May 27, 1998
Mr. Andy S. Counts
Dear Mr. Counts:
Thank you for your letter of March 31, 1998, in which you provided a list of issues which the AFMA would like addressed by the Division of Air Quality (DAQ). This letter provides declarations, clarifications, revisions and statements in response to your letter, and is a result of input from a broad group of staff involved. We make efforts toward being consistent in application of our rules and guidance throughout headquarters and our seven regions. However, due to the number of people and the complexities/uncertainties involved, our goals in this area may not always be attained. Often there are valid reasons for what seems to be inconsistency, particularly where guidance that is valid for one facility will not be appropriate for another. We appreciate your efforts in bringing these situations to our attention. Through distribution of this letter and related communications to our staff, and the letter's subsequent posting on our publicly available Internet web page, we hope to move closer to our goal of being as consistent as is practical. We solicit your continued assistance in this area.
For ease in cross referencing to your incoming letter, the responses below are ordered in a similar manner:
Woodworking Operations - We are generally comfortable with the procedures addressed in Steps 1 through 3 of your letter. However, we recognize that there are many variables and many unknowns which require vigilance to variation from the norm and the need for additional detail so that concise guidance to all situations cannot be provided. We also note that "hog" operations are not addressed.
In step one, for example, yield may change over time, as equipment, products, practices and raw materials evolve. In the longer term, a clear and detailed method for determining yield could possibly be developed and documented by you and/or your constituent companies and provided to me for consideration by our wood task force for approval and use more universally, if warranted.
In step two, we agree that the previously cited percentages are currently accepted when better information is not available. We feel that these data may sometimes be conservative when applied to specific applications, but have no better alternatives to offer. We also recognize that the "% generated from," in Table 1 would be plant-specific and we do not encourage using a universal weighted percentage. It is also necessary to determine the percentage for each type of woodworking on a case by case basis.
We also feel that the industry should soon initiate efforts to better quantify the PM-2.5 proportions of its emissions. Since there is now a new National Ambient Air Quality Standard for this pollutant, we may find areas of the state in a "non-attainment" status after appropriate ambient data are collected. The use of the stated assumption that 100% of PM-10 equals the quantity of PM-2.5 will likely result in a significant overestimation of emissions from the industry and consequently, could result in more stringent control requirements. If an AFMA testing program could be initiated early on, with protocols discussed and approved by DAQ in advance, the question could be addressed more scientifically and objectively. Such test efforts should also improve the information on PM-10. Please alert us as to any plans by AFMA or individual members in this arena well in advance, so that we can allocate resources to participate appropriately in the protocol development, test observation and data review.
Regarding step 3, we recognize that use of the manufacturer's specifications is routine. However, we must also be vigilant to the fact that these numbers are for the circumstances occurring at the time of the tests or under design conditions and that the emissions estimates are for actual conditions. Actual conditions may contribute to deterioration of the control device's efficiency, resulting in higher emissions. Any such deterioration due to maintenance schedules, production and design fluctuations, etc. should be noted and taken into account when calculating the actual emissions. The Division's rules exclude periods of malfunction and upset from estimation of actual emissions for Title V fees. However, this does not excuse a plant to use poor maintenance, "nonstandard" operation practices and/or inadequate design as a means of avoiding adequate control levels. Also, regarding controls, our regions get questions and submittals showing multicyclones as an integral part of a boiler. Please note that multicyclones are considered control devices when used with a boiler and we do not agree that they are an integral part of the boiler.
The changes in our interpretation of the emission factors for wood waste-fired boilers came about following careful review of the recent AP-42 release, dated October 1996 (again revised in February, 1998). At that time, we realized that these factors appeared to have been inappropriately interpreted in the past. Subsequently, we contacted Mr. Roy Huntly of the Emission Factors and Inventory Group at the Environmental Protection Agency (EPA) and his contractor (who did the analysis and review to develop that section of AP-42) to get advice and interpretation. This was done because there were no other factors available specifically for "dry" wood combustion emissions. The result, as documented in Mr. Huntley's August 5, 1997 letter (enclosed), provides basis and support for the judgement that the most sound way to extrapolate from the wet wood basis to dry wood applications is on the basis of Btu content. This clarification is now included as part of the February, 1998 published revision. We and EPA recognize that there are other variables and that there are arguments, which may be valid, that would more than likely lower emission estimates, if applied. We contend that there is no known quantitative way to better reflect these variables in lieu of further test results.
You submitted Attachment 2 as an argument that stack test results (apparently from the FIRE data base) were lower than those in the spreadsheet for miscellaneous hazardous pollutants. The purpose of the FIRE data base, from which they seem to have been derived, is to compile test reports that have not been reviewed by EPA (or DAQ) for reliability and appropriateness for use in estimation of emissions. Also, the values for Unit 2 in your Table 2, are almost always higher (by an order of magnitude) than the "USEPA Average" factors listed. In almost all cases, the average of the three units presented is approximately equal (or even larger than) to that number. The point of reference to this Attachment is not clear and it seems to add little to the resolution of the question of the linear relationship that you contest. We can discuss this further if some information is missing or being misinterpreted, but until we have defensible, representative test data, using approved protocols, we must insist on the use of the acknowledged conservative values in the spreadsheets, taken directly from AP-42, and as indicated in my September 9, 1997 letter to you. Unit specific test data, of course, continue to be preferable, as long as conditions are properly tested, appropriately representative and documented.
In the same section of your letter, you refer tomulticlone particulate control efficiencies. We must also caution against "blind acceptance" of typical control efficiencies being applicable in all situations. The cited particulate control efficiencies were provided by Michael Koerschner of the Asheville Regional Office at the request of Mr. David Stout, Broyhill Furniture Industries, Inc., during a February 4, 1998 telephone conversation regarding control efficiencies of multicyclones. The efficiencies quoted by Mr. Koerschner are from Table C.2-3, "Typical Collection Efficiencies of Various Particulate Control Devices," of AP-42 for a "multiple cyclone w/o fly ash reinjection." During the telephone conversation, Mr. Koerschner expressed his opinion that these numbers seemed high and would not apply specifically to wood-fired boilers. You will also note that in the Section 1.6.4 and Tables 1.6-1 and 1.6-6 of the AP-42 Section that you submitted as an attachment, there is a discussion of different applications and different efficiencies for applications of these types of control devices that range from 50% to 95%. Therefore, the DAQ will not typically accept unqualified "typical control efficiencies," but rather an analysis and justification of the efficiencies by objective means for each application.
Fossil Fuel Boilers - We agree that it is generally acceptable and appropriate to use the most recently published AP-42 emission factors, as are incorporated into our boiler spreadsheets, for estimation of emissions from these units. The Division also incorporates several assumptions into its spreadsheets on the Internet web page which expand upon the specific numbers in AP-42. These spreadsheets are used in permit and inventory evaluation.
Finishing Operations - The mass balance procedures that you describe for volatile emissions are generally adequate and appropriate for these types of operations. We will reserve further comment on the particulate/particle size issue until Mike Aldridge has completed his review of those data.
Gluing Operations - Mass balance procedures are often the best way to quantify these emissions. However, for formaldehyde, we recognize that all such material is not released to the air. We advise using the 12% end of the cited range from your enclosure 4 (Title III, Section 313 Release Reporting Guidance), to insure that "worst case" conditions are employed in these applications, as public health protection must be maintained. Other situations may exist that are specific to particular operations and units, such as: "trace" is not acceptable to quantify vinyl acetate and other toxic (HAPs/TAPs) emissions.
Kilns - The numbers indicated in your letter are not what we believe to be the latest and best available. Our most recent information that we try to use uniformly, as a criterion of acceptability of estimates, indicates VOC emissions from southern yellow pine kilns are on the order of 3.1 pounds per thousand board feet (mbf) for steam heated (as carbon), based on limited NCASI test results. DAQ has been accepting these values for general purposes in absence of more specific test data which we feel are sorely needed. We suspect even these values are likely missing oxygenates and other polar compounds, and the "as carbon" needs to be converted to actual mass. DAQ will provide a supplemental and more definitive letter on these values, as soon as we have further coordinated both internally and with the National Council of the Paper Industry for Air and Stream Improvement, Inc. (NCASI) on this issue. In the interim, you and your industries should make every effort to use the best and most reliable information possible for the CY '97 inventory which meets the stated goals of the inventory effort.
Other - There was a related topic not addressed in your letter which arose from questions received prior to a recent Outside Involvement Meeting. This question concerned what to do when a supplier does not tell you when a reformulation takes place, or the nature of that reformulation. This is not acceptable. The OSHA requirement is that a Materials Safety Data Sheet (MSDS) be attached to every shipment. If a supplier does not provide this information, they should be reminded of this requirement, with a copy of that reminder to the appropriate OSHA office.
B. Keith Overcash, P.E.
Note: This copy of this letter has been corrected to properly provide kiln emission factors in terms of thousands of board feet (mbf) rather than millions of board feet, as incorrectly stated in the original copy mailed to AFMA. The AFMA was notified of this correction by letter dated July 1, 1998, from Jim Southerland of DAQ. No other changes have been made.