Abstract--This paper will describe a novel system for generating dispatchable electric power using wind and solar thermal energy combined through compressed air for transmission and storage and large area solar air heating collector integrated with high heat capacity thermal storage media. The essential subsystems include wind turbines directly coupled to air compressors, high pressure (100 bar) large diameter pipeline, solar collectors with integral thermal storage for air heating, and a turbo-expander driven generator. This hybrid power generation system will be particularly useful in electric systems, such as the Electric Reliability Council of Texas (ERCOT), with very large wind energy potential, severe transmission limitations in very windy locations, areas of high solar radiation, and a significantly higher market value for dispatchable power than for intermittent energy. This hybrid wind / solar system would be a more economical means for achieving zero-emission, firm, dispatchable capacity than independent construction of wind and solar plants. Index Terms—energy storage, compressed air, thermal storage, turbo-expander I. Introduction
Utilities have a growing need for a zero-emission generating system that also meets their requirements for capacity they can count on during peak load periods. Neither wind nor even solar alone can satisfy the requirement of firm capacity during peak periods (late afternoon and early evening during summer months). To make wind and solar dispatchable, a variety of energy storage concepts have been proposed – compressed air, thermal storage (for solar) and electrochemical systems (ie batteries and electrolysis/fuel cells). It should be considerably more economical, however, to integrate wind, solar-thermal, and the appropriate energy storage media in a single system, than to separately design and construct wind and solar plants. That is the objective of the work described in this paper.
Figure 1 illustrates the concept under investigation. The essential subsystems include:
wind turbines directly coupled to air compressors, or conventional wind turbines used to power a central air compressor on the ground
high pressure (100 bar) large diameter pipeline to transport the compressed air from the wind turbines to the solar site which will also contain the expander-generator,
possibly a compressed air storage cavern, which could be sited either near the wind turbines, along the transmission pipeline, or at the solar collector site,
solar thermal collectors with integral thermal storage for heating the compressed air to as high as 1000° C, and
a uniquely designed turbo-expander-driven generator.
III. The Challenges Facing Future Wind Energy in Texas
The Texas electric grid is very weakly connected to the rest of the US grid. Texas has enormous potential for additional wind energy development – wind developers with a total of at least 35,000 MW in projects have filed plans with ERCOT (the Texas power pool and independent system operator). While transmission congestion limits the rate of growth at present, the ultimate limit will be the minimum load, which occurs at the same time of day (ie middle of the night) as the strongest winds. Figure 2 is the load duration curve for ERCOT with baseload generating capacity and planned 2009 wind capacity shown. Note that for over 4000 hours per year, the load will exceed the combined generating level of nuclear, coal and wind. Coal fired generating units have to back off to allow the system to accept the wind energy, and they will have to back off more as more wind capacity is built. One impact of this is to force wholesale energy prices in the middle of the night to lower and lower levels. The absolute ceiling on wind development, however, will be the difference between night time loads and the level of responsive reserve capacity that must be kept operating for system reliability.
Fig.1 Schematic Diagram of Concept
Fig. 2 Load Duration Curve with Baseload and Wind Capacities
Clearly adding energy storage alleviates this problem and facilitates the addition of more wind capacity. Other economic benefits of storage can be computed by viewing the hourly market clearing prices of electric energy and computing the value of the current time profile of wind generation and comparing this to the value of the same daily energy production scheduled to generate during the highest priced hours. Such an analysis was performed for the months of January, February and March 2008 for the Sweetwater 90 MW wind farm (near Abilene, TX), assuming a 6 hour storage system. The value of the scheduled generation from storage was 80 % greater than the value of the conventional wind plant for that period.
A third benefit of adding energy storage is its ability to enable renewables to effectively substitute for conventional generating capacity – ie the value of full dispatchability. As figures 3 and 4 illustrate, neither wind nor solar (without storage) can be counted on during the peak hours in Texas. Even solar falls to zero at the time of day when the ERCOT system load is at 90% of the peak.
Fig. 3 Profile of Typical Wind Plant Output and ERCOT Load
IV. Energy Storage Options
The three categories of utility-scale energy storage technology are pumped hydro-electric, flow batteries, and compressed air with storage in underground caverns and/or large diameter pipes. Pumped hydro-electric has never been built in Texas because the necessary terrain and water resources are lacking in the same location. Currently, compressed air is far more developed (and more economical) than flow batteries.
In a compressed air energy storage system, off-peak electricity is used to power motor-driven air compressors and the compressed air is stored in underground cavern (typically a solution-mined cavern in salt). In the generation mode, the pressurized air is withdrawn from the cavern, preheated using a combustible fuel (typically natural gas) and then expanded through a hot air turbine which drives a generator.
Fig. 3 ERCOT Peak Day Load and Solar Generation Profile
Preheating the compressed air prior to expanding it is necessary for two reasons. One is to significantly decrease the amount of air required per unit of energy produced by the expander; the second is to avoid clogging the air turbine with ice (from moisture in the air) since expansion of highly compressed air has a significant (cryogenic) cooling effect. For example, expanding 20 atm air to 1 atmosphere starting at 20° C would have an exit temperature of minus 115° C. Heating the air to an even higher temperature than 200° C reduces the required flow rate. In the case of the two existing Compressed Air Energy Storage systems (one in Huntorf, Germany, started up in 1978 and the other in MacIntosh, Alabama, started up in 1991) natural gas combustors are used to pre-heat the air prior to expansion in the power turbine. In both plants, approximately 4000 kJ per kWh were required for this preheating . In recently proposed designs, the compressed air would be heated using the exhaust of a conventional gas turbine or an industrial waste heat source.
The essence of the Hybrid Wind / Solar Power Plant is substituting solar-heated thermal storage for natural gas for the pressurized air preheating.
V. Conceptual Design
This section describes the approach taken for calculating the ratings and dimensions of the basic components for a wind / solar hybrid power system capable of generating at the rate of 100 MW for 8 hours in August, assuming a West Texas location.
Analysis begins with the characteristics of the air expander-generator. Assuming that the inlet conditions of the air are 20 atmospheres (280 psig) and 200° C, the air flow rate into the expander would be approximately 14,000 kg per MWH. To generate 100 MW for 8 hours per day, therefore requires 11.2 million kg of compressed heated air per day. This value is used both to determine the size of the compressed air storage component, and the cumulative capacity of the wind-driven compressors.
For our initial conceptual design, we start with 200° C to simplify the design and minimize the cost of the solar collector and thermal storage subsystem. Higher temperatures are being analyzed as this paper goes to press.
B. Pipeline and Storage Reservoir.
Assuming that the air is stored at 70 atm., its density is 84 kg per m3 . To store 11.2 million kg would therefore require a storage “reservoir” volume of 133,000 m3. If the wind plant were connected to the expander by a 1.2 meter (48 inch) diameter compressed air pipeline, this volume would be contained in a pipeline just under 120 km in length. Such a compressed air pipeline could therefore serve both for transmitting the wind-generated energy and storing it. The idea of using compressed air pipelines for energy transmission and storage was explored in a paper given at last year’s Power Gen Renew Conference by Enis, Lieberman and Rubin. Needless to say, if an existing pipeline, perhaps one originally constructed to carry natural gas but no longer needed for that purpose, happened to be available at the right location, project costs could be reduced considerably.
C. Solar collectors and thermal storage.
The specific heat of air is very close to 1.0 kJ/kg C. Therefore, to heat the 11.2 million kg of air from ambient (assume 20° C) to 200° C requires 2,000 GJ per day. Average daily global solar radiation (global horizontal) in Pecos, TX is 6.9 kWh per m2 in August, equal to 24.8 MJ per m2 daily. Assuming a solar collector efficiency of 60% (a reasonable assumption for desired temperature of 200 C) the collector area required would be 130,000 m2. Since solar radiation coincides well, but not perfectly, with the 8 hour peak period during which the expander would operate, the thermal storage component would not have to be sized for the full 8 hours. Rather, 3 hours of thermal storage are incorporated in this preliminary design. Thermal storage requirement would be 750 GJ.
D. Wind Turbines.
To estimate the number and rating of the wind turbines necessary to compress 11.2 million kg of air per day, we assume that a wind plant consisting of 1.5 MW wind turbines with characteristics similar to the GE 1.5 MW machine are used to power a central air compressor that compresses air from 1 to 70 atmospheres. Based on the August wind pattern of the Sweetwater Wind Plant alluded to above, approximately 30 wind turbines would be required to compress 11.2 million kg to 70 atmospheres. Directly coupling smaller air compressors to the shafts of these wind turbines (ie replacing their gearboxes and generators with simplified gearboxes and compressors) would probably result in the same number of wind turbines.
The proposed hybrid system has several advantages over separate wind and solar generating facilities. They are:
a single prime mover – generator and substation
no need for electric transmission connection to the wind farm (only to the expander-generator),
wind turbine gear box requirements are greatly simplified (since the speed/torque characteristics of a wind turbine rotor are better matched to an air compressor than a generator),
no need for natural gas or an industrial waste heat source, and
no need for cooling water (solar thermal plants typically use steam as working fluid so condenser cooling water is needed).
Also, since wind energy production in West Texas tends to be higher in winter and spring, and solar is highest in summer months, seasonal fluctuations in energy supply tend to even out when the two sources are combined in an integrated system.
VI. Future Research
Technical issues to be addressed include:
Perform optimization to determine the best inlet temperature for the hot air expander ; higher temperatures increase the output per kg of pressurized air, but higher temperatures also require more complex solar collector systems (ie sun-tracking systems for concentrating solar)
Analyze power variation under various combinations of air temperature and flow rate entering the expander, in order to quantify the seasonal variation in power output capability
Select the most suitable thermal storage media.
Design the solar collector that is optimal for the inlet temperature.
Perform cost estimates and conduct cost/benefit analyses
Assuming cost/benefit analyses demonstrate advantages over conventional wind and solar power development, begin search for sites for wind and solar within 100 km of each other and also search for existing natural gas pipelines no longer in use
compare sliding pressure vs constant pressure design for the expander
Dr. Ben Enis, Dr Paul Lieberman, Irving Rubin, Duane Bergmann, Randy Dirlam, Septimus van der Linden, Power Generation Sources, Transferline Compressed Air Energy Storage System with Electricity, HVAC and Desalination, Presented at PowerGen Renewable, April 10-12, 2007
ERCOT Market Clearing Price of Energy
Mark Kapne, PE is with Austin Energy, 721 Barton Springs Road, Austin, TX 78704, USA (email: firstname.lastname@example.org)