D. Kearney, E. Zarza, G. Cohen, R. Gee, R. Mahoney, Advances in parabolic trough solar technology, J. Sol. Energ. Eng. 124 (2002) 109e125. [13] B.C. Staley, J. Goodward, C. Rigdon, A. MacBride, Juice from concentrate: Reducing emissions with concentrating solar thermal power. The World Resources Institute (WRI) (2009). http://www.wri.org/publication/juice-from-concentrate (accessed June 2011). [14] V. Quaschning, Technical and economical system comparison of photovoltaic and concentrating solar thermal power systems depending on annual global irradiation, Sol. Energ. 77 (2004) 171e178. [15] C.K. Ho, Software and codes for analysis of concentrating solar power technologies, SAND2008-8053. Sandia National Laboratories, Albuquerque, NM, 2008. [16] M.J. Montes, A. Abanades, J.M. Martinez-Val, M. Valdes, Solar multiple optimization for a solar-only thermal power plant, using oil as heat transfer uid in the parabolic trough collectors, Sol. Energ. 83 (1) (2009) 2165e2176. [17] Solar Radiation Series of Data, http://www.soda-is.com/eng/services/index. html, (accessed June 2011). [18] The National Solar Radiation Data Base, http://rredc.nrel.gov/solar/old_data/ nsrdb, (accessed June 2011). [19] J.E. Pacheco, S.K. Showalter, W.J. Kolb, Development of a molten salt thermocline thermal storage system for parabolic trough plants, J. Sol. En. Eng. 124 (2002) 153e159. [20] R. Tamme, D. Laing, W.-D. Steinmann, Advanced thermal energy storage technology for parabolic trough, J. Sol. En. Eng. 176 (2004) 794e800. [21] H. Michels, R. Pitz-Paal, Cascaded latent heat storage for parabolic trough solar power plants, Sol. Energ. 81 (2007) 829e837. [22] R. Adinberg, D. Zvegilsky, M. Epstein, Heat transfer efcient thermal energy storage for steam generation, Energ. Convers. Manag. 51 (2010) 9e15.

higher the storage capacity, the stronger is the impact of the solar collector factor, resulting in a larger solar fraction increment for a power plant. For instant, taking 4 20 h and setting the rejected energy at 4%, it is possible, by changing the parameter b from 2.4 to 3.0, to raise the solar fraction from a 0.54 to 0.65. 6. Conclusions The process characteristics of thermal storage presented in this study were computed over a wide range of thermal storage capacities varied from a few hours to the order of magnitude of a thousand full load operating hours. The corresponding solar fraction values of hybrid solar-fossil power plants supported by energy storage ranged between 0.2 (no storage applied) and 1.0 (the pure solar operation). Although the modeling was specically developed to simulate parabolic trough systems operated as base load plants, the obtained results can qualitatively be extended to most other CSP systems and also to different continuous load proles (except peaking power generation), since the dominating factor affecting system performance is the intermittent solar power source. Based on the statistical analysis of thermal energy variations in time throughout an annual operating cycle, four distinctive storage scales were identied, referred to as diurnal (4 < 14 h), intermediate (14e102 h), seasonal (102e103 h) and yearly (w103 h) storage capacities. The operational performance of the diurnal-scale storage was shown to be strongly inuenced by hourly and daily solar power conditions so that thermal storage efciency with respect to month-mean solar fractions is considerably smaller in the low-sun season (winter) than in the high-sun season (summer). Correspondingly, the annual solar fraction of hybrid power plants employing thermal storage with a nominal capacity within 14 full load operating hours is limited to below 0.5, so that half or more of the energy input must be provided by burning fossil fuels. An improved diurnal storage efciency resulting in a solar fraction of 0.6 can be achieved by increasing the solar eld size along with losing some a few percent of the total annual amount of solar energy collected. The results of simulations show that a pure solar power plant (solar fraction equal 1.0) should include a storage capacity almost 2 orders of magnitude greater than what is available with the most