PETITIONED HEALTH CONSULTATION NO.2
THERMAL DESTRUCTION FACILITY
LOCK HAVEN, CLINTON COUNTY, PENNSYLVANIA
The comments in this consultation are based on ATSDR staff review of the following documents and the comments provided by an ATSDR modeler :
Drake Chemical Site Incinerator Trial Burn Risk Assessment, Volumes I - III and Appendices .
Trial Burn Plan for the Drake Chemical Superfund Site's Mobile Hazardous Waste Incinerator, Volumes I and II .
Facility Operations Description
The Drake incinerator is a transportable rotary kiln incinerator system owned by OHM Remediation Service Corporation. The soils to be treated are stockpiled in a feed building maintained under negative pressure. The soils are conveyed at a maximum feed rate of 60 tons per hour to the rotary kiln (primary combustion chamber). The organic contaminants are volatilized and the combustion process begins in the rotary kiln which is operated at 1,200°F. The organics and any partially combusted compounds complete the combustion process in the secondary combustion chamber, which is operated at 1,800°F. A thermal relief valve (TRV) is at the top of the secondary combustion chamber. Cyclones between the kiln and the secondary combustion chamber remove larger particulate matter (entrained soil) and combine it with the treated soil from the kiln and the fly ash from the secondary combustion chamber. The combined treated soil and fly ash are sprayed with water to control fugitive emissions and are conveyed to a covered storage area, where they are stored until analysis is completed.
An evaporative cooler after the secondary combustion chamber reduces the gas temperature before the gas goes to the baghouse for removal of smaller particulate matter. The gases are further cooled and scrubbed in a venturi quench before going to a caustic scrubber, which removes and neutralizes any acid gases before releasing the exhaust gases to the atmosphere from the 150-foot stack. Fly ash from the evaporative cooler and from the baghouse is transported in an enclosed conveyor and sprayed with water to control fugitive emissions before it is discharged in the ash storage area. See Figures 1 and 2 .
According to Drake Chemical Site Incinerator Trial Burn Risk Assessment, Volume II, Appendix 3B  and the Air Quality Equivalency Document issued by the Commonwealth of Pennsylvania , the following conditions initiate the automatic shutdown of all waste feed materials to the kiln:
Kiln-firing hood pressure greater than -0.1 inch water column for ten seconds; or greater than zero inches water column instantaneously.
Secondary combustion chamber discharge temperature less than 1,800°F.
Carbon monoxide concentration in stack gas exhaust greater than 100 parts per million by volume (ppmv) for one hour average or greater than 500 ppmv instantaneously or loss of the signal.
Stack gas flow greater than 60 feet per second (or a modified value to be determined from results of the trial burns).
Failure of both secondary combustion chamber burners.
Caustic/water scrubbing solution flow to either of the packed bed acid gas scrubbers at a rate less than 450 gallons per minute.
Acid gas scrubbing solution pH less than 6.0 for one hour.
Differential pressure across the baghouse less than 1.0 inch water column.
Nitrogen oxide emissions greater than 300 parts per million by volume (daily average) or loss of signal.
Kiln discharge gas temperature less than 1,000°F instantaneously or the hourly rolling average less than 1,200°F.
Failure of both kiln burners.
Kiln dry ash conveyor stopped.
Kiln rotation less than 0.4 revolutions per minute.
TRV on secondary combustion chamber not closed.
Solids feed rate hourly average greater than the maximum rate demonstrated during the trial burn.
Even though the waste feed will be discontinued when these conditions occur, the flames will be maintained and the air pollution control equipment will continue to operate; therefore it is unlikely that there would be any increase in stack emissions.
The following conditions will initiate an emergency system shutdown that stops the waste feed, fuel feeds, and induced draft fan and opens the TRV:
Induced draft fan failure.
Evaporative cooler discharge gas temperature greater than 500°F.
Electric power failure.
Acid gas scrubber inlet temperature greater than 250°F.
When the TRV opens, there will be an initial puff of emissions consisting of the gases in the secondary combustion chamber and the rotary kiln. The organics released will likely consist of fully combusted decomposition products, partially combusted organics, and organics just vaporized from the soil. Because the gases will bypass the air pollution control equipment, there will be no removal of metals, particulate matter, or acid gases. As the temperature, pressure, and air flow decrease, the TRV stack emissions will decrease rapidly. The prediction is that they will approach zero within 25 minutes of the TRV opening.
When the gas temperature at the exit of the secondary combustion chamber is greater than 2,600°F, the system automatically shuts down all waste feed and the fuel to both combustion chambers. The hot gases will continue to flow through and be treated in the air pollution control equipment for removal of the particulate and acid gases; however, there will be an increase in organic emissions when the fuel is shut off, causing the flame to go out.
When the kiln discharge temperature exceeds 2,200°F, the system will automatically shut down all waste feed and the fuel to the kiln only; the flame will be maintained in the secondary combustion chamber. Because combustion will continue to occur in the secondary combustion chamber and the gases will still be treated in the air pollution control equipment, stack emissions are not expected to increase.
The Pennsylvania Air Quality Equivalency Document, Section 7, dictates continuous monitoring
and recording of the key operating parameters. All of the conditions that will cause an automatic
waste feed shut off (AWFSO) or TRV opening are continuously monitored and recorded. The
document also dictates that the following stack emissions be continuously monitored and that data
|CO||carbon monoxide concentration,|
|THC||total hydrocarbon concentration,|
|O2||oxygen concentration, and|
|NOx||nitrogen oxides concentration.|
Continuous monitoring and recording of all key operating conditions and parameters allow those who oversee the incinerator operations to verify how the facility was operated and what conditions triggered upset conditions at any point.
The trial burn operating conditions were designed to represent maximum input of soil taken from the contaminated areas and fed to the incinerator operating at minimum combustion temperature and maximum combustion gas flow rate . These conditions represent worst case operating conditions. If the stack emissions measured during the trial burn and risk burn do not pose an unacceptable health risk to the community, then operation of the incinerator under normal operating conditions should not pose a health risk.
The trial burn plan appeared to be well designed and specified standard or updated draft Environmental Protection Agency (EPA) sampling and analytical procedures and quality control measures. If the trial burn and risk burn were conducted in accordance with the plan, the data should be adequate for evaluating the public health implications of the incinerator.
The approach to estimating emissions from the incinerator for the trial burn period was conservative and very thorough.
- The term "hazard index" should be defined and explained on first use. The public may not be familiar with the term.
The air quality dispersion modeling analysis performed in the trial burn risk assessment used the Industrial Source Complex Version 3 (ISC3) model, CALMET/CALPUFF models, and the INPUFF model to predict the ambient air concentrations of contaminants expected to be released from the stack during the trial burn. Dispersion of emissions caused by a possible upset condition, which was defined as a TRV opening, were also modeled to determine the potential ambient impacts. The modeling of stack emissions predicted the monthly average ambient concentrations and ground-level deposition in the area extending 13 miles (20 kilometers) from the Drake site. Modeling predicted one hour and monthly average ambient concentrations of contaminants for the upset release condition. Modeling also provided one hour and eight hour average ambient concentration predictions for carbon monoxide .
Because the modeling was conducted before the incinerator was constructed and operated on site, the designed physical parameters for the incinerator were used. The air quality modeling analysis reflected surface data from the Williamsport, Pennsylvania, National Weather Service station; upper air data from the Pittsburgh, Pennsylvania, National Weather Service station; and precipitation data from Philipsburg, Pennsylvania. Although on-site meteorological data exist, EPA determined the data inadequate for use in the modeling analysis because of instrument malfunctions and improper sensor exposure, which together resulted in unrepresentative wind speed data for the Lock Haven area .
The following comments assume that the same general modeling protocol will be used for the risk assessment for full operation except (1) modeling will cover the full year; (2) the modeling will reflect local meteorology as much as possible; (3) stack emissions data from the trial burn will be modeled; and (4) the trial burn worst case operational data, such as stack gas temperature and velocity, will be used.
- Modeling for potential acute effects should reflect the maximum hourly and maximum daily ambient concentrations.
- A comparison of the Williamsport wind data and the actual site-specific wind data for the months that the data are available will provide additional scientific support to the use of the Williamsport meteorological data. A time histogram or auto-correlation may be useful for comparing the two data sets.
- The use of five years of meteorological data will help assure that likely worst case public exposure concentrations are part of the evaluation.
- Are there other sources of local meteorological data, such as the Piper Memorial Airport, the state college, other industrial air sources, river and stream flow stations, etc.? A definitive statement as to the unavailability or unsuitability of any such data (and what attempts were made to locate other local data sources) would strengthen the full operation risk assessment.
- The uncertainty and sensitivity analyses do not present quantitative assessment of the air quality modeling results. One approach to quantifying the uncertainty and sensitivity issues would be to exercise one of the models at a few locations, systematically varying the various input assumptions. This would yield an ensemble of model predictions that could help health scientists assess the values used in the risk assessment.
- Volume III, Section 5.3, of the risk assessment assumes that the particulate released during a TRV event would be primarily metals (not organics) and that they would be relatively large and thus would deposit on site. This assumption is contrary to the information in Table 2.4-2 of Volume I of the risk assessment and Volume II, Section 3B.3.3. These sections say all the metals except barium, beryllium, and chromium will be 100% volatilized during normal and upset conditions. Barium will be 50% volatilized, and beryllium and chromium will be 5% volatilized. It seems reasonable to assume that most of the metals will be in the vapor state when exiting the TRV and will be dispersed in the plume rather than being large particles that fall on site.
- EPA should provide references to support the following assumptions: (1) the assumption that after an initial puff of one minute duration, the emissions will decrease exponentially to zero within 25 minutes and, (2) the assumption that during the first minute the destruction and removal efficiency will decrease from 99.99% to 99%.
Note: This health consultation evaluates the incinerator operational plans, trial burn protocols, and emissions modeling. Health Consultation #3 will provide an evaluation of data from actual sampling and monitoring of the incinerator during the trial burns and risk burns. In Health Consultation #3, ATSDR will also evaluate the ambient air data for public health implications.