United States Environmental Protection Agency Of flee of Solid Waste Off Ice of Air Off Ice of Research EPA/530-Sw-87.02ici and Emergency Response and Radiation and Development June 1»7 Washington, DC 20460 Washington, DC 20460 Washington, DC 20460 &EPA Municipal Waste Combustion Study Flue Gas Cleaning Technology U.S. Environmental Protection Ag--:.'-•£ region 5, Library (j.L-16) 230 S. Dearborn 3U-r=.r., Room 167Q Chicago, IL 6i'G")4 ------- EPA/530-SW-87-021d June 1987 MUNICIPAL WASTE COMBUSTION STUDY: FLUE GAS CLEANING TECHNOLOGY Prepared by: Charles B. Sedman and Theodore G. Brna Air and Energy Engineering Laboratory U.S. Environmental Protection Agency Research Triangle Park, North Carolina 27711 Prepared for: U.S. Environmental Protection Agency Office of Solid Waste Washington. DC 20460 ------- EPA REVIEW NOTICE This document has been approved for publication by the Office of Solid Waste, U.S. Environmental Protection Agency. Approval does not signify that the contents necessarily reflect the view and policies of the Envir- onmental Protection Agency, nor does the mention of trade names or com- mercial products constitute endorsement or recommendation for use. ------- TABLE OF CONTENTS EPA Review Notice ii List of Figures iv List of Tables v Acknowl edgements vi i CHAPTER 1 INTRODUCTION 1-1 CHAPTER 2 PARTICIPATE MATTER CONTROL 2-1 2.1 Electrostatic Precipitators 2-1 2.1.1 Principles of ESP Design 2-1 2.1.2 Factors Affecting ESP Performance 2-3 2.1.3 Design Considerations 2-6 2.1.4 Performance of ESPs 2-7 2.2 Fabric Filters 2-7 2.2.1 Theory and Principles of Filtration 2-7 2.2.2 Gas Stream Factors that Affect Fabric Filter Design and Operation 2-12 2.2.3 Fabric Filter Systems 2-14 2.3 Wet Scrubbers 2-19 CHAPTER 3 GASEOUS EMISSION CONTROLS 3-1 3.1 Acid Gas Scrubbers (HC1 and HF only) 3-1 3.2 Post Combustion NOX Control 3-1 3.2.1 Selective Catalytic Reduction (SCR) 3-1 3.2.2 Wet NOX Removal Processes 3-6 CHAPTER 4 MULTIPOLLUTANT CONTROL SYSTEMS 4-1 4.1 Wet Scrubbing 4-1 4.1.1 Wet Scrubbing for Acid Gases (HC1, HF, and S02) 4-1 4.1.2 Wet Scrubbing for Acid Gases and NOX 4-3 4.1.3 Performance of Wet Scrubbers 4-4 4.2 Dry and Semi-Dry Scrubbing 4-5 4.2.1 Dry Injection Processes 4-5 4.2.2 Semi-Dry Scrubbing 4-13 4.3 Combination Scrubbers 4-16 4.3.1 Semi-Dry/Dry Scrubbing 4-16 4.3.2 Semi-Dry/Wet Scrubbing 4-20 CHAPTER 5 EFFECTIVENESS OF FLUE GAS CLEANING METHODS 5-1 5.1 Particulate Matter Control 5-1 5.2 Acid Gas Control 5-1 5.3 Post Combustion NOX Control 5-3 5.4 Post Combustion Organic Pollutant Control 5-3 5.5 Heavy Metals Control 5-4 CHAPTER 6 OPERATION AND MAINTENANCE OF FLUE GAS CLEANING SYSTEM 6-1 6.1 Electrostatic Precipitators 6-1 6.2 Fabric Filters 6-5 6.3 Scrubbers 6-5 6.3.1 Lime/Limestone Wet Scrubbing 6-10 6.3.2 Semi-Dry Scrubbing 6-15 REFERENCES m ------- LIST OF FIGURES Figure 2-1 2-2 2-3 2-4 2-5 2-6 2-7 3-1 3-2 4-1 4-2 4-3 4-4 4-5 4-6 5-1 Principles of the electrostatic precipitation process Diagram of a two-field, weighted-wire ESP PM emissions vs. SCA for best-fit equations Municipal waste incinerator equipped with dry flue gas Small shaker-type fabric filter Example of a large reverse-air fabric filter Example of a small pulse-jet fabric filter SCR options for municipal incinerators Circulating fluid bed absorption (dry) process Spray absorption (semi-dry) process Semi -dry/dry scrubber Saturation points of metal and metal compounds Page 2-2 2-8 2-9 2-11 2-15 2-16 2-18 3-2 3-5 4-2 4-6 4-12 4-12 4-17 4-21 5-6 1v ------- LIST OF TABLES Table Page 2-1 Partial list of MWC/ESP applications 2-3 2-2 Plant test data 2-10 3-1 SCR plant designs for sludge Incinerator flue gas 3-4 4-1 Wet scrubber outlet emissions 4-4 4-2 Inlet pollutant concentration and pollutant control with a lime dry injection process 4-7 4-3 Summary of key operating parameters 4-8 4-4 PCDD concentrations (ng/Nm3 @ 82 02) in flue gas and efficiency of removal 4-9 4-5 PCDF concentrations (ng/Nm3 @ 82 02) in flue gas and efficiency of removal 4-9 4-6 Percent removal of other organics 4-10 4-7 Inlet/outlet metal concentrations (yg/Nm3 @ 82 02) 4-10 4-8 Hydrogen chloride concentrations (@ 82 02) and removal efficiencies 4-11 4-9 Sulfur dioxide concentrations (@ 82 02) and removal efficiencies 4-11 4-10 Outlet emission guarantees for municipal waste combustors 4-14 4-11 Acid gas emissions from municipal waste combustor 4-15 4-12 Trace heavy metals control by pilot plant FGC systems... 4-15 4-13 Pollutant emissions and control by semi-dry/dry scrubbing 4-16 4-14 Summary of emission results for tests of September - October, 1986 at the Marion County, OR waste-to-energy facility 4-18 4-15 Permitted emissions from the Marion County, OR waste-to- energy facility 4-19 ------- LIST OF TABLES (cont'd) Table 4-16 Expected pollutant emissions from MWC equipped with semi -dry/wet scrubbi ng 4-20 5-1 Effectiveness of acid gas controls (2 Removal) 5-2 5-2 Spray dryer control of selected organic pollutants 5-4 6-1 Inspections for ESP 6-2 6-2 Typical maintenance inspection schedule for a fabric fi 1 ter system 6-6 6-3 Routine inspection data for reverse air fabric filters.. 6-8 6-4 Routine inspection data for pulse-jet fabric filters.... 6-9 vi ------- ACKNOWLEDGEMENTS The authors express their appreciation to the numerous persons who provided Information for this report. Especially noteworthy were the con- tributions of Claus Jorgensen (Acurex Corporation) regarding European technology and test data; James D. Kilgroe and Julian W. Jones (Air and Energy Engineering Research Laboratory) for their constructive reviews of the initial draft report; Abe Flnkelstein and Raymond Klicius (Environment Canada) for NITEP program data and results; James Crowder and Peter Schindler (EPA's Office of A1r Quality and Planning Standards) regarding emissions data and results; and Steve Green (EPA's Office of Solid Waste) and Glynda Wllkins (Radian Corporation) for their assistance in revising and finalizing the report. Many useful comments were received from Industrial, environmental, and other Interested organizations in response to the release of the draft report for public comment. Most comments have been addressed in the revised report as permitted by the data and time available. Although too numerous to mention the respondents here, their comments are gratefully acknowledged and have, hopefully, resulted in a more com- prehensive and soundly based report. vii ------- 1.0 INTRODUCTION This report presents the results of a recent study of flue gas cleaning technology applied to municipal waste combustors. The Information presented here was developed during a comprehensive, Integrated study of municipal waste combustion. An overview of the findings of this study 1s Included In the Report to Congress on Municipal Waste Combustion (EPA/530-SW-87-021a). Other technical volumes Issued as part of the Municipal Waste Combustion Study Include: o Emission Data Base for Municipal Waste Combustors (EPA/530-SW-87-021b) o Combustion Control of Organic Emissions (EPA/530-SW-87-021c) o Cost of Flue Gas Cleaning Devices (EPA/530-SW-87-021e) o Sampling and Analysis of Municipal Waste Combustors (EPA/530-SW-87-021f) o Assessment of Health Risks Associated with Exposure to Municipal Waste Combustion Emissions (EPA/530-SW-87-021g) o Characterization of the Municipal Waste Combustion Industry (EPA/530-SW-87-021h) o Recycling of Solid Waste (EPA/530-SW-87-0211) Until recently post combustion emission control technology was applied to waste incinerators strictly for particulate matter removal. Early install- ations used either electrostatic precipitators (ESPs) or wet scrubbers to meet local and state emission codes. With the advent of 40CFR60 Subpart E, the New Source Performance Standards for municipal incinerators in the early 1970's, the ESP became the dominant choice. This is reflected by recent studies which Indicate that 75 percent of conventional incinerators in the United States use ESPs and 20 percent wet scrubbers.1 Because several foreign governments are regulating pollutants other than particulate matter (notably Japan and some in Western Europe), many advanced concepts for scrubbing acid gases—sulfur dioxide (503), hydro- chloric acid (HC1), and hydrofluoric acid (HF)—have been commercially applied. State and local regulations 1n the U.S. are requiring the control of these and other pollutants in new municipal waste combustion (MWC) facilities and those now being planned. These control systems Include wet, dry, and semi-dry scrubbers and fabric filters (FFs) or ESPs. More recent concern over trace metals emissions, organics such as dioxins and furans, and nitrogen oxides have reinforced the selection of scrubbers as well as post-combustion NOX control. 1-1 ------- The combustion process is important in controlling organics, such as dioxins and furans. Post combustion emission control technology is not a substitute for good combustion, although it may remove some pollutants resulting from poor combustion. Its use is aimed at controlling acid gases and particulate matter to permitted levels, but it may also effect some control of volatile and semi-volatile organics and trace metals while achieving its primary aims. Post combustion emission control technology also offers an alternative for controlling NOX downstream of the combustor or boiler if such control is required. The major focus of this report is on the use of scrubbers to control pollutants in flue gas. Scrubbers have been classified as c ther wet or dry depending on whether the cleaned flue gas leaving the scrubbing system is saturated or not. With this classification, a wet scrubbing system emits a saturated flue gas (generally with water vapor), while a dry scrubber has a gaseous effluent which is unsaturated. In the U.S., S0£ scrubbers are called wet and dry when the solids from the scrubbers are wet and dry, respectively. Scrubber terminology in other countries often uses "semi-dry" or "wet/dry" and "dry" scrubber. The former usually applies to a spray dryer and fabric filter or ESP system in which the sorbent enters the spray dryer as a slurry or solution, and the cleaned flue gas leaves the particu- late (dust) collector unsaturated. A "dry" scrubber, as often used in the same countries, refers to a dry powdered sorbent being injected into-the flue gas duct or a reactor upstream of the particulate collector (either a baghouse or ESP), with an unsaturated, cleaned flue gas leaving this collec- tor. Because of references to scrubbing systems developed or being developed outside the U.S. in this report, semi-dry and wet/dry scrubbing in this report means that a dirty, hot flue gas from a boiler (steam generator) or a following air preheater contacts a wet reagent in a spray dryer, but the gas stream leaves the spray dryer unsaturated before entering the particulate collector of the flue gas cleaning system. It Is also noted that in the U.S. a scrubbing system with either a spray dryer or a dry-injected sorbent into the flue gas upstream of the particulate collector is commonly called a "dry" scrubber. In-furnace NOX control (thermal DeNOx and flue gas recirculation) is not discussed in this report as it 1s more appropriate to a companion report In the Municipal Waste Combustion series. However, post combustion NO control [selective catalytic reduction (SCR) and wet NOX scrubbing] is addressed In Chapter 3 of this report. The following chapters will describe briefly each generic control sys- tem, design and operating considerations, and control effectiveness on selec- ted pollutants. Control systems which reflect the most prevalent current U.S. practice—that of particulate matter control—are discussed initially, followed by gaseous controls, and finally, the more advanced multipollutant control systems. Chapter 5 summarizes the effectiveness of the control systems discussed in the preceding three chapters for particulate matter, selected acid gas, selected organic pollutants, and selected trace heavy metals. Since the data from commercial municipal solid waste incinerators are limited, pilot plant data were also considered in reporting the control effectiveness of some pollutants. The final chapter, Chapter 6, addresses the operation and maintenance of flue gas cleaning systems. 1-2 ------- 2.0 PARTICULATE MATTER CONTROL 2.1 Electrostatic Precipitators3,4 2.1.1 Principles of ESP Design Review of ESP Fundamentals. The basic steps of the electro- static precipitation process are:(IT development of a current of negative molecular ions, from a high voltage corona discharge, that is used to charge dust particles in the gas stream; (2) development of an electric field in the gas space between the high voltage discharge electrode and the collection electrode that propels the negatively charged ions and particulate matter toward the collection electrode; and (3) removal of the collected particulate matter into hoppers by use of a rapping mechanism. The basic principles of electrostatic precipitation are illustrated in Figure 2-1. The electrostatic precipitation process occurs within an enclosed chamber. A high voltage transformer and a rectifier converts the alternating current (a.c.) electrical power input into direct current (d.c.). Suspended within the chamber are the grounded collection electrodes (metal plates) which are connected to the grounded steel framework of the supporting structure. Suspended between the collection plates are the high voltage discharge (wire) electrodes (corona electrodes) which are insulated from ground and negatively charged with voltages ranging from 20 kV to 100 kV. The large difference in voltage between the discharge electrodes and collection electrodes and the interelectrode space charge of ions and charged particles create the electric field that drives the ions and particles toward the collection electrodes. The particles may travel some distance through the ESP before they are collected or they may be collected, reentrained, and collected again. The last step of the process involves the removal of the dust from the collection electrodes. In dry ESPs, this is accomplished by periodic striking of the collection and discharge electrode with a rapping device that is activated by a solenoid, air pressure, or gravity after release of a magnetic field, or through a series of rotating cams, hammers, or vibrators. The dust is collected in hoppers and then conveyed to storage or disposal. In wet ESPs, the collected dust is removed by an intermittent or continuous stream of water that flows down over the collection electrodes and into a receiving sump. History of MHC/ESP Applications. Lurgi installed the first electrostatic precipitator on a refuse incinerator in Zurich in 1927. Three more installations followed in Hamburg in 1930. By 1968, this manufacturer could claim 40 MWC/ESP installations in Europe and Japan, either in operation or under construction. Measured collection efficiencies of seven precipita- tors started up between 1961 and 1967 ranged from 98.9% to 99.92. A recent list of MWC/ESP installations tested between 1970 and 1985 is given in Table 2-1.5 The trend of lower emission levels in the newer installations is evident in this table. The Munich installation 2-1 ------- EIECTIOOE AT nuatirr Wto ocucro rurittE OIJOU*GE »T ncaATin 7™\w ^ \fr^'^v v ITi w \ » • A\ UNCHUPtD HXTIOIS TO COLLCCTOt aiCTXOOE MO rORHIKC OUST UTEJ M104 Figure 2-1. Principles of the electrostatic precipitation process. 2-2 ------- (Munich North, built in 1984) has only a two-field ESP following a cyclone and spray dryer (for acid gas control). Nevertheless, the particulate emissions are limited to 0.0104 gr/dscf by the 99.52 collection efficiency. In addition, this installation achieves average removal efficiencies of 952 for HC1, 762 for S02, and 96.32 to 98.4% for heavy metals in particulate form. TABLE 2-1. PARTIAL LIST OF MWC/ESP APPLICATIONS5 Plant S. W. Brooklyn, NY South Shore, Brooklyn, NY Dade, FL Braintree, MA Montreal, Canada Chicago, IL Washington, DC Harrisburg, PA Quebec City, Canada Nashville, TN Saugus, MA Bamberg, West Germany Munich, West Germany Lausanne, Switzerland Baltimore, MD Westchester, NY Avesta, Sweden Sample/ Report Date 1970-71 1970 1970 1970/78 1970/71 1971/75 1972 1974 1974 1978 1983 1983 1984 1984 1984 1985 1985 Pollution Control Equipment Particulate Emissions gr/dscf ESP ESP ESP ESP ESP ESP ESP ESP ESP ESP ESP Wet Scr/ESP Dry Scr/ESP Elec Scr/ESP 4 Field ESP 3 Field ESP Cond Scr/ESP 0.114/0.146 0.056 0.027 0.108/0.083 0.013/0.08 0.025/0.03 0.0548 0.06 0.095 0.018 0.025 0.009 0.0104 0.014 0.01 0.016 0.0115 2.1.2 Factors Affecting ESP Performance Several important properties of the gas stream and particulate matter determine how well an ESP will collect a given dust at a given dust loading. They include particle size distribution, gas flow rate, and partic- ulate resistivity, which is influenced by the mineral composition and density of the particulate matter and the process temperature. These factors can also affect the corrosiveness of the dust and gas and the ability to remove the dust from the plates and wires. Brief discussions of these properties are given in the following paragraphs. Dust Particle Size Distribution. 'in ESP the performance is very sensi- tive to the distribution of particle sizes in the combustion gas because of variations in the effectiveness of particle charging by molecular ions. Particle charging is least effective in the range from 0.1 to 1.0 um. The particulate penetration through an ESP (loss to the atmosphere) is typically a 2-3 ------- maximum somewhere in this range. On the other hand, the penetration of any particles larger than a few micrometers is due almost entirely to reentrain- ment. Thus, the variation in collection efficiency with particle size is typically much greater for an ESP than for a baghouse. Dust Electrical Resistivity. Electrical current from the high voltage corona discharge in an ESP passes through the dust layer on the collection plate to reach electrical ground. The product of the average plate current density (A/cm2) and the bulk electrical resistivity of the collected dust (ohm-cm) is a measure of the average electric field (V/cm) within the collected dust layer. When that average electric field becomes too large (greater than about 5 to 10 kV/cm), electrical breakdown occurs within voids in the dust layer. That breakdown leads to sparking and/or back corona that place the practical operating limit on the useful electri- cal power input to an ESP. Thus, the value of dust resistivity bears upon the useful electrification of an ESP. The electrification is usually limited below desirable levels when the dust resistivity exceeds about 6 x lO^O ohm-cm. A dust resistivity that is too low is also undesirable. The lower limit is not well defined, but it is generally believed to lie in the range of 10? to 10^ ohm-cm. Low resistivity dust is easily precipitated, but it also dis- charges quickly on the collecting plate. The electrical force holding the collected dust layer (against the scouring effect of the gas stream) dimin- ishes, and reentrainment problems become severe. In precipitators serving municipal waste combustors, reentrainment problems may be associated with carbonaceous particles resulting from incomplete combustion. Gas Flow Parameters. Uniformity of gas velocity and gas tempera- ture is essential to the optimum performance of a precipitator. Large vari- ations in gas temperature over the face of a precipitator can result in large variations in dust resistivity. The electrical operation of one entire pre- cipitator field energized by a high voltage power supply will be limited by that portion of the field where the collected dust has the highest resistivity. The average gas velocity through a precipitator determines the treatment time for charging and collecting dust particles. However, regions of low gas velocity do not compensate for other regions of high gas velocity. Optimum ESP collection efficiency is achieved with uniform gas velocity. The standard deviation of a matrix of gas velocity measurements over the face of a precip- itator divided by the average value usually does not exceed 0.15 in a modern precipitator. Too low a gas velocity leads to dust deposits in the duct upstream of the ESP. Too high a gas velocity leads to scouring of collected ash from the plates of the ESP and high reentrainment. The optimum average gas velocity usually lies in the range of 3 to 6 ft/s. Considering that the ash from municipal waste combustion may have low electrical resistivity, with the accompanying tendency to reentrain particles from the collecting plates, the average gas velocity should lie at the lower end of this range. The total volume gas flow (cfm) through an ESP is related to the designed average gas velocity by the designed face area of the ESP. Furthermore, the total volume gas flow, compared with the size of the ESP, is mathematically related to the collection efficiency of the ESP. The specific collection 2-4 ------- plate area (SCA, defined as the ratio of the total plate area to the volume gas flow, ft^/lOOO cfm) appears in the exponent of the Deutsch equation which gives the theoretical collection efficiency for dust particles of a specific size. Combustion Parameters. Stable and optimized ESP performance depends on stable and complete fuel combustion. However, control of the combustion process 1s particularly difficult with municipal waste as the fuel. The fuel is usually wet. Depending on the size and time of year, moisture content can vary from 20 to 60 percent by weight. Food wastes, grass and tree clippings, and other wet wastes add to the moisture content of the material. Large quantities of paper products and plastics help off- set this problem somewhat. The effect of a high moisture content is to lower the heating value of the material charged. At a fixed charging rate, this in turn lowers the overall heat release rate and the combustion temper- ature in the furnace. Another problem is the substantial quantity of noncombustible materials that may be charged to the incinerator. These materials include cans, other metallic materials, rocks, dirt, etc. Although paper products and plastics tend to have low ash contents, the overall "ash" content of the charged material is typically 20 to 30 percent by weight. This high ash content also contributes to a lower heating value. Considering noncombustible and moisture effects, heating values varying from 3000 to 6000 BTU/lb are not unusual. Another problem is the variation in size of fuel. Good combustion relies on even distribution of properly sized material on the grate. The underfire air will follow the path of least resistance. If the distribution of material on the grate is uneven, the gas will channel through the path of least resistance and the remainder of the fuel bed will be "starved" for air. In some instances the time/temperature relationship in the furnace may not be sufficient to produce complete combustion. Incomplete combustion can result in ash that is high in carbon content and is easily reentrained from the ESP collection plates. Incomplete combustion can also result in con- densed fumes of submicrometer particles that are more difficult to collect and more likely to create undesirable visible emissions. Finally, the excess air in municipal waste combustion affects ESP per- formance. Increasing the amount of air in excess of stoichiometric require- ments generally Improves combustion and raises combustion temperatures. At some point, however, no real increase in combustion efficiency (or conver- sion of hydrocarbons to C02 and H20) results, and increasing the quantity of combustion air actually decreases the flame temperature (because both the combustion products and the excess air must be heated). This can also increase the flue gas volume leaving the incinerator. Even after the gases are cooled in an evaporative spray chamber or passed through a heat exchanger, the result is a higher gas volume flow to the ESP, which reduces the SCA and treatment time and increases the average gas velocity and the opportunity for dust reentrainment. Another report of this series^ provides a detailed description of combustion processes. 2-5 ------- 2.1.3 Design Considerations The design of a precipitator to clean the flue gas from municipal waste combustion presents no new problems that have not already been faced in applications to coal combustion, cement kilns, and metallurgical refining. In fact, a great deal of ESP operating experience on hundreds of incinerator applications worldwide has been acquired by ESP manufacturers in Europe, Japan, and the United States. Several considerations of special emphasis in the ESP design process are reviewed briefly in the following paragraphs, assuming no acid gas controls on the flue gas entering the ESP. Combustion Efficiency. The main source of difficulty in ESP appli- cations to municipal waste combustors, especially mass burn units, is the var- iability of waste material fired as fuel. Variations in the fuel and in the excess air supplied can cause large variations in the combustion efficiency and in critical parameters of combustion flue gas entering the ESP—gas volume flow, gas temperature, and dust particle size distribution. The precipitator design must take into account worst case conditions. This is a familiar problem in electric generating plants firing a variety of coals. In municipal incinerators, it is essential to maintain and monitor good combustion. Acid Corrosion. The acid dew point depends on the (variable) concentrations of moisture and acid vapors (S02, HC1) in the combustion flue gas. Corrosion problems can be expected when the temperature in the ESP falls below 350*F without acid gas control ahead of the ESP. Heated purge air must be supplied to the high voltage bushings to keep them dry, and special precautions are required during startup and shutdown. Use of corrosion resistant materials, proper insulation, prevention of air inleakage, purging acid gas directly after shutdown, and use of an indirect heating system during shutdown can extend the life of the ESP. Prevention of acid corrosion is a familiar problem in electric generating plants firing high sulfur coal and in pulp and paper mills. Fine Particle Collection. Particular attention must be paid to the collection of submicrometer particles because these particles tend to be enriched in heavy metals. Combustion variability complicates this design consideration because incomplete combustion tends to partition metals to the bottom ash. As combustion improves, submicrometer fumes can be controlled and finer particles can be generated. Experience with modern utility fly ash precipitators has demonstrated that an ESP which is designed for high total mass collection efficiency (99.9%) will also have satisfactory collec- tion of fine particles although lower than for large particles and the total particulate matter. Dust Reentrainment. Higher carbon content in the dust decreases the dust electrical resistivity and leads to increased dust reentrainment from the collecting plates. A uniform gas flow distribution, without large- scale turbulence, will minimize reentrainment problems. Proper design of the aspect ratio (length or depth of the ESP to height of the plate) and the gas flow baffling in the ESP are essential to uniform flow. After the precipitator is brought on line, the electrode rapping schedule and rapping intensity must 2-6 ------- be gradually adjusted (over a period of many weeks) to minimize emissions from rapping reentralnment. Figures 2-2 depicts a typical ESP for existing municipal incinerators in the U.S. 2.1.4 Performance of ESPs6 ESP control data are available for total particulate matter, heavy metals, dioxins, and acid gases (see Reference 7). The particulate matter data have been analyzed, fitted to a modified Deutsch-Anderson equation, and corre- lated with specific collection area (SCA). Figure 2-4 Illustrates this corre- lation for solid waste-fired industrial boilers. The curves shown are based on emission data and may not reflect what vendors would guarantee for a given SCA. Table 2-2 shows the data used to develop the 1986 correlation, reflecting data for high SCA ESP applications which were obtained after 1982. Very limited performance data on control of heavy metals, dioxins, and add gases by ESPs alone are available. Based on data from the Chicago North- west and Andover, MA, tests, an ESP without an appreciable drop in flue gas temperature offers little or no control of volatile organics. Data in Reference 6 show that control of arsenic, chromium, and lead emissions are less than the total particulate control, suggesting that heavy metals are somewhat concentrated in the very fine particles. The limited ESP data show no appreciable reduction in mercury emissions as mercury is mainly a vapor at normal operating temperatures. 2.2 Fabric Filters Fabric filters have not generally been applied directly to flue gas from municipal Incinerators, but rather as a sorbent collector and secondary reactor for dry and semi-dry scrubbers as discussed later in Chapter 4. The basic theory and operation of fabric filtration 1s presented here as background for Chapter 4; no performance data will be discussed in this section. Three reasons that fabric filters have not been applied to Incinerator flue gas are: (1) attack by acid gases upon fabric, (2) fabric blinding by "sticky" particles, and (3) baghouse fires caused by unstable combustion and carryover of sparks into the flue. Electrostatic precipitators and wet scrub- bers have been somewhat more forgiving to these phenomena and have generally been applied. However, with upstream scrubbing of acid gases and sorbent accumulation on fabric materials, all the above concerns are addressed and fabric filters become a very attractive choice for particulate control as well as control of other pollutants. Figure 2.3 shows a municipal solid waste incinerator facility with a flue gas cleaning system having a fabric filter although an ESP may be used In lieu of the baghouse. 2.2.1 Theory and Principles of Filtration8 The five basic mechanisms for particulate collection by fibers In a fabric filter are: (1) Inertia! impaction, (2) Brownlan diffusion, (3) direct interception, (4) electrostatic attraction, and (5) gravitational settling. 2-7 ------- Figure 2-2. Diagram of a two-field, weighted-wire ESP. 9 (Courtesy of Western Precipitation.) 2-8 ------- 0.8- 0.7- 0.6- s ® C.5- i 0.4- jo 0.3- 0.2- 0.1- i ( SO 100 I 200 ,1962 correlation : 1986 correlation 300 SCArn'rUDOOectm) I 400 I 500 I 600 Figure 2-3. PM emissions vs. SCA for best-fH equations. 2-9 ------- TABLE 2-2. PLANT TEST DATA^ (1986 Correlation) PM Emissions, lb/106 Btu 0.266 0.167 0.167 0.092 0.082 0.083 0.071 0.099 0.040 0.059 0.050 0.053 0.036 SCA, ft2/ 1,000 cfm 138 145 134 240 246 260 297 288 343 347 630 616 474 2-10 ------- IX) I A TYPICAL OGDEN MARTIN FACILITY 1 Tipping Floor 2 Re I use Holding Pit 3 Feed Crane 4 Feed Chute 5 Martin Stoker Orate 6 Combustion Air Fan 7 Martin Residue Discharger and Handling System 8 Combustion Chamber 9 Radiant Zone (furnace) 10 Convection Zone 11 Superheater 12 Economizer 13. Dry Gas Scrubber 14. Baghouse or Electrostatic Preclpllalor 15 Fly Ash Handling System 16 Induced Draft Fan 17 Stack 16 Figure 2.4. Municipal waste incinerator equipped with a dry flue gas cleaning system. (Courtesy of Ogden Martin Systems, Inc.) ------- Inertial impaction is the dominant collection mechanism within the dust cake. The forward motion of the particles results in impaction on fibers or on already deposited particles. Although impaction increases with higher gas velocities, these high velocities reduce the effectiveness of Brownian diffusion. Increasing the fabric and dust cake porosity by use of a less dense fabric or more frequent cleaning also reduces diffusional deposition. Except at low gas velocities, gravity settling of particles as a method of collection is usually assumed to be negligible. Electrostatic forces may affect collection because of the difference in electrical charge between the particles and the filter; however, the impact on commercial-scale equip- ment is not fully understood. Sieving, or particle filtering, occurs when the particle is too large to pass through the fabric matrix. It is not a major mechanism for collecting particulate. The combination of all these particle collection mechanisms results in the high efficiency removal of particulate matter. Dust Accumulation on Fabrics. The fabric filtration process or the accumulation of particulate on a new fabric surface occurs in three phases: (1) early dust bridging of the fabric substrate, (2) subsurface dust cake development, and (3) surface dust cake development. The fabric used in a fabric filter is typically a woven or felted material, which forms the base on which particulate emissions are collected. Woven fabrics con- sist of parallel row of yarns in a square array. The open spaces between adjacent yarns are occupied by projecting fibers called fibrils. Felted fabrics are constructed of close, randomly intertwined fabrics that are compacted to provide fabric strength. In the first phase, particles entering a new fabric initially contact the individual fibers and fibrils and are collected by the filtration mech- anisms. These deposited particles, which are essentially lodged within the fabric structure, promote the capture of additional particles. As these particles build up during the second phase, particle aggregates form, bridg- ing of the interweave and interstitial spaces occurs, and a more or less continuous deposit is formed. In the third phase, particles continue to collect on the previous deposit, and the surface dust cake is developed. The cleaning cycle (via shaking, reverse air, or pulse jet) removes some of the surface cake. After a few cleaning cycles, theoretically a steady-state dust cake should be formed, which will remain until the bag is damaged, replaced, or washed. Actually, however, the dust cake can vary significantly from cycle to cycle, particularly in applications involving utility boilers or metallurgical processes. This remaining cake forms a base for the collection of particles when the bag is put back on line after cleaning. 2.2.2 Gas Stream Factors that Affect Fabric Filter Design and Operation A complete characterization of the effluent gas stream is impor- tant in the design and operation of the fabric filter system. It should include the gas flow rate; minimum and maximum gas temperatures; acid dew point; moisture content; presence of large particulate matter; presence of sticky particulate matter; particulate mass loading; chemical, adhesion, and abrasion properties of the particulate; and presence of potentially explosive 2-12 ------- gases or participate matter. These data are used to design a collector with the required degree of control or to optimize the operation of an existing fabric filter, as illustrated by the following considerations: o The size of a fabric filter system is determined by the gas volume to be filtered and the pressure drop at which the filter can be operated, given the fabric type, dust cake properties, and cleaning method. The area of fabric surface (A) is determined by multiplying the total gas flow by the selected air-to-cloth ratio (A/C), which is based on the cleaning method or type of fabric filter. o Penetration is related to the effective A/C ratio in the system, particularly if the A/C ratio is outside the optimum range for the specific application and type of fabric filter. Therefore, the lowest possible face velocity (particle velocity at filter surface) consistent with economic constraints should be specified during the design phase. This parameter should also be considered in the operation of an existing fabric filter, if process flow rates increase significantly or additional sources are added. o Variations in gas stream temperature over time affect the opera- tion and design of a fabric filter. The temperature of gases emitted from industrial processes may vary more than several hundred degrees within short periods of time. It may fall below the gas moisture and acid dew points, or it may exceed the maximum temperature that the fabric will tolerate. The tem- perature extremes must be determined before the filter fabric is selected and during evaluation of fabric filter performance. o The particle size distribution of the dust must be considered in the design and operation of the collector. Particle size distribution affects both the porosity of the dust cake and abrasion of the fabric. The presence of fine particles in the gas stream can create a very compact dust cake and increase the static pressure drop through the cake. These fine particles can also cause fabric bleeding if pulled through the fabric. The presence of large abrasive particles can reduce bag life and may necessitate the use of a precleaner or gas distribution devices in the collection system. Moisture content and acid dew point are important gas composition factors. Operating a fabric filter at close to the acid dew point introduces substantial risk of corrosion, especially in localized spots close to hatches, in dead air pockets, in hoppers, or in areas adjacent to heat sinks, such as external supports. Allowing the operating temperature to drop below the water and/or acid dew point, either during startup or at normal operation, will usually cause blinding of the bags. Acids or alkali materials can also weaken the fabric and shorten its useful life. Trace components, such as fluorine, also can attack certain fabrics. 2-13 ------- 2.2.3 Fabric Filter Systems Although the basic participate collection mechanisms are the same for all fabric filters and gas stream factors affecting their perform- ance are relatively similar, the equipment itself and fabrics used in fabric filter systems may vary widely among vendors and applications. Some of these variations are necessary to meet various performance capability demands and physical characteristics; others are the products of individual contributions of the numerous equipment and fabric vendors. Although fabric filters can be classified in a number of ways, the most common way is by their method of fabric cleaning: shaker, reverse-air, and pulse-jet. Shaker-Type Fabric Filters. A conventional shaker-type fabric filter is shown in Figure 2-5. PTrffcu late-laden gas enters below the tube sheet and passes from the inside bag surface to the outside surface. At regular intervals a portion of the dust cake is removed by manual shaking (small systems) or mechanical shaking (large systems), the preferred method in MWC applications. Mechanical shaking of the filter fabric is normally accomplished by rapid horizontal motion induced by a mechanical shaker bar attached at the top of the bag. The shaking creates a standing wave in the bag and causes flexing of the fabric. The flexing causes the dust cake to crack, and portions are released from the fabric surface. The cleaning intensity is controlled by bag tension and by the amplitude, frequency, and duration of the shaking. Woven fabrics are generally used in shaker-type collectors. Because of the low cleaning intensity, the gas flow is stopped before cleaning begins to eliminate particle reentrainment and to allow the release of the dust cake. The cleaning may be done by bag, row, section, or compartment. Gas flow through shaker-type fabric filters is usually limited to a low superficial velocity (numerically equals A/C ratio) of less than 3 ft/min., typically ranging from 1 to 2 ft/min. High A/C values can lead to excessive particle penetration or blinding, which reduces fabric life and results in high pressure drop. Typical A/C ratios are 2 to 6 cfm/ft2. Mechanical shaker-type units differ with regard to the shaker assembly design, bag length and arrangement, and type of fabric. All sizes of con- trol systems can use the shaker design. Reverse-Air fabric Filters. A large and typical reverse air filter is shown in Figure 2-6. Regardless of design differences, the cleaning principle is the same. Cleaning is accomplished by reversal of the gas flow through the filter media. The change in direction causes the surface contour of the filter surface to change (relax) and promotes dust-cake cracking. The flow of gas through the fabric assists in removal of the cake. The reverse flow may be supplied by cleaned exhaust gases or by ambient air introduced by a secondary fan. 2-14 ------- OVERMOUNTED EXHAUSTER DOOR DOOR INLET HANGER. \r"CHANNEL BAG NOZZLE/ AND-RETAINER NLET CHAMBER AND HOPPER BAFFLE • -SUPPORT DISCHARGE GATE Figure 2-5. Small shaker-type fabric filter. 8 2-15 ------- Figure 2-6. Example of a large reverse-air fabric filter. 2-16 ------- In filters with inside bag dust collection, cleaning is done with com- partments isolated. The filter bags may require anticollapse rings to prevent closure of the bag and dust bridging. Reverse-air filters are usually limited to A/C ratios of less than 3 cfm/ft2 and a range of about 1 to 2.5 cfm/ft2. In general, the appropriate A/C ratio for a reverse-air unit should be about one-third lower than for a similar shaker-type unit application. Pulse-Jet Fabric Filters. In pulse-jet fabric filters, filtering takes place on exterior bag surfaces. A small pulse-jet fabric filter is illustrated in Figure 2-7. The bags supported by inner retainers (usually called cages) are suspended from a tube sheet, an upper cell plate. Com- pressed air for cleaning is supplied through a manifold-solenoid assembly into blow pipes. Venturis are sometimes mounted in the bag entry area to improve the pulse-jet effect and to protect the top part of the bag. The diffuser is placed at the gas inlet to prevent large particles from abrading lower portions of the bag. During cleaning, a brief (generally less than 0.2-second) pulse of com- pressed air injected into the top of the bag creates a traveling wave in the fabric, which shatters the cake and throws it from the surface of the fabric. The dominant cleaning mechanism in a pulse-jet unit is fabric flexing. Fel- ted fabrics are normally used, and the cleaning intensity (energy) is high. The cleaning usually proceeds by rows, and all bags in a row are cleaned simul- taneously. The compressed-air pulse, which is delivered at 80 to 120 psi, results in local stoppage of the gas flow. The cleaning intensity is a func- tion of compressed-air pressure. Pulse-jet units can operate at substantially higher A/C ratios than the previously discussed fabric filters because of their higher cleaning intensity. Typical ratios range from 5 to 10 cfm/ft2. The plenum pulse cleaning method is a variation of the pulse-jet clean- ing mechanism; in this method, an entire compartment of bags is taken off-line and pulsed with compressed air from the clean air plenum. Other Designs and Modifications. Fabric filters can be constructed either as a positive-pressure unit with the fan upstream of the fabric filter or as a negative-pressure unit with the fan downstream of the unit. Recent applications have employed the fan downstream of the fabric filter. The use of a positive-pressure fabric filter eliminates the need for ductwork and a stack downstream of the unit, which reduces requirements for space and other materials but makes sampling to determine particulate loading more difficult. Because positive-pressure units are generally not exposed to as high a static pressure as negative-pressure units, their housings can sometimes be constructed of a bolted light-gauge material. Any leaks from the fabric filter will enter the surrounding air and increase the fugitive emissions from the unit. In negative-pressure units, the fan is located on the clean side of the filter, where it is subject to less wear from dust abrasion. The fabric filter housing must be gas-tight, as any leaks will draw air in from the 2-17 ------- outside. This outside air will normally cool the gas stream. This could reduce the gas temperature below the dew point and cause condensation on the inside of the unit. In some processes, introduction of outside air increases the risk of fire and/or explosion. On the other hand, leaks will not result in fugitive emissions because ambient air is drawn into the unit. 2.3 Wet Scrubbers Wet scrubbers for particulate matter control in municipal incin- eration are not likely to be used in the future. Although accounting for nearly one-fifth of all particle control systems on U. S. incinerators,1 wet scrubbers have the following disadvantages: o cannot meet current or future particulate matter emission requirements without very high pressure losses with accom- panying erosion and increased maintenance requirements, o a liquid waste is generated, and o water scrubbers will absorb acid gases to some extent and, if not designed to handle acids, will have significant operating problems. In short, any wet scrubber applied to an incinerator will be designed for gas absorption/multipollutant control, and any particulate matter con- trol will complement a fabric filter or electrostatic precipitator. For this reason wet scrubber descriptions and functions are discussed in Chapters 3 and 4. 2-19 ------- oo o i: \ \. • _w_ • A i. - ' * * ^""^\ ^ ^""" •f*^*"*^^"^M^*^^l*"^^^>^^^^^*^Tpa**^!^'?^*1l^^a**^tyfc^^TB W - j ^^ •*• » ', ^rjf ^W s ex § K I CM ------- 3.0 GASEOUS EMISSION CONTROLS Gaseous emission controls are designed for capture of a particular gas or gases and are Intended for use in series with particulate control devices. Although several wet scrubbers are available for specific hydrochloric acid mist control, their use has been confined to special waste incinerator appli- cations. Post-combustion NOX control systems have been applied in Japan to non-incinerator combustion scrubbers and more recently to municipal inciner- ators. Consequently, post-combustion NOX controls are also discussed below. 3.1 Acid Gas Scrubbers (HC1 and HF only)10 Add gas scrubbers are generally found in small waste incinerator applications in convenient packaged form. Using water or very dilute sodium solutions, hydrochloric acid and hydrofluoric acids are easily absorbed with little regard to operation and maintenance. However, in municipal incinerator applications with particulate matter, sulfur oxides, and condensible organics, exclusive control of acid gases becomes more complex. Trace alkali in scrubber water can react with flue gas components to form insoluble salts; therefore, the pH must be controlled to less than 4 to limit absorption of gases other than HC1 and HF and precipitation of insolubles. Calcium scrubbing is not likely for this reason, while clear liquor (sodium) scrubbing is also potentially troublesome due to erosion and corrosion. Commercial practice in Europe has been to use water only scrubbing with minimal internals—spray towers or venturi scrubbers—at very low (0 to 1) pH to minimize absorption of other gases. Water is introduced at roughly one liter per 20-25 mg HC1 to be absorbed. Water injection rate is controlled by monitoring solution conductivity. Scrubber blowdown is neutralized with lime before being discharged or further treated. HC1 and HF removals over 90 percent are routinely achieved. Scrubbers are made of corrosion-resistant materials with mist eliminators and rubber-lined stacks to minimize acid attack by the exhaust gases. Figure 3-1 illustrates a typical system of this type. Effluent regulations in Europe are already forcing Installations of this type to evaporate liquids and send residue and solids to special waste storage. Newly enacted S02 removal requirements and restrictions on liquid wastes are rendering single- stage acid gas wet scrubbers obsolete. Multistage scrubbers with effluent fed to spray dryers upstream are currently considered state-of-the-art in Europe. These are discussed in Chapter 4. 3.2 Post Combustion NOX Control 3.2.1 Selective Catalytic Reduction (SCR)H For post combustion NOX control, selective catalytic reduction is the most advanced process, with ammonia reducing NOX to nitrogen, N2, and water vapor in the presence of a catalyst. This approach differs from the 3-1 ------- LEGEND: 1 Gas-Gas Heat Exchanger 2 YentuH Scrubber 3 Stack * Lime 511o 5 Lime Slater ... 6 Neutralization Tank ± 3 Cleaned Flue Gas Waste Liquor Treatment and 01 sposal Figure 3-1. Add gas scrubber. 10 3-2 ------- selective non-catalytic reduction (SNR) process tcommonly called Thermal De-N0x) discussed 1n the combustion report of this series in several ways: o Ammonia is injected into the flue gas at lower temperature [150-400*C (302-752"F)] zones, o A catalyst is used to compensate for the lower driving force of reaction and to better utilize ammonia, o Higher NOX removal Is possible than for Thermal De-N0x, o SCR can be adversely affected by certain metals and acid gases, and o Secondary ammonia emissions are higher with Thermal De-N0x than with SCR. Since virtually all of the NOX in combustion gases is in the form of nitrogen oxide (NO), a small amount of oxygen -promotes the reaction as follows: 4NO + 4NH3 + 02 > 4N2 + 6^0 (3-1) For typical applications on fossil-fuel-fired boilers, SCR removes 60-85 percent of the NOX using 0.61-0.90 mole 1^3 per mole NOX, leaving a few ppm unreacted NH3 (although monitoring methods for NH3 are not avail- able to confirm this on a continuous basis). The addition of larger amounts of NH3 increases NOX removal but produces larger amounts of unreacted N«3 (ammonia slip). The optimum temperature for Reaction (3-1) is 300-400'C (572-752'F), so SCR is generally applied to flue gas at the economizer outlet. Earlier SCR problems such as catalyst poisoning by SOX, plugging, ammonium bisulfate deposition, production of $03, and erosion of cata- lyst have generally been overcome. However, attack of conventional SCR catalysts (which use base metal with titanium dioxide) by hydrochloric acid is still a major problem. This may be overcome by either (1) develop- ment of HCl-resistant catalyst or (2) location of SCR downstream of HC1 controls. Two commercial applications of SCR on municipal incinerators in Japan began operation in late 1986. Both use a special low temperature, acid-resistant catalyst developed by Mitsubishi Heavy Industries (MHI). One, a 150 ton/day plant in Tokyo, is a retrofit application with very limited space for the catalyst. Consequently, the design NOX removal is only 30-40 percent from flue gas having 120 ppmv NOX and 500-800 ppmv HC1. The SCR unit Is located downstream of an ESP and upstream of a sodium-based wet scrubber. The other SCR system, also developed by MHI, has been integrated with two new 65 ton/day Incinerators in Tokyo and follows the lime spray dryer/baghouse systems. This SCR system will treat a gas stream [^ 200*C ( ^392*F)] containing about 50 ppmv HC1, 50 ppmv SOX, 30 mg/Nm3 dust, and 150 ppmv NOX. The system is designed to remove 73 percent NOX (40 ppmv at outlet), but the guarantee is only 33 percent NOX removal or an emission NOX concentration of 100 ppmv. 3-3 ------- SCR 1s also being applied to sludge incinerators in Japan. At least 15 SCR plants on municipal sludge incinerators have been constructed by MHI. Normally, impurities such as HC1 and trace metals degrade SCR catalysts, so the incinerator gas is typically subjected to sodium wet scrubbing and, in some cases, a wet ESP to remove HC1 and trace metals to acceptable levels. Then a special titanium-based honeycomb catalyst developed by MHI is used to remove NOX by 80-90 percent. Flue gas volumes for this application typically range from 2,500 to 108,000 Hm3/h. Design data for some sludge incinerator installations are presented in Table 3-1. As shown, uncontrolled NOX levels of 100-150 ppmv are reduced by 80-90 percent with only 5 ppmv of ammonia slip. Performance data are not available at this time. TABLE 3-1. SCR PLANT DESIGNS FOR SLUDGE INCINERATOR FLUE GAS11 Gas Treated, Nm3/hr S02, ppmv NOX, ppmv Temperature, *C NH3/NOX molar ratio NOX removal, % NH3 slip, ppmv 108,000 10 100 350 1.0 90 5 40,000 50 150 400 1.0 80 5 24,000 20 130 400 1.0 90 5 2,500 20 130 400 1.0 90 5 Three SCR options for incinerators are shown in Figure 3-2.12 option I shows conventional SCR as practiced in coal-fired utility applications. Option II uses SCR after the metals and acids have been removed, but the catalyst operates in a lower temperature range. Such catalysts are in ad- vanced development in both Europe and Japan, and two commercial applications of low temperature catalysts in Japan were noted above. Option III is one of several possible schemes to utilize current commercially demonstrated technology as discussed earlier. Note that air preheat is replaced by reheat of SCR inlet gas, incurring a substantial energy penalty. Other schemes could use auxiliary reheat, but in no case can bypass reheat be used, Noteworthy are the temperatures of operation. Depending on the catalyst used, temperatures as low as 200*C (392*F) may be used, and reheat costs are reduced (see Figure 3-2, Option III). Another benefit is that noxious and difficult to scrub gases, such as mercaptans and sulfides, are decomposed by oxidation by about 80 percent. 3-4 ------- 1 \/ ^{ *tpn \ ^. ^- r""™"\ ^A ) PM/HCI f CONVENTIONAL SCR (WITH MCI RESISTANT CATALYST) A/P b. to I en SCR WITH LOW TEMPERATURE CATALYST A/P c. „ SCR WITH EXTENSIVE GAS REHEAT I = INCINERATOR A/P= AIRPREHEATER PM/HCI = ESP/SCRUBBER, ETC. GGH= GAS GAS HEAT EXCHANGER B = REHEAT BURNER Figure 3-2. SCR options for municipal incinerators. 1 ? ------- Although this SCR process Is commercially used, Improvement efforts focus on two areas: low temperature catalysts (Option II, Figure 3-2) and high temperature catalysts resistant to HC1 attack and metals poisoning. Operation and maintenance techniques to ensure reliable operation include periodic inspection of the catalyst, with the partial replacement of the bed during annual outages. Routine monitoring of NOX emissions and ammonia slip by grab sampling is necessary to determine the catalyst act- ivity and potential for buildup of ammonium sulfate on internal surfaces downstream. The pressure drop across the catalyst bed should be monitored to determine plugging and/or channeling, both of which result in poor performance. The gradual loss in catalyst activity is inevitable during SCR opera- tion. If a minimum NOX removal is mandated, the NH3/NOX molar ratio can be gradually increased to compensate for the degradation of the catalyst. This approach entails increased ammonia leakage and potential sulfate buildup downstream. Alternatively, the catalyst may be partially or totally replaced depending on the tradeoffs of ammonia cost, increased maintenance, and lower reliability versus catalyst cost. 3.2.2 Wet NOX Removal Processes11 Wet processes for NOX removal have several drawbacks: (1) most nitrogen oxides are present as nitric oxide (NO) which is relatively un- reactive; (2) N02 is relatively soluble but somewhat less reactive than $03 or other acid gases; (3) oxidation of NO to N0£ is difficult; and (4) nitrate and nitrite byproducts have limited use, requiring expensive processing to produce a salable product, such as fertilizer. The processes which have been used in Japan and are planned in Germany for incinerators include oxidation/reduction and complex absorption. In oxidation/reduction, a strong oxidizing agent such as sodium chlorite or ozone is added to flue gas to convert all NOX to N02 prior to wet scrubbing (usually sodium-based). The complex absorption process uses ethylenediamine tetracetic acid (EDTA) and ferrous ions which promote NO absorption by forming a complex compound. Another wet NOX process involves addition of ammonia upstream of a sodium hydroxide spray dryer in Japan. At temperatures of 400 to 500*C (752 to 932T), ammonia 1s Injected at a molar ratio of 0.35 NH3/NOX to get 30 percent NOX reduction In the spray dryer-ESP system. Sodium sulfite is the only product waste; hence a reduction of NOX to N£ is suspected, similar to the SCR reactions. Because the effect of HC1 on this process is not known, the applicability of this process to municipal solid waste inciner- ation 1s uncertain. Since the above wet NOX processes involve scrubbing where more reactive species (SOg, HC1) than NOX are present 1n the flue gas, all wet NOX schemes are essentially multiple pollutant control scrubbers. They are discussed in more detail 1n Chapter 4. 3-6 ------- 4.0 MULTIPOLLUTANT CONTROL SYSTEMS European and Japanese regulations generally require control of gaseous and particulate matter emissions; hence control systems which simultaneously remove several pollutants are commonly used. In addition, a growing number of local and state permitting agencies in the United States recognize the need for multipollutant control and, in essence, require the installation of systems similar to those in Europe. This section will review the various types of multipollutant control systems commercially available today and discuss their performance in controlling po"1 utants of concern. 4.1 Wet Scrubbing The use of wet scrubbers for multipollutant control has been practiced since the early 1970s, notably in Europe and Japan. By addition of alkali—sodium or calcium—a particulate scrubber also significantly reduces the level of acid gases—S02, HC1 and HF—and converts them into an aqueous salt solution or slurry requiring treatment and disposal. More recently wet scrubbers designed for acid gas removal have been augmented by addition of chemicals to enhance NOX removal as well. These augmenta- tions involve expensive chemicals and increase the volume of scrubber waste. 4.1.1 Wet Scrubbing for Acid Gases (HC1, HF, and S02) 10 Wet scrubbing for acid gases differs markedly from wet scrubb- ing for particulate matter in several ways: (1) emphasis is on gas/ liquid contact and not impingement of particles, (2) chemical additives to the scrubber liquor require careful attention to the system chemistry and accelerate the potential for corrosion and fouling of scrubber internals, and (3) the volume of waste is substantially increased. As a result, sig- nificantly more attention is paid to process conditions and effluent/exhaust chemical composition with wet scrubbing for gases. Many types of wet scrubber are used for removing acid gases—spray towers, centrifugal scrubbers, venturi scrubbers. Scrubbers with internals, such as packed-beds and trays, are less commonly used. Figure 4-1 illus- trates a typical wet scrubber installation for gas absorption. Gas enters the absorber where It is contacted with an alkaline solu- tion. For this discussion lime Is used as the reagent. The lime solution reacts with the add gases to form salts, which are generally insoluble and may be removed by sequential clarifying, thickening, and vacuum filtering. The dewatered salts or sludges are then landfilled. In coal-fired utility boiler applications, the process features potential closed-loop operation, wherein all water 1s recycled to the process except for evaporative losses and residual moisture in the solid waste. Further, the suspended solids usually contain mostly calcium sulfite which may be oxidized in an intermediate step by blowing air through the liquor to form calcium sulfate or gypsum which is a poten- tially salable by-product, as well as a more manageable, drier solid waste. 4-1 ------- Fresh Mater Liquid Waste Flyash L1me 11 Lime from Vehicles LEGEND: 1 Flue Gas Inlet 2 Absorber 3 Hold Tank 4 Hash Tray Tank 5 Cl ar1 f1 er 6 Thickener 7 Vacuum Filter 8 Pug Mill 9 L1me Storage 10 Line Slaker 11 Dilution Tank 12 Heat Exchanger 13 Stack Figure 4-1. Line scrubbing for add gas and S02 removal 13 4-2 ------- For incinerator lime scrubbing, the following chemical reactions occur: Ca(OH)2 + S02 —> CaS03 • 1/2H20 + 1/2H20 (4-1) CaS03 • 1/2H20 + 1/202 + 3/2H20 —> CaS04 • 2H20 (4-2) Ca(OH)2 + 2HC1 —> CaCl2 • 2H20 (4-3) Ca(OH)2 + 2HF —> CaF2 • 2H20 (4-4) However, the presence of substantial HC1 in incinerator flue gas results in a significant portion of calcium chloride (CaCl2) in waste solids as well as a buildup of chlorides in the liquor such that waste handling and liquor recycle are more difficult. Solids become more difficult to settle and dewater with CaCl2, and generation of a useful waste material such as gypsum is impractical. The buildup of chlorides in liquor necessitates a purge stream to retain good scrubber performance and minimize plugging, scaling, and corrosion. In the case of incinerator flue gas, a single scrubber would operate in essentially an open loop, with large amounts of water consumption, treatment, and waste. A more practical scheme is to install a prescrubber just upstream of the absorber in Figure 4-1 which operates independently, with a separate liquor loop, primarily as a chloride/fluoride scrubber. In this way the main absorber may act as an S02 scrubber (or S02/NOX scrubber in more advanced schemes) and generate more stable waste with liquor recycle. This will be discussed in more detail in Section 4.3 when the hybrid scrubber systems are discussed. 4.1.2 Wet Scrubbing for Acid Gases and NOX11 In more recent advances, NOX control is possible by either adding a chemical which absorbs nitric oxide (NO), the primary constituent of NOX, or oxidizes NO to N02 (nitrogen dioxide) which is more readily absorbed in scrubber liquor. In the former case, a dissolved catalyst, ethylenediamine tetraacetic acid (EDTA), is used in a sodium or ammonia solution with ferrous ion. The sulfur oxides present are absorbed by either sodium sulfite or ammonium hydroxide, and these compounds enter into a complex series of reactions that take place after the NO is absorbed. The overall reaction is as follows for ammonia scrubbing: 2NO + 5S02 + 8NH3 + 8H20 > 5(NH4)2S04 (4-5) Since this process has not been operated on a municipal incinerator applica- tion, it is assumed that chlorides and fluorides will be removed in a separate loop to minimize adverse effects. This process is offered by one German vendor, Saarberg-Hoelter-Lurgi, and is scheduled for installation on utility boilers. Little information has been disclosed about process details, but 4-3 ------- it is suspected that the EDTA process may well be one of the seven wet NOX absorption processes recently announced for installation on incinerators in West Germany.14 The other wet NOX absorption technique uses an oxidizing agent such as ozone, chlorine dioxide, or sodium chlorite to oxidize NO to N02. Then the N0£ is scrubbed in a conventional sodium or magnesium scrubber along with other acid gases. The reactions are: NaC102 + 2NO —> NaCl + 2N02 (4-6) ZNaOH + 2N02 + 1/202 —> 2NaN03 + H20 (4-7) If a calcium scrubber is used, a catalyst must be added, and the result- ing waste stream requires special treatment to avoid gaseous ammonia evolu- tion from the waste liquor. These processes are discussed further in Section 4.3. 4.1.3 Performance of Wet Scrubbers10 The performance of wet scrubbers is dependent on many factors, and the data base is too limited on incineration applications for detailed discussions of the effects. Theoretically, the reaction of strong acid gases (HC1, HF) proceeds rapidly with alkaline solutions and even mildly acidic solutions. Hence, HC1 and HF removal should be high (greater than 90 percent) in every case, assuming proper operation. The reaction of S02 proceeds more slowly and over a limited pH range, the limiting factors being the rate of S02 absorption and, for calcium systems, the dissolution rate of solid caustic particles. Thus, S02 removal may vary greatly over a limited range of operation, depending on pH control, inlet S02 concentra- tion, and many other factors discussed under operation and maintenance. Since wet scrubbers operate at saturation [40-50*C (104-122*F) typically], the quenching of flue gas by some 150 to 200'C (302-392*F) should reduce considerably the volatile compounds, including organics and trace metals. Reported European outlet emissions for a wet scrubber designed for multi- pollutant control, preceded by an electrostatic precipitator, are shown in Table 4-1. TABLE 4-1. WET SCRUBBER OUTLET EMISSIONS10 HC1 5-10 ppmv HF 1 ppmv S02 25 ppmv Particulate Matter <0.01 gr/dscf Heavy Metals3 <0.001 gr/dscf Hg 0.00002 gr/dscf ^Classes 1-3 of West German regulations. Includes Cd, Tl, Hg, As, Co, Ni, Se, Te, Sb, Pb, kCr, Cu, Mn, V. 4-4 ------- It would be expected that volatile organic pollutants would also be removed by wet scrubbers because these pollutants would be condensed to form particulate matter which would be collectible in the scrubber. However, no data are available to test this hypothesis. 4.2 Dry and Semi-Dry* Scrubbing Because of the difficulty in managing the buildup of chlorides necessitating dual wet scrubbing and the burden of a liquid waste disposal, the most popular flue gas scrubbers are the dry and semi-dry types. Dry scrubbing involves the injection of a solid powder such as lime or sodium bicarbonate into the flue gas where acid gas removal occurs in the duct and continues in the dust collector as sorbent and ash particles and condensed volatile matter are captured. In a semi-dry process, better known as spray drying, the sorbent enters the flue gas as a liquid spray with sufficient moisture to promote rapid absorption of acid gases but yet produces only dry solid particles entering the particle collector. 4.2.1 Dry Injection Processes10 Dry injection of alkaline sorbent into circulating flue gas followed by a particle collector has been recently developed in Europe, and more than 20 installations on incinerators are known in Europe and Japan. The best known and studied process of this type 1s the Malmo, Sweden-, install- ation as shown in Figure 4-2. A cyclone precollector removes a large amount of coarse fly ash prior to the waste heat boilers but does not play a role in the process. Dry calcium hydroxide (hydrated lime) is pneumatically injected into a reactor where the gas rises and mixes with the sorbent. The reactor also provides additional solids residence time to allow reactions to occur. An older electrostatic precipitator removes about 95 percent of the dust, while a newer pulse-jet fabric filter removes the remainder. In a new facility, it is likely that a fabric filter would be used, although several units with ESPs are reported. One interesting fact about the Malmo installation is the recent attempt to increase acid gas removal. A heat exchanger was Installed to lower inlet flue gas from 250*C (482'F) to 160-180*C (320-356'F), and the ESP was downgraded to allow more sorbent to enter the fabric filter. It is well known that fabric filters may remove substantial amounts of acid gas in proportion to the ratio of sorbent on the fabric to the acid gas flow rate. Also, a lower flue gas temperature or higher relative humidity in the flue gas increases acid gas removal. *A semi-dry scrubber 1s also known as wet/dry scrubbing and spray drying. It consists of a spray dryer and a dry solids collection device (ESP or fabric filter). In the U.S., dry scrubbing refers to a process 1n which solids from the process are dry (I.e., the cleaned flue gas is unsaturated with water vapor). Using the U.S. terminology, both a dry sorbent injection plus par- ticulate collection system and a spray dryer plus particulate collection system are dry scrubbers. 4-5 ------- 1. FURNACE AND BOILER 2. PRECOLLECTOR 3. WASTE HEAT BOILER NO. 1 4. REACTOR 5. ELECTROSTATIC PRECIPITATOR 6. FABRIC FILTER 7. WASTE HEAT BOILER NO. 2 8. LIME SILO 9. LIME FEEDING 10. LIME RECIRCULATION 11. COARSE DUST CONVEYING 12. FINE DUST CONVEYING 13. DUST SILO 14. DUST HUMIDIFIER 15. DUSTBIN Figure 4-2. Dry absorption system, Malmo, Sweden. 10 4-6 ------- Performance of Dry Injection/Fabric Filter Systems The following normal control device inlet pollutant concentrations and removal data have been reported at Malmo:10 TABLE 4-2. INLET POLLUTANT CONCENTRATION AND POLLUTANT CONTROL WITH A LIME DRY INJECTION PROCESS** Inlet Concentration Pollutant (mg/Nm3) Removal (%) Parti oil ate Matter 10 Not reported HC1 200 80 HF 0.2 98 S02 150 50 Cd 0.002 99+ Pb 0.04 99+ Zn 0.17 99+ Hg 0.04 90 It should also be noted that dioxin emissions were reported below the detectable limit (0.1 ng/Nm3), but this is likely due to very efficient combustion efficiency in the incinerators (reported as 99.8 percent). Recent data from Canada on a slipstream pilot plant indicate very high control of pollutants in municipal solid waste flue gas with dry scrubbing (lime injection followed by fabric filtration) for temperatures from 110-140'C (230-284*F) when humidification precedes the scrubber. Effective pollutant control was also obtained with a wet/dry or semi-dry (spray dryer plus fabric filter) system at 140'C (fabric filter inlet temperature with and without recycle from this flue gas cleaning system). Only the mercury and S0£ removals appear to vary with flue gas temperature. Mercury capture was 90 percent or better at 110-140*C but essentially nil with flue gas at about 200*C (392*F) into the fabric filter. S0£ capture ranged from 96 percent at 110'C to 29 percent for the 200*C fabric filter inlet temperature. Removal efficiencies for other trace organics (chloro- benzenes, polychlorinated biphenyls, and chlorophenols) were generally 95 percent or higher over the temperature range of 110-140*C. Polycyclic aromatic hydrocarbons removal was about 85 percent over the same tempera- ture range but rose to 98 percent at 200*C, while the removal of the other trace organics fell to the 55-60 percent neighborhood at 200'C. Tables 4-3 through 4-9 summarize the test conditions and results reported by Environment Canada for the Quebec City test program.15 A second type of dry injection process is a circulating fluid bed-absorp- tion process shown in Figure 4-3. The process is offered semi-dry for utility applications and totally dry for incinerator applications. A few 4-7 ------- TABLE 4-3. SUMMARY OF KEY OPERATING PARAMETERS^ Dry System4 Incinerator Steam flow, kg/h x TO3 Gas temperature 9 boiler outlet, "C Fabric Filter Pressure drop, cm water gauge Flow Rate Lime, kg/h Dry Flue Gas Inlet, Nm3/hb Midpoint, Nm3/h& Outlet, Nm3/hb Temperature Inlet to pilot plant, *C Inlet to fabric filter/C Flue Gas Composition at Outlet to Pilot Plant C02, % 02. * CO, ppm THC, ppm Particulate Loading Inlet to pilot plant, mg/Nm3 @ 81 02 no'c 32.8 303 15.7 3.7 3600 4400 4230 267 113 7.1 12.7 140 5 7710 125*C 33.4 287 14.4 3.6 3730 4430 4350 258 125 7.4 12.4 180 4 7260 140'C 33.2 291 14.9 3.7 3520 4120 4170 261 142 7.5 12.5 220 5 6250 >200'C 33.0 287 14.7 3.6 3010 3650 3600 253 209 7.3 12.9 160 4 5740 Wet/Dry System3 140*C 32.1 278 15.2 3.5 3650 4180 4220 254 140 8.3 11.8 130 4 5560 140"C "Recycle" 31.3 293 15.9 3.5 3560 4110 4090 263 141 7.5 12.5 170 4 7190 aTemperatures shown are nominal values at the midpoint (fabric filter inlet) of the pilot plant flue gas cleaning system. ^Reference conditions are 25*C and 101.325 kPa. 4-8 ------- TABLE 4- 4-4. PCDD CONCENTRATIONS (nq/NmJ @ 82 0?) IN EFFICIENCY OF REMOVAL^ FLUE GAS AND Inlet, ng/Nm3 Midpoint, ng/Nm3 Outlet, ng/Nm3 Efficiency Inlet/midpoint, X Overall , X 110'C 580 310 0.2 47 >99.9 Dry 125*C 1400 570 NDC 60 >99.9 System3 140*Cb 1300 540 NDC 57 >99.9 >200'C 1030 1140 6.1 (11) 99.4 Wet/ Dry 140'C 1100 840 NDC 24 >99.9 System4 140'C "Recycle" 1300 1270 0.4 2 >99.9 aTemperatures shown are nominal values at the midpoint (fabric filter inlet) of the pilot plant flue gas cleaning system. ^Based on one test. CND = Not detected. TABLE 4.5 PCDF CONCENTRATIONS (ng/Nm3 Q 8X 02) IN FLUE GAS AND EFFICIENCY OF REMOVAL** Inlet, ng/Nm3 Midpoint, ng/Nm3 Outlet, ng/Nm3 Efficiency Inlet/midpoint, X Overall, X no'c 300 270 2.3 11 99.3 Dry 125'C 940 440 NDC 54 >99.9 System3 140'Cb 1000 630 1.0 37 99.9 >200'C 560 490 1.2 13 99.8 Wet/Dry 140'C 660 690 NDC -4 >99.9 System3 140'C "Recycle" 850 1030 0.9 -21 99.9 3Temperatures shown are nominal values at the midpoint (fabric filter inlet) of the pilot plant flue gas cleaning system. ^Based on one test. CND = Not detected. 4-9 ------- TABLE 4-6. PERCENT REMOVAL OF OTHER ORGANICSlS Dry Systerna Wet/Dry System3 140'C 110'C 125*C 140'cb >200*C 140'C "Recycle1 Chlorobenzenes Polychlorinated blphenyls Polycyclic aromatic hydrocarbons Chlorophenols 95 72 84 97 98 >99 82 99 98 >99 84 99 62 54 98 56 >99 >99 >99 99 99 >99 79 96 a Temperatures shown are nominal values at the midpoint (fabric filter inlet) of the pilot plant flue gas cleaning system. bBased on one test. TABLE 4-7. INLET/OUTLET METAL CONCENTRATIONS (Ug/Nm3 @ 82 02)15 Dry System3 Metal Zinc (Zn) Cadmium (Cd) Lead (Pb) Chromium (Cr) Nickel (N1) Arsenic (As) Antimony (Sb) Mercury (Hg) Location Inlet Outlet Inlet Outlet Inlet Outlet Inlet Outlet Inlet Outlet Inlet Outlet Inlet Outlet Inlet Outlet 110'Cb 99000 7 1300 0.4 41000 4 3100 0.4 1000 1.3 150 0.02 2000 0.2 440 40 125'Cb 108000 5 1300 0.4 44000 3 1900 0.4 1800 0.4 100 0.04 800 0.4 480 13 140*C 93000 6 1500 NDd 34000 5 2000 1 1300 0.7 130 0.04 1000 0.6 320 20 >200*C 91000 10 1000 0.6 35000 6 1900 0.5 800 2 80 0.07 1500 n.5 450 610 Wet/Dry 140*C " 77000 5 1200 ND<1 36000 1 1400 0.2 700 1.3 110 0.04 1000 0.3 190 10 System3 140"C Recycle" 88000 6 1100 NDd 34000 6 1700 0.7 2500 2 130 0.03 2200 0.6 360 19 aTemperatures shown are nominal values at the midpoint (fabric filter inlet of the pilot plant flue gas cleaning system. bBased on one test, except for mercury which is based on two tests. concentrations are rounded off for simplicity. = Not detected. 4-10 ------- TABLE 4-8. HYDROGEN CHLORIDE CONCENTRATIONS (9 82 02) AND REMOVAL EFFICIENCIES!* Dry System3 Stoi chi ome trie Ratio Inlet, ppm Midpoint, ppm Outlet, ppm Eff. to midpoint, % Eff. Overall, X no*c 1.16 423b 15b 7b 96 98 125°C 1.03 464C 69b 9b 85 98 140'C 1.04 475b.e 129b,e 2gb,e 73 94 >200*C 1.49 392C,e 196b,e 9ic,e 50 77 Wet/Dry System3 140*C 1.19 366d I49b 29b 59 92 140'C "Recycle" 1.10 470b 152b 42b 68 91 3Temperatures shown are nominal values at the midpoint (fabric filter inlet) of the pilot plant flue gas cleaning system. bBased on continuous gas monitors. cBased on manual sampling (15-30 minute tests). dBased on a combination of continuous gas monitors and manual sampling (15-30 minute tests). test only. TABLE 4-9. SULFUR DIOXIDE CONCENTRATIONS (0 8% 02) AND REMOVAL EFFICIENCIES15 Dry System3 Stoichlometric Ratio Inlet, ppm Midpoint, ppm Outlet, ppm Eff. to midpoint, X Eff. Overall, X no*c 1.16 119 24 4 80 96 125*C 1.03 118 65 10 45 92 140*Cb 1.04 99 64 41 35 58 >200*C 1.49 117 103 83 11 29 Wet/Dry System3 140*C 1.19 106 67 35 37 67 140"C "Recycle" 1.10 106 70 43 35 60 3Temperatures shown are nominal values at the midpoint (fabric filter inlet) of the pilot plant flue gas cleaning system. bBased on one test. 4-11 ------- LIME FLUE GAS 1. LIME SILO 2. REACTOR 3. CYCLONE 4. OUST COLLECTOR 5. STACK 6. WASTE SILO DRY WASTE Figure 4.3 Circulating fluid bed absorption (dry) process. 10 1. LIME FEEDER 2. LIMESLAKER 3. FEEDTANK 4. HEAD TANK 5. SPRAY ABSORBER 6. DUST COLLECTOR 7. STACK DRY WASTE Figure 4-4. Spray absorption (semi-dry) process. 10 4-12 ------- installations on incinerators are currently in operation in Western Europe, but the process has some significant advantages for an ESP retro- fit application discussed in the following paragraphs. Flue gas enters the vertical reactor equipped with a venturi-type gas disperser, where gas velocity is reduced and dry lime introduced. The reactor forms a fluidized layer of fly ash and sorbent reacting with acid gases. Eventually all dry solids are entrained and enter the dust separator just upstream of the electrostatic precipitator. A substantial portion of this collected product is reintroduced into the fluid bed. Important features of this process are high gas/solid mass transfer rates and recycle which improves the normally low sorbent utilization of dry injection processes. Also, reactors and precollectors are small volume vessels compared to wet and semi-dry process vessels. The important control parameter is the pressure drop across the entrained bed, which is maintained at around 20 mbar by controlling the recycle solids. Recycle product is mechanically conveyed to the venturi section and fed by gravity to the system. The fresh reagent rate is controlled by stack gas monitors. It should be mentioned that the circulating fluid bed has found only limited use for incinerator applications, primarily because emissions of heavy metals cannot be adequately controlled to meet many European require- ments. Addition of moisture to the flue gas, which could condense heavy metals, enlarges the fluid bed considerably, making spray drying more attractive for new sources. It is also possible that high CaCl2 concen- trations may complicate the bed dynamics considerably. As a retrofit process for ESPs, this system has several desirable features including: o Substantial acid gas removal upstream of the ESP, o Simple operation of small vessels, o No wet solids handling, and o Better utilization of sorbent than simple dry injection. Typical outlet emissions data^ for the variables studied are as follows: HC1 75 mg/m3 (90S removal) HF 0.06-0.13 mg/m3 (98-99% removal) Cd 0.3 mg/m3 4.2.2 Semi-Dry Scrubbing 10 Semi-dry scrubbing is also called spray dryer scrubbing, dry scrubbing (a misleading title), or wet/dry scrubbing. In this process,the sorbent is injected as a liquid or liquid slurry and the product collected as a dry solid. For the purposes of this discussion, it will be referred to as semi-dry scrubbing to differentiate among true dry scrubbing (dry sorbent injection plus dust collection) and true wet scrubbing. 4-13 ------- Semi-dry scrubbing has many variations: sorbent may be injected through liquid nozzles or rotary atomizers, the sorbent may be screw-fed or pneumatically blown in dry and rewetted by water-only nozzles, or it may be injected wet or dry into a fluidized bed with overhead water sprays. The ensuing discussion will focus upon one of these systems—the spray dryer absorbei—because it is the most common and successful. Figure 4-4 illustrates a typical spray drying process for incinerator applications. Lime 1s slaked, mixed with water, and then pumped as a slurry to a head tank, an option for spray drying. Depending on the inlet concen- tration of pollutants, slurry 1s metered into the spray absorber (shown with a rotary atomizer in Figure 4-4). Flue gas heat is sufficient to dry the slurry into a solid powder within the reactor vessel, and part of the solids are collected 1n the bottom of the absorber vessel while the remainder are collected In the particle collector. Recycle of solids back to the feed tank may be selected as an option if sorbent utilization is very low or higher removals of gaseous pollutants are desired. The control of this process is relatively simple for incinerator applications. The spray dryer outlet flue gas is controlled at temperatures well above its saturation value. This precludes any sorbent from contact- ing downstream surfaces as a wet powder leading to solids buildup. It also assures operation well above the dew points of any acid gases. Because of the presence of calcium chloride (a very hygroscopic, difficult-to-dry solid), temperatures are typically controlled at 110*-160'C (230-320'F) by limiting the amount of water injected. In view of the Environment Canada data15, it appears important that fine atomization of the feed slurry and good gas/slurry droplet mixing be attained to achieve reliable spray dryer operation at a temperature below 140*C (284"F). A second control loop is usually provided based on the pollutant emission levels in stack gases to regulate the addition of reagent to the system. Therefore with liquid and solids regulated by separate control loops, the solids composition of the slurry may vary somewhat. It is typ- ically low (~ 10 percent) for incinerator applications and easily managec In many cases, the sorbent rate is fixed at a conservatively high rate to ensure low stack emissions, but waste disposal plus sorbent operating costs are increased. Many additional details of operation of spray drying are given in References 10, 16, and 17. Compared with other systems, considerable emissions testing has been performed on both pilot- and full-scale spray dryer Installations. Outlet emission guarantees for spray dryer plus fabric filters and spray dryer plus ESP are as follows:10 TABLE 4-10. OUTLET EMISSION GUARANTEES FOR MUNICIPAL WASTE COMBUSTORS10 FGC System Spray Dryer + Fabric Filter Spray Dryer + ESP Flue Gas Component Concentration Removal,? Concentration Removal,% HC1 20-60 ppmv 88-99 20-60 ppmv 90-98 HF 2-5 ppmv 50-80 2-5 ppmv 50-80 S02 20-70 ppmv 67-90 50-175 ppmv 15-50 Particulate 0.013-0.02 gr/scf 0.015-0.04 gr/scf 4-14 ------- From these guarantees, performance data seem similar for either appli- cation except for S02 and participate. Results from these installations at 130-140'C (266-284*F) flue gas outlet temperatures show these similarities. However, when the heavy metal and dioxin/furan removals were examined at the pilot plant level, dramatic differences were noted. First, actual removal of acid gases and particles has substantially exceeded guaranteed levels as shown in a typical case using a fabric filter at 140*C (284*F): TABLE 4-11. ACID GAS EMISSIONS FROM MUNICIPAL WASTE COMBUSTORlO Acid Gas Guarantee, mg/Nm3 Actual Emissions, mg/Nm3 HC1 55 12-35 HF 5 0.2 S02 70 20-70 On a pilot plant, the following heavy metal removals were observed for ESP and fabric filter applications at the same flue gas conditions: TABLE 4-12. TRACE HEAVY METALS CONTROL BY PILOT PLANT FGC SYSTEMS^ Spray Dryer + ESP, % Spray Dryer + Fabric Filter. % Hg* 35-40 75-85 Pb 65-75 95-98 Cd 95-97 95-97 As 93-98 95-98 Particulate 99.2-99.4 99.8-99.9 aVapor only. Further, when dioxins and furans were measured in the pilot plant, from 48-89 percent of the dioxins and 64-85 percent of the furans were controlled by the spray dryer plus ESP. However, a spray dryer plus fabric filter im- proved removal in every case, usually by a substantial margin. In fact at 110'C (230*F), over 99 percent control of both dioxins and furans was observed out of the fabric filter. Recent Canadian data*5 (see Tables 4-3 through 4-9) on a slipstream pilot plant show high pollutant capture for a lime spray dryer/fabric filter. The removal of dioxins/furans approached 100 percent for a fabric filter inlet gas temperature of 140"C (284"F). S02 and HC1 removals were 60 and over 90 percent, respectively. Metals recovery was over 99 percent except for mercury which was about 95 percent at 140*C. The overall removal efficiencies for other trace organics in the MWC flue gas (chlorobenzenes, chlorophenols, polychlorinated biphenyls, and polycyclic aromatic hydro- carbons) were 99 percent or more at 140"C without recycle. With product recycle, the removal of polycyclic hydrocarbons and chlorophenols fell to 79 and 96 percent, respectively. The control efficiencies need to be verified on full-scale operating facilities. 4-15 ------- Tests at the Marion County, OR, MWC facility in 1986 were made to determine compliance with permit conditions. Results for units (boilers) 1 and 2 are shown in Table 4-13, and the discharge permit emission limits are shown in Table 4-14 so that a comparison of test results with the discharge permit emissions may be made.18 Each unit has a rated capacity of 275 tons/day (tpd) and is equipped with a semi-dry/dry scrubber/fabric filter system (see Figure 4-5). Lime slurry is fed to the quench reactor (spray dryer) and a cement kiln dust is supplied to the dry reactor (venturi) just upstream of of the fabric filter. Annual compliance tests for the Oregon Department of Environmental Quality are also slated at this facility in June 1987. Comparison of values in the preceding tables show that NOX emissions from the Marion County facility were about 15Z higher than that specified in its discharge permit. All of the other test results showed values lower than the permit limits. 4.3 Combination Scrubbers Several innovative processes have been recently developed spe- cifically for incinerator applications which utilize combinations of existing scrubber technology. These Include a dry/semi-dry process and wet/dry processes. 4.3.1 Semi-Dry/Dry Scrubbing19 Figure 4-5 depicts the seml-dry/dry scrubbing process which has been commercially operated In Japan and 1s currently installed on three U. S. facilities. The process is relatively simple, c.nsisting of a quench reactor, dry venturi reactor, and baghouse. The quench reactor is essen- tially an upflow spray dryer with multiple sprays of lime slurry used to ensure reliability. An upflow system is claimed to ensure against large droplets (due to sprayer malfunction) Impinging on downstream surfaces. The quenched gas then enters a venturi reactor where a dry powder of calcium silicate/lime composition 1s Introduced. This addition reportedly Increases the particle size of fly ash and sorbent such that the baghouse pressure loss (and hence need for bag cleaning) is minimized. This also results In the Hrne and calcium silicate sorbents being retained on the filters for very long times, reportedly up to eight hours. With the combination of a more reactive sorbent (calcium silicate) and very long retention time 1n the gas stream, high removals of acid gases would be expected. Reported data19 for a commercial system are summarized below: TABLE 4.13. POLLUTANT EMISSIONS AND CONTROL BY SEMI-DRY/DRY SCRUBBING19 Pollutant Emission Concentration Removal, % Particulate 0.0004-0.039 gr/dscf or 1-9 mg/Nm3 98.7 - 99.1 HC1 1-4 ppmv 95 (average) S0£ 0-5 ppmv Hg 85 g/Nm3 4-16 ------- FLUE GAS SOLIDS L*- LIME CALCIUM SILICATE AIR 1. QUENCH REACTOR (SPRAY DRYER) 2. DRY VENTURI 3. BAGHOUSE 4. STACK Figure 4-5. Semi-dry/dry scrubber. 19 ------- • .. *»vfliiiniM UI- i j i \j<\ 1 1. j i o ur I t-MDtK " ULIUtJLK AT THE MARION COUNTY, OR, WASTE-TO-ENERGY FACILITY^ Flue Gas from Boiler 1 Flue Gas from Boiler 2 Flue Gas In Main Stack oo Pollutant NOX S02 CO TSP Pb normal Pb bypassb Be TCDD VOC Fluorides Hg HCl Corrected TSP gr/dscf @ 12% C02 Avg Ib/hr 60.1 10.3 1.9 3.7 0.003 — 2.4x10-7 1.9x10-8 0.2 0.046 0.026 0.61 0.016 Other Requirement Max Ib/hr 69.0 12.8 2.0 4.7 — 0.02 2.7x10-7 2.3x10-8 0.2 0.071 0.73 0.021 Visible Em1 Avg Ib/hr 48.4 10.6 2.2 0.8 — — <2. 0x10-7 0.1 0.034 2.7 0.004 sslons Max Ib/hr 53.2 17.9 2.6 1.2 — — — — 0.2 2.9 0.006 0% Avg Ib/hr 108.5 20.9 4.1 4.5 0.006 <4.4xlO"7 < 3.8x10-8 0.3 0.092 0.060 3.3 091. 5%d tpy 435 83.8 16 18 0.02 cl. 8x10-6 1.5x10-7 1.2 0.37 0.24 13 Max Ib/hr 122.2 30.7 4.6 5.9 0.04 5.4x10-7 4.5x10-8 0.4 0.14 0.068 3.6 eiooz* tpy 535 134 20 25.8 0.2 2.4x10-6 2.0x10-7 1.8 0.61 0.30 16 aFac111ty availability. &Data based on one test of 5 minutes duration during which the baghouse was bypassed. ------- TABLE 4-15. PERMITTED EMISSIONS FROM THE MARION COUNTY, OR, WASTE-TO-ENERGY FACILITY18 Flue Gas from Boiler 1 Flue Gas from Boiler 2 Pollutant NOX S02 CO TSP Pb Be TCDD VOC Fluorides Hg HCl (Ib/hr) 47.0 36.5 27.5 10.0 00.26 1.45x10-6 8.5x10-7 1.65 0.8 0.085 < 11.5 (Ib/hr) 47.0 36.5 27.5 10.0 00.26 1.45x10-6 8.5x10-7 1.65 0.8 0.085 < 11.5 Other Requirement Emission Limits Main Stack for Flue Gas from Both Boilers (gr/dscf) (Ib/hr) at 12? CO?) (tpy) 94.0 73.0 55.0 20.0 00.52 2.9xlO-6 1.7x10-6 3.1 1.6 0.17 < 23 0.030 290 220 170 61 1.6 8.8xl016 5.1x1016 9.6 4.8 0.51 < 69 Opacity Visible Emissions 10Z 4-19 ------- Design conditions for this system are typically 90 percent HC1 removal, 80 percent S02 removal, and 0.008 gr/dscf particulate ratter emission. 4.3.2 Semi-Dry/Wet Scrubbing10 A flow sheet of this process is shown in Figure 4-6. As dis- cussed earlier, a wet scrubber system designed for multipollutant control will likely require separate treatment for HC1 and S02 if high removals of both are desired and liquid wastes are to be minimized. The primary feature of this process is separate scrubbing for HC1 and S02, and the significant innovation is location of an upstream spray dryer which disposes of liquid effluent from both wet scrubbers. Flue gas enters a spray chamber in which spent scrubber liquor is introduced through a series of nozzles, or a rotary atomizer. The reported retention time for gases is 4-5 seconds. Gas then enters a particulate collector (usually an ESP, but it is not clear that a fabric filter would not work as well) and then to the two-stage wet scrubbing system. The first stage is essentially an HC1 scrubber (described in Section 3.1) operated with water injection in a venturi column at a pH of between 0 and 1. Water injection is controlled by monitoring the solution conductivity, and the blowdown stream is neutralized with lime externally. A second option is to raise the pH to 3-4 and inject a sodium chlorite (NaC102) solution, which results .n oxidation of nitric oxide (NO) to nitro- gen dioxide (N02), described in Section 3.2.2. The second stage scrubber, also a venturi, is operated at a higher pH (5-6) with sodium hydroxide for S02 removal (and NOX removal if required). Both stages are designed to intentionally minimize internal surfaces to avoid scaling and erosion. Blowdown from the second stage is sent to a sludge holding tank prior to being sent back to the spray dryer. No plant is currently known to be operating with this system. However, several plants were in the design or construction stage at the end of 1986 and incorporate all features, including NOX removal in the wet scrubbers. Since no data for this system are available, the performance is assumed to be at least equivalent to that described for two-stage wet scrubbing:10 TABLE 4-16. £XP£CT£D POLLUTANT EMISSIONS FROM MWC EQUIPPED WITH SEMI-DRY/ WET SCWJBBINGiO Pollutant Emissions Concentration HC1 10 mg/Nm3 (8 ppmv) SO? 50 mg/Nm3 (25 ppmv) HP 0.5 mg/Nm3 (0.78 ppmv) Particulate 10 mg/Nm3 (0.0057 gr/dscf) Heavy Metals (group 1-3) 1 mg/Nm3 (0.00057 gr/dscf) Hg (vapor) 0.5 mg/Nm3 (0,000028 gr/dscf) 4-20 ------- 1. FLUE GAS 2. EXHAUST GAS 3. SPRAY DRYER 4. ELECTROSTATIC PRECIPlTATOR OR FABRIC FILTER 5. GAS GAS HEAT EXCHANGER 6. VENTURI SCRUBBER 7. NEUTRALIZATION TANK 8. SLUDGE TANK 9. LIME SILO 10. LIME SLAKER 11. SODIUM HYDROXIDE STORAGE 12. SODIUM AIR TANK 13. DRY WASTE Figure 4-6. Sem-dry/wet scrubber. 10 ------- As discussed in Section 3.2.2, NOX removal by oxidation/absorption has been practiced in Japan with 30-90 percent NOX removal when sodium or magnesium is the second stage scrubber reagent. These systems use chlorine dioxide, C102, rather than NaC102, however. The mixing of streams containing chlorides, fluorides, sulfites, sulfates and perhaps nitrates appears to be a complicating factor. Problems relating to corrosion are dealt with by using Has telloy steel and rubber linings in critical scrubber components. However, when the mixed streams are handled, co-precipitation (scale) due to saturation must be addressed. Also, the amount of water to be spray dried is limited by the evaporative capacity of the flue gas and the minimum approach to saturation desired. Thus, waste liquor must necessarily be very concentrated with respect to dissolved and suspended solids. 4-22 ------- 5.0 EFFECTIVENESS OF FLUE GAS CLEANING METHODS 5.1 Participate Matter Control Participate matter control for Incinerator flue gas applica- tions 1s a simple matter of choice of control device and proper operation of that device. Wet scrubbers are relatively Ineffective for particle control, removing 80 to 95 percent at normal operating ranges. Very high pressure losses are required to remove fine particles, and the erosion and corrosion potential 1n acidic gas streams makes this a poor choice from economic and reliability standpoints. Electrostatic precipitators are the most widely used and are the most versatile control systems. Very low emission levels are achievable (<0.02 gr/dscf) at high ratios of collector plate surface area to gas flow volume in the range of 500 min./ft or greater. Fabric filters are seldom used without upstream sorbent injection due to a potential for fires and/or blinding by sticky particles. However, fabric filters are also capable of control to less than 0.02 gr/dscf and are less sensitive to operational upsets that disrupt ESP performance. 5.2 Acid Gas Control Control of acid gases (HC1, HF, and $03) requires scrubbing or devices for gas/liquid or gas/solid contact. Water alone is a reasonably effective sorbent for very reactive acid gases such as HC1 and HF, but an alkali sorbent (or control of liquid pH to the 5 or greater range) is necessary for substantial S02 control. Totally dry sorbents require sub- stantial residence time in the gas for effective acid gas control. Injec- tion of sorbent into a duct must be complemented by either a fluid bed reactor, humidiflcation, a fabric filter dust collector, or combinations of these to be effective. Spray drying or semi-dry Injection of sorbent is more effective than dry injection, with Increasing acid gas control as the approach to satura- tion temperature is decreased, either by waste heat recovery or water Injection/humidificatlon. The most effective control of acid gases is by alkali scrubbers operating at saturation, or wet scrubbing, but this has to be weighed against the amounts of waste water generated. Pilot plant Canadian data show dry lime injection to be effective for removing over 90S of the inlet HC1 for a fabric filter inlet tempera- ture of 140*C (284*F) or less. In the same plant, the S02 removal was over 90% for 125'C (257'F), but only 58% at the 140*C filter inlet temp- erature. The stoichiometrie ratio was about 1.1 (based on both HC1 and S02) in these tests. Combination dry, semi-dry scrubbers control add gases perhaps more effectively than once-through spray drying and are probably similar in effec- tiveness to spray drying with recycle, depending on approach to saturation 5-1 ------- TABLE 5-1. EFFECTIVENESS OF ACID GAS CONTROLS (% REMOVAL) Pollutant Control System HC1 HF SQ2 Dry Injection + Fabric Filter3 80 98 50 Dry Injection + Fluid Bed Reactor/ESP^ 90 99 60 Spray Dryer-ESP 95+ 99 50-70 (Recycle)c (95+) (99) (70-90) Spray Dryer-Fabric Filter 95+ 99 70-90 (Recycle)C (95+) (99) (80-95) Dry/Spray Dryerd 95+ 99 90+ Wet Scrubber6 95+ 99 90+ Wet/Dry Scrubber6 95+ 99 90+ a T = 160-180'C (320-356'F) b T = 230*C (446'F) C T = 140-160'C (284-320*F) d T = 200'C (392*F) e T = 40-50*C (104-122'F) T is the temperature at the exit of the control device. 5-2 ------- temperature. Combination wet-dry systems are the potentially most effective system for add gas control but are Increasingly complex as the number of pollutants targeted Increase. Table 5-1 summarizes the above discussion. The reader 1s cautioned that the reagent requirements and solid/liquid wastes are not factored 1n, and this table only reflects systems as oper- ated. Any of these techniques may be enhanced by more reactive sorbents or operation at more favorable temperatures. In summary, effective acid gas control is possible with dry, semi-dry, and wet scrubbers. HC1 and HF are relatively easy to control, while S02 control 1s more difficult and 1s favored by wet or semi-dry systems with lower flue gas temperatures. Although not discussed due to lack of data, very effective sulfur trioxide control was reported on a spray dryer pilot plant. Should $03 control also become a concern, systems which contact the gas with wet or dry sorbent prior to a particulate control device should be encouraged, since after scrubbing, $03 apparently becomes an aerosol and 1s amenable to capture. Control systems with particle collectors upstream of the scrubber have historically reported poor $03 control effectiveness. 5.3 Post Combustion NOX Control Probably the most difficult and expensive pollutant to control 1s NOX, primarily due to the unreactivity of NO which is 95 percent or more of the total uncontrolled NOX. The most effective control is selective, catalytic reduction (SCR) which currently must be preceded by acid gas and heavy metals control to be effective. (However, a low temperature, acid-resistant catalyst has recently been applied to three small municipal incinerators 1n Japan, but performance data are not yet available.) If the thermal penalties are accept- able, then SCR can remove 80-90 percent of NOX with a NH3/NO molar ratio of 1.0 and about 5 ppmv NH3 slip. Use of special lower temperature, HCl-resistant catalysts in the future can make SCR more attractive. Potentially less effec- tive and more complicated NOX control may be .chieved by an oxidation step integrated into sodium- or magnesium-based wet scrubbing. Due to the liquid waste potential, this may be best applied to the combination wet-dry scrubber system described in Section 4.3.2. Using SCR, a NOX control of 30 to 50 percent would be expec*ed. 5.4 Post Combustion Organic Pollutant Control Control of dioxins and furans, as well as other trace organic compounds, Is not well understood because the mechanism of capture is not known. Likely, condensation and capture as a particle is significant, and attack and capture by caustic reagents is also probable. These capture phe- nomena are best addressed by lowering flue gas temperatures, subjecting the VOCs 1n the flue gas to caustic sorbent, and collecting the product on a highly efficient particle collector. Limited data (mostly for pilot plants) show that spray drying followed by fabric filtration is very effective for YOC control and superior to spray dryer/ESP control. Also lower flue gas temper- atures favor Increased YOC control. Reference 13 is a good discussion of 5-3 ------- these observations. The results are summarized In Table 5-2, where COD refers to chlorinated d1benzo-paradioxlns and CDF to chlorinated dibenzofurans. TABLE 5-2. SPRAY DRYER CONTROL OF SELECTED ORGANIC POLLUTANTS17 Control System (% Removal) Compound D1oxlns: tetra CDD penta CDD hexa CDD hepta CDD octa CDD Furans; tetra CDF penta CDF hexa CDF hepta CDF octa CDF SD + ESP SD + FF @ High Temp. 48 51 73 83 89 65 64 82 83 85 <52 75 93 82 NA 98 88 86 92 NA SD + FF 0 Low Temp. >97 >99.6 >99.5 >99.6 >99.8 >99.4 >99.6 >99.7 >99.8 >99.8 Reference 7 notes that only limited data have been collected on control device efficiencies for dioxins and furans, with only outlet concentrations being reported for most tests. Unfortunately, test data and methodologies are lacking to compare the effectiveness of various control systems on organic pollutants. However, the superiority of a sorbent on a fabric filter for control is evident from Table 5-2. The data shown were based on tests in a single pilot plant, and thus should be used with caution. The results of Canadian tests, based on an Incinerator flue gas slip- stream and the use of dry lime Injection (110-140*C) and lime spray drying (140"C), each with a downstream fabric filter for dust collection, show high overall removal efficiencies (99+Z) for dioxins and furans for the fabric filter Inlet temperatures noted. With dry lime injection, the removal of other organlcs (chlorobenzenes, chlorophenols, and polychlorinated biphenyls) fell markedly at 200'C compared with inlet fabric filter temperatures of 110-140'C (230-284'F). 5.5 Heavy Metals Control The control of heavy metals 1s similar to organic pollutant control in that the effective control of particles and low flue gas temperatures are major factors. Sorbents, however, are not suspected to play a major role. 5-4 ------- Toxic metals enter the collectors as solids, liquids, and vapors, and as the flue gas cools, the vapor portion converts to collectible solids and liquids. Figure 5-1 illustrates various heavy metals as they appear in flue gas and their relative theoretical concentrations (vapor pressures) as a function of flue gas temperature. From Figure 5-1, it can be deduced that reduction of flue gas tempera- tures below 200°C (392°F) and high efficiency particulate collection should result in a very large reduction of metals, except for mercury (Hg), arsenates (As2d3)2, and selenium (Se02 and See). Corresponding reductions of these compounds proceed dramatically as temperatures are lowered. With the metals at there saturation temperatures, each is expected to be reduced by 90 percent for each additional temperature drop of 11 to 17*C (20 to 30*F). If this temperature effect is true, then wet scrubbing or wet/dry scrubbing which operates at saturation [ <\, 40"C (104'F)] will be most effective for total heavy metals control, while most dry and semi-dry systems will be effective for practically all metals except mercury, arsenic, and selenium. Reported metals control data generany show 95-98 percent control or greater for most heavy metals except mercury. Vapor phase mercury control has been reported as follows: 75 to 85 percent control with spray dryer plus baghouse; 35 to 45 percent control with spray dryer plus ESP.1^ This is im- portant in that vapor control is possible with fabric filters and ESPs, although limited data show the former to be clearly superior. Wet scrubbers would appear to be ideal for mercury control, but the collection of mercury vapors via condensation and capture is not well documented. Therefore, the choice of the most effective mercury control is still the subject of contro- versy (see Reference 10). 5-5 ------- ttOOOO 10000 Concentration mg/rn^ Measured He, Concentrations hi raw fas Figure 5-1. Saturation points of metal and metal compounds. 16 5-6 ------- 6.0 OPERATION AND MAINTENANCE OF FLUE GAS CLEANING SYSTEM This section is intended to summarize good operating and main- tenance (04M) practices for flue gas controls typically applied to municipal waste incinerators. Advanced controls, such as selective catalytic reduction and wet oxidation-absorption NOX control, are not addressed here, as exper- ience with these systems is too limited to warrant a discussion. 6.1 Electrostatic Precipitators9 Poor performance of an electrostatic precipitator can be divided into fundamental problems, mechanical problems, and operational problems. Fundamental problems include design inadequacies such as poor gas flow distri- bution, inadequate collector area, or unstable energization equipment. These are best addressed by replacement or redesign of problem areas and are inde- pendent of the 0AM program. Mechanical problems include electrode misalignment, wire breakage, cracked collector surfaces, air inleakage, cracked insulators, plugged hoppers, and dust deposits. Defective components should be replaced once the cause of the problem has been identified. Operational problems, which are those last addressed by O&M programs, include process upsets, inadequate power input, electrical problems, rapper failures, and dust removal valve failures. A good performance monitoring program for ESPs is recommended, which includes measurement of key operating parameters, performance tests, and monitoring and recordkeeping of key operating parameters. These parameters include gas volume flow, velocity, and temperature; chemical composition of gases and particles; particle concentrations, size distribution, and resis- tivity; and power input to the ESP. Use of on-line instrumentation (such as voltage and current meters, spark meters, rapper monitors, transmissometers, and hopper level indicators) are essential for proper ESP operation. Periodic performance testing for particulate concentrations, such as Reference Methods 5 or 17 and Method 9, are useful tools in evaluating long-term or gradual effects not easily monitored. O&M practices include development of procedures for startup, shutdown, and routine operation. Inspections on a daily, weekly, and annual basis are recommended as shown in Table 6-1. Inspection procedures and detailed 0AM guidance for ESPs may be found in Reference 9. 6-1 ------- TABLE 6-1. INSPECTIONS FOR ESP9 Dail; o Corona power levels (i.e., primary current, primary voltage, secondary current, secondary voltage) by field and chamber, (twice a shift) o Process operating conditions [i.e., firing rates, steam flow or load (Ib/h), flue gas temperature, flue gas oxygen, etc.]. The normal operator's log may serve this purpose, (hourly) o Rapper conditions (i.e., rappers out, rapper sequence, rapper intensity, rapping frequency by field and chamber). o Dust discharge system (conveyors, air locks, valves for proper operation, hopper levels, wet-bottom liquor levels). o Opacity (i.e., absolute value of current 6-minute average and range of magnitude of rapper spiking) for each chamber duct if feasible. (2-hour intervals) o Abnormal operating conditions (i.e., bus duct arcing, T-R set control problems, T-R set trips excessive sparking), (twice a shift) o Audible air inleakage (i.e., location and severity). Weekly; o Trends analysis (plot gas load V-I curves for each field and chamber and other key parameters to check for changes in values as compared with baseline). o Check and clean or replace T-R set cabinet air filters and insulator purge air and heating system filters. o Audible air inleakage (i.e., location and severity). o Abnormal conditions (i.e., bus duct arcing, penthouse and shell heat systems, insulator heaters, T-R set oil levels, and temperature). o Flue gas conditions exiting the ESP (i.e., temperature and oxygen content). o More extensive rapper checks (also optimize rapper operation if needed). 6-2 ------- TABLE 6-1 (Cont'd) Annually: o Transformer Enclosure HV line, insulators, bushings, and terminals Electrical connections Broken surge arresters o High-Voltage Bus Duct Corrosion of duct Wall and post insulators Electrical connections o Penthouse, Rappers, Vibrators Upper rapper rod alignment Rapper rod insulators Ash accumulation Insulator clamps Lower rapper rod alignment Support insulator heaters Dust in penthouse area Corrosion in penthouse area Water inleakage HV connections HV support insulators Rapper rod insulator alignment o Collecting Surface Anvil Beam Hanger rods Ash buildup Weld between anvil beam and lower rapper rod o Upper Discharge Electrode Frame Assembly Welds between hanger pipe and hanger frame Discharge frame support bolts Support beam welds Upper frame levelness and alignment to gas stream o Lower Discharge Electrode Frame Assembly Weight guide rings Levelness of frame Distortion of the frame 6-3 ------- TABLE 6-1 (Cont'd) o Stabilization Insulators Dust buildup and electrical tracking Broken Insulators o Collecting Electrodes Dust deposits; location and amount Plate alignment Plate plumbness Plate warpage o Discharge Electrode Assembly Location of dust buildup and amount Broken wires Wire alignment Weight alignment and movement o Hoppers Dust buildup Level detectors Heaters Vibrators Chain wear, tightness, and alignment Dust buildup in corners and walls o Dust Discharge System Condition of valves, air locks, conveyors o General Corrosion Interlocks Ground system Turning vanes, distribution plates, and ductwork 6-4 ------- 6.2 Fabric Filters8 Poor performance of fabric filters can be categorized by (1) problems that affect all fabric filters, regardless of type, and (2) problems that are characteristic of a particular cleaning system design. The first category includes fabric failure, dust discharge problems, corrosion, and improper maintenance considerations. Fabric failures that occur Immediately after the baghouse goes on line are generally caused by improper Installation or manufacturing defects. With proper design and operation, these failures are usually isolated in early stages of operation. Generally reverse-air and shaker-type fabric filters are more prone to initial failures than pulse-jet systems. Other fabric failures may be caused by high temperatures, condensation, chemical degrada- tion, too high air-to-cloth ratio, high pressure drops, and bag abrasion. Dust discharge failures generally result from cool spots 1n the dust hopper, air leakage, or failure (either mechanical or human error) to keep the dust level in the hopper manageable. Also bag cleaning systems have failures generic to the system type. Failure to clean bags properly results in increased pressure drop, bag failure, and reduced bag life. An effective monitoring program that Includes performance evaluation and data collection is a necessity. The most important parameters to record and evaluate are the opacity of exhaust gases and pressure drop across an individual compartment or the entire baghouse. As a general rule, these are checked daily to determine if operation is within the normal range for that system. 04M procedures include established startup, operating, and shutdown procedures which emphasize avoiding dew point conditions through cold gas bypass, auxiliary heat, and system purges. Preventive maintenance practices include periodic inspections of components as shown in Tables 6-2 through 6-4. Inspection procedures and detailed 04M guidance for fabric filters may be found in Reference 8. 6.3 Scrubbers Although wet scrubbers with liquid wastes are not expected to be Installed in quantity, operation and maintenance procedures are presented briefly here. The more likely to be used dry scrubber systems are discussed separately from wet scrubbers. Considerations are based on calcium-based scrubbing for both systems, sodium-based systems being unlikely to be used because of soluble waste disposal restrictions and cost considerations.^ 6-5 ------- TABLE 6-2. TYPICAL MAINTENANCE INSPECTION SCHEDULE FOR A FABRIC FILTER SYSTEMS Inspection frequency Component Procedure Dally Stack and opacity monitor Manometer Compressed air system Collector Weekly Damper valves Rotating equipment and drives Dust removal system Filter bags Cleaning system Hoppers Check exhaust for visible dust. Check and record fabric pressure loss and fan static pressure. Watch for trends. Check for air leakage (low pressure). Check valves. Observe all indicators on control panel and listen to system for properly operating subsystems. Check all isolation, bypass, and cleaning damper valves for synchronization and proper operation. Check for signs of jamming, leakage, broken parts, wear, etc. Check to ensure that dust is being removed from the system. Check for tears, holes, abrasion, proper fastening, bag tension, dust accumulation on surface or in creases and folds. Check cleaning sequence and cycle times for proper valve and timer operation. Check compressed air lines including oilers and filters. Inspect shaker mechanisms for proper operation. Check for bridging or plugging. Inspect screw conveyor for proper operation and lubrication. 6-6 ------- TABLE 6-2 (Cont'd) Inspection frequency Component Procedure Monthly "uarterly Shaker mechanism Fan(s) Monitor(s) Inlet plenum Access doors Shaker mechanism Semi- annually Annually Motors, fans, etc, Collector Inspect for loose bolts. Check for corrosion and material buildup and check Y-belt drives and chains for proper tension and wear. Check accuracy of all indicating equipment. Check baffle plate for wear; if appreciable wear is evident, replace. Check for dust deposits. Check all gaskets. Tube type (tube hooks suspended from a tubular assembly): Inspect nylon bushings in shaker bars and clevis (hanger) assembly for wear. Channel shakers (tube hooks suspended from a channel bar assembly): Inspect drill bushings in tie bars, shaker bars, and connecting rods for wear. Lubricate all electric motors, speed reducers, exhaust and reverse-air fans, and similar equipment. Check all bolts and welds. Inspect entire collector thoroughly, clean, and touch up paint where necessary. 6-7 ------- TABLE 6-3. ROUTINE INSPECTION DATA FOR REVERSE AIR FABRIC FILTERS8 o Routine Inspection Data Stack Fabric Filter Average Opacity Opacity During the Cleaning Cycles (for each compartment) Inlet and Outlet Static Pressures Inlet Gas Temperature Rate of Dust Discharge (Qualitative Evaluation) Presence or Absence of Audible Air Infiltration Presence or Absence of Clean Side Deposits Ripping Strength of Discarded Bags o Baseline and Diagnostic Inspection Data Stack Fabric Filter Stack Test Fan Average Opacity Opacity During the Cleaning Cycles (for each compartment) Date of Compartment Rebagging Inlet Static Pressure (Average) Outlet Static Pressure (Average) Minimum, Average, and Maximum Gas Inlet Temperatures Average 02 and C02 Concentrations (Combustion Sources Only) Time to Complete a Cleaning Cycle of all Compartments Length of Shake Period Length of Null Period Bag Tension (Qualitative Evaluation) Rate of Dust Discharge (Qualitative Evaluation) Presence or Absence of Audible Air Infiltration Presence or Absence of Clean Side Deposits Emission Rate Gas Flow Rate Stack Temperature 02 and C02 Content Moisture Content Fan Speed Fan Motor Current Gas Inlet and Outlet Temperatures Damper Position 6-8 ------- TABLE 6-4. ROUTINE INSPECTION DATA FOR PULSE-OET FABRIC FILTERS^ o Routine Inspection Data Stack Fan Fabric Filter Average Opacity Duration and Timing of Puffs None Inlet and Outlet Gas Temperatures Inlet and Outlet Static Pressures Presence or Absence of Clean Side Deposits Air Reservoir Pressure Audible Checks for Air Inleakage Qualitative Solids Discharge Rate o Diagnostic Inspection Data Stack Fan Fabric Filter Average Opacity Peak Opacity During Puffs Duration and Timing of Puffs Inlet Gas Temperature Speed Damper Position Motor Current Inlet Gas Temperature Outlet Gas Temperature Inlet Static Pressure Outlet Static Pressure Inlet 02 and C0£ Content (Combustion Sources) Outlet 02 and C02 Content (Combustion Sources) Qualitative Solids Discharge Rate Air Reservoir Pressure Frequency of Cleaning Presence or Absence of Clean Side Deposits Audible Air Infiltration 6-9 ------- 6.3.1 Lime/Limestone Wet Scrubbing*3 Extensive surveys of lime/limestone wet scrubbing technology have shown the eight major areas of failure listed below in order of decreasing likelihood of failure: o mist eliminators o ductwork o absorber o stack o fans o pipes and valves o thickener o dampers Design considerations allow ease in addressing these problems but are too detailed to be presented here. They are discussed extensively in Reference 13. Inspection procedures for each major unit operation are also detailed in this reference. 04M practices, preventive maintenance procedures, and unscheduled maintenance procedures are presented in detail in the following discussion. Standard 04M procedures for lime/limestone scrubbers include: o Variable Load Operation. With each change in load, the operator must check the system to verify that all in-service modules are operating in a balanced condition. As the acid gas concentrations in the inlet flue gas change, the FGD system should be able to accommodate and compensate for such change. Operator surveillance of system performance is needed, however, to verify proper system response (e.g., slurry recirculation umps can be added and removed from service as the acid gas concentration increases or decreases). o Verification of Flow Rates. The easiest method of verifying liquid flow rates is for an operator to determine the discharge pressure in the slurry recirculation spray header with a hand-held pressure gauge (permanently mounted pressure gauges frequently plug in slurry service). Flow 1n slurry piping can be checked by touching the pipe. If the piping 1s cold to the touch at the normal operating temperature of 52-54*C (125-130*F), the line may be plugged. o Routine Surveillance of Operation. Visual inspection of the absorbers and reaction tanks can identify scaling, corrosion, or erosion before they seriously impact the operation of the system. Visual observation can identify leaks, accumulation of liquid or scale around process piping, or discoloration on the ductwork surface resulting from inadequate or deteriorated lining material. 6-10 ------- o M1st Eliminators. Many techniques have been employed to Improve mist collection and minimize operational problems. The mist eliminator can be washed with a spray of process makeup water or a mixture of makeup water and thickener overflow water. Successful long-term operation without mist eliminator plugging generally requires continuous operator surveillance, both to check the differential pressure across the mist eliminator section and to visually inspect the appearance of the blade surface during shut- down periods. o Reheaters. Inline reheaters are frequently subject to corrosion by chlorides and sulfates. Plugging and deposition can also occur but are more rare. Usually, proper use of soot blowers prevents these problems. o Reagent Preparation. Operational procedures associated with handling and storage of solid reagent are generic to solids handling. Operation of pumps, valves, and piping in the slurry preparation equipment 1s similar to that in other slurry service. o Pumps, Pipes, and Valves. Operating experience has shown that pumps, pipes, and valves can be significant sources of trouble in the abrasive and corrosive environments of lime/limestone FGD systems. The flow streams of greatest concern are the reagent feed slurry, the slurry recirculation loop, and the slurry bleed streams. When equipment is temporarily removed from slurry service, it must be thoroughly flushed. o Thickeners. Considerable operator surveillance is required to minimize the suspended solids in the thickener overflow so that this liquid can be recycled to the system as supplementary pump seal water, mist eliminator wash water, or slurry preparation water. For optimum performance, the operator must maintain surveillance of such parameters as underflow slurry density, flocculant feed rate, inlet slurry characteristics, and turbidity of the overflow. o Waste Disposal. For untreated waste slurry disposal, operation of both the discharge to the pond and the return water equipment requires attention of the operating staff. In addition to normal operations, the pond site must be monitored periodically for proper water level, embankment damage, and security for protection of the public. Landfill disposal Involves the operation of secondary dewatering equipment. Again, when any of the process equipment is temporarily removed from service, it must be flushed and cleaned to prevent deposition of waste solids. For waste treatment (stabilization or fixation), personnel are required to operate the equipment and to maintain proper process chemistry. 6-11 ------- o Process Instrumentation and Controls. Operation of the FGD system requires more of the operating staff than surveillance of automated control loops and attention to indicator readouts on a control panel. Manual control and operator response to manual data indication are often more reliable than automatic control systems and are often needed to prevent failure of the control system. Many problems can be prevented when an operator can effectively Integrate manual with automated control techniques. Preventive Maintenance Programs. Preventive maintenance is the practice of maintaining system components in such a way as to prevent malfunctions during periods of operation and to extend the life of the equipment. The goal of preventive maintenance is to increase availability of the FGD system by eliminating the need for emergency repair ("reactive maintenance"). The term preventive maintenance is synonymous with periodic maintenance. Such procedures may be as simple as lubrication of a pump or as complex as complete disassembly for inspection and overhaul. Some of the more important preventive maintenance procedures by subsystems are summarized in the following sections: o Absorbers. Of primary concern in the absorber module is the integrity of the structural materials. Maintenance personnel should enter and inspect the absorber module at least semi-annually. o Mist Eliminators. Scale deposits typically are the chief maintenance factor with mist eliminators. The mist eliminator may be subject to nonuniform flow or a faulty wash system. Wash spray pressure should be monitored. Mist eliminators should be inspected during forced or scheduled outages. o Reheaters. Both inline and indirect reheaters are subject to scaling and corrosion. In addition to visual inspection, pressure testing and measurement of heat transfer efficiency are useful in quantifying the magnitude of a reheater problem. In a direct reheat system, the mixing chamber and the air heating equipment must be checked routinely. o Dampers. Fans. Ductwork, and Chimneys. All points in the system must be checked for integrity of lining materials and for damage resulting from collection of condensation products in stagnant air spaces (e.g., duct elbows and corners). Components located in the wet portion of the system are subject to scaling and corrosion. Upstream fans and ductwork may be subjected to erosion. 6-12 ------- o Reagent Preparation. Reagent preparation subjects the ball mill or slaker to abrasive wear. Because the equipment sees intermittent service, it should be inspected visually each time it is placed in service. Annual disassembly is also needed to check for excessive wear. o Reagent Feed. Maintenance of the reagent slurry feed system is critical because failure of this equipment strongly impacts the FGD system operation. The slurry storage tank should be checked daily for leakage and associated equipment inspected for proper operation. o Pumps, Pipes, and Valves. Slurry pumps are normally disassembled at least annually.The purpose of the inspection is to verify lining integrity and to detect wear and corrosion or other signs of potential failure. Bearings and seals are checked but not necessarily replaced. Pipelines also must be periodically dis- assembled or tested in other ways (e.g., hand-held nuclear and ultrasonic devices) both for solids deposition and for wear. Valves must be serviced routinely, especially control valves. o Thickeners. Thickener coatings should be inspected periodically to prevent corrosion. Drag rakes, torque arms, and support cables must also be inspected for wear. o Waste Disposal Equipment. Secondary dewatering devices, mixing components, and transport equipment must also have periodic main- tenance to check for abrasive wear and solids deposition. Vacuum filters, both drum and belt type, require periodic replacement of the filter media. In a centrifuge, both the scroll coating and the bowl surfaces are subject to wear. o Process Instruments and Controls. All electronic equipment (pH, flow, pressure, temperature, level, vibration, noise, and continuous monitors) must be calibrated periodically. Numerous installation and maintenance techniques have proved beneficial in ensuring the reliability of sensors. Ease of access to the sensors is very important. The sensors should be cleaned and calibrated routinely. Experience with process instrumentation and controls in FGD systems has shown that a good preventive maintenance program begins with daily operating procedures. Proper use of instruments will include daily flushing of most instrument lines in slurry service just before monitoring of process variables. Routine comparison of the instruments in a process stream with similar instruments in parallel streams can point out incipient failures. Operating data, especially from the startup test program, can also indicate potential problem areas. 6-13 ------- Unscheduled Maintenance. Even the most rigorous preventive maintenance program will not prevent random failures to which the maintenance staff must respond. Most malfunctions are correctable by unscheduled (reactive) maintenance. In some situations, usually during initial system startup, design modifications may be required to bring the system into compliance with operating standards. Each subsystem of the FGD system is subject to malfunctions from a variety of causes. The discussion that follows Introduces these problems and the probable responses. o Absorbers. Structural failure of absorber internals and recycle pump suction screens have occurred as a result of excessive vibra- tion, uncorrected corrosion damage, or high pressure differentials. These malfunctions must be repaired Immediately before operation is resumed. o Mist Eliminators. Failure of the mist eliminator is typically due to scaling and plugging. The scale may be removed either by thorough washing or by mechanical methods, in which maintenance personnel enter the absorber and manually chip away the scale deposits. o Reheaters. Reheater malfunctions include tube failures in inline reheaters, damper problems in bypass reheat, or nonuniform flows in indirect reheaters. Correction of these problems • 11 probably necessitate changes in equipment design. o Fans. Fans can develop vibrations resulting from deposition of scale in wet service or from erosion of blades in dry service. The cause of the vibration must be eliminated and the fan repaired and rebalanced. o Ductwork. Most problems associated with ducts develop over a long period. Sudden or gross failures, such as a major leak, call for immediate repair. Temporary repair or patching may suffice until the next scheduled outage. Acid condensation in a chimney can cause lining deterioration and subsequent damage to the base metal. These problems are usually Identified during preventive maintenance inspections and require long-term solutions. o Reagent Feed. Malfunctioning components such as slakers must be repaired in accordance with the manufacturer's instructions. Some facilities have experienced trouble with plugging of the lime/lime- stone feeder due to intrusion of moisture. Correction of these problems will probably necessitate changes in equipment design. o Pumps. Pipes, and Valves. Excessive wear of the impeller or separation of the lining from the pump casing is a common problem. Operation of a slurry pipeline with Insufficient flow velocity can cause clogging. High flow velocity or extended service can cause erosion. Malfunction and binding of a valve actuator are typically caused by wear-induced misalignment. 6-14 ------- o Thickeners. The thickener underflow can become plugged because of excessive solids in the slurry or failure of the underflow pump. A plugged underflow or rapidly settling waste solids will produce a heavy "blanket" in the bottom of the thickener. The rake must then be raised so that the torque remains within acceptable limits. If the torque cannot be kept within limits, the thickener must be drained and the sludge blanket removed manually. 6.3.2 Semi-dry Scrubbing20 Although little information is published on OiM of dry or semi-dry scrubbing systems, some vendor-specific information has been compiled and condensed into a general guide for the spray dryer/absorber unit operation. Keep in mind that the particle collector is an integral part of the overall system and information in 6.1 or 6.2 should be added to the guidelines below in developing an integrated OiM program. The following operations are key OiM considerations in spray dryer operation: Temperature Measurements - A key operating parameter is the approach to adiabatic saturation temperature. The "approach" is the difference between the dry and wet bulb temperatures at the exit of the spray dryer (or fabric filter as appropriate). While instrumentation is available to monitor the wet bulb temperature, the reliability of such instrumentation is not as high as desired, and frequent manual wet bulb measurements are recommended for systems operated close to the adiabatic saturation temperature. Also, the spray dryer outlet thermocouples should be regularly checked for proper calibration. Slurry Preparation. S02 removal performance depends greatly on lime reactivity.While most lime vendors produce a consistently high quality product, the quicklime reactivity should be checked frequently using the ASTM C110-10, 3-minute temperature rise test. The slaker outlet tempera- ture should also be checked routinely to verify exit temperatures in the desired range of 77-89*C (170*F-190'F). This is a good indicator of proper slaker operation, suitable quicklime reactivity, and proper slaker exit solids content in the product slurry. The lime and atomizer feed preparation systems require handling of slurries with high solids concentrations. Settling of solids can lead to buildup and pluggage of the handling equipment. Routine checks should be made for proper mixing of solids in the slurry and for buildup of solids 1n piping, on tank walls, etc. In particular, screens or strainers need to be checked frequently for pluggage. 6-15 ------- The atomizer feed slurry solids content can be an important parameter for assuring proper solids drying and to assure that maximum recycle ratios are being employed. The solids contents are generally continuously measured by and controlled to some type of density meter reading. These meters are notorious for calibration drift and should be checked for cali- bration regularly. Some systems installed on utility boilers check density meter calibration as often as once per day. Atomizers. If rotary atomizers are used in the spray system, the atomizer drive system requires continuous monitoring. The atomizer drive lubrication and cooling system should be continuously monitored and maintained according to the manufacturer's instructions. The vibrations produced by the drive system should also be monitored either continuously or at regular intervals. Improper operation of the atomizer drive system at normal operating speed can result in immediate and usually catastrophic failure. The atomizer should always be dynamically balanced prior to use. The atomizer wheel or disk should also be periodically inspected for nozzle blockage, solids buildup, and excessive nozzle surface wear. Two-fluid nozzle or pressure atomization is mechanically a much simpler process than rotary atomization and requires less rigorous monitoring. Nozzles should be periodically inspected for blockage, wear, and corrosion. Nozzles should be properly positioned inside the reactor vessel, and all fittings should be free of leaks. Liquid and air pressure should be monitored and kept at the appropriate levels to assure proper atomization. The spray dryer vessel should be periodically inspected to ensure that a nozzle or several nozzles are not partially plugged and producing a coarse spray which has caused a buildup of wet solids on vessel walls. 6-16 ------- REFERENCES 1. Municipal Waste Combustion Study - Data Gathering Phase. Preliminary Draft. Radian Corporation, October 16. 1985, Part I, Chapter 4.0. 2. Municipal Haste Combustion Study; Combustion Control of Organic Emissions!EPA/530-SW-87-Q21C, June 1987. 3. DuBard, J. L.. Southern Research Institute. Private Communication to Dale L. Harmon, EPA/AEERL, October 8, 1986. 4. DuBard, J. L., Southern Research Institute. Private Communication to Dale L. Harmon, EPA/AIERI, October 28, 1986. 5. Clarke, M. J., "Emission Control Technologies for Resource Recovery." Presented at the Symposium on Environmental Pollution in the Urban Area, Booklyn, NY, Subsection of the American Chemical Society, March 15, 1986. 6. Review and Update of PM Emissions Database for Solid Waste-Fired Industrial Boilers. Memorandum. Radian Corporation, August 29, 1986. 7. Municipal Waste Combustion Study; Emission Data Base for Municipal Waste Combustors. EPA/530-SW-87-021b. June 1987. 8. Operation and Maintenance Manual for Fabric Filters. EPA-625/1-86-020, June 1986, pp. 2-1 through 2-17. 9. Operation and Maintenance Manual for Electrostatic Preci pita tors. -625/1-85-017, September 1985, pp, 5-1 through 6-72. Ope re EPA-( 10. Assessment of Flue Gas Cleaning Technology for Municipal Waste Combustion. Draft Report. Acurex Corporation, September 1986. 11. Ando, J. Recent Developments 1n S02 and NOX Abatement Technology for Stationary Sources in Japan, JJraft Report, (to be published as an EPA report in 1987). 12. Ellison, William H. "Status of German FGD and DeNOx." Presented at the Third Annual Pittsburgh Coal Conference, Pittsburgh, PA, September 9, 1986. 13. Flue Gas Desulfurizatipn Inspection and Performance Evaluation. EPA-625/1-85-019, October 1985, pp. 99-222. 14. Scrubber-Adsorber Newsletter. Mcllvalne Co., Northbrook, IL, July 30, 1986, No. 145, pp 3-6. 15. The National Incinerator Testing and Evaluation Program: Air Pollution Control Techno!oqylEnvironment Canada, Ottawa, Ontario.Report EPS 3/UP/2, September 1986. R-l ------- 16. Moeller, J. T., Jorgensen, C., and Fallenkamp, F., "Dry Scrubbing of Toxic Incinerator Flue Gas by Spray Absorption." Presented at ENVITEC 83, Dusseldorf, W. Germany, February 21-24, 1983. 17. Neilsen, K. K., Moeller, J. T., and Rasmussen, S., "Reduction of Dioxins and Furans by Spray Dryer Absorption from Incinerator Flue Gas." Presented at Dioxin 85, Bayreuth, W. Germany, September 16-19, 1985. 18. Environmental Test Report: Marlon County Solid Waste-to-Energy Facility. Ogden Projects, Inc., Emeryville, CA. Report No. 107, November 21, 1986. 19. Teller, A. J., "The Landmark Framingham, Massachusetts, Incinerator." Presented at the Hazardous Materials Management Conference, Philadelphia, PA, June 5-7, 1984 20. Design and Selection Considerations for Spray Dryer Based Flue Gas Desulfurization Systems.EPA Contract 68-02-2994, WA 1/081. Radian ^ Jm Corporation, September 30, 1986, pp. 5-1 to 5-3. R-2 ------- |