OCR Text |
Show 18 Modified Method 5 runs for semivolati1e organic speciation. Samples were taken for normal routine operation, as well as for a range of induced upset conditions, and during sootblowing. For several runs, samples were taken at up to four stations ranging from the near flame region out to the flue gas. Simultaneous mini-VOST and Modified Method 5 sampling was done at two stations to detect differences due to residence time or temperature during upset runs. Samples were also taken at specified intervals after waste firing was stopped to detect the amount and duration of any hysteresis emission "tail." Tables 2 and 3 summarize the overall test average OREs for volatile and semi volatile POHCs, respectively. The tables also include a total POHC plus PIC result. For these tables, PICs were specified as volatile or semi volatile chlorinated organics. This designation was based on the fact that the POHCs were all chlorinated. Some chlorine was also contained in the parent fuels however, and that could yield some nonwaste-related PICs. Also, the chlorinated POHCs could produce nonchlorinated PICs. These factors preclude a precise identification or closure of PICs. The overall mass-weighted ORE average was 99.9993 percent for the three volatile POHCs and 99.9997 percent for the semivolatile POHC. These destruction efficiencies are higher than the 99.998 percent average for the field tests, possibly because the pilot-scale system was intentionally clean of POHC deposition at the outset of the testing. The differences cited in ORE among the three fuels and the volatile POHCs were not major; and, in all cases, ORE was well above 99.99 percent, even when total chlorinated POHC plus PIC compounds were included as an aggregate. The ORE results on Tables 3 and 4 do not show a strong effect in going from normal operation to upset operation. This observation requires some interpretation in relation to what constitutes an upset. In this regard, Figures 4 and 5 show ORE as a function of average flue gas CO emissions (corrected to 7 percent oxygen) during a run. Figure 4 shows total chlorinated volatile compounds, POHCs plus PICs, and Figure 5 shows only the carbon tetrachloride results. These results show some correspondence between higher destruction at low CO and fewer high destruction results at high CO. These plots are useful to discuss the concept of upset operation. Some of the very low CO results shown in Figures 4 and 5 were operationally classified as upsets since they involved atomization maladjustments, waste flow surges, and other off-spec or transient settings. In some cases, these settings did not result in CO or other emission excursions and would not be classified as upsets if an emission based criteria were used to segregate "normal" operation from "upset" operation. Figures 6 and 7 show the normal and upset results when the upset designation is arbitrarily assigned to any run with an average CO level above 200 ppm. This procedure thereby reclassifies as normal the runs initially designated as upsets but with low CO. This approach to segregating runs by CO emissions shows a generally significant and consistent trend of lowered destruction efficiency resulting from upsets producing combustible emissions. This result for the pilot-scale tests is more pronounced and consistent than was observed in the field tests, possibly because of higher background emissions and hysteresis effects with the field boilers. In Figure 7, "C," "G," and "Oil refer to coal, gas, and oil, respectively, and "U" and "N" refer to upset and normal. The only situation that deviated from the trend cited 6 |