PollutionEngineering June 2012 : Page 34
Sodium Injection for MATS and Boiler MACT Using sodium dry sorbent injection technology to control HCl limits can achieve compliance with MATS and Boiler MACT regulations. By YOUGEN KONG and MICHAEL WOOD R ecent pilot-plant tests con-limit (0.002 lb/mmBTU). Either sorbent ducted at an independent was able to meet the required emissions site incontrovertibly prove levels. Depending on the particular appli-that dry sorbent injection cation, one may work more effectively (DSI) of sodium bicarbon-than the other, but now it is clear that ate or trona can achieve greater than 99 sodium DSI is fully capable of attaining percent removal of HCl and is able to compliance. The new MATS, as well as the Industrial meet the HCl limit in the Mercury and Air Toxics Standards (MATS). The tests Boiler MACT rules, focus on removal also effectively demonstrate the selectivity of HCl in the presence of SO 2 . Well-of sodium sorbents to remove HCl in a documented testing demonstrates that medium to high sulfur environment. trona injection is able to remove SO 3 , The lack of regulations on HCl has which is often a requirement for mercury meant little published data has been avail-reduction, also a part of MATS. Cross able on HCl removal ability and effi-State Air Pollution Rule (CSAPR) contin-ciency when using DSI in utility boilers. ues to focus on SO 2 reduction where the Despite the long-standing demonstrated capabilities of sodium DSI have long been performance of sodium DSI in a high-established. Sodium DSI can remove HCl chloride environment, such as at medi-and, if needed, SO 2 and SO 3 , to meet all cal or municipal waste incinerator sites, rules’ limits. This is good news for those concerns had existed regarding its ability needing to achieve compliance with the to mitigate HCl to the levels required in latest standards. a coal-fired boiler, especially when sub-In a sodium DSI system, a sorbent is stantial quantities of SO 2 are present. injected directly into the hot flue gas Therefore, pilot boiler tests were per-formed at an independent site (the EERC facility at the University of North Dakota) to develop the knowledge to answer these concerns. Utilizing coal that is typically employed in utility boilers from the Central Appalachian Basin (CAPP), tests were conducted with sodium bicarbonate and trona using an electrostatic precipitator (ESP) and a baghouse. The tests dem-onstrated that DSI with trona or sodium bicarbonate is able to meet the MATS HCl Figure 1: Pilot Test Setup 34 Pollution Engineering JUNE 2012 duct where it rapidly reacts with acid gases such as HCl, SO 2 , SO 3 and HF, and in some cases even reduces NO X . Due to its low capital cost and ease of operation, dry injection of sodium sorbents is being used at many locations. Such systems are used in glass manufacturing, chemical processing, coal-fired utility, pulp and paper, and many more industries. In fact, sodium DSI is considered BACT in Europe. While regulations may change, the per-formance of sodium products in a DSI system continues to remain steady and proven. Test setup A pilot-scale pulverized coal combustor with an approximate firing output of 550,000-Btu/hr was used for the test. A schematic of the test setup is shown in Figure 1 . The entire system is designed to generate flue gas and fly ash repre-sentative of that produced in a full-scale
Sodium Injection For MATS And Boiler MACT
Yougen Kong And Michael Wood
Using sodium dry sorbent injection technology to control Hcl limits can achieve compliance with MATS and Boiler MACT regulations.<br /> <br /> Recent pilot-plant tests conducted at an independent site incontrovertibly prove that dry sorbent injection (DSI) of sodium bicarbonate or trona can achieve greater than 99 percent removal of Hcl and is able to meet the Hcl limit in the Mercury and Air Toxics Standards (MATS). The tests also effectively demonstrate the selectivity of sodium sorbents to remove Hcl in a medium to high sulfur environment.<br /> <br /> The lack of regulations on Hcl has meant little published data has been available on Hcl removal ability and efficiency when using DSI in utility boilers.Despite the long-standing demonstrated performance of sodium DSI in a high chloride environment, such as at medical or municipal waste incinerator sites, concerns had existed regarding its ability to mitigate Hcl to the levels required in a coal-fired boiler, especially when substantial quantities of SO2 are present.Therefore, pilot boiler tests were performed at an independent site (the EERC facility at the University of North Dakota) to develop the knowledge to answer these concerns.<br /> <br /> Utilizing coal that is typically employed in utility boilers from the Central Appalachian Basin (CAPP), tests were conducted with sodium bicarbonate and trona using an electrostatic precipitator (ESP) and a baghouse. The tests demonstrated that DSI with trona or sodium bicarbonate is able to meet the MATS Hcl Limit (0.002 lb/mmBTU). Either sorbent was able to meet the required emissions levels. Depending on the particular application, one may work more effectively than the other, but now it is clear that sodium DSI is fully capable of attaining compliance.<br /> <br /> The new MATS, as well as the Industrial Boiler MACT rules, focus on removal of Hcl in the presence of SO2. Well documented testing demonstrates that trona injection is able to remove SO3, which is often a requirement for mercury reduction, also a part of MATS. Cross State Air Pollution Rule (CSAPR) continues to focus on SO2 reduction where the capabilities of sodium DSI have long been established. Sodium DSI can remove Hcl and, if needed, SO2 and SO3, to meet all rules’ limits. This is good news for those needing to achieve compliance with the latest standards.<br /> <br /> In a sodium DSI system, a sorbent is injected directly into the hot flue gas Duct where it rapidly reacts with acid gases such as Hcl, SO2, SO3 and HF, and in some cases even reduces NOX. Due to its low capital cost and ease of operation, dry injection of sodium sorbents is being used at many locations. Such systems are used in glass manufacturing, chemical processing, coal-fired utility, pulp and paper, and many more industries. In fact, sodium DSI is considered BACT in Europe.<br /> <br /> While regulations may change, the performance of sodium products in a DSI system continues to remain steady and proven.<br /> <br /> Test setup <br /> <br /> A pilot-scale pulverized coal combustor with an approximate firing output of 550,000-Btu/hr was used for the test. A schematic of the test setup is shown in Figure 1. The entire system is designed to generate flue gas and fly ash representative of that produced in a full-scale Utility boiler. The time/temperature profile of the pilot unit is typical of that of a full-scale system. To simulate real-world applications, both an ESP and a fabric filter or baghouse (FF) were tested during these sorbent injection tests.<br /> <br /> For all tests, the flue gas from the combustor was routed through a series of heat exchangers followed by a particulate control device. The single-wire tubular ESP was maintained at 325°F, and the actual flow rate of flue gas through the ESP was about 196 acfm, providing about2. 4 seconds of residence time. The FF vessel was a 20 inch ID chamber equipped with three FF bags (new Ryton-type bags, 13 feet in length and 5 inches in diameter).The FF was operated at 325°F with a nominal 3.7 ft/min air-to-cloth ratio.Each bag was periodically cleaned separately with its own diaphragm pulse valve.<br /> <br /> The analyses of the CAPP coal used for the test included proximate, ultimate and chlorine content. The analyses are listed in Table 1. The average chlorine concentration of the tested coal was 962 ppm (average, dry basis). Based on the proximate and ultimate analysis data obtained and the chlorine content in the coal, a theoretical combustion calculation indicates that approximately 0.106 g/dNm3 of total Hcl was expected in the gas stream at a 3 percent oxygen level.<br /> <br /> The continuous Hcl monitor detected 0.10 to 0. 121 g/dNm3 Hcl (3 percent O2) in the baseline flue gas, which is within 5. 7 to 14.2 percent of the theoretical value.<br /> <br /> The sorbents selected were trona commercially available for air pollution control (30 ìm) and that same trona milled (15 ìm); two milled sodium bicarbonate, commercially available milled product for air pollution control at 40 ìm; a finely milled sodium bicarbonate at 17 ìm; and commercially available hydrated lime used for air pollution control.The sodium sorbents were chosen to represent actual plant experience. Many customers do not mill trona and use it as supplied (30 ìm); some customers use pin mills to further reduce the size of the trona to enhance sorbent utilization (15 ìm). Sodium bicarbonate must always be milled whether it is milled by the supplier (40 ìm) or milled on-site to a finer particle size (17 ìm). The sorbents were evaluated in the ESP and the FF configuration. <br /> <br /> Table 2 summarizes sorbent details used for the test.<br /> <br /> For the ESP configuration, the sorbent was injected into the flue gas upstream Of the ESP where the flue gas temperature was 650°F. During the ESP testing, the precipitator was cleaned out off-line between each sorbent change to recover baseline. For the FF testing, sorbents were injected at the inlet of the FF where the temperature was at 325°F. The FF bags were pulsed between each feed rate change as well as for each sorbent to recover baseline and minimize residual effect.<br /> <br /> The unit started with an 8-hour heatup period on gas and then with steady state operation firing coal. The coal combustion baseline condition was established for several hours, and EPA Method 26 (M26) measurement of Hcl was taken during this baseline test to verify the continuous Hcl analyzer measurement. The ESP ash was removed prior to the first sorbent injection.<br /> <br /> Composite coal samples were collected during the course of testing by collecting a portion of the pulverized coal each time the coal feed hopper was refilled. The collected coal sample was then analyzed. It should be noted that the moisture content for samples of the pulverized coal feed was on an as-fired basis since some drying of the coal occurred during coal preparation and pulverization.<br /> <br /> Routine flue gas samples were taken and analyzed at two locations: the outlet of the combustor and the outlet of the PM control device (ESP or FF). After the flue gas passed through sample conditioners to remove the moisture, a pair of multi component continuous gas analyzers were used for analysis of O2, CO and CO2. NOX was determined using a pair of NOX chemiluminescent analyzers. SO2 was measured using a pair of photometric gas analyzers. Each of these analyzers was regularly calibrated and maintained to provide accurate flue gas concentration measurements.<br /> <br /> Flue gas Hcl concentrations at the outlet of the PM control device (ESP or FF) were measured using an Hcl analyzer that utilizes a gas filter correlation nondispersive spectrometer to measure the specific Hcl infrared absorption at 3.4 ìm and has served as the primary instrument for Hcl measurement in flue gas. The analyzer was calibrated at 5 and 90 ppmv on a daily basis. Meanwhile, measurement validation was performed using EPA M26 and compared with the estimated chlorine concentration in the flue gas based on chlorine content in the coal.<br /> <br /> Acid gas reductions are calculated according to Equation 1, which uses the pre-injection ESP/FF outlet acid gas concentration as the baseline value. Acid gas reductions calculated In this way then represent acid gas capture attributable solely to sorbent injection.<br /> <br /> Also, by referencing the pretest conditions, the inherent coal variability was compensated for. Averaged values of Hcl, SO2, and NOX concentrations were computed And used for the acid gas reduction calculations.<br /> <br /> The injection rate of sorbent is represented with total normalized stoichiometric ratio (NSRT) in order to compare performance of different sorbents.NSRT, as shown in Equation 2, is defined as dividing the actual sorbent injection rate by stoichiometric required sorbent to neutralize all acid gases in flue gas.<br /> <br /> Where F is the actual sorbent injection rate; Xi is the actual acid gas flow rate in flue gas; ..i is the stoichiometric required sorbent to neutralize acid gas Xi.<br /> <br /> Results <br /> <br /> EPA M26 samplings were performed at various times throughout the ESP test series to validate the measurement by the Hcl analyzer. For each instance, the results between EPA M26 and the Hcl analyzer Were within 2 to 4 ppm. EPA M26 samples were not taken during the FF test series because of the close agreement of the verification during ESP operation.<br /> <br /> Three different sorbent federates were tested for each sorbent by estimating the amount needed to achieve Hcl removal rates of 80 percent, 90 percent and 99 percent. As shown in Figure 2, the Hcl was easily mitigated below the MATS limit of 0. 002 lb/mmBTU with all sodium sorbents including un-milled trona. The results confirmed that smaller sorbent particle size enhances the performance and in the presence of SO2 there is no issue with meeting the target emissions levels.<br /> <br /> As expected, the injected sodium sorbent also reacted with SO2, but their reaction rate with SO2 is slower than with Hcl.Sodium bicarbonate is more reactive with SO2 than trona (Figure 3).<br /> <br /> Because longer residence time in the baghouse enables more sorbent to be consumed by slower-reacting SO2, the Hcl mitigation performance estimates were not as good in the FF case as in the case of ESP (Figure 4).Neither the un-milled trona nor the coarser sodium bicarbonate brought the Hcl concentrations down to the MATS limit at the quantities injected simply due to a miscalculation of the amount needed. However, the trends clearly show that addition of just slightly more sorbents would mitigate the Hcl below the limit. On the other hand in these tests, hydrated lime was not able to meet the limit, even at a high NSR number.<br /> <br /> Due to long residence time on the FF, baghouse configuration results in more mitigation for slower-reacting SO2, as shown in Figure 5.<br /> <br /> The reaction rate of sodium sorbents was much higher with Hcl than for SO2 in the cases of both ESP and Baghouse configurations, as shown in Figure 6 and 7.<br /> <br /> Conclusion <br /> <br /> The high removal efficiencies of SO2, SO3, Hcl and HF with trona and sodium bicarbonate have been demonstrated at many power plants and other industrial users for over the last 20 years. Its low capital cost makes DSI more attractive in today’s difficult economic environment.