Elsevier

Solar Energy

Volume 108, October 2014, Pages 390-402
Solar Energy

A cost and performance comparison of packed bed and structured thermocline thermal energy storage systems

https://doi.org/10.1016/j.solener.2014.07.023Get rights and content

Highlights

  • Thermocline-type thermal energy storage systems are much more cost effective than currently-used systems.

  • Packed-bed thermoclines are most efficient in terms of cost and performance.

  • Structured thermoclines offer low-cost, efficient storage with no concern of thermal ratcheting.

Abstract

A structured concrete thermocline thermal energy storage (TES) system is proposed as an alternative to currently-used TES systems. The issues of material settlement and thermal ratcheting found in packed bed thermocline TES systems is avoided by replacing the packed aggregate bed with structured high-temperature concrete. A summary of all utility scale TES systems with integrated TES in existence today is provided and discussed. Cost reduction options such as replacing two-tank systems with single-tank systems and replacing liquid storage media with solid storage media are discussed along with limitations of both options. Numeric models are developed to simulate the performance of utility scale packed bed and structured thermocline TES systems; efficiencies of 92.37% and 84% are modeled for packed-bed and structured systems. A complete cost analysis of utility-scale, 2165 MWh packed bed and structured systems is conducted; capacity costs of $30/kWh and $34/kWh are determined for packed bed and structured systems respectively. A structured concrete thermocline is deemed to be a viable TES option due to its low cost and the fact that there are no concerns of thermal ratcheting of the tank.

Introduction

Increasing global energy demands and diminishing fossil fuel resources have raised interest in renewable energy resources. Sufficient solar irradiance is incident on the Earth’s surface to fuel the production of electrical energy for all of its inhabitants (Goswami et al., 2000). However, efficient and reliable methods of harvesting solar energy for power production are needed. Solar energy is harvested and converted to electrical energy using one of two technologies: photovoltaic (PV) panels or concentrating solar power (CSP) plants.

PV panels convert solar energy to electric energy directly; this method of conversion is mature, scalable, and relatively low cost: $4740/kW for utility installments (>100 kW) (Feldman et al., 2012). Two primary drawbacks are associated with PV: electric power output is subject to rapid fluctuations and is only available when the sun is shining. Considering the first drawback, power production can rapidly shift from peak to minimal and back to peak as clouds drift in front of the sun. It is challenging and costly to design control systems and backup generators capable of responding to these fluctuations rapidly enough to maintain a steady supply of power to the grid (Denholm and Mehos, 2011). Considering the second drawback, since PV panels convert solar energy directly to electrical energy, it follows that electricity can only be produced when the sun is shining. Unfortunately, times of peak power demand (shortly after sunset) do not coincide with periods of peak irradiance; though excess electrical energy can be produced during the day and utilized during evening hours, the high cost of large-scale electrical energy storage is not viable (Table 1).

Though CSP is a younger and more costly conversion technology, approximately $5500/kW for utility installments (>50 kW) (Feldman et al., 2012), it is a more viable option than PV at the utility scale because CSP plants can be easily integrated with thermal energy storage (TES). Excess solar energy can be collected and stored as thermal energy. This energy can then be dispatched at will for conversion to electrical energy in a traditional Rankine steam power cycle. This effective “thermal battery” allows the plant to continue power production well into the hours of peak demand after sunset.

A summary of all CSP plants with integrated TES that are in operation or are under construction is provided in Table 2 (NREL, 2014). From this list, it can be seen that the current standard in TES is the two-tank indirect (TTI) configuration with nitrate salt as the energy storage media, and the operating temperature range of 290–390 °C. Fig. 1A illustrates the TTI TES system configuration and a parabolic trough solar collector field. Solar energy is collected in the solar field, where parabolic troughs focus irradiance on receivers, through which thermal oil heat transfer fluid (HTF) is circulated. The oil is then circulated to an oil-to-salt heat exchanger, where the collected energy is transferred to nitrate salt. Heated salt is stored in a “hot” tank until the energy is dispatched, at which point the oil and salt are circulated through a salt-to-oil heat exchange. The hot oil then circulates to the power block and the “cold” salt is stored in the “cold” tank until it is re-heated.

Salt is utilized as energy storage media primarily because it is much more cost-effective than using thermal oil as storage media. To put this in perspective, the thermal mass of solar salt (ρcp) of solar salt and Therminol VP-1 (thermal oil) are relatively comparable at 2.485 MW/m3 °C and 1.865 MW/m3 °C (Van Lew et al., 2011). However, the difference in operating temperature differentials of the two fluids is significant, at 275 °C and 100 °C respectively. This means that on a volumetric basis, solar salt has 3.664 times the energy storage density of thermal oil. The consequence of this increased volume is that more storage media and larger storage systems are needed corresponding higher system costs. Therefore, although salts are corrosive and challenging to work with, the cost reduction benefit outweighs the associated complications.

Referencing Table 1, it can be seen that the TTI configuration is considerably more costly than the other TES configurations (last two entries in Table 1). Therefore, it is logical that one would question why a more expensive technology is still the standard. Referencing Table 2, it can be seen that the majority of CSP facilities utilize parabolic troughs to collect solar energy; these systems have a lower concentration ratio than tower-type systems, meaning that the operating temperature is lower. Utilizing the TTI configuration means that oil is circulated through the solar receiver instead of salt, thereby eliminating concerns of HTF solidification in the collector (solidification of salt occurs at approximately 222 °C). Furthermore, the TTI configuration affords the near-perfect efficiency of 97% and a constant discharge fluid temperature (Medrano et al., 2010).

The primary drawback associated with the TTI configuration is its high cost. Several of the primary cost-inflating factors are as follow: steel tanks, large oil and salt inventories, and oil-to-salt heat exchangers. The two storage tanks of a TTI system are constructed from stainless steel to avoid being corroded by the nitrate salt storage media, therefore, they are very costly. Though nitrate salts are chemically stable in the presence of air up to 565 °C, the energy storage temperature is limited by the chemical stability of thermal oils (around 400 °C) to peak operating temperature of 390 °C. Therefore, a much larger nitrate salt inventory is required than would be necessary if the salt’s heat capacity was fully utilized.

An alternative to the TTI configuration that takes full advantage of the salt’s storage capacity is the two-tank direct (TTD) configuration. In the TTD configuration, the thermal oil and oil-to-salt heat exchanger are eliminated, and the salt is circulated directly to the collector field to be heated before entering storage. Additionally, the TTD system affords high efficiency of 97% similar to that of the TTI system. Although the TTD configuration provides storage at lower cost to than the TTI configuration, the presence of two large storage tanks and a large salt inventory still leaves the system cost relatively high.

A cost-reducing alternative configuration to the TTD system is the direct thermocline (DTC) configuration, a schematic of which is provided in Fig. 1B. In a thermocline type configuration, hot and cold media are stored in the same tank, thereby eliminating one of the costly storage tanks required by the TTI configuration. Nitrate salt, serving the dual function of HTF and storage media, is heated directly in the solar receiver then circulated to storage, thereby eliminating the costly oil inventory and oil-to-salt heat exchangers. Finally, the energy storage temperature is no longer limited by the oil’s chemical stability, therefore the energy storage temperature can be increased from the limit of 390–565 °C. Consequentially, the required quantity of nitrate salt is reduced. Further reduction of the nitrate salt inventory can be achieved by filling a significant volume of the thermocline tank with a “packed bed” of low-cost solid storage media forming a packed-bed thermocline (PBTC) TES system. Reportedly, the PBTC configuration provides TES at 38% the cost of the TTI configuration and 68% of the cost of the TTD configuration (EPRI, 2010, Pacheco et al., 2002). Furthermore, properly-designed PBTC TES systems afford high efficiency, reportedly in excess of 93% (Van Lew et al., 2009).

Though the DTC configuration offers TES at a lower cost than the TTI configuration, two primary concerns prevent its large-scale incorporation. The first concern is shared for all direct TES systems incorporating nitrate salt as storage media: solidification of nitrate salt in the solar collection field. The solution to this problem is relatively simple, as heaters can be incorporated to prevent the salt temperature from falling below a set temperature (300 °C) for a direct two-tank configuration or by adding heat traces to the piping for DTC configurations.

The second and primary concern for DTC TES systems is thermal ratcheting of the tank’s walls. To date, the solid media tested and used in thermocline TES systems is a “packed-bed” of aggregate or aggregate and sand (Pacheco et al., 2002, Brosseau et al., 2005, Herrmann and Kearney, 2002). During the charging phase, hot salt is added to the top of the tank; the tank is made from stainless steel which has a considerably higher thermal strain rate than does the aggregate composing the bed. As the tank’s walls heat and expand, aggregate settles to occupy this new volume. During discharging, cold salt is added to the base of the tank; consequently, the tank’s walls cool and contract. However, the settled aggregate prevents the walls from returning to their initial position, resulting in the introduction of residual stresses to the tank walls. Cyclic charge and discharge processes can lead to catastrophic rupture of the tank (Flueckiger et al., 2011). A recent study investigates thermal ratcheting in the thermocline storage tank at the Solar One facility (operated from 1982 to 1986), comparing results from a finite element model to data from strain gauges on the tank (Flueckiger et al., 2012). A relatively small operating temperature range is considered: 204–304 °C; the model accurately predicts stress in the tank walls to be 190 MPa, less than half of the reported yield strength for the steel tank walls, 400 MPa. However, as energy storage temperature ranges increase, such as the 300–585 °C range considered in this work, the difference in thermal strain (which is directly proportional to temperature) between the stainless steel tank and aggregate bed becomes more significant. This leads to an accelerated and more severe level of thermal ratcheting.

Three approaches have been proposed to mitigate thermal ratcheting: utilizing a “corrugated liner” (Flueckiger et al., 2011), utilizing a novel tank design that encourages the aggregate to expand vertically (Zanganeh et al., 2012), and replacing the packed aggregate bed with structured concrete filler (Brown et al., 2012). Though the first concept is innovative, the author does not clearly explain how the liner will resist the cyclic stresses of the ratcheting without being damaged. Furthermore, the author does not propose a material from which the liner could be made. The second concept, while also innovative, is developed for an air-packed bed thermocline, and the bed walls are constructed from concrete. Testing reported by Skinner (2011) finds that even very dense concrete is permeated by molten salt; therefore it is likely that a concrete containment tank is not viable for a thermocline utilizing salt as liquid media. Construction of a large-scale steel tank having the novel frustum configuration would likely be quite costly due to the engineering and custom fabrication required. The third approach is very straightforward: structured high-temperature concrete replaces the packed aggregate bed, thereby eliminating concerns of thermal ratcheting altogether.

Structured concreted has been proposed as a low-cost energy storage media, reportedly providing sensible heat storage at costs as low as $1/kWht (Herrmann and Kearney, 2002). TES systems employing concrete as storage media typically consist of concrete prisms with embedded heat exchangers; a HTF is circulated through the heat exchanger to transfer energy to and from the system. The DLR research center reports much work on this concept, utilizing both modeling and large-scale test modules (Laing et al., 2006, Laing et al., 2008, Laing et al., 2009, Laing et al., 2010). Energy is stored up to the temperature limit of 400 °C in large-scale test modules and concrete is developed that is compatible with energy storage up to the temperature limit of 500 °C. Work performed at the University of Arkansas reports energy storage tests utilizing prisms composed of high-performance concrete cast around stainless steel heat exchangers up to the temperature limit of 450 °C with nitrate salt as HTF (Skinner et al., 2014). Also reported is the development of concrete compatible with energy storage up to the temperature limit of 600 °C (John et al., 2013). Although the cost of the high-performance concrete is low, (Skinner et al., 2014) conclude that the concrete modules are not a cost-effective system configuration due to the high cost of the stainless steel heat exchangers required.

It is proposed that the packed aggregate bed typically used in the thermocline TES concept be replaced with the low-cost, high-performance concrete reported in Skinner et al. (2014) (see Fig. 3, Fig. 4, Fig. 5). In this work, numeric models are presented and utilized to compare the performance of PBTC and structured concrete thermocline (SCTC) TES systems. The systems considered are utility-scale, having 2165 MWh storage capacity. Subsequently, cost analysis of the two systems is performed using the model laid out by (EPRI, 2010), and the unit capacity cost (cost per unit energy storage) of both systems is given. Finally, conclusions are drawn regarding the viability of a SCTC.

Section snippets

Thermocline models

In this section, the numeric models developed to study both thermocline configurations are presented and discussed briefly. Both models are finite difference (FD) based heat transfer models, with the PBTC model being 1D and the SCTC models being 2D. Properties of the molten salt are provided in Table 3 for the lower, upper, and median temperatures considered in the modeling. In some of our previous work (Strasser, 2012), comparisons of models utilizing constant and temperature-dependent fluid

Performance of thermocline models

A significant drawback associated with solid TES media is that the temperature of the fluid discharged from the storage system declines in the latter stages of the discharging process leading to decreased power block efficiency and wasted energy (Energy retrieved below the limiting minimum temperature) (Laing et al., 2009). This decline in outlet temperature can be attributed to the development of thermal gradients within the solid media. Utilizing small solid elements reduces the magnitude of

Installed cost of utility PBTC and SCTC

Pacheco et al. (2002) present what appears to be the first cost analysis of a liquid-based PBTC. Considering an indirect configuration, temperature differential of 84 °C, and solar salt and quartzite media, a cost of $13,900,000 is calculated for a 688 MWh system, corresponding to a capacity cost of $20/kWh. However, the cost analysis considers few system components and is for a very low temperature range. EPRI (2010) reports a much more thorough and detailed cost analysis procedure. Cost

Discussion

Based upon numeric modeling of PBTC and SCTC systems, it can be seen that the PBTC is about 8.5% more efficient than the SCTC (92.37% vs. 84%). Furthermore, from Table 6, it can be seen that the capacity cost of the PBTC is more than 12% less than that of the SCTC. This study indicates that the PBTC configuration is superior to the SCTC TES system both in terms of cost and performance. However, the SCTC configuration bears no concerns of thermal ratcheting, which is a significant concern for

Conclusions

Based upon numeric modeling, a PBTC is 8.37% more efficient than a SCTC (92.37% vs. 84%). Cost analysis of utility-scale 2165 MWh PBTC and SCTC TES systems indicates that a PBTC is 12.5% less costly than a SCTC ($30/kWh vs. $34/kWh). However, for the high temperature differential considered, 275 °C, thermal ratcheting of the tank’s walls by the aggregate bed is a significant concern. Novel tank liners and configurations have been proposed as countermeasures for thermal ratcheting, but these

Acknowledgments

This research work was supported by a grant from the U.S. Department of Energy (Grant # DE-FG36-08GO18147) through the University of Arkansas. The opinions expressed in the paper do not reflect those of the research sponsor.

The authors would also like to thank the reviewers and editors for their thorough critique of this work which resulted in significant improvement of the reporting of the work.

References (34)

  • Brown, B., Strasser, M., Selvam, R.P., 2012. Proceedings of ASES. Development of a Structured Thermocline Thermal...
  • Y. Çengel

    Heat and Mass Transfer: A Practical Approach

    (2007)
  • Denholm, P., Mehos, M., 2011. Enabling greater penetration of solar power via the use of CSP with thermal energy...
  • EPRI, 2010. Solar Thermocline Storage Systems: A Preliminary Design Study....
  • Feldman, D., Barbose, G., Margolis, R., Wiser, R., Darghouth, N., Goodrich, A., 2012. Photovoltaic (PV) pricing trends:...
  • S.M. Flueckiger et al.

    Thermomechanical simulation of the solar one thermocline storage tank

    J. Sol. Energy. Eng.

    (2012)
  • D.Y. Goswami et al.

    Principles of Solar Engineering

    (2000)
  • Cited by (69)

    • High-temperature heat and power storage

      2023, Future Grid-Scale Energy Storage Solutions: Mechanical and Chemical Technologies and Principles
    View all citing articles on Scopus
    View full text