THERMOSOLAR:EOS PLANTAS TERMOSOLARES

Manuel Silva1, José A. Vélez2, José M. Barea1, Alejandro Barón1, Javier López2, Concepción Ruiz1,

1 AICIA / Grupo de Termodinámica y Energías Renovables. Address: Avda. de los descubrimientos s/n,41092 Sevilla (Spain). Phone: (+34) 954487233. Fax: (+34) 954 487233, e-mail: silva@esi.us.es

2 Aguas y Estructuras S.A. (Ayesa), Avenida Marie Curie, No. 2, Parque Tecnológico de la Cartuja, 41092, Seville, Spain,Tel: +34 954467069, Fax: +34 954462491

A sensitivity analysis of 50 MW parabolic trough solar thermal power plants has been carried out with the help of EOS, a tool for detailed simulation. The parameters analysed have been: latitude, yearly DNI, thermal storage size and solar field size. A total of more than 630 simulations have been performed. The analysis of the results permit to estimate the effect of the parameters selected on the plant technical and economical performance. The important effect of latitude is clearly displayed by these results. Another important conclusion from the result is that the optimum plant sizes from the efficiency and electricity cost (LEC) points of view are not coincident. This optimum is displaced to the right side (larger size) for the LEC.

Keywords: Solar thermal power, parabolic trough, sensitivity analysis, thermal storage.

Spanish renewable energy market has represented a good opportunity for investors during recent years due to the favorable legal situation created, in particular for those with interests in solar thermal sector. As a result of the previously mentioned circumstances, a significant number of 50 MWe solar thermal electric generating plants have been planned in Spain, which has been translated, in some cases, in the beginning of several erection processes. The main characteristics of this kind of plants can be summarized in the adoption of parabolic troughs as solar concentrating system and, in a significant number of cases, the availability of a thermal storage system, mainly based on the use of mixtures of molten nitrate salts. Aspects related with plant location and sizing of the solar field and thermal storage systems are part of the main decisions that designers need to make in order to improve the economical profile of the associated investment. Because of the high number of physical variables involved in this optimization process the most common approach to solve this problem is represented by the use of modeling software. These tools are oriented to design and simulate operation of solar thermal electric generating plants during certain periods, typically one year. Results from different sensitivity analyses, which are based on different physical magnitudes, are presented in this paper for 50 MWe parabolic trough plants with different thermal storage capacities.

The purpose of EOS software is to carry out global simulations of solar thermal electric generating plants. Although they also could be used to predict plant operation once erected, mentioned simulations are particularly oriented to generate data required during plant development and design. In other words, the developed simulator enables plant function characterization and main equipment sizing.

Parabolic trough power plants with thermal energy storage system can be simulated with EOS simulator. Two-tank molten salt storage concept has been selected as thermal energy storage system technology. Previously mentioned plant operation characterization enables EOS simulator to be used in parametric studies depending on different variables, such as solar field aperture or thermal storage capacity. These studies can be integrated in optimization processes oriented to improve solar power plant financial profile, although work presented here has a more general purpose.

The simulator is the result of the continuous collaboration established since 2006 between AYESA and Seville University the Group of Thermodynamics and Renewable Energies of the University of Seville. The association of both parts resulted in the analysis of solar thermal power plants from different points of view, which significantly improved the quality of the results, considering each engineering discipline involved in simulation process.

In general terms, EOS simulator main function is to calculate instantaneous power plant output by means of plant energy and mass balances resolution. This task requires, apart from other details, modeling the complex operation of parabolic trough solar thermal power plants with thermal energy storage, which implies an important effort due to the changing nature of solar resource. EOS philosophy is based on the fragmentation of the simulation in autonomous modules which represent different plant functional parts, as shown in Figure 1. The mentioned modularity permits simulator units to be improved independently, without changing remaining modules.

A numeric language oriented to matrixes and vectors (MATLAB) has been applied to develop EOS simulator, which provides simulations with a relevant resource saving in spite its relative high computation time. Due to simulator absolute flexibility, almost all variables that take part in plant simulation are accessible to the user, which allows simulated operation conditions to be very similar to the actually characteristics associated to the project. The previous feature is also applicable to results generated by EOS simulator, since these can be modified to fit user necessities. Simulations of periods shorter than one year are interesting to analyze some variables evolutions.

Both meteorological and design data are required by EOS to simulate solar thermal power plants. Concerning meteorological information, which has an extremely high influence on result quality, temporal evolution of different variables (direct normal irradiance, ambient temperature, wind speed…) has to be provided to the simulator. Simulation step automatically suits to available meteorological data. On the other hand, the second set of information comprises different plant systems design parameters, that enable to simulate the characteristics of the plant.

Every simulation is preceded by a design phase, necessary to define plant systems features. After characterizing the plant layout, simulations are carried out as a function of meteorological data evolution.

For the purpose of this work, state-of-the-art parabolic trough solar thermal power plants have been considered. The main blocks of the plant are

· The solar field, H-type layout, with 4 subfields, all of them with the same number of loops. Each loop consists of four 150 m length Solar Collector Assemblies (SCA) connected in series. Heat transfer fluid (HTF) is Therminol VP1 [1]. Receiver tubes have been modeled according to data existing in the literature.

· The thermal storage system, consisting of two tanks (“hot” and “cold”) using molten nitrate mixtures as storage medium. The storage system is connected to the solar field by means of the oilto-salt heat exchangers.

· The power block, including the steam turbine and condenser, is connected to the solar field by the steam generator.

· The auxiliary gas burner. This burner is designed to provide up to 30% of the design power of the turbine, and is placed on the HTF side (this is, it is used to heat up the HTF rather than directly generating steam).

The HTF enters the solar field at 293 ºC and is collected at 393 ºC at the main collectors. When operating from the thermal storage –discharging the hot tank- the HTF is heated up to 379 ºC.

An operating strategy has been devised to optimize the use of natural gas –which is limited to generate a maximum of 15% of the plant yearly electricity output- and to minimize the number of plant startups and shutdowns. For the purpose of this study, a total of 14 outage days –of which 10 programmed- has been considered.

Nine meteorological data sets have been generated, corresponding to DNI values of 1800, 2000 and 2200 kWh/m2 and 3 latitudes (37º, 39º, 41º) representative of the southern, central and northern regions in Spain. These data sets contain hourly values1 of DNI and other meteorological variables generated with Meteonorm [2].

Equipment costs have been obtained, when possible, directly from the providers or from project costs. When this has not been possible, costs referenced in the literature have been used.

These costs have been estimated from actual prices when possible, or estimated as accurately as possible

1 EOS full capabilities are only exploited when using 5 or 10-minute time steps, but the characteristic of this work, which is mainly of qualitative nature, made it advisable to use these synthetically generated data sets. From other data. It is to be noted that the results presented in this work should be interpreted in a qualitative way, since the spectacular take-off of the solar thermal power sector in Spain is very likely to be producing significant distortions on the actual costs of important systems or supplies.

For this purpose, 630 simulations have been performed (3 latitudes, 3 DNI sets / latitude, 7 storage capacities,10 solar field sizes (apertures)/storage capacity.

EOS provides very detailed results of the plant performance simulation. In this work, we will focus on three of them: net electricity production, global net efficiency and levelized electricity cost. The main results of the sensitivity analysis are presented in the following.

Annual electricity generation increases with DNI, storage capacity and mirror area for all the latitudes considered. This is illustrated in figure 2 for latitude 37º and 2000 kWh/m2 DNI. Similar trends can be observed for all latitudes considered and DNI levels.

A more detailed analysis of the results presented in the figure 2 shows that the curves corresponding to 2 hours and 4 hours of storage capacity are more separate than the rest. This is mainly due to the different operation strategy implemented for plants with 2 hours storage or less.

For a given storage capacity, electricity generation increases with solar field size, but less than proportionally, due to saturation effects (the storage tanks cannot admit the thermal energy produced in excess by the solar field).

The curves shown in figure 2 suggest that, for each storage capacity, there is a minimum value of the solar field size under which the electricity generation is very close to the electricity generated by the immediately lower storage capacity. It is logical to think that for smaller field sizes, the electricity generated is virtually the same regardless of the storage.

Figure 2. Electricity generation for different storage capacities as a function of solar field size (Lat 37º)

Figure 3. Electricity generation for different latitudes as a function of solar field size

Regarding the dependence on the latitude, figure 3 shows the results for 6 h storage capacity. The effect of the unfavorable incidence angle for high latitudes becomes evident: a plant of 140 loops at 37º latitude, with 2000 kWh/m2 DNI, will generate roughly the same electricity as the same plant at 41º with 2200 kWh/m2 DNI2. For other storage capacities the results are similar, although the effect of latitude increases with storage size.

Global Net Efficiency is defined as the relation between the net electricity generated and the product of DNI times mirror surface and the energy supplied by natural gas.

Figure 4 suggests that there is an optimum field size for every storage size from the point of view of efficiency, and that the optimum efficiency increases with storage size. Similar trends can be observed for all latitudes and yearly DNI values

2 It is reasonable to wonder if, above a certain latitude, E-W oriented collectors (N-S tracking) could be an interesting option.

The effect of DNI and latitude are displayed in figure 5 (no thermal storage). For a given latitude and solar field size, efficiency always increases with DNI. The effect of latitude is illustrated, for example, by the fact that the efficiency of a plant with 80 loops at 37º latitude, with 2000 kWh/m2 yearly DNI is slightly larger than the efficiency of a similar plant placed at 39º latitude with 10% more DNI (2200 kWh/m2). The larger the solar field, the larger the difference. Again, a similar effect is observed for all the storage capacities analyzed.

Figure 4. Global net efficiency for different storage sizes as a function of solar field size

Figure 5. Global net efficiency for different latitudes as a function of solar field size

Levelized cost is the constant annual cost that is equivalent on a present value basis to the actual annual costs, which are themselves variable. The Levelized Electricity Cost (LEC) is a cost comparison methodology proposed by the International Energy Agency [3] that considers the total electrical energy that the power plant will produce and the total investment, financial and O&M costs during its lifetime to calculate the cost of the kWh generated by a certain power plant.

LEC calculations presented here must be considered with caution, since we have not included in the calculations some costs which are difficult to evaluate (for example, the EPC cost –most of the projects in Spain are being promoted under the Project Finance scheme) and other costs are somehow volatile, given the likely distortion of the prices of some of the supplies.

The results of the LEC calculations for 37º latitude and 2000 kWh/m2 yearly DNI are represented in figure 6. The minimum LEC is obtained with a plant of roughly 300000 m2 mirror area and no storage. The larger the storage, the higher the LEC of the optimum size plant, although the increase is relatively small. Similar results are obtained for other yearly DNI values and latitudes, the effect of storage being more negative for higher DNI values.

A more detailed analysis of the results show two clear trends: for 0, 1 and 2 hours of storage the LEC increases rapidly, while for larger storage capacities the optimum LEC remains almost constant, with a slight increase. The reason may be again the effect of the operating strategy for small storage capacity.

For low DNI values, and high latitudes (figure 7) the optimum LEC seems to be fairly independent of the storage capacity, with the exception of the 1 and 2 equivalent hours.

Finally, the comparison of figures 4 and 6 show that the optimum plant sizes differ when we consider the efficiency or the electricity cost as optimization criteria. For example, the optimum efficiency for a 50 MW plant without storage is achieved with a solar field of approx. 80 loops (250000 m2) while the optimum LEC is achieved with a field of approx. 310000 m2. This fact highlights the importance of the correct choice of the variable to be optimized.

Figure 6. Levelized Electricity Cost for plants with different thermal storage capacities at 37º latitude and 2000 kWh/m2 yearly DNI as a function of solar field size

Figure 7. Levelized Electricity Cost for plants with different thermal storage capacities at 41º latitude and 1800 kWh/m2 yearly DNI as a function of solar field size

The sensitivity analysis performed with EOS permits to study the effect of the different variables considered (latitude, yearly DNI, thermal storage capacity and solar field size) on the technical and economical performance of 50 MW parabolic trough power plants in Spain. Although the results must be interpreted in a qualitative way, the important effect of the plant site latitude becomes evident in all analyzed aspects. The importance of the selection of the variable to be optimized is also emphasized by the fact that the resulting optimum sizes when we select different optimization variables (for example, efficiency and LEC) are considerably different.

DOWNLOAD THE COURSE OF PRESSURE EQUIPMENT REGULATIONS

A few months ago completely changed the old REGULATION OF PRESSURE, THROUGH THE TRANSITION TO NEW REP.

SGS and RENOVETEC put to you a free course, practical, and clear summary of the main aspects of the new regulation and Complementary ITC Technical Instruction EP-1 that affects the central thermal

Sensitivity Analisis of csp plants in Spain using EOS