Solar Towers with Supercritical Steam Parameters - is the Efficiency Gain worth the Effort? Academic research paper on "Materials engineering"

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Energy Procedia 69 (2015) 1123 - 1132

International Conference on Concentrating Solar Power and Chemical Energy Systems,

SolarPACES 2014

Solar towers with supercritical steam parameters -is the efficiency gain worth the effort?

J. H. Peterseima* and A. Veeraragavanb

aPhD, Institute for Sustainable Futures, University of Technology Sydney, Level 11, UTS Building 10, 235 Jones Street,

Ultimo NSW 2007, Australia, 2PhD, Lecturer Mechanical and Mining Engineering, The University of Queensland

Abstract

The concentrating solar power (CSP) industry is working intensively on cost reductions to increase the economic competitiveness of CSP plants. Efforts include new power plant concepts, optimised component manufacturing, as well as higher steam parameter and these indeed lead to cost reductions over the last years. However, further improvements are required and this paper analyses the impact of supercritical steam parameters in CSP plants. A few decades ago the coal industry moved from subcritical Rankine cycle power plants, currently all CSP plants operate at subcritical conditions, to supercritical steam parameters to improve cycle efficiency. Many supercritical coal plants are now in commercial operation with significant engineering experience available in regards to plant design, construction and operation. The CSP industry can use this expertise to reproduce the benefits in their plants but some challenges also exist. Currently, supercritical CSP steam has only been shown at demonstration scale and the upscale to the smallest turbine size for supercritical parameters, being 250 MWe steam turbine capacity, is very significant. This paper compares three different 250 MWe (net) solar tower scenarios, one with subcritical and two with supercritical steam parameters. One supercritical power plant scenario is based on a novel high temperature stable molten salt that allows steam parameter of 620 °C at 280 bar and the second uses current molten salt and natural gas to reach supercritical conditions. The analysis shows that the net plant cycle efficiency can be raised from 41.3 % in a subcritical to 44.2% in a supercritical concept, which translates into a levelised cost of electricity reduction of 4.3%.

© 2015TheAuthors. Publishedby Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer review by the scientific conference committee of SolarPACES 2014 under responsibility of PSE AG

* Corresponding author. Tel.: +61 (0)2 9514 4636; fax: +61 (0)2 9514 4941. E-mail address: JuergenHeinzMartin.Peterseim@uts.edu.au

1876-6102 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer review by the scientific conference committee of SolarPACES 2014 under responsibility of PSE AG doi:10.1016/j.egypro.2015.03.181

Keywords: Solar tower; supercritical steam parameter; Rankine cycle;

1. Introduction

Over the last decade concentrating solar power (CSP) plants continuously increased cycle efficiency and reduced cost through innovation, improved equipment and optimized Rankine cycle design. Today's modern solar tower plants operate at subcritical conditions with steam parameters reaching 565 °C at 160 bar in direct steam units and 540 °C at 120 bar in molten salt units. These parameters are similar to utility scale coal fired plants from the 1970-80s. To raise efficiency the designers of coal fired units moved to supercritical steam (SCS) parameters and is likely that the CSP industry will follow the same trend in the future to also raise cycle efficiency and lower generation costs. Large plant capacities, >250 MWe, are required to use available steam turbine equipment and justify the additional complexity of supercritical power plants. While SCS CSP plants are more complex in regards to steam temperature and pressure the technology also provides benefits as two-phase flow related issues become obsolete.

Recently, CSIRO in Australia demonstrated SCS parameters in a demonstration facility [1], which was an important step but also shows the long road ahead for SCS CSP plants. SCS plants with direct steam generation (DSG) have a shorter implementation pathway than SCS plants using a primary molten salt cycle but they also compete with utility scale photovoltaic systems as energy storage for DSG plants is not yet commercially mature. Research is ongoing in high-temperature stable molten salts, up to 700 °C [2], and considering the limited SCS expertise to date such salts should be available within the next 5-7 years when the technology is ready for significant up-scale.

To analyse if the cost reductions justify the additional complexity of SCS CSP plants three different plant concepts are compared in a case study approach. The capacity and technology selected is a 250 MWe (net) solar tower with 7 h molten salt thermal energy storage (TES) and single steam reheating. The three assessed options are based on currently available steam turbine technology:

• Scenario 1: Currently available subcritical Rankine cycle plant with steam parameters of 545 °C at 165 bar,

• Scenario 2: Molten salt solar tower heating 280 bar supercritical steam to 545 °C with additional natural gas superheating to reach a steam temperature of 620 °C,

• Scenario 3: Supercritical Rankine cycle plant using pre-commercial 700 °C molten salt [2] with steam parameters of 620 °C at 280 bar.

The base case scenario 1 is a tower technology that is deployable immediately with various suppliers available to build such a facility. Commercial proponents for a 250 MWe already exist, such as 2x 250MWe Palen project in the US [3], but no proposal is near construction yet.

Scenario 2 is also immediately deployable as existing molten salts can heat supercritical steam to 545 °C without modification to the receiver and TES system. The molten salt/steam heat exchangers would have to be designed for the higher pressure of 280 bar but this is a straight forward mechanical design solution with available engineering expertise for such steam parameters. The additional steam superheating is done with natural gas, which has already been realised in plants, such as the 100 MWe Shams One parabolic trough plant [4] and the 27 MWe Moerdijik energy from waste plant [5]. The external superheater design for scenario 2 is possible with today's modern engineering tools. Part of the external superheater energy would be used for feedwater and combustion air preheating to maximize cycle efficiency. While the plant concept has not been realised at this capacity the technical and finance risk is comparatively low due to the use of available technologies.

Scenario 3 is be based on a pre-commercial molten salt able to withstand temperatures of 700 °C [2]. This would allow the generation of supercritical steam parameters without an additional fuel. The molten salt has not been tested yet in a commercial plant, which increases technology and finance risk, but significant research efforts are being made to raise the maximum molten salt temperature levels.

2. Method

The assessment analyses the benefits of supercritical steam parameters in CSP tower plants by investigating the technical, economic and environmental performance of two different options and comparing these to a standard subcritical tower. A case study approach is chosen to compare the options differences but the assessment is transferable to other locations when adapting site and cost conditions.

2.1. Modeling

All scenarios will be thermodynamically and economically modeled with Thermoflex and Steam Pro versions 23. The software is well established for modeling supercritical Rankine cycle systems, mainly coal- and gas-fired generation, and is widely used in academia and industry for actual CSP and conventional power plant modeling.

Despite higher investment and lower cycle efficiency all scenarios will be modeled with air rather than water-cooling as water is typically a scarce commodity in high solar irradiance areas. The investment data was derived from the Thermoflex and Steam Pro database but the cost assumptions were verified by meeting 110 MWe Crescent Dunes solar tower cost. From this base case the cost uncertainties to move to supercritical parameters is low as this is a key feature of the Thermoflex and Steam Pro software packages. Recent studies from the Savannah River National Laboratory (SRNL) indicate that certain metal alloys (such Haynes-230) and ceramics (SiC) can be suitable materials for the high temperature salt TES. It is also worthwhile to note that, Thermoflex has the ability to consider material costs for such high temperature and pressure resistant alloys. All process diagrams are provided in the appendix. They include all relevant information to reproduce this analysis, including feedwater heating arrangement, live and reheat steam conditions, primary working fluid temperatures etc.

The scenarios are based on equal finance cost for all plants, plant commissioning in 2025, 30 year operation and a discount rate of 7.9%. The current solar field and tower costs are lowered by 20% to reflect the expected cost reduction to 2025. All scenarios are expected to have a net annual generation of 985.5 GWh, which equals a capacity factor of 45%.

Key technical data for the plant design include 120 mbar condenser inlet pressure, a design ambient air temperature of 27.9 °C, a design DNI of 800 W/m2, a solar multiple of 2, 7 h of TES and steam reheating at 40 bar. The three process diagrams are provided in the appendix.

2.2. Case study

The site selected for this analysis is near Longreach in Queensland, Australia, as the site has an annual average DNI of 2,564 kWh/m2 that is ideally suited for a utility scale grid-connected solar tower plant. The direct normal irradiance as well as temperature data are based on information from the Australian Bureau of Meteorology [6], [7].

3. Results and discussion

This chapter provides the technical, economic and environmental analysis of the three scenarios as well as a discussion about technology risk, upscale, and future efficiencies.

3.1. Technical

As expected the efficiency of the supercritical scenarios 2 and 3 is higher than the subcritical scenario 1. Table 1 presents the various steam parameters and its impact on net cycle efficiency. The modeling shows that the net cycle efficiency increases by 7.1 % from scenario 1 to 3, which is a significant increase. Despite the same steam parameters the cycle efficiency of scenario 2 is slightly lower than in scenario 3, which is explained by: a) the energy losses associated with the 90°C exhaust gas leaving the external superheater unused and b) additional parasitic losses for pumping and the operation of draft fans.

The receiver capacity decreases with rising cycle efficiency in scenarios 2 and 3. While scenario 1 requires a capacity of 605 MWth, scenario 2 has a significantly smaller capacity of 472 MWth, and scenario 3 of 565 MWth.

The receiver capacity in scenario 2 is significantly smaller because natural gas for steam superheating replaces some of the solar energy input. The increased cycle efficiency also has an impact on solar tower height and required heliostat field land area. The tower height decreases from 280 m in scenario 1 to 245 m in scenario 2 while the

heliostat field land area decreases from 11.6 km2 to 9.1 km2.

Table 1: Technical comparison for a 250 MWe (net) solar tower plant

Steam temperature, Steam Gross efficiency, Net efficiency, Gross capacity,

°C pressure, bar % % MWe

Scenario CSP technology

1 Subcritical tower 545 165 43.8 41.3 264.8

2 SCS tower + gas 620 280 46.9 43.9 265.4

3 SCS tower with new salt 620 280 47.3 44.2 267.7

3.2. Economic

The paper will provide investment and levelised cost of electricity (LCOE) for the different options. A key feature of the analysis will be the variations in cost for the different plant components, e.g. in a supercritical plant the solar field is likely to be lower cost but the steam turbine higher cost. The modeling software used is capable of providing this information and showing these differences are important to identify the cost drivers in supercritical units

The investment and levelised cost of electricity (LCOE) results are provided in Table 2. Scenario 2 has the lowest cost as part of the solar energy input is replaced by natural gas, which lowers the number of heliostats required, tower height and receiver capacity. With thermal energy storage (TES) being a key benefit of CSP the impact of higher temperature TES on plant cost is interesting. The TES cost of scenario 1 amount to AU$72m and while the TES cost in scenario 3 decrease to AU$65m. However, this 10 % cost reduction requires the availability of higher temperature stable molten salts.

While scenario 2 does best in regards to plant cost its LCOE is, at a gas price of AU$ 6.5/GJ, with AU$133/MWh identical to scenario 3. The LCOE reduction compared to scenario 1 is 4.3 %, which might not be as high as initially expected in particular considering the Australian 2020 LCOE target of AU$ 120/MWh [8] and the US target of US$ 60/MWh [9]. The reason for the higher LCOE is that this analysis assumes a more conservative cost reduction trajectory than other studies due to new market uncertainties, e.g. disappearance of the Spanish CSP market and slower growth in the US market. The solar field and tower costs are lowered by 20 % to reflect future cost reductions but all other component costs are not expected to decrease significantly except for the benefits obtained through higher cycle efficiency. Some minor cost reduction are likely through learnings with the installation of steam turbines, condensers etc. but the cost of this equipment is unlikely to fall significantly as it has been manufactured, supplied and installed for many decades with high competition from various vendors. New CSP plant concepts, such as Brayton and supercritical CO2 cycles, might lead to higher cost reductions but are less mature, in particular supercritical CO2 cycles.

It has to be highlighted that the LCOE of scenario 2 strongly depends on the natural gas price. At below AU$ 6.5/GJ the LCOE would be lowest of all scenarios but at a gas price of AU$ 11/GJ the LCOE is equal to scenario 1. Natural gas prices are currently increasing in Australia due to the availability of new liquefied natural gas export terminal coupling the local gas price to the higher price world market. This has to be kept in mind when considering CSP plants with natural gas superheating.

Table 2: Economic comparison for a 250 MWe (net) solar tower plant

Scenario CSP technology Plant cost, AU$m Specific cost, AU$m/MWe AU$/M

1 Subcritical tower 1,360 5.4 139

2 SCS tower + gas

3 SCS tower with new salt

1,200 1,300

4.8 5.2

133 133

3.3. Environmental

Based on the net annual generation of 985.5 GWh and Australia's 2010 CO2 intensity of electricity generation of 841 kg CO2/MWh [10] the scenarios 1 and 3 can offset CO2 emissions of up to 830,000 t annually. The annual abatement potential for scenario is due to its use of natural gas slightly lower, up to 755,000 t. This leads to a carbon intensity of scenario 2 of 77 kg/MWh, which is very low compared to any current fossil fuel plant and the 80% carbon emissions reduction target by 2050.Water and land use are other criteria but have not been investigated here in detail as they will vary depending on the project location. In general the water and land use of scenario 2 is smallest as the natural gas superheater reduces the solar field size and number of heliostats.

3.4. Risk

Applying the supercritical Rankine cycle concept to the solar tower technology has an inherent risk as the technology has not yet been proven in this field. First on-sun demonstration trials were successful in Australia [1] but the pathway to a commercial 250MWe plant is long and intermediate steps are required for scale-up. Supercritical steam turbines are not available for scale-up to capacities of 50 or even 100 MWe and therefore the only option to demonstrate the technology at a larger commercial scale is the integration of a SCS CSP tower system into a supercritical coal plant. Many modern coal plants with supercritical steam parameters operate worldwide with some of them being in suitable high DNI regions, such as 750 MWe Kogan Creek power station in Australia. The hybridisation of CSP with coal plants is contentious [11] but seems to be the only option to up-scale SCS CSP systems as smaller supercritical steam turbines are not available and combined cycle natural gas plants do not operate at such conditions. This would not only be a low risk approach - the fossil fuel plant can take over at any time - but also low cost approach as major high cost items, such as steam turbine and condenser, could be used. Considering that the solar field, receiver and TES contribute to approximately 60% of the solar tower plant cost [12] an immediate cost reduction of 40% is possible. Also the receiver could be mounted on the stack of the coal fired plant as these have heights of typically >180 m. This would obviously require modifications to the stack but the significant saving of not having to erect a new tower would offset these. This concept has already been investigated for integrated solar combined cycle plants [13], [14] and Figure 1 provides an example of a 100 MWe equivalent SCS CSP retrofit with 5 h TES to a 2 GWe coal plant.

Fig. 1. Example of a 100MWe equivalent supercritical solar tower retrofit with 5 hours of advanced thermal storage to a 2GWe coal plant.

The technical risk mostly applies to the primary molten salt cycle as the design of the supercritical Rankine cycle can be based on significant experience from supercritical coal plants. Obviously, special attention has to be given to the plant cycling but these issues can be addressed with modern engineering tools and back-up boilers. The use of molten salt as the primary working fluid is also well understood with various commercial plants in operation. The increased scale and working fluid temperature in a 250 MWe SCS CSP plant has its inherent challenges in regards to optimised storage tank and heat exchanger design but experience with current 540 °C systems will be very useful and the technology risk can be considered moderate.

An important non-technical aspect is the availability of qualified personnel to operate a SCS CSP plant. In Australia only one coal fired power station operates at supercritical parameters and finding personnel to operate a SCS CSP plant would be an additional challenge. Globally, more experience is available and special training can address this barrier.

3.5. Future steam parameter

The supercritical steam parameters used in this assessment reflect current state-of-the-art technologies as applied in several modern supercritical coal fired power plants. To improve cycle efficiency and lower emissions current research and development aims to implement new materials that can withstand steam parameter of 700 °C at 350 bar [15]. Once material tests are successful and first demonstration plants commence operation the same equipment could be used in CSP plants. Assuming steam parameter of 700 °C and 350 bar in a plant configuration similar to scenario 3 cycle efficiency would increase to 49.1 % (gross) and 45.6 % (net). Due to the absence of publicly available data it is not possible to reliably estimate the overall cost impact of such materials. However, considering that the solar field is a high cost part of any CSP plant it is likely that the increased cost for the final superheater, high temperature/pressure piping, and the high-pressure steam turbine section do not offset the solar field savings.

4. Conclusions

As shown in the analysis, raising the steam parameter of CSP plants to supercritical conditions can increase net cycle efficiency to 44.2 % and improve the LCOE competitiveness by around 4.3 % to achieve a LCOE of AU$133/MWh. This analysis assumes a conservative CSP cost reduction trajectory and the results strongly vary dependent on the deployed CSP capacity within the next years. The upscale from current SCS demonstration facilities to commercial installation has its specific challenges as SCS turbines are not available at capacities smaller 250 MWe. This poses a significant up-scale challenge which could be addressed by retrofitting SCS CSP systems to modern SCS coal plants. However, this should only be an intermediate step to test the technology at commercial scale and further upscale to utility scale standalone CSP plants.

Standalone SCS CSP plant would require new high-temperature stable molten salts to be available and research is already ongoing in this field. To quickly upscale the SCS technology this analysis investigated a high solar share CSP concept which uses natural gas to raise steam parameters above the temperature limits of current molten salts. This concept achieves comparatively low plant cost and LCOE, reduces technical risk, but its economic viability strongly dependent on future natural gas prices.

The CSP industry is working on cost reductions through innovation and the move to supercritical steam parameters is certainly one option to increase cost competitiveness. While some challenges exist, such as availability of high temperature stable molten salts and minimum 250 MWe steam turbine capacities, significant engineering expertise is available from the design of coal fired power plants, which commenced to move to supercritical steam parameters over the last few decades. This ensures the availability of experienced supplier and lowers technology risk for future supercritical CSP plants.

Appendix

The appendix contains the three process diagrams this analysis is based on and provides all relevant technical assumptions to reproduce this analysis.

J.H. Peterseim and A. Veeraragavan /Energy Procedía 69 (2015) 1123 - 1132 A.1. Process diagram of the subcritical solar tower design - scenario 1

J.H. Peterseim and A. Veeraragavan / Energy Procedía 69 (2015) 1123 - 1132 A.2. Process diagram of the supercritical solar tower design with natural gas superheating - scenario 2

J.H. Peterseim and A. Veeraragavan /Energy Procedía 69 (2015) 1123 -A.3. Process diagram of the supercritical solar tower design - scenario 3

References

[1] CSIRO, "Supercritical solar - new frontier for power generation," 2014. [Online]. Available: www.csiro.au/Portals/Media/Supercritical-solar-new-frontier-for-power-generation.aspx. [Accessed: 24-Jul-2014].

[2] J. W. Raade, B. Elkin, and J. Vaughn, "Novel 700 °C molten salt for solar thermal power generation with supercritical steam turbines," in

SolarPACES Conference, 2012.

[3] P. M. de Molina Melendez, "Abengoa and BrightSource Energy partner to build the world's two largest solar power towers," Abengoa Solar, S.A., 2013. [Online]. Available:

http://www.abengoasolar.com/web/en/acerca_de_nosotros/sala_de_prensa/noticias/2013/solar_20130315.html. [Accessed: 22-Jul-2014].

[4] O. Goebel and F. Luque, "Shams One 100 MW CSP Plant in Abu Dhabi - Update on Project Status," in SolarPACES Conference, 2012.

[5] J. Toebes and M. Beker, "External steam input and process steam output at the Moerdijk (Netherlands) combined-cycle plant," VGB POWERTECH, no. 7, Essen, Germany, pp. 21-27, Jul-1998.

[6] Bureau of Meteorology, "Australian Monthly Average Hourly Global and Direct Normal Solar Exposure Gridded Data," Melbourne, 2012.

[7] Bureau of Meteorology, "Climate statistics for Australian locations - Longreach," 2014. [Online]. Available: www.bom.gov.au/climate/averages/tables/cw_036030.shtml. [Accessed: 22-Jul-2014].

[8] CSIRO, "Australian solar thermal research initiative," 2013. [Online]. Available: http://www.csiro.au/en/Organisation-Structure/Flagships/Energy-Flagship/ASTRI.aspx. [Accessed: 04-Apr-2014].

[9] SunShot Initiative, "SunShot Vision Study," U.S. Department of Energy, Washington, DC, 2012.

[10] International Energy Agency, "CO2 Emissions from Fuel Combustion 2012 - Highlights," Paris, 2012.

[11] J. H. Peterseim, S. White, A. Tadros, and U. Hellwig, "Concentrating solar power hybrid plants - enabling cost effective synergies," Renew. Energy, vol. 67, pp. 178-185, 2014.

[12] IRENA, "Renewable Energy Technologies: Cost analysis Series Volume 1 Issue 2/5 Concentrating Solar Power," International Renewable Energy Agency, Abu Dhabi, 2012.

[13] J. H. Peterseim, S. White, A. Tadros, and U. Hellwig, "Integrated Solar Combined Cycle plants using solar power towers to optimise plant performance," in SolarPACES Conference, 2012.

[14] J. H. Peterseim, A. Tadros, U. Hellwig, and S. White, "Integrated solar combined cycle plants using solar towers with thermal storage to increase plant performance," in ASME Power Conference, 2013.

[15] S. J. Patel, J. J. DeBarbadillo, B. a. Baker, and R. D. Gollihue, "Nickel Base Superalloys for Next Generation Coal Fired AUSC Power Plants," Procedia Eng., vol. 55, pp. 246-252, Jan. 2013.