ML16048A218
ML16048A218 | |
Person / Time | |
---|---|
Site: | Diablo Canyon ![]() |
Issue date: | 11/10/2015 |
From: | Pacific Gas & Electric Co |
To: | Office of Nuclear Reactor Regulation |
Shared Package | |
ML16048A230 | List:
|
References | |
DCL-15-142, CAC MF4019, CAC MF4020 | |
Download: ML16048A218 (87) | |
Text
{{#Wiki_filter:Compressed Air Energy Storage (CAES)For My Home About Contact Us Safety EnglishPage 1 of 2Go Log inSearchPG&Es EnvironmentalComnirnntCompWhat We're DoingRecreatoa Areas PrograrrPutting Energy Efficiency onNvmFirstOnNvmPacific GasClean Energy Solutionswork on aCWaveConnect"M utility's incrCompressed Air Energy The grantvStorage (CAES) (ARRA). OFighting Climate Change (CPUC) aplGreening Vehicles project. ThPG&E withPromoting Stewardship technology,Innovative Community reduce grerProgramsenergy souBuildings & Operations support forEarning RecognitionPaperless BillingWhat You Can DoTaking ResponsibilityNEXT100 Blogressed Air Energy Storage (CAES)nOverviewber 24, 2009, the U.S. Department of Energy (DOE) awardedand Electric Company (PG&E) a $25 million grant to fund initial3ompressed Air Energy Storage (CAES) project to support theeased use of intermittent renewable energy (e.g., wind, solar).N'as given under the American Recovery and Reinvestment Act)n January 21, 2010, the California Public Utilities Commissionproved PG&E's request for matching funds of $25 million for thele CPUC found that the CAES demonstration project will providea better understanding of a promising energy storage,which has the potential to lower costs for customers andenhouse gas emissions through greater integration of renewablerces. The California Energy Commission (CEC) has also shownthe project with conditional approval of a $1 million grant.Related LinksNew Type of Clean EnergyDOE Energy Storage Systems ProgramAdvanced Compressed Air Energy StorageDemonstrationFinal Environmental Assessment andFinding of No Significant Impact,Compression Testing Phase, San JoaquinCounty, CAFinal Environmental AssessmentEnvironmental Assessment AppendicesFinding of No Significant ImpactDraft Environmental Assessment,Compression Testing Phase, San JoaquinCounty, CA
Dear Reader LetterDraft Environmental AssessmentDraft Environmental Assessment AppendicesUnderground Injection Control (UIC) Permitissued by EPA for Public CommentUnderground Injection Control PermitPG&E is exploring this project in 3 primary phases:
Phase 1 : Reservoir feasibility including site control, reservoir performance,economic viability, and environmental impacts.Phase 2: Commercial plant engineering, procurement and construction,and commissioning.Phase 3: Operations monitoring and technology transfer.Only Phase 1 is currently funded. It is envisioned that at the end of this firstphase, in the 4th Quarter of 2015, a decision will be made on whether tomove forward with seeking the required regulatory approvals to proceed toPhase 2. This decision will be based on information learned in Phase 1.The commercial-scale project has a nominal output capacity of 300megawatts (MW) -similar to a mid-sized power plant- for up to ten hours. Itis estimated that a commercial plant could come on-line in the 2020-2021http://www.pge.com/en/about/environment/pge/cleanenergy/caes/index.page11/10/2015 Compressed Air Energy Storage (CAES)Page 2 of 2timeframe.About CAES TechnologyThe proposed CAES project would use excess off-peak energy tocompress air and inject it into a depleted natural gas reservoir and then usethe compressed air to power a generator during peak periods when theenergy is needed most.California's Renewables Portfolio Standard (RPS) will require PG&E, otherinvestor-owned utilities, electric service providers, and community choiceaggregators to increase procurement from eligible renewable energyresources to 33% of total procurement by 2020. Storage facilities such asCAES will be increasingly important in managing the intermittent nature ofrenewable energy. Energy storage will contribute to grid efficiency,reliability, and sustainability by compensating for the fluctuations in poweroutput from intermittent renewable generators, like wind and solar plants. Itwill also reduce the need to build additional fossil-fueled generation byshifting excess generation in periods of low demand to peak demandperiods.Contact the CAES TeamEmailCAES@pge.comMailing AddressCompressed Air Energy StoragePacific Gas & Electric Company245 Market Street, Mail Code N13WSan Francisco, CA 94105ou In dark?s= p=About Careers Contact Us Privacy Newsroom Regulation Accessibility"PG&E' refers to Pacific Gas and Electric Company, a subsidiary of PG&E Corporation. © 2015 Pacific Gas and Electric Company. All rights reserved.VEI1Nhttp://www.pge.comlenlabout/environmentlpge/cleanenergy/caes/index.page 11/10/201511/10/2015 FOR IMMEDIATE RELEASE: July 28, 2011Contacts:Robert Schulte Phone: (612) 804-5363Executive Director, ISEPA e-mail: rhs@schulteassociates.comIowa Stored Energy Park Project TerminatedDes Moines, IA -July 28, 2011The Iowa Stored Energy Plant Agency (ISEPA, www.isepa.com) announced today thetermination of the Iowa Stored Energy Park (ISEP) project.ISEP is a proposed 270 Megawatt (MVV), $400 million compressed air energy storage(CAES) electric generation facility to be located at Dallas Center, Iowa, near DesMoines. In-service was planned for 2015. The project would take renewable wind andother resources available on the electric transmission grid during off-peak weeknightand weekend hours when customer electric loads and prices are low, and store theenergy as compressed air in a unique sandstone aquifer geologic structure 3000 feetunderground. Then, the compressed air would be used during higher value, on-peakhours on weekdays to generate electricity and deliver it back to the grid whencustomers need it most.As an illustration, ISEP's 270 MW generation output would be more than twice the sizeof the electric load of downtown Des Moines on a hot summer day. Its 220 MWcompression (storage) cycle could absorb the output of 100 to 150 large wind machineswhen operating at their rated wind speed."We have learned a great deal about what it takes to do utility-scale energy storage andcoordinate it with regional wind energy resources", said Bob Schulte of SchulteAssociates LLC, a management consulting firm retained by ISEPA to lead the duediligence assessment of the project. "The economic studies performed for ISEP showthat an innovative CAES project like this can be cost-effective compared to conventionalgeneration alternatives, and supportive of additional wind energy development in theregion.However, the geology studies of the project site show the storage reservoir is notsuitable for the scale of project that was envisioned. While the CAES project conceptand potential long-term economics are sound, due to geology limitations specific to theDallas Center site the ISEPA members have easier, less expensive and less riskyconventional alternatives to meet their customers' future electric needs."317 SIXTH AVENUE / SUITE 950 / DES MOINES, IA 50309-4128PHONE (515) 243-3122, www.isepa.comn The Iowa Stored Energy Plant Agency (ISEPA) owns ISEP. ISEPA represents 57municipal electric utilities located in Iowa, Minnesota and the Dakotas. The decisiontoday affects only the ISEP project. The ISEPA members will determine at a later datewhether they will consider or pursue additional projects.About $8.6 million (or about 2% of the total projected project cost, if completed) hasbeen spent on the ISEP project to-date. In addition to investments by the ISEPAmembers, the majority of ISEP development funding has been provided by the U.S.Department of Energy's (DOE) Energy Storage Program, with additional support fromthe Iowa Power Fund.2 Princeton Environmental InstitutePRINCETON UNIVERSITYEnergy Systems Analysis GroupCompressed Air Energy Storage: Theory, Resources,And Applications For Wind Power8 April 2008Samir Succar and Robert H. Williams AcknowledgmentsThe authors wish to thank all those who contributed in drafting and producing this report.In particular, we would like to gratefully acknowledge the following personsAlfred Cavallo, Charles Christopher, Paul Denholm (NREL), David Denkenberger (CUBoulder), Jeffery Greenblatt (Google), Gardiner Hill (BP), Kent Hoist (ISEP), GrahamHowes (BP), Aleks Kalinowski (Geoscience Australia), Vello A. Kuuskraa (AdvancedResources International Inc), Septimus van der Linden (Brulin Associates LLC), James P.Lyons (GE), James Mason (Hydrogen Research Institute), Michael J. McGill (Electricityand Air Storage Enterprises LLC), Robert B. Schainker (EPRI), Robert H. Socolow(Princeton University) and Ian R. VannGenerous financial support from BP and the William & Flora Hewlett Foundation2 Compressed Air Energy Storage, Succar and Williams April 2008Table of ContentsACKNOWLEDGMENTS ............................................................................................... 2TABLE OF FIGURES ..................................................................................... 5PREFACE .................................................................................................................. 6EXECUTIVE SUMMARY......................................................................................... 71. BACKGROUND ..................................................................................... 11.1. EVOLVING MOTIVATIONS FOR BULK ENERGY STORAGE................................................... 121.2. CAES OPERATION ................................................................................................ 151.3. SUITABLE GEOLOGIES FOR CAES ........................................................................... 171.3.1. Salt......................................................................................................... 181.3.2. Hard Rock................................................................................................. 191.3.3. Porous Rock.............................................................................................. 211.4. EXISTING AND PROPOSED CABS PLANTS .................................................................... 221.4.1. Huntorf.................................................................................................... 221.4. 2. Mclntosh.................................................................................................. 231.4.3. Norton ..................................................................................................... 241.4. 4. Iowa Stored Energy Park............................................................................... 251.4.5. Proposed Systems in Texas............................................................................. 262. CAES OPERATION AND PERFORMANCE............................................................ 272.1!. RAMPING, SWITCHING AND PART-LOAD OPERATION....................................................... 272.2. CONSTANT VOLUME AND CONSTANT PRESSURE.......................................................... 292.3. STORAGE VOLUME REQUIREMENT............................................................................. 302.3.1. Case 1; Constant Cavern Pressure ................................................................... 322.3.2. Case 2: Variable Cavern Pressure, Variable Turbine Inlet Pressure ........................... 322.3.3. Case 3: Variable Cavern Pressure, Constant Turbine In let Pressure ........................... 332. 3.4. Cavern Size............................................................................................... 332.4. PERFORMANCE INDICES FOR CAES SYSTEMS................................................................ 362.4.1. Heat Rate ................................................................................................. 372.4.2. Charging Electricity Ratio ............................................................................. 372.4.3. Toward a Single CAES Performance Index.......................................................... 372.4.3.1. Primary Energy Efficiency .................................................................................... 382.4.3.2. Round Trip Efficiency ......................................................................................... 382.4.3.3. Additional Approaches......................................................................................... 393. AQUIFER CAES GEOLOGY AND OPERATION........................................................ 423.1. MOTIVATIONS ...................................................................................................... 423.2. APPLICABILITY OF INDUSTRIAL FLUID STORAGE EXPERIENCE ........................................... 443.2.1. CO2 Storage .............................................................................................. 443.2.2. Natural Gas Storage .................................................................................... 443.2.2.1. Site Characterization and Bubble Development ........................................................... 443.2.2.2. Applicability to CAES......................................................................................... 463.2.2.3. Differences....................................................................................................... 473.3. GEOLOGIC REQUIREMENTS...................................................................................... 473.3.1. Porosity, Permeability and Thickness ................................................................ 483.3.2. Reservoir Dimensions................................................................................... 493.3.3. Pressure Limits and Caprock Characteristics........................................................53.3.4. Residual Hydrocarbons ................................................................................ 523.4. OXIDATION CONSIDERATIONS................................................................................... 533.5. CORROSION......................................................................................................... 543.6. FLOW INAQUIFERS................................................................................................ 563.7. PARTICULATES ..................................................................................................... 573 Compressed Air Energy Storage, Succar and Williams Arl20April 20084. WINDICAES SYSTEMS IN BASELOAD POWER MARKETS....................................... 584.1. METHODOLOGY .................................................................................................... 584.2. GENERATION COSTS FOR ALTERNATIVE BASELOAD POWER SYSTEMS OPERATED AT SPECIFIEDCAPACITY FACTORS........................................................................................................614. 2.1. Dispatch Competition in Base/cad Power Markets ................................................ 624. 2.2. Dispatch Duration Curves..................................................................... 635. ADVANCED TECHNOLOGY OPTIONS .................................................................. 666. A WAY FORWARD .............................................................-.................. 697. CONCLUSIONS................................................................................................. 718. REFERENCES .............................................................................................. 73APPENDIX A THEORETICAL EFFICIENCY OF COMPRESSED AIR ENERGY STORAGEFOR ALTERNATIVE CONFIGURATIONS................................................................. 804 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Table of FiguresFIGURE 1 AREAS WITH GEOLOGIES FAVORABLE FOR CAES AND CLASS 4+ WINDS ..................................8FIGURE lIITHE WIND/CAES SYSTEM SCHEDULED TO BEGIN OPERATION IN 2011 NEAR DES MOINES, IOWA(IAMU, 2006) ........................................................................................................ 9FIGURE 1 GLOBAL WIND CAPACITY 1995-2007 (GWEC, 2008) ................................................... 12FIGURE 2 ONSHORE WIND RESOURCES AND POPULATION DENSITY IN THE CONTINENTAL US (US CENSUS2000; NREL, 2001, 2002, 2006) ................................................................................. 14FIGURE 3 CAES SYSTEM CONFIGURATION ............................................................................... 15FIGURE 4 AREAS CLASSIFIED FOR SUBSURFACE STORAGE OF FLUIDS. FROM THE NATIONAL PETROLEUMCOUNCIL REPORT OF THE COMMITTEE ON UNDERGROUND STORAGE FOR PETROLEUM, APRIL 22,1952; UPDATED IN OCT 1962 B C.T. BRANDT, UNDERGROUND STORAGE AND MINING CONSULTANT,BARTESVILLE, OK; ADDITIONAL CHANGES REFLECT COMMENTS IN KATZ AND LADY, 1976 ........... 17FIGURE 5 STRUCTURE OF HUNTORF CABS PLANT SALT DOME STORAGE WITH CAVERNS AND PLANT ON SAMESCALE [25]............................................................................................................... 18FIGURE 6 COINCIDENCE OF HIGH WIND POTENTIAL AND SALT DOMES IN EUROPE. RED CIRCLES INDICATEAREAS INVESTIGATED FOR CABS DEVELOPMENT [28]........................................................ 19FIGURE 7 AREAS WITH GEOLOGIES SUITABLE FOR MINED STORAGE (RED) AND HIGH-QUALITY WINDRESOURCES (BLUE) [23, 3 1-33]................................................................................... 20FIGURE 8 AERIAL VIEW OF THE HUNTORF CABS PLANT [25]....................................................... 22FIGURE 9 HUNTORF MACHINE HALL [40] .............................................................................. 23FIGURE 10 MCINTOSH CABS SYSTEM COMPRESSOR TRAIN (LEFT) AND COMBUSTION TURBINE (RIGHT) ....24FIGURE 11 A RENDERING OF THE PROPOSED 2700 MW CABS PLANT BASED ON AN ABANDONED LIMESTONEMINE IN NORTON, OH [45]........................................................................................ 25FIGURE 12 DIAGRAM OF THE IOWA STORED ENERGY PARK [48] ................................................... 25FIGURE 13 TURBINE PERFORMANCE CHARACTERISTICS FOR AQUIFER CABS BASED ON EPRI DESIGN FORMEDIA, ILLINOIS SITE [50] ........................................................................................ 27FIGURE 14 CONSTANT PRESSURE CABS RESERVOIR WITH COMPENSATING WATER COLUMN. (1) EXHAUST (2)CABS PLANT (3) SURFACE POND (4) STORED AIR (5) WATER COLUMN [41]............................. 29FIGURE 15 THE ENERGY PRODUCED PER UNIT VOLUME FOR CABS WIThf CONSTANT PRESSURE RESERVOIR(CASE 1), VARIABLE PRESSURE RESERVOIR (CASE 2) AND VARIABLE PRESSURE RESERVOIR WITHCONSTANT TURBINE INLET PRESSURE (CASE 3). INSET SHOWS THROTTLING LOSSES ASSOCIATED WITHCASE 3 RELATIVE TO THE VARIABLE INLET PRESSURE SCENARIO (CASE 2)- FIGURE FROM [22]........34FIGURE 16 THE RATIO OF STORAGE ENERGY DENSITY BETWEEN A CONSTANT VOLUME CABS SYSTEM WITHCONSTANT TURBINE INLET PRESSURE (CASE 3) AND A PRESSURE COMPENSATED CABS RESERVOIR(CASE 1 ) AS A FUNCTION OF THE RATIO BETWEEN THE OPERATING PRESSURES OF THE CASE 3 SYSTEM(Ps2/Ps1) .................................................................................................................. 35FIGURE 17 A COMPARISON OF AREAS OF HIGH QUALITY WIND RESOURCES AND GEOLOGY COMPATIBLE WITHCABS (AREAS SUITABLE FOR MINED ROCK CAVERNS OMITTED DUE TO THE HIGH ESTIMATED COST OFDEVELOPING SUCH FORMATIONS FOR CABS) [23, 27, 31-33]. LOCATIONS OF THE EXISTING MCINTOSHCABS PLANT, THE RECENTLY ANNOUNCED DALLAS CENTER WIND/CABS SYSTEM AND THE PROPOSEDMATAGORDA PLANT ARE INDICATED AS WELL.................................................................... 42FIGURE 18 AQUIFER DIMENSIONS RELEVANT TO TOTAL CLOSURE RATING [50] .................................. 48FIGURE 19 VISCOSITY AND GAS DEVIATION FACTOR OF AIR VERSUS NATURAL GAS [69] ..................... 49FIGURE 20 POROUS ROCK CABS STORAGE VOLUME [13]........................................................... 50FIGURE 21 MEASUREMENTS OF THRESHOLD PRESSURE AS A FUNCTION OF PERMEABILITY [ 13 ]................ 51FIGURE 22 THIS PHOTOGRAPH, FROM THE HUNTORF CABS FACILITY IN GERMANY, SHOWS WHERE THEPROTECTIVE FIBERGLASS-REINFORCED PLASTIC TUBING FRACTURED. [25] ................................ 55FIGURE 23 DISPATCH COSTS FOR THE FOUR ALTERNATIVE POWER SYSTEMS FOR TWO VALUATIONS OF GHGEMISSIONS................................................................................................................ 64FIGURE 24 A POSSIBLE TECHNICAL CONCEPT FOR AN AA-CAES SYSTEM UNDER DEVELOPMENT [28]......665 Compressed Air Energy Storage, Succar and Williams Arl20April 2008PrefaceThis report reviews the literature on compressed air energy storage (CAES) andsynthesizes the information in the context of electricity production for a carbonconstrained world.CABS has historically been used for grid management applications such as load shiftingand regulation control. Although this continues to be the dominant near-term marketopportunity for CABS, future climate policies may present a new application: theproduction of baseload electricity from wind turbine arrays coupled to CABS.Previous studies on the combination of wind and CABS have focused on economics andemissions [1-10]. This report highlights these aspects of baseload wind/CABS systems,but focuses on the technical and geologic requirements for widespread deployment ofCABS, with special attention to relevant geologies in wind-rich regions of NorthAmerica.Large penetrations of wind/CABS could make substantial contributions in providingelectricity with near-zero GHG emissions if several issues can be adequately addressed.Drawing on the results of previous field tests and feasibility studies as well as the existingliterature on energy storage and CABS, this report outlines these issues and frames theneed for further studies to provide the basis for estimating the true potential ofwind/CAES.6 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Executive SummaryCompressed Air Energy Storage (CAES) is a commercial, utility-scale technologysuitable for providing long-duration energy storage with fast ramp rates and good part-load operation. CAES works by using electricity to compress air, which is subsequentlystored in a large reservoir (typically in an underground geologic formation). Electricity isregenerated by recovering compressed air from storage, burning in this air a smallamount of fuel (typically natural gas), and expanding the combustion products through aturbine (see section 1.2, page 15).This report is intended to analyze the potential of CAES for balancing large penetrationsof wind energy. The economic analysis of wind coupled with CAES for providingbaseload power indicates that it will likely be competitive in economic dispatch under thesame range of greenhouse gas (GHG) emissions price needed to make carbon capture andstorage (CCS) economic for new coal integrated gasification combined cycle (IGCC)systems (~-$30/tCO2). However the potential for wind/CABS is contingent on theavailability of geologies suitable for CAES in windy regions. Thus the central focus ofthe report is on the geologic and technical requirements for CABS as they relate to thepotential for large-scale deployment of this technology.The CAES storage reservoir can often be constructed in pre-existing formations (e.g. asalt cavern, aquifer or abandoned mine). As a result, the capital cost of adding anincremental amount of storage capacity can be much lower than for other comparablestorage technologies. This makes CAES especially well suited for bulk storageapplications.The total capital cost of a CAES unit tends to be dominated by the cost of theturbomachinery. The low total capital cost can be understood by noting that theturbomachinery is essentially a gas turbine for which the compression and expansionfunctions are separated in time--and gas turbines are characterized by relatively lowcapital costs.In the 1 970s, CAES began to attract attention as a way to store inexpensive baseloadpower produced during off-peak periods for use later when demand is higher andelectricity is more valuable.Shifts in market conditions led to diminished interest in CABS. However, the sustainedrapid growth of wind power has catalyzed a renewed interest in this technology as anoption for making wind power dispatchable (see section 1.1, page 12). Additionally,because CABS consumes significantly less fuel than a conventional gas turbine per unitof energy delivered, the GHG emissions from wind/CABS systems can be quite low.Although the global wind resource can theoretically satisfy the demand for electricityseveral times over, the variability of wind and the typical remoteness of high-qualitywind resources from major electricity demand centers (e.g. in the U.S.) must beaddressed for wind to serve a large percentage of electricity consumption (>20-30%).CABS offers the potential for overcoming these challenges by both smoothing the outputfrom wind and enabling the cost-effective operation of high capacity, high-voltagetransmission lines carrying this power at high capacity factors.7 Compressed Air Energy Storage, Succar and Williams Arl20April 2008The ultimate potential of wind in satisfying electricity needs via wind/CAES depends onthe availability of geologies suitable for CABS in regions with high-quality windresources (for a description of geologic options for CABS reservoirs see section 1.3, page17). In the continental US, high-quality wind resources overlap more closely with porousrock geology than any other storage geology (see Figure i). Thus, in this region at least,widespread deployment of CABS in connection with wind power implies a considerablerole for aquifers.Figure iAreas with geologies favorable for CAES and class 4+ winds (see Section 3,"Aquifer CABS Geology and Operation" on page 42)Although two commercial CABS plants have been built, neither uses aquifers as thestorage reservoir (see section 1.4 "Existing and Proposed CABS Plants" on page 22).However, previous studies and field tests have confirmed that air can be successfullystored and withdrawn using a saline aquifer as a storage reservoir. Furthermore, arecently announced wind/CABS plant in Iowa will use an aquifer [a porous sandstoneformation (see Figure ii)]. Once built, this project will provide important informationabout these systems in terms of both the utilization of aquifers for air storage andcoupling of CABS to wind. The system is being designed to enable wind power to bedispatched in electric load-following transmission support applications, which is likely tobe the most important near-term use of wind/CABS systems.Although there has been no commercial experience with aquifer CAES, much can begleaned from what is already known about natural gas storage in aquifers. The natural gasstorage industry has vast experience with porous rock formations under conditionssimilar to those for CABS (see section 3.2.2, page 44). As such, the theory of natural gasstorage provides a useful point of departure for understanding CABS, and many of themethodologies and data amassed for identifying natural gas storage opportunities maywell prove useful for assessing CABS sites.8 Compressed Air Energy Storage, Succar and Williams Apis20April 2008Relative to methane however, air has both different physical properties (e.g., air has ahigher viscosity than methane) and different chemical properties (e.g., introducingoxygen underground can lead to various oxidation reactions, corrosion mechanisms, andthe promotion of bacteria) that could pose challenges for air storage (see sections 3.4 and3.5 on page 53). While it is expected to be often feasible to mitigate the effects of thesefactors, it will be essential to test the viability of aquifer CAES under a wide variety ofgeologic conditions and to carefully determine the impact of local geology on CAESsystem planning and design.The use of CAES in an intermediate load application such as that envisioned for the Iowawind/CABS plant will provide a valuable demonstration of wind/CABS integration.However, demonstration of much more closely coupled systems capable of servingbaseload power markets is also needed to understand better the potential of wind/CABS,because although bulk storage may be valuable for serving a broad range of gridmanagement applications, ultimately the role of wind as a tool for climate changemitigation will depend on the extent to which it will be able to supplant new baseloadcoal-fired capacity.Figure ii The wind/CABS system scheduled to begin operationin 2011 near Des Moines, Iowa (IAMU, 2006)A dispatch cost analysis suggests that a natural gas-fired wind/CABS system would oftenbe able to compete against coal and other baseload power options, especially under aclimate change mitigation policy sufficiently stringent to make CO2 capture and storagecost-effective for coal power (see section 4, "Wind/CABS Systems in Baseload PowerMarkets" on page 58). Thus, the wind/CABS hybrid could give both wind and natural gasentry into baseload markets in which they would otherwise not be able to compete.9 Compressed Air Energy Storage, Succar and Williams Arl20April 2008The storage capacity of CAES systems designed to deliver baseload power wouldtypically be several times that for other grid management applications, but even so the"footprint" of a 1 0-in thick aquifer capable of providing baseload wind/CAES powerwould occupy a much smaller ('-14%) land area than that of the corresponding wind farmunder typical conditions (see section 2.3 on page 30).A better understanding is needed of the performance of CABS over a wide range ofconditions. In particular, use of CABS for wind balancing will require CAES to adjustoutput more frequently and to switch between compression and generation modes morequickly than has been required of CABS in applications such as storing off-peak power atnight and generating peak electricity during the day (see section 2.1, page 27).Understanding well the impacts of these operational demands requires further study.Determining the ultimate potential of baseload wind/CABS as a climate changemitigation option also requires knowledge of the prevalence of suitable geologies.Although porous rock formations seem to be prevalent in high wind areas, understandingthe full potential of this technology will require in-depth assessments of the extent offormations with anticlines suitable for containment and, for promising structures, theirgeochemical and geophysical suitability for CABS. Data on local geology from US andstate geological surveys including natural gas storage candidate site evaluations might aidin further characterizing these areas, but new data will also be needed, especially inregions where natural gas storage is not commonplace (see section 3.3 "GeologicRequirements" on page 47).CABS appears to have many of the characteristics necessary to transform wind into amainstay of global electricity generation. The storage of energy through air compressionmay enable wind to meet a large fraction of the world's electricity needs competitively ina carbon constrained world. If the needed steps are taken soon, it should quickly becomeevident just how large this fraction might be.10 Comnpressed Air Energy Storage, Succar and Williams Arl20April 20081. BackgroundCompressed Air Energy Storage (CABS) is a low cost technology for storing largequantities of electrical energy in the form of high-pressure air. It is one of the few energystorage technologies suitable for long duration (tens of hours), utility scale (100's to1000's of MW) applications. Several other energy storage technologies such as flywheelsand ultracapacitors have the capability to provide short duration services related to powerquality and stabilization but are not cost effective options for load shifting and windgeneration support [I11, 12].The only technologies capable of delivering several hours of output at a plant-level poweroutput scale at attractive system costs are CAES and pumped hydroelectric storage (PHS)[ 13-17]. Although some emerging battery technologies may provide wind-balancingservices as well, typical system capacities and storage sizes are an order of magnitudesmaller than CABS and PHS systems (--10 MW, <10 hours) with significantly highercapital costs (see Table 1).PHS does not require fuel combustion and has a greater degree of field experiencerelative to CAES, but it is only economically viable on sites where reservoirs atdifferential elevations are available or can be constructed. Furthermore, theenvironmental impact of large-scale PHS facilities is becoming more of an issue,especially where preexisting reservoirs are not available and sites with large, naturallyoccurring reservoirs at large differential elevations where environmentally benign,inexpensive PHS can be built are increasingly rare.Table 1 Capital Costs for Energy Storage Optionis [11, 12, 18]Technology Capital Cost: Capital Cost: Hours of Total CapitalCapacity ($/kW) Energy (S/kWh) Storaee Cost ($/kW)CAES (300MW) 580 1.75 40 650Pumped Hydroelectric 600 37.5 10 975(1,000MW) _______Sodium Sulfur Battery 1720-1860 180-210 6-9 3100-3400(10MW) _______ _______ _______Vanadium Redox Battery 2410-2550 240-340 5-8 4300-4500(10MW) _______ _______ _______In contrast, CABS can use a broad range of reservoirs for air storage and has a moremodest surface footprint giving it greater siting flexibility relative to PHS. High-pressureair can be stored in surface piping, but for large-scale applications, developing a storagereservoir in an underground geologic formation such as solution mined salt, salineaquifer, abandoned mine, or mined hard rock are typically more cost effective. Thewidespread availability of geologies suitable for CABS in the continental US suggeststhat this technology faces far fewer siting constraints than PHS, which is especiallyimportant for the prospect of deploying CAES for wind balancing.One of the central applications for CABS is for the storage of wind energy during timesof transmission curtailment and generation onto the grid during times of shortfalls inwind output. Such wind balancing applications require not only large-scale, long duration11 Compressed Air Energy Storage, Succar and Williams Arl20April 2008storage, but also fast output response times and siting availability in wind-rich regions.Prior studies indicate that suitable CAES geologies are widely available in the wind-richUS Great Plains. Furthermore, CAES is able to ramp output quickly and operateefficiently under partial load conditions making it well suited to balance the fluctuationsin wind energy output. Finally, the low greenhouse gas (GHG) emissions rate of CAESmakes it a good candidate for balancing wind in a carbon constrained world.Among the geologic options for air storage, porous rock formations offer the mostwidespread availability and potentially the lowest cost. Moreover, geographicaldistributions of aquifers and good wind resources are strongly correlated in the US.Therefore the potential for CABS to play a major role in balancing wind output andproducing low greenhouse gas (GHG) emitting power will depend to a large degree onthe availability of aquifer structures suitable for CAES.1.1. Evolving Motivations for Bulk Energy StorageCABS emerged in the 1970s as a promising peak shaving option [19]. High oil pricestogether with an expanding nuclear power industry sparked an interest in energy storagetechnologies such as CABS to be used in load following applications. The high price ofpeak power and the perceived potential for inexpensive baseload nuclear power madeattractive the option of storing inexpensive off-peak electricity and selling this electricityduring peak demand periods [20, 21 ].I 00,000 .40%9 0 ,0 0 0 ..... ...... ... .......... .... .. ... .... ......... .. ...3 5 %8i 0,000 .... .. ... .,070,000 2.. ...%. .50,000 210% 19519 97 9819 00 01201 00 0520 0%'MW4,800 6,100 7,600 10.200 13,600 17,40023,900 31.100 39.431 47,620I 59.09l 74,133 94.,112Figure 1 Global Wind Capacity 1995-2007 (GWEC, 2008)These conditions initially fueled a strong interest in CAES among many utilities, but asthe nuclear power industry lost momentum and oil prices retreated from their peaks, themarket conditions for CABS began to change. During the 1 980s the gas turbine andcombined cycle generation emerged as the leading low cost options for peaking and load-12 Compressed Air Energy Storage, Succar and Williams Arl20April 2008following markets. This together with overbuilt generating capacity on the grid and theperception that domestic natural gas supplies were abundant led to erosion of marketinterest in energy storage.Recent trends in gas price and wind power development have fostered new interest inenergy storage, not as a way to convert baseload power into peak power, but as a waymitigate the variability of wind energy [8, 10]. Global wind power capacity has grownrapidly in recent years from 4.8 GW in 1995 to 94 GW by the end of 2007 (see Figure 1).The variability of wind output requires additional standby reserve capacity to ensureoutput during times of peak demand. Gas turbines can respond quickly to shortfalls inwind output and so gas fired spinning reserve units are good candidates for dispatch tomeet the challenge of balancing this growing wind segment of the power mix.Energy storage represents an alternative wind balancing strategy, and the low fuelconsumption of CABS makes it especially relevant in the face of high gas prices.Although wind balancing has long been acknowledged as a potential application for bulkenergy storage [22], it is only recently that wind penetrations have reached levels thatrequire additional balancing measures for maintaining system stability [23]. Howeverrecent studies have shown that bulk storage can reduce the integration costs for windenergy even at relatively low penetration levels [24].I The use of storage for balancingwind and for serving other grid management applications will be especially valuablewhere the supply of flexible generating capacity (e.g. hydroelectric) is limited [10, 25].The continued increase of wind penetration on the grid and the need to reduce greenhousegas emissions may create an incentive to use storage systems directly coupled with windto produce baseload power rather than as independent entities to provide grid supportservices (see below). Further, because the fuel consumption of CAES is less than half ofthat of a simple cycle gas turbine, using CAES would provide a hedge against natural gasprice volatility [26].A further reason for considering wind farms coupled to CABS storage (henceforthreferred to as wind/CABS) stems from the fact that most high quality onshore windresources are often remote from load centers. The exploitable onshore wind potential inclasses 4 and above in North America is huge--equivalent to more than 12 times totalelectricity generation in 2004 [27, 28].2,3 However the resources in the US areconcentrated in the sparsely populated Great Plains and Midwest States (see Figure 2)which account for over half of the exploitable US wind generation potential in class 4+[29]. Bringing electricity cost-effectively from the Great Plains to major urban electricity'The cited report indicates that removal of bulk storage (pumped hydroelectric storage in this case)increases integration costs for wind by approximately 50% for a wind penetration level of 10%. Also,doubling of storage capacity lowered integration cost by -$1 .34/MWh in the 20% penetration case.2 The Greenblatt (2005) estimate is based on the assumption that various land use constraints limit thetechnical potential for wind to what can be produced on 50% of the land on which class 4+ wind resourcesare available.SThe technical wind power potential at the global level is also huge. Considering only class 4+ windsexploited on 50% of the land on which these resources are available, as in the North American case,Greenblatt (2007) estimated that the global technical wind energy potential is 185,000 TWh/y on land plus49,400 TWh.year offshore. For comparision the global electricity generation rate in 2004 was 17,400TWh/year.13 Compressed Air Energy Storage, Succar and Williams Arl20April 2008demand centers requires that it be transmitted via GW-scale high-voltage transmissionlines that are baseloaded. CAES systems coupled to multi-GW-scale wind farms couldprovide such baseload power.As will be shown, wind/CABS systems have good prospects of being able to compete in acarbon constrained world directly with other low carbon baseload power options such asthe coal integrated gasification combined cycle (IGCC) with carbon capture and storage(CCS) (see Section 4).Because the incremental capital cost for increasing CABS storage volume capacity isrelatively low, it is well suited for providing long-duration storage (>80 hours) needed toproduce baseload power. Although seasonal storage of wind is also possible, it wouldrequire much larger storage volumes [30].Wind/CAES also gives natural gas a role in baseload power markets that are often out ofreach due to the relatively high dispatch costs of natural gas generation. Thus,wind/CABS gives both wind and natural gas an entry into large baseload power marketsto which they would not otherwise have access.While typical capacity factors for wind farms are approximately 30-40% [31 ],wind/CABS systems can achieve capacity factors4 of 80-90% typical of base load plants.4~ Capacity factor in this case is on the basis of a constant demand level. The rated capacity of the wind parkwill be "oversized" relative to this demand level and the CAES turboexpander capacity matched to it such14 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Therefore, the coupling of wind to energy storage enhances utilization of both existingtransmission lines and dedicated new lines for wind. This can alleviate transmissionbottlenecks or obviate transmission additions and upgrades.In the case that transmission capacity is limited, it will be advantageous to site the storagereservoir and wind turbine array as closely as possible to exploit the benefits describedabove. If this is not the case however, there is no need to co-locate the storage system andwind array. Independently siting these components would allow added flexibility forsimultaneously matching facilities to the ideal wind resource, storage reservoir geologyand the required natural gas supplies.1.2. CAES OperationCAES systems operate much in the same way as a conventional gas turbine except thatcompression and expansion operations occur independently and at different times (seeFigure 3). Because compression energy is supplied separately, the full output of theturbine can be used to generate electricity during expansion, whereas conventional gasturbines typically use two thirds of the output power from the expansion stage to run thecompressor.CABS systemCompressor train Expander/generator trainAir EhasPc = Compressor Fueleupraopower in Ful(e.g. natural gas, distillate)PG= Generator ---[ ---power outI 06 asalt cavern, a P;or hard rock(aP)Figure 3 CAES System Configurationthat excess wind can be stored to balance subsequent shortfalls. While it is possible to produce constantoutput (i.e. 100% capacity factor) from a wind/CAES plant, it would require a significantly larger storagevolume capacity.15 Compressed Air Energy Storage, Succar and Williams Arl20April 2008During the compression (storage) mode operation, electricity is used to run a chain ofcompressors that inject air into an uninsulated storage reservoir, thus storing the air athigh pressure and at the temperature of the surrounding formation. The compressionchain makes use of intercoolers and an aftercooler to reduce the temperature of theinjected air thereby enhancing the compression efficiency, reducing the storage volumerequirement and minimizing thermal stress on the storage volume walls.Despite the loss of heat via compression chain intercoolers, the theoretical efficiency forstorage at formation temperatures in a system with a large number of compressor stagesand intercooling can approach that for a system with adiabatic compression and airstorage in an insulated cavern (see the discussion of compression efficiency in AppendixA).5During the expansion (generation) operation mode, air is withdrawn from storage andfuel (typically natural gas) is combusted in the pressurized air. The combustion productsare then expanded (typically in two stages), thus re-generating electricityFuel is combusted during generation for capacity, efficiency and operationalconsiderations. Expanding air at the wall temperature of the reservoir would necessitatemuch higher air flow in order to achieve the same turbine output -thus increasing thecompressor energy input requirements to the extent that the charging energy ratio wouldbe reduced by roughly a factor of four [32]. Furthermore, in the absence of fuelcombustion the low temperatures at the turbine outlet6 would pose a significant icing riskfor the blades because of the large airflow through the turbine, despite the small specificmoisture content for air at high pressure. There is also the possibility that the turbinematerials and seals might become brittle during low temperature operation.5Adiabatic CAES designs capture the heat of compression in thermal energy storage units (see discussionof AA-CAES in section 5, Advanced Technology Options)6 For example assuming air recovered from storage at 20°C, adiabatic expansion, and a 45x compressionratio, T=-174°C at the turbine exhaust16 Compressed Air Energy Storage, Succar and Williams Arl20April 2008V~mOu I~cK5U ~~ ~ we Pms woca.Figure 4 Areas classified for subsurface storage of fluids. From the National Petroleum Council Reportof the Committee on Underground Storage for Petroleum, April 22, 1952; updated in Oct 1962 b C.T.Brandt, Underground Storage and Mining Consultant, Bartesville, OK; additional changes reflectcomments in Katz and Lady, 1976.1.3. Suitable Geologies for CAESGeologies suitable for GALS storage reservoirs can be classified into three categories:salt, hard rock, and porous rock. Taken together, the areas that have these geologiesaccount for a significant fraction of the continental United States (see Figure 4). Priorstudies indicate that over 75% of the U.S. has geologic conditions that are potentiallyfavorable for underground air storage [33, 34].However, those studies carried out only macro scale analyses that did not evaluate areasaccording to the detailed characteristics necessary to fully estimate their suitability forCAES. While the large fractions of land possessing favorable geologies is encouraging,broad surveys such as the data presented in Figure 4 can only serve as a template foridentifying candidate areas for further inquiry and detailed regional and site-specific datawill be necessary to determine the true geologic resource base for CAES.17 Compressed Air Energy Storage, Succar and Williams Arl20April 20081.3.1. SaltThe two CABS plants currently operating use solution-mined cavities in salt domes astheir storage reservoirs (see Figure 5 and section 1.4 "Existing and Proposed CABSPlants"). In many ways, such formations are the most straightforward to develop andoperate. Solution mining techniques can provide a reliable, low cost route for developinga storage volume of the needed size (typically at a storage capital cost of -$2.00 perkWh of output from storage) if an adequate supply of fresh water is available and if theresulting brine can be disposed of easily [11, 12]. Furthermore, due to the elasto-plasticproperties of salt, storage reservoirs solution-mined from salt pose minimal risk of airleakage [33, 36].Large bedded salt deposits are available in areas of the Central, North Central and NorthEast United States while domal formations can be found in the Gulf Coast Basin [37].Although both bedded and domal formations can be used for CABS, salt beds are oftenmore challenging to develop if large storage volumes are required. Salt beds tend to bemuch thinner and often contain a comparatively higher concentration of impurities whichpresent significant challenges with respect to structural stability [37]. Caverns minedfrom salt domes can be tall and narrow with minimal roof spans as is the case at both theHuntorf (see Figure 5) and Mclntosh CAES facilities. The thinner salt beds cannotsupport long aspect ratio designs because the air pressure must support much larger roof18 Compressed Air Energy Storage, Succar and Williams Arl20April 2008spans. In addition, the presence of impurities might further compromise the structuralintegrity of the cavern and further complicate the development a large capacity storagesystem.Although the locations of domal formations in the United States are not well correlatedwith high quality wind resources (see Figure 17), there are some indications the prospectsmay be more favorable in Europe (see Figure 6).Figure 6 Coincidence of high wind potential and salt domes in Europe. Red circles indicateareas investigated for CAES development [38]1.3.2. Hard RockAlthough hard rock is an option for CABS, the cost of mining a new reservoir is oftenrelatively high (typically $30/kWh produced). However in some cases existing minesmight be used in which case the cost will typically be about $10/kWh produced [11, 39,40] as is the case for the proposed Norton CAES plant, which makes use of an idlelimestone mine (see section 1.4).19 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Detailed methodologies have been developed for assessing rock stability, leakage andenergy loss in rock-based CAES systems including concrete-lined tunnels [44-46].Several such systems have been proposed [47] and known field tests include two recentprograms in Japan: a 2 MW test system using a concrete-lined tunnel in the formerSunagaawa Coal Mine and a hydraulic confinement test performed in a tunnel in theformer Kamioka mine [11 ].In addition, a test facility was developed and tested by EPRI and the Luxembourg utilitySociete Electrique de l'Our SA using an excavated hard-rock cavern with watercompensation [48]. The site was used to determine the feasibility of such a system forCAES operation and to characterize and model water flow instabilities resulting from therelease of dissolved air in the upper portion of the water shaft (i.e. the "champagneeffect").Hard rock geologies suitable for CAES are widely available in the continental US andoverlap well with high-quality wind resources (see Figure 7). However, because thedevelopment costs are currently high relative to other geologies (especially given thelimited availability of preexisting caverns and abandoned mines [36]), it is unlikely thatthis option will be the first option pursued for a large-scale deployment of CABS.Although future developments in mining technology may reduce the costs of utilizingsuch geologies, it appears that other geologies may currently offer the best near-termopportunities for CABS development.n Chim4+ n Mined StorageFigure 7 Areas with geologies suitable for mined storage (red) and high-quality windresources (blue) [33, 41-43]20 Compressed Air Energy Storage, Succar and Williams Arl20April 20081.3.3. Porous RockAlso suitable for CABS are porous rock formations such as saline aquifers. Porousreservoirs have the potential to be the least costly storage option for large-scale CAESwith an estimated development cost of-S$0.11l/kWh for incremental storage volumeexpansion [11]. In addition, large, homogeneous aquifers potentially suitable for CAESoperation can be found throughout many areas of the central US. Because this areacoincides with areas of high quality wind (see Figure 17) and because of the limitedavailability and/or cost-effectiveness of other options, aquifer CABS will be especiallyrelevant to the discussion of energy storage for balancing wind. Despite its potential forlow cost development, utilization of an aquifer for CABS requires extensivecharacterization of a candidate site to determine its suitability (see section 3, "AquiferCABS").A 25 MW porous rock-based CABS test facility operated for several years in Sesta, Italy.Although the tests were successful, a geologic event disturbed the site which led toclosure of the facility [11]. In addition, EPRI and the U.S. Department of Bnergy haveconducted tests on porous sandstone formations in Pittsfield, Illinois to determine theirfeasibility for CAES (see section 3, "Aquifer CABS"). Construction of the firstcommercial CAES plant with a porous rock reservoir is scheduled to begin in DallasCenter, Iowa in 2009 (see section 1.4)21 Compressed Air Energy Storage, Succar and Williams Arl20April 20081.4. Existing and Proposed CAES Plants1.4.1. HuntorfThe 1Huntorf CAES plant, the world's first CABS facility, was completed in 1978 nearBremen, Germany (see Figure 9 and Figure 8). The 290 MW plant was designed andbuilt by ABB (formerly BBC) to provide black-start services7 to nuclear units near theNorth Sea and to provide inexpensive peak power. It has operated successfully for almostthree decades primarily as a peak shaving unit and to supplement other (hydroelectric)storage facilities on the system to fill the generation gap left by slow-respondingmedium-load coal plants. Availability and starting reliability for this unit are reported as90% and 99% respectively.Because the Huntorf plant was designed for peaking and black start applications, it wasinitially designed with a storage volume capable of two hours of rated output. The planthas since been operationally modified to provide up to three hours of storage and hasbeen used increasingly to help balance the rapidly growing wind output from NorthGermany [35, 49].cavern NK2power p~4tFigure 8 Aerial view of the Huntorf CABS plant [35]The underground portion of the plant consists of two salt caverns (310,000 m3 ttldesigned to operate between 48 and 66 bar. The air from the salt caverns was found tocause oxidation upstream of the gas turbine during the first year of operation, leading tothe installation of fiberglass reinforced plastic (FRP) tubing. Because the turbineexpanders are sensitive to salt in the combustion air, special measures were taken toensure acceptable conditions were met at the turbine inlet as well [35].7 Black start refers to the ability of a plant to start up during a complete grid outage. Because nuclear powerstations require some power to resume operation, the Huntorf CABS plant was built in part to provide thisstart up power.22 Compressed Air Energy Storage, Succar and Williams Arl20April 2008The compression and expansion sections draw 108 and 417 kg/s of air respectively andare each comprised of two stages. The first turbine stage expands air from 46 to 11 bar.Figure 9 Huntorf Machine Hall [50]Because gas turbine technology was not compatible with this pressure range, steamturbine technology was chosen for the high-pressure (hp) expansion stage. Due to theincrease in heat transfer coefficient at elevated pressure and temperature and to ensureproper cooling (and to control NOx emissions as well), the hp turbine inlet temperaturewas held to only 5500 C compared to 8250 C for the lp turbine (typical for a gas turbinewithout blade cooling). Moderate combustion inlet temperatures also facilitate the dailyturbine starts needed for CAES operation [50].Although the plant would be able to operate at a lower heat rate if equipped with heatrecuperators (so as to recover exhaust heat from the Ip turbine for preheating the gasentering the hp turbine), this addition was omitted in order to minimize system startuptime [51, 52].1.4.2. MclntoshAlthough high oil and gas prices through the early 1 980s continued to draw the attentionof utilities to CABS as a source for inexpensive peak power [47] it was not until a decadelater that a CABS facility began operating in the United States. The 110 MW Mclntoshplant was built by the Alabama Electric Cooperative on the Mclntosh salt dome insouthwestern Alabama and has been in operation since 1991 (see Figure 10). It wasdesigned for 26 hours of generation at full power and uses a single salt cavern (560,000in3) designed to operate between 45 and 74 bar.23 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Figure 10 Mclntosh CAES system compressor train (left) and combustionturbine (right)The project was developed by Dresser-Rand, but many of the operational aspects of theplant (inlet temperatures, pressures, etc) are similar to those of the BBC design for theHuntorf plant. The facility does, however, include a heat recuperator that reduces fuelconsumption by approximately 22% at full load output and features a dual-fuelcombustor capable of burning No. 2 fuel oil in addition to natural gas [11].Although the plant experienced significant outages in its early operation, the causes ofthese outages were addressed through modifications of the high pressure combustormounting and a redesign of the low pressure combustor [53]. These changes enabled theMcIntosh plant, over 10 years of operation, to achieve 91.2% and 92.1% average startingreliabilities as well as 96.8% and 99.5% average running reliability for the generationcycle and compression cycle respectively [54].1.4.3. NortonA proposal has been under development to convert an idle limestone mine in Norton,Ohio into the storage reservoir for an 800MW CABS facility (with provisional plans toexpand to 2,700 MW [9 x 300 MW] see Figure 11). The mine, purchased in 1999, wouldprovide 9.6 million cubic meters of storage and operate at pressures of between 55 and110 bar. The project, initially approved by the Ohio Public Siting Board in 2001, wasgranted a five-year extension in 2006. Project negotiations are currently underway and itappears that the project will move forward [52, 55-57].24 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Figure 11 A rendering of the proposed 2700 MW CAES plant based on an abandonedlimestone mine in Norton, OH [55]1.4.4. Iowa Stored Energy ParkThe Iowa Association of MunicipalUtilities (IAMU) is developing an aquiferCAES project in Dallas Center, Iowa thatwill be directly coupled to a wind farm (seeFigure 12). The Iowa Stored Energy Park(ISEP, a 268 MW CABS plant coupled to75 to 100 MW of wind capacity, wasformally announced in December 2006.This is the only publicly announced projectto date directly linking CAES with windenergy and the only one using a porousrock storage reservoir. The CABS facilitywill occupy 40 acres located within 30miles of Des Moines, Iowa and use a 3000ft deep anticline in a porous sandstoneformation to store wind energy generatedas far away as 100 to 200 miles from thesite. This was the third location studied forISEP after an initial screening of more thanFigure 12 Diagram of the Iowa StoredEnergy Park [58]25 Compressed Air Energy Storage, Succar and Williams Arl20April 200820 geologic structures in the state. Studies of the chosen formation have verified it hasadequate size, depth and caprock structure to support CABS operation. Construction isdue to begin in 2009, with completion and operation scheduled for 2011 [58].1.4.5. Proposed Systems in TexasSeveral factors make Texas and the surrounding region attractive for CABS development:First, the rapid growth of wind power in Texas (currently the largest and fastest growingwind market of any US state) is putting increasing burdens on existing load-followingcapacity in the region. Second, there are considerable transmission bottlenecks and fewinterconnection points with neighboring grids presenting a significant curtailment risk forwind developers as wind penetrations continue to increase. Lastly, domal salt formationssuch as those used at the existing Huntorf and McIntosh CABS sites exist in the state.This geology has been proven to work well under CABS operating conditions and thusposes limited risk.Consequently, Ridge Energy Storage & Grid Services L.P. have announced plans todevelop several CABS projects throughout Texas, including a 540 MW (4x135MW)system in Matagorda County, TX based on the McIntosh Dresser-Rand design andutilizing a previously developed brine cavern.8Ridge also prepared two CABS studies focused on the Texas panhandle and surroundingregion. The first, commissioned by the Texas State Energy Conservation Office (SECO)and led by the Colorado River Authority, analyzed the alleviation of transmissioncurtailment through the use of CABS [7]. The second addressed the broader economicimpacts of CABS in Texas, Oklahoma and New Mexico (the study area comprised thecontrol area of SPS, an operating company of Excel Energy) [8]. The studies foundcompelling reasons for pursuing CABS in this region--including improved deliveryprofile for renewable energy on the system, reduced ramping of other system capacitydue to wind energy, and transmission cost offsets. Furthermore, the study estimated a netvalue of $10 million per year to SPS for developing a 270 MW CABS unit with 50 hoursof storage. The report also claims that such a system could enable the development of anadditional 500 MW of wind without any additional ramping burdens on the system.More recently, Shell and TXU have announced they intend to explore the possibility ofadding CABS to a proposed 3,000 MW wind farm in the Texas Panhandle [59].8At the time of the release of this report, it appears that this project is not moving forward.26 Compressed Air Energy Storage, Succar and Williams Arl20April 20082. CAES Operation and PerformanceCAES Media SiteTurbine TrainDesign Point(T =5 Hours)T 10 Hours I Initial OperationI IiI350r 5000r800 r-300 1-700 I250 I-600 1-4500 120-S0.500 I-400!IIIAI4000 t300 1-50-421004* I A A A I IIIia II0 100 200 300 400 500 600 700 800 900Regenerator Flow Rate, LB/SECThe above turbine output and heat rate values are as calculated and do not include margins.Figure 13 Turbine performance characteristics for Aquifer CAES based on EPRI designfor Media, Illinois site [60]2.1. Ramping, Switching and Part-Load OperationThe high part-load efficiency of CAES (see Figure 13) makes it well suited for balancingvariable power sources such as wind. The heat rate increase at part-load is small relativeto a conventional gas turbine because of the way the turboexpander output is controlled.Rather than changing the turbine inlet temperature as in a conventional turbine, the CAES27 Compressed Air Energy Storage, Succar and Williams Arl20April 2008output is controlled by adjusting the air flow rate with inlet temperatures kept constant atboth expansion stages. This leads to better heat utilization and higher efficiency duringpart-load operation [51].The McIntosh CAES plant delivers power at heat rates of 4330 kJ/kWh (LHV) at fullload and 4750 kJ/kWh (LHV) at 20% load [53]. This excellent part-load behavior couldbe further enhanced in modular systems such as the proposed Norton plant where the fullplant output would be delivered by multiple modules. In this case, the system could rampdown to 2.2% of the full load output and still be within 10% of the full load output heatrate.The ramp rates for a CAES system is also better than for an equivalent gas turbine plant.The McIntosh plant can ramp at approximately 18 MW per minute, which is about 60%greater than for typical gas turbines. The Matagorda Plant proposed by Ridge EnergyStorage is designed to be able to bring its four 135 MW power train modules to fullpower in 14 minutes (or 7 minutes for an emergency start)--which translates to 9.6 to 19MW per minute per module. These fast ramp rates together with efficient part loadoperation make CAES an ideal technology for balancing the stochastic variations in windpower.To initiate compression operation, the turbine typically brings the machinery train tospeed. After synchronization, the turbine is decoupled and shut off and the compressorsare left operating. This means that the turbines are called upon to initiate bothcompression and generation. In the case of the Huntorf CAES system the switch fromone operating mode to another is completely automated and requires a minimum of 20minutes during which time the system is neither generating power nor compressing air[50]. The switchover time could have a significant impact for balancing rapid fluctuationsin wind output. It is possible alternative startup designs, such as use of an auxiliarystartup motor could reduce this interval further [60].Operation switchover time limitations could even be eliminated altogether with newsystem designs that decouple the compression and turboexpander trains. By separatingthese components rather than linking them through a common shaft via a clutch as in theMcIntosh and Huntorf system, direct switching between compression and expansionoperation is possible. This change also means compressor size can be optimizedindependently of the turboexpander design and permits standard production compressorsto be used in the system configuration [52].28 Compressed Air Energy Storage, Succar and Williams Arl20April 2008A;tt .Figure 14 Constant pressure CAES reservoir withcompensating water column. (1) Exhaust (2) CAESPlant (3) Surface Pond (4) Stored Air (5) WaterColumn [51]2.2. Constant Volume and Constant PressureA CAES system can operate in a number of different ways depending on the type ofgeology being utilized for the storage reservoir. The most common mode is to operate theCAES system under constant volume conditions. This means that the storage volume is afixed, rigid reservoir operating over an appropriate pressure range.9 This mode ofoperation offers two design options: (1) it is possible to design such a system to allow thehp turbine inlet pressure to vary with the cavern pressure (reducing output) or (2) keepthe inlet pressure of the hp turbine constant by throttling the upstream air to a fixedpressure. Although this latter option requires a larger storage volume (due to throttlinglosses), it has been pursued at both of the existing CAES facilities due to the increase inturbine efficiency attained for constant inlet pressure operation. The Huntorf CABS plantis designed to throttle the cavern air to 46 bar at the hp turbine inlet (with cavernsoperating between 48 to 66 bar) and the Mclntosh system similarly throttles the incomingair to 45 bar (operating between 45 and 74 bar).A third option is to keep the storage cavern at constant pressure throughout operation byusing a head of water applied by an aboveground reservoir (see Figure 14). The use of9 Although aquifer bubbles are not rigid bodies, the time scale at which the air-water interfaces migrate ismuch longer than CAES storage cycles and therefore porous rock systems can be approximated as fixed-volume air reservoirs in this context (see section 3.6)29 Compressed Air Energy Storage, Succar and Williams Arl20April 2008compensated storage volumes minimizes losses and improves system efficiency, but caremust be taken to manage flow instabilities in the water shaft such as the so-calledchampagne effect [61 ].This technique is incompatible with salt-based caverns since a continual flow of waterwould dissolve walls of the cavern. Brine cycling with a compensating column connectedto a surface pond of saturated brine could be implemented, but biological concerns andground water contamination issues would need to be addressed [51]. Since pressurecompensated operation cannot be employed in aquifer systems (see Flow in Aquifersbelow), the use of constant-pressure CAES operation is primarily limited to systems withreservoirs mined from hard rock.2.3. Storage Volume RequirementAlthough several CAES systems have been successfully implemented and even thoughsuitable geologies appear plentiful, the realistic potential for large scale worldwidedeployment will not be known until there is much better understanding of the geologicresources available to support many plants deployed under a wide variety of conditions.One of the keys to assessing the geologic requirements for CAES is to understand howmuch electrical energy can be generated per unit volume of storage cavern capacity(EGEN/Vs). The electrical output of the turbine (EGEN) is given by:ETEN= "M G J 0T WCV,0T dt (1)where the integral is the mechanical work generated by the expansion of air and fuel inthe turbine,wcv,T~oT = total mechanical work per unit mass generated in this processrhT= air mass flow ratet = time required to deplete a full storage reservoir at full output powerlIM =mechanical efficiency of the turbine (which reflects turbine bearing losses)i = electric generator efficiencySince all CABS systems to date are based on two expansion stages, the work output canbe expressed as the sum of the output from the two stages. The first term reflects thework output from the hp turbine that expands the air from the hp turbine inlet pressure(p1) to the Ip turbine inlet pressure (p2). Likewise, the second term reflects the expansionwork derived from the expansion from P2 to barometric pressure (Pb).C'TT= WC1+ wcV2 =---f" v dp-fhvd 2Consider first the work output from the first expansion stage. Assuming adiabaticcompression and that the working fluid is an ideal gas with a constant specific heat (sothat P'vk = c, a constant, where k1 Cp1/Cv1) the work per unit mass is:=ptldk-,lkf1(3)30 Compressed Air Energy Storage, Succar and Williams Arl20April 2008-k1-1l[(C'1/k1~lp7 ? = k1--1 [plvl -p2v2] (4)Combining with a similar expression for the second stage gives the total work per unitmass for the process (Wcv,ToT):W C TOT Cp274 [ [1 ( Pblk2JJ] (7)Furthermore, the total mass flow through the turbine can be expressed as separate air andfuel input terms:in7 = rA + r = mA1( +- (8)Sinceconstant (9)The result is:EGEN~ m .fl-jz+/- )dt (10)whereandSk1-1"31 Compressed Air Energy Storage, Succar and Williams Arl20April 20082.3.1. Case 1: Constant Cavern PressureFirst consider the case of a CABS system with constant cavern pressure such as a hardrock cavern with hydraulic compensation (see Figure 14). In this case, the mass flow ofair is constant throughout the process and can be expressed as a simple ratio:mA 4 p5T/5(13)t R~tLikewise, since the inlet pressures and temperatures are constant in time, equation (10)reduces to the following:EGEN =a /3+/-1 [ _ dl (14)Combining these expressions,VS R7 P S [ P(-\ jJ)2.3.2. Case 2: Variable Cavern Pressure, Variable Turbine InletPressureIn the case of a variable pressure CABS system, the pressure at the turbine inlet isallowed to vary over the operating range of the storage volume (from ps2 to psi).However, since the pressure ratio across the hp turbine (p2/pi) remains constant, thepressure ratio across the lp turbine is proportional to the cavern pressure ps [32]:P___b = P__ Pb_ constant (16)P2 P2 PPs Pswhere cp is a correction factor that accounts for the pressure loss from the storagereservoir to the turbine inlet (~-0.90).____ d__ __1_mA ks R~s I dt (8Substituting equations (16) and (18) into (10), the energy storage density is:ks- k-1 ___ k2-1'\EGE M s -p5 ,, Pi[31. Pb (19V5-R7~ k5f~s2 jkj/3+1 Ps p5(932 Compressed Air Energy Storage, Succar and Williams Arl20April 2008k5-1 F ks-1 k2-1 -lk- ]M " (51)fffs ( dp __Pb___ k,-.21RT5 k\j3+ Psi k d Ps~;2) k2 f "ZPl \ dps{ (20)Ts s /+ s p FP PS2-k2-PP2s2 ] Ps2k2/I25 1 /ks Jzs 2q kh- 2 -sIMvlwPs2 (ips +]iI 1 I P4 fPs22(/S1) 1- -{(22)aM s2 L Psi]) ( )Ps b 2 1s( _L~] Psi122R7~ (/+/-1) I P 'P52 ks( +kL) I s2.3.3. Case 3: Variable Cavern Pressure, Constant Turbine InletPressureThe third case we consider is one in which the air recovered from storage is throttledfrom the reservoir pressure Ps to the hp turbine inlet pressure pi such that the mass flowand expansion work output are constant in time. As in case 1, the integral representingthe mechanical work in turbine expansion can be reduced to a simple time average, but inthis case, the net air mass withdrawn from storage is a function of the storage pressurefluctuation over the range ps2 to psi:_m g= s g -s _ (24)Substituting these into equation (10) yields______-____1_____12.3.4. Cavern SizeFigure 15 shows the energy storage density for the above three cases as a function of themaximum reservoir pressure, and, for cases 2 and 3, as a function of the storage pressureratio as well.For all three cases, the electric energy storage density EGEN/Vs increases approximatelylinearly with increasing reservoir pressure Ps2 (or equivalently with mass per unit volumeps2*Mw/RTs2). In some cases however, this might result in large heat loss in theaftercooler depending on the thermal constraints of the cavern [62].33 Compressed Air Energy Storage, Succar and Williams Arl20April 2008The use of a constant-pressure compensated cavern requires the smallest cavern by far.Zaugg estimates for a configuration similar to the Huntorf design (with a storage pressureof 60 bar), a constant pressure cavern could deliver the same output with only 23% of thestorage volume required for a constant volume configuration with variable inlet pressure(ps2psl=1 .4) [32]. If hard rock reservoirs are unavailable or too costly, pressurecompensated systems will.most likely not be an option, so that a case 2 or a case 3 designwould be required.16( 1EenSO 60Fig. -Determining the size of the reservoirEoen = Generator energy' 5 = Storage volume.=Upper storage pressure= Lower storage pressure-... = Reservoir, case I-= Reservoir, case 2... Reservoir, case3TErry = 825 °K, TEzND- 1100 0KFigure 15 The energy produced per unit volume for CAES withconstant pressure reservoir (case 1), variable pressure reservoir (case 2)and variable pressure reservoir with constant turbine inlet pressure(case 3). Inset shows throttling losses associated with case 3 relative tothe variable inlet pressure scenario (case 2)- figure from [32].34 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Notably, although the throttling losses incurred in case 3 relative to the variable turbineinlet pressure system (case 2) implies a required larger storage volume, the penalty is notlarge (see Figure 15 inset). In particular the throttling losses are small with large initialpressures (ps2>60 bar) that is consistent with all known existing and proposed CABSsystems. Because this small penalty is offset by the benefits of higher turbine efficiencyand simplified system operation, it is often optimal to operate a CABS system in thismode (as is the case at both the Huntorf and McIntosh plants).However, in some cases it might be advantageous to allow the inlet pressure to varydepending on the geologic characteristics of the system. For aquifer Systems for example,due to the large amount of cushion gas needed, the storage pressure ratio pps1/s isrelatively small (<1.5) such that the hp turbine can operate over the full storage reservoirpressure range with relatively small penalties relative to the design point performanceU ,8U,1 1,* Ii I1 i*' I* S* S* I"I "Hunotor f:, ..,ps2/Ps-l=38 'Si II- I--,I *S.! " 1I' I.1 I, I ...'. .., .I..Als tom (2004).-ps2/psi=2.0-1* 0.9* 0.8* ¸¸o:.7* o.6-0.5,0.4*,0.30.2* 0,,10[11. (I-. * .i , .l0 0.5 I1.5 2 2. 5I :3.2I44.5Pressu'e Ratio (ps2/psl)Figure 16 The ratio of storage energy density between a constant volume CAES system with constantturbine inlet pressure (case 3) and a pressure compensated CAES reservoir (case 1) as a function of theratio between the operating pressures of the case 3 system (ps2/Psl). 101oHere we assume ks=l.4 and (Ps2/Ts2)/ (ps1/Ts1) -135 Compressed Air Energy Storage, Succar and Williams Arl20April 2008(see Figure 13) [50, 60],Although a variable pressure reservoir CABS system requires a larger storage volumethan a compensated reservoir, volume requirements might be reduced substantially by anappropriate design of the storage volume pressure range, to the extent that so doing isconsistent with the pressure limits of the reservoir and the turbomachinery. The ratio ofthe energy storage density for case 3 relative to case 1 is given by (compare equations(25) and (15)):)(26)This term increases with ps2/Psl as shown in Figure 16. Thus selecting formations thatcan accommodate large pressures swings and high maximum reservoir pressures willreduce land area requirements for CABS through increased storage energy density.Typical numbers for EGEN/Vs are 2-4 kWh/in3 for lower pressure ratios such as those atHuntorf (ps2/ps1=1.38, ps2=66 bar, EGEN/Vs=3.74) and 6-9 kWh/in3 for the newer designssuch one proposed by Alstom, which is designed with higher operating pressures andlarger pressure ratios (ps2/Psr=2.0, ps2=1 10 bar, EGEN/Vs=8.44) [11, 63].In section 4, "Wind/CABS Systems in Baseload Power Markets", a CABS system designis described which converts wind power into baseload electricity. The systemconfiguration includes a storage reservoir capable of supporting 2 GW of baseload powerfor 88 hours (176 GWh of storage). The land area requirement for the wind turbine arrayis 860 km .For a system with an electricity storage density consistent with a formationdepth similar to the Dallas Center, Iowa CAES plant (depth 880m, discovery pressure80 bar, EGEN/Vs5 5 kWh/in3) the total pore volume needed for the cycled air would be 35million cubic meters.11 Assuming the ratio of total air mass (cushion air plus cycled air)in the reservoir to the mass of cycled air is 5 [64], and assuming an average reservoirheight of 10 meters and an effective porosity of 15%, the "footprint" of the reservoirwould occupy an area of land equal to approximately 14% of the land area of the windturbine array.2.4. Performance In dices for CAES SystemsThe energy performance of a conventional fossil fuel power plant is easily described by asingle efficiency: the ratio of electrical energy generated to thermal energy in the fuel.The situation is more complicated for CABS due to the presence of two very differentenergy inputs. On the one hand, electricity is used to drive the compressors and on theother natural gas or oil is burned to heat the air prior to expansion. This situation makes itdifficult to describe CABS performance via a single index in a way that is universallyuseful--the most helpful single index depends on the application for CABS that one hasin mind. Before turning to a discussion of alternative options for a single CABS11 This volume corresponds to a gas volume that is of the same order as the working gas capacity of thelargest porous rock natural gas storage sites in the US and Canada, but is considerably larger (by about anorder of magnitude) than the mean capacity among these facilities (AGA, 2004).36 Compressed Air Energy Storage, Succar and Williams Arl20April 2008performance index, it is worthwhile considering, the two performance indices that applyto each energy input separately: the heat rate and the charging electricity ratio.2.4.1. Heat RateThe heat rate (HR) or fuel consumed per kWh of output for a CABS system is a functionof many system design parameters, but the design choice that most critically affects theheat rate is the presence of a heat recovery system. The addition of a heat recuperatorallows the system to capture the exhaust heat from the lp turbine to preheat the airwithdrawn from the storage reservoir. Heat rates for CAES systems without a heatrecovery system are typically 5500-6000 kJ/kWh LHV (e.g., 5870 kJ/kWh LHV forHuntorf). Heat rates with a recuperator are typically 4200-4500 kJ/kWh LHV (e.g., 4330kJ/kWh for McIntosh). By comparison, a conventional gas turbine has at least twice thislevel of fuel consumption (-~9500 kJ/kWh LHV) because two thirds of the electricaloutput is used to run the compressor. Because the CABS compression energy is suppliedseparately, the system can achieve a much lower heat rate [11, 51].The addition of the heat recuperator reduced the fuel consumption at McIntosh by 22%relative to operation without this component [53], but a high pressure combustor was stillrequired in this case. Newer CABS designs feature higher inlet temperatures at the lpturbine. The added heat generated at this stage facilitates the removal of the hp combustorfrom the design altogether (as for the CABS unit shown in Figure 3). In addition tofurther reducing fuel consumption, these systems also offer significant NOx emissionsbenefits relative to prior designs [63].2.4.2. Charging Electricity RatioThe second performance index for CABS is the ratio of generator output to compressormotor input--the charging electricity ratio (CER). Because of the fuel input, the CER isgreater than unity and will typically lie in the range of 1.2 to 1.8 (khupt/~ipt [11,32, 65]. The CER also takes into account piping and throttling losses as well ascompressor and expander efficiencies. Throttling loss is a function the reservoir pressurerange (see Figure 15). Turbine efficiency is especially important in the low-pressureexpansion stage, in which most of the enthalpy drop occurs and where approximatelythree quarters of the power is generated [66]. Increased turbine inlet temperatures (e.g.,by using expander blade cooling technologies) would enhance the turbine, and CABSelectrical efficiencies as well [67].2.4.3. Toward a Single CAES Performance IndexSeveral single-parameter performance indices have been proposed for CABS. Thesimplest possible index is an efficiency rl defined as the ratio of energy generated by theturbine (ET) to the sum of electrical energy delivered to the compressor motor (EM) andthe thermal energy in the fuel (EF)r- ET(27)EM + FTypical HR and CBR values of, respectively, 4220 kJ/kWh and 1.5 imply rl 54%.However, because of the substantial difference between the energy qualities of the37 Compressed Air Energy Storage, Succar and Williams Arl20April 2008thermal energy in the fuel and the electrical energy supplied to the compressor, their sumis not a meaningful number. In order to estimate the total energy input to CABS, it isnecessary to express both the fuel and compressor electricity on an equivalent energybasis. One approach is to express the electrical input as an equivalent quantity of thermalenergy.2.4.3.1. Primary Energy EfficiencyWhen CABS is used to convert baseload thermal power into peaking power (in place ofgas turbines or other peaking units) one can' introduce a primary energy efficiency defined in terms of the thermal efficiency of the baseload plant (OIT). Here compressormotor energy input EM is replaced by an expression for the effective thermal energy inputrequired to produce EM. Thus, the overall efficiency value reflects the system (grid +CABS) efficiency of converting primary (thermal) energy into electrical energy:Er (28)This methodology has been applied to CAES units charged by nuclear and fossil fuelpower plants [32], CHP plants [62], as well as grid-averaged baseload power [68].Assuming riT = 40% (as might characterize a modem supercritical steam-electric plant)and the same other parameters as considered in the earlier calculation of v, implies riPE=35%.In principle, this formulation of system efficiency can be applied to a wind/CABS systemby using the atmospheric efficiency of the wind turbines YlwT" in place of the thermal plantefficiency riT. This formulation, proposed by Arsie et al, gives rise to a system efficiencyof 39% [69]. However, the use of atmospheric efficiency in this case does not serve thesame function as the thermal efficiency. In the case of fossil fuel or nuclear power as thesource of compressor energy, use of the thermal efficiency provides a measure of theamount of primary fuel needed to deliver a quantity of electrical energy EM. In contrast,the extraction of "fuel" in the case of wind energy does not affect the environmentalimpact or overall cost of the plant. Consequently, this measure of the amount ofatmospheric kinetic energy captured in providing EM is not very helpful and in the case ofwind/CABS systems and therefore this is not the optimal formulation for CABSefficiency.2.4.3.2. Round Trip EfficiencyA CAES unit powered by wind energy will be compared to other electrical storageoptions that might be considered for wind back up such as electrochemical or pumpedhydroelectric storage. Such alternative storage systems are typically characterized by aroundtrip electrical storage efficiency viRT" defined asTiRT -- (electricity output)/(electricity input).To facilitate comparisons of CABS to other electrical storage devices, a round tripefficiency can be introduced that employs an "effective" electricity input =EM + TING*EF.The second term is the amount of electricity that could be have been made from the38 Compressed Air Energy Storage, Succar and Williams Arl20April 2008natural gas input EF, had that fuel been used to make electricity in a stand-alone powerplant at efficiency TING instead of to fire a CAES unit. The round-trip efficiency r1RT,1 sodefined is:1R1- ET (29)r,, =EM+/- +rNG ]EFThis methodology has the advantage of providing an electricity-for-electricity roundtripstorage efficiency that isolates the energy losses in the conversion of electricity tocompressed air and back to electricity. Several values for 'riNG have been proposedincluding the hypothetic Carnot cycle efficiency [65] as well as the efficiencies ofcommercial simple cycle and combined cycle power plants [2, 70]. For typical naturalgas power systems, (heat rates in the range 6700-9400 kJ/kWh) CABS roundtripefficiencies are in the range of 77-89% assuming a 1.5 ratio of output to input electricityand a heat rate of 4220 kJ LHV per kWh. An exergy analysis of conventional CABSsystems indicates that 47.6% of the fuel energy input is converted into electrical work[71]. For this measure of the thermal efficiency, the roundtrip efficiency is 81.7%.An alternative formulation 11rT,;2 of an electrical roundtrip storage efficiency introduces anoutput correction term EF*TING. Instead of expressing the fuel input as an effectiveelectrical input, the electrical output is adjusted by subtracting the assumed contributionto the output attributable to the fuel. Correspondingly the output attributable to theelectrical input is ET -EF*T1NG [72].E/r,2 -ETG(30)EMUsing the same assumptions as for 'rliR-,I with the Zaugg efficiency for fuel conversion,'rING = 47.6%, the round trip efficiency is 66%/.Thus, depending on the index chosen for its measure, the roundtrip efficiency for CABSis typically in the range 66-82%. This is in the same range as the roundtrip efficienciescited for other bulk energy storage technologies such as pumped hydroelectric storage(74%) and Vanadium flow batteries (75%) [70].2.4.3.3. Additional ApproachesStill another measure of the efficiency of CABS proposed by Schainker et al might beuseful for an economic evaluation of CABS in load leveling or arbitrage applications.This approach is similar to in that it adjusts the fuel input by a correction factor:fA E E (31)A EF/c R/ + EMIn this case however, the fuel input is converted to equivalent electricity not by using theprimary energy conversion efficiency for natural gas but rather the cost ratio CR =(off-peak electricity price)/(fuel price) [73] .Although this index might be helpful in decidinghow to operate a given CABS unit over time, the measure varies significantly both overtime and with geographical region and so is not a useful general plant characterization.39 Compressed Air Energy Storage, Succar and Williams Arl20April 2008A final description of CAES efficiency compares the CABS output to the electricaloutput of a thermodynamically ideal CABS plant operating between ambient temperatureTo and Tmax [65]:=ETETRvETRlEv =EM +EF -Toe*AS :=EMt+ EF -TO eEF/TM~(32)(33)Analysis of a conventional CABS system yields a second law efficiency Ofrjii=68% witha recuperator and 59-61% withoutlzUltimately, the choice of efficiency measure remains an open question because thermalenergy and electrical energy quantities cannot be combined by algebraic manipulation.The formulations provided in this Section help only to provide a basis for comparisonwith other storage technologies, but~as indicated above, the relevant expression isdetermined in large part by the application one has in mind.12 The range of efficiencies for the system without recuperator reflects change in system performance dueto varying storage pressures (Ps =20 to 70 bar). The change in efficiency was < 1% for the system withrecuperator.40 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Table 2 Selected CABS Efficiency* Expressions and Values in The LiteratureParameter Definition Reported Value________No Recuperator With Heat Recup~eratorHeat Rate Er 6000-5500 kJ/kWh 4500-4200 k J/kWhOJF = E--- (-60-65%) (-80-85%)Charging ET 1.2-1.4 1.4-1.6Energy Ratio Th'sE = -Primary Energy ET CABS Charged From Nuclear Power (Thr=33%) [32]Effcinc rPE M/lr+ F Charged From Fossil Fuel Power Plant (tiT=42%) [32128.2% 34.4%Charged from Combined Heat and Power Plant (tiT=35%)[62].. .. ....... .. ....35.1-41.8%Charged from grid-averaged Baseload Power (flT=35%,CER=1.4) [68]____________________42-47%Roundtrip ET 4220 kJ LHV/kWh, CER=1.5, flNo=47.6%, [2]Efiiny(= lT 81.7Effciecy(1) rl~,1EM + rING EFRoundtrip ET -- EFVING 4220 kJ LHV/kWh, Eo/Ei=l1.5, 1lNG=47.6% [72]Efficiency (2) VIRT.2 -- 66.3%E___66.3%ESecond Law ET To=l15 C, TMAX=900 C, ps=20 bar [65]Efficiency Thiz = Ermv5.%68.3%41 Compressed Air Energy Storage, Succar and Williams Arl20April 20083. Aquifer CAES Geology and Operation3.1. MotivationsInterest in aquifer CAES technology stems from the widespread availability of thisformation type and the expected relatively low development costs. Furthermore, Figure17 shows that onshore wind resources in the US of class 4 and above correlate well withaquifers.While solution-mined salt domes offer advantages in terms of reliability and flexibility ofdesign, the supply of salt domes is limited in the U.S. to the Gulf Coast region (see Figure17). However, most of this region has very poor wind resources (typically wind classes 2and below) that are not economically exploitable. If the aim of storage is to providebackup for large quantities of wind power, salt domes will not play a large role in theUnited States. While bedded salt formations might be used, their development will likelybe more challenging and costly than the salt dome CAES systems that have beendeployed (see section 1.3.1).i clasus 4+ Wind ResourcesFigure 17 A comparison of areas of high quality wind resources and geology compatible with CAES(areas suitable for mined rock caverns omitted due to the high estimated cost of developing suchformations for CAES) [33, 37, 41-43]. Locations of the existing Mclntosh CABS plant, the recentlyannounced Dallas Center wind/CABS system and the proposed Matagorda plant are indicated as well.Figure 17 indicates areas favorable for air injection into porous rocks overlaid with areaswith wind resources of class 4 and above (today, class 5 winds are economical, and class4 resources are considered marginally viable). The overlap includes large areas in the42 Compressed Air Energy Storage, Succar and Williams Arl20April 2008southern tier states that extend from New Mexico to Arkansas, and includes large areas ofColorado, Wyoming, Montana, Kansas, Iowa, and Minnesota and Iowa, and most of theDakotas. Although resource maps such as Figure 17 can be useful in helping to decidewhere to site a CAES storage unit, a detailed geologic site characterization is needed toascertain whether a site is actually suitable for CAES development.Although the total cost of developing a porous rock formation for CAES will depend onthe characteristics of the storage stratum (e.g. thinner, less permeable structures willrequire more wells and therefore a higher development cost), it appears that this type ofgeology is often the least cost option. Prior CAES cost estimates (see Table 3) indicatethat total development costs are in the range $2-$6 million per Bcf of total volume(working gas and base gas) which is similar to development cost estimates for natural gasstorage in porous rock [74]. This implies a capital cost of $2.0-$7.0 per kWh of storagecapacity depending on the site characteristics and assuming a five-to-one base gas toworking gas volume ratio [641. These costs are somewhat lower than those estimated forsalt cavern storage ($6-$i10 per kWh of storage capacity) which is the next cheapestoption.Table 3 Estimated Well and Reservoir Development Costs for Aquifer CAESaSite 1 : Oneida Site 2: Rockland County Site 3: BuffaloDepth 910 460 610CAES Well, Each ($) 775,000 480,000 520,000Well Lateral, Each ($) 100,000 100,000 100,000Gathering System ($) 2,600,000 2,600,000 2,600,000Number of Wells 18 -38 80 -107 40 -71Total Cost ($ per kWh 2.0 -2.2 5.6 -7.0 2.7 -3.4of storage capacity)b'¢' ____________ ____________ ____________a. Costs based on a 1994 survey of CAES plant sites in New York State [64] inflation-adjusted to a $2006basisb. Wells, laterals and gathering system account for 90% of total cavern development costs. Remaining costsinclude reservoir characterization activities such as a seismic monitoring array for the candidate site.c. Storage costs assume a five-to-one ratio of base gas volume to working gas volume. Actual base gasvolume ratios will depend on the characteristics of individual sites.Aquifer CAES has the further advantage that the cost of incremental additions to storagecapacity is significantly lower than for alternative geologies. Assuming sufficient wellsare in place to ensure adequate air flow to the surface turbomachinery, the cost ofincreasing the storage capacity of the aquifer is simply the compression energy requiredto increase the volume of the bubble [60]. This cost (-~$0.11I/kWh) is an order ofmagnitude lower than the equivalent marginal costs of solution mining salt and more thantwo orders smaller than excavating additional cavern volume from hard rock [11].Because this combination of low cost and potential for widespread availability is uniqueamong the options for storage reservoirs types, it will be essential to pursue developmentof aquifer-based systems if CABS is to serve more than a niche role in balancing U.S.wind capacity.43 Compressed Air Energy Storage, Succar and Williams Arl20April 20083.2. Applicability of Industrial Fluid Storage ExperienceTo gauge the potential for aquifer CABS, much can be gained from existing studies onother underground fluid storage applications. To date the storage of natural gas has beenthe principal commercial application for storage of fluids in porous rock strata, butstorage of other materials such as liquid fuels, propane and butane have been pursued aswell.3.2.1. CO2 StorageMore recently, storage of supercritical CO2 in deep formations has garnered significantattention in the context of carbon capture and storage (CCS) technology development forclimate mitigation.Assessments of CO2 storage are somewhat less relevant to CABS however. Theminimum depth required for CO2 to become supercritical (--800m) is typically at the highend of acceptable limits for CABS (see Geologic Requirements below). In addition,because CO2 is stored permanently rather than being cycled, the presence of an anticlineis not necessary. Flatter caprock layers are in fact more desirable for storage of carbondioxide, since they promote further migration and faster dissolution of the injected CO2 inthe brine. In addition, the higher viscosity of CO2 under storage conditions and the loweraverage permeability of deep aquifers imply that flow behavior relevant to carbon storagewill be different than for CABS.3.2.2. Natural Gas StorageIn contrast, natural gas is stored under conditions much closer to those needed for CABS.Consequently, consideration of natural gas storage provides a valuable starting point foran analysis of air storage in porous rock formations.The extensive industrial experience with natural gas storage provides a theoretical andpractical framework for describing underground storage media and assessing candidatesites for seasonal storage of natural gas [75]. Field tests and prior studies discussed belowindicate that this theory is applicable to CABS site analysis and operational planning.Seasonal storage of natural gas began as an industry in 1915 when the Natural Fuel GasCompany used a partially depleted natural gas reservoir in Ontario, Canada to meet peakwinter demand for gas. By 2004 the working gas capacity of the natural gas storageindustry in the U.S. and Canada had grown to 4.1 trillion standard cubic feet in 428facilities spread over 30 U.S. states and 5 Canadian provinces. This storage capacitycorresponds to roughly 17% of the total annual demand for natural gas in the U.S. andCanada for 2002 [76, 77]. Over 95% of this capacity is held in porous rock formations(mostly in depleted gas fields) making this industrial experience base especially relevantto the understanding of aquifer CABS systems.3.2.2.1. Site Characterization and Bubble DevelopmentWhile there are important differences in the details of storing air versus natural gas inunderground formations, the methodologies developed for evaluating natural gas storagesites are directly applicable to CABS.44 Compressed Air Energy Storage, Succar and Williams Arl20April 2008High-resolution seismic surveys can help to define the shape of a geologic structure, thethickness of a zone of interest and presence of viable cap rock. Also, pump tests can beused to measure critical flow properties of the reservoir. Following successful sitecharacterization, the reservoir is developed over the course of several months.By injecting fluid above the discovery pressure (the hydrostatic pressure in the formationprior to well drilling), the brine can be displaced from the porous stratum with gas -initially fingering through the stratum and eventually resulting in formation of acoalesced bubble. The bubble is developed to the point that bubble volume and closurerating are deemed sufficient (for further discussion of closure rating see GeologicRequirements section and Figure 18 below). From this point forward, the reservoir canbegin storage operations.During operation the mean pressure in the reservoir is kept at the discovery pressure toensure that the bubble volume remains constant and so that there is no long-termmigration of the bubble walls (migration of water interface is more pertinent to seasonalnatural gas storage than to high frequency reservoir cycling for CABS, see section 3.6,"Flow in Aquifers").Formation flow (injectivity and deliverability) is critical for determining the suitability ofa candidate storage site. The analytical description of reservoir flow begins withcalculations of steady state flow, which is described by Darcy's Law:q =k dp (34)A pdLwhereq =flow rate (cm3/s)A = cross-sectional area (cm2)k = permeability (darcy)gt=viscosity (centipoises)dp/dL = pressure gradient in the direction of flow (atm/cm).Assuming radial laminar flow near a well (injection well or recovery well) through anaquifer [described as a homogeneous formation of thickness h (with A = 22trh) andpermeability k], the flow rate for a single well can be expressed as.2srrhk dp (5~udrFrom the real-gas equation-of-state, the number of gas moles n is given by:pVn -(36)ZRTwhere:Z = gas deviation factorThe flow rate q at temperature T and pressure p can be expressed in terms of the flow rateqsc at standard conditions (Psc, Tsc) by:45 Compressed Air Energy Storage, Succar and Williams Arl20April 2008pV _ PscVsc (37)zT Ts(*and soqsc -qz cP (38In English units, Ts = 519.67 °R (60 0F) and Ps = 14.7 psia, so that the flow Qsc (inMiMscfd) is:=0 .447x10-6.r k h p dp (9paT Zdr/ rBecause the total radial flow rate is independent of the radial distance from the well, Qsccan be evaluated by integration from the wellbore radius to the formation radius.Assuming the temperature in the reservoir is constant, the deliverability equation is:=0.703x10-6 k h [p2 _ ps] (0p TZ ln['*'rJwhere:rw = welibore radius (ft)rF = formation radius (ft)ps = pressure at the wellbore (psia)PF = pressure at the formation edge (psia)h =formation height (fi)k = permeability (millidarcy)T = temperature in the reservoir (°R)(centipoises)Q =gas flow rate (MMcfd)--which is positive for flow out of the reservoirThis equation is widely used to describe the flow capacity of natural gas fields [78].Additional terms are needed to reflect effects of turbulence, but field studies indicate thatthe assumption of laminar flow is adequate to describe CABS operation [79].*133.2.2.2. Applicability to CAESThe applicability of this methodology for describing airflow in aquifer-based CABSsystems was verified during the Pittsfield Aquifer Field Test, which took place at thePittsfield-Hadley Anticline in Pike County, Illinois from 1982-1983. Prior to conductingdeliverability measurements of the site, data sources such as core sample analysis, pumptests, injection tests, and earlier geophysical tests were sampled. These providedestimates of formation thickness and permeability data that were used to calculateSteady state flow equations are useful for evaluating reservoir deliverability, but time-dependentunsteady-state and pseudosteady-state flow expressions are required to adequately describe the evolution offlow during bubble development (see section 3.6, "Flow in Aquifers")46 Compressed Air Energy Storage, Succar and Williams Arl20April 2008predicted deliverability rates. Ultimately, the deliverability measurements acquiredduring site operation corresponded closely with the predicted values based on thegeophysical data:During the process of reviewing and analyzing the multitude of operating data for thePittsfield experiment, most of the questions and apprehensions regarding the Pittsfieldreservoir were answered satisfactorily. The flow behaviors of the Green and White St. Peterare now understood to the extent necessary to conduct an underground storage operation.Natural gas equations have been shown to be applicable to air flow. There is no questionthat the experiment proved that CABS in porous media is feasible in terms of storage andflow of air [7/9].The applicability of natural gas storage formation analysis techniques extends beyondporous rock formations (aquifers). In the case of salt dome storage, the fact that both theHuntorf and Mclntosh CAES facilities are located adjacent to natural gas storagefacilities mined from the same formation'4 suggests that the conditions favorable forCAES development and natural gas development might often overlap. Since a largevolume of test data is available from state geological surveys on potential natural gasstorage facilities, it is likely that this body of knowledge will be useful in identifyingpotential sites for CAES.3.2.2.3. DifferencesWhile natural gas storage provides an important departure point for a discussion ofCABS, several important differences must be considered. First, the differences in thephysical properties of air relative to natural gas have important implications for thegeologic requirements for aquifer CABS. Second, a CABS system used for arbitrage orbacking wind power will likely switch between compression and generation at least oncea day and perhaps several times a day. In contrast, most natural gas storage facilities areoften only cycled once over the course of the year to meet the seasonal demandfluctuations for natural gas. Third, several oxidation processes might take place in thepresence of oxygen from the air depending on the mineralogy of the formation. Also,introduction of air into the formation might promote propagation of aerobic bacteria thatmight pose a significant corrosion risk. Finally, additional corrosion mechanisms mightbe promoted due to the introduction of oxygen into the formation. These considerationsand their impact on system design and operation are discussed in the following sections.3.3. Geologic RequirementsThe requirements for air storage in a porous rock reservoir encompass a broad range ofgeologic features. In general terms, CAES operation requires an anticline consisting ofpermeable, porous media such as sandstone capped by an impermeable caprock (seeFigure 20). Other important considerations during site selection are the volumerequirement of the storage application, the pressure requirements of the surfaceturbomachinery, the homogeneity of the formation and the detailed mineralogy.14 The Huntorf CABS facility was built adjacent to a preexisting natural gas storage facility consisting offour caverns solution-mined from a Permian salt dome. The Mclntosh Salt dome natural gas storage facilitywas completed three years after the CAES facility began operating.47 Compressed Air Energy Storage, Succar and Williams Arl20April 2008One of the most complete studies on the feasibility of aquifer-based CABS systems,prepared by the Public Service Company of Indiana and Sargent and Lundy Engineers forthe Electric Power Research Institute (EPRI) in 1982, explores the potential benefits ofthese systems [60]. Although no field tests were conducted as part of this EPRI study, adetailed methodology was presented for identifying formations with the necessarygeologic requirements. A score-based system was developed to evaluate candidate siteson the basis of geologic, economic and environmental considerations (see Table 4). Theparameters used to evaluate the geologic aspects of the formation include permeability,depth, porosity, closure, geology type, and caprock properties.3.3.1. Porosity, Permeability and ThicknessEach parameter will impact different aspects of CAES operation including reservoircapacity, compressed air deliverability and compatibility with operating pressures forstandard turbomachinery. The permeability and reservoir thickness will determine thedeliverability of the reservoir (see section 3.2.2.1) and together with the porosity willdetermine the pore volume per unit land area and the number of wells needed to achievethe desired total flow./ (IA%.%IW, CO RT0e¢o oua I/\//I!TOP VIEW OLSIII 0%F B OTTO.. ........ .4% IIIV~O~SIDE VIEWh V/AWHERE~ h -AVERAGE HEIGHT 0F AIR RESERVOIRV *VOLUME OF AIR RESERVOIRA *UIOTTOM AREA OF AIR RESERVOIROSAIR RESERVOIRFigure 18 Aquifer dimensions relevant to total closure rating[60]Air has a viscosity approximately twice that of natural gas over a wide range of pressuresand temperatures as well as a higher gas deviation factor (see Figure 19). Therefore in48 Compressed Air Energy Storage, Succar and Williams Arl20April 2008order to achieve the same flow rate, a formation for storing air must have a higher flowcapacity1 than a natural gas storage facility operated under similar conditions (seeequation 40).This underscores the importance of careful site characterization, including seismicmonitoring, core sample analysis, injection tests, pump tests, and careful wellobservation. A reliable permeability value for the formation is essential for predictingbubble development and deliverability characteristics of a reservoir for air storage.Porosity indicates the percentage of the media that consists of voids and interstices. Alower porosity implies a larger areal expanse is needed to contain the necessary volumeof air. In the context of the 1982 EPRI study, 13% was deemed the minimum porosityneeded for CABS operation. All of the aquifers screened for this study met this criterionand 12 of 14 candidate sites exceeded 16% porosity.a-0oI1.0g£Temperature (°F) Pressure (psia)Figure 19 Viscosity and Gas Deviation Factor of Air versus Natural Gas [79]3.3.2. Reservoir DimensionsThe total void volume of the aquifer above the spill point contour (VR) must be at least asgreat as the volume needed for CABS operation (Vs). But if VR is much bigger than isneeded for CABS operation, excessive land rights acquisition costs might be incurred andhence values of Va./Vs greater than 3 receive a reduced score.15 "Flow capacity, " the product of formation thickness and permeability (kh), is a parameter used tocharacterize the flow properties of geologic formations used for underground storage of fluids.49 Compressed Air Energy Storage, Succar and Williams Arl20April 2008The total closure rating is defined as the ratio of the total thickness of the formation (H)to the thickness of the fully developed air bubble (h) (see Figure 18). This parameter isimportant with regards to water encroachment into the weilbore.Water might be drawn up into the well during extended air withdrawal periods due to theradial pressure gradient created as air is withdrawn. To avoid this condition sufficientdistance between the bottom of the well perforations and the air-water interface should bemaintained at all times. Typically, the reservoir will be developed such that 10 to 15 feetof air is maintained below the well perforations, but the actual distance depends on thepressure relative to the discovery pressure of the formation as well as the permeabilityand porosity of the structure.Figure 20 Porous Rock CABS Storage Volume [19]It would be optimal to develop the air bubble to the extent that it spans the full formationthickness (h/H=1.0), in which case the possibility of water encroachment is eliminated.This is more easily accomplished in thinner anticlines with larger curvature so that asmaller volume of air is needed to displace the air/water interface sufficiently. In the caseof flatter and thicker reservoirs, it might not be possible to develop the bubble to thisextent.50 Comipressed Air Energy Storage, Succar and Williams Arl20April 20083.3.3. Pressure Limits and Caprock CharacteristicsPressure limits presented in the EPRI study were based on considerations related tocaprock integrity and turbomachinery operational limits. For the 1982 EPRI study, theallowed pressure range was set at 14-69 bar.16 However, to make best use of existingturbomachinery and to ensure optimal performance, the desired range was 39-50 bar.Both the McIntosh and Huntorf systems operate in this range (45 and 46 bar inletpressures, respectively). The pressure limits or depth limits in a new CAES applicationmight be substantially different from these values, depending on the caprockcharacteristics and the CAES turbomachinery design.The caprock layer must be a relatively impermeable stratum immediately over the porousstorage reservoir. The rock, usually shale, siltstone or dense carbonate, must be thickenough to prevent fracturing and have low permeability together with large capillaryforces in order to prevent air from migrating through the media. As a rule of thumb, thepressure of injection is not allowed to exceed the discovery pressure of the formation bymore than 0.16 bar per meter depth to avoid caprock fracture [19].An important measure for determining the adequacy of the caprock layer is the thresholdpressure, which is defined as the pressure at which air begins to displace water from a16Based on the turbomachinery available at the time, the maximum allowable turbine inlet pressure andmaximum compressor discharge pressure was 62 bar and 76 bar respectively. The minimum turbine inletpressure was 10 bar and a 3.4 bar pressure drop from the storage reservoir to the surface turbomachinerywas assumed.51 Compressed Air Energy Storage, Succar and Williams Arl20April 2008porous rock. A sufficiently high threshold pressure is needed to ensure that air will notmigrate through pore spaces in the caprock in response to pressure fluctuations duringCAES operation. This threshold pressure reflects the wetability of the rock and is afunction of the surface forces at the water-rock interface. These forces are ultimatelyresponsible for the water-filled caprock layer's ability to act as an impermeable barrier toair migration [75]. Threshold pressure and its relationship to caprock permeability can bedetermined by measurements of water migration through core samples subject todifferential pressures (see Figure 21).3.3.4. Residual HydrocarbonsIn addition to using saline aquifers for CABS, it is also possible to use depleted oil andgas reservoirs, which are fundamentally aquifers. Since the bulk of natural gas storageexperience is in depleted fields, many issues related to residual hydrocarbons have beenextensively studied; however the injection of oxygen would present challenges notencountered when storing natural gas.For example, residual hydrocarbons in the pore spaces of the formation might lead to theformation of permeability-reducing compounds and corrosive materials. Anotherpossibility is that the presence of residual hydrocarbons may introduce the risk offlammability and insitu combustion upon the introduction of high-pressure air.The flammability of the natural gas/air mixture may be a concern for CABS operation,but displacement of natural gas away from the active bubble area can mitigate this riskconsider'ably. In some cases, nitrogen injection may be desirable to further minimizeair/natural gas mixing. Previous studies indicate that these methods adequately addressthe challenge of using depleted natural gas fields for CABS and that these structures canprovide a suitable air storage medium [79].52 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Table 4: Ranking Criteria for Candidate Sites for Aquifer CABS [60]Score 1 2 3 4 5Score Interpretation Unusable Marginal OK Good ExcellentPermeability (md) < 100 100- 200 200- 300 300-500 > 500Porosity (%) < 7 7-10 10-13 13-16 > 16Total Reservoir Volume <0.5 0.5 -0.8 0.8 -1.0 1.0 -1.2(VR/Vs) or or______ > 3.0 1.2 -3.0Total Closure Rating (h/H) < 0.5 0.5-0.75 0.75-0.95 0.95-1.0Depth to Top of Reservoir < 137 140-170 170-260 260-430 430 -550(x7or or or>760 670-760 550-670Reservoir Pressure (bar) < 13 13-15 15-23 23-39 39-50or or or> 69 61-69 50-61Type of Reservoir Highly Moderately Reefs, Channel BlanketDiscontinuo vulgar highly sandstones sandsus limestone & vulgardolemite limestone &dolemiteResidual Hydrocarbons (%) > 5% 1-5% < 1%Caprock leakage Leakage No data Pumping test shows no leakageevident availableCaprock Permeability (md)>10 <0.Caprock Threshold Pressure 2 1-55 >5(bar)Caprock Thickness (in) < 6 [ > 6,3.4. Oxidation ConsiderationsThe Pittsfield CAES experiment, conducted during the period 1981-85 in Pike County, ILunder EPRI/DOE sponsorship, involved extensive field tests to determine the feasibilityof using aquifers for air storage [79]. One of the important findings of the study was thatintroduction of air into the reservoir leads to the reaction of oxygen with native speciesthat in turn leads to a reduction in the 02 concentration in the stored air. These oxidationreactions proceed with a characteristic time scales of the order of months.18 The observedoxygen depletion was largely attributed to the presence of sulfide minerals in theformation and subsequent reactions that were catalyzed by the injection of air into theformation. The presence of oxygen can lead to reactions among several mineral specieswith various outcomes.The primary reactant in the Pittsfield case was pyrite, a sulfide of ferrous iron (FeS2). Theoxidation of pyrite ultimately leads to the formation of hematite (Fe203):4 FeSz +11 02-2 Fe03+ 8 SO2(41)The products of this reaction do not present significant problems for reservoir operability.However, if this process does not proceed to completion the presence of intermediatespecies might lead to serious formation damage. Partial oxidation might lead to the17 Depth limits are based on a hydrostatic pressure of approximately 0.09 bar per meter.iThsoxygen depletion was not observed in short duration (several day) storage tests53 Compressed Air Energy Storage, Succar and Williams Arl20April 2008presence of ferrous sulfate or Fe(OH)SO4, which can result in the production of colloidalferric hydroxide and melanterite'9 respectively.2FeS2 + 702 + 2H20 ---,2FeSO4+/-+2H2S04 (42)4FeSO4 + 02 +2H20 --,4Fe(OH)SO4 (43)These species swell to as much as 500% of the original pyrite volume and result inconsiderable permeability decline in the reservoir. This expansion, together with thecollection of these products on pore interiors could impact the permeability of thereservoir substantially. In addition, the volume increase due to oxidation of pyrite andcarbonates might lead to deteriorating expansive stresses on caprock layers.Another problematic oxidation product is gypsum (CaSO4
- 2 H20), which mightprecipitate through dissolution of carbonate minerals. Gypsum forms scale deposit thatmight occlude porosity and impair CAES system performance [79].The degradation of reservoir permeability is not the only challenge which oxidation posesfor aquifer CABS systems. Because the withdrawn air is combusted with fuel, thedepletion of oxygen might result in impaired combustion efficiency downstream.However, because current CABS systems do not utilize all the oxygen in the air stream,some depletion can be tolerated without any loss in combustion efficiency [79].Oxidation might have significant impacts on CABS operation and as such it is essential tofully characterize the mineralogy of a candidate site. It might be possible in some cases tocontrol the rate of reactions by dehumidification of incoming air. Such dehumidificationmight have additional benefits, as discussed below.In addition, if the formation cement between sand grains consists predominantly ofcarbonates and/or sulfides, the dissolution of these materials through oxidation mightrelease particulates. If this happens in the vicinity of the well bore, it is likely that theseparticles can find their way to the turbomachinery (the effect of p articulates on surfaceturbomachinery will be covered below). For this reason and for reasons related to theeffects mineralogical reactions described above, reservoirs having high sulfide contentshould be avoided [79].3.5. CorrosionThe deterioration of weilbore tubulars and casing cement through corrosion is animportant problem to consider for CABS applications. Prominent corrosion types includebiological (esp. bacterial), uniform, galvanic, crevice, pitting, erosion, intergranular,stress corrosion cracking, fatigue, and fretting corrosion. The promotion of corrosion byair injection might be further exacerbated by high-pressure and high-temperatureconditions, especially if significant moisture is present.19 Melanterite (FeSO4 e 7 H20) is a hydrated form of ferrous sulfate often formed from oxidation in pyriticore zones54 Compressed Air Energy Storage, Succar and Williams Arl20April 2008While many corrosion types (e.g. erosion corrosion, corrosion fatigue, fretting corrosion)might be prevented by suitable choice of materials or might simply not be relevant to theconditions in a CAES reservoir (e.g. intergranular corrosion), some might presentparticular problems for air storage applications. Controlling electrochemical corrosionprocesses such as uniform corrosion and pitting corrosion might require internal coatingsof piping and wellbore tubulars. Although such coatings and linings might mitigate someof the effects of corrosion, even the most corrosion-resistant materials might ultimatelysuccumb to deterioration, and care must be taken to carefully monitor the condition of allpiping and well materials (see Figure 22). Because water might form an electrolyte andenhance the corrosion rate, it might be desirable to dehydrate the injected air. In the oiland gas industry, use of dehydrated natural gas streams has been shown to controlcorrosion and stress corrosion cracking.General aerobic bacteria (GAB) such as Thiobacillus thioxidans (sulfur oxidating) canflourish in a CAES environment. Such species might oxidize native sulfur to sulfuricacid, which might have detrimental effects on wellbore tubulars and casing cement.Presence of these bacteria can result in localized corrosion and pitting of steel surfaces.Free-floating planktonic species might be present as well, which could be detrimental toFigure 22 This photograph, from the Huntorf CAES facility inGermany, shows where the protective fiberglass-reinforced plastictubing fractured. [35]formation permeability. Care must be taken to avoid contamination of the reservoirduring drilling operations including careful choice of drilling fluids. To controlpopulations of preexisting bacterial species, biocides might be injected into the air streamonce relevant species have been identified. Comprehensive reviews of reservoir analysistechniques for the detection of corrosion causing bacteria are available in the literature[80].55 Compressed Air Energy Storage, Succar and Williams Arl20April 20083.6. Flow in AquifersThe dynamics of air flow are important for determination of storage energy density andprediction of air -water interface evolution during initial bubble development andsubsequent storage operation. The deliverability calculation outlined above (see section3.2.2.1) is simply a static calculation of airfiows, but in reality the flow conditions willevolve as the bubble size fluctuates. This in turn will impact the storage energy densityand reservoir volume requirement for CABS. While detailed analysis of aquifer flowbehavior is outside the scope of this report, it is useful to highlight some basic conceptsand discuss the impact on aquifer dynamics on CABS design and operation.Use of aquifers for air storage differs greatly from other storage options due to the limitedmobility of fluids through porous media. Hard rock caverns and solution mined saltformations can be described as rigid, open-space containers where pressure changesquickly equilibrate throughout the volume. However, flow through porous reservoirsresults in dynamic pressure gradients throughout the formation that evolve over hours,days or weeks. Steady-state deliverability estimates are useful, but operational planningmust take into account the effects of unsteady-state and pseudosteady-state air flowswithin the reservoir. The dynamics of these flow modes and the deviations of airflowbehavior from steady state conditions are determined by the propagation rates of pressuregradients through the reservoir.The injection or withdrawal of air at the wellbore introduces pressure pulses within theformation that propagate according to the viscosity of the fluid, the size of the pressuregradient, as well as the permeability and porosity of the reservoir. As a pressure gradientpropagates through the formation, the pressure within the formation varies as a functionof both time and location. This condition, called unsteady-state flow, persists until a flowboundary is reached.When airflow is impeded (e.g. by the air-water interface, a permeability pinch-off, thepresence of an adjacent well or some other flow constraint) the pressure throughout thereservoir will vary uniformly with time. This flow condition is called pseudosteady-stateflow and the edge of this advancing pressure gradient is called the radius of drainage (rd).Under pseudo steady-state flow the rate of change of pressure is uniform within theformation (i.e. independent of radial distance from the wellbore).Van Everdingen and Hurst developed expressions for the evolution of aquifer pressuresunder unsteady-state conditions subject to constant terminal pressure and pseudosteady-state in a finite reservoir [81]. The radius of drainage is described in terms of thestabilization time (hours) for the reservoir to transition to a pseudosteady-state flowcondition [75] and the time for the radius of drainage to reach a radial distance r isexpressed as:kpwherer = radial distance from the well borela. viscosity (cp)56 Compressed Air Energy Storage, Succar and Williams Arl20April 20084)= porosityk =permeability (md)p = mean pressure between the weilbore and the radius of drainage (psia)Typical values for tstab over small distances are of the order of hours. Over significantfractions of a kilometer, tstb will typically be of the order of days. The speed at which thepressure gradient evolves impacts the relevant flow regime at a given time. Moreimportantly, it is clear that whether the reservoir is managed under "unsteady-state" or"pseudo-steady-state" flow conditions, true "steady-state" flow cannot occur in aquifersand hence aquifers cannot operate efficiently as compensated, constant-pressure systems(see section 2.2 above) [79].The flow of water through the formation follows the same behavior described above, butdue to the much larger viscosity of water and forces acting at the water interfaces, thestabilization time will be 20 to 100 times longer. Thus, the bubble movement will occurover time scales of days/weeks and the initial bubble development will typically takeseveral months. Consequently, the impact of air-water interface migration will typicallybe most relevant during initial bubble development and for seasonal storage applications.Such considerations imply that over the time scales necessary to balance wind, the bubblewill not change appreciably in shape or extent [19]. Aquifer CAES systems can thereforebe approximated as rigid, constant-volume systems when determining the storage volumenecessary to provide a given storage capacity (see section 2.3, "Storage VolumeRequirement").3.7. ParticulatesWhen particulates are 'generated around the wellbore, they can be carried in the air flowto the CABS turbomachinery where they might damage the turbine blades and othersensitive equipment. The ability of the air to transport particles depends on the air flowrate, the particle size distribution, and the distance of particle formation from thewellbore. Previous studies have shown that because of the high flow rates that would betypical for CABS, the air stream will be able to pick up particles of nearly any size thatare generated within a few feet of the wellbore [60].The generation of particulates in the reservoir can come about via a number of differentmechanisms. As mentioned above, the dissolution of minerals that act as cement betweensand grains can generate free particles that can be entrained into the air stream. Inaddition, injection of air, especially at elevated temperatures, can lead to dehydration anddestabilization of clays that might lead to particulate formation.Several approaches can be taken to mitigate particulate damage on turbomachinery.Particle filtration units are available for any size particle, but the capital cost and energypenalty increases steeply for small particle sizes. Alternatively, injecting a silica solutioninto the formation can cement the grains in the structure. This is commonly done in thenatural gas storage industry to preclude the formation of particles in loosely heldsandstones. The procedure gives rise to only a slight change in permeability and costsonly about $25,000 ($1982) per well [60].57 Compressed Air Energy Storage, Succar and Williams Arl20April 20084. WindlCAES Systems in Baseload Power MarketsThis section addresses the emissions, and economics of baseload wind/CABS systems toillustrate the prospective importance of developing CAES, and especially aquifer CABS,for baseload power applications based on wind. These systems are compared to baseloadpower systems, giving emphasis to economics under a climate change mitigation policy.Baseload power is typically provided by technologies such as conventional coal andnuclear generation. Although wind has a low, stable short run marginal cost, thevariability of wind implies that it is unable to deliver firm power at similar capacityfactors (@70-90%) without some form of backup generation. However a baseload powersystem made up of wind power plus dispatchable backup generation can be compared toother baseload generation options.Two options for backing wind are utilizing dedicated stand-alone natural gas capacity andCABS. Natural gas capacity is chosen as the stand-alone backup generation technologydue to its low capital costs and its fast ramping rates that are well suited to balancingrapid fluctuations in wind power output.To illustrate the potential benefits of these baseload wind options, costs are comparedwith those of three other baseload power systems: coal integrated gasification combinedcycle (IGCC) with CO2 vented (IGCC-V), coal IGCC with CO2 captured and stored(IGCC-C) and natural gas combined cycle (NGCC).A~lthough coal IGCC power is currently more costly than coal steam-electric power, theincremental cost of CO2 capture and storage (CCS) is less for IGCC plants (via pre-combustion CO2 capture) than for steam-electric plants (via post-combustion CO2capture). Furthermore, the total generation cost of coal IGCC power with CCS tends to beless than that of coal steam-electric power with CCS--at least for bituminous coals [82].Thus coal IGCC-C is likely to be the major competitor that wind/CABS will face in aworld with a climate policy in place.Costs are presented for greenhouse gas (GHG) emissions prices of $0 and $31 per tonneof CO2 equivalent --the first carbon price for the current situation where there is noclimate change mitigation policy and the second carbon price representing a GHGemissions valuation that is likely to characterize a climate change mitigation policy. (AGHG emissions price $3 1/tCO2 is the minimum price on GHG emissions needed tomake a coal IGCC-C plant with storage of CO2 in deep saline aquifers competitive with acoal IGCC-V plant (see Table 8) [83, 84].4.1. MethodologyLevelized generation costs for alternative baseload power systems are estimated using thefinancing model in the EPRI Technical Assessment Guide [85]. The assumed financingparameters are 50% debt (9%/y nominal return) and 45% equity (12%/y nominal return),a 30-year (20-year) plant (tax) life, a 3 8.2% corporate income tax rate, a 2%/y propertytax/insurance rate, and a 2.3 5%/yr inflation rate. Under these conditions the discount rate(real weighted after-tax cost of capital) is 6.72%/year, and the levelized annual capitalcharge rate is 13.3%/year. It is assumed that plant construction requires four years(except wind capital which is built over one year), with the capital investment committed58 Compressed Air Energy Storage, Succar and Williams' pil20April 2008in equal annual payments, so that interest during construction factor (IDCF) is 1.0687with Base Case financing.20 All costs are expressed in 2006 inflation-adjusted U.S.dollars.Table 5: Coal IGCC System ParametersaIGCC-V IGCC-CFate of CO2 Vented CapturedCapacity Factor 80%Levelized Annual Capital Charge Rate (%) 13.3Coal Price ($/GJ HHV) 1.65Installed Capacity (MWe) 640.3 555.7CO atr rcin()0.00 90Fixed Operation and Maintenance ($ikW-yr) 34.81 43.16Variable Operation and Maintenance ($/MWh) 6.40 7.98Efiiec LH/HV )(39.6/38.2) (33.7/32.5)CO2 Transport/Storage ($/tCO2) 0 5.0Overnight Construction Cost ($/kWo) 1789 2358aAll IGCC performance/cost estimates are for a water-slurry-fed single-stage GEE gasifier, which iscurrently the least cost 1GCC option with CO2 capture and storage. Data adapted from NETL 2007 [84]and expressed in 2006$.20 Televelized annual capital charge = LACCR*IDCF*OCC, where LACCR = 13.3%/year, IDCF=1.0687, and OCC = overnight construction cost.59 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Table 6: Wind System ParametersWind/CAES [ Wind/GasInstalled Baseload Capacity (MWe) 2000Levelized Annual Capital Charge Rate (%) 13.3System Capacity Factor (%) 85Natural Gas Price ($/GJ HHV) 6.00Wind Farm Rated Power (MWe) 3130 2000CAES Expander Capacity (MWe) 1270 0CAES Compressor Capacity (MWe) 1130 0SC Capacity (MW0) 0 234CC Capacity (MWe) 0 1700Storage Capacity at CABS Expander Capacity(Hours) 88 0Wind Turbine Specific Rating [86] 1.21 1.36Transmission Loss Over 500 km (%) b 3.39 3.06Transmission Line Capacity Factor AfterLosses for 85% System Capacity Factor (%) 85 42.2Wind Energy Transmitted Directly for 85%System Capacity Factor (TWh/y) 10.3 7.40Wind Energy Input to CABS at 85% SystemCapacity Factor (TWh/y) 2.97 0CAES Output Power (TWh/y) 4.46 0SC Power Output (TWh/y) 0 0.239CC Power Output (TWb/y) 0 7.26CABS Charging Energy Ratio (CER) 1.5 0CAES Heat Rate (k J/kWh) 4220 0SC Heat Rate (kJ/kWh) 0 9020CC Heat Rate (kJ/kWh) 0 6680Wind Capital Cost at Nominal Rating $/kWea $1241/kW $1241/kWCABS Capital Costa ___________Cost of CABS surface turbomachinery andbalance of plant capital ($/kWe) a 6100Capital cost of incremental storage capacity(S/kWh) 1.95 0SC Overnight Construction Cost ($/kWe) a -0 410CC Overnight Construction Cost ($/kWe) a 0 611a Wind turbine costs based on [31], CABS costs based on [11, 12], SC and CC costs based on [87]Installed Capacity for systems with dedicated transmission lines reflects the discharge capacity at the endof the transmission line after losses.b Transmission losses are expressed as a fraction of transmitted energy at the source of generation. Sincetransmission here reflects a differential in transmission distance, converter losses are not included. Suchlosses would add an additional 0.75% of loss at each terminal.Energy quantities are expressed on a lower heating value (LHV) basis, except energyprices are on a higher heating value (HI-IV) basis--the norm for US energy pricing.Energy prices are assumed to be $1.65/GJ for coal and $6.00/GJ for natural gas [87]. TheGHG fuel emissions include the CO2-equivalent upstream GHG emissions (3.66 kg CO2per GJ of coal and 10.4 kg CO2 per GJ of natural gas [88]), resulting in total CO2-equivalent GHG emissions rates of 93.0 kg CO2 and 66.0 kg CO2 per GJ of coal andnatural gas, respectively.60 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Coal IGCC plant performances, capital costs, and O&M costs are based on 2007 NETLdata [84]. CO2 transport and storage costs are estimated using the model developed byOgden et al [89] (see Table 5). Cost modeling of wind energy systems and transmissionas well as optimization methodology for variable scaling of wind turbine components (i.e.derating) are as described in previous studies unless otherwise noted (see Table 6) [2, 86,90].Although assumptions in this report relating to capital costs reflect the most recentnumbers published in the open literature, the escalation of construction costs continues[91], so that estimated absolute costs may differ from actual realized cost levels for plantsthat might be built. However, construction cost escalation is a phenomenon affectingessentially all energy technologies, and it is reasonable to assume that continuingconstruction cost escalation will not appreciably affect the relative economics among thealternative baseload options considered or the conclusions of this analysis.The cost of electricity (COB) or generation costs is estimated two ways. For the first setof COB estimates presented in Table 8, it is assumed that the power systems are operatedat specified capacity factors. Subsequently, economic dispatch is discussed, which, in realmarkets, has the effect of reducing the capacity factors of systems with high dispatchcosts.4.2. Generation Costs for Alternative Baseload PowerSystems Operated at Specified Capacity FactorsThe CO~s for alternate baseload power systems are presented in Table 8 disaggregatedinto components. The CO~s are compared under three sets of conditions: The first set ofcosts are evaluated without a valuation on GHG emissions , the next set applies a CO2-equivalent GHG emissions price of $3 1/tCO2 and the third includes the cost oftransmitting remote wind supplies 500 kmn to demand centers.Table 7 CO2-equivalent GHG Emission Rates for Alternative Baseload Power Systems (kgCO2/MWh)IGCC-V I IGCC-C Wind/CAES I Wind/Gas I NGCC829 132 86.5 224 440In the absence of a GHG emissions price, IGCC-V is the least costly baseload poweroption, while the cost for wind/CABS is a few percent higher than that of IGCC-C. WhenGHG emissions are valued at $3 1/tCO2, the wind and natural gas options become morecompetitive with the coal options. In this case, wind/gas and NGCC are the least costlybaseload power options. At this GHG emissions price (the breakeven price for IGCC-Cwith respect to IGCC-V), wind/CABS is now has a nearly equivalent cost as both coaloptions. The addition of transmission line costs adds approximately 10% to the levelizedcost of energy to both baseload wind options.The generation cost estimates presented in Table 8 underscore the sensitivity of theresults to the stringency of the climate change mitigation policy and the wind resourceremoteness.61 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Table 8: Disaggregated Generation Costs for Coal IGCC, Baseload Wind and NGCC ($/MWh)IGCC-V IGCC-C Wind/CAES Wind/Gas NGCCCptl36.37 47.94 65.15 39.66 13.49Maintenance 4.95 6.15 3.90 3.95 1.75Variable Operations andMaintenance 6.38 7.99 8.98 5.42 1.94Fuel 15.55 18.27 8.43 22.68 44.53CO2 Transport andStorage 0.00 4.21 0.00 0.00 0.00TotalDispatch Cost 21.93 30.47 17.40 28.09 46.47Total Generation Cost atZrCabnPie63.25 84.55 86.45 71.70 61.72GHG Emissions Costs@$31/tCO2.25.35 4.04 2.64 6.86 13.48Total Dispatch Cost 47.28 34.51 20.04 34.96 59.95Total Generation Cost @ 88.60 88.60 89.09 78.56 75.19$31/tCO2Cost of 500kmn DedicatedTL for Remote Wind a 0.00 0.00 7.23 7.25 0.00Transmission Losses b 0.00 0.00 3.29 1.29 0.00Total Generation Cost'Including TL Cost for 88.60 88.60 99.61 87.11 75.19Remote Wind @ $31/tCO2 ____ ____ _____ ____a This is the TL cost per total MWh of electricity production. Allocated only to the electricitytransmitted, the TL cost for the Wind/Gas option is 95% greater than the TL cost for wind/CAESbecause of the lower TL capacity factor.b Transmission costs based on 500kV bipole technology [92]. Since transmission distance isregarded as differential rather than absolute only the cost of the 500kmn increment are included (i.e.no convertor costs).4.3. Dispatch Competition in Baseload Power MarketsThe ordering of the total generation costs presented in Table 8 does not represent theordering that would occur in real-world power markets, in which capacity factors cannotbe assumed to be fixed at a specified rate. Rather, capacity factors are determined bymarket forces to reflect the relative dispatch costs of the competing options on the electricpower grid.For a given set of power generating systems connected to the electric power grid, the gridoperator determines the capacity factors of these systems by calling first on the systemwith the least dispatch cost. Under this condition, deployment in sufficient quantity of thetechnology with the least dispatch cost can lead to a reduction of the capacity factors andthus an increase in the COEs of the competing options on the grid.The impact of dispatch competition on capacity factors is well known. For example, as aresult of the recent increases in natural gas prices in the U.S. this phenomenon has62 Compressed Air Energy Storage, Succar and Williams Arl20April 2008resulted in reducing capacity factors for natural gas combined cycle plants originallydesigned for baseload operation to average utilization rates in the range 30-50% wherecoal plants are available to compete in dispatch [82].In principle, this downward pressure on capacity factors for options with high dispatchcosts could be avoided with "take-or-pay" contracts that require the generator to providea specified fixed amount of electricity annually. However, uncertainties about future fuelprices, technological change, and future electricity demand make such contracts rare.4.3.1. Dispatch Duration CurvesTable 8 presents average dispatch costs for the options considered. The table shows thatin the presence of a GHG emissions price of $31/tCO2 the total average dispatch cost (i.e.the sum of all short-run marginal costs on average: fuel + variable operations andmaintenance + GHG emissions cost) is the lowest by far for wind/CABS systems.Since dispatch costs determine the relative suitability of alternative options for baseloadoperation, it is necessary to examine closely the dynamics of dispatch. Although to goodapproximation one can assume that the dispatch costs for coal IGCC plants are constant,the dispatch costs for wind-based power systems cannot be treated as simple averages.Dispatch costs for wind-based systems vary from the minimum value (corresponding totimes when all electricity is provided by wind--i.e., when fuel expenditures are zero) andincrease significantly as backup generation comes on line to balance shortfalls in windoutput. Thus, it is important to analyze the variations in dispatch costs for these options,not simply their average value as reported in Table 8.Figure 23 shows the variation in dispatch costs in a manner similar to a "load-duration"curve or, more precisely, as an inverse cumulative probability curve counting from thetop end of the distribution. The choice of horizontal axis (in reverse order from 1 to 0)can be useful since horizontal axis values at the intersection of the wind curves with eachconstant-cost IGCC line indicate the percent of time that it can deliver power at a lowerdispatch cost. These dispatch cost curves are evaluated at both pGEIG=$0/tCO2 and$31l/tCO2 (this is the break-even greenhouse gas emissions price for IGCC-C relative toIGCC-V as is evident from Table 8).4.3.2. ResultsDispatch costs are the same lowest value for both the wind/gas and wind/CABS systemswhen all power comes directly from the wind array (right portion of each plot in Figure23), but dispatch costs rise at very different rates as the fraction of power coming fromthe backup system increases (left portion of each plot). In addition, the wind/CABSsystem has an intermediate dispatch cost regime where CABS compressors are running tostore wind energy that cannot be transmitted; this appears as a step in intermediate rangeson the wind/CABS dispatch cost curve.Figure 23 shows that wind/gas has the highest dispatch cost of all the coal and windoptions when natural gas generation is dispatched in significant quantities to balancewind output. For the portion of the dispatch duration curve corresponding to zero windoutput, the dispatch cost matches the dispatch cost of NGCC as expected. Theserelationships hold true at both valuations of GHG emissions assumed in Figure 23. At63 Compressed Air Energy Storage, Succar and Williams Arl20April 2008$0/tCO2 wind/gas cannot compete in economic dispatch relative to the lowest cost coaltechnology for more than 35% of the time and even at $31/tCO2 it will be competitiveless than 40% of the time. Hence a baseload-level capacity factor cannot be sustainedwith wind/gas if either coal or wind/CABS capacity is available in significant quantity onthe grid. Thus in light of current and prospective high natural gas prices, it is unlikely thatPGfIG %GHG = $31iC'O2-Wind/Gas-W'm4/CAES60. -"-ICIGCGCCwCCS Vet60--'. ---""' .. .. ...--- NGCC50 50.1 0.8 0.6 0.4 0.2 0 1 0.8 0.8 0.4 0.2 0Cumi Prob, Dispatch Cost~s Y Cuml Prob, Dispatch Cost s YFigure 23 Dispatch costs for the four alternative power systems for two valuations of GHGemissionswind/gas will be a viable baseload power option for the near future.2'In contrast, because wind/CAES systems have a lower heat rate (4220 k J/kWh) andbecause direct energy from wind accounts for a larger fraction of the output (see Table6), they are able to run at a lower dispatch cost than both coal options more than 70% ofthe time at $0/tCO2 and more than 85% of the time at $3 1/tCO2.Thus, via dispatch competition, wind/CAES systems can be highly competitive with coalpower systems--especially in the presence of a substantial valuation of GHG emissions.An economic model of the entire electric power system is needed to determine thecapacity factors of coal power plants on the grid resulting from dispatch competition.Although such modeling is beyond the scope of this report, it is clear that the averagecapacity factor for coal systems would decline as more and more wind/CAES power isadded to the grid. At a GHG emissions valuation of $31/ItCO2, the COB for awind/CABS system at 85% capacity factor would be lower than for an IGCC-C system21 Wind power backed by existing reserve capacity might still be cost-effective in serving intermediate loadapplications, especially where diurnal variations in wind speed are positively correlated with electricitydemand. However, analysis of intermediate load markets is outside the scope of this report.64 Compressed Air Energy Storage, Succar and Williams Arl20April 2008when the latter has a capacity factor less than 79% when both systems are equally distantfrom major electricity markets or less than 71% when the wind supply is more distant by500 km.The coupling of wind farms to large scale storage technologies such as CAES opens thedoor to participation in baseload markets for both wind and natural gas--especially in thepresence of a strong climate change mitigation policy. The variability of wind makes itimpossible for a "pure" wind system to provide baseload power. Moreover, current andprospective high natural gas prices exclude natural gas combined cycle power technologyfrom providing baseload power if there is a substantial amount of coal power on the grid.But coupling wind to CABS makes it possible for wind to deliver firm power. And theuse of wind to provide compressor energy results in fuel consumption that is sufficientlylow for wind/CABS to be competitive with coal in economic dispatch. This represents animportant opportunity for both wind and natural gas to compete in baseload powermarkets, and opens the door to an important option for realizing cost-effectively deepreductions in GHG emissions from the power sector.65 Compressed Air Energy Storage, Succar and Williams Arl20April 20085. Advanced Technology OptionsAlthough commercial CAES plants have been operating for several decades, thetechnology is still in an early state of development. This is reflected in the fact that thetwo existing plants are based largely on conventional gas turbine and steam turbinetechnologies. Consequently, various technological improvements might be pursued toenhance performance and reduce costs over relatively few product cycles.start (stop)LP discharge150 MW cmrso2OO gis omprnor250 MW20"C i ~Prda. 620"C (565"C)oomprssof450 kgls (500 kgls)Figure 24 A possible technical concept for an AA-CAES system under development [38]One option that has attracted interest is to reduce (and perhaps eliminate) the CABS fuelrequirements and associated GHG emissions by recovery and storage of the high-qualityheat of compression in thermal energy storage (TES) systems. Heat recovery could beimplemented at some or all compression stages, which would then allow stored heat to beused in place of fuel to reheat air withdrawn from the CABS cavern thereby eitherpartially or completely eliminating the need for natural gas [65]. In order to be economic,the fuel cost reductions must offset the additional capital cost associated with the TESsystem. Early studies found that very high fuel prices would be required to justify suchsystems making adiabatic CABS too costly for commercial use [93-97].More recent studies however suggest that new TES technologies, together withimprovements in the compressor and turbine systems might make so-called AdvancedAdiabatic CABS (AA-CAES), economically viable [9, 98]. One such AA-CAES conceptwith a high efficiency turbine and a high-capacity TES, achieves a round trip efficiencyof approximately 70% with no fuel consumption (see Figure 24) [38]. But it should benoted that the efficiency gain of adiabatic systems over multistage compression withintercooling is small (see Appendix A), and both the fuel use and GHG emissions forwind/CABS systems are already very modest (see Table 7).66 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Table 9 The main thermal energy storage (TES) concepts considered for AA-CAES [98] aConcept Solid TES Liquid TESRock Cowper- Concrete Cast 'Hybrid'- Two 1-Tank Air-bed Derivative Walls Iron phase- Tank Thermo- LiquidSlabs change clinematerialsContact Direct Direct Direct Direct Direct Indirect Indirect IndirectStorage Natural Ceramics Concrete Cast Ceramics, Nitrate Nitrate NitrateMaterials Stone Iron Salt Salt, Salt, Salt,Mineral Mineral Mineral___ __ _ __ _ __ __ _ __ __ __ __Oil Oil Oila. Storage technologies chosen on the basis of the capability to deliver 120-1200 MWh (thermal), maintainhigh consistency of outlet temperature, and cover the full temperature range (50 to 650°C)Another proposal is to use biomass-derived fuels to reheat the air withdrawn fromstorage. This could reduce GHG emissions and decouple the plant economics from fuelprice fluctuations [99]. This might also allow CAES to be run on fuel produced locally,thereby facilitating the utilization of energy crops in remote, wind-rich areas andeliminating the need to secure natural gas supplies. However, as in the adiabatic case, theemissions benefit would be small because the emissions level of wind/CABS is alreadyquite low (~- 2/3 the rate for a coal IGCC plant with CCS, see Table 7). Moreover, abiofuels plant dedicated to a wind/CABS system would require fuel storage, becausebiofuels must be produced in large-scale plants that are run flat-out in order to be costeffective, while CABS expander capacity factors for backing wind will typically bemodest (see Table 6)A CABS variant proposed for wind applications is to replace the electrical generator inthe wind turbine nacelle with a compact compressor. So doing would enable the windturbine to generate compressed air directly, thereby eliminating two energy conversionprocesses.2 However, the reduced losses and potential drop in turbine capital cost wouldhave to offset the added capital cost of the compact compressors and the considerablecost of the high pressure piping network needed to transport the compressed air fromeach turbine to the storage reservoir.In contrast to the option of coupling intermittent wind to CAES to enable the provision ofbaseload electricity, CABS might also be coupled to baseload power systems to facilitatethe use of such systems to provide load-following and/or peaking power, the functionoriginally envisioned for CAES--e.g., by coupling CABS to a coal IGCC plant [100,101].Improving CABS turbomachinery is a promising area for innovation. CABS turbineoperating temperatures might be increased, thereby increasing their efficiency byintroducing turbine blade cooling technologies routinely deployed in conventional gasturbines but not in commercial CABS units. Other advanced CABS concepts includevarious humidification and steam injection schemes which can be used to boost thepower output of the system and reduce the storage requirement [ 102]. The CABS22The company General Compression is currently developing this technology.67 Compressed Air Energy Storage, Succar and Williams Arl20April 2008combined cycle is still another option that allows the system to generate electricity evenwhen the compressed air storage reservoir is depleted [ 103, 104].A recent hybrid CAES system design incorporates a standard combustion turbine in placeof the turboexpander chain in a traditional CABS design. The air withdrawn from storageis heated by means of a recuperator at the turbine exhaust instead of by way of fuelcombustors as in a conventional CAES plant. The heated air is then injected into theturbine to boost the output. The use of commercial technology and the elimination of fuelcombustors could reduce the capital cost of the system substantially and provide a lowrisk option for early adoption of bulk storage. Such an Air-Injection CABS (Al-CABS)plant could also include a bottoming cycle and TES system to reduce fuel consumptionfurther [52, 105].Although it is possible that new CABS concepts will bring important changes to the wayair storage operates or the way wind power is stored, performance/cost gains are mostlikely to arise in the near term as a result of marginal improvements in existing CABSdesigns as a result of learning by doing. Thus, after technology launch in the market,costs for new technologies such as CABS can be expected to decline at faster rates thanfor mature technologies and more quickly the faster the rate of deployment. Thisphenomenon bodes well for wind/CABS as a baseload power climate change mitigationoption if there is a way forward that offers opportunity for substantial early marketexperience.68 Compressed Air Energy Storage, Succar and Williams Arl20April 20086. A Way ForwardAlthough the exploitable global wind potential is sufficient to meet total electricitydemand several times over, the future role of wind will ultimately be determined by theextent to which the temporal variability and resource remoteness challenges can beaddressed. Compressed air energy storage is a potential solution, but to evolve from thetwo commercial-scale CABS plants in the field today to wide-scale deployment of thistechnology requires clarification of several issues.Widespread deployment of CAES will depend on the availability of suitable geologiesthat can be developed economically to provide the needed storage capacity. The twoexisting commercial CABS plants at Huntorf and Mclntosh both use salt dome storagebut, as Figure 17 shows, regions with domal salt formations do not have significantoverlap with high quality wind resources. Bedded salt and hard rock geologies overlapwell with windy areas (see Figure 7 and Figure 17), but there are challenges associatedwith each, namely structural issues in the case of salt beds and the high cost of miningnew caverns in the case of hard rock (see section 1.3). Developments in miningtechnology may reduce the cost of using hard rock storage reservoirs making this geologya viable option for future CAES systems. However, porous rock formations can currentlybe developed at a much lower cost and appear to be available in many windy areasthroughout the continental US and thus are the most likely candidate for coupling CABSwith wind capacity in the near term.Although the geographical distributions of good wind resources and potential aquiferstorage opportunities seem to be well correlated (see Figure 17), this broad-brushjudgment must be buttressed by detailed assessments of specific aquifers and local,facility-sized structures in the aquifers. In the necessary detailed resource assessments,clarification is needed of the extent of anticlines with suitable characteristics(permeability, caprock thickness, etc) among the porous rock formations of the regionswhere there are good wind resources and of the geochemical suitability of variousformations for storing air. Data on local geology from US and state geological surveysincluding natural gas storage candidate site evaluations might aid in furthercharacterizing these areas, but new data will also be needed especially in regions wherenatural gas storage is not commonplace.The planned wind/CABS system in Iowa will help to establish the viability of aquiferCABS, but as indicated in section 3, the suitability of a porous rock formation for CABSdepends on a host of geologic factors. As such, it will be important to demonstrateseveral commercial scale systems to ensure that CABS technology can be developed in asufficiently broad set of geologic conditions as to have the potential for widespreaddeployment.Finally, direct coupling of CABS with wind farms will present challenges not faced intoday's CABS systems. The system at Huntorf is primarily used for peaking services andthe Mclntosh system charges storage at night and provides output during the day. This isin contrast to the higher frequency fluctuations imposed by wind power and the morerapid switching between compression and generation modes needed to back up windpower.69 Compressed Air Energy Storage, Succar and Williams Arl20April 2008The use of CAES in an intermediate load application such as that envisioned for the Iowawind/CABS plant will provide a valuable demonstration of wind/CABS integration.However, demonstration of a much more closely coupled system capable of servingbaseload markets is also needed to understand better the potential of wind/CABS fordisplacing new coal capacity in a carbon constrained future. Ultimately the role of windas a tool for climate mitigation will depend on the extent to which it will be able tosupplant new baseload coal-fired capacity.70 Compressed Air Energy Storage, Succar and Williams Arl20April 20087. ConclusionsTraditionally, CAES technology has been used for grid operational support applicationssuch as regulation control and load shifting. But a new major possibility that is especiallyrelevant for a carbon constrained world is to enable exploitation at large intermittent windresources that are often remote from major electricity demand centers. CAES appears tohave many of the characteristics necessary to transform wind into a mainstay of globalelectricity generation.Backing wind to produce baseload output requires short response times to accommodatefluctuations in compressor power and turbine load. The ability of a CAES system to rampoutput quickly and provide efficient part-load operation make it particularly well suitedfor balancing such fluctuations--key performance characteristics that are not often calledupon at existing CAES plants that simply store low-cost off-peak electricity for use whenelectricity is more valuable.Air storage volume requirements translate into a geologic footprint -15% of the windfarm land area, so that CAES will have relatively limited impact on land use and ecology.The wide availability of potentially suitable geology in wind-rich areas points to CABSas a technology well-suited for making baseload power from wind--thereby making itfeasible to provide wind power at electric grid penetrations far greater than 20%+/-penetration rates that are feasible without storage. And, to the extent that wind-richregions are remote from major electricity markets, such baseload power can often bedelivered to distant markets via high voltage transmission lines at attractive costs.Aquifer CABS seems to be the most suitable storage geology for wind/CABS in the USdue to the potential for low development costs and because regions with porous rockgeologies are strongly correlated with the onshore wind-rich regions of the US.Aquifer CABS technology has been studied for nearly three decades, but the firstcommercial plant was only recently formally announced. Nevertheless, a great deal ofcommercial experience can be gleaned from the natural gas storage industry, which usesgeologies similar to those needed for CABS to meet seasonal heating demandfluctuations. The methodologies for evaluating natural gas Storage reservoirs have beenshown to be directly applicable to aquifer CABS development, but several differencesbetween use of methane and air as a storage fluid must be taken into account. Care mustbe taken to carefully characterize local mineralogy, existing bacterial populations andrelevant corrosion mechanisms in order to anticipate any problems resulting from theintroduction of air into porous underground media. Methods for mitigating the impact ofthese factors such as air dehydration, particulate filtration or biocide application couldhelp to expand the number of suitable sites. Despite the various issues that must be takeninto account, none obviously diminish CABS as a strong candidate option for windbalancing..The planned wind/CABS plant in Iowa will provide valuable experience both with anaquifer as a storage. medium and with operating a CABS system under conditionssomewhat different from those at Huntorf and McIntosh due to the coupling of CABSwith variable wind power.71 Compressed Air Energy Storage, Succar and Williams Arl20April 2008However, understanding the large-scale deployment potential of CABS will require botha more detailed characterization of existing porous rock formations as well as operationalexperience from multiple plants over a wide variety of geologic conditions.An economic analysis of wind/CABS systems shows that its costs would be very similarto costs for other baseload power options offering low GHG emissions. The dispatch costof wind/CABS systems is low enough to defend a baseload (-85%) capacity factoragainst other low carbon generation technologies such as coal IGCC with CCS.Furthermore, the fact that few commercial CABS systems exist suggests that significant-cost reductions are likely to be realizable over relatively few product cycles of experiencevia "learning by doing"The storage of energy through air compression offers the potential to enable wind to meeta large fraction of the world's electricity needs competitively in a carbon constrainedworld. If the needed resource assessments and system studies are completed soon, itshould quickly become evident just how large this fraction might be.72 Compressed Air Energy Storage, Succar and Williams Arl20April 20088. References[1] A. J. Cavallo, "High-capacity factor wind energy systems," Journal of SolarEnergy Engineering, Transactions of the ASME, vol. 117, pp. 137-143, 1995.[2] J. B. Greenblatt, S. Succar, D. C. Denkenberger, R. H. Williams, and R. H.Socolow, "Baseload wind energy: modeling the competition between gas turbinesand compressed air energy storage for supplemental generation," Energy Policy,vol. 35, pp. 1474-1492, Mar 2007.[3] J. F. DeCarolis and D. W. Keith, "The economics of large-scale wind power in acarbon constrained world," Energy Policy, vol. 34, pp. 395-4 10, Mar 2006.[4] I. Arsie, V. Marano, G. Rizzo, and M. Moran, "Energy and Economic Evaluationof a Hybrid Power Plant with Wind Turbines and Compressed Air EnergyStorage," in 2006 ASME Power Conference, Atlanta, GA, United States, 2006.[5] P. Denhotm, G. L. Kulcinski, and T. Holloway, "Emissions and energy efficiencyassessment of baseload wind energy systems," Environmental Science andTechnology, vol. 39, pp. 1903-1911, 2005.[6] A. Cavallo, "Controllable and affordable utility-scale electricity from intermittentwind resources and compressed air energy storage (CAES)," Energy, vol. 32, pp.120-127, Feb 2007.[7] N. Desai, S. Nelson, S. Garza, D. J. Pemberton, D. Lewis, W. Reid, S. Lacasse, R.Spencer, L. M. Manning, and R. Wilson, "Study Of Electric Transmission InConjunction With Energy Storage Technology," Lower Colorado RiverAuthority, Texas State Energy Conservation Office August 21, 2003.[8] N. Desai, S. Gonzalez, D. J. Pemberton, and T. W. Rathjen, "The EconomicImpact of CAIES on Wind in TX, OK, and NM," Ridge Energy Storage & GridServices L.P., Texas State Energy Conservation Office June 27, 2005.[9] G. Salgi and H. Lund, "Compressed air energy storage in Denmark : a feasibilitystudy and an overall energy system analysis," in World renewable energycongress !X Florence, Italy, 2006.[10] D. J. Swider, "Compressed air energy storage in an electricity system withsignificant wind power generation," Ieee Transactions on Energy Conversion,vol. 22, pp. 95-102, Mar 2007.[11] EPRI-DOB, "Handbook of Energy Storage for Transmission and DistributionApplications," EPRI, DOE, Palo Alto, CA, Washington, DC 2003.[12] EPRI-DOE, "Energy Storage for Grid Connected Wind Generation Applications,"EPRI, DOE, Palo Alto, CA, Washington, DC 1008703, 2004.[13] S. M. Schoenung, "Characteristics and Technologies for Long-vs. Short-TermEnergy Storage: A Study by the DOE Energy Storage Systems Program," SandiaNational Laboratories, Albuquerque, New Mexico SAND2001-0765, March2001.[14] F. R. McLarnon and E. J. Cairns, "Energy-Storage," Annual Review of Energy,vol. 14, pp. 241-271, 1989.[15] A. Gonzalez, B. 03. Gallachdir, E. McKeogh, and K. Lynch, "Study of ElectricityStorage Technologies and Their Potential to Address Wind Energy Intermittencyin Ireland," Sustainable Energy Research Group, Department of Civil andEnvironmental Engineering, University College Cork RE/HIC/03/001, May 2004.73 Compressed Air Energy Storage, Succar and Williams Arl20April 2008[16] L. W. M. Beurskens, M. d. Noord, and A. F. Wals, "Analysis In The FrameworkOf The Investire Network: Economic Performance Of Storage Technologies,"Energy Research Centre of the Netherlands (ECN), Petten, The Netherlands ECN-C--03-132, December 2003.[17] CERA, "Energy Storage An Emerging Competitor in the Distributed EnergyIndustry," Cambridge Energy Research Associates, Cambridge, Massachusetts,USA 2002.[18] EPRI, "Wind Power Integration: Energy Storage for Firming and Shaping," EPRI,Palo Alto, CA 1008388, 2005.[ 19] D. L. Katz and E. R. Lady, Compressed air storage for electric power generation.Ann Arbor, Michigan: Ulrich's Book Inc., 1976.[20] Z. S. Stys, "Compressed air storage for load leveling of nuclear power plants," inProceedings of the 12th Intersociety Energy Conversion Engineering Conference.vol. II Washington, DC, USA: American Nuclear Soc, 1977, pp. 1023-30.[21] K. G. Vosburgh, D. C. Golibersuch, P. M. Jarvis, J. A. Bast, J. H. Eskesen, A. S.Mitchell, B. J. Sosnowicz, S. Serata, and J. Pepper, "Compressed air energystorage for electric utility load, leveling," in Proceedings of the 12th IntersocielyEnergy Conversion Engineering Conference. vol. II Washington, DC, USA:American Nuclear Soc, 1977, pp. 10 16-22.[22] B. Sorensen, "Dependability of Wind Energy Generators with Short-TermEnergy-Storage," Science, vol. 194, pp. 935-937, 1976.[23] H. Holttinen, B. Lemstr~im, P. Meibom, H. Bindner, A. Orths, F. v. Hulle, C.Ensslin, L. Hofmann, W. Winter, A. Tuohy, M. O'Malley, P. Smith, J. Pierik, J.0. Tande, A. Estanqueiro, J. Ricardo, E. Gomez, L. S~der, G. Strbac, A. Shakoor,J. C. Smith, B. Parsons, M. Milligan, and Y.-h. Wan, "Design and operation ofpower systems with large amounts of wind power: State-of-the-art report "VTTTechnical Research Centre of Finland, Vuorimiehentie, Finland VTT-WORK-82,October 2007.[24] EnerNex, "Wind Integration Study for Public Service Company of Colorado,"EnerNex Corporation May 22 2006.[25] E. A. DeMeo, G. A. Jordan, C. Kalich, J. King, M. R. Milligan, C. Murley, B.Oakleaf, and M. J. Schuerger, "Accomodating wind's natural behavior," IEEEPower and Energy Magazine, vol. 5, pp. 59-67, 2007.[26] R. DeCorso, L. Davis, D. Horazak, J. Molinda, and M. DeCorso, "ParametricStudy Of Payoff In Applications Of Air Energy Storage (CAES) Plants -AnEconomic Model For Future Applications," in Power Gen International Orlando,FL, 2006.[27] IEA, "World Energy Outlook 2006," International Energy Agency, Paris, France2006.[28] J. B. Greenblatt, "Wind as a source of energy, now and in the future,"InterAcademy Council, Amsterdam, Netherlands October 1 2005.[29] D. L. Elliott, L.L. Wendell, and G. L. Gower, "An assessment of the availablewindy land area and wind energy potential in the contiguous United States," U.S.Department of Energy, Pacific Northwest Laboratory, Richland, WashingtonPNL-7789, DE91 018887, August 1991.74 Compressed Air Energy Storage, Succar and Williams Arl20April 2008[30] A. J. Cavallo and M. B. Keck, "Cost effective seasonal storage of wind energy,"Houston, TX, USA, 1995, pp. 119-125.[31 ] R. Wiser and M. Bolinger, "Annual Report on U.S. Wind Power installation,Cost, and Performance Trends: 2006," U.S. Department of Energy, EnergyEfficiency and Renewable Energy, Washington, D.C. DOE/GO-102007-2433,May 2007.[32] P. Zaugg, "Air-Storage Power Generating Plants," Brown Boveri Review, vol. 62,pp. 33 8-347, 1975.[33] K. Allen, "CAES: the underground portion," IEEE Transactions on PowerApparatus and Systems, vol. PAS-104, pp. 809-12, 1985.[34] B. Mehta, "CAES geology," EPRI Journal, vol. 17, pp. 38-41, 1992.[35] F. Crotogino, K. U. Mohmeyer, and R. Scharf, "Huntorf CAES: More Than 20Years of Successful Operation," in Solution Mining Research Institute MeetingOrlando, Florida, USA, 2001.[36] W. F. Adolfson, J. S. Mahan, E. M. Schmid, and K. D. Weinstein, "Geologicissues related to underground pumped hydroelectric and compressed air energystorage," in 14th Intersociety Energy Conversion Engineering Conference,Boston, MA, USA, 1979, pp. 452-4.[37] K. L. DeVries, K. D. Mellegard, G. D. Callahan, and W. M. Goodman, "CavernRoof Stability For Natural Gas Storage In Bedded Salt," United StatesDepartment of Energy National Energy Technology Laboratory Topical ReportRSI-1829, DE-FG26-02NT41651, June 2005.[38] B. Calaminus, "Innovative Adiabatic Compressed Air Energy Storage System ofEnBW in Lower Saxony," in 2nd International Renewable Energy Storageconference (IRES II) Bonn, Germany, 2007.[39] K. Sipila, M. Wistbacka, and A. Vaatainen, "Compressed air energy storage in anold mine," Modern Power Systems, vol. 14, pp. 19-2 1, 1994.[40] 5. Shepard and S. van der Linden, "Compressed air energy storage adapts proventechnology to address market opportunities," Power Engineering (Barrington,Illinois), vol. 105, pp. 34-37, 2001.[41] M. Schwartz and D. Elliot-t, "Remapping of the Wind Energy Resource in theMidwestern United States," in Third Symposium on Environmental Applications,Annual Meeting of the American Meteorological Society Orlando, FL, 2001.[42] M. Schwartz and D. Elliott, "Validation of Updated State Wind Resource Mapsfor the United States.," in Proceedings of the World Renewable Energy CongressVIII (WREC 2004) Denver, Colorado: Elsevier, Ltd., 2004.[43] D. Elliott and M. Schwartz, "Validation of New Wind Resource Maps," in 2002Conference Proceedings of the American Wind Energy Association WindPower2002 Conference Portland, Oregon: American Wind Energy Association, 2002.[44] B. R. Mehta, "Siting compressed -air energy storage plants," in Proceedings ofthe American Power Conference, Chicago, IL, USA, 1990, pp. 73-78.[45] Y. Zimmels, F. Kirzhner, and B. Krasovitski, "Energy loss of compressed airstorage in hard rock," in Fourth International Conference on Ecosystems andSustainable Development ECOSUD IV, Siena, Italy, 2003, pp. 847-857.75 Compressed Air Energy Storage, Succar and Williams Arl20April 2008[46] T. Brandshaug and A. F. Fossum, "Numerical Studies Of Caes Caverns In HardRock," in Mechanical, Magnetic, and Underground Energy Storage 1980 AnnualCon tractors' Review. 00255, Washington, DC, USA, 1980, pp. 206-212.[47] N. Lihach, "Breaking New Ground With Caes," EPRI Journal, vol. 7, pp. 17-21,1982.[48] EPRI, "Compressed-air energy storage using hard-rock geology: test facility andresults. Vol. 1-2," Electr. Power Res. Inst., Palo Alto, CA, USA, Copyright 2005,IEE EPRI GS-6620, 1990.[49] S. van der Linden, "Bulk energy storage potential in the USA, currentdevelopments and future prospects," Energy, vol. 31, pp. 3446-3457, Dec 2006.[50] A. J. Karalis, E. J. Sosnowicz, and Z. S. Stys, "Air Storage Requirements For A220 MWe Caes Plant As A Function Of Turbomachinery Selection AndOperation," IEEE Transactions on Power Apparatus and Systems, vol. PAS- 104,pp. 803-808, 1985.[51] 0. Weber, "Air-Storage Gas Turbine Power Station At Huntorf," Brown BoveriReview, vol. 62, pp. 332-337, 1975.[52] 5. van der Linden, "Review Of CABS Systems Development And CurrentInnovations That Could Bring Commercialization To Fruition," in ElectricalEnergy Storage Applications & Technology (EESAT) Conference San Francisco,CA: Energy Storage Association, 2007.[53] V. De Biasi, "110 MW McIntosh CABS plant over 90% availability and 95%reliability," Gas Turbine World, vol. 28, pp. 26-8, 1998.[54] L. Davis and R. Schainker, "Compressed Air Energy Storage (CABS): AlabamaElectric Cooperative McIntosh Plant -Overview and Operational History," inElectricity Storage Association Meeting 2006: Energy Storage in ActionKnoxville, Tenn.: Energy Storage Association, 2006.[55] OPSB, "Staffs Report of Investigation and Recommended Findings," Ohio PowerSiting Board, Public Utilities Commission of Ohio, Columbus, OH March 202001.[56] OPSB, "Staff Investigation Report and Recommendation," Ohio Power SitingBoard, Public Utilities Commission of Ohio, Columbus, OH March 8 2006.[57] J. A. Strom, "NORTON ENERGY STORAGE, LLC ANNUAL PROJECTPROGRESS STATUS REPORT TO OPSB STAFF," NORTON ENERGYSTORAGE, LLC, Columbus, OH February 22 2007.[58] IAMU, "Site for ISEP Development is Officially Announced," in IOWAASSOCLIATION OF MUNICIPAL UTILITIES NEWSLETTER. vol. 2006, 2006,pp. 1,10.[59] TXU, "Luminant and Shell Join Forces to Develop a Texas-Sized Wind Farm,"July 27 2007.[60] EPRI, "Compressed-Air Energy Storage Preliminary Design and SiteDevelopment Program in an Aquifer," Electric Power Research Institute EM-235 1, November 1982.[61] A. J. Giramonti and B. B. Smith, "Control of the champagne effect in CABSpower plants," Atlanta, GA, USA, 1981, pp. 984-8.[62] B. Elmegaard, N. Szameitat, and W. Brix, "Compressed air energy storage(CABS) -possibilities in Denmark," in 18th International Conference on76 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Efficiency, Cost, Optimization, Simulation and Environmental Impact of EnergySystems Trondheim, Norway, 2005.[63] I. Tusehy, R. Aithaus, R. Gerdes, and P. Keller-Sornig, "Evolution of gas turbinesfor compressed air energy storage," VGB Powertech, vol. 85, pp. 84-7, 2004.[64] EPRI, "Evaluation of Benefits and Identification of Sites for a CABS Plant inNew York State," Energy Storage and Power Consultants, Electric PowerResearch Institute (EPRI), Palo Alto, CA TR-104268, August 1994.[65] E. Macchi and G. Lozza, "Study Of Thermodynamic Performance Of Caes Plants,Including Unsteady Effects," in Gas Turbine Conference and Exhibition,Anaheim, CA, USA, 1987, p. 10.[66] D. R. Hounslow, W. Grindley, R. M. Loughlin, and J. Daly, "The development ofa combustion system for a 110 MW CABS plant," Journal of Engineering for GasTurbines and Power-Transactions of the Asme, vol. 120, pp. 875-883, OCT 1.998.[67] I. Tuschy, R. Althaus, R. Gerdes, and P. Keller-Sornig, "Compressed air energystorage with high efficiency and power output," VDI Berichte, pp. 57-66, 2002.[68] Y. S. H. Najjar and M. S. Zaamout, "Performance analysis of compressed airenergy storage (CABS) plant for dry regions," Energy Conversion andManagement, vol. 39, pp. 1503-1511, OCT 1998.[69] I. Arsie, V. Marano, G. Nappi, and G. Rizzo, "A model of a hybrid power plantwith wind turbines and compressed air energy storage," in 2005 ASME PowerConference, Chicago, IL, United States, 2005, pp. 987-1000.[70] P. Denhoim and G. L. Kulcinski, "Net energy balance and greenhouse gasemissions from renewable energy storage systems," Energy Center of Wisconsin,Madison, WI 223-1, June 2003.[71] P. Zaugg, "Energy Flow Diagrams For Diabatic Air-Storage Plants," BrownBoveri Review, vol. 72, pp. 179-183, 1985.[72] P. Denholm and G. L. Kulcinski, "Life cycle energy requirements and greenhousegas emissions from large scale energy storage systems," Energy Conversion andManagement, vol. 45, pp. 2153-72, 2004.[73] R. Schainker, M. Nakhamkin, J. R. Stange, and L. F. Giannuzzi,"Turbomachinery Engineering And Optimization For 25 MW And 50 MWCompressed Air Energy Storage Systems," Amsterdam, Neth, 1984.[74] FERC, "Current State of and Issues Concerning Underground Natural GasStorage," Federal Energy Regulatory Commission AD04-1 1-000, September 302004.[75] D. L. Katz and K. H. Coats, Underground Storage of Fluids. Ann Arbor, MI:Ulrich's Books, Inc., 1968.[76] lEA, "World Energy Outlook 2004," Inernational Energy Agency, Paris, France2004.[77] AGA, "Survey of Underground Gas Storage Facilities in the United States andCanada," American Gas Association, Washington, DC 20001 Catalog No.XU0404, June 2004..[78] D. L. Katz, D. Cornell, J. A. Vary, R. Kobayashi, F. H. Poettmann, J. A. Vary, J.R. Elenbaas, and C. F. Weinaug, Handbook of natural gas engineering NewYork, NY: McGraw-Hill, 1959.77 Compre~ssed Ai" Energy Storage, Succar and Williams Arl20April 2008[79] EPRI, "Compressed-air energy storage: Pittsfield aquifer field test. Test data:engineering analysis and evaluation," Electr. Power Res. Inst., Palo Alto, CA,USA, Palo Alto, CA, USA EPRI GS-6688, 1990.[80] B. Y. Chen and R. B. Chen, "Monitoring Microbial Corrosion In Large OilfieldWater Systems," JPT, Journal of Petroleum Technology, vol. 36, pp. 1171-1176,1984.[81] A. F. Van Everdingen and W. Hurst, "Application of Laplace transformation toflow problems in reservoirs," American Institute of Mining and MetallurgicalEngineers -- Journal of Petroleum Technology, vol. 1, pp. 305-326, 1949.[82] K. Thambimuthu, M. Soltanieh, J. C. Abanades, R. Allam, 0. Bolland, J.Davison, P. Feron, F. Goede, A. Herrera, M. Iijima, D. Jansen, I. Leites, P.Mathieu, B. Rubin, D. Simbeck, K. Warmuzinski, M. Wilkinson, R. H. Williams,M. Jaschik, A. Lyngfelt, R. Span, and M. Tanczyk, "Capture of C02 " in CarbonDioxide Capture and Storage, a Special Report of the Intergovernmental Panel onClimate Change, B. Metz, 0. Davidson, H. Coninck, M. Loos, and L. Meyer, Eds.New York, NY: Cambridge University Press, 2005.[83] T. Kreutz, R. Williams, S. Consonni, and P. Chiesa, "Co-production of hydrogen,electricity and C02 from coal with commercially ready technology. Part B:Economic analysis," International Journal of Hydrogen Energy, vol. 30, pp. 769-784, Jul 2005.[84] DOE/NETL, "Bituminous Coal and Natural Gas to Electricity," National EnergyTechnology Laboratory DOEiNETL-2007/1 281, May 2007.[85] EPRI, "Technical Assessment Guide (TAG)," Electric Power Research Institute,Palo Alto, CA TR-102275, 1993.[86] 5. Succar, D. Denkenberger, J. B. Greenblatt, and R. H. Williams, "An IntegratedOptimization Of Large-Scale Wind With Variable Rating Coupled ToCompressed Air Energy Storage," Proceedings of A WEA Windpower 2006,Pittsburgh, PA, 2006.[87] EIA, "Annual Energy Outlook 2007 With Projections to 2030," Office ofIntegrated Analysis and Forecasting, Energy Information Administration, USDepartment of Energy, Washington, D.C. DOE/EIA-0383(2007), February 2007.[88] M. Q. Wang, "GREET 1.5 -Transport Fuel-Cycle Model," Center forTransporation Research, Energy Systems Division, Argonne Natinoal Laboratory,Argonne, IL ANL/ESD-39, Vol. 2, August 1999.[89] J. M. Ogden, "Modeling infrastructure for a fossil hydrogen energy system withCO2 sequestration," Proceedings of the 6th International Conference onGreenhouse Gas Control Technologies, Kyoto, Japan, October., 2002.[90] D. Denkenberger, J. B. Greenblatt, and R. H. Socolow, "Optimal Wind TurbineRated Speed Taking Into Account Array Effects and Capacity Factor," ElectricPower 2004.[91] M. W. Chupka and G. Basheda, "Rising Utility Construction Costs: Sources andImpacts," The Edison Foundation, Washngton, D.C. September 2007.[92] M. P. Bahrman and B. K. Johnson, "The ABCs of HVDC transmissiontechnologies," IEEE Power &," Energy Magazine, vol. 5, pp. 32-44, 2007.[93] I. Glendenning, "Compressed air storage," Physics in Technology, vol. 12, pp.103-10, 1981.78 Compressed Air Energy Storage, Succar and Williams Arl20April 2008[94] D. Kreid, "Analysis Of Advanced Compressed Air Energy Storage Concepts," inthermophysics and heat transfer conference Palo Alto, CA: American Society ofMechanical Engineers 1978, p. 11.[95] 5. C. Schulte, "Economics Of Thermal Energy Storage For Compressed AirEnergy Storage Systems," in Proc of the Mech and Magn Energy StorageContract Rev Meet Washington, DC, USA: DOE, (Conf-790854), Washington,DC, 1979, pp. 191-198.[96] R. B. Schainker, B. Mehta, and R. Pollak, "Overview of CAES technology,"Chicago, IL, USA, 1993, pp. 992-997.[97] R. W. Reilly and D. R. Brown, "Comparative Economic Analysis Of Several CaesDesign Studies," in Proceedings of the Intersociety Energy ConversionEngineering Conference. vol. 1 Atlanta, Ga, USA: ASMIE, New York, NY, USA,1981, pp. 989-994.[98] C. Bullough, C. Gatzen, C. Jakiel, M. Koller, A. Nowi, and S. Zunft, "AdvancedAdiabatic Compressed Air Energy Storage for the Integration of Wind Energy,"in European Wind Energy Conference London, UK: EWBA, 2004.[99] P. Denholm, "Improving the technical, environmental and social performance ofwind energy systems using biomass-based energy storage," Renewable Energy,vol. 31, pp. 1355-70, 2006.[100] M. Nakhamkin, M. Patel, E. Swensen, A. Cohn, and B. Louks, "Application of airsaturation to integrated coal gasification/CAES power plants," San Diego, CA,USA, 1991, pp. 1-5.[101] M. Nakhamkin, M: Patel, L. Andersson, P. Abitante, and A. Cohn, "Analysis ofintegrated coal gasification system/compressed-air energy storage power plantconcepts," San Diego, CA, USA, 1991, pp. 1-6.[102] M. Nakhamkin, E. Swensen, P. Abitante, R. B. Schaiker, and R. Pollak,"Technical and Economic Characteristics of Compressed-Air Energy StorageConcepts with Air Humidification," in Second International Conference onCompressed Air Energy Storage San Francisco, CA,' 199.2.[103] K. Yoshimoto and T. Nanahara, "Optimal daily operation of electric powersystems with an ACC-CAES generating system," Electrical Engineering in Japan(English translation of Denki Gakkai Ronbunshi), vol. 152, pp. 15-23, 2005.[104] M. Nakhamkin, E. Swensen, R. Schainker, and R. Pollak, "Compressed air energystorage: survey of advanced CAES development," San Diego, CA, USA, 1991,pp. 1-8.[ 105] M. Nakhamkin, "Novel Compressed Air Energy Storage Concepts," in ElectricityStorage Association Meeting 2006: Energy Storage in Action Knoxville, Tenn.:Energy Storage Association, 2006.79 Compressed Air Energy Storage, Succar and Williams Arl20April 2008Appendix A Theoretical Efficiency of Compressed AirEnergy Storage for Alternative ConfigurationsThe storage efficiency of adiabatic compressors and storage in an insulated cavern iscompared to that of intercooled compressors and storage at ambient temperature.The theoretical maximum efficiency of compressed air energy storage is ratio of themaximum work b (the exergy, in kJ) that can be extracted from 1 kmol of air stored attemperature Ts and pressure P5 to the work wc required to compress 1 kmol of air fromambient temperature To (= 300 K) and pressure P0 (= 1 atmosphere):Vis =b/wc (45)b (Ps,Ts) =h (Ps,Ts) -h (Po,To) -To*[S (Ps,Ts) -s (Po, To)], (46)whereh =air enthalpy, ands =air entropy.Suppose that air is compressed from Po, To to Pc, Tc. Assuming air is an ideal diatomicgas with constant specific heats:k =c%/c = 7/5 = 1.4 (47)where:Cp= specific heat at constant pressure,,,= specific heat at constant volume,the exergy per kmol of compressed air is:b (Ps, Ts) = cp*(Ts -To) -cp*To*ln (Ts/To) + RTo*ln (PsiPo) (48)= RTo*[k/(k -1)]*[[(Ts/To -1) -In (Ts/To)] + [(k -1)/k]*ln (Ps/Po)], (49)where R is the universal gas constant (R = 8314 kJoules/kmole/K).Moreover, assuming a compressor with an efficiency with N stages of adiabaticcompression, with perfect intercooling between stages, and with the optimal compressionratio per stage =(Pc/Po)l/N, the work required to compress a kmnol of air from pressure P0to Pc is:wc = RTo*[Nk/(k -l)]*[(Pc/Po)(k-1)/Nk -1]/ric (50)and Tc is given by:Tc =To*(Pc/Po)(k-1)Nk (51)The theoretical maximum efficiency of storage is thus:TIls (= [(Ts/To- 1)-in (Ts/To)+[(k -1)/k] *ln(Ps/Po)]/[(Pc/Po)(kl)/Nk-1] (52)Case I: Consider first a system with one stage of adiabatic compression (N =1) andperfect insulation of the air storage reservoir, so that Ts = Tc and Ps =Pc. In this case, Tls80 Compressed Air Energy Storage, Succar and Williams Arl20April 2008= Tic, and the highest possible storage efficiency is realized. However, this is not a goodrepresentation of the actual situation where the air in storage is typically cooled to theambient temperature.Case II: Consider next a system with N stages of compression, perfect intercoolingbetween stages, and poor insulation of the storage reservoir so that Ts --) To beforeenergy recovery is attempted. In this case,Ps --) Pc*(To/Tc) = Pc*(Pc/Po)-(kIl)/Nk = P* (Pc/Po)1-(k-1)/Nk (53)b(Ps, Ts) "-) RTo*[1 -1/N + 1/(Nk)]*ln (Pc/Po) (54)andTis = (TIciN)*[(k -1)/k]*[1 -1/N + 1/(Nk)]*In (Pc/Po)/[(Pc/Po)(k-I)/Nk -1] (55)For example, suppose air is compressed to Pc -- 100 atmospheres and N =1, so that Tc =1118 K, and at the time of energy recovery, Ps = (300/11 18)' 100 = 26.8 atmospheres. Inthis case Tls = 0.345"rlc.But if Pc = 100 atmospheres and N =5, Tc =390 K and Ps =77.0 atmospheres at thetime of energy recovery, so that Tis = O.824"aqo.In the limit of an infinite number of stages of compression with perfect intercooling, thecompressor work is isothermal, and the compressor work required is:23wc --) (RTo*/Tic)*ln(Pc/Po), P5 --* Pc, so that Tis "- Tic (5,6)This is the same as for Case I. Thus, via the use of large number of intercoolers, thetheoretical efficiency of a CABS unit with storage at ambient temperature can approachthat of a CABS unit compressing air adiabatically and storing air in an insulated cavern.23 Note that (Xr- 1)/a"-) In Xas a --)O.81}}