EMISSIONS STANDARDS FOR MUNICIPAL WASTE COMBUSTORS:

 

Discussion and Critique about Recently Proposed Standards

 

 

Marjorie J. Clarke

Fellow, Center for Applied Studies of the Environment

City University of New York

 

 

 

1995 International Conference on Solid Waste Management:

 

Thermal Treatment and Waste-to-Energy Technologies

 

 

 

Washington, DC

 April 18 - 21, 1995

 

 

 

Emissions Standards For Municipal Waste Combustors:

Discussion and Critique about Recently Proposed Standards

 

 

Marjorie J. Clarke

Center for Applied Studies of the Environment

City University of New York

 

 

Abstract

          Over the last several years the EPA and a number of states have been working to develop and modify standards for emissions from Municipal Waste Combustors (MWCs).  In the 1980s, when EPA did not act to enact national standards, some forward-looking states did so, establishing benchmarks, which, along with European standards which were becoming more stringent with time, could serve as models for EPA.  As a result of legal action and requirements in the 1990 Clean Air Act Amendments, EPA, in 1991 and again during this last year, has moved to recommend standards for MWCs.

 

          This paper discusses and evaluates the proposed federal MWC standards in light of emissions data from new and retrofitted MWCs in this country and abroad.  Two interpretations of Section 129 of the Clean Air Act Amendments and their effects on calculation of the Maximum Achievable Control Technology (MACT), and therefore the standards, will be discussed.  The effect on emissions of different emission control designs will be explored briefly.  Particular attention will be paid to the effect of operations (e.g., flue gas temperature, injection rates for reagents such as lime and activated carbon) and maintenance of the MWC on emissions, and the role of optimizing operations as part of standards.  An alternative set of MWC standards, based on EPA’s dataset, additional data, and an alternative MACT interpretation will be presented.

 

INTRODUCTION

          Recently EPA and a number of states have been working to develop and modify standards for emissions from MWCs.  In the 1980s, when EPA did not enact national standards, some forward-looking states did so, establishing benchmarks, which, along with European standards which were becoming more stringent with time, could serve as models for EPA.  As a result of legal action and requirements in the 1990 Clean Air Act Amendments (CAAA), EPA, in 1991 and again during this last year, recommended standards for MWCs.

 

          This paper critiques the proposed federal MWC standards in light of alternative interpretations of Part 129 of the CAAA and emissions data from new and retrofitted MWCs in this country and abroad which EPA excluded from its analysis.  Particular attention will be paid to the effect of operations (e.g., flue gas temperature, injection rates for reagents such as lime and activated carbon) and maintenance of the MWC on emissions, and the role of requiring optimization of operations in the standards.  Alternative emissions standards are proposed.

 

EPA’S REVISION OF LIMITS ACCORDING TO MACT

New Plants

          For new plants EPA interprets MACT to equal the performance level of the average plant in the top 12%-ile in their database of existing US plants, instead of the performance level of the best plant, as Sec. 129 stipulates.  EPA's argument is that the best plant's performance varies over time, and therefore, the basis for this best plant's performance, and therefore, the NSPS, should be based on a range representing the best plant's performance, and the numerical value chosen should be less than the reported performance of the best plant in the database.  This assumes that the all the figures in EPA’s database represent the best possible emission for each plant listed.  But there is no evidence to support this assumption.  In fact, a more supportable assumption is that each emissions database as a whole is based on data from different plants operating under a range of operations, some plants operated well, most average, and some poorly, such as in a normal distribution (bell curve).  Additionally, one would expect that any plant at the top of a performance database would have relatively optimal operations and maintenance, and therefore would have less variation in performance. Thus, any individual datum in a database such as EPA’s is most likely to represent the middle of a distribution of operations for that plant, than one end or the other.  So it is not at all clear, as EPA contends, that the best datum in each of EPA’s emission databases does not represent a typical emission for that plant.  Further, it is certainly not clear that the average of the upper 12%ile of a bunch of different plants is an accurate representation of the emissions from the best plant.

 

          Even if it were a correct interpretation of Sec. 129, to place a NSPS limit at a level less stringent than the performance level of the best plant in a database, variation of an order of magnitude by one plant surely would not reflect the performance of the best plant. But, including data which EPA excluded, it will be shown that EPA’s proposed NSPS is often an order of magnitude higher than their databases’ best plant’s emissions.

 

Existing Plants

          For existing plants EPA is interpreting MACT to equal the average of the top 12%-ile of permits issued over many years by many states.  Part 129 (a) (2) clearly requires EPA to derive floors from lower actual emissions, when data is available:  “Emissions standards for existing units in a category may be less stringent than standards for new units in the same category but shall not be less stringent than the average emissions limitation achieved by the best performing 12 percent of units in the category...”  The statute emphasizes achievement (i.e. performance), not permitted levels, which are considerably higher.  In contrast, for the most part, the states based their permitted emission levels on the capability of older plant designs and older concepts of optimal operating practice.  Just as the older designs alone are not nearly as efficient at removing pollutants as EPA’s current design basis for controlling emissions of organics, metals, and acid gases (i.e., activated carbon, scrubbing, efficient particulate removal), previous concepts of good operating practice are incomplete, and in some cases inaccurate in the light of current experience. Following are additional reasons why permits should not serve as a basis for establishing emission guidelines for existing plants.

 

1)       Since there are no MWC permits in this country for facilities having EPA’s design basis, the database of current permits cannot come anywhere close to the eventual performance of existing MWCs when retrofitted with activated carbon as well as efficient scrubbers and particulate removal devices.

 

2)       EPA's own BID document [1] prepared in support of the 1991 MWC standards show that model plants, equipped with "good combustion and temperature control with best acid gas control and best PM control" would perform better than the currently proposed standards for existing MWC's for almost all types and sizes of plants.  (EPA’s definition of these operating practices include: "exhaust gas temperature control to 300^oF", "best acid gas control - spray dryer", and "best PM control (0.01 gr/dscf)".)  (Note that these model operating conditions do not include activated carbon injection.)   Even without carbon injection, there is considerable disparity between the EPA model plant retrofit achievement levels and EPA's proposed guidelines for large and small existing MWC's (see Figure 1).

 

            It is clear that the EPA model plant retrofits can achieve far lower emissions levels than the proposed guidelines.  The addition of activated carbon to the model retrofit should serve to reduce the model emissions even further than the levels in Figure 1 would indicate.  For this reason alone EPA's proposed guidelines for existing large and small plants are insupportable and should be lowered considerably.

 

3)       EPA in its draft MWC emission guidelines report[2] states that the Camden, NJ tests showed that injection of carbon into the SD/ESP system provides further removal of dioxins/furans of about 60% over SD/ESP systems without the injection.  A 60% reduction of the 5 ng/dscm total model dioxin/furan model plant emission number brings it down to 2 ng/dscm, a more supportable average top 12%-ile performance level.

 

4)       NOx control has usually not been required at all on MWC's in the US, and permits have frequently listed generous, if not completely uncontrolled values for NOx.  But EPA says it based its NOx emission limits on SNCR technology for all new and for large existing MWC's, but uses a database it characterizes as uncontrolled for NOx as a basis for establishing NOx emission guidelines.  This is unreasonable, and will certainly result in a guideline which is considerably higher than the typical emission levels achieved by MWC's equipped with SNCR. 

 

5)       EPA has promulgated a single set of hazardous waste incineration requirements to apply to facilities regardless of their size, type of waste burned, or age.  Since hazardous waste and the incinerators designed to burn it can vary even more than MWC's, EPA should, likewise, not choose to subdivide the universe of MWC's by age, combustor type, and size, or to devise different measures for good combustion practice and emission limit for each. 

 

RECOMMENDED INTERPRETATION OF MWC MACT FLOORS

          A more straightforward and accurate interpretation of Sec. 129 as regards the MACT level for new plants is warranted.  The performance datum for the best performing plant in the world should be the basis for the MACT floor for new plants. It is necessary to enlarge EPA’s database to include MWC’s outside the US because the basis for the proposed standards for new and existing MWC's includes activated carbon injection.  EPA’s interpretation should really have included a database of plants with this design. If activated carbon were installed on all plants in this country, the emissions data (particularly for mercury, dioxin, and acid gases) would show far improved performance than EPA’s current database.  Thus, EPA’s interpretation results in a higher emission limit than its design basis should indicate.

 

          For the reasons stated above, EPA's reliance on permitted values for regulating emissions from existing facilities is not warranted.  Since permitted values are such a poor approximation of the likely performance of MWC's retrofitted with activated carbon, scrubbers, efficient particulate controls, and NOx controls, a more supportable basis for guidelines for all existing MWC's would be the average of the top 12%-ile performance data for existing facilities in this country.  This level of performance has already been achieved by 12% of the plants in EPA's database, most of which don’t use activated carbon injection or NOx controls, and it should certainly be achievable by existing plants once they are retrofitted with these more advanced controls, scrubbers, and efficient particulate control. 

 

The MACT database is incomplete  

          Part 129 directs EPA to revise all of the numerical limits proposed prior to the 1990 CAAA, plus propose some numerical limits for additional pollutants, calculated according to MACT.   EPA compiled emissions databases for all the pollutants they were supposed to revise, except carbon monoxide and PM10.  So MACT was not calculated for these two required parameters.  Further, EPA ignored much relevant data from new and retrofitted European plants with advanced, state-of-the-art mercury, dioxin, and NOx control technologies, by assembling a strictly U.S. database. In addition, EPA’s U.S. database, itself, is not complete, excluding a number of data points which, if included, would alter the selection of MACT for several pollutants.  These exclusions are discussed below, followed by a listing of additional data for each database, and recalculation of EPA’s MACT floor for new and existing plants in the tables at the end.  A new calculation for CO and PM10 is also presented.

 

Ignored Technologies

          In addition to leaving out data from many plants in the US, EPA chose as the design basis for the new standards from less than a complete array of technologies, which have already been in use on MWC's in Europe and elsewhere for some time.  Many of the technologies that are operating, with good results, on European and/or American plants were not seriously considered by EPA for its design basis.  These advanced technologies include:  dual-stage wet scrubbers, Electro-Dynamic Venturi technology (EDV), condensers, fixed deep bed activated carbon technology, injection of combinations of reagents, such as Sorbalit.  Mercury and dioxin emission databases on plants using these technologies were presented earlier.[3]   Instead, EPA focused its design basis strictly on “dry” technology and on technologies which are the least expensive, but not necessarily the most effective.  EPA ignored Selective Catalytic Reduction and flue gas recirculation for NOx control despite the fact that  the California Air Resources Board’s evaluation of NOx control technologies[4]  indicates SCR to be the most effective, and one that they recommend to achieve a 30-50 ppm limit.  And although EPA has stipulated SNCR as the technological basis for its NOx standards, and activated carbon injection as the technological basis for acid gas, mercury and dioxin control, EPA did not base the numerical standards for either on a database of plants using these technologies.  This decision results in a MACT calculation, and therefore proposed standards and guidelines, which are far higher than they should be.

 

Operating Conditions

          Though operating conditions are one of the largest determinants of environmental performance, EPA did not include in its database information about the extent to which the all of the operating conditions at the front and back end of each plant were optimized at the time of testing.  Some of the data in EPA's databases may be reflective of reasonably good operating practice at the plants from which the data were taken.  But it is also true that much of the data in EPA's databases are from older plants whose operators do not always carry out one or more of the following operating practices,  which used together result in optimal environmental performance:

 

1.       Screen wastes at the plant to reduce incineration of pollutant precursor-bearing items

2.       Optimize mixing of waste in pit or on tipping floor (to homogenize moisture and BTU content).

3.       Optimize furnace operation (e.g., optimized grate speeds, underfire and overfire air injection rates, locations, and directions, and operation of auxiliary burner)

4.       Survey combustion equipment regularly to ensure it continues to be properly sealed and operative

5.       Optimize type of nitrogen-reducing reagent used

6.       Optimize injection location and rate for nitrogen-reducing reagent

7.       Control water injection rate to optimize flue gas temperature in control devices (to maximize condensation and capture of pollutants on particulate and reagent)

8.       Optimize type of alkaline reagent used (to maximize absorptive capacity)

9.       Optimize injection location and rate for alkaline reagent

10.     Optimize type of carbon used (to maximize adsorptive capacity)

11.     Optimize injection location and rate for carbon

12.     Optimize voltage and other electrical parameters of an ESP (to maximize capture of particulate)

13.     Control ID fan speed to optimize residence time of flue gases within combustor and control devices (e.g., fabric filters, scrubbers, furnace)

14.     Inspect and calibrate CEMS frequently

15.     Survey Emission Control Devices to ensure they are/ continue to be properly sealed, insulated, and operative

16.     Operate the plant using certified operators at all times.

 

          If any of these are not optimized when emissions data are sampled (and most plant operators do not optimize all of these simultaneously at all times), then it is likely that measurements in EPA's database reflect less than optimal environmental performance.  Since it is known that variation in application of the techniques, practices, and conditions listed above will result in variation in environmental performance, EPA should base its standards on data reflecting known good operating practices, and not on emissions from plants which do not follow optimal operating practices, for to do otherwise encourages less than optimal operations. 

 

CO -- Good Combustion Practice

          EPA has not gathered a database of CO emissions as it has for the other pollutants of concern, and has thereby not complied with the requirements of CAAA Sec. 129; instead it is relying on its 1989 BID document[5].  In the current proposal, EPA has again recommended a triple standard (50, 100, 150 ppm) for new plants and a quintuple standard (50, 100, 150, 200, 250 ppm) for existing plants depending on type of combustor.  And yet in the BID document, EPA states that the first goal of good combustion practice (maximization of in-furnace destruction of trace organics) is accomplished by optimizing waste feeding procedures, achieving adequate combustion temperatures, providing the proper amount and distribution of combustion air, and optimizing the mixing process.  A failure in any one of these components will be accompanied by spikes or bulk increases in flue gas CO concentrations.”  Further EPA states, "Failure to achieve the necessary temperatures and residence times will result in the escape of organics from the furnace, which will lead to elevated concentrations of CO in flue gases", "CO emissions typically increase when insufficient O₂ is available to complete combustion, or when excessive amounts of O₂ quench combustion reactions", and "Failure to distribute combustion air in the correct proportions to primary and secondary supplies can result in elevated organics and CO emissions".  It is clear from these statements that EPA considers that it is the combustion practices which govern CO emissions. Allowing some plants to have 150 or 250 ppm CO emissions would indicate that these plants receive deferential treatment, and can fail to distribute combustion air in correct proportion, or provide insufficient O₂ or too much O₂, or fail to achieve satisfactory combustion temperatures, whereas other plants would be required to conform to the good combustion practices mentioned, and consistently achieve 50 ppm.  These triple and quintuple standards are inherently unfair, and makes a mockery of the term "good combustion practice", since it gives the operators at certain plants latitude in the monitoring and optimization of their combustion operations.  In order to demonstrate that good combustion practice is being achieved at all plants, all plants should be held to the 50 ppm CO emissions level.

 

          In the proposed standards and guidelines, RDF plants are permitted to average 150 ppm, and existing RDF plants to average 200 ppm.  As argued above, good combustion practices, such as correct combustion air, mixing, and temperature, should not be required for some plants and not for others.  That higher CO emissions are permitted from RDF plants means that these plants are not required to achieve as good combustion as some others.  In addition, there is no evidence in the aforementioned background document used to devise these CO standards for RDF plants that concerted attempts were typically made to zero-in on those operating practices which optimized combustion.  In fact, the Penobscot, ME plant, discussed in connection with good combustion at RDF plants in the BID, has no impetus to optimize combustion such that it operates any more efficiently than its 400 ppmv/4 hour permit requires.  How can a standard of good operating practice be based on operations at such a plant?  

 

          Based on information relating combustion efficiency and emissions of PICs to a number of incinerator performance criteria, EPA, in 1987 issued its first good combustion practice guidelines[6], which flatly stipulated one CO limit, an O₂ limit, and other firm limitations for all plants.  This guidance, advising the states regarding good combustion practices, indicated that carbon monoxide emissions of 50 ppm over a four-hour averaging time, along with 6-12% oxygen after combustion, among other requirements, were indicators of good combustion practice in a municipal waste combustor.  Environment Canada also later adopted the 50 ppm CO standard in its standards for good combustion.[7]  New regulations issued by the Netherlands are similar at 44 ppm (corrected from 50 mg/Nm³).[8]

 

          As important as the level of CO emissions maintained in a MWC is the averaging time over which these emissions are evaluated.  It is important to note, in this regard, that the ASME/NYSERDA Pittsfield combustion tests[9] showed that CO levels above 100 ppm were associated with higher dioxin levels.  If several types of new and existing MWC's are permitted to exceed this 100 ppm level routinely, then EPA is not attempting to minimize dioxin emissions in certain combustor designs.  The Pittsfield research demonstrates the importance of minimizing the number, intensity, and duration of CO spikes, and thus, of limiting the length of the averaging period for CO.  Thus, to minimize the opportunity for formation of PICs, it is necessary to require an average limit for CO which would reflect a severe limitation on the frequency, intensity, and duration of excursions.

 

Good Combustion Recommendations

          Since "Good Combustion Practices" (GCP) are, ostensibly, practices which, when utilized, result in good combustion, the more reasonable approach would be to reestablish an across-the-board standard which is reflective of good combustion practices.  Though EPA’s previous guidance stipulated this figure as 50 ppm, a recalculation of the MACT floor for CO shows that this figure is higher than MACT.  In addition to using CO as an indicator of good combustion, EPA's original GCP's included a range of oxygen content (6-12%).  EPA should reinstate this requirement defining good combustion, since it has been demonstrated that lower oxygen values increase the formation of PICs and higher oxygen values result when there is too much excess air, resulting in cool spots and reduced flue gas residence time in the furnace.  Thus, it is recommended that good combustion practices for all incinerators require not only a minimum temperature of 1800^oF across the furnace to ensure destruction of PICs as was recommended by EPA[10] but also a maximum temperature of 2000^oF at fully mixed height to minimize formation of NOx, as well as the requirement that flue gases remain at such high temperatures for two seconds.

 

Exclusion of Startup, Shutdown, and Upsets

          Since it is during these times that extremely poor emissions occur, the exclusion of startup, shutdown, and periods, during which time the plant is malfunctioning, from measurement via CEM or stack testing for purposes of compliance is a huge loophole which winks at suboptimal operations and bad operating practices.   Such a provision does not penalize a poorly maintained or operated plant which is often down, even though the air quality in the plant vicinity is certainly adversely affected by such inconsistent operations.  Since it is assumed that, for good combustion practice, auxiliary burners are required in the furnace to keep temperatures to the correct level, avoiding upsets, there should be no need for EPA to exclude periods of startup, shutdown and upset from compliance monitoring and testing.  Therefore, in keeping with the above recommendations regarding good combustion practice, it is further recommended that there be no exclusion in either stack sampling or CEM measurements, and that CEM measurements be required at all times a MWC is in operation in order to stay in compliance.

 

Flue Gas Temperature

          This is the single most important short-term improvement to maximizing environmental performance at many MWC's operating today.  Maintaining low flue gas temperature will have the dual effects of improving reagent (lime) utilization and increase removal of volatile trace elements, such as mercury and dioxin/furan as well as acid gas emissions (HCl and SO₂) as described below.

 

          In the 1991 MWC standards EPA was only interested in outlet temperature being below 450^oF to avoid secondary formation of dioxins in the emissions control devices.  But in EPA's 1989 BID document[11], p. 3-2), EPA's two most effective emissions control options, both require a temperature of 300^oF.  EPA mentions temperature control as one of the best technologies for retrofit for each type of MWC.  But in the current proposal, there is no specific back-end temperature requirement.  Data produced by Environment Canada[12] showed that temperatures around 285^oF were optimal for plants using both spray dryers and sorbent injection.  Wet scrubbers lower the temperature much more than this, with as good or better results.  Permitting MWC's to maintain high flue gas temperatures at MWC's is at odds with efforts to lower most emissions of concern and should be addressed in the new standards and guidelines. 

 

          In fact, MWC's can be successfully operated at outlet temperatures of 240-260^oF.  In fact, in pilot plant tests, spray dryer absorber outlet temperatures as low as 200^oF have been tested while maintaining a free flowing residue product according to Joy/NIRO.  During start-up and testing of the Zurich SDA system (Zurich has a large ESP), long-term tests were performed at outlet temperatures as low as 220^oF without drying problems. Though some vendors with less experience in low temperature operations, claim that lower flue gas temperatures can cause corrosion and operating problems, there is convincing evidence, presented by Donnelly and Felsvang[13] that these problems can be avoided by proper design and operation.  Four methods, which are inexpensive and easy to implement, are employed to minimize adverse effects on ESP's:  insulation, control of air inleakage, hopper heating, and operating procedures.

 

          In Europe, dual-stage and other wet scrubbers, condensers, and heat exchangers operate at much lower temperatures with good environmental performance.  In a memo from Radian Corp., EPA's consultants, to EPA[14], Radian discusses a few of the many activated carbon pilots and permanent installations in the US and abroad, and the incremental polishing effect which activated carbon has already achieved on MWC's equipped with scrubber/ESP and scrubber/FF.  In addition to demonstrating that activated carbon achieves close to an additional order of magnitude polishing effect over the reduction achieved by the conventional emission control system, the Radian memo also shows that reduction in flue gas temperature (below 250^oF) using activated carbon injection is associated with an even greater polishing effect for dioxins (96%) than use of activated carbon injection at a higher flue gas temperature of 284^oF (41-58%).  This additive effect of lowered flue gas temperature has also been shown to apply to emissions of acid gases and mercury as shown by Environment Canada in the aforementioned report on the Quebec MWC and by Brown and Felsvang[15].  But despite the evidence that MWC flue gas temperatures around 250^oF have been demonstrated to optimize the capture of dioxin, mercury, and acid gases, EPA has neglected to set a specific standard or guideline for flue gas temperature for MWC’s.  In sharp contrast, EPA did set such a requirement in its recently proposed standards for Medical Waste Incinerators.  In fact in order to stay in compliance the plant must demonstrate continued optimization of flue gas temperatures at the particulate control device inlet over time.

 

Reagent Injection Rates

          There are also significant improvements to be gained by optimizing the injection rate for activated carbon as shown by Kane (Ref. 14).  At the Kassel MWC the polishing effect of the activated carbon for decreasing dioxin emissions increased from 78% to 98% as carbon feed rate increased from 25 mg/dscm to 137 mg/dscm while holding flue gas temperature constant at 275^oF.  This effect was also shown at the Zurich MWC and the Borgess MWI.  At several European plants Brown and Felsvang showed the same effect for mercury.  For example, at the Kassel MWC, an increase in carbon injection rate from 9 to 64 mg/m³, while temperature was held constant at 279^oF, resulted in an increase in removal from 48% to 82%.  At the Amager plant in Denmark an increase in carbon injection rate from 19 to 70 mg/m³, at a flue gas temperature of 260^oF, resulted in a increase in average mercury removal from 88% to 97%.  At the Zurich MWC as temperature remained constant at 248^oF, and the carbon injection rate was increased from 7 mg/m³ to 30 mg/m₃ the average mercury removal efficiency increased from 85% to 92% ug/m³.

 

          Similarly, the injection rate for alkaline reagents affects the emission of acid gases[16], and probably also mercury and dioxin.  EPA's aforementioned model emission control systems assume a 2.5 sorbent-to-acid gas stoichiometric ratio.  The location of alkaline reagent injection is also critical to emissions control as seen below.

 

Dayton MWC -- A Case In Point

          In 1989 EPA conducted a large test program on the 1970-vintage Montgomery County South incinerator[17].  Six different operating conditions were tested, three runs apiece, and most of the major pollutants of concern were tested for each operating condition.  All but the sixth were at a furnace mixing chamber temperature of 1750^oF.  The operating conditions were:

 

1.       ESP inlet setpoint 575^oF, no sorbent injection

2.       ESP inlet setpoint 400^oF, no sorbent injection

3.       ESP inlet setpoint 400^oF, furnace sorbent injection 500 lb/hr

4.       ESP inlet setpoint 300^oF, furnace sorbent injection 500 lb/hr

5.       ESP inlet setpoint 300^oF, duct sorbent injection 300 lb/hr

6.       ESP inlet setpoint 525^oF, no sorbent injection; 1500 mixing temp

 

          The results were striking.  The best condition is #5 for most pollutants. For dioxin the differences between conditions are dramatic.  At condition 5, two of the three runs for toxic equivalents are quite low -- 1/5 the level of the European dioxin standard, with the third reading twice that standard.  Considering this finding it is strange that the Dayton incinerator is not being operated at this condition.  At condition #4 the effect of changing from duct to furnace injection and increasing quantity of lime, increased the dioxin toxic equivalents one to two orders of magnitude.  This test points to the value of having duct injection.  Changing to condition #3, holding the lime injection rate constant and increasing temperature to 400, resulted in a 75% increase in toxic equivalents and a marked increase in acid gases (HCl increased roughly 400% and SO₂ increased about 350%).  (This condition is most similar to current operating conditions at Dayton South.)  At condition #2, removing lime injection and holding the temperature at 400 seemed to improve the CO emissions, but doubled HCl emissions.  At condition #1, increasing the temperature to 575 caused one to two orders of magnitude increase in dioxin toxic equivalents over the previous condition (400^oF, no injection).  Finally, operating at the worst condition, #6, without lime injection, at high ESP temperature (525^oF) and at lower furnace mixing chamber temperature (1500^oF, the mercury and cadmium emissions increased somewhat from the previous condition, and the CO, lead, and particulate emissions increased about an order of magnitude.  This latter condition is significant to Dayton because it appears that the incinerator currently seems to operate frequently at suboptimal furnace mixing temperatures.

 

          The results of the testing were clear.  The downstream flue gas temperature at Dayton should be no higher than 300^oC, injection of lime should occur in the duct, and furnace temperature should be maintained at 1800^oF.  Instead the flue gas temperature is closer to 500^oF, lime is still injected into the furnace, and suboptimal furnace temperatures occur frequently, accompanied by wide variation in CO.  Although minor changes in design and operation had an enormous effect on emissions of dioxin and the other pollutants of concern, EPA's study was ignored, and no government authority requires operators at this MWC to optimize operations there.  Since there is less automation at this plant, the training and watchfulness of the operators is even more important here.  There is also evidence of maintenance problems at Dayton South.  This case study illustrates the long-disregarded need, not only for optimization studies at all MWC's, but also enforcement of optimal designs, operations, and timely maintenance.

 

Good Operating Practice -- Recommendations

          The data presented above (Joy/NIRO, Environment Canada) argues for institution of an operations requirement similar to the one promulgated by the New Jersey Department of Environmental Protection in September, 1994, as part of New Jersey's mercury emission regulations for MWC's.  In order to ensure that plants are operated optimally at all times, New Jersey's mercury regulations will require all MWC's to conduct tests to determine the optimized reagent feed rate for mercury emissions control apparatus during the first quarter of the source emissions tests required by the NJDEP, and to adhere to that rate thenceforth.  Since the rate of carbon and alkaline injection has such an impact on emissions and control efficiency, it is incumbent upon EPA to require MWC operators to conduct tests to determine optimal carbon and alkaline reagent injection rates, and then require the operators to adhere to these at all times.  These rules should also require MWC operators to provide regulatory authorities with records verifying regular purchase of each reagent to confirm that reagent optimization is actually occuring over time.  The lessons of Dayton should not be disregarded.

 

          Since good operating practice can be quantified, EPA should include in its NSPS and Guidelines for MWC’s some of the same provisions requiring optimization of operations that it included in its Medical Waste Incinerator standards:

 

·        Initial optimization of reagent injection rates for lime, carbon, and other reagents for all plants,

·        Continued observance of optimized injection rates in order for a plant to stay in compliance with the standard, and

·        Optimization at all times of flue gas temperatures at the particulate control device inlets of all plants in order to maintain compliance.  A good maximum target level for this flue gas temperature level would be 250^oF.

 

MONITORING

          It is important that operators easily monitor all devices and parameters of concern.  Continuous emission and process monitors are designed to assist in this, and as soon as new technology is developed to monitor continuously pollutants of concern, EPA should require their use.  At present EPA does not yet require HCl monitors even though they have been in use in Europe and in the US for years.  In addition, particulate and mercury CEMS have been in use in Europe for years.  Requirement of CEMS benefits operators, operations, and regulators, and is essential.

 

OPERATOR TRAINING AND CERTIFICATION

          In the 1991 NSPS EPA required that chief facility operators (CFO) and shift supervisors (SS) be certified to the first level of the ASME Operator Certification (OC) program, and that each plant have a plant operations manual that each employee was to "review".  In this NSPS EPA has additionally required that the CFO's and SS's be certified to the second, site-specific, level of ASME's OC program, and that control room operators who are to take over for a CFO or SS should also be certified to the first level (optional).  Also, with regard to operator training, the new standard would require that all CFOs, SS and control room operators complete an MWC operator training course approved by EPA within 2 years.  These additions are definite improvements, but more is needed.

 

          EPA has a minimal role in developing the questions for the provisional and site-specific exams.  At present, the ASME QRO committee, many of whom represent the resource recovery industry and its training programs, derives its questions in hodgepodge fashion just from committee members and any others who happen to hear about it through word-of-mouth.  In fact, ASME frequently states at QRO meetings that there is a shortage of questions in several categories.  As a result, the questions on the exams may not be as rigorous or as varied as they should be.  The first two times that the provisional test was given over 90% passed.  Subsequent tests were passed by fewer applicants, but more of these were repeating the test because they had failed it before.  Enough provisional exams have now been given that most of the existing CFOs and SSs have already taken it, so the only way to rectify possible flaws in testing would be recertification testing.  At the present time, no recertification testing has to take place at specified intervals.  Recommendations are detailed below.

 

Operator Training and Certification Recommendations

1.       Limit frequency/period of time that control room operators can fill in for Chief Facility Operators and Shift Supervisors.  Require all control room operators to have full certification if they are to substitute for Chief Facility Operators or Shift Supervisors;

2.       Require operators to take tests on new regulations and new technologies every five years; this would ensure that operators stay up-to-date with the constantly changing technologies and regulations in the field;

3.       Require that no employee of a firm which designs, operates, or constructs municipal waste combustors either create exam questions or have access to exam questions (currently there are potential conflicts of interest on the QRO);

4.       Requirement that no employee of a firm which has designed, operated, or constructed the specific municipal waste combustor at which an applicant is taking a site-specific exam, be permitted to sit on the examining board (this would prevent future conflicts of interest); and

5.       Require a minimum educational requirement for taking the certification exams:  either a technical baccalaureate degree or 60 credits in physical science and/or engineering at an accredited institution.  (Currently the minimum qualification requirement is a high school diploma or equivalent.)

6.       EPA’s Air Pollution Training Institute should be heavily involved in developing questions for the provisional exams, and staff from this Institute should be involved in administering the site-specific exam as well.  This should be done quickly, since ASME is scheduling exams at a fast pace.

7.       Require that EPA's Training Institute approve the operation and training manuals at each incinerator site and that these manuals include specific directions for proper screening of waste, and for AVOIDING and not just dealing with upsets.  In this NSPS, each plant is responsible for designing its own manual according to general guidelines.  This will most certainly result in little uniformity in plant operations manuals across the country.  Employees are only asked to "review" the manual annually.  This lack of implied or enforced rigor will also ensure lack of uniformity in employee training and preparedness.  EPA does not appear to have oversight either in approving the manuals or in making sure operators review these manuals adequately or are properly trained.

 

NUMERICAL LIMITS VS. PERCENTAGE CONTROL

          Part 129 requires numerical limits for all the pollutants of concern.  However, in contradiction of this requirement, the current proposal lists numerical limits for some, and for three pollutants, gives a choice of numerical or percentage reduction -- whichever is LESS stringent.  Practically speaking, the standards for HCl, SO₂, and now, mercury are not numerical standards, since the alternative percentages specified for each are so lax, it is, in practical terms, the percentages that are governing for the acid gases, and will be so for mercury. EPA says it chose the 85% number for Hg because 85% control is still possible even when there is a spike in the Hg inlet value due to a battery or similar.  But this dual standard could discourage active efforts to limit batteries and other items with concentrated levels of pollutant precursors from entering the waste stream. The remedy here is to require application of pollution prevention measures to ensure that those waste items which typically cause the spikes never enter the incinerator in the first place, not to assume that they will always be there and relax the standard to accommodate them.  Pollution prevention measures to address the mercury spike problem are an integral part of New Jersey’s mercury emissions standard and regulations for MWC’s, and include measures to separate certain batteries from the waste stream and ban certain others, and other measures which reduce the concentration of mercury in packaging materials over time. Figure 2 demonstrates that uncontrolled emissions above 525 ug/dscm (a common occurrence) would only need to meet an 85% control level, no matter how high the inlet value.  New Jersey assumed uncontrolled Hg emissions to be 700 ug/dscm in its calculations.  Many uncontrolled mercury emissions exceed 1000 ug/dscm at times when waste items concentrated in mercury are incinerated.  For mercury inlet values of under 525 ug/dscm, the 80 ug/dscm standard would apply, no matter how low the inlet value became.  This means that the control efficiency would be allowed to plummet, and operators would have no impetus to optimize operations.

 

          The effect of a dual standard is similar for acid gases. With respect to SO₂, the uncontrolled emissions would have to be less than 150 ppm for 80% control to be less stringent.  Many uncontrolled emissions of SO₂ range as high as 500 ppm.  As for HCl, the uncontrolled emissions would have to be less than 500 ppm in order for 95% reduction to be less stringent.  The range of uncontrolled HCl is closer to 500 to over 1000 ppm.

 

Recommendation against Dual Standards

          Considering that the CAAA requires numerical, not percentage standards, as well as the adverse effects of having a dual standard both on emissions when pollutant precursor content is high or on operations when they aren’t, it is recommended that the percentage control numbers be dropped entirely from the NSPS and Guidelines for HCl, SO₂ and mercury, and that the numerical limits be operative at all times.  A more protective method would be to have a dual standard which chooses the most stringent option.  In this way if the waste stream is high in pollutant precursors the numerical limit would ensure that high levels of the pollutant are not emitted.  If the waste stream is low in pollutant precursors, then the minimum percentage control requirement would be operative, ensuring that the plant operators must remain alert and operations and maintenance continue to be optimized.

 

DIFFERENT STANDARDS FOR SMALLER PLANTS

          Since the 1991 NSPS, EPA decided that there was no reason to subdivide the standards for new MWC’s, perhaps because it saw no difference in the emission control technologies available to and already used on large and small plants.  However, for purposes of this rulemaking, EPA divided the universe of existing MWC plants into three size categories: large (over 250 tons per day), small (25 - 250 tons per day), and too small to be regulated (under 25 tons per day).  But, as successful retrofits on small MWC’s in this country and Europe have demonstrated, very small plants can be successfully retrofitted and exhibit superior environmental performance similar to larger plants. (See data presented in Ref. 3.)  But despite this fact, EPA is proposing weak or no emission requirements at all for small existing MWC’s.

 

          That EPA's small MWC database is full of small plants which do not perform well is not due to the inherent inability of small plants to perform well.  It is more a consequence of low expectations by regulators, followed by very lax standards and permits, which encourage plant design using less advanced, cheaper technology.  Permits for small plants have often included no requirements for control of emissions.  There is no technological basis for continuing to forgive smaller plants for having higher emissions.  And there is certainly no reason for EPA to excuse the smallest plants (i.e., 25 Mg/day or less, such as apartment incinerators, the new wave of residential MSW incinerators, and other very small units) from adhering to any standards at all.

 

          In its currently proposed Guidelines, (pp. 87-88) EPA states that many of the small plants don't have permits.  For some pollutants, less than 11 small MWC permits were identified, so in those cases, typical uncontrolled emission levels for that pollutant were used for determining the average of the top 12% of emission limitations.  The Clean Air Act Amendments did not specify that a minimum number of plants have to be included in a database for a standard to be set, and it certainly did not contemplate the standard being set at an uncontrolled level.  EPA has claimed that smaller and older plants should not be held to the same standards as new and larger plants.  The basis for this seems to be cost and paperwork.  But there are a number of exemplary MWC’s in Europe which fall into EPA’s “small” MWC category, and which have demonstrated the capability of superior environmental performance comparable to larger plants. (See Ref. 3) A more supportable small plant guideline would be possibly by enlarging the small plant databases by undertaking further data collection, and then applying the average 12%ile criterion to choose a guideline, or alternatively, by taking the best or second best number out of the 11 plants in the current database as the operative guideline.

 

          But EPA has indicated before that it sees no difference in the capability of smaller existing plants to be retrofit and perform as well as larger plant retrofits.  In its 1989 BID document cited above, EPA states that model plant retrofit emission figures in Figure 1 all apply equally to large plants as well as to small plants.  For example, Modular Excess-air units of 100 tpd and of 140 tpd, Modular starved-air reciprocating grate units of 25 tpd and of 50 tpd, as well as small mass burn waterwall units of 100 tpd. Furthermore, using permits as a basis for the small plants guidelines results in limits anywhere from 3 to 16 times what EPA said in 1989 that these small plants could achieve with the best technology for CO, particulates, dioxin, and acid gases.  These facts argue against the need for special treatment for smaller MWC's.

 

          EPA has decided that smaller plants don't need to have NOx controls at all (i.e. emission limit of 500 ppm, which is quite a bit higher than any uncontrolled NOx level).  This is such a high emissions level that it invites the plant operators to become careless in regulating temperature and oxygen conditions in the furnace.  EPA's decision could have been based on the fact that permit levels are uncontrolled for NOx, and control technology has not been required in the past.  But EPA also claims, erroneously, that SNCR is incompatible with smaller, modular combustors. Enercon, which has always included flue gas recirculation in its incinerators, and which has comparatively low emissions of NOx as a result, indicates that SNCR is very compatible with its system because of the flue gas recirculation.  The latter technology results in a stabilization of the temperature in the furnace at the level correct for SNCR injection of ammonia or urea.[18]  So by mischaracterizing SNCR’s wide applicability, and by not even considering flue gas recirculation and other technologies, EPA has mistakenly excluded smaller plants from NOx control requirements.

 

          EPA has also used the argument that a 250 ton per day cutoff is needed to separate large from small plants because the cost of air pollution control devices is increasingly more expensive for smaller plants.  But activated carbon injection is an inexpensive retrofit since it only involves duct work and silo.  The European database is replete with examples of small plants successfully retrofitted and performing well with not only activated carbon injection, but also dual stage wet scrubbing and other technologies (see Ref. 3).

 

          A case in point demonstrates the capacity of small plants in achieving reduced emissions[19].  A small modular MWC in Pittsfield, Massachusetts (3 x 120 tons per day) was retrofitted with the following equipment: a steaming economizer and a trim economizer (reduce flue gas temperature), 4-field ESPs (for particulate removal), condensate economizer (further reduces flue gas temperature -- to 160^oF), wet scrubber (packed tower absorbers for acid removal), multi-cyclone/recirculating flue gas (for combustion air and NOx control, and CEMS for CO, NOx, SO₂, and O₂ (for improvement of combustion and emissions control).  This retrofit was completed in 18 days, during a scheduled shutdown for maintenance.  Emissions were reduced for all pollutants of concern by as much as three orders of magnitude, and often more than one order of magnitude due to this retrofit.  These results show impressive improvements in performance not reflected in EPA's guidelines.  It also  demonstrates the viability of wet scrubbing technology for small as well as large MWC’s.

 

Exempting plants <25 Mg/Day

          The NSPS exempts very small incinerators from these emissions and siting requirements, even though it has been shown that the smallest plants can be responsible for the worst ambient impacts (e.g., in New York City -- apartment house incinerators were antiquated, uncontrolled, badly operated, and emitted at roof level).  Excusing these plants from standards encourages more of them.  An inventor in Marblehead, MA is seeking patents for a new single-family household garbage incinerator.  New Jersey's new mercury standard does not exempt small plants, thereby requiring them to upgrade or shut down.  New York City passed a law to phase out 2200 apartment incinerators and small commercial and institutional incinerators because of their environmental impacts.  Thus even the smallest MWC’s should be included in the guidelines.

 

Recommendation Regarding Plant Size

          All guidelines should apply across-the-board to all sizes of MWC just as do standards for all sizes of hazardous and medical waste incinerators.  The full range of emission control and combustion control technology is available and has been retrofitted on large and small MWC’s alike, and smaller MWC’s, particularly those under 25 tons per day, can produce a disproportionate effect because of short stack height and smaller dispersion.  Excluding these smallest MWC’s just encourages construction of new ones, and should be avoided.

 

Recommendations From the Emission Control Industry

          It is of interest that the Institute for Clean Air Companies (ICAC) stated in its testimony to the docket[20] that EPA’s proposed standards and guidelines are too lax in a number of areas. 

 

Acid Gases. For example, among the recommendations are that “on small plants the required removals could be increased to perhaps 70% for SO₂ and 85% for HCl without making scrubber cost exorbitant.” 

Mercury.  And recognizing that EPA’s design basis for mercury includes activated carbon in tandem with convention technology (SD/FF), ICAC has stated that the “proposed limit on mercury emissions of 0.090 mg/dscm or 85% reduction from both new and existing units can be met using current technology” and that lower limits have been set in Florida, New Jersey, and Minnesota.  ICAC goes on to corroborate this statement by citing a report that “approximately 99% SD/FF mercury control efficiency at a municipal waste combustor”.  Also, “greater than 98% reductions in mercury emissions to a level below 0.050 mg/dscm are achievable using activated carbon injection in combination with a spray dryer/baghouse”.  “Substantial reduction may occur even at low inlet mercury levels.  For example,... injection... yielded greater than a 90% reduction from an uncontrolled level below 0.010 mg/dscm”. 

Dioxin.  With respect to dioxins, ICAC states that “in many cases, use of a dry scrubber ...without carbon injection will be sufficient to meet the proposed limits”.  Thus, even the industry agrees that EPA has underestimated the capability of its own technology design basis in setting its mercury and dioxin standards.  NOx.  Insofar as NOx is concerned, ICAC states “EPA’s proposed limit of 180 ppm is neither a technical limit to the capabilities of NOx control technologies, nor is it a minimum in the cost-effectiveness/removal efficiency curve for these technologies.  In fact, a lower limit can be met comfortably and cost-effectively.  We thus recommend that the Agency promulgate a limit of 150 ppm, with an alternative of a 50% reduction in emissions”.  In addition to these recommendations regarding NOx, ICAC states: “EPA’s analysis of SNCR costs neglects economics of scale at plants with multiple combustors.” 

HCl CEMS.  Regarding continuous monitoring, ICAC recommends EPA require CEMS for HCl since “CEM for HCl are in use at 15 municipal waste combustion plants in the U.S., where there are 60 continuous HCl monitors.  Pennsylvania, Maryland, and New Jersey require HCl CEM on new units (and Pennsylvania on existing units as well).  Because continuous monitoring of NOx, SO₂, CO, and opacity already is required, the incremental cost of continuous monitoring of HCl will be limited to the cost of purchasing and maintaining a detector. 

Compliance Schedule.  Finally, ICAC suggests “that EPA reduce the length of time afforded sources to come into compliance with the limits specified in the emissions guidelines.  A three year timetable to reach compliance might make sense if owners and operators of combustors had no forewarning of impending regulation.  In fact, these owners and operators have the additional two  years between proposal of the guidelines and their implementation by the states, and further should have been aware of the order of magnitude of the standards since at least 1991.”

RECOMMENDATIONS FOR NUMERICAL NSPS AND GUIDELINES FOR EXISTING MWC’S

          EPA’s proposed NSPS and guidelines for existing MWC's are founded on incomplete databases which are biased towards poorer performing plants, older plant designs, and less than optimal operation and maintenance, as a result of excluding some of the newest, best plants in this country and abroad.  EPA should investigate all the best current technologies and current database worldwide, just as it requires of applicants for PSD permits, update its databases, and promulgate standards for new and existing MWC's which are supported by those databases.  As it proposes for medical waste incinerators, EPA should also stipulate and require at all times observation of criteria representing Good Operating Practice, practices which result in efficient combustion and optimized air pollution control operations such as reagent injection rates and low flue gas temperature operation, as well as good maintenance and close monitoring of emissions and process parameters, at all times to maintain compliance.

 

Recalculation of NSPS for New and Existing Plants

          Based on the aforementioned arguments, the top 12% performance average on which EPA has based MACT for new plants, underestimates the performance capability of the best plants currently existing (even those without activated carbon injection).  But Sec. 129 of the Clean Air Act requires that the NSPS be set at the level achieved by the best plant.  Therefore, EPA's NSPS for all parameters should be more stringent.  As EPA did, our MACT calculations involve ranking the performances of plants for each pollutant, resulting in different plants achieving the lowest emission limitations for different pollutants.  To attempt to choose a single plant as best for all parameters would involve subjective judgements.  Tables 1-10 present, for each pollutant, data from U.S. MWC’s not in EPA’s databases, interspersed with EPA’s data in numerical order, with EPA’s original data in bold for the top 12%ile of plants. Also in these tables EPA’s proposed NSPS is contrasted with the average data for the best plant in the combined U.S./European database, which includes EPA’s database plus additional plants.   It is this performance average which is recommended as the NSPS level for each pollutant.  Also presented in each table is the number of plants represented by 12% of the number of plants in the revised database. MACT for existing plants is calculated by averaging the emissions from top 12%-ile of small and large domestic plants, using our list and EPA’s list combined.[21]  (Please see this reference for a full enumeration of the U.S. and European databases and references, since there is insufficient room to list every datum in both our and EPA’s datasets as well as the source of every plant stack test and CEM report). Also, since EPA did not provide data for CO or PM10, MACT is calculated based on our database alone.  Carbon monoxide is sampled regularly in most plants, and it is recognized that the potential CO database would be enormous.  The best operating plants are likely to be well represented by the two plant data shown in the table (CO emissions even lower than these figures have been achieved often by well designed and operated plants).  Figure 3 shows the current NSPS Proposal vs. Average emissions from the best plant in each database.  Note that our recommended NSPS’ are frequently at least an order of magnitude lower than EPA’s proposed NSPS’.  Figure 4 shows that the average plant performances for each pollutant for the top 12%ile of plants from the expanded domestic database result in considerably lower numbers than EPA’s proposed guidelines.  Thus, based on all the arguments presented, it is recommended that EPA include the additional plants in their databases and adopt the recalculations of MACT presented here.

 

REFERENCES

 

 

 Figure 1.  Proposed Guidelines for Small and Large Existing MWC’s  vs. EPA Model Retrofit Data

 

 

 

       Figure 2.  Adverse Effects of a Dual Standard

 

 

 

 

           Figure 3.  Current NSPS Proposal vs. The Best Plant in each Database

            (Note:  the NSPS for CO is represented as the average of the 50 - 150 ppm range)

 

 

 

          Figure 4.  Guidelines for Large MWC’s vs. Average 12%-ile Performance Level for Existing Database

            (Note: The CO Guideline is represented as an average of the 50 - 250 ppm range)  

 

 

Key Words

 

Marjorie J. Clarke

 

 

 

Municipal Solid Waste

Waste-to-Energy

MWC Standards

MWC Guidelines

Dioxin

Mercury