Volume 5, Number 8
A Gossman Consulting, Inc. Publication
August 1999

HWC MACT: Part 1 - Commentary on the Preamble

David Gossman

The following are first pass observations based on our reading of the pre-publication version of the Hazardous Waste Combustor MACT rule and preamble signed on July 30, 1999. Undoubtedly, other issues of concern will come up but this should provide food for thought. (Section headings refer to the preamble organization.)

Part Three, IV. B.
In contrast with what EPA led the industry to believe, the new MACT standards will not get facilities out of site-specific risk assessments (SSRAs). EPA will continue to use omnibus authority to require these expensive tests and modeling. That said, EPA indicates "...the incorporation of site-specific, risk based permit conditions into a permit is not anticipated to be necessary for the vast majority of hazardous waste combustors."

Part Four, II. A. 2. Footnote 38.
This footnote states that the dioxin standard controls APCD inlet temperatures to below 400°F. This is not true. Higher temperatures are allowed if the 0.2 ng TEQ/dscm standard can be met.

Part Four, II. A. 2.
This whole section contradicts itself. The logic provided for not separately regulating some metals is the exact opposite of that provided for why particulates is not an adequate surrogate for all but mercury for the other metals. The good news is, no more feedstream testing of barium, silver, and thallium. Too bad beryllium could not have been included with these.

Part Four, II. A. 2.
EPA classifies thallium and antimony as low volatile metals despite the fact that data shows them to behave very similar to cadmium and lead respectively.

Part Four, III. C.
All standards and limits are based on two significant figures. ASTM E-29-90 is recommended for use. This means, for example, that a temperature limit of 400°F will allow up to 404°F without triggering an exceedence.

Part Four. IV.
DRE testing will continue to be required despite the considerable database showing that, for cement kilns at least, it is virtually impossible to fail with a well designed test. Setting minimum "combustion zone" temperatures based on DRE testing will continue to be problematic in cement kilns.

If CO monitoring is chosen over hydrocarbon monitoring. HC testing will be required during the compliance test. HC continuous monitoring will eliminate the need for any CO monitoring.

Part Four, VII.B.2.
Apparently ignoring significant quantities of data that demonstrate just the opposite, EPA states "non-dioxin and non-furan organic hazardous air pollutants emitted from hazardous waste burning cement kilns have the potential to be greater than those from non-hazardous waste burning cement kilns because hazardous waste can contain high concentrations of a wide variety of organic hazardous air pollutants." The simple fact is, PICs for cement kilns come from the kiln feed, not the fuel. It is not entirely clear why EPA continues to ignore the data that clearly demonstrates this fact.

Part Four, VII.D.2.
EPA states "The 0.4 ng TEQ/dscm standard is the level that all cement kilns, including data from non-hazardous waste burning kilns, are achieving when operating at the MACT floor control level or better." The February, 1996 Volume III Draft Technical Support Document for HWC MACT Standards contains four values exceeding the 0.4 ng TEQ/dscm standard with temperatures at or below 400F. The December, 1996 database does not contain these values. Furthermore, the PC MACT rule preamble recognizes the existence of one value from a non-hazardous waste burning kiln, but dismisses the data as an outlier.

In this same section, EPA states "...hazardous waste burning does not have an impact on dioxin/furan formation, dioxin/furan is formed post-combustion." It's too bad they haven't realized this is also the case for other cement kiln PICs.

Part Five, I.C.
Cement kilns and light weight aggregate kilns do not have to maintain compliance with the HWC/MACT rule when not burning hazardous waste. However, in order for cement kilns to have this option, they have to comply with the PC/MACT rule including all notification requirements. In essence, this places a permanent APCD inlet temperature limit on all cement kilns whether or not hazardous waste is being burned.

Part Five, I.E.
EPA acknowledges that exceedence of an operating parameter does not demonstrate an associated emissions standard was exceeded, but still asserts that it is an "actionable event for enforcement purposes."

Part Five, VII.B.1.
It is no longer satisfactory to monitor chamber pressure "continuously" (every 15 seconds). "It must detect and record pressure constantly...." Footnote 191 suggests every 50 milliseconds. A year's worth of values recorded every 50 milliseconds would have over 63 million recorded values! This requires 300 times the memory storage of 15 second "continuous" data. Section 63.1206(c)(5)(i) addresses combustion system leaks and states "maintaining the maximum combustion zone pressure lower than ambient pressure using an instantaneous monitor."

Part Five, VII.B.2.c
Regulatory officials can require shorter averaging periods than the one hour rolling hourly average used for most parameters.

Part Five, VII.B.3.
EPA assumes in the design of compliance tests that "...the source can readily control..." operating parameters for which limits are required. There are provisions for "conflicting parameters."

EPA excuses their departure from requiring MACT compliance testing under normal conditions and instead requiring worst case conditions for hazardous waste combustors using only what can be described as very flimsy logic.

Part Five, VII.B.5.
Rolling average limits are not initially affective until adequate data is accumu-lated. Start-ups after initial compliance require the use of older data, from just prior to shutdown, to be used to calculate rolling averages.

Part Five, VII.C.2.a.ii.
EPA, in what appears to be an effort to get industry to do its research, is requiring dual sampling trains for all Method 5i particulate sampling and all particulate CEMs correlation work for both Method 5 and 5i. They are recommending dual trains for all regular Method 5 determinations. They are waving federal particulate emission limits during particulate CEMs correlation testing if a CEMs correlation test plan is submitted with the comprehensive performance test plan.

Part Five, VII.C.5.
If particulate CEMs are installed prior to the time they are required for testing purposes, be certain to obtain a written agreement with the agency to prevent any data obtained from being used for enforcement purposes.

Part Five, VII.D.1.b.iv.
The rule does not require limits on certain parameters, but suggests that permit writers may impose limits on parameters such as minimum nozzle pressure, minimum viscosity and maximum solids.

Part Five, VII.D.1.f.
The rule does not establish limits on naturally occurring dioxin/furan inhibitors, but introduces the option for permitting officials to do so during the review of the comprehensive performance test plan.

Part Five, VII.D.3.c.ii.
The agency will require that "If you want to burn wastes with higher than historical levels of metals, you must incur the costs and address the hazards to plant personnel and testing crews associated with spiking metals into your feedstreams during comprehensive performance testing." This self-serving statement completely ignores the fact that it is EPA who has required metal spiking for nearly ten years without ever addressing the risks or safety issues involved. Furthermore, the "extrapolation" option introduced in the rule, which is so bound by conditions it is unlikely to be used, does not eliminate the need for spiking as they note earlier - from the same paragraph quoted above.

Part Five, VII.D.4.c.
The rule allows for an adjustment of the oxygen correction factor for CO and HC monitoring during startup and shutdown because the high O2 values artificially inflate the CO and HC values. For cement and LWA plants that do not burn waste during startup and shutdown, this may have little applicability.

Part Five, VIII.A.
EPA wants facilities to evaluate whether or not individual congener detection limits are adequate to verify compliance with the dioxin/furan standard prior to testing. This was, of course, the issue that a number of commentors raised regarding the inherent accuracy and precision of Method 23 at the proposed rule stage. Now the agency wants facilities to predict detection limits and perform theoretical worst case TEQ calculations in advance of testing.

The very next paragraph is a request for industry to do research for EPA by testing for mono- through tri- chloro dioxin and furan congeners. EPA hopes to develop a surrogate CEMs for dioxin using this data.

This section of the preamble also indicates that Method 26A must be used for HCl and Cl2 and provides the logic for this requirement - opposite the PC/MACT requirements. This is in direct conflict with the rule (63.1208(b)(5)).

Part Five, VIII.B.
This nonsensical statistical requirement is not even worded the same as the regulatory requirement in 63.1208(8). Neither makes technical sense, but they do appear to be a further attempt to ratchet down affective feedstream limits. This idea was never presented for public comment in preview drafts of the regulation. Similar concepts proposed for criteria on remediation projects have been thoroughly rejected by the technical community in the past.

Part Five, X.A.1.b.
EPA states that cement kilns may have increased metal emissions due to elevated chlorine levels in direct conflict with published data that shows chlorine input levels do not impact metal emissions.

Part Five, X.C.
Even if metals and chlorine in input streams are non-detectable, full emission testing will still be required.

Part Five, X.D.
As one of the "excuses" for not allowing SF6 use as a CEM for combustion conditions, EPA suggests that hydrofluoric acid might escape unreacted despite the obvious data and technical reasons for why this cannot occur from a cement kiln. Furthermore, they apparently are unaware of the extremely low levels of SF6 proposed for use in such devices.

Part Five, X.F.
EPA suggests that emission averaging for dioxins will not be allowed despite evidence that captured dust acts as a "sink" for the dioxins which are then released when the raw mill is turned off, no matter what the APCD inlet temperature. The preamble ducks entirely the issue of dioxins found in raw materials.

Part Six, IV.
EPA clarifies that all "blending activities prior to combustion are considered to be a treatment activity. This covers off-site blenders as well as on-site blending activities.

Part Six, VI.A.
Dioxin testing of CKD is now required. This requirement appears to be effective 30 days after publication of the rule. Depending on the compliance mode (statistical comparison versus health based limits), different congeners may need to be determined. A careful reading of the rule is required.

Part Six, VI.B.
EPA clarifies that all Part 266 Appendix VIII residues must be determined in Beville residues. This portion of the rule will become effective 30 days after publication of the rule.

Part Six, VIII.
EPA fails to realize that the ASME standard discussed does not address training for all the positions the regulations require. In fact, it specifically excludes maintenance mechanics, which the rule includes. It will be interesting to see how EPA plans to enforce a regulatory requirement to use a specific ASME standard that literally does not exist.