OCR Text |
Show RESULTS AND DISCUSSION Results are presented in two parts. First, the measured puff is considered without correction for any dilution effects. This represents what must be handled by the afterburner and is of interest to operators and designers of rotary kiln systems. Next, two sets of data are presented that have been corrected for dilution effects. These represent the rates of volatile release from the waste. Set 1: Dilution Effects Not Considered Full quadratic models with respect to air and oxygen flows were fitted to the data from the first set of experiments (Table I) using weighted least squares, where the weights were equal to the number of replicates. These models, obtained using the preprogramrned procedures in SAS (Statistical Analysis System), should not be considered unique since other models might attain fits essentially identical to those reported here. Once determined, these models can be described pictorially through response surface plots. The response surface plots are unique to the data they describe. Figure 3 presents these contour plots for the dependent variables: volatile hydrocarbon peak area (peak area) (Figure 3a), volatile hydrocarbon peak concentration (peak height) (Figure 3b), carbonaceous particulate filter weight (filter weight) (Figure 3c), and CO peak concentration (peak CO) (Figure 3d) versus oxygen and air flows. Full quadratic models were able to explain total response variances (r2) of 0.9485, 0.8992, 0.9993, and 0.9610 for peak area, peak height, filter weight, and peak CO, respectively. The variance is a measure of the dispersion of the response, defined as the square of the standard deviation. Evident from Figure 3 is that, for all the response variables, transient puffs are reduced at higher air and oxygen levels. Peak 9 |