|Gossman Consulting, Inc.|
Continental Cement Trial Burn Strategy Follow-up
Operations Consultant - Gossman Consulting, Inc.
Director of Environmental Affairs - Continental Cement Company
David L. Constans
Engineering Consultant - Gossman Consulting, Inc.
Presented at the AWMA International Specialty Conference on Waste Combustion in Boilers and Industrial Furnaces April, 1997
The Continental Trial Burn strategy, presented at the 1995 BIF Conference, included the use of "data-in-lieu-of" from previous compliance testing conducted at the facility. Since the submission of the Trial Burn Plan and the 1995 presentation, Continental Cement has completed their two campaign trial burn. This paper will update the implementation of the Continental Trial Burn strategy.
Continental Cement Company (Continental) is located near Hannibal, Missouri on the Mississippi River. Continental operates a Portland Cement manufacturing facility utilizing hazardous waste as fuel. The kiln utilizes the wet slurry clinkering process. The kiln was constructed in 1966 and produces about 1,800 tons of Portland Cement clinker each day. Continental began utilizing hazardous waste fuel in 1986. Currently, Continental consumes about 5,000 tons each month of liquid hazardous waste fuel and about 2,500 tons each month of solid hazardous waste fuels.
As explained two years ago when the Continental Trial Burn
Strategy paper was presented, the purpose of the trial burn plan is
to document a comprehensive program to supply sufficient information to
the agency(s) to allow a facility operating permit to be written. This
information must be supplied in conformance to specific regulations,
CFR 270.22 and 270.66. Additionally, the EPA has supplied the permit
writers with a Trial Burn Guidance Document. This guidance document
addresses additional data requirements. The Trial Burn Plan and
Submittal Of "Data in Lieu Of" documentation submitted to the agency in
July of 1994 attempted to supply the information in accordance with the
regulations and the data request generated by the guidance document.
This paper picks up on what happened after the trial burn plan and
submittal of "data in lieu of" documentation was submitted.
ORIGINAL TRIAL BURN PLAN STRATEGY
The applicable regulations for BIF trial burns are found in CFR
270.22. In addition, CFR 270.66 addresses permits for boilers and
industrial furnaces burning hazardous waste. The facility must supply
documentation that it has met the organic emissions standard in
266.104, the particulate emissions standard in CFR 266.105, the metals
emissions standard in CFR 266.106, and the chlorine and hydrogen
chloride emissions standard in CFR 266.107. The facility must also meet
other Part B information requirements such as, automatic waste fuel
cut-off systems and Bevill exclusion requirements. This includes the
section on "data in lieu of trial burn."
Submittal Of Data In Lieu Of Trial Burn
The "Data in lieu of trial burn" section in 270.22 allows the
submittal of data acquired during "previous compliance testing of the
device in conformance with 266.103." In fact, the "data in lieu of
trial burn" can be from a "compliance testing or trial burn or
operational burns of similar boilers or industrial furnaces burning
similar hazardous wastes under similar conditions." What was learned in
the trial burn process was that even though CFR 270.22 (a) (6) states:
"The Director shall approve a permit application without a trial burn
if he finds that the hazardous wastes are sufficiently similar, the
devices are sufficiently similar, and the data from the compliance
tests, trial burns or operational burns are adequate to specify (under
266.102 of this chapter) operating conditions that will ensure
conformance with 266.102 (c) of this chapter," and even though
Continental collected a large volume of analytical data on input feeds
and output streams as well as process condition data to know and
document the similarity, there can still be trouble getting "data in
lieu of a trial burn" accepted. The language of the regulation is very
clear but as many of us have learned, interpretation by industry and
interpretation by the regulators can be miles apart. But, first a note
about what type of data was
not submitted as "data in lieu of a trial burn."
Emissions Data For Which Data in Lieu of Was Not Submitted
As explained in the 1995 paper, data in lieu of trial burn was not
submitted for dioxin/furan emissions. Continental had conducted this
emissions testing during the 1992 COC Testing and had demonstrated
compliance to the BIF emissions standard. Continental decided however
not to submit this data as "data in lieu of" primarily because at the
time of the COC Test Burn, the kiln was operating under the provisions
of the Tier III PICs standard and since then Continental began
operating under the Tier I CO standard. Continental believes that the
results of the dioxin/furan testing in 1992 would not be representative
of the current operation of the kiln.
Trial Burn Guidance Document
In September of 1993, the EPA issued a draft guidance document to its permit writers on criteria for trial burns. This was followed in April of 1994 with a heavily revised document which included revised attachments. The stated purpose of the document was to provide the permit writer with guidance on the criteria for a trial burn. That is, specific selection of emissions testing and analyses and operating conditions that would supply the data needed to perform an indirect, food chain pathway risk analysis.
EPA's concern, as stated in this document, is primarily the emissions of PICs. While there is some discussion of chlorine and metals, these discussions generally are concerned with their impact on PIC emissions.
The document included specific guidance on determining the list of PICs to be analyzed in the emissions. Also, there is guidance on waste feed selection and rates and trial burn operating conditions. It is not the purpose of this paper to critique the Trial Burn Guidance Document, such a critique has been done by Gossman Consulting, Inc. However, the guidance provided in the document includes contradictory operating conditions and a misunderstanding of cement kiln technology. Consequently, Continental was left to propose a test sequence that would acquire the necessary PIC emissions data.
What Continental had chosen to do was to submit to the agency a test strategy that would provide a set of testing campaigns that accurately represented the realistic operating extremes for this cement kiln. During the course of the year the kiln moves between two extreme operating conditions. The Continental design allowed one testing campaign at the extreme condition experienced immediately before the facility performs the annual major maintenance on the kiln. The second campaign was conducted during the operating conditions that exist immediately after this major kiln maintenance. This two campaign strategy was designed to achieve greater condition testing stability than is possible by trying to do it all over a few days time period. In the Trial Burn Plan submitted to the agency the two conditions were described as follows:
Kiln Conditions After a Major Maintenance Shutdown. During a major kiln shutdown, the necessary refractory brick will be replaced in the various zones of the kiln and worn pieces of chain will be replaced. Additional chain may also be added to improve heat transfer, if necessary. After a major shutdown, the kiln refractory will be at its maximum thickness and the mass of the chains will be at a maximum. Less heat will be lost through the refractory to the kiln shell. More efficient heat transfer will occur between the chains and the raw materials, thereby decreasing the temperature of the exit gas stream. Under the conditions that exist after a shutdown, fuel usage will be reduced as the thermal efficiency of the kiln is maximized. The kiln conditions after a major maintenance period represent one of the campaign conditions under which Continental proposed PIC testing. The campaign test periods were planned to reflect two extremes of the normal cycle of operating conditions, instead of subjecting the kiln to extreme conditions and stressing the kiln unnecessarily, regardless of its thermal condition during the test.
Kiln Conditions Prior to a Major Maintenance Shutdown.
During the normal operations of the kiln, the refractory and the chains
will experience wear. More heat will be lost through the kiln shell and
less heat will be transferred between the gases and raw materials in
the chain zone, as the mass of the chain decreases. Fuel consumption
will increase to meet the reaction needs. Since heat transfer between
the chains and the raw materials is at a reduced efficiency, the exit
gas temperature will be at a maximum. The kiln conditions prior to a
major maintenance period represent a second set of campaign conditions
under which Continental proposed to perform PIC testing. Again, the
campaign test periods were planned to reflect the extremes of the
normal cycle of operations conditions without unnecessarily subjecting
the kiln to artificially extreme conditions.
OVERVIEW OF THE TRIAL BURN PLAN AND SUBMITTAL OF DATA IN LIEU OF
One section of the original trial burn plan submittal included discussions on metal volatility, metals partitioning and the insensitivity of system removal efficiencies (SRE) to most operating parameters. These discussions were included to demonstrate the knowledge gained during the 1992 COC Testing of metals emissions under various kiln operating conditions of 24 different kilns. Continental felt this information was important in demonstrating that although each kiln may be unique, the cement clinkering process overwhelmingly predetermines the fate of metals that enter the system. Graphs supporting these statements were included.
Submittal of "data in lieu of a trial burn" includes a demonstration that the device and wastes for which this data was generated and the devices and wastes for which the permit is requested be "sufficiently similar." The kiln was not only sufficiently similar but was the exact same kiln. In addition, the proposed operating limits were the exact same ones established during the 1992 COC Testing and under which the facility had been operating since that time. Continental was not asking for new waste fuel maximum feed rate limits or increase of AWFCO operating limits. "Data in lieu of" was perfect vehicles for this.
One entire section of the trial burn plan was the submittal of a "data in lieu of trial burn." The data was presented in the same order as the emissions standards in BIF. That is, CFR 266.104, standards to control organic emissions through, CFR 266.107, standards to control chlorine and hydrogen chloride emissions.
One section detailed the proposed timing and plan of execution of
the PICs emissions testing. That is, one campaign conducted just before
the annual kiln maintenance and one campaign immediately after. The
first campaign was conducted in February, 1996 and the second campaign
was conducted in June, 1996. The timing of the first campaign and the
interpretation of the "data in lieu of a trial burn" is when the fun
began. Fun in the sense of retrospective comic rationalization.
DERAILING THE "DATA IN LIEU OF" TRAIN
Continental was moving forward with the planned first campaign of the trial burn, scheduled for the first week in February, as Region VII had indicated approval for the trial burn plan as submitted. It was just a few weeks before the scheduled first campaign of the trial burn in February when Continental heard from the state of Missouri. The state of Missouri responded negatively to the "data in lieu of a trial burn" section of the trial burn plan submittal. The MDNR claim was that since they did not have an official observer during the compliance testing that had produced the "data in lieu of a trial burn", that MDNR chose not to accept that data. They agreed that an MDNR observer had been there for the testing, but that it was not in the capacity of an "official" observer.
As mentioned previously, the first campaign was to be performed
towards the end of the brick life of the kiln resulting in the hottest
temperatures coinciding with the least amount of chains left in the
kiln. If the kiln had to go down prematurely, then a re-brick job would
put the trial burn schedule off at least a year. With time pressure on,
the state forced the removal of all references to "data in lieu of a
trial burn." This included the removal of a lengthy portion in the
general information section of the trial burn plan which included
discussions about HCl/Cl2, metals emissions including system
removal efficiencies and POHC destruction and removal efficiencies. In
addition, the state forced the removal of a forty-four page section
which presented the "data in lieu of" documentation. All of this
occurred at the end of January, less than one week before the first
campaign was scheduled to begin. It was last minute and the changes
were grudgingly agreed to under extreme time pressure to meet the trial
THE TRIAL BURN CAMPAIGNS
The first campaign of the Continental trial burn was conducted February 6 & 7, 1996. This was to occur when the chain section was the most worn and the kiln exit temperature would be expected to be the hottest. Test Day One parameters were for particulates, metals and HCl/Cl2. The Day Two test parameters were for dioxins and PICs.
The second campaign of the Continental trial burn was conducted June 18 & 19, 1996, which was approximately four and one half months after the first campaign. Day Three test parameters included dioxins and POHCs DRE. The Day Four test parameters included metals, dioxins and PICs under normal operating conditions. This was to assist in the risk assessment.
In the original Trial Burn Plan submitted in July of 1994, which
contained "data in lieu of a trial burn" test results, previous test
data would have been used in lieu of re-testing for various parameters,
including HCl/Cl2, metals emissions and POHC destruction and
removal efficiencies under worst case test conditions. This would have
been a much safer and more environmentally friendly trial burn which
could have also been conducted at a tremendous cost savings to
Continental. Instead, the state of Missouri denied "data in lieu of a
trial burn", ignoring some very clear language in the BIF regulations
and forced a complete re-test situation. A look at the 1992 metal
system removal efficiencies (SREs) as compared to the 1996 trial burn
SREs in Table 1 and Figure 1 demonstrates the logic of the "data in
lieu of" section of the BIF regulations. The SREs are essentially
The Continental Trial Burn strategy was two fold. To perform testing at two different times, campaigns if you will, that more accurately reflected two extreme kiln operating conditions and then to use "data-in-lieu-of a trial burn", primarily to conduct a more environmentally friendly compliance test and also to reduce testing costs. Continental was allowed to perform their trial burn using the two campaign strategy. That data, submitted in September 1996, is now undergoing agency review. The two campaign strategy appears to have been a success. However, while Continental and Gossman Consulting are quite confident that the data will be acceptable and ultimately a permit will be granted, it is too early to know for sure. Unfortunately in the end, and this is one area of the BIF regulations that is quite clear, the state of Missouri did not allow use of previous compliance testing data resulting in a compliance test that had the usual health and safety risks associated with full blown compliance tests and it was much more expensive than was necessary.
Concerning the safety and health concerns associated with metals
spiking, consider a citation from the proposed hazardous waste
combustor regulations. EPA refers to metals spiking as "expensive,
potentially dangerous to the testing crew that must handle the toxic
metals, and causes higher than normal emission rates during compliance
testing." (FR 17429, 4-19-96) One has to question the wisdom of denying
the opportunity to take advantage of a portion of the BIF regulations
that really do reflect common sense only to force metals spiking for
what certainly appears to be purely political reasons. In addition, the
accident incident rate for stack sampling contractors is estimated by
the industry to be 10 per 20,000 man-hours. This rate is about twice
the accident rate for chemical manufacturing (5.2 per 20,000
man-hours). In comparison, the accident rate for industries such as
crushed stone production and mining are 6.8 to 6.0. (Data on accident
rates were taken from the U.S. Bureau of Labor Statistics for 1993 and
1994.) Clearly then, the impact of manual stack testing must be
considered when the option not to test exists. Where is the concern for
human health and the environment?
Table 1. Continental Cement Company, LLC 1992 COC Versus 1996 Trial
Burn System Removal Efficiencies.
* HCl method does not exclusively represent HCl.
Figure 1. Continental Cement Company, LLC 1992 COC Versus 1996 Trial Burn System Removal Efficiencies Comparison.
"Data in Lieu of"