Underground coal gasification research and development in the United States

D.W. Camp    Lawrence Livermore National Laboratory, Livermore, CA, United States

Acknowledgments

I thank professors Bill Krantz and Bob Gunn, my MS thesis advisor at the University of Colorado and coadvisor at the University of Wyoming, for introducing me to the fascinating technology of UCG. Lawrence Livermore Lab's UCG team of the 1970s and 1980s deserves special recognition, including Chuck Thorsness, my first UCG summer-job mentor, Doug Stevens, who hired me into LLNL, Bob Cena, Dick Hill, Jerry Britten, and George Metzger, the team's instrumentation hardware guy. More recently, Julio Friedmann championed UCG, grew LLNL's 2005–15 program, and gave me the chance to join it. I thank the many people who contributed to this and especially wish to recognize Jeff Wagoner, John Nitao, Josh White, Ravi Upadhye, Xianjin Yang, Steve Hunter, and visiting professor Evgeny Shafirovich. Finally, I thank Burl Davis, technical director of the Rawlins and Rocky Mountain 1 tests, for sharing his UCG knowledge and stories with me and the community.

4.1 Introduction and scope

An intense and productive period of research and development on underground coal gasification (UCG) took place in the United States from the mid-1970s to the late 1980s. This period began with little domestic understanding of UCG. Early steps included the translation and review of Soviet literature, proposals of large-scale UCG schemes, and some early calculations and modeling that were often not well informed by experimental observations. Field tests became the marquis activities around which an integrated multifaceted program was built. Efforts included extensive monitoring and measurements of field tests, scientific analyses, mathematical modeling, and laboratory experimentation. The period ended with an accurate observation-based conceptual understanding of how UCG works and a corresponding predictive multiphysics mathematical model of the process. Innovative methods and technologies were invented and developed that form the basis of many UCG projects around the world today.

This chapter also covers recent US activities and accomplishments during the period 2004–15 and touches briefly on preceding work between 1948 and 1963.

The goals of this chapter are to summarize what has been done on UCG in the US and, more importantly, what has been learned about UCG from these activities. This chapter is largely a condensed version of a recent comprehensive review by this author of US UCG work that is broader in scope and much longer and more detailed than space allows here (Camp, 2017).

4.2 Major contributing institutions and field-test locations

4.2.1 Bureau of Mines, AEC, ERDA, DOE

Several federal organizations in the United States had an important role in UCG's development. Management and funding of much of the UCG program and some of its main technical contributors resided at various times in the Bureau of Mines (within the Department of the Interior), Atomic Energy Commission (AEC), and Energy Research and Development Agency (ERDA) and finally centralized in the Department of Energy (DOE) into which many of the functions of the other institutions were rolled in 1977.

4.2.2 Bureau of Mines station in Tuscaloosa, Alabama

The Bureau of Mines office in Tuscaloosa, Alabama, had a technology division and experimental station that conducted the Gorgas, Alabama UCG tests in the late 1940s to 1950s.

4.2.3 MERC, METC, NETL

A government research center with expertise in coal gasification in Morgantown, West Virginia, became the Bureau of Mines' Morgantown Energy Research Center (MERC) and then the Morgantown Energy Technology Center (METC) under DOE and, more recently, merged with its Pittsburgh, Pennsylvania counterpart to form the National Energy Technology Laboratory (NETL). METC looked at UCG scale-up and economics, operated the Pricetown field test, and helped DOE manage the UCG program.

4.2.4 LERC, LETC, WRI, Universities of Wyoming and Colorado

The Bureau of Mines Laramie Energy Research Center (LERC), in Laramie, Wyoming, conducted the first and successive UCG tests at the nearby Hanna site. This was renamed to the Laramie Energy Technology Center (LETC) under DOE and then privatized in 1983 into the Western Research Institute (WRI), which remained an important part of the UCG program through the Rocky Mountain 1 test and related follow-on work.

4.2.5 LLL, LLNL

Lawrence Livermore Laboratory (LLL), renamed to Lawrence Livermore National Laboratory (LLNL) in 1982, is in Livermore, California, and is named after its founder, Earnest Lawrence. It is a very large multidisciplinary science and engineering research institution whose core mission since it began in 1952 has been in nuclear security. Livermore's original charter and its continuing nature has been innovation—pioneering new and better ways to do things. That can be seen in its series of UCG activities. Every one of Livermore's field tests was different and pioneered something new.

4.2.6 Universities of Wyoming and Colorado

Researchers at the University of Wyoming (Laramie) and the University of Colorado (Boulder) collaborated with LETC and other institutions, analyzing UCG field data and developing UCG models (c.f. Krantz and Gunn, 1983c).

4.2.7 Texas institutions

A modest amount of UCG work was aimed at lignite fields in the state of Texas. This has mostly been funded by private industry with some state support. Organizations involved include a company named Basic Resources, which operated the Tennessee Colony field test, Texas A&M University in partnership with a consortium of companies, and the University of Texas at Austin.

4.2.8 Gulf Research and Development and Energy International

Gulf Research and Development Company, the R&D arm of the major oil company Gulf, had the most active and longest-running UCG program in American private industry. On a cost-shared basis with the DOE, Gulf ran the two Rawlins field tests and had a key role in the Rocky Mountain 1 test. Some of Gulf's UCG principals later joined Energy International, which aimed at larger UCG projects.

4.2.9 ARCO

The major oil company ARCO had a subsidiary coal company that conducted the Rocky Hill field test and made plans for larger scale operations. Some of its UCG staff had been principals in LETC's UCG program.

4.2.10 Gas Research Institute

The Gas Research Institute (GRI) was an independent research institute in Chicago, Illinois, with expertise that includes coal gasification (mainly surface) that operated on a mix of industrial and government grant funding. GRI funded some of the efforts and was a leader in the DOE-assembled consortium that ran the Rocky Mountain 1 test. More recently, it merged with the Institute of Gas Technology to form the Gas Technology Institute (GTI).

4.2.11 Other institutions

Many other research and engineering institutions participated and made contributions to the UCG program. Many researchers at universities and research institutions conducted laboratory experiments and developed models related to UCG. Many engineering and geologic service providers did much of the support work under contracts, including United Engineers for the Rocky Mountain 1 field test and Williams Engineering for some of the earlier field tests.

4.2.12 Locations

Fig. 4.1 shows the locations of many of the field tests and institutions involved with UCG.

f04-01-9780081003138
Fig. 4.1 Location of US UCG institutions and field tests (Map by LLNL and Ergo Exergy).

4.3 Periods of UCG activities

4.3.1 Pre-1970 work

The only significant work before 1970 seems to be by the US Bureau of Mines at Gorgas, Alabama. This began in the late 1940s, included field tests, and ended with summary reports in the early 1960s (c.f. Capp et al., 1963). Unfortunately, little from this project appeared to become part of the 1970s program consciousness.

4.3.2 The main 1970–80s program and activities

US government institutions (see Section 4.2) initiated UCG research in the early 1970s and grew it into a well-funded, sustained, and technically vibrant program. The primary motivation was to advance the domestic energy security of the United States. Oil and gas supplies were appearing limited, the OPEC oil embargo was to hit soon, and US coal resources were and still are enormous. Throughout its course, the long-term program objective was large-scale commercial operations that would have a significant impact on US energy supplies.

The program began with paper studies and the creation of large-scale conceptual design schemes. An important early activity in the program was the translation and study of Soviet reports on UCG, thus taking advantage of the greatest previous efforts in UCG technology development. This guided early program thinking and approaches.

The US approach was a multifaceted program that included well-instrumented and monitored field tests of modest scale (100–10,000 Mg), large-scale conceptual designs and economic estimates, scientific analyses and modeling, some focused laboratory experiments, and technology innovation. Concepts began by emulating Soviet technical approaches and evolved into US-developed methods and technologies. Great progress was made in understanding and modeling the process and development of designs and operations that held promise for improved efficiency, operations, cleanliness, and scale-up. A sizable workforce grew in numbers and experience with UCG research, development, and operations. The capstone field test was Rocky Mountain 1, conducted by a DOE- and GRI-led consortium of public and private organizations. In addition to being by far the largest US field test, this provided a second and convincing demonstration of the new controlled retractable injection point (CRIP) technology (described later) for efficient operation and growth of the process and the "Clean Cavern" approach for minimizing groundwater contamination (also described later).

By the late 1980s and early 1990s, American UCG activities tailed off to only a bit of Rocky Mountain 1 technical follow-up and report writing and some groundwater remediation activities. Oil and gas prices had stabilized, and the US government philosophy of energy technology development had shifted toward the private sector. Over the next two and three decades, nearly all the individuals who developed UCG expertise moved to other work, retired, and/or passed on.

4.3.3 A small revival in 2005–15

Growing UCG interest and activities around the world in the early 2000s, largely motivated by increasing and more variable gas prices, rekindled US activities from the mid-2000s to the mid-2010s. In addition to commercial project planning activities, some federal and state government agencies, universities, and nongovernmental organizations took an interest in and studied UCG to various degrees during this period. Wyoming supported a UCG review and detailed study of a notional large project and its costs (GasTech, 2007). The Indiana Geological Survey and Purdue University evaluated the suitability of Indiana coal resources for UCG (Shafirovich et al., 2009). The US Department of Interior, Office of Surface Mining, Restoration, and Enforcement organized regulators from several states and Native American nations to become better versed in UCG.

LLNL revived a modest UCG program at the beginning of this period and motivated largely by the goal of coupling carbon capture and sequestration (CCS) with UCG to affordably reduce the carbon footprint of energy produced from coal (c.f. Friedmann, 2005). This seemed especially relevant for locations with plentiful coal, scarce oil and gas, high energy prices and energy security issues, and anticipated time when economic incentives to minimize CO2 emissions would be in place. The program's UCG technical work emphasized mathematical models, evaluation of resources for UCG suitability, technical education and training, geophysical monitoring, and groundwater contamination.

Interest and activities in UCG and coal diminished in the latter years of this period with plentiful oil and gas production from fracturing operations and greater weighting of the challenge of groundwater cleanliness and the impacts from greenhouse gas emissions.

4.4 Recommended references

The following references are recommended for the “top shelf” of both the serious newcomer to UCG and the working professional.

An exceptionally complete chronicle of US UCG activities in the 1970s and 1980s is provided in the Proceedings of the Annual Underground Coal Gasification/Conversion Symposia that ran from 1976 to 1989 (Proceedings, 1976, 1977, 1978, 1979, 1980, 1981, 1982, 1983, 1984, 1985, 1986, 1987, 1988, 1989). LLNL's final program report (Thorsness and Britten, 1989b) contains a concise summary of 18 years of UCG work by Livermore that emphasizes what was learned. Britten and Thorsness (1989b) describe LLNL's conceptual understanding of UCG and its important phenomena. In the mid-1980s, UCG overviews are found in Dockter (1986), Stephens et al. (1985), Stephens et al. (1982), and Krantz and Gunn's “State-of-the Art” collection (1983a).

LERC and LETC's Hanna series of field tests were summarized by Bartke and Gunn (1983) and detailed by Bartke et al. (1985). LLL's Hoe Creek series of field tests and their results were described in detail by Stephens (1981) and more briefly by Thorsness and Creighton (1983). ARCO's Rocky Hill test is described by Bell et al. (1983), the Pricetown test is summarized by Schrider and Wasson (1981), and the tests in Texas are reviewed by Edgar (1983). Both of Gulf's Rawlins tests are covered in detail by Bartke (1985). LLL's Large Block tests at Centralia were summarized by Hill and Thorsness (1983), and their Centralia Partial Seam CRIP test was summarized by Cena et al. (1984). Cena and Thorsness (1981) give, for each DOE-funded UCG field test through 1979, a general description, well layout, simplified chronology, summary tables of flows, compositions, etc. for each major time period and time-history plots of interesting parameters.

The Rocky Mountain 1 test is described very briefly by Dennis (2006). Contemporary descriptions of many specific aspects of Rocky Mountain 1 appear in the Proceedings of the 14th Annual UCG Symposium (1988), including a chronology and description of results (Bloomstran et al., 1988). Reviews and analyses of results are found in Cena et al. (1988a,b) and Thorsness et al. (1988), cross sections of the cavities are found in Oliver et al. (1991), and the Clean Cavern operations and groundwater contamination results and overviews are found in Boysen et al. (1988, 1990) and Lindblom and Smith (1993).

Modern-era reviews include Camp (2017), which provides more detail than this chapter in several areas including narrations for most field tests, covers a few additional topics, and provides more references. Shafirovich et al. (2011) is a convenient reference for factual information from DOE-sponsored UCG field tests. Couch (2009) is an excellent wide-scope review of UCG that includes a generous portion devoted to US contributions. LLNL's recent program began with the critical review of UCG titled “Best Practices in Underground Coal Gasification” (Burton et al., 2006). GasTech (2007) reviewed UCG with emphasis on a modern commercial project. Groundwater contamination phenomena are reviewed in Camp and White (2015), geophysical monitoring in Mellors et al. (2016). Modeling references are in Section 4.6.

4.5 Field tests

4.5.1 Summary

Table 4.1 provides summary information for all of the UCG field tests conducted in the United States between 1948 and 1995.

Table 4.1

Summary of UCG field tests in the United States (parameter definitions appear at the end)

Early Hanna
Test nameHanna I
hydr.frac
Hanna I
main phase
Hanna II Phase 1AHanna II Phase 1BHanna II Phase 2Hanna II Phase 3
Dates3/73–5/735/73–3/745/75–7/757/75–8/7512/75–5/765/76–7/76
OperatorLERCLERCLERCLERCLERCLERC
LocationHanna, WYHanna, WYHanna, WYHanna, WYHanna, WYHanna, WY
BasinHannaHannaHannaHannaHannaHanna
Coal
Consumed (Mg)81843471650769 + RB43114641
RankHV Bit. C
Not swell/agglom
HV Bit. C
Not swell/agglom
HV Bit. C
Not swell/agglom
HV Bit. C
Not swell/agglom
HV Bit. C
Not swell/agglom
HV Bit. C
Not swell/agglom
Htg. value (kJ/kg)20,00020,00020,00020,00020,00020,000
Thickness (m)8.8Top ~ 5 of 8.89.19.19.19.1
Depth (m)11411484848484
Dip (degrees)777777
Process
InjectantAirAirAirAirAirAir
Link methodHydraulic fracturingReverse burnReverse burnReverse burnReverse burnReverse burn
Inj-prod spacing (m)9 and 181818181818
Design and operationsIgnition and injection into a central well. Production from multiple surrounding wells5 vertical inj/prod wells. RB links between many pairs of these. Forward burns between combinations of wellsSimple two vertical wells linked by reverse burn, followed by forward burnRB linked from the HII-1A cavity to a new third vert. well. Forward burn injecting into the new wellSimple two vertical wells linked by reverse burn, followed by forward burnTried to create a broad link between one link and a parallel burn cavity (failed). Did simple two-well forward burn
Results
Days of Ign. and R.B.112814114516
Days fwd (air/ox-st)6016837372745
Consumed fwd. (Mg)8183304162076936804258
Gas HV fwd (MJ/Nm3)4.25.05.55.76.85.5
Gas recov fwd (%)141037812999100
Highlights, accomplishments, observations, comments, problems, conclusionsHydraulic fracturing with sand proppant is not an adequate link. High gas leakage through open boreholes and casing failuresFirst UCG test of this era was successful. RB linking between many wells and burn cavities showed scale-up potentialDid simple two-well test that worked wellRepeated scale-up technique of linking from a mature cavity to a new well, and injecting into the new wellDid another similar test that worked pretty wellAttempted broad front advancement between links and cavities in both reverse and forward modes. Both failed
Later Hanna, Rocky Hill, and Pricetown
Test nameHanna IIIHanna IV-AHanna IV-BRocky HillPricetown I
Dates6/777/7712/77–6/784/79–9/799/78–11/786/79–10/79
OperatorLERCLERCLETCARCO (LETC heritage)METC
LocationHanna, WYHanna, WYHanna, WYReno Junction, WYPricetown, WV
BasinHannaHannaHannaPowder RiverPittsburg seam
Coal
Consumed (Mg)4771503620423270 + RB885
RankHV Bit. C
Not swell/agglom
HV Bit. C
Not swell/agglom
HV Bit. C
Not swell/agglom
Subbit. CHV Bit. A
Swelling and agglomerating
Htg. value (kJ/kg)20,00020,00020,00020,800High
Thickness (m)9.78.58.5Top 20 of 301.8
Depth (m)509898190270
Dip (degrees)777LowLow
Process
InjectantAirAirAirAirAir
Link methodReverse burnReverse burnReverse burnReverse burnReverse burn
Inj-prod spacing (m)1831232318
Design and operationsSimple two vertical wells linked by reverse burn, followed by forward burnMultiple vert. inj/prod wells attempted to link by RB. Forward burns attempted between various well combinationsMultiple vert. inj/prod wells attempted to link by RB. Forward burns attempted between various well combinations3 vert. inj/prod wells in a line linked by RB. Forward burn injected into an end well and produced from the middle well3 vert. inj/prod wells in a line linked by RB and cycled to open links. Fwd burn injected into an end well and produced from the middle well
Results
Days of Ign. and R.B.161078310106
Days fwd (air/ox-st)3858236012
Consumed fwd. (Mg)4734455013343270450
Gas HV fwd (MJ/Nm3)5.54.15.47.46.9
Gas recov fwd (%)928092114
Highlights, accomplishments, observations, comments, problems, conclusionsExtensive groundwater monitoring, but no results reportedSeemed like another similar test, but there were many problemsSeemed like another similar test, but there were many problemsReplicated Hanna-METC methods in a thicker deeper seam of different coal. Monitored overburden subsidence and hydrology effectsOnly US test attempted in a swelling agglomerating coal. RB links were eventually created, but had high resistance. Persistent plugging in rev. and fwd. modes
Hoe Creek and Rawlins
Test nameHoe Creek IHoe Creek IIHoe Creek IIIRawlins IRawlins II
Dates10/76–10/7610/77–12/778/79–10/7910/79–12/798/81–11/81
OperatorLLLLLLLLLGulf R&DGulf R&D
LocationGillette, WYGillette, WYGillette, WYRawlins, WYRawlins, WY
BasinPowder RiverPowder RiverPowder RiverHannaHanna
Coal
Consumed (Mg)1902658403615137770 + RB
RankSubbit.Subbit.Subbit.Subbit. BSubbit. B
Htg. value (kJ/kg)18,96018,96018,96023,55023,550
Thickness (m)7.6\(5)\3.47.6\(4.6)\3.47.6\(5.4)\3.011.411.4
Depth (m)~ 40 m to lower38 to lower54 to lower113155
Dip (degrees)LowLowLow6363
Process
InjectantAir/-Air/oxy-stmAir/oxy-stmAir/oxy-stm-/Oxy-stm
Link methodExplosive fracturingReverse burnBorehole + RB expansionBoreholeBorehole + RB links
Inj-prod spacing (m)101830.5 and 411660
Design and operationsHigh explosive fractured between two vertical process wells. Forward burn between themSimple two vertical wells linked by reverse burn, followed by forward burnHorizontal borehole link between three vertical wells. Expanded by RB. Fwd burn at 30 m space, then extended out to 41 mSteeply dipping bed. Directionally drilled injection and production wells and link. Injection point ~ 16 m downdip from production pointOne vertical production borehole between two injection wells w RB links between them. Injected into each well separately and into both together
Results
Days of Ign. and R.B.11437~ 30
Days fwd (air/ox-st)11/-56/27/4728/5-/65
Consumed fwd. (Mg)190/-2470/55334/36551225/228-/7767
Gas HV fwd (MJ/Nm3)4.0/-4.3/10.54.5/8.46.0/8.4-/12.8
Gas recov fwd (%)93788397
Highlights, accomplishments, observations, comments, problems, conclusionsFirst explosively fractured rubble bed trial in program.
Successful but suboptimal; hard to control pattern
First oxygen-steam UCG in program.
Inj. well broke near top causing heat loss to wet roof rock
First horizontal borehole link and borehole ELW in program.
UCG “burns” through a 5-m interburden to reach the next seam
First successful US test in steeply dipping bed. Used directionally drilled boreholesMany challenges. RB links between boreholes and cavities don't go where expected. Gasified lots of coal but not easy
Centralia and Rocky Mountain 1
Test nameCentralia LBK5,2,3,4Centralia LBK1Centralia PSC (CRIP)Rocky Mountain 1 ELW moduleRocky Mountain 1 CRIP module
Dates11/81–1/821/8210/83–11/8311/87–1/8811/87–2/88
OperatorLLLLLLLLNLGRI consortiumGRI consortium
LocationCentralia, WACentralia, WACentralia, WAHanna, WYHanna, WY
BasinTonoTonoTonoHannaHanna
Coal
Consumed (Mg)25 each (4)302400400010,200
RankSubbit.Subbit.Subbit.HV Bit. CHV Bit. C
Htg. value (kJ/kg)11,77011,77011,77020,00020,000
Thickness (m)Top 8 of 11Top 2 of 11Top 6 of 11Top 5 of 9Top 7 of 9
Depth (m)161652112108
Dip (degrees)15151477
Process
Injectant-/Oxy-stm-/Oxy-stm-/Oxy-stmAir/oxy-stmAir/oxy-stm
Link methodBoreholeBoreholeBoreholeBoreholeBorehole
Inj-prod spacing (m)1811 then 1822, 22, 159090
Design and operationsOne vertical production well. One horizontal injection well and borehole linkOne vertical production well. One horizontal injection well and borehole link. Linear CRIPParallel CRIP w vertical initial production well (1H Inj well and borehole link. 1H Prod well and borehole link. 1 V initial prod well)Horizontal production well and borehole link to two vertical inj. wells. Fwd burn fr distal well. ELW switch to second well failedParallel CRIP w vertical initial production well (1H Inj well and borehole link. 1H Prod well and borehole link. 1 V initial prodn well)
Results
Days of Ign. and R.B.11144
Days fwd (air/ox-st)-/3–6 ea-/4-/307/403/90
Consumed fwd. (Mg)-/~ 25 ea-/30-/24004000 total10,200 total
Gas HV fwd (MJ/Nm3)-/10.7-/10.8-/9.2-/10.3-/11.3
Gas Recov fwd (%)21–6185839189
Highlights, accomplishments, observations, comments, problems, conclusionsExcavation showed cavities filled with rubble, even in early all-coal stagesFirst CRIP maneuver tried in the field was successful.
First linear CRIP! Postburn excavation of young cavity
First CRIP system at full field-test scale.
First parallel CRIP field test.
Postburn excavation of full-sized cavity
Clean Cavern concept minimized groundwater contamination
Injection well completion at top of seam gave poor performance
Parallel CRIPdemonstrated successfully again. Three CRIP maneuvers created three fresh cavities. CRIP repeatedly rejuvenated burn
Clean Cavern minimized contamination
Other field tests
Test nameGorgas seriesFairfield or Big BrownTennessee ColonyAlcoaCarbon County
Dates1948–5919761978–7919801995
OperatorUS Bureau of MinesBasic ResourcesBasic ResourcesTexas A&M ConsortiumWilliams
LocationGorgas, ALFairfield, TXTennessee Colony, TXAlcoa, TXRawlins, WY
BasinHanna
Coal
Consumed (Mg)215 in first test4100Small
RankHV Bit. ALigniteLigniteLigniteSubbit.
Htg. value (kJ/kg)~ = Rawlins
Thickness (m)12.24.5~ = Rawlins
Depth (m)9> Rawlins
Dip (degrees)FlatSteep
Process
InjectantAir, maybe oxy-stmairAir/oxy-stmAir
Link methodR. Brn, hyd.frac?Reverse burn
Inj-prod spacing (m)9
Design and operations“Soviet methods.” Incl. mine addits, RB links, possibly hydraulic fracturing“Soviet”“Soviet”
Results
Days of Ign. & R.B.All the days
Days fwd (air/ox-st)26197 total0
Consumed fwd. (Mg)4100 total0
Gas HV fwd (MJ/Nm3)2+/5 first test4.73.0/8.61.3–5.6 when linking
Gas recov fwd (%)Low
Highlights, accomplishments, observations, comments, problems, conclusionsFirst US field tests!
High gas losses.
Roof collapse
26-day trialHeat loss to overburden, high water intrusion rates21 days of unsuccessful RB linking. Mech failure of well casingsUnsuccessful short operation.
Groundwater contamination from high operating pressures
Table definitions
ParameterParameter definitions
DatesDate range from first ignition to termination of oxidant injection
OperatorInstitution in charge of the test
LocationCity, state in United States
BasinGeologic basin
Coal
Consumed (Mg)Total coal consumed (Mg) for all phases of the entire field test (gas loss corrected and including char)
RankCoal rank
Htg. value (kJ/kg)Coal heating value, as received (kJ/kg). (Reports rarely specified whether they were giving the higher heating value or the lower heating value. More likely it was the higher.)
Thickness (m)True seam thickness. “Top x of y” used the upper x m of the y m-thick seam, leaving the bottom (y-x) m untouched. For two seams separated by a rock interburden, thicknesses are bottom\(interburden)\top
Depth (m)Vertical depth from the surface to the top of the seam at the main cavity or injection point (m)
Dip (degrees)Seam dip, degrees (horizontal is 0 degrees)
Process
InjectantInjection gas composition: air or mixtures of oxygen and steam (oxy-stm). Tests with separate periods of each are shown as: Air/oxy-stm
Link methodMain method of creating a permeable path(s) between injection and production well(s)
Inj-prod spacing (m)Distance between cased injection points and cased production points
Design and operationsDescription of design and operations, including process wells, linking, forward-burn injection and production wells, switching injection points
Results
Days of Ign. and R.B.Number of days for ignition, reverse burn linking, connecting, and link enhancement operations
Days fwd (air/ox-st)Number of days of main forward-burn operation (while injecting air/while injecting oxy-st)
Consumed fwd. (Mg)Coal consumed during the main forward-burn periods (gas loss corrected and including char) (air/oxy-st)
Gas HV fwd (MJ/Nm3)Average higher heating value of dry product gas during forward-burn phases (air/oxy-st)
Gas recov fwd (%)Estimated percent of gas that is recovered through the production well during forward burns (1 minus loss)
Highlights, accomplishments, observations, comments, problems, conclusionsHighlights, accomplishments, observations, comments, problems, conclusions

t0010_at0010_bt0010_ct0010_d

In the first several field tests, several US institutions developed competency in the Soviet-developed approach in which reverse burns are used to link vertical wells, with the main “forward-burn” phase of gasification taking place by injecting air into one of those wells or a sequence of them. Subbituminous and nonagglomerating bituminous coals were gasified successfully. An agglomerating swelling bituminous coal proved very problematic. Lignite beds were gasified, but their low heating value, thin seams, and high water influx produced poor gas.

Oxygen-steam injection was tried successfully and over time became the preferred injectant. The later tests in the series began to take advantage of the newly emerging and improving technology of directional drilling to create in-seam links between injection and production points. The Soviet scheme for gasifying steeply dipping beds was tried successfully, using directionally drilled boreholes for links. Further innovation led to the invention of the CRIP process. By the last field test, US practice and preferences had swung solidly toward using directionally drilled in-seam borehole links and CRIP.

McVey and Camp (2012) calculated the average dry product gas composition, weighted by mass of coal, for all the air-blown periods of US field tests and for all the oxygen-steam-blown periods of US tests. All field tests listed in Table 4.2 were included, from Hanna I to Rocky Mountain 1, with the exceptions of Rocky Hill, the Texas lignite tests, and the Centralia Small Block tests.

Throughout the series, the R&D nature of the program advanced the understanding of the UCG process and its interactions with the environment. This included a growing awareness of problems with groundwater contamination and improved understanding of how to minimize it, including the Clean Cavern concept of operation demonstrated successfully at Rocky Mountain 1. It became expected that deeper projects would be favored because of better isolation of contamination risks from the surface.

4.5.2 Gorgas (USBM)

Initial UCG activities in the United States were carried out by the US Bureau of Mines in 1947–60 near Gorgas, Alabama. The program included multiple field tests, used or explored combustion in mine tunnels, grouted boreholes, hydraulic fracturing and electrolinking to prepare the seam, and tried oxygen-enriched air. The small amount of data available conveniently from secondary sources appears in Table 4.1. There are very few references to these tests in the US literature of the 1970s, and it appears that the Gorgas program did not play a significant role in informing that generation of US UCG researchers.

4.5.3 Hanna series (LETC)

From 1973 to 1979, LERC and its successor LETC operated a series of UCG field tests south of the town of Hanna, Wyoming. The site was favorable, with a single, low-dip-angle seam of high-quality coal overlain by reasonably competent rock of low moisture content. Counting the various defined phases and subphases, there were at least nine identifiable tests that are summarized in Table 4.1. Each of the tests used air as the injectant and reverse burns to link between multiple vertical process wells laid out in various patterns.

The first UCG test of this generation, Hanna I, was a remarkable success. Through perseverance, the LERC team got UCG to work well. They accomplished multiple reverse-burn links, some between wells and some from forward-burn cavities to wells. They operated effectively in forward gasification mode for five and a half months and showed that forward burn was robust, tolerating changes in flows, pressures, and even changes in the injection location. They also showed the feasibility of scale-up by successfully extending the process laterally from the initial well pair out to additional process wells. Three important negative results were learned: neither sand-propped hydraulic fracturing nor pneumatic fracturing created sufficient connectivity between wells to sustain a forward burn; ungrouted process and instrument wells and uncemented boreholes result in large gas losses; and, for a given injection-production well pair, product quality tended to decrease with time. Given all the challenges involved in all the subsequent US field tests, getting the first test of this generation to work well was an impressive accomplishment. Technical failure or a severe safety incident could have derailed the entire US program. The success of Hanna I paved the way for the major US program that followed.

The subsequent Hanna tests were remarkably similar to each other in their main aspects. The well patterns and intended link paths seemed to be the main design/preparation variable, and the sequence of injection and production wells during the main forward gasification phases seemed to be the main operations variable. Many difficulties were encountered, and most test histories ended up being very different than their test plans.

The overriding objective appeared to be to test well and link patterns for their scale-up potential, looking for high production rates, high resource utilization, and extension of the process to large and larger areas. Hanna II Phase 2 and Hanna II Phase 3 were notable for having two adjacent two-well cavities merge into each other to essentially create a larger combined cavity, again showing lateral scale-up potential.

Hanna III had the objective of investigating groundwater contamination, but it was dry and didn't recharge as expected, so little useful data were collected. Hanna IV had a seemingly endless series of problems and revisions.

4.5.4 Rocky Hill (ARCO)

ARCO's Rocky Hill field test was conducted in 1979 in the very thick subbituminous Wyodak seam of Wyoming's Powder River Basin. The design and operations were similar to the simpler Hanna tests, with a vertical injection and production well linked by reverse burn. The gas quality was very good, probably because of a low ratio of heat loss to the roof per volume of coal consumed in the very thick seam. Evidence suggests that the cavity grew upward faster than sideward and that there was a significant degree of sagging and/or fractures extending some distance up into the overburden, although no surface subsidence had occurred 3 years later.

4.5.5 Pricetown (METC)

METC, with operations support from Monsanto's Mound facility, conducted in 1979 the Pricetown field test in northern West Virginia. They targeted a seam of swelling agglomerating bituminous coal that was 2-m thick and 270-m deep. Three vertical process wells were drilled and completed into the lower third of the seam. They were in a line, with 18 m between wells. With great effort, perseverance, and high injection pressures, reverse-burn links were eventually completed between the wells, using reverse-forward alternation with limited success to expand the links. Finally, a proper gasification phase was started, but it only lasted 12 days. A series of flow resistances were encountered underground and/or in production wells. Injection locations were moved around, with mixed success, followed by more flow resistance. The test was stopped after a rupture in the casing of a production well occurred, pressurizing the aquifer at a depth of 62 m with product gas.

The clear conclusion is that trying to do UCG in swelling agglomerating coals is problematic. UCG capabilities must improve considerably before trying to tackle these difficult coals again. Large-diameter drilled boreholes may work better than reverse burns for creating links.

4.5.6 Hoe Creek series (LLNL)

LLL conducted three UCG field tests from 1976 to 1979 at the Hoe Creek site in northeastern Wyoming. Initially intending to use the 30-m-thick Wyodak seam at a depth of 300 m, plans were changed to target a much shallower seam for the first experiment to reduce its drilling costs. As it turned out, all three tests were done there. The shallow depth allowed more tests with more downhole instrumentation for a given amount of R&D funding but contributed to unexpected and unacceptable environmental impacts. A horizontal 8-m subbituminous seam was targeted. Above it were 5 m of weak interburden and a second coal seam that was 3-m thick.

These tests were sponsored by the US DOE and its predecessors, with the GRI cosponsoring the third of these tests. Each test pioneered something new.

The Hoe Creek I was intended to be a first small pilot test of a large-scale explosively fractured packed bed reactor concept that LLL had developed. Two simultaneous 340-kg chemical explosive charges at the bottom of the lower seam, spaced about 7 m apart, were used to create a packed bed of coal rubble and sufficient permeability to gasify coal between two vertical process wells. It worked, producing gas of good but declining quality. Most of the permeability was in the top of the seam resulting in heat loss to the roof, poor resource recovery, and premature air bypass. In part because of the difficulty of creating a uniform permeability field, explosive fracturing was abandoned as a linking/preparation step for UCG.

The Hoe Creek II test retreated to the simple two-vertical-well, reverse-burn linked process, similar to LERC's previous Hanna II-1A and II-2 tests, but in different geology than Hanna. It also fielded intensive instrumentation to define the temperature field and cavity location. Linking went well, but forward burn suffered from a break in the injection well early in the test that caused poor gas quality because of roof heat loss and override. An important lesson was learned when they switched injection to a surviving smaller pipe that reached the bottom of the seam—the gas quality rose directly. Hoe Creek II also included the first period of oxygen-steam injection for the first time in this era and made a medium-quality product gas. Another important observation, also observed in nearly all field tests of the era, is that the cavity grew up faster than sideways, extended far up into the overburden, and was filled with rubble. Process instrumentation and postburn drill backs produced the cavity cross sections shown in Fig. 4.2.

f04-02-9780081003138
Fig. 4.2 Hoe Creek II postburn cross sections. Left: in the plane of the injection well (A) and production well (B). Middle and right: orthogonal cross sections near the injection well (middle) and production well (right). Scale in meters. Figure credit: Stephens (1981).

Experience at both Hanna and Hoe Creek II had shown that the paths of reverse-burn links could not be controlled; they were often multiple and often included a link along the top of the seam. In addition, Hanna had repeatedly been unable to link wells at spacings greater than 23 m, suggesting that scale-up of a reverse-burn linked process would require a large number of vertical process wells. Directional drilling technology was improving, and Livermore decided to use it to control the placement of a long link near the bottom of the Felix 2 seam for the Hoe Creek III test. This had the potential of scaling up with fewer wells and helped the process stay low in the seam for good thermal efficiency.

The Hoe Creek III test pioneered three things. It demonstrated extended operation with oxygen-steam injection to make a medium-quality gas that would be suitable for conversion to transportation fuels; it made use of the emerging technology of directional drilling to create a single link along the bottom of the seam with controlled location; and it pioneered and demonstrated the Extended Linked Well method (the method chosen for one of Rocky Mountain 1's modules 8 years later), in which new UCG cavities could be ignited and burned at the bottom of different vertical injection wells that intersect a directionally drilled horizontal borehole link. The oxygen-steam injection was successfully implemented. Plugging of the injection well, probably from molten slag covering the injection point, forced impromptu switching of injection to a different well, resulting in an increase in gas quality. These experiences were on the evolutionary path toward the invention of the CRIP process.

In general, the process efficiency and gas quality of all the tests at the Hoe Creek site were relatively poor compared with most other sites. In addition to the override problems that all sites wrestle with, the Hoe Creek coal seams that were gasified had a thick interburden layer and a weaker wetter overburden, as well as higher seam and overburden permeability for water influx, all leading to more heat loss per energy content of the consumed coal. Problems with subsidence and groundwater contamination were experienced, especially at Hoe Creek III; these are discussed in Section 4.8.

All the Hoe Creek tests, consistent with Hanna test observations, showed the benefits of having a low-seam injection placement and the rejuvenation of efficiency when the injection point is moved to a new low location. “Based on the experience of the Hoe Creek experiments two things were done. First, a new, more favorable site was sought [with low coal and overburden permeabilities] and a drier and more competent overburden. Secondly, a new method of performing the gasification process, the CRIP process, was conceived” (Thorsness and Britten, 1989b).

4.5.7 Rawlins series (Gulf)

Gulf Research and Development Company operated the two Rawlins tests under shared funding with the DOE. These were done in a 7-m-thick steeply dipping (63°) subbituminous seam west of Hanna.

Rawlins I was an air-blown test with a smaller oxygen-steam phase at the end. As with Hoe Creek III, Gulf used a directionally drilled borehole for the in-seam link. As shown in Fig. 4.3, the injection well, AIW, was directionally drilled to enter the seam from below (to protect it from heat and collapse events) and completed near the bottom of the seam. The production borehole and link, PGW, was directionally drilled along the bottom of the seam to below the injection well and completed and cased to a point about 15 m updip from the injection point. The process ran well and demonstrated the feasibility of gasifying a steeply dipping seam using a directionally borehole link and process wells. The reactor apparently grows via periodic dropping of large chunks of coal and roof rock into the base of the reactor. A rubble bed was established over the base of the injection well, serving as a fire pit. The cavity expanded significantly up into the roof rock that was vertically above the gasified coal cavity.

f04-03-9780081003138
Fig. 4.3 Cross section of Rawlins I test after gasification. The injection well, AIW-1, was cased to about the 6550' elevation. The production well, PGW-1, was cased to about the 6600' elevation and continued down as a borehole in the seam past the injection point. Figure credit: Bartke (1985).

Rawlins II was a larger and more complicated version of Rawlins I that mainly injected oxygen-steam. It used directional drilling to complete two injection wells roughly across dip from each other and a production well and borehole running downdip to the midpoint of the injection points. Reverse burns linked the injection wells to the production borehole. The sequence of events was very complicated and did not go according to plan. Ultimately, with considerable improvisation, various combinations of injection wells were used, and the gasification was successful. As at Rawlins I, with both spalling and collapse events grew the cavity upward through the coal and the overburden, feeding rubble beds at the bottom. Excellent quality gas was produced, largely because the steep dip keeps the upward-growing cavity within the coal seam, minimizing the ratio of roof heat loss to coal consumed.

The Rawlins I and II tests showed that once linked, a steeply dipping bed process can be run with a low injection point and a borehole and/or reverse link extending from it upward to a production well. Steeply dipping bed UCG can produce excellent quality product gas, because of the thermal efficiency of the “fire pit” packed bed nature of the process and the low ratio of roof heat loss to coal consumption. The burn cavity tends strongly to go more up than sideways. Although material balances showed little gas loss, it was observed at an updip surface outcrop of the seam and at the surface from the exterior surface of an imperfectly grouted well.

4.5.8 Centralia series (LLNL)

The Large Block tests (LBK) and the Partial Seam CRIP (PSC) test were conducted in 1981–83 by LLNL near Centralia in southwestern Washington. They were done in cooperation with the Washington Irrigation and Development Company (WIDCO), under the sponsorship of DOE and GRI. These tests are notable for the first field demonstrations of the Controlled Retracting Injection Point (CRIP) system (described in Section 4.8.9) and for being excavated afterward.

These tests were unique in that they were conducted at an exposed face or “high wall” of a subbituminous coal seam at a mine on the side of a hill. This allowed postburn excavation of the cavities. Fig. 4.4 is a cutaway sketch of the site, showing the configuration of the Partial Seam CRIP test. The Large Block tests were done first on an adjacent area of the same exposed coal face. (The name “Large Block” came from an earlier plan for experiments in large isolated blocks of coal from which the actual test evolved.)

f04-04-9780081003138
Fig. 4.4 Sketch of the WIDCO mine site at Centralia showing the configuration of the Partial Seam CRIP test. The terraces on the left are real. The right face of the figure is a cutaway cross section. Figure credit: Hill et al. (1984).

The Large Block tests were five similar small-scale oxygen-steam tests constructed with a vertical production well and a horizontal injection borehole drilled from the face and cased only part way in. Each of the tests lasted 3–5 days and consumed about 20 t of coal.

The first test in the field of LLNL's CRIP process was the fifth of the Large Block tests, LBK-1 (the test numbers did not reflect their sequence), proving CRIP's feasibility to relocate the injection point to a new location upstream in the injection well in unburned coal and ignite the burn there.

After the tests, the cavities were excavated and inspected (Fig. 4.5). The height-to-width ratio was typically 1.3–1.7 to 1. Most of the cavity volume was filled with rubble consisting of dried coal, char, ash, and some slag. Ash and slag are confined to the bottom. Toward the production well, the volume is entirely dried coal and char rubble extending upward from the original borehole. During one of the Large Block tests, an injection point had plugged during a period of high injected oxygen concentration. The excavation inspection revealed this to be plugged with mineral slag. This is another failure mechanism that CRIP can help with.

f04-05-9780081003138
Fig. 4.5 All five of the Centralia Large Block tests were excavated, inspected, described, and documented. This advanced an understanding of the rubble-filled nature of UCG cavities. Photo courtesy of LLNL photo.

The Partial Seam CRIP test was a full-scale oxygen-steam field test, of similar magnitude to most of the others in the past decade. Fig. 4.6 shows the well and borehole configurations. After 12 days of forward burn using the initial injection point, a successful CRIP maneuver was performed, burning a hole in the injection well liner and igniting the coal there and starting a new burn cavity. Forward burn continued from this new injection point for an additional 18 days. In addition to the usual small-scale spalling and fracture growth of the cavity, a major roof collapse even occurred after a total of 20 days of gasification, with the top of the collapse block at 5.5 m above the coal seam. This major addition of wet inert roof rock into the process reduced the gas quality for the rest of the test.

f04-06-9780081003138
Fig. 4.6 Plan (top) and elevation (bottom) views of the Partial Seam CRIP process and instrument wells. The burn was started at the end of the injection-well liner near I-6 and that was the injection point until the CRIP maneuver burned a hole in the liner at the CRIP point near I-7, igniting and injecting there. Initial production was out of vertical well PRD-2 and later through the slant production well, PRD-1. Figure credit: Hill et al. (1984).

Postburn coring was done to delineate and characterize the cavity. Later, the mine operators began to actively mine coal there, allowing excavation of the two (pre- and post-CRIP) burn cavities and the outflow channel at and above the in-seam production borehole. Fig. 4.7 shows the exposed coal face where the cavity was relatively large and a sketch of the cavity based on observations and characterization of thermally modified minerals. The sidewalls of the ~ 20-m-wide cavity were bowl-shaped (concave up) near the bottom, becoming largely vertical and extending far up into the overburden, higher than wide. Any remaining notion that UCG could be pictured as an open cavity and a rock ceiling was dispelled by the excavation of the Partial Seam CRIP test. The excavation found the “cavity” to be full of ash, slag, char, dried coal, and rubblized overburden, with a few gaps. The outflow channel had an upward V shape and was filled with char and dried coal rubble.

f04-07-9780081003138
Fig. 4.7 The Partial Seam CRIP field test, done at the exposed face of a coal seam, was excavated, inspected, sampled, and analyzed to learn the nature of the rubble-filled cavity; (c) is virgin coal, (d) is char, (a) is ash slag, and (b) is thermally affected overburden rubble. Figure credit: Kühnel et al. (1993).

4.5.9 Rocky Mountain 1 (DOE-industry consortium)

4.5.9.1 Overview

Rocky Mountain 1 (Fig. 4.8) was the largest, most successful, and final test of the US UCG program. It was organized by the DOE and cost-shared 50/50 with industry. Major participants included GRI, Stearns-Rogers Engineering (United Engineers), Gulf (later Energy International), LLNL, and WRI. The test was conducted several hundred meters southeast of the Hanna tests in the same low-dip nonswelling bituminous coal seam at a depth of about 110 m.

f04-08-9780081003138
Fig. 4.8 The Rocky Mountain 1 test took place near the old Hanna Wyoming site from November 1987 to February 1988. Photo courtesy of LLNL photo.

There were two adjacent but separate modules, CRIP (which used the Controlled Retracting Injection Point system - see Section 4.8.9) and ELW (which used the Extended Linked Well system). Both used horizontal boreholes for linking. The well configuration of the two modules is shown in Fig. 4.9. ELW was thought to have the best economic potential for relatively shallow seams in which many vertical wells could be afforded. The newly invented CRIP method was thought to have the best economic potential for relatively deep seams for which having fewer longer wells may be more cost-effective. The descriptions below emphasize the CRIP module because it was much more successful than the ELW module and because it appears to be more relevant to modern UCG practice.

f04-09-9780081003138
Fig. 4.9 Rocky Mountain 1's process well layout for its side-by-side ELW and CRIP modules. The artist's cutaway shows the planned cavity sequences; both modules stopped one cavity short of what is shown, and the cavities in both modules extended up higher into the overburden than shown. Figure credit: Thorsness and Britten (1989a).

The modules were ignited, started up, stabilized, and switched to the main phase of oxygen-steam gasification in late November 1987 and run continuously until being shut down in mid-January (ELW) and late February 1988 (CRIP). Perhaps to remind the US team of its Russian predecessors, the weather provided blizzards and windchills to − 36°C and challenges with frozen equipment (Fig. 4.10). The CRIP module ran very well and is described further below. The ELW module ran but was handicapped because its injection wells were completed at the top of the seam by mistake. Groundwater contamination received serious attention and the Clean Cavern approach was used to minimize it. This is discussed more in Section 4.7.1. A postburn drilling, coring, and logging program was done in the years after the test. This was used, along with downhole thermocouple data and material balance information obtained during the test, to delineate the cavity boundaries and extent of thermal alteration of surroundings and describe the cavity contents.

f04-10-9780081003138
Fig. 4.10 Wyoming's winter weather repeatedly challenged US UCG field-test operations, including this final Rocky Mountain 1 test. Photo courtesy of LLNL photo.

4.5.9.2 CRIP module

Fig. 4.11 is a plan view of the CRIP module. The casing endpoint of the vertical production well (CPW-2) was near the bottom of the coal seam. The horizontal injection and production boreholes were in the bottom half of the seam. The injection well was cased to 4 m from CPW-2. The horizontal production well's casing started about 90 m downstream of CPW-2. Injection was always into the horizontal well CIW-1. The injection point was initially at the end of the well casing within about 4 m of CPW-2 but was periodically moved upstream about 18 m at a time by a CRIP maneuver. The burn was initiated using vertical well CPW-2 for production but then was soon switched to the CPW-1 production well that was cased to about 90 m from CPW-2 and open borehole the rest of the way to its intersection with CIW-1 and CPW-2.

f04-11-9780081003138
Fig. 4.11 Plan view of Rocky Mountain 1's CRIP module, including the horizontal production borehole (CPW-1), horizontal injection well (CIW-1), vertical initiation and initial production well (CPW-2), and final cavity boundary (solid and short-dashed perimeter). The initial injection point (end of the CIW-1 casing/liner) was 4 m in unknown direction from CPW-2. The production well (CPW-1) casing ends about 25 m west of this figure. Figure credit: Oliver et al. (1991).

The CRIP module ran very well and had no serious problems over 93 days of forward-burn gasification when it was shut down because of schedule and budget. It produced high-quality gas, with efficiency parameters comparable with surface gasifiers. Four successive cavities were operated, using three CRIP maneuvers to create new ones, with each new injection point located about 18 m upstream of the previous one. For each injection point and cavity, the process efficiency and product gas quality followed the usual decline with time but improved again with each CRIP maneuver (Fig. 4.12). There was a long-term decline in product quality and efficiency, possibly because of the growing ratios of exposed roof area (and corresponding heat loss) and cavity perimeter (and corresponding water influx) to rate of coal consumption. Table 4.3 shows summary performance data for the CRIP module and the ELW module. Table 4.4 shows the energy balance for the steam-oxygen periods of the first CRIP cavity and the entire ELW module.

f04-12-9780081003138
Fig. 4.12 Process efficiency parameter (product heating value divided by injected oxygen rate) as a function of process day for Rocky Mountain 1's CRIP module. Oxygen-steam injection started on day 12. Vertical lines show the times of CRIP maneuvers. Figure credit: Cena et al. (1988b).

Fig. 4.13 shows the movement of the injection points with time and an early estimate of the cavity progression (Cena et al., 1988b). The actual injection points are further west (left) than the intended CRIP points because of thermal attack recession. Figs. 4.14 and 4.15 show the lengthwise cross sections of the final cavity, D-D' along the line of injection points and E-E' along the production borehole. Figs. 4.164.18 show the three cross sections perpendicular to the injection and production boreholes, all looking west (Oliver et al., 1991). These cross sections show that the thickness of the part of the seam that was utilized was about 6 m, probably because that was the elevation of the imperfectly placed injection well and production borehole. Where adequate time for growth had been allowed, the cavity width was roughly 18 m or about 3 times the thickness of the gasified coal. The height of roof rock that fell into the cavity or was thermally altered was typically 1.5–2.0 times the height of the coal below it that was converted.

f04-13-9780081003138
Fig. 4.13 Early estimates of the cavity evolution with time, based on mass balance, downhole thermocouples, and knowledge of the injection point. Each figure represents one of the CRIP periods, with maneuvers in between. Each period shows contours near the period's beginning, middle, and end. The locations of the actual injection points at those three times are shown with open circles. X marks the intended CRIP injection locations at the beginning of each period. Figure credit: Cena et al. (1988b).
f04-14-9780081003138
Fig. 4.14 Rocky Mountain 1's CRIP module final cavity geometry looking north, cross section D-D'. Fig. 4.11 shows the corresponding plan view with a scale. Figure credit: Oliver et al. (1991).
f04-15-9780081003138
Fig. 4.15 Rocky Mountain 1's CRIP module final cavity geometry looking north, cross section E-E'. Fig. 4.11 shows the corresponding plan view with a scale. Figure credit: Oliver et al. (1991).
f04-16-9780081003138
Fig. 4.16 Rocky Mountain 1's CRIP module final cavity geometry looking west, cross section A'-A. Figure credit: Oliver et al. (1991).
f04-17-9780081003138
Fig. 4.17 Rocky Mountain 1's CRIP module final cavity geometry looking west, cross section B'-B. Figure credit: Oliver et al. (1991).
f04-18-9780081003138
Fig. 4.18 Rocky Mountain 1's CRIP module final cavity geometry looking west, cross section C'-C. Figure credit: Oliver et al. (1991).

4.5.9.3 ELW module

For the ELW module, the initial injection well was VI-1 with VI-2 being the planned second injection well. The ELW production well P-1 was cased to about 90 m from VI-1 and open borehole the rest of the way. Produced gas was intended to flow through the open production borehole, thought to be in the bottom half of the seam, and into the cased portion of production well P-1. The casing endpoints of the two vertical injection wells were in the top 2 m of the coal seam instead of the bottom as planned. This handicapped the ELW module from the beginning.

The ELW module ran fairly well initially, but it struggled with override and oxygen bypass that hurt its performance and duration. It was not able to successfully switch injection to the second injection well, VI-2. Unacceptably, high oxygen levels in the product gas began to be seen that could not be remedied, and the module was shut down for safety considerations after 46 days of forward burn. Process results were given in Tables 4.3 and 4.4.

The cavity width that was in the vicinity of the injection well was roughly 15 m or about 3 times the thickness of the gasified coal. The height of roof rock that fell into the cavity or was thermally altered was typically 2–3 times the height of the coal below it that was converted.

4.5.9.4 General observations

The Rocky Mountain 1 test demonstrated: the importance of a bottom-seam injection point; the feasibility of the CRIP process for establishing a series of at least four new injection points and cavities at field-test scale; and the production of a gas with quality and efficiency metrics that are comparable with surface gasification processes (Thorsness and Britten, 1989b).

The ratio of product heating value to injected oxygen, a common efficiency metric, was 50% higher for the CRIP module than the ELW module. The CRIP module's superior performance is directly traceable to differences in geometry. The CRIP method assured that the injection point was always low in the seam, and CRIP maneuvers were able to generate new cavities whenever the cavity grew big enough to involve much overburden. The ELW module suffered from a high injection point, making the burn primarily near the top of the seam, thus involving a proportionally large amount of overburden and correspondingly high heat loss.

The extent of upward roof heating and rubblization was less in the CRIP module than the ELW module, despite having a slightly larger thickness of consumed coal, slightly wider cavity, and much greater duration of operation and quantity of coal processed. This is probably because the CRIP process frequently moved the injection point, reducing the duration of time that any one roof area is exposed to the hottest conditions.

There was more hydrologic connectivity and fluid/pressure interaction between the two modules than had been anticipated based on preburn characterization. This is an important finding that relates to scale-up in a nonhorizontal bed.

Reverse-burn connections were problematic. Drilling technology of the time resulted in boreholes missing their intended intersections by a few meters. Reverse burns were used to make these short connections in both the ELW and CRIP modules. Speaking of the CRIP module but applicable to both modules, Cena et al. (1988a) said that “the linking phase of the test was of short duration but was by far the most taxing phase in terms of the physical plant and personnel. Also, high pressures used for air acceptance tests and linking phases may have produced local, unwanted increases in permeability… Startup and subsequent operation would have been much easier had mechanical connection of the wells existed… .”

In addition to the inconvenience, the high air pressures (4–7 bar over hydrostatic) used to reverse burn the short connections drove product gas outward more than 200 m, mainly in the southwest direction as shown in Fig. 4.19 (Beaver et al., 1988). Contaminants were found updip in excess of permit requirements, but the regulators allowed the test to continue because of prompt discovery, understanding of the cause, and assurances of no more high pressures (Dennis, 2006).

f04-19-9780081003138
Fig. 4.19 High injection gas pressures that occurred during the brief reverse-burn connection-making operations at Rocky Mountain 1 drove gas including pyrolysis and gasification products out and updip more than 200 m. Figure credit: Beaver et al. (1988).

The high pressures used during air acceptance testing that preceded reverse-burn operations caused some of the monitoring wellhead flanges to mechanically fail. The pressure pathway for this is from the injection well into and through the formation to the open/screened bottom of the monitoring well and up the monitoring well.

Drilling navigation has improved greatly since 1987. Now, intersections are likely to be very precise, requiring less effort, pressure, hassle, and time to connect them, hopefully by water-jet erosion.

Minimization of groundwater contamination using the Clean Cavern concept was a high priority for Rocky Mountain 1. This is described in Section 4.7.1.

4.5.10 Other field tests

These commercial field tests were not well reported and had little impact on the main US UCG program.

4.5.10.1 Texas lignite field tests (Basic Resources and the Texas A&M Consortium)

Basic Resources, a subsidiary of Texas Utilities, purchased the US rights of the Soviet UCG technology in 1975. They operated a 26-day air-blown test in 1976 near Fairfield, Texas. In 1978–79, Basic Resources operated a 197-day test near Tennessee Colony in Anderson County, Texas. Table 4.1 contains information about this test. The product has had high CO2 concentrations. Efficiency suffered from heat losses to the overburden and water influx from adjacent sand formations.

A corporate consortium led by Texas A&M University carried out three short and small tests in Texas lignite, all air-blown. Operations and results were poor. The first two lasted 1 and 2 days. Problems included sand control, excessive water production, heat loss due to the thin seam, casing problems, and air/gas bypass. The third test, in 1980, spent 21 days at the Alcoa site trying to achieve reverse-burn linkage at well spacing as small as 9 m. Problems included mechanical failure of well casings due to thermal expansion.

4.5.10.2 Carbon County (Williams Energy)

“In 1995, Williams Energy conducted a UCG pilot project in Carbon County near Rawlins, WY. This test was located adjacent to the Rawlins UCG trials… however, it was performed at deeper levels…. The test was unsuccessful and resulted in groundwater contamination due to poor well linkage and operation of the UCG reactor above hydrostatic pressures” (GasTech, 2007). Organic compounds including benzene increased in concentration after the burn in groundwater within the coal seam and in overlying and underlying sandstone units.

4.6 Modeling

4.6.1 Introduction

Much effort in the United States has been devoted to mathematical modeling of UCG. The best of these have been well informed by field-test observations of the cavity and its nature of growth and an accurate understanding and conceptual model of the UCG process. Two excellent high-fidelity, multiphysics, integrated models, CAVSIM and UCG-SIM3D, were developed that capture the most important phenomena of the full multidimensional dynamic UCG process in its main forward gasification mode. Some much-simplified lumped-parameter engineering models, EQSC, UCG-MEEE, and UCG-ZEEE, were developed that are useful for rough estimates, trade-off studies, or screening of resource suitability.

4.6.2 High-fidelity, multiphysics models of the UCG process and cavity growth

The model of the 1970s and 1980s that most accurately and completely captured the important aspects and phenomena of UCG was LLNL's CAVSIM model (Britten and Thorsness, 1988, 1989). As illustrated in Fig. 4.20 (top), CAVSIM modeled a single cavity in a horizontal seam. The cavity was constrained to be 2D axisymmetric (i.e., varying only in vertical and radial directions), with a fixed injection point on the symmetry axis low in the coal seam. It included the essential chemistry, heat transfer, gas transport, water permeation influx with both a saturated and unsaturated zone in the surroundings, upward and outward cavity growth by spalling of coal and overburden, accumulation of coal, char, ash, and roof rock rubble in the lower fraction of the cavity, pyrolysis, gasification, and combustion of coal at the wall and in the packed bed of rubble. It also had an added module to represent the heat transfer and chemistry in the link between the main cavity and the production well entrance. It accurately reproduced the approximate cavity shape, water influx, and product gas compositions for the Centralia Partial Seam CRIP test and the first two cavities of Rocky Mountain 1's CRIP module (Fig. 4.20, middle and bottom).

f04-20-9780081003138
Fig. 4.20 Top: schematic of CAVSIM model, showing phenomena occurring in zones treated by various submodels. Middle: predicted and measured shape of the Centralia Partial Seam CRIP field test. Bottom: predicted and measured production rates of H2 and CO for the first two cavities of Rocky Mountain 1's CRIP module. Figure credit: Britten and Thorsness (1989).

More recently, a different LLNL team developed a modern high-fidelity multiphysics integrated model of UCG called UCG-SIM3D (Nitao et al., 2011; Camp et al., 2013). It models essentially the same phenomena as CAVSIM but takes advantage of modern computational capabilities, algorithms, and software elements of other state-of-art codes. Important advances over CAVSIM include flexible three-dimensional (3D) geometry that allows for arbitrary spatial variations of geologic properties such as multiple coal seams of different compositions, dip, and varying permeabilities; flexibility to move one or more injection points and production points to locations that can change with time; a sophisticated algorithm that tracks 3D growth of the cavity and rubble boundaries and rubble composition; an improved 3D model of flow, reactions, and heat transfer within the rubble bed and in the open void region; and a 3D nonisothermal unsaturated water and gas flow model for both the near- and far-field surroundings. As with CAVSIM, sideward and upward growth of the cavity in coal and overburden rock is by spalling, with user-specified rate coefficients in a temperature-dependent model. The code structure would allow for interface with a geomechanics code that could predict cavity growth by structural roof collapse, but this was not implemented. After fitting some parameters, UCG-SIM3D accurately calculated for both the Hoe Creek III and Rocky Mountain 1-CRIP field tests: the 3D development of the cavities and their rubble contents; the 3D time histories of the temperature, pressure, and composition fields within the cavities and in their surroundings; and the product gas composition histories (Figs. 4.214.23). UCG-SIM3D development ended before being matured into an engineering tool for use by nonexperts.

f04-21-9780081003138
Fig. 4.21 UCG-SIM3D calculations of cavity and rubble geometry and product composition, compared with measurements for the first 15 days of the Hoe Creek III field test. Figure credit: Camp et al. (2013).
f04-22-9780081003138
Fig. 4.22 UCG-SIM3D calculations of cavity and rubble geometry for the Rocky Mountain 1 CRIP module. Top: plan view at 21 days, bottom: cross sections after 47 days. Figure credit: Camp et al. (2013).
f04-23-9780081003138
Fig. 4.23 UCG-SIM3D calculations of product gas history compared with measurements for the Rocky Mountain 1 CRIP module. Top: heating value rate, bottom: (H2 + CO) per injected O2. Figure credit: Camp et al. (2013).

4.6.3 Simplified engineering models

Simpler engineering models were developed that are easier to use by a competent UCG engineer and are useful for obtaining rough estimates, dependencies/sensitivities, and screening resources. These are typically lumped-parameter models with no spatial or temporal resolution.

LLNL developed EQSC (Upadhye, 1986) to calculate energy and material balances (by species) based on a simple multizone model of UCG, chemical equilibrium, and a set of required inputs. In addition to the coal analysis, inputs included the water influx and the fraction of this that enters the process before and after the water-gas shift equilibrium is set, the methane ratio (since methane is governed more by pyrolysis than equilibrium), the effective temperature for calculating water-gas-shift equilibrium (different than the process temperature), heat loss, and product exit temperature. EQSC was extended in recent years by LLNL to a spreadsheet-based model called UCG-MEEE (material, energy, and economics estimator) (Upadhye et al., 2013). Its core is the EQSC model for a single (but reproducible) UCG module. UCG-MEEE provides side calculations to help estimate some of the required input parameters. It envisions many UCG modules operating in parallel on industrial scale for a long project duration, and so, it requires and calculates industrially relevant engineering parameters. In addition to a full material (by species) and energy balance and large-scale flow and resource parameters, UCG-MEEE estimates a selling price for the product gas to achieve a desired rate of return. The economics are based on the GasTech (2007) study and standard scaling factors. UCG-MEEE's utility for determining parameter sensitivities and trade-offs was illustrated in Burton et al. (2012).

4.6.4 Models of narrower scope

Krantz and Gunn (1983b) and Dobbs and Krantz (1988) review US models of the reverse-burn mode of UCG. Models that describe narrower set(s) of phenomena have been developed or applied to UCG, such as detailed chemical reaction models, multiphase hydrologic reactive flow and transport models, or state-of-art geomechanical codes. These are not reviewed here.

4.7 Environmental aspects

4.7.1 Groundwater contamination

US field tests demonstrated repeatedly that the risk of groundwater contamination from UCG is real. The Hanna tests resulted in small amounts of contamination. Relatively minor remediation was needed at two of the four Hanna test areas. The Hoe Creek tests, especially Hoe Creek III, seriously contaminated the site, requiring an extensive expensive remediation. The reported gas escape, large upward extent of cavity growth, fracturing, subsurface subsidence, and overlying aquifer at Rocky Hill are all indirect indications of the spread of contaminants although ARCO's paper on this test did not say so. There was upward transport of product gas to the surface in the steeply dipping coal seam and in the periphery of a well at Rawlins. The rupture of a production well at Pricetown pressurized an overlying aquifer and required shutdown. The extent of research on UCG-produced contaminant species by Texas-based workers of the time (c.f. Humenick and Mattox, 1978, 1982) suggests that it was a concern in the Texas lignite tests. Rocky Mountain 1 lost about 10% of its gas overall, and early in the test, it spread product gas hundreds of yards updip, exceeding contamination limits, during their high-pressure reverse-burn connecting operations. Williams' short-lived Carbon County testing at greater depths in the steeply dipping G seam near Rawlins operated at pressures higher than surroundings and contaminated groundwater in the seam and overlying and underlying sandstone units with benzene and other organic compounds, requiring remediation. There were significant differences in observed groundwater contamination between tests, with only minor changes in groundwater over limited areas found and reported after some tests and serious contamination over broader areas following other tests, with imperfect correlation to operating practices and estimated gas losses.

Gas escape is the major mechanism for transporting contaminants away from the immediate process area. Material balances, tracer tests, and other observations (c.f. Cena and Thorsness, 1981; Cena et al., 1988a; Bell et al., 1983; Davis, 2011) indicate that losing 10%–20% of the produced gas was common throughout the entire field-test program all the way through Rocky Mountain 1.

Factors that contribute to groundwater contamination were generally known at the beginning and early phases of the main US field-test program, but they appear to have been given lower priority over technical success, process efficiency, and project costs. Ironically, the early environmental failures motivated significant efforts that resulted in improved approaches for operating more cleanly. As the reality and importance of minimizing groundwater contamination impacts became more apparent, more attention was devoted to this, resulting in a greater understanding of the processes and the development of mitigating practices. WRI was perhaps the largest contributor in this area; additional contributors included LLNL, GRI, and researchers from Texas.

By the time of Rocky Mountain 1, this understanding and a growing body of research led to a set of recommendations called the Clean Cavern concept. Mainly advanced by WRI researchers, these made sense, were adopted by Rocky Mountain 1's management, and facilitated successful permitting (Covell et al., 1988; Boysen et al., 1990). Clean Cavern recommendations included maintaining cavity pressures below the hydrologic confining pressures, actively monitoring surrounding hydrologic pressures to inform the control of cavity pressures, postburn venting and steam flushing of cavities to evacuate pyrolysis vapors, cooling of the cavity to minimize further pyrolysis, and assuring subsequent inflow and production of groundwater from the cavity. Except for the pressure excursions of reverse-burn connections and their associated small contamination levels, Rocky Mountain 1 operations followed these rules and the resulting magnitudes and spatial extents of contamination were low.

Groundwater contamination was a major area of emphasis in LLNL's recent (2005–15) UCG program. Burton et al. (2006) summarize the Hoe Creek investigation and made general recommendations for cleaner practices. Camp and White (2015) describe the phenomena involved with groundwater contamination, contaminant transport scenarios and pathways, and practices to minimize the magnitude and spatial extent of contamination and its impacts. Fig. 4.24 summarizes some of the phenomena, pathways for transport, and opportunities for early detection of potential contaminant transport. Fig. 4.25 shows a semiquantitative estimate of the relative propagation rates of temperature, condensation, and dissolution that would be expected if gas escaped from a UCG cavity along a pathway of fixed diameter.

f04-24-9780081003138
Fig. 4.24 Possible UCG contaminant transport pathways and opportunities for using gas detection or sampling to discover them early. Figure credit: Camp and White (2015).
f04-25-9780081003138
Fig. 4.25 Qualitative profiles of temperature and condensable species concentration after UCG gas has leaked out along a permeable channel of fixed cross section for a short and a long period of time. For typical UCG situations, the thermal front will travel outward only 1/1000th or 1/100th as far as the gas front. The rock heat capacity will cool the gas, and the condensable species will be deposited onto cool rock after flowing past the thermal front. Figure credit: Camp and White (2015).

An understanding of groundwater contamination mechanisms and scenarios includes the following:

 Pyrolysis occurs in UCG. It produces many toxic organics. Unlike many surface gasification processes, a large fraction of these organics is not completely converted to simple gases. They remain in the gas product in the cavity, link, and production well.

 Inorganic contaminants can result from secondary processes driven by higher temperatures and/or geochemical changes such as pH change from increased CO2, NH3, or SOx concentrations.

 Gas escape from the process is the primary vector for transporting these contaminants away from the process, although the lower-volatility and higher-solubility contaminants will condense into their liquid or solid phase or dissolve into less-mobile groundwater along the way and go out much more slowly than the uncondensable and insoluble components of the gas.

 Transport of low-solubility contaminants by groundwater flow is slow and retarded by adsorption. Char, coal, and carbonaceous species in sedimentary strata tend to adsorb organic contaminants, which can greatly retard their transport by both gas and aqueous flow.

 Gas will tend to escape if the cavity gas pressure exceeds the pressure of its water-saturated surroundings. In a permeable dipping coal seam with impermeable roof rock, gas may escape updip unless inward permeation gradients overwhelm the dip. The magnitude of gas escape will be greater if the seam and surroundings have high permeabilities or high-permeability paths.

 UCG cavities can grow upward a long distance into the overburden, with fractures extending above those, and these will bring contaminant-laden UCG process gas to these higher elevations where the surrounding hydrostatic water pressure may be lower than the cavity pressure.

 Drawdown (cone of depression) from water flowing from the surroundings into the process cavity can reduce the water pressure of the surroundings, and, if the cavity pressure is not correspondingly reduced with time, the cavity pressure will exceed that of the changed surroundings.

 Gas can escape to shallower surroundings by flowing up uncemented boreholes, the outside of poorly grouted boreholes, natural faults, and permeable pathways such as dipping permeable strata or coarsely filled stream beds. Gas can escape to shallower (and therefore lower-pressure) surroundings by leaking at joints or failure points from wells that are open to the cavity (at its pressure) but closed or restricted at the surface, including the main production wells and instrument wells.

The following activities and choices will reduce the risk of unacceptable groundwater contamination. These are consistent with the Clean Cavern procedures but are broader in scope and stated more generally:

 A site should minimize exposure to sensitive environmental receptors. This means minimizing nearby, downgradient, and updip uses of groundwater and surface water; proximity to potentially usable aquifers; and proximity to residences, businesses, and recreational activities of people and of valued animal habitats.

 A site should be chosen that has barriers to transport of contaminants from the immediate and contaminated UCG process area to overlying or nearby sensitive environmental receptors. UCG should be deep and in horizontal or very low-dip strata. Thick and wide-extending continuous strata of low permeability should exist between the sensitive receptors and the highest possible zone of fracturing or subsidence-induced permeability above the UCG cavity. There should be no faults that could provide a transport path through the low-permeability barrier strata.

 Operations should not create or increase transport pathways, such as with uncemented or poorly cemented boreholes, hydraulic, or pneumatic fracturing that extends up or out of the immediate future cavity or leaky process or instrumentation wells. The design and construction of production wells and their external grouting must be of the highest quality to withstand severe temperature changes and gradients, erosive and corrosive particulate-laden gas flows, and mechanical stresses from ground movement. These wells are the barrier between high-pressure, contaminant-laden product gas and shallower, low-pressure aquifers.

 Operations must assure that the cavity pressure (and therefore the pressure in all gas-connected volumes) is lower than the pore water pressure in the surroundings of those gas-connected volumes that include the highest connected fractures above the cavity. The hydrologic pressure field must be actively monitored (by a combination of measurements and modeling) to assure that potential gradients are always and everywhere inward toward the cavity. The operator must be vigilant in accounting for upward growth of the cavity and its gas-connected fractures and drawdown of surrounding aquifer pressures.

 Methods should be deployed to detect gas escape from the cavity. Early detection of gas escape allows process adjustments to be made quickly to and reduce the distance and extent that contaminants are transported.

 The quantity of contaminants left underground in and near the cavity and product exit pathway should be minimized. It will help to continue the gasification until the downstream link areas, in which tars accumulate early in the process, have been consumed. The Clean Cavern shutdown protocol or an evolved version of it should be followed.

 Groundwater and pore gas in the immediate process area and the surroundings, especially in and near aquifers and sensitive receptors, must be characterized before operations begin, during, and afterward for several years.

4.7.2 Subsidence and changes to the overburden permeability field

Roof collapse and subsidence were analyzed in the main era of UCG research and in LLNL's recent (2005–15) UCG program. As with removing coal from the underground by mining, UCG can cause overburden to collapse, rubblize, fracture, and strain. One unwanted outcome of this is surface subsidence. But long before there is much surface subsidence, collapse, movement, fractures, and strains underground are likely to change the permeability field. In large-scale UCG operations, it is more likely that higher permeabilities would be created far above the seam.

This can make groundwater contamination worse. To isolate the UCG operation environmentally from shallower sensitive contaminant receptors, the zone of enhanced permeability must stay below the impermeable barriers that are counted on to protect shallower aquifers.

Modern geotechnical modeling and engineering must be used to help assure that UCG operations don't affect the subsurface above them in unacceptable ways. Model calibration/validation must be consistent with the tall cavities observed in US UCG field tests. A large degree of conservatism is needed for UCG because of thermally accelerated drying/fracturing/spalling of the roof rock and lesser control over and knowledge of cavity geometry.

4.7.3 UCG and greenhouse gases

LLNL's recent UCG program in the 2000s was motivated in large part by a perceived opportunity to reduce emissions of carbon dioxide (Section 4.3.3). Effort was devoted effort to understanding UCG's advantages and disadvantages with respect to greenhouse gas emissions.

Gasification is technically amenable to efficient separation and capture of carbon dioxide that could then be sequestered. This has been advanced for surface gasifiers in recent developments and demonstrations. UCG could work similarly, although the methane in UCG's product gas cannot easily be water-gas shifted, limiting the extent of carbon capture that can be achieved. The utility of carbon capture and sequestration (CCS) at a scale that is large enough to matter assumes that sequestration technology matures and becomes accepted. CCS consumes significant additional energy and adds significant cost. If carbon “utilization” is substituted for sequestration, the specifics must be analyzed to assure that the carbon is kept out of the atmosphere for the long term.

In the early 2000s, the notion was advanced of using UCG cavities for sequestering captured carbon dioxide from UCG-produced gas. Considered analysis shows this to be a poor idea for several reasons, and it should be dropped for the foreseeable future.

Methane comprises a significant fraction of UCG product gas. Methane is a much more potent greenhouse gas than carbon dioxide. Potential leaks of methane-containing product gas from UCG operations need to be included in analyses.

As with other coal energy technologies, UCG will produce more greenhouse gases per unit of useful energy than many other energy sources. If UCG tilts the energy mix toward more coal and less other sources, then this creates more greenhouse gases that will enter the atmosphere or need to be captured and sequestered.

But if the same amount of coal will be used anyway because of local circumstances, UCG, like surface gasification, may have carbon capture advantages over combustion. Coupled with CCS, it could approach the carbon footprint of natural gas (without CCS). Doing so economically at large scale would require incentives for reducing greenhouse gas emissions and development, maturation, and acceptance of both UCG and CCS technologies.

4.8 Process technology, characteristics, and performance

4.8.1 Ignition

Several methods of igniting the coal downhole were used successfully. Sometimes ignition went easily on the first try. Other times, it took many tries over many days with many modifications. No test was canceled because the coal could not be ignited, but sometimes, failure to ignite when and where desired required changing the operation plans.

Two general types of methods were used to ignite exposed coal in a borehole, generally after pumping out free water. In the first type, which only works at the bottom of vertical wells, crushed coal and/or charcoal was placed (dropped) to cover an electric igniter, and air (sometimes enriched with oxygen and/or methane or propane, staying outside of explosion limits) is fed to the location. In the second type, a fluid that autoignites in air (e.g., tetraethyl borane or silane in argon) is fed to the ignition point where it contacts injected air (sometimes with oxygen and/or methane/propane enrichment) to get the initial flame, followed by flow of easy-burning hydrocarbon fuel (ranging from methane to liquid diesel fuel) and more air (or oxygen-enriched air).

Ignition proved challenging as late in the program as the final Rocky Mountain 1 field test, which took several tries and modifications in both modules to succeed. This is detailed by Thorsness et al. (1988). They used a silane-argon mixture, air, methane, a special ignitor tool, and special nozzles. Laboratory testing before and after the Rocky Mountain 1 field test found that qualitative differences in ignition behavior and the ease of successful ignition depended on the relative directions of flow of the fluids in the borehole or well and nozzle jets, and the orientation of the borehole/well (vertical or horizontal) depended on the relative directions of flow of the gases in the well and nozzle jets, and the orientation of the borehole or well (vertical or horizontal) and whether the borehole was vertical or horizontal.

4.8.2 Forward gasification requires a link, not coal permeability

Forward-burn gasification requires an open or highly permeable link or pathway from the burn area to the production well. Repeated tries, during several of the Hanna phases, to get a forward burn to proceed out from a well-ignited injection well into either a virgin coal seam or a hydraulically or pneumatically fracked coal seam were never successful.

UCG forward burn does not operate in unlinked or unrubblized coal according to some of the early conceptual and mathematical 1D “permeation” models of UCG. These erroneous conceptual models contributed to the incorrect general assertion that high permeability was desirable for a UCG coal seam. Except for reverse-burn linking, high coal permeability is generally bad for UCG because of water influx, gas escape, and groundwater contamination.

Sufficient links were shown to include the following and combinations thereof: an open borehole; a char-filled reverse-burn channel; an initial borehole or reverse-burn channel that has become full of char and dried coal rubble and surrounded by fissured dried coal; an open burn cavity volume; a burn cavity volume that is filled with rubble of ash, char, dried coal, and rock pieces; and an explosively fractured bed of rubblized coal.

4.8.3 Reverse-burn links

Most US field tests through 1978 and some in 1979 used reverse burns to link process wells and boreholes. Details of field-test experiences and technical practices are documented in most of the field-test reports. An excellent simple sketch (Fig. 4.26) and description of reverse-burn linking appeared in Bell et al. (1983). “After igniting the coal at the base of well P, air injection is then introduced to well I that causes reverse combustion links (RCLs) to propagate from well P toward well I. These link channels are not open conduits, but are very permeable regions of char, approximately 1 m in diameter, which form along the paths of greatest oxygen supply. More than one RCL may form, and they may propagate at different rates. Eventually, one of the links breaks through to injection well I. In practice, the idealized case does not usually occur. RCLs may follow irregular paths from one well to another. The flow path may rise to the top of the seam.”

f04-26-9780081003138
Fig. 4.26 Sketch of ignition and reverse-burn linking, showing two wells completed into the lower zone of a coal seam. I and P are the injection and production wells for the reverse burn (and often but not always for the subsequent forward burn). (A) Ignition—start burn, (B) reverse combustion, and (C) link established. Figure credit: Bell et al. (1983).

Even with the drilled borehole links used in most tests from 1979 on, reverse burns were often used to make short connections between wells/boreholes that missed each other by a few meters.

The US field-test experience with reverse-burn linking was mixed. Sometimes, it worked smoothly, and a link could be accomplished in days. Sometimes, it did not go smoothly and required a great deal of fussing and trials, even in the hands of experienced operators. Hanna IV and Rawlins II had the worst experiences. Even with all the experienced personnel present at Rocky Mountain 1, using reverse burn to make short connections between boreholes “was by far the most taxing phase in terms of the physical plant and personnel. … Startup and subsequent operation would have been much easier had mechanical connection of the wells existed” (Cena et al., 1988a).

Links were preceded by pumping/blowing water out of wells and followed by air acceptance testing between candidate well combinations. When connectivity was inadequate, pneumatic or hydraulic fracturing was tried (sometimes producing unwanted permeability pathways), and even then, closer-spaced wells had to be drilled. The linking process itself always used air injection pressures that exceeded surrounding hydrostatic pressure. Sometimes, product gas was observed to move out a long distance during this operation. Links up to 23 m in length were usually successful, although some required fracturing. Links 30 m and longer were never successful, even with fracturing. Reverse links were sometimes enhanced by alternating forward periods by switching injection and production back and forth.

Reverse burns were successfully drawn from a large active burn cavity through the coal seam to a linking injection well. When the source burn was broad and the injection air source was also broad and parallel to it across the seam, only one or a few narrow links will be made—a reverse burn will not advance as a broad front.

Reverse-burn linking accomplished, with great difficulty, in a swelling agglomerating bituminous coal seam in the Pricetown field test, but the link(s) permeability was very low and repeatedly plugged up.

4.8.4 Directionally-drilled links

LLL's Hoe Creek III field test in 1979 was the first of the US program to replace reverse-burn links with a directionally drilled horizontal borehole. (To improve the conductivity of the small-diameter borehole that was drilled in this first experiment, it was expanded by a reverse burn along its length, a practice not done in future tests using larger-diameter boreholes.) Gulf's Rawlins 1 test 2 months later also used a directionally drilled in-seam borehole for its link. With the exception of the failed Alcoa test, all US field tests from 1980 on used directionally drilled boreholes for links.

Drilled links facilitate keeping both the injection points and product gas pathway low in the seam. They provide tight spatial control that will be beneficial for scale-up to multiple modules. Directionally drilled boreholes allow for a variety of completion designs such as the location and details of casing, liners, tubes, devices, and instruments. This makes scalable and efficient process schemes such as CRIP and ELW possible, as making possible other schemes not yet invented.

4.8.5 Characteristics of the main forward-burn phase of gasification

Given the preparation of a link between the main injection well and a production well, either by reverse burn or a drilled borehole (or even by explosive rubblizing), forward-burn operations could always be established easily and reliably. In field tests, the forward-burn period was almost always easier than ignition and linking.

Forward-burn operations repeatedly proved to be very robust. It operated well over a wide turndown range in injection rate. Operations could be stopped for days or weeks and resume quite easily. UCG operations in forward-burn mode responded quickly, stably, smoothly, and predictably to changes in injection gas pressure and composition, such as switching back and forth between air and oxygen-steam mixtures. They even tolerated major events such as collapsing volumes of roof coal and/or rock with only modest and conceptually reasonable changes in behavior.

Given suitable links and a mature forward burn associated with one injection point, a new forward burn can be initiated and grown at a different injection point by stopping injection to the first location and starting it at the second location. Sometimes, this happened unintentionally, as when an injection well failed and the injection point moved from the bottom of the seam to a break near the top of the seam. Moving the injection point intentionally, either by plan or improvisation, was common and successful.

This ability to switch locations of injection point was found to be useful in at least two ways. If there is a mechanical problem with one injection well, another can be substituted to “rescue” an operation. Moving to a new injection point at the bottom of the seam in new coal will improve gasification efficiency and performance.

The natural tendency of a forward burn is to ride up and consume coal near the top of the seam or in the upper half of the seam. Not only is this bad for resource utilization, but also it hurts thermal efficiency and gas quality (see Section 4.8.7). The best way to keep the burn from overriding into the top of the seam is to assure that the injection point is at the bottom of the seam and that it is moved to a new location at the bottom in new coal when heat losses to the roof significantly erode efficiency, such as when it is predominantly consuming coal near the roof. In field tests where the performance was declining due to override or excessive heat loss to the roof, moving the injection point to a new place at the bottom of the seam in ungasified coal usually restored performance. This understanding motivated the invention of CRIP.

4.8.6 A conceptual model for the UCG process and cavity growth

Most reports on field tests contain sketches of the best estimate of the final cavity geometry (c.f. Singer et al., 2012). These are based on a combination of postburn drill backs, in situ temperature measurements, geophysical monitoring, and material balances.

The field-test observations and model calculations evolved into a conceptual model and scientific understanding of the UCG process. This is described well in the phenomenology sections of Stephens et al. (1982), Thorsness and Britten (1989b), and Britten and Thorsness (1989) and formed the basis for the CAVSIM model.

To a first approximation, cavities are symmetrical with respect to a vertical axis through the injection point. They are rubble-filled and have nearly vertical walls. More precisely, they grow about twice as fast in the direction of the product gas exit as in the backward direction and they are perhaps more elliptical than cylindrical in a vertical cross section through the middle of the cavity taken perpendicular to the injection-production line.

In cross section, perpendicular to the flow path, cavities are usually taller than wide, both in early stages when the cavity is still within the coal and late stages when the cavity extends into the overburden. Every US field test of significant duration and coal consumption showed time and again that the cavity extended far up into the overburden, often by many or tens of meters.

The “cavity” was consistently found to be filled with rubble, often with a small void volume near the top. This rubble consists of slag, ash, char, dried coal, and thermally altered overburden rubble. This was seen clearly and conclusively during postburn excavations at Centralia following the very short-duration Large Block tests and the full-duration Partial Seam CRIP test.

Cavity growth can be complicated but seems to always involve thermal spalling and sometimes involve structural collapse events and intermediate-scale fracturing and block falling. Spalling grows the cavity in the coal seam and in the overburden, both sideward and upward. Nonswelling coals and sedimentary rocks overlying all the US tests tended to spall upon heating and drying. Large collapse events were occasionally observed and inferred.

The rubble bed covers the injection point, dispersing injected oxidant and evolving gases through the bed and out to the side walls and small voids near the ceiling. Gas reactions with the coal, char, and pyrolysis gases in this bed define much of the activity in the system.

The spalling of coal and char pieces into a rubble bed is likely very good for UCG efficiency as it increases surface area. But spalling of roof rock is bad because it increases the rate at which roof rock becomes heated, causing heat loss and water influx. An ideal site would have coal that spalls easily and roof rock that does not.

Exit channels that began as horizontal boreholes were discovered to evolve into a steep-sided upward “V” shape containing thermally affected dried coal and char rubble filling the V. The permeability to gas flow was greater near the top of the V channel. This presumably applies also to exit channels that began as reverse-burn channels of permeable char. When coal above was dried and pyrolyzed, it apparently shrunk in volume, fractured, spalled, and fell, opening void and fractures above and outward. Downward growth of this exit channel was limited by thermal conduction.

4.8.7 Energy balance determines process efficiency and gas quality

For a given coal, the process efficiency and the product gas quality are largely determined by an energy balance. (c.f. Thorsness and Creighton, 1983). Four losses are usually considered: sensible heat of overburden (and interburden) rock that remains underground, sensible heat of the gas product, vaporization and sensible heat of water that flows into the process underground, and vaporization and sensible heat of the pore water that had been in overburden (and interburden) rock that was heated. Only the first of these is a true loss (stays underground), but it is impractical to recover useful energy from the other product streams—one wants the energy to be in the chemical heating value of the product gas.

In virtually every field test, for a given injection point, thermal efficiency and gas quality started high and tapered down as the burn progress. The decrease was associated with the gasification cavity reaching the roof. The ratio of coal consumption to the amount of roof rock that is heated and dried (and falls into the cavity by spalling) decreases. If the roof rock stayed in place, the slow conduction of heat through the rock would not progress far, and only a small volume of rock would become heated. But spalling keeps exposing new surface and enhances the overall rate at which rock is dried and heated (Camp et al., 1980).

The best way to minimize heat loss to overburden rock and its pore water is to choose a site whose overburden does not readily spall, is resistant to structural collapse, and has a low water content. But in all the US tests, the cavity extended far up into the overburden and that volume of rock ended up as hot rubble in the cavity. Stephens et al. (1985) put it this way, “Site selection plays a major role in gasification quality. Sites with relatively dry, strong overburden, and at least moderately thick coal produce favorable results. Sites with thin coal or containing wet, weak overburden produce less favorable results.”

For a given geology, these overburden losses can be minimized by moving the injection point into an adjacent volume of new coal (e.g., by doing a CRIP maneuver) as soon as the roof rock is exposed in the currently burning cavity. Many field tests, including Rocky Mountain 1's CRIP module, demonstrated that moving the injection point to a new location “rejuvenates” the process efficiency and product quality.

Gas (and produced water) sensible heat can be minimized by having the hot gas leaving the cavity flow through a long channel in coal that will be preheated now and consumed later. Long links made by in-seam boreholes will help accomplish this.

Experience has shown that for a given geology, reducing the cavity pressure far below the water pressure in the surroundings so that the gradients in potential are more strongly inward tends to increase water influx. However, increasing the cavity pressure so that the gradients in potential are outward does not reduce water influx by much and results in gas losses that are detrimental to both process efficiency and groundwater contamination. The best practice is to operate the cavity pressure just low enough to assure that the hydraulic gradients are everywhere inward. Therefore, for a given geology (and its permeability field) and a requirement of no gas loss, the only way to minimize permeation influx is to minimize the vertical extent of the gas-connected cavity. This could be done by moving the injection point to new coal before the overburden roof spalls or collapses much (e.g., by doing a CRIP maneuver).

The best way to minimize water influx by permeation is to choose a site with low-permeability underburden, coal, and overburden, with that priority order set by the relative magnitude of the inward gradient in potential.

These energy-balance considerations provided much of the motivation for LLNL to use in-seam drilling to put a long borehole link at the bottom of the seam (Hoe Creek III) and for inventing CRIP to assure bottom-seam injection and conveniently move the injection point into new coal (Centralia).

4.8.8 Prediction of gas composition

Product gas compositions from different UCG field tests vary quite widely, even from tests at the same site and often between different periods of the same test. Much effort was put into finding simple ways to predict the composition of the product gas from a UCG operation. Simple equilibrium and kinetic models were ineffective. Correlations were repeatedly looked for with little success. There is no simple “UCG assay” that can be done on a coal to predict a UCG product gas composition. This is because the composition of the product gas exiting from a UCG operation results from its entire detailed spatial and temporal history of chemical species and energy concentrations and fluxes. The best simple predictive methods, described in Section 4.6.3, involve several physically based but adjustable/fitted parameters. Unless enough process information is available to justify using such a model, the average gas compositions of Table 4.2 are reasonable first approximations.

4.8.9 CRIP—the Controlled Retracting Injection Point system

Arguably, the most outstanding, useful, and enduring product of the entire US UCG effort was the invention of the Controlled Retracting Injection Point (CRIP) system. CRIP provides positive control of the gasification process—a means to extend the process spatially into successive volumes of new coal while keeping the injection point low in the coal seam, roof involvement low, and efficiency high. The CRIP system consists of compatible well design and completions, CRIP-specific down-hole hardware, and operating technique.

The first publication describing CRIP was in the 7th UCG Symposium (Hill and Shannon, 1981). Its abstract reads as follows. “The underground coal gasification process, in practice, is subject to various problems that make it difficult to maintain and control an efficient long-term operation. One of the major problems is the need to move the injection point to new areas of unburned coal as the burn progresses. To achieve better control of the gasification process, we recommend the controlled retracting injection point or CRIP system. With this technique, the operator can choose the optimum time and distance to move the injection point and consequently the burn zone, to get the best possible performance from the gasification process.” The essence of CRIP was illustrated in Hill and Shannon (1981), shown here in Fig. 4.27.

f04-27-9780081003138
Fig. 4.27 Hill and Shannon's original sketch of CRIP. Their caption read: “Basic design of the controlled retracting injection point (CRIP) system. As the cavity burns toward the left, the injection point is moved to the left also, step by step, by cutting off or perforating the injection pipe, which can be done remotely from the surface. Thus the injectant gas is always being fed to a zone of the coal seam where unburned coal remains to be gasified.” Figure credit: Hill and Shannon (1981).

CRIP was conceived and developed by the LLL team after the Hanna and Hoe Creek field tests showed the importance of keeping the injection point low in the seam, the viability of a directionally drilled borehole to be a link, and the benefit of being able to inject at different points along such a borehole into new coal to create successive “young” cavities.

CRIP requires the construction of a long cased or lined injection well in a coal seam, and it requires a specialized downhole tool fed by a gas line that can be moved inwards and outwards. CRIP provides a way to ignite the coal at the end of the lined portion, inject at that point, start and continue a burn cavity at that point, and then later melt a hole in the liner at an upstream location, ignite the coal there, inject there, burn a cavity there, and continue repeating this process. The injection well would be drilled along the bottom of the seam to always assure the injection point would be low in the seam.

Fig. 4.27 shows exactly how CRIP was first demonstrated in the field in LLL's Centralia Large Block test number LBK-1. Comparing this figure to a cross section of LLL's Hoe Creek III test shows that Hoe Creek III was a stepping stone in the evolution of CRIP. The borehole and production well are the same, but Hoe Creek III used two vertical wells to reach the two injection points.

While Fig. 4.27 shows CRIP in what has now become known as the “linear CRIP” configuration, the CRIP system has never been wed to this geometry; from the beginning, CRIP was intended to be usable in a variety of configurations. “There are several possible geometries for the gas production well (as shown in Fig. 4.1.), or one can use another horizontal hole in the seam, parallel to the injection hole. A third possibility, with particular application to thick seams, is a horizontal (production) hole at the top of the coal seam, vertically above the horizontal injection hole (well)” (Hill and Shannon, 1981). Original sketches of these, with captions, are shown in Fig. 4.28. Hill and Shannon describe how these can be scaled up to long injection well lengths and arranged parallel to each other with the spacing set by trading off resource recovery against subsidence.

f04-28-9780081003138
Fig. 4.28 Hill and Shannon's original illustrations of two possible CRIP configurations. Top: what now would be called double parallel CRIP. Their caption read: “Plan view of stream method adaptation of the CRIP system, with both injection and production holes drilled at the bottom of the coal seam. This version would be advantageous for dipping seams where slag accumulation in the bottom of the cavity might plug a down-dip production well.” Bottom: vertically-stacked parallel CRIP. Their caption read: “Vertical cross-sectional view of a module of a commercial CRIP system, in which the injection hole is at the bottom of the seam, while the production hole is at the top of the seam.” Figure credit: Hill and Shannon (1981).

LLL's Partial Seam CRIP field test in Centralia, Washington, provided the first full-size field-test demonstration of CRIP. This was deployed in what is now known as “parallel CRIP” configuration, similar to Fig. 4.28 (top) but with only one production borehole, angled slightly differently. Rocky Mountain 1's CRIP module also used CRIP in the parallel configuration. While not needed in theory, the practice at that time for parallel CRIP used a vertical production well to get the process initiated, and the first cavity begun in the linear configuration.

Thorsness et al. (1988) describe in detail the tool and procedure used in the Rocky Mountain 1 field test to perform a CRIP maneuver. This uses a silane torch to melt the steel liner and ignite the coal at a new location. When not in use, it is retracted a distance back upstream in the injection well, out of harm's way. When a CRIP maneuver is needed, it is pushed out to the desired position, operated, and retracted again.

4.8.10 Steeply dipping beds and thick seams are efficient but risk gas loss

UCG in steeply dipping beds will tend to be thermally more efficient than in flat beds. Burns and cavity growth tend to go upward, largely because overlying spalled material falls from its location down into the burned cavity. In a steeply dipping bed, such a cavity will have a higher ratio of coal consumption to thermal loss to roof rock. The high heating values obtained in the steeply dipping Rawlins I-air and Rawlins II tests and the very thick-seam Rocky Hill test demonstrate this.

Steeply dipping beds make it more challenging, if not impossible to avoid updip gas loss and its associated transport of contaminants toward the surface where pollutant receptors are more sensitive. Product gas escaped up the seam to the surface at Rawlins, as evidenced by carbon monoxide readings. Even the 7° dip at Rocky Hill enabled a few days of high-pressure operation to push gas many hundreds of meters up dip. Vertically tall cavities, operated at pressures below the water pressure of the surroundings at the top, will have a large inward gradient at the bottom, resulting in high inward permeation of water there.

UCG in thick seams will tend to be thermally more efficient than in medium or thin seams because of the ratio of roof rock involvement to coal consumed. Water permeation influx at the bottom or gas escape at the top may be a problem because of the large vertical extent of the cavity. Gasification of thick seams may result in a greater vertical extent of roof collapse (goafing). This can open up gas connectivity far up into the fractures of a sagging overburden, making it more likely for product gas to escape up into shallow strata. Even though it consumed a modest 3300 Mg of coal, the thick-seam Rocky Hill field test produced a high vertical extent of roof collapse and fracturing.

4.8.11 Monitoring of UCG operations and cavity evolution

The philosophy of the government-funded UCG program recognized that for UCG to be effectively developed into an efficient large-scale industrial process, it had to be understood. Understanding requires detailed knowledge of conventional chemical process parameters such as flow rate, composition, pressure, and temperature. Understanding also requires knowledge of what the process is like underground—how the cavity grows and what is it like, where the fluids flow, how the temperature field evolves, etc. A great deal of effort was spent on this, and it was valuable in understanding the process. These monitoring and postburn characterization efforts helped to evolve a better conceptual model of how UCG proceeds and consequentially guided mathematical model development. Monitoring details are in the technical reports of the time, found mainly in the Annual UCG Symposia proceedings.

4.8.12 Technical maturity and scale-up

The US program demonstrated UCG's feasibility many times at the single- or few-cavity scale-up to about 10,000 Mg. But even after dozens of tests at this scale, UCG design, construction, and operation had not become routine. As late as Rocky Mountain 1, experienced operators were not able to have the project, and operations go smoothly. UCG was still young in its maturation.

Many sketches were produced that envisioned how UCG techniques or systems might be scaled up by adding multiple modules, but none of these were tested. Nor were the sketches subjected to a detailed geotechnical design analysis that would look quantitatively at geomechanical aspects such as roof collapse, goafing, room-and-pillar resource recovery, and subsidence (not only at the surface but also of the strata that provide environmental isolation) or hydrologic aspects such as groundwater depletion, multimodule interactions, and gas escape.

Technologies and approaches that would facilitate scale-up were conceived and tested. Directionally drilled boreholes and wells for linking and gas injection/production were tried successfully. CRIP was invented and has excellent potential for helping scale-up. Use of successive reverse burns to extend the process laterally by linking many closely spaced vertical wells was demonstrated successfully.

One of the biggest problems needing more development is the design and construction of well completions. There were many failures from heat, structural failure, leaks, poor grouting between the well and the formation, and plugging by slag accumulation. A number of factors in different tests caused the air or oxygen-steam to be injected high in the seam instead of at the bottom as intended. Mundane process engineering technologies, such as particulate erosion of pipes, were also still a challenge and needed maturation.

At the end of the US field-test program, UCG remained low on the curve of technology development toward routine operations at large scale.

4.9 Conclusions

4.9.1 Technical accomplishments

Successful field tests at scales up to 10,000 Mg of consumed coal were conducted on several different subbituminous and nonswelling bituminous coals. Successful tests were done with both air and with oxygen-steam mixtures as the injectant, producing gas with heating values of 4.3–7.4 and 8.4–12.5 MJ/Nm3, respectively. The tests in lignite appeared to struggle operationally and/or produce low-quality gas, but other than its heating value, there is nothing to suggest that UCG could not work well with lignite. One test in an agglomerating swelling bituminous coal went poorly, with struggles to establish and maintain a sufficiently permeable link; expectations are low for such coals.

Successful tests were run in flat, slightly dipping, and steeply dipping seams; in seams with thicknesses between 6 and 30 m; and at depths between 40 and 190 m. Steeply dipping and very thick seams show good efficiency and gas quality, but they are more prone to gas escape and contaminant transport. One cause of poor gas quality from the lignite field test was its thin (2 m) seam. Deeper seams provide better environmental isolation from valuable and sensitive near-surface receptors, but wells must be engineered to contain the higher-pressure gas.

Several methods of ignition were used successfully, although ignition was sometimes challenging. Directional drilling of in-seam boreholes, reverse burns, and even high-explosive fracturing produced adequate initial links to initiate and grow the main forward-burn process. Unfractured, pneumatically fractured, and hydraulically fractured coal seams did not provide sufficient communication to get a successful forward burn going.

Once started, forward gasification was found to be robust and easy to manage. It responds in a stable and predictable manner to changes in operating parameters and the underground environment. It tolerated high turndown ratios and even stopping of injection for days or weeks. As long as a link existed to a production well, a new injection point could readily be ignited and started.

The UCG “cavity” was found to be full of rubble, not void. The rubble consisted of ash (and sometimes slag for hot-burning oxygen-steam-fired tests), char, dry coal, and rock. These enter the cavity largely by small-scale spalling of dry and fractured coal and rock from the sidewalls and ceiling. The cavity is usually taller than wide and grows toward the production well about twice as fast as backward, with sideward being in between. Downstream links, even simple boreholes, evolve into an upward “V-shape” cross section of rubblized char and dried coal.

An improved conceptual model of UCG evolved, along with scientific understanding of its important phenomena. Mathematical models of UCG and its phenomena were developed that allow accurate calculation of cavity growth, temperature and composition fields, and product gas composition. Simpler models were also developed to assist with engineering estimates.

For a given coal, the process efficiency and product gas quality are largely determined by an energy balance, with two of the main losses going to the drying and heating of roof rock (that ends as hot rubble), and the vaporization of water flowing into the process by permeation. Strong, spalling-resistant, low-water-content roof rock, and low coal and rock permeabilities are desirable.

It was demonstrated repeatedly that the highest quality gas is produced when the injection point is low in the seam, the burn is low in the seam, and the injection point is surrounded by coal and char and their rubble, and there is little involvement of roof rock.

The CRIP system was invented to assure that the injection point is always low in the seam and can be easily moved into a new volume of coal. It was demonstrated successfully, in both linear and parallel configurations, in two of the last three field tests, showing great promise for efficiently and cost-effectively scaling up the process.

While most UCG tests ended up being successful, they did not always go smoothly as planned. Hardware issues and operational challenges, even as late as Rocky Mountain 1, were a frequent reminder that UCG remained low on the technological development curve toward mature industrial practice. A high degree of watchfulness and creative ingenuity were needed for most field tests.

Some field tests resulted in unacceptable groundwater contamination, and some of these proved very difficult and expensive to remediate. It became apparent that the hazard of groundwater contamination from UCG operations was real and serious. After groundwater contamination became a priority, research led to a better understanding of this problem and approaches to minimize it. Rocky Mountain 1's “Clean Cavern” approach resulted in only minor and local contamination that was reduced to de minimis levels after a period of pumping. It remains to be seen if subsequent UCG operations, especially ones at larger scale, can be operated with acceptably low environmental impacts. It will take great care and commitment to reduce this risk to an acceptable level.

4.9.2 Programmatic aspects

The advances during the 1970s and 1980s resulted from intense activity by a critical mass of participants that produced a fast pace of development and learning. Some of the keys to its technical success were the continuity of institutions working on it, government funding, sharing of results in public conferences and reports, and determination to understand UCG and continuously make innovative improvements.

While the many field tests formed the centerpiece of the program, they were not isolated activities. There was iteration between field-test observations, scientific understanding of phenomena, modeling, and lab experiments, with each informing and improving the other. Field tests were first and foremost experimental trials and innovation test beds; they were highly instrumented and monitored, and drill backs were common. Learning, understanding, and technical innovation were emphasized.

Program participation was well rounded. Government research institutions led much of the field test and modeling work. Large energy companies and small UCG-niche companies also had programs that typically included field tests, sometimes with government support and sometimes not. University researchers were involved with laboratory experiments and model development.

One of the outstanding aspects of the US program of the 1970s and 1980s was the extent to which results were communicated and documented. The Annual UCG Symposia were especially effective at fostering communication among researchers, and their written proceedings left a rich legacy.

4.9.3 Closing

Beginning with very little domestic knowledge or experience in UCG, the United States caught up and made its own important contributions. The technical feasibility of UCG was demonstrated convincingly in many configurations in several coal seams. UCG field tests were designed, constructed, started, operated, and shut down safely with no significant accidents. Researchers from multiple organizations working at different sites developed a breadth and depth of competence and understanding of UCG and used this expertise to experiment and innovate. Approaches and technologies were developed that allow UCG to be done better and more easily than ever before. The reader is encouraged to immerse themselves in the references below, as this summary can only scratch the surface in describing all that has been done.

Auspices and disclaimer statements

This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

This document was prepared as an account of work sponsored by an agency of the United States government. Neither the United States government nor Lawrence Livermore National Security, LLC, nor any of their employees makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States government or Lawrence Livermore National Security, LLC. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States government or Lawrence Livermore National Security, LLC, and shall not be used for advertising or product endorsement purposes.

Reference Sources

Most UCG literature is in the form of topical reports and papers in conference proceedings. The following suggestions may make it easier to find and acquire references.

Reports by government organizations, reports on government-funded work, and conference proceedings organized by government organizations that have been approved for public release may be obtained from the US Department of Commerce's National Technical Information Service (NTIS, https://ntrl.ntis.gov/NTRL/) and/or (for documents funded or authored by DOE or its institutions, especially more recent documents) DOE's Office of Scientific and Technical Information (OSTI, https://www.osti.gov/scitech/). These are the first places to look, as they should have whatever LLNL and NETL have.

If not available from NTIS or OSTI, LLNL's library (https://library-ext.llnl.gov/) may have a copy of LLL/LLNL reports that have been approved for public release. NETL's library (https://www.netl.doe.gov/library) may have a copy of reports by MERC/METC or METC-managed contracts and proceedings of conferences they organized, including some of the annual symposia.