An Analysis of the Effect of Molten Salt Thermal Storage on Parabolic Trough Concentrated Solar Power Plant Efficiency

Christopher Hickin1, Henry Li2 and Sharnan Kemp3

Faculty of Science, University of Technology Sydney, P.O. Box 123, Australia

1 E-Mail: Christopher.J.Hickin@student.uts.edu.au

2 E-Mail: Henry.Li-1@student.uts.edu.au

3 E-Mail: Sharnan.Kemp@student.uts.edu.au

Received: 25/5/2015 / Accepted: 29/5/2015 / 1/6/2015:

Abstract: In the development of renewable energy sources, there has been a trend toward increasing and stabilising the power output of Concentrated Solar Power Plants (CSPPs) during times of reduced solar resource through the use of Thermal Energy Storage Devices (TESDs). This study investigates whether the use of a molten salt TESD decreases the efficiency of a parabolic trough CSPP due to additional system energy losses despite prolonging the operational time of the CSPP. A theoretical analysis of a simplified CSPP was made to determine if a TESD would impact the efficiency of the CSPP. This was followed up by a survey of currently active parabolic trough CSPPs both with and without molten salt TESDs. The theoretical analysis illustrated that a TESD would have no effect on the efficiency of a CSPP. However, the survey revealed that the use of a TESD improved the efficiency of a CSPP. The results of the study don’t support the theoretical analysis or the hypothesis suggesting that a property has been overlooked. This property is most likely to be that generators tend to operate best within a certain temperature range, and in a CSPP the optimum temperature range cannot be maintained. This results in a generator being selected capable of operating for the longest period with the lowest amount of excess solar energy. When a TESD is implemented, the excess solar energy is stored for later use, prolonging the generator’s running time and increasing the useable energy. The realisation of the ability of a TESD to increase the efficiency of a CSPP as well as extending its operating time shows a promising area of development in CSPP technology and increasing its application in electricity generation.

Keywords: Concentrated Solar Power; Thermal Energy Storage; Molten Salt; Parabolic Trough; Efficiency

Copyright: © 2015 by the authors. This article is distributed under the terms and conditions of the Creative Commons Attribution license (https://creativecommons.org/licenses/by/4.0/).

DOI: http://dx.doi.org/10.5130/pamr.v2i0.1392

1. Introduction

Throughout the last century, fossil fuels have been a major part of the world energy market. However, as the world’s energy demand has increased [1], so too has its burden on the Earth, leading to a need to implement renewable and environmentally friendly forms of power generation. Concentrated Solar Power (CSP) is one such power generation method being utilised and developed. Power is generated through the reflection of sunlight onto a receiver, the heat from this is then transferred to steam to power a turbine. While these setups may be attractive from an environmental point of view, there are key detractors from using this as a primary source of power generation, two of the larger being power generation during times of low light, and low efficiency. A method of combating the former utilises Thermal Energy Storage Devices (TESD) to store energy for these downtimes. This study seeks to investigate the impact of TESDs on the efficiency of the Concentrated Solar Power Plants (CSPPs) they are attached to.

139201.jpg

Figure 1. Example of a Parabolic Trough CSPP Setup [2]

CSP technology, having been first put into practice in the 70s, has flourished recently, with dozens of new CSP projects coming to completion in the new millennium [3]. Many improvements in design have occurred since the first plant was constructed. These include better solar tracking, usage of better materials in construction, new materials to act as better thermal transfer fluids, and the focus of this study, TESDs [4]. These upgrades make the technology more viable through a variety of means; reducing both the initial and ongoing costs of the plant, increasing the efficiency of the electricity generation, and partially overcoming the major hurdle stopping CSP from being a major energy provided: low light power generation.

There are four main types of CSP plants that are currently in use. These are Parabolic Troughs, Linear Fresnel Reflectors, Sterling Dish, and Power Tower systems. While these all have different setups, their primary mechanism is the same; reflection of sunlight onto a central receiver. Of these systems, the Sterling dish is the most efficient, reaching an efficiency of 27% [5]. The Linear Fresnel Reflector, parabolic trough, and tower system all have efficiencies of close to 13% [6]. It is worth noting photovoltaics in this comparison, as despite being a different technology and having a different mechanism of action, they also generate power from sunlight. While having reached peaks of 45% efficiency in laboratory testing environments, it is more likely to see an efficiency of 18% when used commercially [7]. While these results may fall short of electricity generation technologies such as coal and nuclear, they show that a viable and clean alternative energy source for electricity production is available.

The CSP technologies this study will be focusing on are the parabolic trough and the molten salt TESD. The parabolic trough is the most prevalent of the CSPPs as it was the first design. It consists of a series of curved, rectangular reflectors that reflect the light onto a central line containing the transfer fluid that is suspended above the trough. This variation of CSP runs at between 200 and 300 degrees. Once the heat is collected, it is either transferred to a TESD or used to generate steam for the turbine. Though there are varying setups for the TESD, they all consist of a hot and cold reservoir, heat exchanger, and a pump. When heat from an external system is available, the molten salt is pumped from the cold reservoir through the heat exchanger and stored in the hot reservoir. When the external system requires heat the process may be reversed, pumping the heated salt back through the heat exchanger and stored back in the cold reservoir [8].

139202.jpg

Figure 2. Schematic diagram of a CSPP with a TESD [9]

The hypothesis of this paper proposes that by looking at the efficiency of a CSP plant with a molten salt TESD attached, it would be expected that if the theoretical net daily heat of a thermal storage device is zero, then the annual efficiency of a CSP plant without a TESD would be greater than one that does due to energy losses associated with the efficiency of thermal storage devices and heat exchangers. It would follow that if the CSP plants with Thermal Storage (TS) have a lower efficiency than their TS lacking counterparts, that it would be more efficient to store energy electrically than thermally. Knowing this can allow manufacturers of CSP plants to more accurately decide on the benefits of either having more efficiency in their plant compared to having longer uptime with solar-based power generation.

2. Methods

In order to test the validity of the proposed hypothesis a meta-study was performed, analysing the efficiency of CSPPs that utilised a TESD against CSPPs that did not, and a theoretical analysis of the two systems was undertaken. The theoretical analysis was performed by applying the first law of thermodynamics with the assumptions that:

In order to limit variables and minimise error, the study was restricted to parabolic trough CSPPs that are currently operational that use a steam Rankine cycle to generate power. When gathering data for the TESD equipped CSPPs the TESD was confined to a two-tank molten salt system. The required data for the analysis was collected from the National Renewable Energy Laboratory databases.

The analysis also included gathering data from currently active parabolic trough CSPPs using molten salt TESD (as seen in Table 3 and 4). The main accumulation of data included:

By applying the accumulated data, the annual solar resource captured by the CSPPs was calculated. This was done by multiplying the annual solar resource by the aperture area, and the annual efficiency (η) of each CSPP was also calculated. This was done by:

139203.gif

This process was repeated for operational parabolic trough CSPPs that does not utilize a TESD. Calculations for all CSPPs efficiencies were then averaged. A comparison was made between the averaged efficiencies from each category to determine if a TESD would affect the annual efficiency of a CSPP.

3. Results and Discussion

3.1. Theoretical Analysis

In order to simplify the analysis a CSPP can be separated into three closed systems; the solar receiver (SR), generator (steam Rankine cycle) and TESD. For this analysis, we shall assume that the energy losses are negligible, the steam Rankine cycle of the generator remains the same for both setups, and the internal energy of the solar receiver remains constant.

139204.jpg

Figure 3. Simplified Schematic of the Solar Receiver Thermal Circuit

The key system to focus on is the solar receiver, and specifically the heat transferred to the generator as this is what will determine the output energy. An expansion of the following derivation of the following equations can be found in the appendix. Applying the first law of thermodynamics to this system gives us the equation:

139205.gif

This equation states that the heat from the solar resource being added to the system is equal to the heat leaving the system into the thermal storage and the generator. It is important to note that the heat from the TEDS can be either negative or positive depending on whether it is charging from the system or discharging into the system. When the TESD is discharging into the system the equation becomes:

139206.gif

During the period of time when the TESD is recharging and removing heat from the system the equation is written as:

139207.gif

For this analysis, it will also be assumed that the TESD undergoes a full recharge-discharge cycle in a day in order to prolong the effective operation time of the CSPP as would be done in an actual plant. This assumption then gives the daily net TESD heat as zero, which in turn yields an annual heat as zero. This result is equivalent to that of a CSPP which does not use a TESD.

139208.gif

With the assumption that the steam Rankine cycle remains the same in each scenario, the annual efficiency of each of the CSPPs would be equal. In a real world situation, it is likely that the efficiency of a TESD CSPP would be lower than its non-TESD counterpart. This lower efficiency may be due to the energy losses associated with storing and transferring the thermal energy with the additional processes introduced by the addition of a TESD.

3.1. CSPP Data

Table 1. Parabolic Trough CSPPs Efficiency using a Molten Salt.
Plant Aperture Area (m2) Solar Resource (kWh/m2 Yr) Available Power (MWh/Yr) Electricity Generation (MWh/yr) Efficiency (%)
Andasol-1 (AS-1) 510120 2136 1089616.32 158000 14.50
Andasol-2 (AS-2) 510120 2136 1089616.32 158000 14.50
Andasol-3 (AS-3) 510120 2200 1122264.00 175000 15.59
Archimede 31860 1936 61680.96 9200 14.92
Arcosol 50 510120 2097 1069721.64 175000 16.36
ASE Demo Plant 3398 1527 5188.75 275 5.30
Aste 1A 510120 2019 1029932.28 170000 16.51
Aster 1B 510120 2019 1029932.28 170000 16.51
Astexol II 510120 2052 1046766.24 170000 16.24
Extresol-1 510120 2168 1105940.16 158000 14.29
Extresol-2 510120 2168 1105940.16 158000 14.29
Extresol-3 510120 2168 1105940.16 158000 14.29
La Africana 550000 1950 1072500.00 170000 15.85
Manchasol-1 510120 2208 1126344.96 158000 14.03
Manchasol-2 510120 2208 1126344.96 158000 14.03
Termesol 50 510120 2097 1069721.64 175000 16.36
Average 14.60

 

Table 2. Parabolic Trough CSPPs Efficiency without a TESD.
Plant Aperture Area (m2) Solar Resource (kWh/m2 Yr) Available Power (MWh/Yr) Electricity Generation (MWh/yr) Efficiency (%)
Helios 1 300000 2217 665100.00 97000 14.58
Helios II 300000 2217 665100.00 97000 14.58
Ibersol Ciudad Real ( Puertollano) 287760 2061 593073.36 103000 17.37
ISCC Kuraymat 130800 2431 317974.80 34000 10.69
La Risca 352854 2174 767104.60 105200 13.71
Lebirja 1 412020 1993 821155.86 120000 14.61
Majadas 1 372240 2142 797338.08 104500 13.11
Palma del Río I 372240 2291 852801.84 114500 13.43
Palma del Río II 372240 2291 852801.84 114500 13.43
Saguaro Power Plant 100000 2636 263600.00 2000 0.76
Shams 1 627840 1934 1214242.56 210000 17.29
Average 13.05

139209.jpg

Figure 4. CSPP without TESD Efficiency Distribution with Normal Distribution Curve Overlay

139210.jpg

Figure 5. CSPP with TESD Efficiency Distribution with Normal Distribution Curve Overlay

4. Conclusions

The distribution of the CSPPs with the TESD presents a possible bimodal distribution. This type of distribution is usually an outcome when the results of two different processes are combined in the same set of data. When comparing the distribution of the systems using TESDs with the expected normal distribution excluding the outlier, the potential for a bimodal distribution becomes more distinct. However as the available data set is limited this cannot be confirmed. Both distribution graphs also highlight a low-end outlier by a significant amount. This again suggests that there are different processes being used though the data gathered remains insufficient to support this. When the outliers are excluded from the normal distribution curve it more closely matches the accumulated results.

The hypothesis that the efficiency of a CSPP with a TESD would be less than one without it is proven to be false from the gathered results. We found that, when compared, the average efficiency of the CSPPs with a TESD was greater than those without one. These findings would infer that the addition of a TESD in a CSPP can use more of the available solar resource. This ability may be due to the specifications of the turbine in the steam Rankine cycle used to convert the thermal energy to electrical energy.

In a CSPP, the heat transferred to the working fluid varies. This variation does not allow the turbine to operate at its maximum capacity resulting in a loss of power output. Because of this a turbine needs to be selected that is capable of working for the longest period with the lowest amount of excess solar energy. In a CSPP that uses a TESD, the excess solar energy can be stored for later use. By doing this, the generator can run at its maximum capacity for longer, and the amount of useable energy is increased.

As the difference in the efficiencies is relatively small more data must be obtained to solidify this result. To do this, a wider variety of CSPP configurations may be sampled as well as the specifications of the power block used. The calculated averages may also be misrepresentative of the actual efficiency difference due to the limited sample size available for analysis.

Following from this analysis other varieties of CSPP should be analysed along with the effects of differing power blocks and TESDs. This wider range of data should provide more conclusive results in order to distinguish more clearly the reason behind and future use of TESDs in CSPPs.

Acknowledgments

We would like to express our sincere thanks to Dr Jurgen Schulte for sharing expertise and valuable guidance extended to us. We also place on record our gratitude to the reviewers for their helpful comments on the manuscript.

References and Notes

1. U.S. Energy Information Administration. Available online: http://www.eia.gov/cfapps/ipdbproject/IEDIndex3.cfm?tid=44&pid=44&aid=2

2. Parsons, D. Solar Thermal System at the Jeffco Jail, 1996. Available online: http://images.nrel.gov/viewphoto.php?imageId=6308924 (accessed on 3 May 2015)

3. Py, X. Concentrated Solar Power: Current Technologies, Major Innovative Issues and Applicability to West African Countries. Renewable and Sustainable Energy Reviews 2013, 8, pp. 306-315.

4. Mills, D. Advances in Solar Thermal Electricity Technology. Solar Energy 2004, 76, pp. 19-31.

5. Liao, T. Lin, J. Optimum performance characteristics of a solar-drive Stirling heat engine system. Energy Conversion and Management 2015, 9, pp. 20-25.

6. Cau, G. Cocco, D. Comparison of Medium-size Concentrating Solar Power Plants based on Parabolic Trough and Linear Fresnel Collectors. Energy Procedia 2014, 45, pp. 101-110.

7. Hao, H. Shi, Wei. Chen, J. Lu, M. Mass production of Si quantum dots for commercial c-Si solar cell efficiency improvement. Materials Letters 2014, 133, pp. 80-82.

8. Jonemann, M. Advanced Thermal Storage System with Novel Molten Salt. NREL, 2013.

9. National Renewable Energy Laboratory. Concentrating Solar Power. Fact Sheet

10. NREL. Available online: http://www.nrel.gov (accessed on 3 May 2015).

11. Menendez, R.P.; Martinez, J.A. A Novel Modeling of Molten-Salt Heat Storage Systems in Thermal Solar Power Plants. Energies 2014, 7, pp.6721-6740

12. Wagner, M.J.; Blair, A. A Detailed Physical Trough Model for NREL’s Solar Advisor Model, SolarPACES 2010, Perpignan, France, September 21-24 2010; National Renewable Energy Laboratory: Golden, United States of America.

13. Wagner, M.J.; Zhu, G. A Generic CSP Performance Model for NREL’s System Advisor Model, SolarPACES 2011, Granada, Spain, September 21-24 2010; National Renewable Energy Laboratory: Golden, United States of America.

13. National Renewable Energy Laboratory. Electricity, Resources, & Building Systems Integration: Concentrating Solar Power, 2010.

14. International Energy Statistics. Available online: http://www.eia.gov (accessed on 3 May 2015)

Appendix - Theoretical Calculations

Symbol Definitions
ΔU Total System Internal Energy
ΔQ Total System Heat
ΔW Total System Work
QSR Solar Receiver Heat
QSR TESD Heat
QSR Generator Heat

Looking only at the solar receiver system

The first law of thermodynamics:

139211.gif

Expanding the first law to the solar receiver and assuming no energy is stored in the system, and no work is done by or to the system gives:

139205.gif

Having state one as the period of time when the TESD is discharging into the system the equation can then be written as:

139206.gif

State 2 is the period when the TESD is charging which yields:

139207.gif

If state 3 is taken over the period of a day and it is assumed that the TESD goes through a recharge-discharge cycle in a day then QTS,3 = 0. With this assumption the equation then becomes:

139208.gif

This equation is the same as that which would model a CSPP without a TESD, and when the period is extended to a year the equation and relationship still applies.