Page 2 of 3
Technical Comments on the Revised Application to Modify International Paper, Ticonderoga Mill's Title V Permit to Conduct a Two-week Tire-Derived Fuel trial.
Submitted by the Board of People for Less Pollution to the NYDEC
4. The revised permit application does not accurately characterize/ analyze the composition of the raw TDF. No statistical analysis of the composition of the TDF is presented in the application. The inability of IP-Ti to adequately characterize the composition of TDF compromises any attempt to predict emissions during the test burn. Statistical tests and sampling methods exist to characterize heterogeneous materials. IP-Ti has failed to provide a statistically valid analysis of TDF composition. This information must be provided before any test burn. If a trial should occur, statistically-valid TDF feed samples should be taken from the material entering the boiler at regular intervals. Furthermore, a feed specification should be developed for TDF and a careful feed monitoring program put in place to track feed properties. It is important to have good values and ranges for the key feed parameters: heat content, ash at the very least and hopefully also carbon, hydrogen, nitrogen, and oxygen. For example, knowing not only the heat content but the variability in the heat content of the TDF fed would aid in assessing the heat content of the feed mix and the degree to which the variability in the TDF might affect the boiler's performance.
5. The revised permit application assumes a 73% control efficiency for all metals. It does not make sense to treat zinc the same as the other metals in terms of removal efficiency. The previous application assumed an 86% removal efficiency for all metals. The current application assumes a 73% removal efficiency for all metals based on a 1999 stack test for nickel emissions from oil. The actual nickel control efficiency during the 1999 stack test varied between 51 and 88% depending on nickel feed rate. No statistical analysis beyond a simple average is presented for the assumed 73% control efficiency. Zinc is molecularly dispersed in the tires (it is added in the manufacture of synthetic rubber - it is a homogeneous reaction catalyst that is not removed from the final product). Most of the metal from the radials in the tires is removed to make TDF. Much of the other metals are essentially all 'chunks' of residual radial. It does not make sense to treat zinc the same as the other metals in terms of removal efficiency. Indeed, the other metals could be expected to be more concentrated in the grate area and to be disproportionately removed with the grate ash. However, the molecularly-dispersed zinc may provide a nucleation site in the formation of particulates and could be expected to be removed at a much lower efficiency. Specifically, how does nickel particle size compare to the particles containing the molecularly-dispersed zinc in the tires and to the chunks of residual radials? Were other control efficiencies calculated during the 1999 stack test? Those should be tabulated in the permit application, along with any estimations/information concerning the size of the particulates containing that metal. These data would enable better estimation of potential control efficiencies to expect when burning TDF.
6. The revised permit application does not adequately consider the potential for high zinc emissions during the test burn. Zinc oxide is homogeneously dispersed in the rubber; therefore burning the rubber will release very small particulates of zinc. These can do various things, including agglomeration but also including adsorbing products of incomplete combustion (PICs) and then passing right through the current air pollution control equipment (because of their very small size) and exiting the smokestack. Note the following zinc emissions results for ALL the pulp/paper facilities in the 1997 EPA Report, listed by facility> letter (before --> after 1 to 30% TDF, in lb/MM BTU): L: 15 --> 250; M: 2.5 --> 16; O: 0.715 lb/hr --> 0.85 lb/hr; P: 0.25 --> 40 to 86; Q: 0.23 --> 3.4. These are all (except for Facility O) MAJOR increases in zinc emissions. Moreover, the outlier facility, (facility O) is the only one with an ESP! This makes sense, since an ESP would be able to remove a significant fraction of these very fine particulates that would just go right through a scrubber. (Note that Facilities L, M, P, and Q all have Venturi scrubbers (most also state that they, like IP-Ti, have multiclones in front of the scrubber). And Venturi scrubbers are generally more efficient at particulate removal than IP-Ti's 'spinning basket-style scrubber (the latter design is optimized for SOx, not particulate, removal).
7. Boiler temperature will not be adequately monitored during the proposed test. Dioxin-family compounds are known to form in the post-combustion region of boilers. Temperature should be measured in the post-combustion region. Where is temperature measured in the boiler? In how many places? How? It is logical to assume that adding TDF to the fuel will affect the combustion. For example, a larger percentage of the fuel is being added at the bottom of the boiler when exchanging tires (fed below) for oil (fired in at maybe 10-20 feet above the grate). Therefore, one would expect the boiler temperature to be hotter in the lower region and cooler in the upper region (relative to the base case). Is this temperature difference significant? It has been communicated to us by IP that they feel that the hotter burn temperatures near the grate will promote more complete burning of the wood chips and therefore actually reduce products of incomplete combustion (PIC's). However, it is also well known that increasing boiler temperature increases NOx. It is also well known that dioxin family compounds form at intermediate temperatures (ca. 400 to 800 deg. F - in the post-combustion region). Could operating with TDF cause more NOx in the lower region and more dioxin family compounds in the post-combustion region? This possibility has not been seriously addressed in the application.