Renewable Energy xxx (2012) 1e5

Contents lists available at SciVerse ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Inuence of solar energy resource assessment uncertainty in the levelized electricity cost of concentrated solar power plants in Chile
Matas Hanel, Rodrigo Escobar*
Department of Mechanical and Metallurgical Engineering, Ponticia Universidad Catlica de Chile, Vicua Mackenna 4860, Macul, Santiago, Chile

a b s t r a c t
Keywords: Solar power Chile Parabolic trough

The deployment of renewable energy power plants is a priority of the Chilean government. A mandatory quota system requires that 5% of the electricity generated in the country must come from renewable energy sources, gradually increasing to 10% by 2024. As of 2010, solar energy has received attention only for small-scale future demonstration projects. Concentrated solar power (CSP) plants are an interesting option for the country, especially when considering the high levels of solar radiation and clearness index that are available in northern Chile. Here we present a thermal and economic analysis of CSP plants of the parabolic trough type, comparing ve different congurations including thermal energy storage and fossil fuel backup. The electricity yields are obtained from hourly simulations that consider radiation levels, solar eld, and power plant characteristics. An economic model that includes the costs of construction, operation and maintenance allows predicting the levelized electricity cost (LEC) as a function of plant conguration and location. The results indicate that the plants can produce dispatchable electricity at a cost that is competitive and inversely proportional to radiation levels. A sensitivity analysis is conducted in order to determine the inuence of solar eld area and radiation levels, and the optimal plant conguration and solar eld area are obtained as a result. 2012 Elsevier Ltd. All rights reserved.

1. Energy in Chile The main energy sources that the country utilizes are oil and its derivates, coal, and natural gas. The country does not produce any of them in signicant quantities, and it does not hold any meaningful reserves that could be explored and exploited in the future. As of 2009, Chile relies on fuel imports to meet its growing energy demand, which combined with limited fossil fuel resources make Chile a growing net importer of energy. Renewable energy sources in use by the country comprise only hydroelectricity and woodbased biomass, accounting for 24% of primary energy consumption, while non-renewable sources account for the other 76%. The electricity sector has begun to rely heavily on coal-red power plants, with up to 3 GW of capacity being planned to enter the system in the next three to ve years. Thus, Chile is not only staying dependent on imported energy, but is also switching to more expensive sources such as liqueed natural gas, and to fuels of greater environmental impacts such as coal. These two concrete actions that Chile is taking in order to secure energy supply go directly against the sustainable development denition. Therefore,

* Corresponding author. Tel.: 56 2 3545478. E-mail address: rescobar@ing.puc.cl (R. Escobar). 0960-1481/$ e see front matter 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.renene.2012.01.056

it is of critical importance for the country to achieve three primary strategic goals: rst, to provide adequate energy supplies in order to continue its economic growth; second, to ensure that imported energy is accessed through international markets to satisfy any requirements that cannot be met by indigenous production; and third, to ensure the development of indigenous energy sources at a sufcient rate such as needed for the substitution of imported energy resources in order to rapidly achieve energy security and a degree of energy independence. Starting on 2010, a new law has been passed which requires electricity distributors to provide 5% of their energy sales from renewable energy sources, at average bided prices, increasing this contribution to 10% by 2024. The government hopes to promote the use of renewable energy for electricity generation, as a result of modifying the electricity sector law, effectively removing barriers for the incorporation of renewable energy plants. The law has resulted in several wind and biomass energy power plants being planned and entered into the environmental impact assessment mechanisms. In general, Chile is thought to be abundantly endowed with renewable energy but no large scale renewable energy resource assessment has been conducted, and in particular for wind and solar. Therefore, any energy planning effort that considers these renewable sources is seriously impeded for the time being. In the case of solar energy, large scale systems are not being planned

Please cite this article in press as: Hanel M, Escobar R, Inuence of solar energy resource assessment uncertainty in the levelized electricity cost of concentrated solar power plants in Chile, Renewable Energy (2012), doi:10.1016/j.renene.2012.01.056

M. Hanel, R. Escobar / Renewable Energy xxx (2012) 1e5

or even discussed. Regarding the power generation sector, the solar thermal power plant technology is scarcely known. Solar energy development in Chile is small, mostly focusing on water heating applications for the residential sector. The total contribution of solar energy to the primary energy consumption of Chile is negligible [1]. 1.1. Parabolic trough plants It has been argued that Concentrating Solar Power (CSP) technologies are the most convenient in economics terms, with their cost projections being such that they are becoming competitive with traditional power plants even today [2,3]. The Parabolic Trough Collector (PTC) power plant is the most developed and proved, with commercial use in the US since 1985, and new plants being built in Spain, USA, Morocco, Algeria, and plans to build in other countries as well. The PTC concept is simple, and basically focuses direct solar radiation in an absorber tube, which is located at the center of an evacuated glass tube for minimizing thermal losses. The concentration factors can be as large as 90 [4], and commercial systems are available from 14 to 80 MW. The HTF is heated in the absorber element of the concentrator array to temperatures close to 400  C, and then transfers thermal energy to the power block through a heat exchanger. A fossil-red boiler can supply backup heat during non-sunlight hours, or can stabilize the steam thermodynamical state in order to maintain a constant power generation. Given sufcient solar input, the plants can operate at full rated power using solar energy alone. Most of the research on PTC plants is focused on achieving higher plant availability and efciency, by means of thermal energy storage (TES), improvement of the HTFs, better prediction of available solar radiation, and direct steam generation (DSG) techniques [2e5] (Fig. 1). The PTC plants are becoming very competitive in the actual context of high energy prices and environmental pressures [6e8]. A great effort has been devoted to the development of adequate steps for cost reduction, dening several research priorities with an estimated cost reduction potential of up to 40%: increase the scale to plants larger than 50 MW, improve concentrator structure and assembly, utilization of advanced energy storage schemes,

advanced reectors, increase HTF temperature, and reduction of parasitic loads [9,10]. Improved assessment of solar energy resource is perceived to be the rst step in developing a technical program that will lead to proper installation of CSP plants with economical feasibility. An erroneous input for direct solar radiation can lead to an erroneous size of collector eld, which can result in severe nancial difculties for the plant during operation. In this context, computational simulation is perceived as the best tool for properly estimating collector eld size during the design stages of CSP PTC plant planning [10,11]. Since all research and commercial application of CSP plants has been done in geographical regions that receive a lower average of solar energy than Chile [4,12] (Fig. 2), this seems to indicate that the country presents a distinctive advantage for CSP utilization, and that PTC plants installed in Chile could display a better performance than what has been achieved in other countries.

2. Potential in Chile for CSP plants The potential in Chile for CSP plants has not being determined on a large scale. It is possible to afrm that the Atacama Desert in the northern part of the country is one of the best regions for solar energy, based on energy density data from several sources [4,12,13]. Chilean skies in the northern part of the country exhibit the highest number of clear days of any region in the world, and as such have attracted many astronomical observatories. Consumption centers in the northern part of the country are mostly mining industries, which consume the highest share of power generation [1] with fundamentally constant demand. And the region is a desert, with ample plains and at, unused terrain. Therefore, all three basic conditions for the development of solar thermal power plants are met in the northern region of Chile: high levels of direct solar irradiance during most of the year, availability of at terrain, and short distance to consumption centers [2]. However, the rst necessary step in order to adequately perform energy planning activities and especially solar energy conversion systems design is to have precise solar energy availability databases of low uncertainty, which unfortunately is not the case in Chile. Although several data sources exist, they either lack on spatial and

Fig. 1. Schematic of a PTC CSP plant of the SEGS/LUZ systems, with thermal energy storage and auxiliary, fossil fuel red backup boiler. From [16].

Please cite this article in press as: Hanel M, Escobar R, Inuence of solar energy resource assessment uncertainty in the levelized electricity cost of concentrated solar power plants in Chile, Renewable Energy (2012), doi:10.1016/j.renene.2012.01.056

M. Hanel, R. Escobar / Renewable Energy xxx (2012) 1e5

temporal resolution, or exhibit high levels of uncertainty that makes them unsuitable for hourly simulation of solar power plants [13]. As seen in Fig. 3, which depicts several data sources for the city of Calama in Northern Chile, there are signicant differences in the data: according to different sources for measurements, satellite estimations, and weather simulation models, the maximum daily values of solar radiation can be as high as 10.5 kWh/m2day, or as low as 6.7 kWh/m2day. This same situation is repeated for locations throughout the country, where at least two or more data sources are available. Thus, the question is: What data can a designer select for dimensioning a CSP plant? And also, what is the impact on selecting one data source instead of others? In what follows, we briey present the thermodynamical model that allows us to perform hourly simulations of the CSP plant operation. This will result in predictions of the total amount of electricity generated in a year at a given location, for ve different plant congurations. Then, an economic model is also briey described, which results in a levelized electricity cost as a function of solar collector eld area, and radiation level. 3. Thermodynamic and economic model A CSP plant can be of one of several different congurations, depending on the solar eld connection to the steam cycle, and the presence of both a thermal energy storage and fossil-fuel backup systems. Here we consider ve basic congurations that include most combinations present in a CSP plant. The rst plant conguration is direct power production, where a heat transfer uid passes through the solar collectors and then through a series of heat exchanger in order to produce superheated steam, which is in turn injected to the power block. The second model uses indirect

Fig. 2. (right) e Annual mean of global horizontal radiation in Chile (from [13]; units in kWh/m2day).

Fig. 3. Several sources of solar radiation data in Northern Chile display signicant differences.

Please cite this article in press as: Hanel M, Escobar R, Inuence of solar energy resource assessment uncertainty in the levelized electricity cost of concentrated solar power plants in Chile, Renewable Energy (2012), doi:10.1016/j.renene.2012.01.056

M. Hanel, R. Escobar / Renewable Energy xxx (2012) 1e5

Fig. 4. A schematic depiction of the economic model used to predict generation costs.

multiple of 2, and multiplied by the relation between the yearly DNI in the city of Antofagasta (w1.8 MWh/m2) and Kramer Junction (w2.1 MWh/m2). The collector aperture in meters is the one for the LS-3 structure. The hourly DNI and temperature are obtained from databases [14]. For the thermal model the electricity production is computed only for the rst year; it is assumed that following years behave in the same manner. The economic model includes construction and operational costs as indicated in Fig. 4. The result of the economic model is the levelized energy cost, or LEC, which is the cost of energy that makes the present value of the project zero. If the price of electricity is higher than the LEC, the project is feasible. In other cases an economic incentive from state should apply to make it interesting to investors. 4. Results The simulations were done for Antofagasta (1800 kWh/m2 year), Calama (3200 kWh/m2 year), and Santiago (1400 kWh/m2 year), three cities in Chile that offer a range of annual radiation levels, and which are located in the coast, desert, and central agricultural regions. Fig. 5 shows the hourly electric energy produced with solar energy for the 6th to the 11th of January at the three cities and for different plant congurations. Both TES congurations atten the power production curve and translate energy to hours after sunset. Fig. 6 shows the monthly average of hourly power production, with bars indicating the best and worst months. Not surprisingly, Calama is the best location for installing a solar trough plant of any conguration. All the plant congurations with direct energy production show a better average with a large range between maximum and minimum. A TES system modulates the energy production, making the plants produce energy with a smaller range between max and min. Finally, direct TES has a better performance than indirect TES, since it doesnt uses a heat exchanger between the storage system and power block. Fig. 6 also displays the computed LECs for each location, this time adding the city of Copiap (2500 kWh/m2 year). It can be observed that the lower costs are associated to the Direct and Indirect TES congurations using fossil backup. The highest cost corresponds to the direct production strategy, which, as a solaronly mode, is limited by the availability of sun hours and therefore has the least energy production for a similar investment as the other congurations. As mentioned, the sources for solar radiation in Chile display a wide range of different data, which in occasions can reach a 40% uncertainty or more as shown in Fig. 3. Thus, it is very difcult to

Fig. 5. Hourly simulation results for CSP plants in Antofagasta during six days in January.

thermal energy storage and adds a fossil backup. In this conguration the heat transfer uid in the solar collector circuit heats a secondary uid which is used for thermal storage, which in turn delivers the thermal energy to the power block. The fossil backup can be used in order to maintain continuous power production when solar radiation is temporarily unavailable or at night. A variation of this conguration lacks the fossil backup. The fourth conguration id the direct storage with fossil backup, in which the uid in the solar collector circuit is the same uid used for thermal storage. This uid delivers its energy to the power block via a series of heat exchangers. Finally another direct storage conguration lacks the fossil fuel backup. Details of the thermodynamic and economic models can be obtained from [15], which is readily available by email upon request to the authors, and in [16,17]. For the solar eld model the following inputs are necessary: the aperture area measured in square meters, which is set to 1,400,000 m2, computed based on the SEGS area that have in average 6150 m2 per installed MW of aperture, with a solar

Fig. 6. Average daily energy production for different plant congurations at three locations (left) and LEC for CSP plants at different locations in Chile (right).

Please cite this article in press as: Hanel M, Escobar R, Inuence of solar energy resource assessment uncertainty in the levelized electricity cost of concentrated solar power plants in Chile, Renewable Energy (2012), doi:10.1016/j.renene.2012.01.056

M. Hanel, R. Escobar / Renewable Energy xxx (2012) 1e5

Fig. 7. LEC and energy production considering uncertainty in the solar radiation data.

uncertainty and thus the nancial risk of a CSP project in the country. In what was presented, ve different congurations of CSP plants were analyzed for the local conditions in Chile, and an economic model implemented in order to predict the LEC and optimal plant size. It was seen that the LEC is inversely proportional to available radiation, and that the best locations for CSP are in the Atacama Desert. However, both the LEC and optimal solar collector eld for minimum LEC display a signicant dependence on solar radiation, which is larger for small-scale plants. Considering that the government is actively promoting the deployment of a 5e10 MW plant, it is concluded that a strategy that minimizes uncertainty in the LEC could gain an advantage by designing larger plants.

argue that one has chosen the correct data until a good solar energy resource assessment of validated, quality data has been developed for the country. Fig. 7 shows the LEC and yearly Energy production in Copiap, as a function of the plant conguration, solar radiation, and solar eld area. It can be seen that, rst, the optimal solar eld area depends on the radiation level for minimum LEC. Second, the LEC variation as a function of solar radiation is signicant, and can result in a serious nancial risk to investors if the uncertainty is not taken into account. Also, considering terrain availability constraints, the uncertainty in solar radiation data can also impact site selection if one location is deemed to be too small when the radiation is underestimated.

References
[1] Balance Nacional de Energa 2008 (national energy balance 2008), downloadable from www.cne.cl [2] Price H, Lupfert E, Kearney D, Zarza E, Cohen G, Gee R, et al. Advances in parabolic trough solar power technology. Journal of Solar Energy Engineering 2002;124:109e25. [3] Sargent and Lundy consulting group. Assessment of parabolic trough and power tower solar technology cost and performance forecasts. NREL/SR-55035060; 2003. [4] Dufe JA, Beckman AW. Thermal engineering of thermal processes. 3rd Ed. New York, USA: Wiley & Sons, INC; 2006. [5] Zarza Rojas, Gonzlez Caballero, Rueda INDITEP. The rst pre-commercial DSG solar power plant. Solar Energy 2006;80:1270e6. [6] Trieb F, Langniss O, Klaiss H. Solar electricity generation - A comparative view of technologies, costs and enviromental impact. Solar Energy 1997;59:89e99. [7] Kalogirou Lloyd, Ward. Modelling, optimization and performance evaluation of a parabolic trough solar collector steam generation system. Solar Energy 1997;60:49e59. [8] Thomas A. Solar steam generating systems using parabolic trough concentrators. Energy Conversion Management 1996;37:215e45. [9] Mills D. Advances in solar thermal electricity technology. Solar Energy 2004; 76:19e31. [10] Pitz-Paal Robert, Dersch Jrgen, Milow Barbara, Tllez Felix, Ferriere Alain, Langnickel Ulrich, Steinfeld Aldo, Karni Jacob, Zarza Eduardo, Popel Oleg. Development Steps for Parabolic Trough Solar Power Technologies with Maximum Impact on Cost Reduction. Journal of Solar Energy Engineering 2007;129:371e7. doi:10.1115/1.2769697. [11] Quashning V, Kistner R, Ortmanss W. Inlfuence of direct normal irradiance variation on the optimal parabolic trough eld size: a problem solved with technical and economical simulations. Journal of solar energy engineering 2002;124:160e4. [12] Goswami Y, Kreith F, Kreider F. Introduction to Solar Energy Engineering. 1st. ed. USA: Taylor & Franciss; 2004. [13] Sarmiento P. Energa Solar: Aplicaciones e Ingeniera. 3a Ed. Ediciones Universitarias de Valparaso; 1995. [14] Ortega A, Escobar R, Colle S, Luna de Abreu S. The state of solar energy resource assessment in Chile. Renewable Energy 2010;35(11):2514e24. [15] Hanel, M. Levelized Electricity Cost of Concentrated Solar Power Plants in Chile. MsC thesis, Ponticia Universidad catlica de Chile (2010) [16] Patnode, A. (2006). Simulation and Performance Evaluation of Parabolic Trough Solar Power Plants. MSc thesis. University of Wisconsin-Madison. [17] IEA.. Guidelines for the economic analysis of renewable energy technology applications; 1991.

5. Conclusions Chile is not a fossil energy producer; the country satises its internal consumption based mainly on imported fuels. This makes the country dependent on international markets in order to secure its energy needs, which makes it vulnerable against supply disruptions and price volatility. The Chilean government is actively seeking to promote the deployment of renewable energy plants by a mandatory quota system and also nancial incentives. The successful deployment of renewable energy in the country will depend on providing an adequate investment environment, which in turn is affected by the availability and quality of the renewable energy resources. In this respect, a proper assessment of the solar energy resource has not yet been performed in Chile, which results in some regions of the country where there is simply no data available and others where plenty of data exist, although with wide dispersion and often even contradictory. This uncertainty in the data introduces a large uncertainty in the nal cost of the solar electricity, and it potentially can have a signicant impact on the nancial side of an operation. Therefore, it is necessary for Chile to improve the quality of available data, and also to derive means of at least reducing the data

Please cite this article in press as: Hanel M, Escobar R, Inuence of solar energy resource assessment uncertainty in the levelized electricity cost of concentrated solar power plants in Chile, Renewable Energy (2012), doi:10.1016/j.renene.2012.01.056