United States National Risk Management Environmental Protection Research Laboratory Agency Research Triangle Park NC 27711 Research and Development EPA/600/SR-96/062 June 1996 «>EPA Project Summary Prototype Scale Testing of LIMB Technology for a Pulverized- Coal-Fired Boiler G.C. England, D.K, Moyeda, Q. Nguyen, and B.A. Folsom The report summarizes results of a project conducted to evaluate furnace sorbent injection for control of sulfur dioxide (SOj) emissions from coal-fired utility boilers. The project, sponsored by the U.S. Environmental Protection Agency (EPA) and the Electric Power Research Institute, is one of three sor- bent injection projects conducted on full-scale, coal-fired utility boilers in the U.S. The emphasis of the project was on evaluating a wide range of process parameters during relatively short-term periods of operation to enhance under- standing of the process, as opposed to a demonstration of sorbent injection over an extended period of time under optimum sorbent injection conditions. The process parameters which were evaluated included sorbent type, boiler operating conditions, and injection pa- rameters. A very flexible sorbent Injec- tion system was installed at the host boiler and tested over a wide range of conditions. A flue gas humidification system also was installed and tested to enhance electrostatic precipitator performance during sorbent injection. Energy and Environmental Research Corporation conducted the project. This Project Summary was developed by EPA's National Risk Management Research Laboratory, Research Triangle Park, NC, to announce key findings of the research project that is fully docu- mented in a separate report of the same title (see Project Report ordering infor- mation at back). Program Goals Primary goals of the program were to provide information on parameters that af- fect calcium utilization in tangentially fired boilers and to document the impacts of sorbent injection on the boiler, ancillary equipment, and pollutant emissions. Sec- ondary goals were to document humidifi- cation system performance and to establish an information base for future combined sorbent injection/humidification designs that are optimized for S02 re- moval. Host Boiler The host boiler for the program was Richmond Power and Light's Whitewater Valley Unit 2 in Richmond, Indiana. This unit, one of the smallest existing tangen- tially fired utility boilers, is a Combustion Engineering (CE) VU-40 steam generator with a nominal rating of 61 MWe. Pendant and spaced superheaters in combination with a baffleless boiler bank result in a single-pass cross-flow arrangement of heat transfer surfaces. It has a finned tube economizer and a rotating regenerative air heater. At full load, 540,000 Ib/hr* of steam at 1320 psig and 955° F is gener- ated. Unit 2 was commissioned in 1972 and is a balanced draft design. Three elevations of burners are located at the furnace corners (12 burners total). The burners were previously modified by CE with a low nitrogen oxide (NO„ concentric firing system. The plant fires a medium sulfur (2.5 lb sulfur per 10® Btu) bitumi- nous coal blend. Unit 2 is equipped with a Lodge-Cottrell cold-side electrostatic precipitator (ESP), erected with the boiler. The ESP treats 227,000 acfm of flue gas at 285°F. The design specific collection area of the ESP is 198 ft2/1000 acfm. It has two mechani- * Readers more familiar with metric units may use the factors listed at the end of this Summary to convert to that system. ------- cal fields, four equal electrical fields in series, and a design particulate removal efficiency of 99.0%. Baseline Tests Baseline tests were performed on the host boiler to establish normal performance prior to operation of the prototype furnace sorbent injection and flue gas humidifica- tion system. The test results were used to establish the impact of sorbent injection/ humidification on boiler and ESP perfor- mance over a wide range of operating conditions. The baseline tests included characterization of air pollutant emissions; boiler thermal performance, including slagging and fouling tendencies; ESP op- eration and performance; and furnace flow and temperature characteristics. Baseline Emissions Average emissions of NOx, S02, and particulate were determined on the host boiler over a range of boiler loads and coal sulfur contents. NOx emissions ranged from 0.65 lb/106 Btu at full boiler load to 0.53 lb/10s Btu at low load. The baseline NOx levels are significantly below the un- controlled emissions from the host boiler prior to the low NOx concentric firing sys- tem retrofit. During the baseline tests, the coal sulfur content varied from 1.7 to 3.1% on an as-received basis. S02 emissions were found to correlate well with coal sul- fur content. Typical of bituminous coals, essentially no inherent sulfur capture in the coal ash was detected. Particulate emissions were 0.04 to 0.13 lb/106 Btu over the range of conditions tested. Baseline Boiler Efficiency An on-line boiler performance monitor- ing system developed by Energy and En- vironmental Research Corporation (EER) was used to determine boiler efficiency, which ranged from 88.1% at the nominal low load condition (40 MWe) to approxi- mately 87.7% at maximum boiler load (64 MWe). The measured efficiency was ap- proximately equal to the manufacturer's predicted design efficiency. Baseline Gas-Side Fouling Fouling of gas-side tube surfaces in the superheaters and convective passes is expected to increase during sorbent injec- tion because the spent sorbent will result in a large increase in the total amount of particulate matter entrained in the gas. Baseline fouling characteristics were de- termined so that the impact of sorbent injection could be quantified. A surface "cleanliness coefficient" was defined for each boiler section as the ratio of actual heat absorbed to average heat absorbed at a given boiler load. This was monitored on-line during operation with the boiler performance monitoring system. The stan- dard deviation of cleanliness coefficient was used to represent the range of baseline gas-side fouling. Measurements were also made using an air-cooled foul- ing probe to quantify the rate of deposit buildup in the upper furnace and super- heater sections. Baseline ESP Performance Baseline performance of the ESP was determined while firing the baseline and high sulfur coals. The baseline coal had a sulfur content of 2.0 to 3.0%, while the high sulfur had between 3.5 and 5.3% sulfur. The collection efficiency of the ESP was slightly below the design value, even when firing the high sulfur coal. However, stack opacity remained well below regu- lated levels. Fly ash resistivity was normal for the coal type and flue gas conditions occurring during the tests. No significant difference in fly ash resistivity between baseline and high sulfur coals was ob- served. Process Design Measurements Flow and temperature fields were ex- tensively characterized to provide infor- mation needed to support the detailed design of the sorbent injection system and flue gas humidifier. Tests included mea- surement of gas temperature, velocity, gas species concentration, and radiative flux at locations in the furnace and superheat- ers. Additionally, physical measurements were taken of ductwork and areas of in- terference to accommodate the process equipment that would be installed during the sorbent injection/humidification system retrofit. Baseline Test Conclusions The baseline tests indicated that the host boiler and ESP for the sorbent injec- tion/humidification prototype were perform- ing essentially as designed, within normal limits. Relatively minor problem areas (ex- cessive air heater leakage, air heater bas- ket corrosion, missing tube shields, ESP plate misalignment) were corrected during the tests to ensure that sorbent injection/ humidification was evaluated under condi- tions representative of normal boiler ther- mal operation. The tests provided baseline data for air pollutant emissions, boiler ther- mal performance, fouling trends, and per- formance of the ESP, fans, and coal pulverizers. Additionally, historical data pro- vided by the plant were used to establish baseline reliability and availability of the unit. Prototype System Design The sorbent injection and humidification systems were designed in two steps: • Process design studies were con- ducted to establish design specifi- cations expected to result in optimum system performance on the host boiler and to develop boiler specific conceptual designs. • Engineering design studies were conducted to translate the concep- tual design into engineering designs that included specification of equip- ment, controls and instrumentation, and process equipment arrange- ment. Sorbent Injection System The approach to the design of the sor- bent injection system was based on a generalized methodology developed by EER, which involves (1) the application of various experimental and analytical meth- ods to determine boiler locations at which temperature levels are optimum for sor- bent injection, and (2) to define injection conditions that will produce uniform dis- persion of the sorbent material. The opti- mum injection configuration developed for full boiler load operation consisted of eight 3-in. diameter injectors located near the plane of the boiler nose with an injection velocity of 140 ft/sec. Sorbent transport air corresponded to 2.5% of total combus- tion air. The optimum injection configura- tion for low boiler load operation consisted of four 2-in. diameter injectors with an injection velocity of 326 ft/sec. Sorbent transport air corresponded to 2.0% of to- tal combustion air. Sorbent Selection Bench-scale testing of 15 candidate sor- bent materials was conducted to establish their relative reactivity towards S02. Four sorbents representing a range of avail- able materials were selected for use in the field evaluation studies. These included highly and moderately reactive calcific hy- drates, a highly reactive atmospheric do- lomitic hydrate, and a limestone of relatively low reactivity characteristic of this sorbent class. Humidification System Sorbent injection was expected to in- crease particulate emissions due to (1) an increase in total particulate matter enter- ing the ESP by a factor of 2 to 3, (2) a decrease in the mean particle size at the inlet, and (3) an increase in dust resistivity that would degrade ESP performance. The methodology for designing the humidifica- tion system included experimental and 2 ------- computational efforts. Twin-fluid atomiz- ers were selected and installed as an ar- ray of 28 atomizers in the existing flue gas duct close to the air heater. Com- pressed air was used as the atomizing medium. Spacing was chosen to minimize overlap of adjacent sprays and impaction on the duct walls while achieving rapid and uniform mixing of the water droplets and flue gas. The design was capable of achieving complete evaporation of water down to a 75°F approach to adiabatic saturation temperature. Waste Management The product of S02 removal by sorbent injection is solid calcium compounds, pri- marily unreacted calcium oxide and the reaction product, calcium sulfate. These solids are removed from the flue gas, with the fly ash creating a solid waste stream. Modifications to normal fly ash handling and disposal practices necessary to ac- commodate this additional solid waste bur- den included • Removal of the ash mixture through the dustless unloader into a 20-ton dump truck; • Segregation of spent sorbent/fly ash mixtures from conventional fly ash and disposal at the landfill; and • Treatment of the ash system sluice/ Hydroveyor water with sulfuric acid to neutralize pH. Sorbent Injection Results The test program was limited to para- metric tests of short duration that focused on establishing data quality, sorbent utili- zation data, and the effects of injection parameters and boiler operating conditions on S02 removal. Three sorbents were se- lected for evaluation during the paramet- ric tests program: Marblehead hydrate, Linwood hydrate, and Marblehead hydrate containing a calcium lignosulfonate addi- tive. Sorbent injection parameters investi- gated included • Injection location (upper and lower furnace), • Number of sorbent injectors, • Sorbent jet velocity at the injector exit, • Sorbent nozzle tilt and yaw, and • Sorbent type. S02 Removal Marblehead hydrated lime was used as the baseline sorbent during the paramet- ric tests. S02 removal ranged from 23 to 48% at a Ca/S (molar) ratio of 2.0, de- pending on injection configuration and boiler load. S02 removal of 50% was achieved at a Ca/S of 2.5 with the boiler operating at 50 MWe (82% of maximum continuous rating). At near full load (60 MWe), a Ca/S of 3.0 or greater was re- quired to achieve 50% S02 removal. In addition to load, other boiler operat- ing parameters, (e.g., excess oxygen (02) and burner adjustments) were varied. S02 removal generally increased with increas- ing excess 02 especially near full load. Increasing furnace exit gas temperature caused S02 removal to decrease, sug- gesting that the gas temperature at the point of sorbent injection in the upper fur- nace was higher than optimum, particu- larly at full load. The capability of favorably modifying the burner air distribution to af- fect the upper furnace thermal environ- ment was also evaluated. At full load, the impact of burner adjustments on S02 re- moval was estimated to be 2 to 4%. In general, sorbent injection parameters such as injection velocity and nozzle tilt/ yaw had only a second order impact on S02 removal. Calcium utilization for the three sorbents was evaluated at Ca/S of 1.7 and 2.3 and 60 MWe boiler load. Ironi- cally, the highest calcium utilization was not achieved at the design configuration, but with sorbent injection higher up in the furnace and with higher injection velocity. Injection velocity was increased by increas- ing the sorbent transport air flow with con- stant injector diameter. Attempts to further optimize sorbent injection were limited by the available furnace wall penetrations. Shifting more of the sorbent above the nose plane of the superheater region re- sulted in lower calcium utilization. This was likely due to a combination of less than optimum temperatures and reduced residence time. Calcium utilization was similar for all three sorbents. ESP and Boiler Impacts ESP performance was substantially de- graded by sorbent injection. However, hu- midification of the flue gases to moderate levels (approximately 130°F approach to adiabatic saturation temperature) restored performance. In most tests, opacity re- mained near baseline levels and was be- low 20% for nearly all the tests. Boiler performance was carefully moni- tored throughout the tests, using the on- line boiler performance monitoring system. Sorbent injection resulted in significant superheater fouling, but this was easily controllable by operating the six retract- able sootblowers in the superheater and boiler banks on a 1-hour cycle. There was no impact on furnace fouling or slagging. The host boiler is equipped with a finned- tube economizer and a rotary regenera- tive air heater. There are no sootblowers in the economizer, and decreased heat absorption was observed in the econo- mizer during sorbent injection, although unacceptable buildup of deposits did not occur. Air heater performance was also satisfactory, and there was no evidence of permanent fouling. Heat loss efficiency was slightly lower with sorbent injection, primarily due to the increase in deposits which resulted in increased stack gas tem- perature. Discussion S02 removal for the nominal design con- ditions was considerably below the level predicted during the design phase. A con- tributing factor is thought to be boiler clean- liness and its impact on the furnace gas temperature at the injection locations. Tem- perature and velocity probing of the test unit disclosed a very complex flow tem- perature and flow field. Significant gradi- ents were observed. These factors were considered in the design phase and were accommodated to the extent possible in the actual installation of injectors subject to physical limitations and interferences. Sorbent injection clearly tended to increase gas temperatures at the injection plane, which contributed to a large part of the difference between predicted and mea- sured results. However, it is not sufficient to explain the entire difference. Tests were conducted to estimate the degree of sor- bent dispersion. The measurements re- vealed considerable variation in the local Ca/S ratio, indicating non-uniform sorbent dispersion, which in turn contributes to a decrease in S02 removal. The analysis suggests that the complex flow and tem- perature field can potentially contribute to low S02 removal. Conclusions The efficiency of sorbent utilization in small tangentially fired boilers may be af- fected by complex flow and temporal char- acteristics, requiring higher rates of sorbent injection to achieve S02 control targets than less complex systems. The maxi- mum rate of sorbent practically achiev- able on a long-term basis will be boiler specific and most probably limited by foul- ing tendencies. In applications of sorbent injection to tangentially fired boilers where substantial temperature and velocity gra- dients may exist in the upper furnace, the design of the sorbent injection system must recognize the potential influence of local conditions in order to maximize sorbent utilization. In addition, sorbent injection it- self may have an impact on gas tempera- tures at the injection location, which may affect S02 removal if the impact is large. 3 ------- Metric Conversions Readers more familiar with metric units may use the factors in Table 1 to convert to that system. Table 1. Metric Conversion Factors Nonmetric Multiplied by Yields Metric acfm 4.719 x 10* m>s Btu 1.055 x 10' J "F S/S(°F-32) °c ft 3.048 X 10 ¦' m ff 9.290 x 10* m* fP/1000 acfm 1.968 X 10* rrf/1000 mVs in. 2.540X10* m to 4.536 X 10 ¦' kg psig 6895 (psig + 14.7) Pa (absolute) ton 9.072x 10' kg G.C. England, O.K. Moyeda, Q. Nguyen, and B.A. Folsom are with Energy and Environmental Research Corp., Irvine, CA 92718. David G. Lachapelle is the EPA Project Officer (see below). The complete report, entitled "Prototype Scale Testing of LIMB Technology for a Puh/erized-Coal-Fired Boiler,' (Order No. PB96-1838S0; Cost:$21.50, subject to change) will be available only from National Technical Information Service 5285 Port Royal Road Springfield, VA22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at National Risk Management Research Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 United States Environmental Protection Agency National Risk Management Research Laboratory (G-72) Cincinnati, OH 45268 Official Business Penalty for Private Use $300 EPA/60CVSR-96/062 BULK RATE POSTAGE & FEES PAID EPA PERMIT No. G-35 ------- |