OCR Text |
Show FIELD EXPERIENCE In July of 1985, the MIS started to incinerate dioxin contaminated soils from the Denny Farm Site in McDowell, Missouri. This incineration consisted of mostly soils and some "trash" and continued into February 1986. "Trash" consisted primarily of contamjnated materials generated on site. Such materials included: plastic storage liners, protective clothing, wooden pallets, plastic and metal drums and other debris (18). In the operation of a large scale incinerator, the quantities of these products would not be significant compared to the normal waste feed. However, due to the smaller capacity of the MIS, these wastes were often fed in relatively large proportions. These materials were ram-fed in the rotary kiln intermittently. Gases evolved from these materials burned rapidly, producing sharp increases in kiln temperature and sharp decreases in excess oxygen. Before the implementation of the oes, the MIS had difficulty in burning these combustible materials smoothly, partly due to its relatively small capacity, but primarily due to the responsiveness of the incineration system (20, 21). Shown in Figure 3 are data obtained during the period of air burner operation while combustible materials were being burned. Even though the normal excess sec oxygen level was about 8% (dry), occasional feeding practice upsets caused puffs to occur as evidenced by the drop in the O2 concentration close to 0 percent and by the CO spikes measured at the exit of sec. Although the MIS is designed so that the waste feed is automatically cut-off whenever, at the exit of sec, the O2 level is below 4 percent and/or whenever the co level is about 500 ppm instantaneously or 100 ppm for 6 minutes, the waste materials already in the kiln can continue to release combustible gases for a few minutes. During this time the complete destruction of hazardous materials was not assured. Extreme caution by the MIS operators to limit the waste feed rate and to adjust the air flow rate generally avoided these upset conditions. This was still a significant operational constraint. However, immediately after the complete implementation of the oxygen feedback control feature, the response of the MIS to the transient combustion behavior associated with combustible solids was dramatically smoother. A typical example of how this system worked is given in Figure 4 and 5, where kiln and sec data logged during a period of two and a half hours, are shown. As illustrated in Figure 4, the oxygen level of the gas leaving the kiln was controlled to within ± 2 percentage points of the target of 9 percent O2 (wet) and the kiln temperature was reasonably smooth. Note in the lower chart of Figure 4 that the oxygen feed to the kiln was responding dynamically to the transient oxygen demand. As a result, as shown in Figure 5, the sec O2 level was consistently above 5% (dry), and carbon monoxide spikes were not detected (with the analyzer detection level at 2 ppm). Since the incineration system's responsiveness to these transient emissions was significantly improved, a much larger amount of combustible materials could be incinerated without feed shutdowns. Note, however, that this system required that the kiln oxygen analyzer be kept functional continuously to achieve such an effective control. This task was quite demanding, as discussed in the following section. A statistical study was conducted on the number of transient-caused shutdowns in two comparable periods before and after the oes. As summarized in Table 1, the study showed, with a high confidence level, that a very significant reduction in frequency (58%) was achieved by the oes in routine operation. - 9 - |