Applied Energy 88 (2011) 32283238

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Performance analysis of an Integrated Solar Combined Cycle using Direct Steam Generation in parabolic trough collectors
M.J. Montes a,, A. Rovira a, M. Muoz a, J.M. Martnez-Val b
a b

Department of Energy Engineering, Universidad Nacional de Educacin a Distancia (UNED), C/Juan del Rosal No. 12, 28040 Madrid, Spain Department of Energy and Fluid Mechanics Engineering, Madrid Polytechnical University, C/Jos Gutirrez Abascal No. 2, 28006 Madrid, Spain

a r t i c l e

i n f o

a b s t r a c t
The contribution of solar thermal power to improve the performance of gas-red combined cycles in very hot and dry environmental conditions is analyzed in this work, in order to assess the potential of this technique, and to feature Direct Steam Generation (DSG) as a well suited candidate for achieving very good results in this quest. The particular Integrated Solar Combined Cycle (ISCC) power plant proposed consists of a DSG parabolic trough eld coupled to the bottoming steam cycle of a Combined Cycle Gas Turbine (CCGT) power plant. For this analysis, the solar thermal power plant performs in a solar dispatching mode: the gas turbine always operates at full load, only depending on ambient conditions, whereas the steam turbine is somewhat boosted to accommodate the thermal hybridization from the solar eld. Although the analysis is aimed to studying such complementary effects in the widest perspective, two relevant examples are given, corresponding to two well-known sites: Almera (Spain), with a mediterranean climate, and Las Vegas (USA), with a hot and dry climate. The annual simulations show that, although the conventional CCGT power plant works worse in Las Vegas, owing to the higher temperatures, the ISCC system operates better in Las Vegas than in Almera, because of solar hybridization is especially well coupled to the CCGT power plant in the frequent days with great solar radiation and high temperatures in Las Vegas. The complementary effect will be clearly seen in these cases, because the thermal power provided by the solar eld compensates the gas turbine part load performance due to the high temperatures. The economical analysis points out that this hybrid scheme is a cheaper way to exploit concentrated solar energy, although it is limited to a small fraction of the combined cycle power. The analysis also shows that the marginal cost of solar electricity is strongly inuenced by the goodness of coupling, so this cost is lower in Las Vegas than in Almera. 2011 Elsevier Ltd. All rights reserved.

Article history: Received 23 December 2010 Received in revised form 25 February 2011 Accepted 25 March 2011 Available online 16 April 2011 Keywords: Integrated Solar Combined Cycle Direct Steam Generation Parabolic trough Solar thermal power plants

1. Introduction Integrated Solar Combined Cycle (ISCC) power plants have been widely studied as an alternative to the conventional arrangement of parabolic through collectors coupled to a Rankine power cycle (that will be referred as solar-only plant in this work). In this paper, a specic conguration of ISCC plant is proposed, and the coupling between the solar eld and the combined cycle is studied. For that, the annual performance of the proposed conguration and its corresponding reference Combined Cycle Gas Turbine (CCGT), dened in next sections, have been carried out in two locations, comparing both the electricity production and the costs, for each of the selected locations.

Corresponding author. Tel.: +34 913986465; fax: +34 913987415.

E-mail addresses: mjmontes@ind.uned.es (M.J. Montes), rovira@ind.uned.es (A. Rovira), mmunoz@ind.uned.es (M. Muoz), mval@etsii.upm.es (J.M. Martnez-Val). 0306-2619/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2011.03.038

There are several advantages of the ISCC plants over the solaronly plants [1,2]. For instance, overall solar-to-electric conversion efciency is higher in the ISCC systems. Also, integration arrangements overcome thermal losses associated to the daily start-up and shutdown processes of the steam turbine. Finally, if solar integration is accomplished on an existing CCGT power plant, incremental costs of replacing the steam turbine by a greater one are lower than the overall unit cost in a solar-only plant. Besides the advantages mentioned above, it may be added that this conguration does not require any storage system (but it may be included in the scheme) to ensure a manageable range of electricity output, which is one of the crucial issues in electricity production by renewable energy. Dersch showed that, whether the strategy adopted is a solar dispatching or a scheduled load operation mode, there is always a power level of the plant that is guaranteed [1]. The difference between the two performance strategies is based on the use of an auxiliary back-up fossil boiler to supply energy when solar energy is not available.

M.J. Montes et al. / Applied Energy 88 (2011) 32283238

3229

Nomenclature CCGT DNI DSG HRSG Combined Cycle Gas Turbine direct normal irradiation Direct Steam Generation Heat Recovery Steam Generator ISCC LCE TMY Integrated Solar Combined Cycle Levelized Cost of Energy typical meteorological year

Several thermal power plants with this scheme are being constructed, like the ISCC projects in Egypt, Morocco and Algeria [3]. The conguration proposed in those projects basically consists of a parabolic trough solar eld coupled to the bottoming steam cycle of a CCGT power plant. Parabolic troughs are cooled by synthetic oil, so an intermediate oil-to-water/steam heat exchanger is needed between the solar eld and the power block. This particular conguration is called Heat Transfer Fluid (HTF) technology, and it is the most used in parabolic trough thermal plants, where the power block is usually a Rankine cycle. A conventional ISCC scheme with a HTF solar eld is shown in Fig. 1. Instead of that approach, another scheme is proposed in this paper. This scheme consists of a Direct Steam Generation (DSG) parabolic trough plant coupled to the bottoming steam cycle of a CCGT power plant. This layout presents several advantages compared to conventional ISCC layouts. For example, there is no need of an intermediate heat exchanger between the solar eld and the steam cycle, resulting in lower investment. Thermal and exergy losses associated to that system are also avoided, improving overall efciency. Furthermore, coupling with DSG presents another important advantage for the deployment of this technology: the ability of guaranteeing the electricity production during non-radiation periods without using a thermal storage, a system that is still under development for the DSG technology [4,5]. A similar arrangement to the one studied in this paper is proposed in [6]. There, two different DSG solar elds are used to produce high pressure (100 bar) and low pressure (5 bar) steam. However, in the conguration proposed in this work, only the high-pressure level (at 90 bar) is considered for the solar coupling, with the aim of using the solar thermal energy at the highest exergy level, as it will be explained later. Furthermore, not only the boiling process is accomplished in the parabolic troughs but also the preheating, which reduces the average temperature of the

water in the troughs and, therefore, the thermal losses from the tubes to the environment decreases. Thus, the water is conducted to the solar eld before the inlet to the Heat Recovery Steam Generator (HRSG). Another similar arrangement is the one proposed in [7]. In this case, preheating, boiling and superheating are accomplished in the parabolic trough collectors, to supply steam to the high pressure steam turbine. The proposed design aims to avoid the steam superheating in the parabolic troughs, which introduces several technical difculties [8]. 2. Integrated Solar Combined Cycle proposed The ISCC scheme proposed has been already presented by the authors of this paper in [9]. That work was focused on the analysis of different solar hybridization sizes for the same location (Canary Islands, Spain). In this paper, only one hybridization size (50 MWth approximately, at the design-point conditions, although the steam turbine has only been oversized in 25 MWth, as it will be explained later) has been simulated, but in two different locations with different warm climates. The scheme proposed is based on a conventional CCGT power plant around the 220 MWe power rating. The CCGT consists of a 2 1 conguration: two gas turbines of 72.5 MWe each, two HRSGs that produce steam at high pressure (90 bar) and at low pressure (8 bar) and a steam turbine of 75 MWe. As it can be seen in Fig. 2, there is only one extraction line from the steam turbine, and it is directed to a deareator. Liquid water from the deareator is pumped to the low and high pressure economisers of the HRSG, and to the solar eld. DSG parabolic trough plant is coupled to the high-pressure level in the HRSGs. Integration is carried out at the higher pressure level. This pressure is in the optimum working pressures range to oper-

Fig. 1. Conventional ISCC with a HTF parabolic trough eld.

3230

M.J. Montes et al. / Applied Energy 88 (2011) 32283238

Fig. 2. Integration scheme for a DSG plant in the two pressure bottoming steam cycle of a CCGT.

ate a DSG solar eld, because pressure drop in the collector loop is reduced [10] and overall efciency of the solar thermal power plant is signicantly improved [11,12]; besides that, the solar thermal power is used at the highest temperature level. Feedwater from the deareator is pumped to the solar eld, where it is preheated and evaporated. Wet steam from the parabolic trough collectors is introduced into the high pressure drums of the Rankine cycle. The steam from the high pressure drum is then superheated by the exhaust gases in the HRSGs before feeding the steam turbine. The strategy adopted by the power plant is a solar dispatching mode. With this operation mode, it is not necessary to consider either auxiliary fossil-red boiler or a thermal storage system. Also, the gas turbines performance is not affected by the solar integration; these turbines operate according to the ambient temperature and pressure from the typical meteorological year chosen for the simulation. Finally, the steam turbine is over-powered in 25 MWth, compared to the conventional CCGT, in order to assume the solar thermal power. It was observed in [9] that if the turbine is boosted to absorb all the solar thermal power at the design-point conditions, it operates at pressures quite below the design pressure during the whole year, because design conditions for the gas turbines and the solar eld are not simultaneously reached in any time. For that reason, the suitable oversizing was investigated in this work, resulting in 25 MWth, although the nominal thermal hybridization is greater, around 50 MWth. In this search, it was intended that the steam pressure of the high-pressure level was near the design point along the year with a maximum overload below 10 bar. The following subsections are devoted to explain the different models used to obtain the daily operation and annual performance of the plant.

Table 1 Design parameters of the gas turbine of the combined cycle. Parameter Compression ratio Air mass ow (kg/s) Gas mass ow (kg/s) Turbine inlet temperature (K) Turbine outlet temperature (K) Power (MWe) Efciency (%) Value 16 210 214 1450 828 72.6 35.1

Table 2 Design parameters of the HRSG of the combined cycle. Parameter High pressure steam temperature (K) High pressure (bar) High pressure pinch point (K) Low pressure steam temperature (K) Low pressure (bar) Low pressure pinch point (K) Solar thermal power (MWth) Value 818 90 10 566 5 10 12.5

2.1. Combined Cycle Gas Turbine model The design data selected for the CCGT are similar to those of actual 220 MWe combined cycle power plants of two levels of pressure [13]. Design parameters of the gas turbine are summarized in Table 1, for the ISO conditions (15 C and 1.013 bar). HRSG sizing is based on the design parameters shown in Table 2. The simulation program for combined cycles has been developed, validated and used in previous works [1315]. That software enables the pre-design and simulation of every single component

of the combined cycle. These components are joined together yielding a particular plant conguration. Therefore any combined cycle conguration can be simulated, for both design and off-design conditions. The full load simulation is achieved by means of mass and energy balances applied to every component of the CCGT. The results are the thermodynamic properties of the steam and the gas at every point of the cycle, the CCGT power and efciency and the heat exchanged at each element of the HRSG. CCGT part load and off-design simulation requires the prediction of the performance of the cycle elements at every operating condition. Once the full load simulation has been done the following data must be introduced or calculated: the characteristic curves of every turbomachine, the UA product (global coefcient of heat transfer by the heat transfer area) and the ambient conditions (pressure and temperature). El-Gammal [16] and Stamatis et al. [17] supply dimensionless curves that can be used to extrapolate the performance of gas turbines in whatever condition. The equations that describe the HRSG performance can be found, for

M.J. Montes et al. / Applied Energy 88 (2011) 32283238

3231

example, in [14]. The variation of the heat transfer coefcient U as the gas turbine load is modied is described by Valds and Rapn [18]. Finally, Cotton [19] gives some correlations for the steam turbine performance prediction. For the work presented in this paper, the solar hybridization has been incorporated to the combined cycle program, through a series of correlations and matrices shown in Section 2.3. Owing to the variable ambient conditions, the power plant generally works at off-design conditions during the annual simulation [13], although it always maintains the gas turbine inlet temperature (after the combustion chambers) at its design value. Because of that, drum pressure is variable: as the steam mass ow decreases, required turbine pressure also decreases and, consequently, drum pressure decreases. For this reason, it was necessary to consider the inuence of the outlet solar eld pressure as it will be explained in next sections. 2.2. Solar eld conguration and performance at the design-point conditions The solar eld design-point adopted is solar noon on June 21st. Table 3 shows the main design point parameters at this time, for the two selected sites: Almera and Las Vegas. Every collector loop is composed of eight collectors: three collectors approximately are devoted to water preheating whereas the other ve are used for water evaporation. Each collector consists of eight modules of 12.27 m, so total loop length is 785 m. Geometrical and optical parameters of a collector loop are summarized in Table 4 [20]. These parameters are very similar to the corresponding in the Eurotrough-150 (ET-150) collector model. The main difference is the absorber tube thickness, which is greater in the DSG parabolic trough collectors, compared to the oil refrigerated collectors, owing to the higher working pressure. Outlet design pressure of the wet steam from the solar eld has been set to 90 bar. This value implies that solar eld inlet design conditions must be set over 92 bar. Pressure drop is slightly greater in Las Vegas because of the section devoted to evaporation is greater, as design DNI is also greater. The simulation model for the parabolic trough collectors using water-steam as heat transfer uid, has already been developed and validated in other works [10,21,22]. Table 5 summarizes design-point results for one collector loop, for both sites. In Table 5 energy efciency is dened as

Table 4 Geometrical and optical parameters for the collector loop considered. Geometrical parameters for the collector loop Absorber tube outer diameter (m) Absorber tube inner diameter (m) Glass envelope outer diameter (m) Glass envelope inner diameter (m) Number of collectors (preh + boiling) Number of modules per collector Length of every module (m) Mirror length in every module (m) Optical parameters for the collector ET-150 Intercept factor Mirror reectivity Glass transmissivity Solar absorptivity Peak optical efciency Thermal emissivity (C)

0.07 0.055 0.115 0.109 8 8 12.27 11.9 0.92 0.92 0.945 0.94 0.75 0.04795 + 0.0002331 T

Table 5 Design-point simulation results for one collector loop. Parameter Mass ow per loop (kg/s) Heat gain (MWth) Energy efciency Heat loss (kWth) Pressure drop (bar) Outlet quality Almera (Spain) 1.44 2.684 74.81 159.8 2.355 0.63 Las Vegas (USA) 1.44 2.848 74.79 170.5 2.689 0.71

the solar eld size is the same for both locations, which is better suited for the following comparative analyses. 2.3. Solar eld performance under different irradiation values and outlet pressures For annual simulation, the same strategy that in [20,21] has been adopted: it has been considered that incidence angle remains constant (0) and only DNI changes from 350 to 950 W/m2. Incidence angle effect must be taken into account when annual performance is calculated, considering that hourly DNI is reduced by the incidence angle modier determined in the center of every interval. The main difference between the strategy presented in [20,21] and the one used in the present work is that not only DNI is taken into account, but also solar eld outlet pressure must be considered. As it was explained in Section 2.1, several simulations with different outlet pressures are necessary to match DSG solar eld model to combined cycle program. Simulations of one parabolic trough loop have been performed for different outlet pressures, from 50 bar to 100 bar, giving a solar matrix as result for every considered parameter: mass ow through the absorber tube, heat gain per collector loop and solar efciency. The results show that mass ow is nearly independent of the outlet pressure, whereas it depends linearly on the DNI. For that reason, mass ow can be modeled by a rst-order polynomial function of the DNI

genergy  


collector loop

 Q th;net   DNIincident collector

1
loop

where genergy is the energy efciency; Qth,net (kWth) is the net heat gain per collector loop; DNIincident (kWth) is the incident DNI normal to the collector aperture area, i.e. taking into account the incidence angle value at the design point. The heat gain shown in Table 5 is referred to one loop, and is greater in Las Vegas than in Almera. For a nominal solar hybridization of 50 MWth, the number of parallel loops has been set to 17, so
Table 3 Design point parameters for the ISCC power plant located in Almera and in Las Vegas. Design point parameters Direct normal irradiance (W/m2) Altitude (m) Longitude () Latitude () Ambient temperature (C) Incidence angle (NS axis orientation) () Almera (Spain) 850 366 2210 1900 W 37050 27.800 N 25 13390 14.400 Las Vegas (USA) 900 664 115100 1200 W 3640 4800 N 35 12380 34.800

_ jcollector m

loop

1:68 DNIincident 0:0495

_ (kg/s) is the uid mass ow per collector loop; and In Eq. (2), m DNIincident (kWth) is the incident DNI normal to the collector aperture area. The values of the heat gain per collector loop t to a rst-order polynomial function of DNI and outlet solar eld pressure

 Q th;net collector

loop

3:4 103 DNIincident

100 Poutlet 105 2

3232

M.J. Montes et al. / Applied Energy 88 (2011) 32283238

It is also important to analyze the occurrence of high solar radiation at high temperatures. In this case, the number of hours in which DNI is greater than 900 W/m2, and ambient temperature is greater than 30 C is equal to 2 in Almera, whereas it is equal to 303 in Las Vegas. 3.1. Annual simulation results Annual simulations have been carried out for both locations and for the two congurations considered: the ISCC and the conventional CCGT. The main results are shown in Table 6. It can be observed that annual electricity production of the conventional CCGT conguration is greater in Almera than in Las Vegas, mainly because temperatures are more moderate in the former location. However, in the case of hybridized scheme, the power plant located in Las Vegas presents a higher annual electricity production because there is a greater number of hours at which the solar thermal power available is high. In Table 6, the annual ISCC efciency is dened as the annual average of all the hourly efciencies (gISCC) during the year, which are calculated by means of

Fig. 3. Solar efciency per collector loop as a function of the DNI and the outlet solar eld pressure.

where Qth,net (kWth) is the net heat gain per collector loop; DNIincident (kWth) is the incident DNI normal to the collector aperture area; and Poutlet (bar) is the uid pressure at the outlet of the solar eld. The coefcients of Eqs. (2) and (3) have been obtained by a multiple linear regression. For the mass ow, the coefcient of multiple determination (R2) is 0.99, and the error variance is 3.84e07. For the heat gain, the coefcient of multiple determination (R2) is 0.99, and the error variance is 4.69e06. Finally, efciency is a more complex function of the DNI and the solar eld outlet pressure, so a two-dimensional interpolation is necessary. Fig. 3 shows the solar matrix for the solar efciency, referred to one loop. 3. Annual simulation: comparative analysis and results As in CCGT power plants, ISCC performance is considerably inuenced by ambient conditions, like temperature and pressure. Therefore, it is fundamental to use annual performance calculations for ISCC plant analysis. Meteorological data for the annual simulation were obtained from the Meteonorm Database, by means of the simulation program Trnsys 17 [23]. Hourly data for DNI, ambient temperature and pressure were carefully analyzed for the two locations considered: Almera and Las Vegas. Ambient pressure is similar in both locations (around 1.01 bar in Almera and 0.94 bar in Las Vegas) and it does not change much during the year for each site. It is more interesting to analyze the number of hours at high DNI (in which the solar thermal power is high), and at high temperature (in which the gas turbines power is low). Both parameters are shown in Fig. 4.

gISCC

NISCC mf Hc Ac DNIincident

where NISCC (kWe) is the power produced by the ISCC power plant; mf (kg/s) is the fuel mass ow consumed; Hc (kJ/kg) is the fuel low

Table 6 Annual simulation results. Parameter Annual electricity production ISCC power plant (GW he) Annual electricity production CCGT power plant (GW he) Annual fuel consumption (GW hth) Solar thermal power supplied to the power block (GW hth) Solar thermal power incident on the parabolic trough eld (GW hth) Electricity production increment ISCC versus CCGT power plants (GW he) Annual global efciency ISCC power plant (%) Annual net incremental solar efciency (%) Annual net electrical solar fraction (%) Almera 1880.54 1857.43 3496.32 78.12 107.51 23.10 52.18 21.49 1.23 Las Vegas 1891.34 1845.9 3478.01 121.97 166.44 45.44 51.90 27.30 2.40

Fig. 4. Inverse cumulative function of the direct normal irradiation and ambient temperature in Las Vegas and in Almera.

M.J. Montes et al. / Applied Energy 88 (2011) 32283238

3233

heating value; Ac (m2) is the collector area; and DNIincident (kW/m2) is the incident DNI normal to the collector aperture area. This denition is referred to 1 h, which is the time step chosen in the annual simulation. This annual efciency is greater in the ISCC power plant located in Almera than in Las Vegas because the thermal power from the solar eld is greater in this last location, and the solar-to-electricity conversion efciency is minor than the fossil-to-electricity conversion efciency. Because of the gas turbine cycle in the scheme proposed is identical for the conventional CCGT and for the ISCC power plant, the annual fuel consumption is the same for both congurations in every location. It can be observed that this consumption is lower in Las Vegas than in Almera, which in fact means that conventional CCGT annual electricity is lower in Las Vegas than in Almera, as it was mentioned before. Nevertheless, in the ISCC system, the lower annual fuel consumption does not yield to lower electricity production, owing to the solar contribution. The net incremental solar efciency [1] (gnet_inc_solar) is dened as the ratio between the annual increment of net electricity produced by the ISCC system compared to the CCGT power plant (the same gas turbine and using the same amount of fuel), and the annual solar thermal power. The value presented in Table 6 is the annual average of all the hourly net incremental solar efciencies, calculated by

gnet

Inc: solar

N ISCC NCCGT Q th solar

In Eq. (5), NISCC (kWe) is the electricity produced in the ISCC power plant; NCCGT (kWe) is the electricity produced in the conventional CCGT reference plant; and Qth_solar (kWth) is the thermal power from the solar eld. This denition is referred to 1 h, which is the time step chosen in the annual simulation. The last parameter in Table 6 is the net electrical solar fraction [1], xnet_elec_solar, and it is based on the net incremental solar efciency, in such a way that a higher net incremental solar efciency also means a higher net electrical solar fraction, dened in

xnet

elec solar

gnet

Inc: solar

Q th NISCC

solar

In Eq. (6), gnet_inc_solar is the net incremental solar efciency, Qth_solar (kWth) is the thermal power from the solar eld; and NISCC (kWe) is the electricity produced in the ISCC power plant. This denition is referred to 1 h, which is the time step chosen in the annual simulation. As expected, Table 6 shows that the last two parameters are substantially higher in Las Vegas, since the difference in the annual electricity production between the ISCC and the CCGT congurations is higher than in Almera. Figs. 57 present the electricity production by the CCGT and ISCC schemes considered during some characteristic days for both locations: the day with highest ambient temperature and the day with highest solar radiation. The objective of these graphics is to show the effect of ambient conditions in the CCGT and ISCC performance. Fig. 5 shows the CCGT performance during the days of highest temperature in both locations. These 2 days have been represented in the same graphic to point out that the maximum temperatures reached in Las Vegas are considerably higher than those of Almera. In particular, the maximum temperature in Almera is 37.14 C. There are 397 h in Las Vegas that exceed this value, and the maximum temperature is 44.38 C, as shown in the graph. The CCGT operation represented in Fig. 5 is the worst performance that can occur in each of the localities, with a minimum of electricity production. This minimum is much lower in Las Vegas than in Almera: 85% of the nominal output of the plant, compared to the 88% in Almera. This difference is because the maximum temperature is much higher in Las Vegas and, therefore, the gas turbine works worse. Fig. 5 only analyzes the conventional CCGT performance, thus the DNI has not been displayed in this gure, as it does not affect the operation of this conguration. The days simulated in Figs. 6 and 7 are characterized by de highest DNI in Las Vegas and Almera, respectively. In this case DNI is also signicantly higher in Las Vegas than in Almeria, 967 W/m2 versus 940 W/m2. It can also be observed that temperatures at which DNI is higher are relatively moderate in Almera, around 25 C, while in Las Vegas, most of the days with high DNI are also days of very high temperature. This is explained because in Almera, being a seaside location, a mist is formed at high temperatures, causing days not so clear. These factors cause that the

Fig. 5. Electricity produced by the CCGT located in Las Vegas on June 27th (the day with highest ambient temperature) and the CCGT located in Almera on July 21st (the day with highest ambient temperature).

3234

M.J. Montes et al. / Applied Energy 88 (2011) 32283238

Fig. 6. Electricity produced by the ISCC and the CCGT power plants on June 20th (the day with highest solar DNI in Las Vegas).

Fig. 7. Electricity produced by the ISCC and the CCGT power plants on May 25th (the day with highest solar DNI in Almera).

difference between the daily production between the ISCC and the CCGT congurations is higher in Las Vegas than in Almeria for days of high DNI. This is shown graphically by the area enclosed by the solid lines representing electricity production, which is greater in Fig. 6 than in Fig. 7. Finally, it is important to point out that, when there is no solar radiation, the production of the ISCC plant is slightly lower than the output of the CCGT plant, as it is shown in Figs. 6 and 7. This is because the steam turbine of the ISCC plant is slightly oversized (25 MWth) to accept the solar thermal power. In the absence of this power, the steam turbine of the ISCC operates further from the design point than that of the conventional CCGT, so ISCC performance is slightly worse.

4. Economic analysis An economic analysis of the power plants studied has been carried out, calculating the Levelized Cost of Energy (LCE) for each one of them. In order to do an adequate comparison between the LCE from the ISCC and the CCGT power plants, the escalation rate of the fuel cost must be taken into account [24]. This parameter cannot be neglected in this calculation, as one of the main advantages of renewable energy versus fossil fuels would not be otherwise considered: its independence from the uncertainties associated with future cost of fossil fuels. All the parameters and data used in the economic analysis are summarized in Table 7:

M.J. Montes et al. / Applied Energy 88 (2011) 32283238 Table 7 Cost data used for the economic evaluation of the ISCC and the CCGT power plants. Investment Specic investment cost for solar eld (/m2) Specic investment cost for power block (/kWe) Specic land cost (/m2) Surcharge for construction, engineering and contingencies (%) Operation and maintenance Solar eld specic O&M costs (/m2 year) Combined cycle specic O&M costs (/kWe year) O&M equipment cost percentage of investment per year (%) Financial parameters Effective discount rate (%) Economic lifetime (years) Ination rate (%) Escalation rate O&M (%) Escalation rate fuel cost (%) Fuel cost Fuel type Natural gas price (c /kW h)

3235

Table 8 Total costs at the rst year, levelized costs and LEC values for the four congurations analyzed. ISCC Las Vegas Investment cost (Mio ) Annual O&M cost (Mio ) Annual fuel consumption (GW hth) Annual fuel cost (Mio ) Levelized costs (Mio ) Annual electricity production (GW he) LCE (/MW he) LCEsolar,marginal (/MW he) 261.94 7.48 3478.01 74.97 152.30 1891.34 80.52 79.65 ISCC Almera 261.94 7.48 3496.32 75.37 152.87 1880.54 81.29 156.64 CCGT Las Vegas 245.20 6.39 3478.01 74.97 148.68 1845.90 80.55 CCGT Almera 245.20 6.39 3496.32 75.37 149.25 1857.43 80.35

200 984 2 10 9 17.9 1 10 15 3 1 2.5 Natural gas 2.3199

The data of the costs of the different systems are all referenced to the same year (2010). The data for the solar eld have been estimated from [25,26], taking into account that the needed investment of the DSG system is about 7% lower than for conventional HTF solar elds [27]. The Natural Gas prices are referred in [28]. The specic costs of the conventional combined cycle have been estimated using the procedure shown in [29]. The total costs of a 528 MWe and a 1000 MWe CCGT power plant have been taken from [30,31]. Specic cost for a 220 MWe have been calculated by means of Eq. (7), taking into account the particularities shown in Fig. 8: the total cost of the combined cycle depends linearly of the nominal power of the plant, but the specic cost follows an hyperbolic variation.

steam turbine is slightly oversized. The value shown in Table 7 (984 /kW he) is for the 220 MWe power plant, and it is consistent with the values calculated in [32,33], for the same reference year. The economic evaluation has been done following the guidelines of [34] and it is based on the data summarized in Table 7. The specic equations for the calculation of the LCE are summarized in the Appendix A. Table 8 shows the results of this economic analysis. In Table 8, the LCE (/MW he) is calculated by

LCE

LCInv LCO&M LCFuel Eannual

x=kW he

360 106 466:1 528 103 y y

In Eq. (7), x (/kW he) is the specic cost of the combined cycle, and y (kW he) is the nominal power of the plant. For the CCGT power plant, the nominal power is 220,000 kW he and for the ISCC scheme, the nominal power is slightly higher, around 233,000 kW he, as the

where Eannual (MW he) is the annual electricity production by the power plant; and LCInv (), LCO&M () and LCFuel () are the levelized costs of the inversion, the operation and maintenance, and the fuel, respectively. Fist of all, it is observed that all the LCE values shown in Table A1 are very similar. This is because the solar hybridization is very small and does not cause large differences between the two congurations. If the solar hybridization were higher, it could have major differences. Regarding the ISCC plants, the LCE value is lower in Las Vegas than in Almera, mainly because the fuel consumption is lower in the former location whereas the annual electricity production is higher. As it was explained in the last section, the worse gas turbine performance is offset by the better operation from the

Fig. 8. Total cost and specic cost of a combined cycle power plant as a function of the nominal power.

3236

M.J. Montes et al. / Applied Energy 88 (2011) 32283238

Table A1 Numerical values of the parameters used in the economic calculation. Economic parameters calculation Capital-recovery factor (CRF) (%) Nominal escalation rate O&M (%) Nominal escalation rate fuel cost (%) k O&M k Fuel cost Constant-escalation levelization factor O&M (CELF) Constant-escalation levelization factor fuel cost (CELF) 13.15 4.03 5.58 0.95 0.96 1.30 1.44

solar eld, which has a higher net incremental solar efciency than in Almera. This effect is better quantied by the LCEsolar,marginal, dened in

LCEsolar;marginal

LCjISCC LCjCCGT DEannual

where LC () is the levelized cost for the ISCC and the CCGT plants, respectively, and DEannual (MW he) is the annual increment of electricity production, due to the solar hybridization. The LCEsolar,marginal represents the marginal cost of the additional kilowatthour generated in the oversized steam turbine of the ISCC conguration, due to the solar thermal power. In this case there are signicant differences in the value of this parameter for the two sites. This is because the annual electricity produced by this small solar eld is almost twice in Las Vegas than in Almeria (45 GW he compared to 23 GW he). It is important to point out that this marginal cost of the electricity produced from the solar eld is cheaper than the global cost of the electricity generated by de conventional CCGT in Las Vegas, so the hybridization size could be increased; whereas in Almera, this comparison is reversed, so the solar hybridization is slightly higher than this optimum value. All these considerations have been made without taking into account any special tariff for the electricity production from renewable sources. Although in this paper there is no an analysis of a solar-only plant, i.e., a parabolic trough eld coupled to a Rankine cycle, according to other references [35], it is clear that the solar electricity produced by the ISCC conguration is cheaper than the solar electricity generated by the solar-only plant, that is around 150 /MW he. In respect to the comparison between the ISCC and the CCGT power plants, the LCE value is slightly higher in the CCGT than in the ISCC for the case of Las Vegas, whereas the LCE of the CCGT located in Almera is lower than the LCE of the ISCC in the same loca-

Fig. 10. LEC values for the CCGT power plant as a function of the fuel cost at the rst year and the escalation rate of the fuel cost.

tion. These results are in line with the marginal cost of the additional solar kilowatthour that have been obtained and discussed above. The small differences between the LCE values in this case are mainly due to the small percentage of solar hybridization in the ISCC conguration analyzed in this work, which has a minor inuence on the total inversion and the annual electricity production. Fig. 9 shows the variation of the LCE values for the ISCC plant located in Las Vegas, as a function of the specic investment cost of the solar eld. At last, Fig. 10 shows the great inuence of the fuel cost at the rst year, as well as the escalation rate, on the LCE values. Only LCE values for the CCGT power plant in Las Vegas have been represented. According to the conguration analyzed, the fossil fuel consumed by the ISCC plant is the same for the same location, thus LCE values for this case follow a parallel variation. 5. Conclusions This work analyzes the annual operation of an ISCC system, in comparison to a conventional GTCC plant, assessing the importance of the solar thermal power to improve the gas turbine performance in days with high sun radiation and ambient temperature. For that, annual simulations have been conducted for two sites: Almera and Las Vegas. Although both locations are suitable for placing a solar thermal power plant in them, the benecence of the coupling solar eld-combined cycle is more evident in the latter location, which has a very hot and dry climate. To demonstrate this general effect of the ISCC systems, the particular selected arrangement is a DSG parabolic trough eld which coupled to the high-pressure level of the steam turbine in the bottoming Rankine cycle. The DSG solar eld ts very well to that pressure level. Besides that, this arrangement does not require of two critical elements for the DSG deployment: the thermal storage, as there is always a minimum load that is guaranteed by the fossil fuel, and the solar superheating section, because steam superheating is accomplished in the HRSG. It is important to indicate that the solar hybridization size has been set to 50 MWth while the steam turbine has not been boosted to accommodate all the nominal solar hybridization, but a smaller power (25 MWth). This value was selected taking into account the results of a previous work [9], in which an excess of solar hybridization made the steam turbine to work at part load during the whole year. The selected upgrade allows working pressures close to the nominal one along the year without high overloads.

Fig. 9. LEC values for the ISCC and the CCGT power plants as a function of the specic cost on the solar eld.

M.J. Montes et al. / Applied Energy 88 (2011) 32283238

3237

Finally, an economical analysis has been done, showing that LCE of the ISCC plant in Las Vegas is less than LCE of the ISCC plant in Almera. In the former location, this LCE value is lower than the LCE of the reference CCGT, whereas in Almera, the LCE of the ISCC is higher than the corresponding to the CCGT plant. For both cases, there are only slightly differences, as the solar hybridization size is small. The difference in the performance of the two solar elds are also patent by the calculus of the cost per additional kilowatthour produced by the coupled solar eld, that is much lower in Las Vegas than in Almera, although these values are highly dependent of the solar radiation data, for the small solar hybridization considered. In summary, hybridization between natural gas and solar power can be considered in two alternative ways: a small amount of gas complementing some deciencies in the performance of a solar plant; or a small percentage of solar energy complementing the drop in efciency of gas turbines in very hot atmospheric conditions. The former seems to be better suited for the conventional deployment of concentrated solar power with feed-in tariffs; while the latter seems a good way for advancing in the solar power learning curve, using different types of trough collectors for testing them in suitable conditions, without needing to build a full plant. Moreover, in a future scenario with very high gas prices, the boost provided by the solar unit can rise to much higher values, and a performance model could be devised with the following rationale: to reduce the power of the gas unit in the hours with maximum solar irradiation, in order to keep the total nominal power, and to take advantage of the synergism between both sources of heat. This idea is still to be developed, but it can be included in the portfolio of future work, to assess the actual potential of alternative roads and roles for concentrated solar power. Appendix A The levelized cost of the megawatthour is dened in

1 rn 1 ieff

A6

rn is the nominal escalation rate, which represents the annual change in cost and includes the effects of both the real escalation rate rr and the ination ri. The real escalation rate for the O&M and for the fuel cost, as well as the ination rate are shown in Table 7 in the paper. The values of the parameters dened above are summarized in Table A1.

References
[1] Dersch J, Geyer M, Herrmann U, Jones SA, Kelly B, Kistner R, et al. Trough integration into power plants a study on the performance and economy of integrated solar combined cycle systems. Energy 2004;29(56):94759. [2] Kelly B, Herrmann U, Hale MJ. Optimization studies for integrated solar combined cycle systems. In: Proc of the ASME int sol energy conf: the power to choose, Washington, DC; 2001. [3] SolarPaces; 2009. <http://www.solarpaces.org/>. [4] Birnbaum J, Eck M, Fichtner M, Hirsch T, Lehmann D, Zimmermann G. A direct steam generation solar thermal power plant with integrated thermal storage. In: Proc of 14th int solarPACES symp on sol therm conc technol, Las Vegas, USA; 2008. [5] Laing D, Bahl C, Bauer T, Lehmann D, Steinmann W-D. Thermal energy storage for direct steam generation. In: Proc of 15th int solarPACES symp on sol therm conc technol, Berlin, Germany; 2009. [6] Rheinlnder J, Ratzesberger R, Hahne, E. Direct solar steam-generation for combined power cycles. In: Proc of 7th int symp sol therm conc technol, Moscow, Russia; 1994. [7] Nezammahalleh H, Farhadi F, Tanhaemami M. Conceptual design and technoeconomic assessment of integrated solar combined cycle system with DSG technology. Solar Energy 2010;84:1696705. [8] Zarza E. Generacin directa de vapor con colectores solares cilindro parablicos. Proyecto direct solar steam (DISS). PhD thesis, Universidad de Sevilla, Spain; 2002. [9] Montes MJ, Rovira A, Muoz M, Martnez-Val JM. Proposal of an integrated solar combined cycle system using direct steam generation technology. In: Proc of 15th int solarPACES symp on sol therm conc technol, Berlin, Germany; 2009. [10] Montes MJ, Abnades A, Martnez-Val JM. Thermouidynamic model and comparative analysis of parabolic trough collectors using oil, water/steam, or molten salt as heat transfer uids. J Solar Energy Eng 2010;132(2):0210017. [11] Eck M, Benz N, Feldhoff F, Gilon Y, Hacker Z, Mller T, et al.. The potential of direct steam generation in parabolic troughs results of the German project diva. In: Proc of 14th int solarPACES symp on sol therm conc technol, Las Vegas, USA; 2008. [12] Zarza E, Valenzuela L, Len J, Weyers H-D, Eickhoff M, Eck M, et al. The DISS project: direct steam generation in parabolic trough systems. Operation and maintenance experience and update on project status. J Solar Energy Eng 2002;124(2):12633. [13] Rovira A, Muoz M, Snchez C, Valds M, Durn MD. Thermoeconomic optimisation of heat recovery steam generators of combined cycle gas turbine power plants considering off-design operation. Energy Convers Manage; 2010. doi:10.1016/j.enconman.2010.11.016. [14] Rovira A. Desarrollo de un modelo para la caracterizacin termoeconmica de ciclos combinados de turbinas de gas y de vapor en condiciones de carga variable. PhD thesis, Universidad Politcnica de Madrid, Spain; 2004. [15] Valds M, Rovira A, Fernndez JA. The inuence of atmospheric conditions on the performance of combined cycle gas turbine power plants. GT2006-90204. In: Proc of GT2006 ASME turbo expo: power for land, sea and air, Barcelona, Spain; 2006. [16] El-Gammal AM. An algorithm and criteria for compressor characteristics real time modelling and approximation. Trans ASME J Eng Gas Turb Power 1991;113(1):1128. [17] Stamatis A, Mathioudakis K, Papailiou KD. Adaptative simulation of gas turbine performance. Trans ASME J Eng Gas Turb Power 1990;112(2):16875. [18] Valds M, Rapn JL. Optimization of heat recovery steam generator for combined cycle gas turbine power plants. Appl Therm Eng 2000;21:114959. [19] Cotton KC. Evaluating and improving steam turbine performance. 2nd ed. New York, USA: Cotton Fact Inc; 1998. ISBN-13: 978-0963995513. [20] Montes MJ, Abnades A, Martnez-Val JM, Valds M. Solar multiple optimization for a solar-only thermal power plant, using oil as heat transfer uid in the parabolic trough collectors. Solar Energy 2009;83:216576. [21] Montes MJ, Abnades A, Martnez-Val JM. Performance of a direct steam generation solar thermal power plant for electricity production as a function of the solar multiple. Solar Energy 2009;83:67989. [22] Montes MJ. Anlisis y propuestas de sistemas solares de alta exerga que emplean agua como uido calorfero. PhD thesis, Universidad Politcnica de Madrid, Spain; 2008. [23] Trnsys 17; 2010. <http://sel.me.wisc.edu/trnsys/>. [24] Masters GM. Renewable and efcient electric power systems. USA: Wiley Interscience; 2004. ISBN: 0-471-28060-7.

LCE

LCInv LCO&M LCFuel Eannual

A1

Eannual (MW he) is the annual electricity production by the power plant; and LCInv (), LCO&M () and LCFuel () are the levelized costs of the inversion, the operation and maintenance, and the fuel, respectively. The levelized cost of the investment, LCInv (), is determined by

LCInv CRF INV

A2

where INV is the total inversion () and CRF is the capital-recovery factor, dened as

CRF

ieff 1 ieff n 1 ieff n 1

A3

ieff (%) is the effective discount rate, 10% in Table 7, and n (years) is the economic life o span period of the power plant. The levelized cost of the operation and maintenance, LCO&M () and the fuel cost, LCFuel (), are determined by Eqs. (A4) and (A5), respectively.
&M O&M LCO&M C O CELFO&M C 0 0

kO&M 1 kO&M CRF 1 kO&M

n

A4

CELFFuel C Fuel LCFuel C Fuel 0 0

kFuel 1 kFuel CRF 1 kFuel

A5

CELF is the constant-escalation levelization factor, which depends on the effective annual cost of the money, ieff, and the nominal escalation rate, rn, by means of the parameter k dened in

3238

M.J. Montes et al. / Applied Energy 88 (2011) 32283238 [30] Scottish Power; 2010. <http://www.scottishpower.com/>. [31] Electricity Supply Board; 2010. <http://www.esb.ie/main/home/index.jsp>. [32] Skone T, James R. Life cycle analysis: natural gas combined cycle (NGCC) power plant. Tech rep. DOE/NETL-403/110509. USA: Natl Energy Tech Lab; 2010. [33] Poullikkas A, Hadjipaschalis I, Kourtis G. The cost of integration of parabolic trough CSP plants in isolated Mediterranean power systems. Renew Sustain Energy Rev 2010;14:146976. [34] Bejan A, Tsatsaronis G, Moran M. Thermal design and optimization. USA: John Wiley and Sons; 1996. ISBN: 0-471-58467-3. [35] World Bank. Assessment of the world bank/GEF strategy for the market development of concentrating solar thermal power. Rep no GEF/C.25/Inf.11. Washington, USA: Glob Environ Facility; 2005.

[25] Solar Advisor Model (SAM). Natl Renew Energy Lab; 2010. <http:// www.nrel.gov/analysis/sam/>. [26] Turchi C. Parabolic trough reference plant for cost modeling with the solar advisor model (SAM). Tech rep NREL/TP-550-47605. USA: Natl Renew Energy Lab; 2010. [27] Feldhoff JF, Benitez D, Eck M, Riffelmann K-J. Economic potential of solar thermal power plants with direct steam generation compared with HTF plants. J Solar Energy Eng 2010;132. 041001-1 [9 p]. [28] Spanish Natural Gas Prices. BOE-A-2010-20004. <http://www.mityc.es/Gas/ Seccion/Precios/>. [29] Kehlhofer R, Hannemann F, Stirnimann F, Rukes B. Combined-cycle gas steam turbine power plants. 3rd ed. Tulsa, Oklahoma, USA: PennWell Corp; 2009. ISBN: 978-1-59370-160-0.