EPA TECHNOLOGY TRANSFER 62362-77 FABRIC FILTER PARTICULAR CONTROL ON COAL-FIRED UTILITY BOILERS: NUCLA CO. AND SUNBURY PA. U.S. ENVIRONMENTAL PROTECTION AGENCY '. ENVIRONMENTAL . RESEARCH INFORMATION CENTER ------- ------- EPA TECHNOLOGY TRANSFER FABRIC FILTER PARTICUATE CONTROL ON COAL-FIRED UTILITY BOILERS: NUOLA, CO. AND SUNBURY, PA. U.S. ENVIRONMENTAL PROTECTION AGENCY ENVIRONMENTAL . RESEARCH INFORMATION CENTER EPA-625/2-77-013 ------- Sunbury plant profile .:-,: ;,,7:_ _ -S :=-n:-!kLi;::'r!!?j«?~-r,E.T--ss«-r;;».j ------- Fabric filters are one of the oldest means of controlling aerosol particles. Since the 1950's, they have been used widely as air cleaning devices in asphalt plants, cement manufacturing, carbon black plants, glass furnaces, and a variety of ferrous and nonferrous foundry operations. To date their use in the electric utility industry has been limited, primarily because of reservations about the durability of the filter bags under harsh stack gas conditions. The use of new glass fabrics and the development of high- temperature lubricants to enhance bag life, however, has rekindled electric utility interest in fabric filters as a means of controlling particulate emissions. As shown in Table 1, four utilities currently use fabric filters, and eleven more are under construction. Table 1. Status of Fabric Filter Installations in the Electric Utility Industry Unit ^Pennsylvania Power and Light Co. =r~ Pennsylvania "Power and Light Co. 5- Colorado Ute - Electric Association ^Public Service Co. _of Colorado Southwestern Public ^ Service Co. g.r , " - ' .. ~-- Texas Utilities - Service Inc. I^Board of Public Utilities " Location . * fc_ - PP-- Sunbury, Pa. s Mr rHoItwood, Pa. . Nucla, Colo. Cameo, Colo. I?***' plAmarillo, Texas Mount Pleasant, jexas nsas City, Kansas of Colorado Springs f Colorado Springs, r' " " ' jpcoio. i, .: .: : -. ".. .- I m - '- Colorado Ute , |^Meatrose, Colo. £ Electric Assn. i^Crisp County Power Co. _ bituminous coal estern coal ybbituminous .ignite }1 Minnesota Power f^& Light Co. Nebraska Public Power Public Service of : Colorado ^ Texas Utilities Tbrdele, Ga. "tohasset, Minn. Jellevue, Nebr. palisade, Colo. Bpbertson City, xas Note: N/A = Information not available. Combination of etroleum coke hd anthracite Combination of etroleum coke anthracite, , . ,p^ Operational J FT since 1973 1 -,_;J 1 ^Operational if* <;inrf> Innp ' te ;since June, 1975 Operational ' 1974 N/A ^Startup: Mid-1977 Startup: jjune 1978 " Startup: March 1978 Startup: 1979 ^Startup: 1980 -j- Startup: 1977 Startup: 1975 Startup: 1978 Startup: 1977 -i.^ Startup: 1977 / Startup: 1980 ------- Electrostatic precipitators are by far the pre- ferred control devices for participates in the electric power industry. Their low maintenance requirements and high control efficiencies have resulted in their widespread acceptance. Recently, however, a num- ber of circumstances have enabled fabric filters to compete with electrostatic precipitators. The increase in demand for solid fuels has caused many utilities to depend upon variable coal supplies, which in turn results in a varying chemical composition of the fuels. The abundant low-sulfur fuel in the Western states may soon replace some dwindling supplies in the Eastern coal fields. With dwindling fuel supplies, increased attention is being paid to the use of mixed fuels; Pennsylvania Power and Light Company, for example, uses anthracite waste and petroleum coke at its Sunbury Station. Each of these circumstances can affect the performance of electrostatic precipi- tators. Variations in fuel composition, for instance, alter the ash surface chemistry which affects the electrical resistivity of the particulates. Electrostatic precipitator efficiency is a direct function of particle resistivity. Fabric filter efficiencies, on the other hand, are much less dependent upon ash surface chemistry. For high sulfur fuels, the condensation and oxidation of sulfur oxides on paniculate surfaces decreases particle resistivity and hence enhances precipitator efficiency. Thus, the trend toward lower sulfur fuels could adversely affect electrostatic pre- cipitator performance. Fabric filters, on the other hand, are well suited for low-sulfur fuels. In fact, lower sulfur content aids bag filter operation, since high concentrations of sulfur trioxide can accelerate fabric deterioration. Table 2 summarizes the per- formance characteristics of fabric filters and electro- static precipitators with respect to selected power plant operating conditions and fuel properties. Table 2. Performance Characteristics of Fabric Filters and Electrostatic Precipitators c 1 Power Plant Variable : Maximum Collection ; Efficiency i, Fuel S Concentrations , Ash Metallic Qxi.de ; Concentration Paniculate Loading Flow Rate Temperature Hot Side ESP "It Fabric Filter Cold Side 99.0 to 99.8 * F ' 99.0 to>99.T*W I I P- " -- " ~ s -S f- Minor ^Dependence r Very Important i - I. . r '- No Dependence ' Very Important o Dependence3 «. ' ^ %^ v f, , l «i* 3^1ov Dependence " Loading Swings Decrease Collec- _ . Ip'et Loading Does Not Signifi- f tion Efficiency Significantly L cantly Affect Efficiency. All i- i ' " i ^1 |2?riatioris fn Outfet tConcentra- ^ ^ In. - . ,.,-,,", ,.,.T ,| are^ampened Considerably J t- > .' i,r »-> J ^S**"1 -«r»»iiic^ ass (,*"jf.'vj - ^ r^s, Mt t Nonuniform Velocity Distribution _ ^Wide Velocity Range Does Not i Causes Decreased Collection , SiSn'^'cant'y Affect Performance ' ' ffiq 1 ..... UN ' I Non . ^ ».. .. » . onuniform Flue Gas Temperature ^.No Temperature Dependence for , f. Causes Decreased Collection m Collection Efficiencyb ^.Efficiency -s _J aHigh SOX concentration in gas streams causes high deterioration rates in many fabrics. Multiple temperature excursions through the acid dew point will increase fabric deterioration rates. ------- As previously mentioned, two coal-fired stations have been using fabric filters since 1973 the Nucla Plant of the Colorado Ute Electric Station and the Sunbury Station of the Pennsylvania Power and Light Company. Each of these operations is described in Table 3. The Nucla fabric filtration facility was designed to control the entire particulate load from the three j stoker-fired boilers of the 39 MW facility. This instal-' lation, shown on the cover, has been able to meet ! the Colorado air pollution regulations with ease. j Every boiler has its own baghouse which consists of six independent compartments, each containing 112 bags for a total of 672 bags per baghouse. The design gas flow through each baghouse is 86,240 acfm at 360°F. Bag cleaning is accomplished by a combination of reverse air and gentle shaking. The Table 3. Description of the Nucla and Sunbury Fabr air used to clean the bags is cleaned flue gas recycled from the baghouse discharge. Cleaning all six com- partments requires about 30 minutes, with the cleaning cycles initiated automatically by preset limits for baghouse pressure drop. The Sunbury powerplant operated for 2 years using the original bags with only slight problems and without appreciable bag failures. Tests per- formed on samples of the two year old fabric and new fabric indicated the used fabric to be nearly as strong as the new material. At Nucla, changes in the baghouse thimble plate improved bag wear. One baghouse containing over 600 fiberglass bags, oper- ated 6 months without a bag failure. c Filtration Facilities SiLocation ^Total MW Capacity fiCombustion Units =rInstallation Date jiCoal Type _.... , _...... ........ , ^Manufacturer H . , , -- tCJeaning Technique Fabric Type :Number of Baghouses ENumber of Bags Per ^Baghouse HBag Size LAir/cloth ratio (net) Pressure Drop at Full Load ucla, Colorado MW 15^ ^ mj ^ u ^ ^* ~ 1 r Three Spreader Stokers '-":- - -"1 )ecember 1973 Bituminous, f(0.6 to 1.8% S) ^A/heelabrator-Frye (Baghouse) |W,W. Crisweil Co. (Bags) Combination of shaking and '"'everse air flow |,Graphite/silicone coated fiberglass r3 ;2 ft. long x 8 in. diameter K, ,; . +* ^ * 4 5 in. of water Sunbury 1 -SJOsimokin Dam, Pennsylvania S^ ^v, _ T" j- * *, H^ * ;>s MW " J* iFpur pulverized fuel boilers Sfebruary 1973 ^Anthracite and Petroleum Coke (1.2 to 3.2% S) , Western Precipitation (Baghouse) f^Menardi-Southern Co. (Bags) Reverse air flow 1 Teflon coated fiberglass I j 30 ft. long x 12 in. diameter ^3 in. of water on 2 year old bags ------- The pressure drop across the Sunbury baghouse at full load was a nearly constant 2.5 in.h^O. During the two years of operation, baghouse increase in pressure drop was less than 1.0 in.H2O. The average pressure drop across the Nucla bags was approxi- mately 4.5 in. H2O. The Sunbury steam electric station, shown on the inside cover, replaced an electrosatic precipitator collection system which was unable to meet the particle control efficiency required by state regula- tions. The fuel, a combination of low sulfur anthracite and high sulfur petroleum coke in proportions of 15 to 35 percent coke by weight, produced a high resistivity in the ash which led to decreased precipi- tator efficiency. The replacement bag filter system has met the Pennsylvania regulations since its initial operation in 1973. Each Sunbury baghouse handles 222,000 acfm at temperatures on the order of 325°F. Individual baghouses contain 1260 bags which are divided equally among 14 compartments. Every 30 minutes, the entire system is cleaned by a sequential cleansing of each individual compartment. The bag filtration system installed at these utilities is depicted schematically in Figure 1. The only major differences in the designs of these instal- lations is the variation in the fabric cleaning method employed by these facilities and the coarse particle removal system installed upstream from the fabric filter. The Nucla facility employs a combination of gentle shaking and reverse air flow for cleaning, while the Sunbury system uses only reverse air flow as depicted in Figure 2. The Sunbury facility also differs in the application of a mechanical collector which removes the larger particulate matter in the flue gas (about 70 percent) prior to bag filtration. At Nucla, a single deflection baffle is used to remove the coarser particles from the filter influent Figure 1. Schematic Diagram of a Flue Gas Cleaning System Incorporating a Fabric Filter Baghouse Figure 2. Gas Flow Through Baghouse Compart- ments During Normal Operation and Cleaning MECHANICAL COLLECTOR 1-GAS INLET DAMPER-OPEN 2-CAS INLET DAMPER-CLOSED 3-BAC COLLAPSING DAMPER-OPEN 4-BAC COLLAPSING DAMPER-CLOSED 5-OUTLET DAMPER-OPEN ------- Nucla plant profile ------- Sunbuiy - view of upper walkway, bag tension springs and bag caps ------- The U.S. EPA Industrial Environmental Research Laboratory in Research Triangle Park, North Carolina sponsored field sampling programs at both the Nucla and Sunbury facilities. Emission testing and analysis was performed by CCA/Technology Division, Bedford, Massachusetts. A major objective of the project was to characterize the performance of fab- ric filter systems used to remove particulates from boiler flue gases.* Total mass emission rates were measured using the standard EPA Method 5 techni- ques. Isokinetic sampling of both the inlet and outlet ducts, which provided accurate determination of particulate mass concentrations, allowed for precise measurement of bag filter efficiencies. Thirty-one tests were run at the Sunbury facility. The average mass emission rate of 0.0&46 pounds of particulates per 106 Btu of coal fired Corresponded to an average weight collection efficiency of 99.91 percent. Twenty- two runs at the Nucla Plant indicated an average mass emission rate of 0.01 pounds per 106 Btu input to the boiler and a collection efficiency of 99.84 percent. Both facilities easily satisfied the local particulate emission regulations that the previously installed I Sunbury electrostatic precipitators and the Nucla mechanical collectors had been unable to meet. Note that the emissions form both facilities are also considerably less than the New Source Performance Standards of 0.1 pounds per 106 Btu input for fossil fuel-fired steam generators. Filter efficiency as a function of particle diameter was also determined using inertial impactors. Daily isokinetic impactor sampling combined with condensation nuclei counter and diffusion denuder data on submicron particle concentrations provided accurate determi- nation of the mass mean diameter, defined to be the diameter at which 50 percent of the mass emitted is composed of particles with larger diameters. The mass median diameters at the outlets of the bag- houses from Sunbury and Nucla were 3.9 and 8.3 microns, respectively. The test programs measured emission rates for a variety of fuel compositions and boiler and bag filter operating conditions. No significant deviations from the average emission rates were noted for variations in firing rate, fuel sulfur content, or ash content of the coal-coke mixture. Only the condition of the filter bags was found to affect the mass mean diameter measurements. Measurements when new bags were in use indicated a slightly lower (about 10 percent) mass median diameter for the effluent particulate matter than did the results of tests on used bags. No other test caused any significant degradation of the fabric filter performance. The Industrial Environmental Research Laboratory's Utilities and jndustrial Power Division has issued reports on the Nucla and Sunbury Plants. More detailed information is available in the follbwing reports: EPA 600/2-76-077a: Fractional Efficiency of a Utility Boiler Baghpuse Sunbury Steam-Electric Station. March 1976. EPA 600/2-75-013a: Fractional Efficiency of a Utility Boiler Baghjause Nucla Generating Plant. August 1975. ------- Sunbury - view of exterior walkways to baghouse and exhaust ducts ------- Available information on the capital investments required for the fabric filter installations at Nucla j and Sunbury are summarized in Tables 4 and 5. Because both Nucla and Sunbury were retrofit con- structions replacing original mechanical and electro- static precipitator equipment, a direct cost comparison between the two facilities is difficult. The Sunbury construction costs, for example, include the installa- tion of an ash slurry pumping and piping facility for a settling pond more than two miles away from the plant. The data in Tables 4 and 5, however, do provide a useful frame of reference for judging some of the costs encountered in fabric filter installation. Table 4. Sunbury Steam Electric Station Bag Filter In Item "I Capital investments for ESP are comparable to baghouse costs. Figure 3 presents the estimated range of fly ash precipitator costs for both hot and cold side ESPs designed to remove high resistivity particulates. Cost data for a typical ESP installation at the TVA Gallatin Steam Plant is presented in Table 6. This ESP controls the particulate emissions from one 300 MW unit at a design efficiency of 95 percent. Annual operating costs of the Nucla and Sunbury baghouses are summarized in Tables 7 and 8. For 1977, the Nucla and Sunbury facilities had estimated annual operating costs of $1.15 per acfm and $0.82 per acfm, respectively. Using an electro- static precipitator operating at 92 to 95 percent, the TVA Gallatin Steam Plant operating costs were estimated to be $0.61 acfm. stallation Cost Breakdown (1972 $) aterial Cost PFour baghouses sJQesJgn and Engineering £ Vacuum cleaning system % Platforms and ladders : Supplements and contingencies Upland.and land rights -. Foundation and site preparation Ash slurry pump house KAsh removal system-baghouse sh slurry system ^Additions and improvements Electrical equipment ^Overhead ~ Total capital expenditure ST.: Total capital expenditure fe (1977 estimate)3 aSca!e factor from Chemical Engineering M & S equipment cost index 1972 - 4th quarter 1,266,985 30,415 95,105 39,800 90,800 241,500 343,000 v3"5a 55,400 35,600 Total Cost 2,286,985" " 563,140 74T22.5 116,310 161,030 " 1,500 192,500 179'60° 414,700 221,100 ' r 72,000 662,800' ,500,100' \ 1976. ------- Table 5. Nucla Steam Electric Station Bag Filter Installation Cost Breakdown (1974 $) h I Item : I Three baghouses ; Ash conveyor system aiiiRetrpfit items k Overhead tT Engineering and fee ,.. P P 3 i Total capital expenditure F Total capital expenditure j: (1977 estimate)3 *Tptal Cost B 1,740,000 250,000 210,000 , 120,000 300,000 Sz 2,620,000 3,188,000 t -, f 1 I 1 i i i i aScale factor from Chemical Engineering, M & S equipment cost index 1972 - 4th quarter 1976. Figure 3. Estimated range of fly,ash precipitator costs hot and "enlarged" (cold). (From JAPCA September 1974 article by N. W. Frisch and D. W. Coy of Research-Cottrell, Inc.) 20 o 15. 1000 MW plant subbituminous coal 99.5% efficiency 10' Cold precipitator operating resistivity, ohm-cm Table 6. Average Costs of ESP Operation at the Gallatin Steam Plant j: Capital Cost ~ * i '* i , i i t ESP & ! , ' t * ' i f s Direct Operating Costs 1 !, :.. , , ! I;; ; : I | Utilities 1 ESP maintenance *' Ash handling fuel consumption 2 Ash handling maintenance labor 1 and material j Indirect Operating Costs ! Total Annual Operating Cost 7 Total Annual Operating Cost (1977 Estimate)3 - IT- HI IS 1- P" r L ".fi~ " 1" ' " "^ "1,324,000 ~* f. ^ I^R tf3^ a, ai^, p (Annual average 1969-1971 "" »IS-St"lSSEttll^--fW1P 1? "* 7,835 "* " ' - ; ' "%5iD8( "" 2,200 58,980 r . i "2 a'j 164,100 235,215 343,470 , J f- s| r»« "f figures) rfSMSftS s* 1 t f _, 1 : 1 ""-> I sfTtST hi i i I i i s aSca!e factor from Chemical Engineering, M & S equipment cost index 1972 - 4th quarter 1976. ------- 1 Nucla - view of rappers located on top of baghouse ------- Sunbury - view of hoppers capturing flyash ------- Table 7. 1976 Nucla Bag Filter Installation Operating Cost Estimate K & iSi-^.---* ... . _ /year ^Percent j ia£^....r^.:^ Sft^'n', w>*w^T-i aScale factor from Chemical Engineering, M & S equipment cost index 1972 - 4th quarter 1976. Table 8. Average Annual Operating Costs at Sunbur^ S.E.S. Bag Filter Installation (1973, 1974, 1975) Direct Costs Operation and maintenance labor Maintenance material ^Utilities |. Ash handling Tndirect Costs l^pepreciation 4.9% t 3.4% ^Insurance 0.1 % s,TTaxes tjotal Annual Operating Cost , 1977 estimate3 aScale factor from Chemical Engineering, M & S equi Percent ^'"^"1 ^?,337 JASPS. complete baghouse Placement each year) jment cost index 1972 - 4th quarter 1976. ------- The successful application of fabric filtration to the coal-fired utility boilers at Nucla and Sunbury Indicate their potential for further electric utility installation. Bag filters should be considered for all new and retrofit particulate control systems when the following conditions are present: control efficiencies exceeding 95 percent are required effluent gas flow rate varies widely due to frequent load changes or regular cycling fuel composition is variable resulting in changing chemical concentration on the fly ash surfaces high ash resistivity causing inefficient particle collection by electrostatic precipitators. In some cases, high-volume systems using ESPs could benefit from the installation of a baghouse in parallel to the ESP to reduce the volume flow and increase its efficiency. New fabric materials have eliminated the excess maintenance requirements of the original baghouse designs, resulting in operating costs comparable to electrostatic precipitators. Initial capital investments are also similar and lower in situations where collec- tion efficiency requirements are greater than 99 percent. Therefore, as new, more restrictive regula- tions are promulgated and the use of low-sulfur, Western coals increases, the fabric filter will become the first choice for many utility boiler operations firing solid fuels. Sunbury - filter bag support frames and tension adjusting device This capsule report has been prepared jointly by Technology Transfer and the Utilities and Industrial Power Division, Industrial Environmental Reasearch Laboratory. For further information write to: Particulate Technology Branch Utilities and Industrial Power Division, IERL Research Triangle Park, N.C. 27711 ------- |