Since the enactment of the Clean Air Act Amendments of 1990 by Congress, the U.S. Environmental Protection Agency (EPA) has a renewed interest in the regulation of oxides of nitrogen (NOx) generated by fuel combustion in boilers such as those used in the generation of electricity and industrial processes such as the manufacture of portland cement (Walters et al, 1999).

On Dec. 28, 1998, EPA promulgated a final rule to require 22 states and the District of Columbia to submit state-implementation-plans' revisions relating to emissions of NOx. On the same date, EPA signed a notification of intention to promulgate federal implementation plans (FIPs) to reduce the regional transport of ozone if a state within the Northeastern Ozone Transport Region fails to promulgate an acceptable revision.

The proposed FIP for the cement industry will require either the operation of an indirect-coal-firing system with a low-NOx burner or burning of tire-derived fuel during the "ozone season" (typically the summer months).

The justification of EPA's proposed FIP is the Alternative Control Techniques Document-NOx Emissions from Cement Manufacturing (ACT Document), prepared by the Office of Air Quality Planning and Standards of EPA March 1994 (EPA-453/R-94-004). The ACT Document reports a "20% to 30% reduction of NOx with the conversion to indirect firing with a low NOx burner," and it reports a "20% to 30% reduction of NOx with the mid-kiln firing of tires in long kilns."

In a letter to Carol M. Browner, administrator of EPA, Feb. 3, 1995, Russell MacMann and Mallory May wrote, "[the ACT document] does not employ 'tried and proven' scientific methods to demonstrate the achievability of the alternative control techniques. The emissions quantities contained in the report should flow from empirical observations and be expressed in calculated statistical parameters with acceptable confidence limits. However, they are not so expressed. Unlike the information in the document, achievable control levels should reflect research conclusions following a testing of null hypotheses. Such a procedure would involve collecting empirical data during both the untreated and treated phases in the application of control techniques." (MacMann and May, 1995)

In fact, the technical conclusions in support of EPA's proposed FIP are based upon interoffice memos, facsimiles, letters, and telecommunications. We feel there are no data or technical conclusions derived from adequately measured NOx emissions before and after a treatment, i.e., the use of a low-NOx burner or mid-kiln firing of tire-derived fuel presented in the ACT Document.

PSM International, Inc. has conducted an investigation of each of these applications to determine whether there is a reduction of NOx correlated with the application.

Low-NOx burner The potential effect of this modification was evaluated using NOx emissions data obtained from the Dragon Products Co. plant located in Maine, where an indirect-fired coal system with a Pillard Rotaflam Kiln Burner was installed in March 1995 during the annual maintenance program. Modifications were made to the coal system and a Pillard Rotaflam Kiln Burner was installed at Dragon's portland cement plant in Thomaston, Maine. Kiln operators trained and experimented with the new burner for six months following the installation. Beginning in September 1995, optimum operating conditions with the new burner were obtained.

Prior to its modification, the coal system at Dragon was a semi-direct system. The primary air from the system was 22% to 25% of the total combustion air. Following the modification to the coal system (installation of an indirect coal mill) and the installation of the Rotaflam, the primary air is 8% to 10% of the total combustion air.

NOx emission data, collected at the Dragon plant with continuous emission monitors, was analyzed using statistical t-tests. Analysis of NOx emission data from 1990 through November 1995 from the Thomaston plant was performed to determine if any significant reduction in NOx emissions had been achieved by the modifications made to the coal system and installation of the Pillard Rotaflam Kiln Burner.

The 1990 data was only available in 24-hour average measurements. The first comparison examined the difference between the post-installation mass emission rate with the pre-installation mass emission rates of NOx for 1990, 1991, 1992, 1993, and 1994.

The second statistical test compared the 1-hour emission rate values from the post-installation period (Sept. 6, 1995) with the pre-installation conditions (March 11, 1994 through Feb. 18, 1995). Seventy-two values were selected at random from each of these groups for the statistical test. This number was used so that the confidence of the test would be similar to that of the pre-installation data.

The null hypothesis for testing the difference between the pre-installation and post-installation means of the mass emission rates of NOx is:

HO-Post-Rotaflam-installation-NOx-emissions as lb/hr are equal to pre-installation-NOx-emissions; and

HA-Post-Rotaflam-installation-NOx-emissions as lb/hr are less than or greater than pre-installation-NOx-emissions (two-tailed test).

No significant differences were observed between the 1995 post-installation mean NOx emissions and the annual means for 1990, 1991, 1993, and 1994 based on 24-hour average measurements. The mean for 1992 was greater than the mean for the post-installation period. The Rotaflam burner and indirect coal mill was installed between Feb. 18 and March 13, 1995. The NOx emission data for the period March 13, 1995 to Sept. 5, 1995 could not be used because stable operation of the kiln and fuel system had not been established.

Comparison of the 1-hour average NOx concentrations for 1994, and the period Sept. 5, 1995 to Nov. 30, 1995 indicates that no statistically significant difference in NOx emissions resulted from the installation of the Pillard Rotaflam Kiln Burner at Dragon.

In both of these comparisons, there were no statistically significant differences in NOx emissions resulting from the installation of the indirect-fired-coal system and a low-NOx burner in terms of mass emissions per unit of time. Further investigation revealed that an increase of production had occurred following the installation. Therefore, there was a reduction in NOx emissions per unit of production. Consequently, at this plant, the annual rate of generation of NOx is not reduced as a result of this installation.

Mid-kiln firing of tire-derived fuel Pre-installation and post-installation NOx emission data were obtained from Ciment St. Laurent, where tire-derived fuel (TDF) as whole tires is injected into four long, dry kilns. These data were used to determine the statistical parameters necessary to test a null hypothesis as to whether NOx emissions have significantly (95% confidence interval and greater) changed for either application.

About 80,000 datum records, such as burning-zone temperature; kiln-feed rate; oxygen percentage; CO, SO2, and NOx concentrations in the stack gases were analyzed. Physical and chemical factors such as kiln-revolution rate, secondary-air temperature, and free-lime percent were inspected for ascertaining optimum operating conditions.

A comparison was made between the mean parameters of NOx before and after the commencement of burning of the tire-derived fuel. The null hypothesis for testing the difference between the means of the mass emission rates of NOx for pre-installation and post-installation is:

HO-Post-TDF-firing-NOx-emissions as lb/ton are equal to pre-TDF-firing-NOx-emissions; and

HA-Post-TDF-NOx-emissions as lb/ton are less than or greater than pre-TDF-firing-NOx-emissions (two-tailed test).

The mean for the rate of NOx emission before burning TDF is 11.8 lb/ton and after is 10.2; the standard deviation is 2.4 and 1.7, respectively; and the number of measurements is 1,154 and 760, respectively. The null hypothesis is rejected in favor of the alternative hypothesis. This computes to a 13.9 % reduction in the rate of NOx emission, which is significant at the 99 % confidence level.

Conclusion Based upon these two studies, it appears that the burning of TDF is the only effective alternate control technique proposed by EPA in its FIP. However, neither of these alternate control techniques achieve the percent of reduction reported in the ACT Document. Therefore, if EPA insist on the application of indirect-coal-fired system with a low-NOx burner, it will merely cause an increase equipment cost without effectively reducing NOx during the "ozone season."