United States Environmental Protection Agency Industrial Environmental Research Laboratory Research Triangle Park NC 27711 Research and Development EPA-600/S7-81-018 July 1981 Project Summary Second Survey of Dry SO2 Control Systems Mary E. Kel y and S. A. Shareef This repor: is an update of the first survey report on the status of dry flue gas desulfurization(FGD) processes in the United States published in February I960. This updated assessment of dry FGD systems is based on review of current and recently completed research, develop- ment, and commercial activities. Dry FGD systems covered include spray dryers with a fabric filter or an electro- static precip tator (ESP), dry injection of alkaline material into flue gas combustion of a pulverized coal/alkali in an ESP or a fabric filter, and combustion of coal/alkali fuel mixtures. Almost all new systems use a lime sorbent and include a fabric filter for particulate collection. Removal guarantees for SOz range between 62 and 85 percent, depending on coal sulfur content. Two full-scale indus- trial spray diying systems are current- ly in operation, with the first large utility system scheduled for start-up in the Spring of 1981. A number of pilot-scale demonstra- tion programs funded by vendors and/or utilities have been completed in the past year. The Environmental Protection Agency (EPA) is currently funding three demonstration pro- grams (two spray drying and one dry injection). The Agency is also funding development of two combustion processes for SO2 control: combus- tion of coal/limestone fuel pellets and combustion of a pulzerized coal/alkali fuel mixture in a low-NOx burner. Favored sorbents for continuing pilot test programs of dry injection include nahcolite, trona, and upgraded trona (90 percent NaHCOs). This Project Summary was develop- ed by EPA's Industrial Environmental Research Laboratory, Research Tri- angle Park. NC. to announce key findings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at back). Introduction This report is a semi-annual updateon dry flue gas desulfurization (FGD) processes in the United States for both utility and industrial applications. Forthe purposes of the report, dry FGD is defined as any process which involves contacting a sulfur-containing flue gas with an alkaline material and which results in a dry waste product for dis- posal. This includes (1) systems which use spray dryers for a contractor with subsequent baghouse or electrostatic precipitator (ESP) collection of waste products; (2) systems which involve dry injection of alkaline material intotheflue gas with subsequent baghouse or ESP collection; and (3) other varied dry systems which are primarily concepts that involve addition of alkaline material to a fuel prior to combustion. Also since the open loop, spray dryer contractor portion of the Rockwell process had been adapted for a "throwaway" system, it has been included here. The report discusses the commercial and developmental activities for each type of process (spray drying, dry irtjec- ------- tion, and -combustion of a coal/lime- stone fuel mixture); an assessment of the state-of-the-art for each type of process; a brief comparison of the advantages and disadvantages of dry and wet FGD systems; and, areas for further research. Recent Developments Spray drying continues to be the only commercially applied dry FGD process. Since the last survey was conducted in the Fall of 1979, eight new utility andtwo new industrial spray drying systems have been sold, bringing the total to ten utility and four industrial commercial systems (Table 1). All but one of the new systems use a lime sorbent and all include a fabric filter for particulate collection. (One industrial system uses a sodium carbonate solution in a spray dryer-based system designed to remove SOz and HCI from high temperature gases.) Removal guarantees for S02 range between 60 and 90 percent on the systemssoldtodate.Thefull-scale utility systems are designed for relatively low sulfur coals (0.5 to 1.3 percent sulfur), while higher sulfur coals (1.5 to 2.5 percent sulfur) are burned at two indus- trial systems. Systems designs are similar, except for variations between spray dryer designs, atomization and slaking (lime preparation) techniques, and the use of gas bypass or solids recycle. The major spray dryer system vendors offer a basic system design with recycle or gas bypass. Ongoing bench-scale studies aim at better understanding the mechanisms of the reaction occurring in the spray dryer. Sodium-based sorbents are favored for dry injection; nahcolite, trona, and upgraded trona appear to be the most viable choices. Sorbent availability and waste disposal, as well as high stoichio- metric requirements and somewhat limited S02 removal capability (about 80 percent maximum in tests to date), continue to restrict the commercial application of dry injection. Development of coal/limestone pellets and combustion of a coal/alkali fuel mixture in a low NO, burner are continuing, under EPA-funding, on development- and pilot-scales, respec- tively. The minimal equipment require- ments of these techniques make them particularly attractive, but large-scale demonstrations and full characteriza- tion of the effects on boiler design and operation, as well as particulate control, remain. Current Status of Dry FGD Processes Spray Drying Two commercial industrial spray drying systems are operational at this time. The Rockwell/Wheelabrator-Frye system at Celanese Fibers Company's Amcelle plant passed Maryland State S02 compliance tests in February 1980. The Mikropul system at Strathmore Paper has been in operation since July 1979. Although there were some initial problems with atomization, availability during the past 11 months of about 90 percent based on boiler demand has been reported. Both of the operational industrial systems burn medium sulfur eastern coals (1.5 to 3 percent sulfur) and are designed for 70 to 85 percent S02 removal with lime sorbents. Both designs include a spray dryer followed by a fabric filter. The Mikropul system uses four two-fluid nozzles for atomization and a pulse-jet fabric filter. A single rotary atomizer and a pulse-jet fabric filter are included in the Rockwell/ Wheelabrator-Frye system. The first large scale commercial utility system start-up is scheduled for the Rockwell/Wheelabrator-Frye system at Otter Tail Power Company's Coyote Station. The 110-MW Joy/Niro retrofit system at Northern States Power is scheduled to begin operation in the Fall of 1980, but will be operated initially as a demonstration unit. Testing at about half the maximum gas flow rate is scheduled to begin this fall. Arvexisting ESP will be used until the baghouse portion of the system is completed in early 1981. The utility systems guarantee S02 removals from 61 to 87 percent. Almost all of the utility systems use a lime sorbent and include a fabric filter for collection of fly ash and the dried product solids. Exceptions are the Coyote system which will use a sodium- based sorbent (initially commercial soda ash) and a Babcock & Wilcox system .at Laramie River that will use four ESPs rather than a fabric filter. Some designs include hot or warm gas bypass and/or recycle of spent solids mixture. The differences between system design include the use of nozzle or rotary atomizers and the atomizer configura- tion in the spray dryer. However, several vendors claim that there are more plug- gage or erosion problems with nozzles, especially for lime slurries. Nozzles, however, generally have lower capital and operating costs. Some vendors offer a "multiple atomizers per dryer" design. This allows the absorber to remain operational even when a particular atomizer has to be taken out of service. The use of multiple atomizers in non- FGD applications, however, is not very common. Other basic system design differences include: 1. use of an ESP instead of a fabric filter for particufate collection (most use fabric filters), 2. variations in size and shape of spray dryer (top, side, or bottom gas entry; single- or two-point discharge; horizontal dryer or cylindrical tower with conical bottom, concurrent, or counter- current flow), and 3. reagent preparation techniques for lime (less costly paste slakers with grit removal or ball mill I slakers that produce more finely ™ ground product). In addition to developing a capacity for supplying a commercial spray drying system, many firms, as well as EPA, are involved in large scale demonstrations and fundamental research on the finer points of the technology. There are six major demonstration programs that have been recently com- pleted or are underway (Table 2). These systems range in sjze from 8500 to 120,000 acfm. Many of the studies are investigating various portions of the spray-dryer-based process including: 1. fabric filter vs. ESP collection with regard to collection efficiency and effect on SOz removal, 2. atomization technique, 3. reagent preparation techniques, 4. reactivity of various sodium- and calcium-based sorbents, and 5. waste solids disposal. Key process parameters that are varied I to characterize and optimize the process ------- Table 1. Key Features of Commercial Spray Drying Systems Sold to Date System (Vendor) Coal/SOz Removal Guarantees System Description Status Otter Tail Power, Coyote Station, Unit 1. 410 MW. (Rockwell/ Wheelabrator- Frye) Basin Electric Power Coop, Antelope Valley Station, Unit 1. 440 MW. (Joy/Niro) Basin Electric Power Coop, Laramie River Station, Unit 3, 500 MW. (Babcock & WHcox) Northern States Power, Riverside Station, Units 6 & 7, 110 MW. Retrofit. (Joy/Niro) Tucson Electric, Springer- ville Station, Units 1 & 2 350 MW each. {Joy/Niro) United Power Association, Stanton Station, 65 MW. (Research-Cottrell) Plan River Power Author- ity, Rawhide Station, Unit 1, 250 MW. (Joy/Niro) Colorado Ute Association, Craig Station, Unit 3, 450 MW. (Babcock & WHcox) North Dakota lignite, 0.78% S average, 7050 Btu/lb, 7% ash, 70% SOi removal for all fuels. North Dakota lignite, 0.68% S average, 1.22% S maxi- mum. 62% SOs removal for average coal, 78% for maximum S coal. Wyoming subbituminous, 0.54% S average, 0.81% S maximum, 8140 Btu/lb, 8% ash. 82% SOz removal for average S coal, 90% for maximum S coal. 1% S Montana Coal, 3.0 to 3.5% S Illinois coal. SO2 removal varying between 70 and 90% during demon- stration tests. New Mexico coal, 0.69% S. 61% SOz removal. Low and intermediate sulfur Montana subbituminous coal. Western subbituminous- coal. 1.3% S 80% SOt J/. removal. 0.70% S, 8950 Btu/lb, 14% ash design coal; 0.40% S, 10250 Btu/lb, 8% ash per- formance coal. 87% SOa removal for design coal. Four parallel spray dryers with 3 centrifugal atomizers each, followed by fabric filter with Dacron bags. Will initially use commercial soda ash. Sorbent utilization guarantee of 80%. Five parallel spray dryers (one spare), single rotary atomizer per dryer, followed by fabric filter with Teflon- coated fiberglass bags. Lime sorbent with partial recycle of solids. Ball mill slaker. Four parallel reactors (one spare) with 12 fluid nozzles each. Each reactor followed by as ESP. Lime sorbent, no solids recycle. One spray dryer with rotary atomizer. Will initially be demonstrated at 300,000 acfm with ESP. Full flow with fabric filter. Ball mill and attrition slaker for lime sorbent. Spray dryer/fabric filter design. Lime sorbent. Rotary atomization. Spray dryer/fabric filter rotary atomizers, possibly multiple atomizers per dryer. Lime sorbent. Spray dryer/fabric filter design. Rotary atomizers Lime sorbent. Horizontal spray dryers with nozzle atomizers, followed by fabric filter. Solids recycle. Ball mill slaker for lime sorbent. Start-up scheduled for mid- 1981. Start-up scheduled for April 1982. Start-up scheduled for Spring 1982. Testing with existing ESP scheduled to start Fall 1980. Fabric filter on-line in early 1981. Unit 1 scheduled to start up in late 1984; Unit 2 in 1986. Start-up scheduled for 1981. Start-up scheduled for 1983. Initial operation in November 1982. Commercial operation in April 1983. Sunflower Electric Coop. 'olcombe Station, Unit 1, '31O MW. (Joy/Niro) Western subbituminous coal 80% SOi removal. Spray dryer/fabric filter. Rotary atomization. Lime sorbent. Start-up scheduled for 1983. ------- Table 1. (continued) System (Vendor) Coal/SOz Removal Guarantees System Description Status Industrial Celanese Fibers Com., Amcelle Plant, 65000 acfm. (Rockwell/Wheela- brator-Frye) Strathmore Paper, Woronco, MA, 40000 acfm. (Mikropuf) University of Minnesota, 2 units at 120,000 acfm each. (Kennecott- Development Co., Environ- mental Products Division) Calgon, KY, 57000 acfm. (Joy/Niro) 1.5 to 2.0% S eastern coals. SOZ removal. 70% for 1.0% S coal and 87% for 2% S coal. 2.3 to 3% S eastern coal. 75% SOz removal. Subbituminous coal, 0.6 to 0.7% S. 70% S02 removal. 6000-8000 ppm SOa 8000 ppm halides. 75% S02 removal, 90% HCI removal. Spray dryer with single rotary atomizer followed by fabric filter with felt/fiber- glass bags. Paste slaker for lime sorbent. No solids recycle. Spray dryer with four two- fluid nozzles, followed by fabric filter with specially finished acrylic bags. Spray dryer with single rotary atomizer followed by fabric filter with fiberglass bags. Lime sorbent. Spray dryer/fabric filter. Rotary atomizer. Soda ash sorbent. Removing SOi HCI from 1700°F gases. Solids recycle. Operational. Passed Maryland State compliance tests in February 1980. Has achieved guaranteed removal. Operational. Now achieving removal guarantee. Commercial operation in Fall 1981. Under construction. Table 2. Major Spray Drying Demonstration Activities Vendor Location Size Comments Babcock & Wilcox Buell Envjrotech/ Anhydro Combustion Engineering Combustion Engineering Ecolaire Systems, Inc. Research-Cottrell (Cottrell Environmental Sciences) Pacific Power & Light Jim Bridget Station Colorado Springs-Martin Drake Station Northern States Power Sherburne County Unit #1 Alabama Power-Gadsden Station (under construction} Nebraska Power - Gerald Gentlemen Station Public Service of Colorado Comanche Station 120,000 acfm Testing in progress. 8,500 acfm Also EPA-funded dry injection program at same location. 20,000 acfm Testing complete. 100,000 acfm Testing to start in September 1980 soon after construction is completed. 10,000 cfm Testing in progress. mobile pilot plant 10,000 acfm EPA-funded, test in progress ------- Include SOz inlet concentration, sorbent 'stoichiometry, flue gas temperature drop over the spray dryer, gas residence time in dryer, approach to saturation, atomizer disc speed (rotary atomizers), sorbent slurry or solution concentration, and source (spray dryer fallout or col- lected solids) and amount of solids recycle. An additional objective of the programs is todemonstratethe system's capability to achieve the desired SOz removal on a sustained basis. Objectives of smaller scale funda- mental research also being conducted include obtaining a better understand- ing of the reaction mechanisms and definition of the most important vari- ables affecting the rate and degree of completion, characterizing the effects of atomized droplet size, sorbent particle size, and fly ash alkalinity, and charac- terizing the chemical and physical stability properties of sodium- and calcium-based waste solids. These activities are geared toward increasing the applicability of spray drying to high sulfur coal-fired units. Dry Injection Dry injection is a very attractive alter- native for combined removal of SOz and 'ly ash with minimized equipment requirements. But the commercial development of thistechnology has been constrained due to the lack of the pre- ferred sorbent (nahcolite) and accept- able disposal practices for the sodium- based waste solids. Studies have shown that dry injection of nahcolite intothedirtyfluegasstream and collection of the solids in a fabric filter results in 60 to 80 percent SOz removal at moderate gas temperatures (300 to 350 °F) and inlet SOzConcentra- tions «2000 ppm). Nahcolite utilization is generally less than 80 percent but it is not available in the quantities required for large scale commercial operations. Trona is less reactive than nahcolite and thus higher stoichiometries are required to achieve the same SO2 re- moval. Lime and limestone achieve significant SOz removal only at much higher gas temperatures (<600°F). Despite these constraints, several dry injection development studies are being conducted (Table 3). Objectives of these programs include improvement of sorbent reactivity (utilization), particu- larly for the more available sorbents and characterization of the waste solids for >acceptable disposal techniques. The following process parameters were varied during the te method (continuous combination of thos ing cycle time, inje sorbent particle size tration, and sorb Sorbents being inve lite, trona, upgradec sodium bicarbonate bonate. Combustion of Fuel Mixtures Variations of comb include test work or coal/limestone pe stokers, which is b Battelle Laboratorie; 3.5:1 Ca:S (mole scheduled to begin ir a 60,000 Ib steam/I Motors' Indianapoli scale tests havesho retention of the avail the 3.5:1 Ca:S pellet Energy and Em/in Inc., (EERC) is alsov tion of a pulverize mixture in a low-NO, Btu/hr scale. Limes both been tested, wi ing higher sulfur r percent retention i ratio of 2 to 3). The been demonstrated cant SOa removal. F be carried out on the Btu/hr scales. Research and dev with both processes characterizing effec operation, and maim strating the degre< achievable, and (3 effects of the resulti ulate loading. State-of-the-Ar Dry FGD isanattr st: sorbent feeding batch, precoat, or a 3 methods), clean- ;tion temperature, S02 inlet concen- nt stoichiometry. itigated are nahco- trona, commercial , and sodium car- ?oal/Alkali ustion modification the combustion of lets in spreader 3ing conducted by i. A 1 4-day test of a ratio) pellet was November 1980on ir boiler at General j plant. Laboratory wn 50 to 75 percent ablefuel sulfur with inmental Research orkingoncombus- d coal/alkali fuel burner on a 70,000 one and trona have h limestone show- stention (50 to 70 t a stoichiometric technique has also to achieve signifi- jture test work will 1,10, and 50x106 elopment activities are focused on (1) s on boiler design. enance, (2)demon- of SOz removal determining the g increased partic- Assessment ctive alternative to conventional wet scrubbing because it produces a dry, easy-to-handle waste rather than a wet-sludge. There are also potential capital and operating costs savings resulting from reduced equipment requirements, lower energy and water requirements, and relative process simplicity. Advantages/Disadvantages of Dry FGD vs Conventional Wet Scrubbing The advantages and disadvantages of the three technologies discussed here are based on pilot plant data, limited reported operating experience, and conceptual studies. Technically, spray drying and wet FGD can be compared in four areas: reagent requirements, energy requirements, operation and maintenance require- ments, and waste disposal requirements. Dry systems require a higher stoichi- ometric ratio of sorbent on the basis of moles of sorbent required per mole of SOz removed. Stoichiometric ratios for dry systems, are based on moles of sorbent required per mole of SOz in the inlet flue gas. Thus, a reported stoichio- metric ratio (SR) of 1.22 for a dry system achieving 80 percent SOz removal would translate into a SR of 1.5 under the conventional definition for wet systems. Dry systems also require an increased SR to achieve a given SOz removal at increased inlet SOz concentrations. This factor may pose a technical limit to application of spray drying to high sulfur coals, since the amount of liquid (and therefore, sorbent) that can be sprayed into the gas is limited by the available flue gas temperature drop over the spray dryer. This temperature drop is in turn fixed by the inlet gas temperature, the margin of safety (in terms of degrees above the- adiabatic saturation temperature) that must be maintained, and the overall SOz removal efficiency required (which limits warm or hot gas bypass). The maximum attainable solution concen- tration or weight percent solids in the sorbent slurry also limits the amount of sorbent that can be added per unit time. The energy consumption of the dry system should be less than for the wet scrubbing because of lower pumping requirements (lower L/G) and reduction or elimination of the need for flue gas reheat. Several vendors claim that the spray- dryer-based systems will have lower maintenance requirements and more operational flexibility than .comparable wet systems. Spray dryer system designs do not include sludge handling equipment or large slurry recirculation equipment. There is no wet/dry interface in the spray dryer system other than that in the gas suspension, making the process operation more flexible with respect to variations in boiler load and inlet SOz concentration. Economics appear to be one of the major driving forces behind selection of spray drying over conventional wet systems for low sulfur coal applications. Basin Electric evaluated the cost of a dry ------- Table 3. Current Dry Injection Programs Vendor Location Size Comments EPA/Buell-Envirotech Colorado Springs - Martin Drake Station DOE/Grand Forks Energy GFETC Labs Technology Center DOE/Pittsburgh Energy Technology Center EPRI/KVB PETC Labs Public Service Company of Colorado - Cameo Station 4500 acfm Testing completed in May 1980, EPA funded. 200 acfm Testing complete. Report expected in Fall 1980. 500 Ib coal/hr Testing in progress. furnace 20MWe Testing in progress. system to be 15 to 25 percent less over a 35-year plant life than a wet system for Laramie River and Antelope Valley, respectively. The Tennessee Valley. Authority (TVA) found the cost of a lime- based spray drying system to be con- siderably less than for a wet system. The basis for this estimate was a new 500- MV plant burning a low sulfur western coal with a 70 percent S02 removal requirement. The higher reagent re- quirements and the reagent cost differ- ential between lime and limestone may result in a significant economic disad- vantage for spray drying for high sulfur applications. The same advantages with respect to energy, and operation and maintenance requirements should apply for a dry injection system. The equipment re- quirements for dry injection are lessthan for conventional wet scrubbing or spray drying. However, the dry injection system has a distinct disadvantage with regard to reagent utilization and reagent-related operating costs. Nahco- lite utilization has been relatively low in tests conducted to date (60 to SOpercent), leading to fairly high reagent require- ments to achieve high 862 removal. Also, sodium-based sorbents are much more reactive than lime or limestonebut are considerably more expensive. The present dry injection technique would then be limited to relatively low sulfur coals. Sodium-based wastes, being readily soluble in water, also entail high disposal costs relative to the stabilized lime and limestone-based wastes. The combustion of coal/alkali fuel mixtures to control SO2 has obvious economic potential because of minimal equipment requirements and the fact that significant S02 removal has been demonstrated with limestone. However, both processes (combustion of coal/ limestone pellets and firing a pulverized coal/limestone mixture in a low-NO* burner) are still in the early stages of development, and the effects on boiler design, operation, and maintenance have yet to be fully characterized. Also, these technologies are currently limited to specific boiler types; i.e., spreader stokers for the pellets and dry bottom pulverized coal boilers for thecoal/lime- stone fuel mixture. Fuel sulfur content may be limited by the sheer volume of reagent required for higher sulfur applications and the resulting effects on boiler and particulate control device operation. Areas for Further Research There are several areas requiring further research if dry FGDtechnology is to become a widely applicable alterna- tive to conventional SO2 control techniques. Concerning spray drying, research effort in the following areas could serve to increase process applicability for units firing high sulfur coal: 1. Improved understanding of the absorber and downstream SO2/ sorbent reaction mechanisms, 2. improved reagent preparation techniques, 3. improved understanding of fly ash alkalinity utilization and investiga- 4. tion of various sorbent recycle schemes, and development of a limestone spray drying process. In addition to research in areas (1) through (3) above, dry injection work may also need to focus on improving the reactivity of sorbents that are more readily available than nahcolite and do not pose the same waste disposal prob- lems as those with sodium compounds. Research in combustion of coal/alkali fuel mixtures will need to define the important process variables, suchasbed temperature, and their effect on SOz retention; and further evaluate the long- term effects of firingfuel/alkali mixtures on boiler operation. 4 U.S GOVERNMENT PRINTINO OFFICE. 1801-757.012/7166 ------- ------- United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 Postage and Fees Paid Environmental Protection Agency EPA 335 Official Business Penalty for Private Use $300 PS 0000329 ------- |