DOE/FETC-98/1059 Results of Stoker Testing on CONSOL Fuels CRADA PC90-003, Final Report December 18,1992 U.S. Department of Energy Pittsburgh Energy Technology Center 626 Cochrans Mills Road Pittsburgh, PA 15236 and Consolidation Coal Co. 4000 Brownsville Road Library, PA 15129 DISCLAIMER This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsi- bility 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. Refer- ence herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recom- mendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
Transcript
DOE/FETC-98/1059
Results of Stoker Testing on CONSOL Fuels CRADA PC90-003, Final Report
December 18,1992
U.S. Department of Energy Pittsburgh Energy Technology Center
626 Cochrans Mills Road Pittsburgh, PA 15236
and Consolidation Coal Co. 4000 Brownsville Road
Library, PA 15129
DISCLAIMER
This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsi- bility 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. Refer- ence herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recom- mendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
DISCLAIMER
Portions of this document may be illegible electronic image products. Images are produced from the best available original document.
CONSOL PC90-003
Results of Stoker Testing on CONSOL Fuels
Combusnon and emissions data are reported for the following fuels relevant to the CONSOL CRADA:
Pittsbureh Bailev Mine) Parent Stoker Parent Stoker Extrusion w/ paper additive Extrusion w/ paper additive - dried
Briquette 5 cc (no sorbent) Briquette 1 cc w/ Hydrated Lime
(Ca/S = 3) Briquette 1 cc w/ Coke Fines 0
All of these fuels have been tested in the warm-air furnace at a firing-rate of approximately 150,000 Btdhr . Two fuels (Pocahontas briquettes - sorbentless and hydrated lime) were tested in dx cast-iron boiler at a firing-rate of approximately 1.8 MMBtu/hr.
Tables 1 a d 2 summarize the pertinent analyses for these fuels. In general, the data are consistent although there appears to be discrepancies. particularly for the heating values and moisture content of the RDF-containing fuels.
STOKER TEST FACILITIES
Residentiai Warm-Air Furnace (WAF)
The WAF is a typical 225,000 Btu/hr (thermal input) residential warm-air furnace which is equipped with an underfeed stoker to normally deliver 16 lb/hr of coal solids, with two additional speeds at 12 and 20 lb/hr. At a typical warm air blower pressure of 0.1 ki H20, 1 100 scfin of hot air in the 150-225 O F range can be produced depending upon the air intake temperam and furnace firing rate. The furnacehderfeed stoker was manufactured by the Will-Burt Company. A cross-sectional diagram of the commercially-avadable system is presented in Figure 1. The WAF test facllity has a two-fold purpose: first, to establish a residentiai emissions data base under continuous, cyclic, and hold-fire operations; and second, to serve as a screening furnace for limited quantities of reconstituted coal fuels. On the latter 'sue, the WAF requires about 120 lb of fuel per test day. The WAF furnace was modified and heaviiy instrumented in order to provide research quality measurements whde operating he furnace in a realistic manner.
Before c e d i n g the various system modifications, it Is usettul to summarize the basic principles of residentiai stoker operation. In typical off-the-shelf underfeed stoker designs.
(a) fuels analyzed by Consol. All others were analyzed by PE TC.
(b) fuel analysis is pending. For now, calculations are based on the parent coal determinatlons.
( e ) denotes an estimate based on the reported percentayes of fuel components.
( f ) iisstiiiic:, t l u t all sullur IS converted to SO, and that all fuel nitroyen IS converted t o NO,.
4.1 1 85.29 1.07 0.83 1.98 6.72
14539
6.0
21 76 2319 2375 2422
1 .I3 2.39
Table 2 : Pittsburgh Stoker and Agglomerated (Extruded) Fuel Analyses
Pg h/RDF2"I Ex t r I I s ion
( : I l l ISOl
Pgh/RDF2 Pgh/RDF5'" Pgh/RDF5 Pgh/Pap (ar) Pgh/Pap (dry) Puh F x t r I i sion Ex t r I ]sic in Ex t r 11 si on Ex t r I I si0 n E x t r I is ion St ohoi
o oil sol (:ollsol (:01 ISOl c:Ol IS01 ( : ( J l l ! ~ l > l ( : I N I!.Ol
t1tW (Btu/lb AH) 10105 e 10770 9064 e 10563 10125 12520 13764
F s I NA 3.5 NA 0.5 3 .O 8 .O
A S I Profile ('F) Illltlal Softening t leni i lllld
NA NA NA
~ NA
205 1 2092 2144 21 79
NA NA NA NA
201 8 2151 2077 2190 2133 221 1 2324 2236
2036 2093 2127 2243
Lb/MMBtu (inax) so, (fl NO, ( f )
2.44 1.94
1.92 3.06
1.97 2.27 3.03 3.08
2.88 3.73
2.28 3.18
Table Notes
(a) fuels analyzed by Consol. All others were analyzed by PETC.
(bl fuel analysis IS pending. For now, calculations are based on the parent coal determinations.
(4 deriotes JII estiiiidte based (111 tlle lepoi ted percentayes of tuel componelits.
( 1 ) i l ~ ( i i i ~ e b tlidt all sulfur is coiivertutl to SO, and tllat all fliul i i itrouen is converted t o NO,.
--
. .
air flow is regulated by the arrangement of several dampers in key locations to ensure satisfacrory combustion air flow based upon fuel bed resistance. Coal is conveyed through a screw tube and up into the inner hollow pit (retort) whereby it undergoes increasingly severe changes beginning with devolatilization and ending wth combustion, characterized by a \sllow-orange flame. Combustion air enters the retort about 1/3 down from its b m througn several rows of small air holes known as tuyeres. The key characteristic of the underieed is its ever-changmg, irregular fuel bed as coke constantly forms, falls, and bums. This behavior is dependent on stoker design, operating conditions, fuel particte size, and most irnportantly, fuel type. In general, the highest portions of the fuel bed can be one foot above the top of the retort. The combustion air is mechanically linked to the stoker motor to accommodate both the coal feed and burning rate; thus, as the fuel bed resistance increases, the combustion air back pressure increases. This causes the damper to open more, thereby allowing more combustion air to bum off the excess inventory. Overfire air occurs via air leakage through small slats in the firebox door. Typical recommended draft for these systems is 0.03-0.06 in of H20. Unfortunately, the delicate balance of gas pressure variation does not permit the inclusion of additional flow resistances, including metering.
To provide flow metering capabllity, the built-in combustion air blower was replaced with a more powerful blower to permit installation of a venturi and pneumatically-controlled but tedy valve. Similarly, an induced draft fan on the stack permits flue gas flow metering (orifice) and control (pneumatic valve). A calibrated rotary meter was used to standardize the oririce and venturi meters to ensure consistent flow monitoring. With the modified systems. draft and overfire air flow control is extremely consistent even under cyclic (combustion air on/off) operation which characterizes residential heating. Figure 2 presents the overall flow schematic of the WAF test facility.
Other modifications included the installation of a 1) variable-speed motor to improve fuel firing consistency and minimize the effects of fuel density variability (coals vs. briquettes) on auger feed rates, 2) nitrogen purge line to the firebox in order to quench h e fire at the end of testing, and 3) sight port for visual observations, video recording, and dual-color pyrometer measurements of the firebox during testing.
Commercial Cast-Iron Boiler (CB)
The CIB is a H.B. Smith M450L 11-sectional commercial cast-iron boiler capable of producing hot water or low pressure steam with a designed firing rate of 3.24 MMBtu/hr for gas and fuel oil. The CIB was originally used to study coal water mixture firing using a PETC-patented rotary cup burner. The CIB was retrofitted with a Worley HW-9 underfeed stoker tthree speed 175 lb/hr. 120 lb/hr, and 75 lb/hr) to achieve a designed coal-firing rate of 2.6 MMBtu/hr which is the equivalent of the H.B. Smith 1 I-sectional Model 45 1 stoker boiler.
The CXB has a closed loop water system in which the boiler inlet and exit temperatures are
controlled at 160 O F and 180 O F , respectively. A site cooling tower provides external heat exchange for the boiler captive water loop. The CIB is essentially similar to the WAF in terms of the basic components and operation. Both units have variable speed auger drives to allow testing at constant firing rates. Nitrogen purge lines similarly allow quenchmg of the fire at the end of each test.
Figure 3 is an overall flow schematic of the CIB system which includes a cyclone and baghouse for particulate control. CIB instrumentation is sim13ar to the WAF in that the combustion air and flue gas flow are metered, having been standardized with a calibrated turbine meter. Unlike the WAF, the CIB has a separate, metered source of overfire air that, during operation, is injected directly into the firebox.
Analvtid Instrumentation
The WAF and CIB are well instrumented in gas analyzer monitoring. A Beckman Model 865 nondispersive inftared analyzer was used to track carbon monoxide (0-5000 ppm) cont inudy while a Beckman Model 880 analyzer provided additional CO measurement in the 0-500 pprn range. This arrangement allows accurate, continuous monitoring of CO without changing analyzer ranges during WAF operations. Similarly, two total hydrocarbon (THC) analyzers were employed during tests: a Beckman Model 400A was set at the 0-10 ppm range during continuous tests and a Beckman Model 400 analyzer provided coverage at 0-100 ppm. A Thermo Electron Model 10 chemiluminescent analyzer was used to track NO, emissions in the 0-loo0 ppm range. A Beckman Model 755A paramagnetic oxygen analyzer was employed at the 0-25% level while carbon dioxide (0-20% range) was measured with either a Beckman Model 865 or Model 880 infrared analyzer. Sulhr dioxide measurements are accomplished in combination at medium/high SO, concentrations with a Beckman Model 865 (0-2000 ppm) and/or Horiba PIR-2000 (0-10oO ppm and 0-5000 ppm) analyzers, and at low SO, concentrations with a moisture-compensated Horiba VIA-200 (0- 200 ppm) nondispersive infrared analyzer.
The WAF gas sample was conditioned prior to analyzer measurement by continuously withdrawing the sample through a Mott Metallurgical 2-in diameter cylindrical, sintered stainless steel mter of 5-micron pore size for removal of particulates. To remove moisture, the sample passes through coiled tubing submerged in a 32°F ice bath prior to flue gas analysis. Vapor-liquid equilibrium calculations (SOl/H,O) and ion chromatography measurements of the condensate revealed an acceptably low removal of SO, during moisture knockout, However, because of the inherent interference between H,O and SO, with IR absorption. a Permapure drier was installed to remove residual sample moisture from the ice bath-conditioned stream prior to the SO, analyzers. Otherwise, a 32°F dewpoint could produce an interference of about 6 ppm SO, based on the typically accepted interference ratio (100 ppm H,O = 1 ppm SO,) with infrared SO? analyzers.
WAF C Jmbustion/Emissions Screening Procedure
Screening tests are conducted in the WAF under continuous h g based on the desire to mimic *e conditions found in large commercial and industrial boilers. Typically, a 1-hr feedscrew test was performed to reveal fuel grindability and determine auger feed (rpm) settings - needed to aciueve an approximate constant finng rate of 150,000 Btu/hr (input). This point became important when earlier studies revealed that some fuels fed as much as 50% faster than the stoker coals. Feedscrew tests also reveal the degree of fuel degradation which is an important consideration in comparing the novel coal-based fuels with stoker coals.
After the feedscrew test, fresh fuel (2-4 lb) and wood lundling were placed over the retort. After lighting, combustion air was initiated for 10-15 minutes to establish the fire before initiation of coal feed. Typically, a warmup period was conducted to establish the proper amount of combustion and overfire air. Initial conditions were set at 50% excess combustion air plus an additional 50 lb/hr of overfire air and then altered, if necessary, to achieve more optimum combustion for the subsequent formal 6-hour test period. During this period, air flows ~ z r e generally constant and ~ n i y occasional adjustments were made to ensure a good test bum. Bed temperatures were obtained with a Capintec Instruments, Inc. Ratio-Scope Model ID dual-wavelength infrared pyrometer.
At the end of each test, combustion air and fuel feed were turned off; the firebox was allowed to cool for 10 minutes before flooding with nitrogen to extinguish the h e . The firebox ash was then partitioned into three categories according to its stage of combustion - powder ash, gray ash, and black ashkoke - to more carefully scrutinize combustion results.
Powder ash consisted of the most well-burned, finely divided (-1/4 in) material relative to the other two fractions. Gray ash consists of larger pieces (plus 1/4 in) while black ashkoke results irom the least combustion. The partitioned ash samples are then separately weighed, crushed. riffled, and analyzed.
The selected 150,000 Btu/hr h g rate was found to produce intense combustion conditions without overheating the carbon steel internal furnace components. The principal guideline here was to ensure that the flue gas exit temperature was below 950 OF. While average flue gas temperatures were in the 700-800 O F range at the 150,OOO Btu/hr firing rate, significant tempexamre excursions can occur, partmAarly for coking coals which can exhibit wide variances in burning rate. With respect to the WAF, several points are noteworthy. First, the older domestic U.S. stokers were not optimally designed for thermal efficiency, as the WAF is approximately 55% thermally efficient. Thermal efficiency can be dramatically improved upon considering the 80% values reported for newer European (e.g. Britain) domesnc coal-fired furnaces. Second, domestic stokers are really meant for cyclic (on/off) operation to meet variable heating demands throughout the days and years of the homeowner residence. In several older stoker studies, on cycle operation (in minutes) at full load was gearea towards achieving a 900 O F flue gas exit temperature before the onset of the off cycle. Finally. WAF operation at 150,000 Btu/hr did not produce an excessive quantity of firebox
ash over Lie test period.
The iniuailv targeted (about 50% excess air equivalent) combustion air flow through the retort was found to be sufficient for the better burning fuels. In reviewing the avadable literature. stoker testing has been reported at combustion air levels ranging from 20% to 300% exess air equivalent, although the majority of testing has occurred at the lower excess combusnon air levels. In general, higher combustion air flows have been proposed as a means of improving combustion of cakmg stoker coals.
WAF testing, however, revealed that high combustion air flows could produce two unfavorable conditions in the fuel bed. One is the possible formation of large quantities of sparklers (burning fine particulate), depending upon the extent of feedscrew degradation. which effectively poses an upper limit on "realistic" combustion air flow settings. Vendor operation instructions advise against operations resulting in sparkler formation due to various considemions, such as excessive particulate carryover and dangers associated with sparks escaping from the firebox door which is under only a very slight (0.03-0.06 in H,O) draft.
Another undesirable result of excessive combustion air is the tendency for intense combustion to occur deep inside the retort cup rather than the vendor recommendation of 4-6" above the retort brim. This problem can be particularly troublesome for hghly caking fuels for several reasons. In caking fuel beds, combustion rates are substantially less than that based on the fuel feed nte for significant periods of time. Since the most intense combustion naturally tends to occur at the retort periphery (in the vicinity of the tuyeres) and the highly caking fuels produce large/tall coke trees in the center of the bed, it is difficult to prevent the bed from burning deep inside the periphery of the fuel bed. where local air-to-(burning) fuel ratios are high. CIearly, the worst condition is a highly caking fuel which is heavily weighted with fmes.
The initiaiiy targeted overiire air (about 50 lb/hr) was essentially determined after adjusting the furnace draft and firebox slat openings to a "typical" condition as judged by a visiting Will-Burt representative during early WAF testing. This arrangement corresponds to approximately 20-25 % of the total air flow based on operation at a 150,000 Btu/hr firing rate using the above combustion air (50% excess air equivalent) flow. This level of overfire air is reasonable based on the majority of existing literature. Thel study of overfire air effects has not been aggressively pursued in the WAF due to the fact that results might be dependent on many factors (e.g., gas flow patterns, temperature, composition) related to the localized burningic&ing zones of the fuel bed.
CIl3 Combustion/Emissions Screenine Procedure
CIB screening tests are conducted in similar fashion to the WAF, although a longer 1-1.5 hour warm-up period is needed before testing at the desired continuous firing rate. This gradual u r n - u p period minimizes the possibiIity of cracking refractory surfaces and ensures
I good combustion as the firing rate is increased. Unlike the WAF, the CIB can not be operated at high loads from a relatively cold start as fresh fuel will rapidly pile up in the retort. The 1-1.5 hour warmup period, in which the firing rate is slowly increased, has been shown to alleviate this problem. The CIB operational procedure is summarized below.
Feedscrew tests were performed to reveal fuel grindability and determine auger feed (rpm) settings needed to achieve an approximate constant firing rate of 1.8 MMBtu/hr (input). One point oi interest was that the feed variability (Ib/hr per auger rpm) was much less in the CIB (highlow values within 15% spread) than that found (high and low values within 50% spread) in the WAF.
After the warmup period, initial formal test conditions were set at 50% excess combustion air plus an additional 150-250 lb/hr of overfire air and then altered, if necessary, to achieve more optimum combustion for the subsequent formal 6-hour test period. During this period, air flows were generally constant and only occasional adjustments were made to ensure a good test bum. Bed temperatures were obtained (in similar fashion as the WAF) with a Capinnw: Instruments, Inc. Ratio-Scope Model HI dual-wavelength infrared pyrometer.
It should be noted that the relative overfire-to-total-air-flow ratio is lower for CIB operations (in the 10% range) than that employed in the WAF screening tests in the 30% range. In prior CIB tests at higher overfire air flows, sight port observations revealed that the fire could become considerably cooled and deflected (towards the front of the boiler) in certain instances. Such visual observations are used to supplement calculated (e.g., burning rate, emissions) values to determine appropriate operating conditions. In comparing the CIB and WAF overfire air systems, differences in overfire air source (CIB uses 15 psig compressed air; WAF uses ambient draft air), nozzle diameter/number (CIB has four 1-in nozzles; WAF has twelve 0.5-in holes), and nozzle height (CIB nozzle 15 in above retort; WAF slat openings 20 in above retort) probably contribute to these observations.
At the end of each test, combustion air and fuel feed were turned 0% the firebox was allowed to cool for 10 minutes before flooding with nitrogen to extinguish the fire. The firebox ash was then vacuumed out of the furnace. It appears that the coking characteristics of the p e n t coals are more suppressed than that observed previously in the WAF. In addition. the CIB does not seem to produce large, blocky coke (as was the case in some WAF mts) but does tend to produce large clinkers (fused ash) for some of the lower AST coals.
The selected 1.8 MMBtu/hr firing rate was found to produce intense combustion conditions and achieve the vendor-stated 65% overall thermal efficiency. A factor in f h g rate seiecnon was the baghouse miet temperature which must be maintained below 425 "F to prevent damage to the Nomex bags. It should be noted that baghouses are not used for units of this size in actual installations: typically, particulate control is achieved with (cheaper. simpier. lower maintenance) cyclones. Of course, as developments in higher temperature filtration media (e.g., ceramic) proceed, baghouse-type filters could be developed for small
commercial stokers. At present, the PETC system relies on a nitrogen/water spray nozzle humidification system, located about 10 ft upstream of the baghouse mlet, which can provide a maximum of 180,OOO Btu/hr of coohng which could reduce the flue gas temperature by approximately 250 F". This translates into a maximum boiler exit temperature of about '7.50 O F (accounting for other pipe heat losses between boiler exit and baghouse inlet) to maintain a 400 "F inlet baghouse temperature and allow for normal fluctuations in CIB burning rate.
Another key consideration in selecting the firing rate was firebox ash accumulation. While the CIB has a peak firing rate approximately 14 times that of the WAF, the CIB design does not appear to be as accommodating with respect to ash buildup capacity. For example, the CIB firebox floor area is only about 5 times that found in the WAF; in considering the area surrounding the retort, the CIB is only 4 times greater than the WAF. One possible reason for this apparent size discrepancy is that an actual CLB installation would likely have a boiler operator to perform periodic ash cleanout while a residentially-sized unit (e.g., WAF) may be designed for less frequent ash removal by the homeowner.
As was the case with the WAF, units in the CIB's size range are similarly meant to cycle (on/off) in order to match variable heating demands at the boiler site. For this reason, continuous operation at reduced load (approximately 70%) seemed appropriate as was found previously from materials considerar;,ons in the WAF.
During CIB testing, the baghouse was pulsed periodically (usually several times per test) when the overall baghouse pressure drop exceeded 5 in H,O. This translates into a 3 in H,O drop across the filter cake given the normal baseline 2 in H,O drop across the baghouse at typical flow rates. Because of periodic pulsing and size considerations, the CIB is operated at higher drafts than the WAF. Typically, CIB testing was performed at 0.15-0.40 in KO draft to prevent flue gas seepage from the boiler front panels during pulsing periods and burning rare fluctuations. This level of draft produced measured air in-leakages of about 80 to 200 lbrhr for the 0.15 to 0.40 in H20 draft range, respectively, which represents only about 3 4 % of the flue gas flow rate as determined from the sum of the combustion air, overfire air, and fuel (gas equivalent) burning rate. In CIB cold flow tests at (nearly) zero draft, gas flow agreement was generally 98% or better, upon comparing the boiler exit (venturi) flow to the sum of the combustion and overfire air flow orifice measurements. For this reason. it is felt that the small air in-leakage has minimal influence on the most important CIB experimental findings.
FEEDSCREW DEGRADATION RESULTS
Table presents the WAF feedscrew test results. Samples were sieved at the following sizes: : inch, 3/4 inch, 1/2 inch, 1/4 inch, 8 mesh, 28 mesh, and pan. To simplify the resuits. size distributions are grouped into the most important fractions which are taken at the 1,: inch and 28 mesh sizes.
Table 3 : Size Distributions after WAF Feedscrew Test
The PgM2DF2 and PghRDFS extrudates were excellent as very little size degradation was observed. Size distributions of the two extrudates were comparable to those observed with typical U.S. stoker coals. These extrudates were superior to the PETC coal-based briquettes owing KO the combination of their slightly larger size and apparently (no measurements provided) greatex compressive strengths; compressive strengths for the PETC coal-based briquenes were in the 14-28 lb range. These factors also likely explain the degradation resistance of the large 5 cc Pocahontas briquette.
In conrrast to the extruded coaURDF fuels, the Pocahontas/coke briquette underwent considerably more breakage, resulting in a size distribution on the lower side of what has typically been observed with briquetted fuels tested at PETC. However, the size distribuaons of briquetted fuels after feedscrew tests have varied widely as shown in Table 3. Indeed, for one briquette (Pocahontas/hydrated lime), excellent structural integrity was observed. This suggests that some "fine tuning" of the briquetting process can achieve dramanc results, particularly when considering the closeness in crushed coal particle size (14 mesh wth 20% less 100 mesh), binder addition (3-4 wt%), and also sorbent dosage (0-1 1 wt%) used in the PETC briquetted coal-based fuels.
The Pittsburgh codpaper extrudates were excellent in WAF feedscrew tests. Both the as- received and dried material fed with less than 10% breakage. This was determined from PETC hand screening at 1/8 inch which was performed (in lieu of sending the samples to an outside contractor) because of the limited quantity of fuel and the desire to conduct additional WAF tests.
Screening of the CIB feedscrew samples (by Warner Labs) has not been completed as yet. However. a few observations are noteworthy. The CIB appears to be significantly more destructiye of the briquetted fuels which entailed 1, 5, and 7 cc briquette sizes as compared to the WAF. The main reason for the increased crushing is the larger force exerted at the screw inlet when the feed hopper is filled. This can be alleviated somewhat by feeding the fuels at a lower hopper inventory. In a few crude tests, breakage was reduced from about 90% to about 45% by maintaining low hopper inventory.
COMBLSTION/EMISSIONS RESULTS
Table 4 summ&s the combustioniemissions results in the WAF and CIB facrfities. Emissions data were calculated based on results obtained during the last 4-5 hours of each test, where operations were the most stable. In the WAF, considerable difficulty was encountered in h g the combustion/emissions screening tests for the Pittsburgh stoker, Pocahontas #3 stoker, Pocahontas 5 cc briquette, and Pocahontas/Coke 1 cc briquette. These fuels resulted in highly irregular fuel beds which produced poor combustion conditions as indicated by the low carbon conversions (64-82%) in Table 4. Aidfuel conditions were adjusted during the test to try to optimize combustion; however, success was limited. In general. these caking fuels seem to bum better at hgher firing rates and total air flow rates, although furnace temperature constraints can become a factor here. Some success was found (in the Pocahontas series) by lowering the combustion air and increasing overfire air to help bum tall coke trees. For this reason, the Pocahontas stoker and sorbentless 5 cc briquette were actually conducted at sub-stoichiometric conditions in terms of retort combustion air (overfire air not included) as shown in Table 4. Particularly troublesome was the PocahontasKoke I cc briquette which produced considerable quantities of fines which effectively limit air flow rates due to sparkler formation.
In sharp contrast, good combustion in the WAF was achieved for the RDF- (Pittsburgh cod), paper- (Pinsburgh), and lime- (Pocahontas #3) containing agglomerated fuels which exhibited carbon conversions in the 9598% range. These fuels burned with virtually no adjustment in aidfuel flow during the tests. The principal reason for the improved combustion was a sharp reduction in coke tree formation. In contrast to the above fuels which produced large, blocky coke, these fuel beds consisted of very porous and brittle coke which naturally Fractured and burned. PETC's results with other parent stoker coals which underwent considerably less coking when agglomerated in the presence of even low (3-5 wt%) dosages of sorbent are described in a recent paper which is enclosed.
Table 2 : WAF/CIB Screening Results of CONSOL Stoker Coals and Agglomerated Fuels
Cust-Iron Eoilur Tests 4ir Fiimoc
t’bl I EXT-dry PAPER
-- Warm-Air Furnace Tests _I ----
)‘CY PC Y P‘CY t’CY STOKER BRQ BRQ BRQ
Ce/S=O COKE CalS=3
16.0 16.9 16.9 11.4 234 240 227 151
Tests ---- Pcill E X T
RDF 2
--- t’t i l I E X T
RDF 5
-- I
- I
-_I
I’C Y l’(. Y BRQ BRQ
CalS=O CaIS = 3
121 152 1750 2050
1950 2100 150 200
2100 2300
45 35 58 47
9
4 2 - 7
4.8 . .--
10.9 1 1 1
42 68 392 348 187 142 2.1 4.3
9.2 8.1
19.4 18.2 0.0 0.0 74 111
697 568 333 232 3.7 7 0
14.6 157
15.1 160
12.7 158
FUEL FEED RATE (LBIHR) FIRING RATE IN 1000 BTUIHR
COMBUSTION AIR (LB/HR) OVERFIRE AIR (LBIHR) TOTAL AIR (LBIHRI
% EXCESS RETORT COMBUSTION AIR O h EXCESS TOTAL AIR O h OVERFlREflOTAL AIR
20 1 196 80 60
28 1 256
53 95 114 154 28 29
3 .O 3.7
6.6 6.9 13.5 13.1 1013 1027 659 476 155 158 42 13
18.5 18.3 0.0 0.0
2825 2731 1839 1265 432 41 9 118 36
205 69
274
64 119 23
3.1
184 53
237
54 98 22
183 50
233
51 92 21
164 116 280
- 7 58 41
168 94
262
- 10 40 36
3.5
303 224
37 23
3.2 3.1 ~- FLUE GAS MOISTURE (mol%)
AVERAGE FLllE GAS (dry)
0, (Yo) - ACTUAL CO (PPM) SO, IPPM) NO, (PPM) THC (PPM)
co, (I%)
5.6 4.6 4.0
7.6 12.05
607 502 148 8.3
9.8 9.8 115 575 120 1.3
8.1 11.4 178 376 117 3 .O
7.3 12.7 851 189 128 4.4
9.2 10.5
1185 361 112 4.1
12.5 7.2 435 56 1 115 9.9
3.1 10.0 162 171 68 1 .o
AVERAGE FLUE GAS (dry) co, I % ) 0, (XI - COhHECTED T O 0% CO (PPM) SO, (PPM) NO, (PPM) THC (PPM)
LBIMMBTlJ (FUEL ) co
NO, us NO, THC
so,
18.4 0 .o 21 6
1078 225 2.4 _..-
17.9 0.0
1423 1177 346
20
18.8 0 .o 655 853 172
15
17.7 0.0 389 823 256 6.6
18.4 0.0
2370 722 224 82
18.5 17.4 9.0 0.0
21 53 309 478 326 324 130
11 1.9
2.12 0.22 0.50 0.55 0.26 0.16 0.0020 o.oooa
0.99 1.84 0.39 0.0077
0.35 1.88 0.30 0.0031
0.48 1.12 0.30 0.0041
1.50 0.96 0.22 0.0047
0.44 1.25 0.20 0.0060
2.36 1.94 2.71 2.01 0.54 0.49 0.055 0.015
0.055 0.083
0.0016 0.003
0.020 0.072 .-.
CARBON CONVERSION (So) /I FLYASH (GRISCF) + 0.020
97 98 70 1 :3 I 64 I ,%:* 0.022 0.1 15
96
0.030 0.036 0.030
During Psting, it has become clear that the CIB facllity is much better suited for the combustion of caking fuels. This is dustrated in Table 4 upon comparing the sorbentless vs. lime-containing Pocahontas briquette results. Clearly, the marked improvement found in the WAF operations was not nearly as distinguishable during operations in the CIB. Avadable Literature indicates that the underfeed stoker design is critical for successful combustion of caking coals, as evidenced by the widely variant accounts of caking coal compatibiiity with underfeed stokers in the abundant pre- 1960 studies. One reason for this effect may be that the CIB fuel bed appears to be about 100-300 F" hotter than the WAF which typically exhibited furnace temperatures in the 2000-2700 O F range, with the more common average values in the 2100-2400 O F range. The higher CIB bed temperatures probably accounts for the lower (1520%) SO, capture of the lime-containing briquettes relative to the 40% removals observed at Ca/S=3 for briquetted fuels in the WAF.
In comparing lb/MMBtu SO2 emissions (Table 4) with predicted fuel analyses (Tables 1 and 2), it appears that approximately 20% sulfur capture has occurred for the Pgh/RDF2 fuel and about 40% removal for the Pgh/RDFS fuel. PETC performed elemental ash analyses on these samples with the following results
29.48 16.73 7.04 0.97 0.00 16.84 1.20 2.34 1.16
13.26
27.03 15.47 7.52 1.03 0.00 19.52 1.07 1.63 1.14
16.93
These results translate into a CdS mole ratio of 1.5 for the Pgh/RDFS briquette and 1.3 for Pgh/RDE briquette. The elevated calcium content of these fuels could explain the observed sulfur capture, although the 40% sulfur removal for the Pgh/RDFS seems high given this Ievel of dc ium.
NO, emissions were in the 0.26-0.54 lb/MMBtu range during all tests. This corresponds to about 515% of the fuel nitrogen equivalent, given the standard assumption that thermal NO, is minimai under stoker combustion conditions. Other emissions (THC, particulate) were in the Same range as other tests i n h e WAF. One interesting note is that the CO emissions were the lowest in the CIB tests, probably reflecting the higher bed temperature in the CIB as compared to the WAF.
c I
Induced Draft Stack
Natural Draft Stack Vented
Hot Forced Air ID Fan I
I
, I
8 ' 0
To Sample Train
Door Ope
) - r ) I': - -
Coal Hopper
Ambient AIr To Be Heated
Venturi r Warm Air Furnace
0 Combustion )-$ AirBlower
Figure 2 : Flow Schematic of Warm-Air Furnace Test Facility
I * Bag
House
lsokineticl Particulate Sire Sampling Port I.D.
Fan 1
ZI r I ’ Gas Sample Underfeed
-
Venturi Cast Iron
Boiler
1 Orifice I- T -
Return Hdr. I
-L Overf ire Air
- Orifice 3
I-< Corn bust ion Air
Figure 3 : Flow Schematic of Cast-Iron Boiler Test Facility
Emissions f3alua tion u i Eriaue rted Cm/Sorbent Fuels
!,faric C. Freeman, S e m y W. P2mi ine . James 1. Joubert Pittsburgh Eze rF TzcSaology Center
U.S. Department or Ezergy, ?ittsburgh, P.4 15236
The in-aouse rlue gas cieanup program at me ?imourgn Energ Technology Center has c o n d u c ~ a studies to evaluate novel coaVsorbent bnauetred fuels aimed at meeting stringent emsions standards in conventionai coal-frred small-scae comDustors. Results to date have been encouragng, as typical SO2 removals of approximatelv 40% for briauened coallsorbnt fuels were observea at calcium-to-sulfur molar ratios of about w e e :o one. For one fuel. SOz removals exceeded 50%. Marked improvements in combustion charactenstics were also observed for the briquetted fuels conraining sorbent. even at dosages as low ;is 3-5 wt%, for five parent coals. Other sigiiicant findings are presented along with an overview or tix expenmental facilities.
-
INTRODUCTION
The Pimburgh Energy Technoiogy Center i P E T 0 of the U.S. Department of Energy has supported itle deveiopmenr oi novei coal-based fueis m a combustors that couid displace natural gas ana fuel xi in snail-scale (c 100 MMBtu/'nr) hestlng systems. In 1988. the PETC Rue Gas Clzmup ?ro_mam issued a soiicitation enarled "Z.mssions Csnnoi for Smaii-Scale Combustors" Li which :missions oi 0.4 Ib S02i?tIMBtu, 0.3 !b NO.,3,i,tfBtu. ma 0.02 Ib pamcuiawMMBtu were targeted for new combustors while industnai reoont ( 10-100 MMBtltltltltltltltll rargets were set somewhat iigher a 1.3 :5 SOfllMBtu, 0.6 ib NO,I?IIMBtu. x a 0.03 IS pamcuiate/MMBtu (11.
The current in-house PETC research effort seeks to characterize the emissions of traditional stoker coals ana novel coal/sorbent fuels in actual smail-scale furnaces under redistic operating conditions. These characterizations are needed to assess present emissions and potenrial emissions controls with stoker technology. Several motivations are behind the evduaaon of reconstituted coal fueis for stoker-rired combustors. At present. a signiiicant opportunity exists for agglomeration or by-product fmes (14-28 mesn) from conventional cod-cleaning operanons. This opporruniry is magnuie-ci in advanced coal cleaning processes, which produce much finer (-325 mesh) marerial. whife cod reconsnrution efforts are typically aimed at uriiity appiicauons, Lie smaii-scaie stoker market couid provide early penenanon potentiai for such fueis based on the more favorable economics associ3red with the cost premiums (approximately S2-7/ton) charged for stoker-sized versus utility-sited CCU j3j. Finaiiy, tfie inciusion of small amounrs or sorbent in the reconsdtutea fuels offers the potennal 'io iower SO, emissions. To achieve the 0.4 lb SO,i?cLMBru lwei for a 13,000 Bwib coa. SO, nmovais need to be in the 48 to 73% rmge for coal suLfur contents of 0.5% to 1.0 srt"7o. respectiveiy. These ieveis of SO? capture mgnt t e ricnievaole using coahoroent r;:eb ~ i a zornmercial stoker technoiog.
.4 Cscxon was made to conduct screenings wiLh actuai stoker rumaces rather than building a batch i3r seziondnuous simuiator furnace. To minirmze ccaiisorbent fuel fabrication costs, underfeed stoker its were seiected based on their commercii avaiability at low thermai inputs. It was felt .,'la[ ~-2 zest imponant combustionlemissions expenmenral resuits would be transiemble to the ravecg-grate and spreader stokers more comnoniy icuna in large commeraaVindustrial boilers. Tfiis 2s:;mprion was based on me bulk cr :he existing stoker iiterme.
...
A 2 X 9 0 Bm/hr (input) resiaenual warm-air furnace (WAF) was obtltlned from the Wi-Bun Compny to perform the emissions evaluations. Under continuous operation, the WAF R ~ U K ~ S only I20 Ib of fuel per test day. The WAF was rnoaified to provlde accurate data under realistic f u m e operation. For exampie, the built-in combusnon air blower was repiaced with a more powe*i biower to pemit instailation of a ventun and pneumancally controlled bunenly valve. S i m h i s , an induced drdi fan, onfice, ana control system were instailed in the flue gas stack. X caiibmed rotary meter was used to standardize the ontlce and venturi meters. Overrire air is r e g u k u by adjusung fiebox door slat-openmgs ana the furnace drait Figure 1 presents an overrtll tiow xnematic of the WAF test iacliiry. Orher WAF features include 1) vanable-speed auger motor drive :J adjust feed rates, 2) ninogen purge for fire quenching, and 3) sight pon for visual obsen-mons, video recording, and pyrometer rmdings. An existing natural draft stack was not unii7PI, in this study. Furrner de& on the WAF have been given eisewnere (31.
Beclr;=...an Model 865 nondispersive infrarea (IR) vlaivzers are used to track carbon dioxide (0-20%) ma cz3on monoxide (0-5000 ppm) emissions. A Beckmm Model 880 IR analyzer provides aesuiiment or 0-500 ppm CO which allows more precse continuous momtoring without chanpg miyzcr ranges. This capabiiity pemirs accurare measurement of CO ermssions that can exceed 3 m or' magmtuae aunng a test. In similar iasnron. two total hydrocarbon (THC) analyzers are tmpicya. h Beckman Model 400A provides coverage in me range or' 0-10 pprn during connnuous screez~ig tests while a Beckman Model 400 allows (3-100 ppm monitoring. .\ Themo Elecaon Mxic: :O chemiiummescqnt NO, analyzer (O-lOOO ppm) is used as weil as a Beckman Model 755A prmzgnetic oxygen analyzer (0-25%). SO? is momtored wirh Beckman Model 865 (0-2OOO ppm), 9 Hct1x PIR-2000 (0-1000 ppm), and a moarure-compensated Horiba VIA-200 (0-200 ppm) nonrikmmve infrared analyzers. A Kave Corporanon dam logger transmits gas compositions and other rzaings (e.g., temperatures, pressure drops) to a DEC PDP 11/70 for data storage every 24 heconas. F o m programs calcuiate the relevant engmemng quannaes - for example, lbhr of SO?, Bwix ci flue gas sensible heat. etc. - which are then mtegated over the iength o i the test to prow& absolute quannries for matenal ana energy baiances. Such emissions integrations nave not 3een -coned previousiy in the iitemure.
A ikt $as samule was connnuatlv withdrawn through a >¶OK .Metallurgical 2-in diameter cylinr=lcai, srnteria stainiess steei I'ilter ( 5 microns) for parncuiate removal prior to sampie momire h c c K a u t in a 32°F ice bath. Vapor-liquid zcuiibriurn calculations (SO&O) and ion chmxography measurements or' the condensate reveaied an acceptably low effect on SO, aunng noism knockout. .I\ Perma Pure drier rimher rcauces residual sample moisture wnich would otheru-se increse r e d n g s sigmiicantiy at low ieseis (< 100 ppmr or SO1. A 32°F dewpoint couid proauce an mtert'erence of about 6 ppm SOz based on the typicaly acceprea interierence mu0 i 100 ppm E:O = i ppm SO-) wirh intiarea SO2 ma~~ze:s.
Tabie. I ;.resents the anaiyses for the coais that i x i u d e Hazard $4 (Kentucky), Brooknlle i,Pennsyivaniaj, Twin Branch (West Virginia'), Blue G e n (Kznruckyj, m?Ci Pocahontas 83 (Virgmta). The CXL differ in freeswelling mdex. ash soiteni~g ~apenrure, TU aci chemical content The Blue G e z coal was procured in both a raw (4% asn) ad clemeij (1% ash) stoker form to simuiate h e c:ocxc.s from advanced cod cleaning processes: t;.,us. ihese coals are reierred to as Blue Gem bledium i s h ana Blue Gem Low Ash, respecdveiy.
Coausomnt briquened fuels (3/4" x 3/8" x 5/16") were prepared wth a Fenotech Model WP-10 Briauemz using a srarch-baed BREWEX (RDE Incorporated) binder. The selected sorbens for the mriai screenings mciuded hydrated lime (Mercer Lime Company), a low-magnesium dolomitic limestone {Sationd Lime and Stone Co.), and a high-punty caicitic limestone (Dravo Corporation). E x n b a t a of briquettes was formed from il mesh crusned cod to simuiate conventionally cfevled fines. i*&ugh it is recognized that conventionally cleaned fmes couid differ in other characteris&. .A h a r m high-speed mixer was used to blend crusned cod, binder, and sorbent; in several :nstances. xater (0.5-2.0 wt%) was added IO rne n~~xmre to minimue dustiness. Table 2 presents bricuexe compositions, laboratory strengh cetermmations, and heating values. 8riqueuing procede2 smoothly for the heated (155 O F to 185 OF) blends, resulting in satisfactory compressive ( r y p i c d ~ aver 18 Ibs) and six-foot drophhatter (rlpicaily 90% plus 1/4") strengths. Because of the !ow sment and binder dosages, the briquettea fuel heanng vdues were acceptable (over 89%) r?!isnve := the parent stoker coals. The CuS rmo for rhe fuels was based on the caicium content or t;?e s m n t and on the ruifur content of the cod.
Scre,o?mg :ests were conducted in the WAF under conunuous iiring based on the aesire to rmrmc h e C C ~ = I O I I S found &I Iarge commercial and indusmai borlers. Typ~caily, a I-hr feeascrew tesr xas p x f m e d to reveai fuel grindabiiiry ana a z t e m e auger feed (rpm) settings needed to achieve ..n approxmate consrant firing rate of i90.000 B w h r (input). This point became important when Zxuer sxiies reveaied that briquettes red 3s mucn zs 508 faster han the stoker coals. F i w 2 m n p m me average size dismbutions ior the fed bncietted heis and fed stoker coais along with rzw S:CE coals. Briauetted fueis were crescreenea to remove fines prior to feedscrew ana ern~~sions screening tests.
.After x c feedscrew test. fresh fuel (2-4 Ib) ana ivooa kinaling were piaced over the reton After lignnng. :ombustion ;iv was initiated for 10-15 minutes to establish the tire before initiation of cod kea. Tmrcaliy, . - 3 warmup period was conducted to establish the proper amount of combusuon ana ovenire a. M u d conditions were set at 50% excess comousaoq air pius an additionai SO Ib/hr or overi-2 air and then altered if necessary, io acnieve more opmurn cornbusuon ior the suoseauex formd 6-hour test period. D U M ~ rhrs penod. air flows were generally constant ana onis crcczsionai adjustments were made to ensure 3 gooa test burn. Bed temperatures were ootamed with a Cqmtec Insmrnenrs, Inc. Ratio-Scope Moaei III dud-wavelength inbred pyrometer.
. i t IT'.: 2:~ of e x n rest, combusuon air and fuei feec :sere turned oif: t'le firebox was ailoweu IO cooi fcr :I minutes before tlooding with nitrogen to e ~ ~ g u i s h the fie. The firebox ash was men ~arnccma into three categories according IO its stage o i comousnon - powder ash. gray an. ma j lacs 3; coke - 10 more carefully smmze comDusnon resuirs. ?owaer ash consisted oi rhe most w e & h r x i . h e i v divided (-1/4 in) matenal reianve !o me other nvo frxnons. Gray asn corntea
a i 1~3:: jieces (plus ii4 1111 d13t general! ma unaeqone mreneaiate cornousuon. Black asnjcoke res~1'fi irom the least comDusnon, conssung pnrnaniy o i h e iuei ied near the end or the test The zat-zcsed ash samuies were then separately weighea. crusned. zfilea, ana d y z e d .
RESL'LTS .AND DISCC'SSION : Screerzix with Hztzizard =A
Take I surnmanzes Ferment WAF comDustiowemIssions resuits for Hazard # stoker ma bnccerzd fuel analogs. The purpose of these ESKS was to determrne ir' So, removah were snongiy dewcent on the type of sorbent (hydrared lime, caicinc hes tone , and dolomitic limestone) at a c&xm-to-suifur ratio of approximateiy 3:l. A s indicated in Tabfe 3, sulfur removals for all three briaurres ranged between 3643% as calculated from the inregrated mass rlow of sulfur (as Sod over ::e entire test iength and the amount ci sulfur in the cod feed Bed temperarures rlucruated c o b b l y in the 2000-2?00 O F range ma averages were in the 2100-2400 "F range.
Of ECZ was the sulfur capture for the Hazard +$ stoker and sorbentiess briquette, which were in the I-; ;5 range due to the incomplete ComDusuon o i suifur. Approximately 30-90% oi this 'suifur
b l a ~ x e / a s h are quite close (within 2 307i'C) to the parent, indicating similar Ievels of combustion for rzt sulfur- and carbon-containing species. Lltimateiy, these trends wid aid projecnons of SO, rernmitls based on a fixed combustion efficiency in future repom.
-0 . b r l r C . n, e.- ;I - ' actually resides in the black asdcoke collected in the test. The sulfur/caroon m o s oi the
D u ~ z a i i tests found in Table 3 for Hazard st stoker demonsnate WAF repeatability. Of pmcular agrezent were the key combustion indicators. such ;?s caroon conversion and fuel bed (ash breaiiowns) caking tendencies, and key emissions indicators, such as moiar average (pprn) and inregzza mass (Ib/MMBtu) values. In comparing the Hazard #4 stoker and briquetted coaVsoroent heis. 1: must be remembered that a srrcng emphasis was placed on conducting screening tests at nt~-:znmurn combustion conditions w~thm each tesL cay. For this reason, several operating p a r z t ~ r s (e.g., excess d) are not fixed m c s s rhe Hazard series in Table 3.
As a n t e d in Tabie 3, all the NO, emissions were low, in the 0.39-0.55 Ib NO,@MBtu range, ana amin the DOE _god of 0.6 Ib/MMBm ior reuorit combustors. NO, emissions averaged from 1 7 6 - 3 upm range on a dry basis. This range of NO, is comparable to about 10-15% fuel niwgen convemon during combustion, assuming that all of the SO, resulted from fuel nitrogen pathways.
Whik 50, emissions were fairiy constant uxougn the iiazara senes. CO and THC zmissions were more vmable, in the 181-932 ppm (0.3-1.50 IbMMBtur and 3-16 ppm (0.004-0.016 lb/MMBtu) range. respectively. CO ana THC are extremely sensldve, varyng up to an order oi maptucie dunce 3 est Other species (e.g.. CO,, O1. SO?, XO,) rypically rlucruate within 2 2040% of the inem-xcause of their more linear relation to the bumng rate. CO and THC zrnissions. however, vary - ~ ~ a e l y as coke forms, fractures, ana begins to burn.
' Ta5k 3 weals that fly ash occurred in a range of about 2- 10% of firebox ash levels as judged from isor.-.nc sampiing durrng rhe Hazara smes tests. Uniorrmatelv, particulate sampiing in the WAF is szmniicated by a small diameter SECK t8 in'), fluctuanons in stack temperatures. ma most i m F a u y , fly ash deposition at the r e s of the furnace ana other unavoidable low veiocity (less ihan - 3 Pisec) areas. Stack cieanines mer i0-15 tesfs reveaied hat deposiaon was cornparaole to
ioc-zz: smpiing vaiues. Ciasequentiy, x p m d ~arr!c.::sx iosaings mciude a multi-test average ieprxzsn factor. Particuiare emissions m e e d born 0.3-1.: 1S;MMBtu in the Hazard screening -fits. i?d here was no apparent nend w i h sorgent inclusion. 7 : s is not e n k i y unexpected given *e izzz number of variabies, sucn as ruei-kea-parncie size. sn content. cornousdon air flow me, =a xzmstion efficiency, which influence r?y ash emissions.
tZcTr:mon/Emissions Trecds - a l u e Gem. T*.vin Branch. 2rooK\iiIe. and Pccahontas #3
? r o c :it Hszard Nsorbent sGeening results. hydrared lime was seiected as h e principal sorbent ior ~t x x t senes or’ briquened coaVsorbent iueis, based on the &sm to maximize briquened fuel heanq values by minimizing sorbent load with (high-cdciurnl hydrated h e . Briquerted ~ o & j ~ ~ n t fuels at Ca/S=3 were prepared for the Brookviile, Pocahontas #3, Twin Branch. Blue Gem L3w Ash, and Blue Gem Medium Ash: Sriquetted fuels at Ca/S=2 and Ca/S=O were ais0 p e p a m for the iatter three coais. WAF rem were then conducted for all of these fuels.
In s e v d respects, these WAF teas resulted in trends simriv to hose found in the Hazard +4 series, Cor example, SO, emissions inferred about the same degree (10-15%) of fuei ninogen convexon. Other similarities included the range of bed temperamres, try ash loadings, and suifur :erencca levels for the stoker coak and sorbenuess briquettes. In general, these findings were in 3geezzem with the literature i4-51. Some oi L!e more important nends are highlighted below.
Fi=gtm 3 summarizes the observed SO? emissions for these stoker coais and their briquetre analogs, althccp it is important to redize the suifur rsention factor md its combustion dependence. Resuis :&ere simiar to the Hazard 84 senes wxh observed SO2 removals in the range of 3545% for ti? 3:vdrated lime-containing briquened h e i s at Ca/S=3. X norable exception occurred wrh the Blue gem Low Ash Ca/S=3 fuel, which exhibited a hisher SO, removai of 54%. There is no ~ p p a x c : explanation for this enhancement. The sorbent to coal ash ratio of 6.4 for this Blue Gem Low .%a briauetted fuei is substantially sesrer. however. than hat found (0.5-1.6 range) in &e xne: xi/so;ixnt fuels. nis may nave :sa to improvea !-iyarated lime-so, interamon oy sLlDCZ5SIng physical encapsuiarion or‘ C3iCiUm oy inert ash or by reducing chemicai scavenfpg due :o rex=ms or calcium oxide with an species such as silica or dumina
From t zrocess standpoint, the high SOz removai associated wirh the low-ash Blue Gem bnauene is encc&gmg ii advanced coal cleaning couid be integrated wirh reconstitution schemes to proauce 9 hie cuaiity cod-based fuei. On an energy equivaienr basis, the observed SO, emission of 0.48 3/?,1Y3m is very dose to the 0.4 lb/MMBtu DOE program goai for new combustors. The Twin 3raiiz2 fCrl/S=3) briquerre also approached this goal ;fs shown m>r”igun: 3.
The zcuetred coailsorbent fuels or Ca/S=2 had lower SO, removals. in the 2535% range. Of note is RE xi coai/sorbenr fuels resulted in SO, emissions well below the 1.2 !b/MMBtu god suggested far ~:cr;su-iai renorit combustors. From Figure 3, 0.6 lb SO~,’zrfilBtu is an achievaoie goai w i n low : ~ d s of sorbent addition for low-sulfur sroker coals.
.-is -AX.. h e Hazard $4 senes. NO. emissions Ivere iairi]: iow in most tests, ger,erdv in the 0.2-0.6 Y 3 L M t u (as SO4 range, ma within [ne rarget god or‘ 0.6 lb SO,%hMBtu for reuont comoustors, >ut s ~ g x i y higher man the 0.2 !b SO&lMBtu goai sougnt iix ~iew cornouston. The exception was . i p _.- . SI--- ---,onras ~3 briauene (0.23 1DIMMBtu 3s SO4 and was probably due to the low iuei-niuogen
Tie ccst signu'icant result of this WAF test senes was h e dramanc mprovemenr in combusnon Aiix-tics of the briquetted codsorxnr ?As x!ative to ik parent stoker coais. These ;owsacent fueis had much more staole ruei beds and had :ernarkably bener i p t i o n and a m m u o n qualines as judged from vsuiFI o0SeNanons. While me suppression of coal caking xnez:c:s in the presence of sorbent has been repoma in the literature, it was surprising to observe sucn :z:=cts even at dosages as low as 3-5 wt% sorwnt. To lllusuare these observations, severai f i g u s nave been prepared. In each case, the bnquened coal/sorbent fuels seem superior to even the ueri-behaved (in the WAF) Hazard stoker coai reported in Table 3.
.-
F i - m f iIlusnates the suppression in apparent coking tendency w k h sorbent inclusion, as judged by tk clack asNcoke formed as a percent of the fed fuel. Figure 5 illustrates the dramatic rise in the uzli-burned powder fraction o i tFle firebox ash with sorbent inclusion; this is a consequence of Ae s-Lppressed coking tendency coupied with the unproved igmtion rates. In Figures 4 and 5, it I s Lmortant to note that these enhancernenls far s-lrrpass what would be expectea (Le., mass b a l k , from simple physical addition of firieiy divided hydrated lime to the parent coals. This points strongly conveyed in Figure 5, which shows the improved carbon conversions exhibited by the 'kauetted coal/sorbent fueis. Somewhat surprisingiy, the increase in carbon conversion actually cffstzi corbent addition in terms of the net fireoox asn generated ;is shown in Figure 7.
A s &::suit of the improved combustion, the codisorbent briquetxd fuels produced lower emissions of C3 a d THC as shown in Figures 8 and 9, respecuvely. X aerarled study of these emissions is urzder way to develoD projected CO and THC umssions reducnons based on pseudo-steady srate ~ S S U I C D ~ O ~ . Prelimin& estimates reveal that CO ana THC reductions on the order or 3040% are x i i e v k with the briquetted coalhorbent fuels. -. r:rA:s 2-9 appear somewhat contradictory in assessing sorbentless briquened fuei perfomance x k v e :o the stoker coals and briauerted coaYsorbent analogs. The combustion indicators (Figures 4-71 minirively revealed intermediate improvements over the stoker coals. It is not clear whetner ths is 3 particle size effea5n thaa briquened fuels produce more fines upon feeding as shown in F i -m 2. or a binder effect. In Figures 8 and 9, nowever, CO and THC emissions from the j o r ~ c ~ < s s briquettes were higher than that observed with the parent coals. Visual observations indiczza that more stable fuel beds were obtained with sorbentcontaining briquened fuels.
From idbie 3, it is interesting to note that while siight carbon conversion improvements (24 ;ercmqe points) were observed for the bnquenea coUsorbent fuels in the Hazard series, some of ~e other trends (e.g., reduced coke formation) were absent These subdediscrepancies are p-obwwiy the result of the inherent iow-caking (FSI=i.S) tendency of the Hazard parent stoker coai.
Inre.Fmtion of carbon conversions and other Faramerers in Figures 4-9 should not be used to e x x F i a t e direct parent coai comparisons. There s no issue more divided in the smail-underfeed stoKer :irerature an the "reponed" companbiiitv of c h g coais. >¶any reports note the mihence or zany vanabies S U C ~ as fuei pamcie size aismbunon. heat-release rate, fuei moisnue, and most impcrzxiy, sroKer [feeascrew. reton) dts;gn. For L-~S reaon, iiiterpretations need to be maae ?n 3 gs-zx i context ivithin each parent ccai series. .A. key point here was that. in every case, ;?e
jnac=rzs coaUsorbent f:ei priormance resuitea 2, sim.pier siolier qersnon 111 terms oi fuei ligt- aif ~-.a zed stability: ii snort tyese fueh burned more e m & , ::quti-g less operator amnuon.
The z:jo1tite carbon conversions ior ail bnauetred codsoroent f ~ e i s were greater than 92%, and in CP 95-98% range for most cases. which is exceilent for a smrul-\mderieed stoker, Lager-scaie ana icnger-term tesnng is needed to assess reat-life benefits ror commercii feasibility. In econormc :crms. :ncremenral fuei cost wouid be reasonable 3t about 51.5,'ton ana SO.14NMBt-u for a p-picai S35/to3. 13000 Btujlb coal and 5 wt% hydrated Qrne sddition. The improved combusaon charaaenstics as well as modest emissions reductions in SO,, CO, and THC are worth considering in furi i i fine coal recovery/reconsatuaon schemes with stoker mara t intentions.
WAF zissions testing will be completed for remaining coaVsorcent briquettes. Larger-scaie emissions tests are planned in a commercii 2.6 MMBWhr cast-iron boiler for stoker coais and any prormslng coaVsorbent fuels idennfied in the WAF screening study. X detailed topicai report is being ?repared.
DISCLUMER
Refexnes to any specific commercial product. process. or sewice 1s for understanaing purposes oniy uia does not impiy endorsement or ravonng by the United Sues Department or' Enerq.
REFERESCES
i6 I
3OEPETC Program Research and Development Announcement NO. DE-RA22-88PCS8869, Ynissions Controi for Small-scale Combustors. .4Drii 6. 1988.
:?ivate communications with various US. c o d compames.
'.l.C. Freeman. H.W. Penniine. 3.1. Jouben, md P.A. Vore, "Emissions Screening o i C,xd/Sorknt Fueis." Proceedings oi the 22"" lnsnnrte of Briquetting and Agglomeration Csnference, San Antonio, Texas, November 3-6, 1991 (in press).
-
3.W. Rising, H X . Conkle. W.J Dawson, and R.D. L i z "Advanced Development OK C x i - Limestone Fuel Pellet for hdusmal Boilers," Barteile-Columbus Laboratories, EPA Connacr YO. 68-02-3 189, September. 1983.
.?.E. Dougias, G.E. Wasson. and W.C. Corder, "Stoker Boiler Operations with Briquetted Fines," Proceedings, Current Deveiopmenrs in Solid Fuel Combusaon Syste,?ls.
_':mcil of Indusxiai Boiler Owners Conference, llq 8-9. 19%.
fnissions and Their Control from Indusmal Ir,srailanons L'sine Solid Fueis: Propress ReDorr Yn. 3. JuIv-December. 1990. British Coal Cxporanon. Coal Research Establishment, Gloucestershire. United Kingaom. 1990.
Table 2: Briauetted CoaUSoment Fomuiatlons and Characteristlcs
Composition (w%) taboratow Strength HHV BtuAb Aazard #4 I Coal Starcn Soro I Cornores Drop (6 ft) 1 % of Parent i ‘ NacK 96.0
Hyarz;;ed Lime, WS=3 89.3 Doiomtic LS, Ca/S=3 85.5 c a m LS, ca/s=3 85.5
