OCR Text |
Show r 70 50 so1 1 S5 40 *| ro 30 •o a5 20 >- 10 o 9 8 7 ^ 6 4 • - , A - ' ? " " - I T A o.--' Pit. i i 1 ! i i A A A Weight Loss - Total Aerosol- • • • o .*"' ..--• Soot .>'"' - 8, 15,000 K/s, 1 atm i • i i t • CO " • ^ 0.2 0.4 0.6 Soot Fraction Figure 2. Evaluation of the predicted distributions of major products (top), oxygenated gases (middle), and hydrocarbons (bottom) for secondary pyrolysis of a Pit. #8hv bituminous coal (Chen, 1992). is within experimental uncertainty through the first half of secondary pyrolysis, but low by five percent during the later stages. However the sum of the yields of tar, soot, and oils (labeled as Total Aerosol" in Fig. 2) are predicted within experimental uncertainty throughout. Fig. 2 also shows that the soot fractions were accurately matched to measured values throughout this companson. The predicted yields of the oxygenated gas species exhibit the correct tendencies for uniform yields of C 0 2 and H20 throughout secondary pyrolysis, with progressively increasing yields of C O as tar expels its oxygen dunng soot formation. Quantitatively, the predictions are within experimental uncertainty for C O and C02 , but too low for H 2 0 by 2 vt. %. The predicted yields of the major hydrocarbons also exhibit the observed tendencies Methane yields pass through a maximum near the mid-point of secondary pyrolysis, whereas C2H2 and HCN yields increase monotonically. Predicted yields of C H 4 and HCN are within expenmental uncertainty throughout the last half of secondary pyrolysis. Whereas the predicted C2H2 yield surges throughout secondary pyrolysis, it reaches only about half the measured value by the end of the process Although H2 yields were not measured, the predicted values show that H2 is clearly a major fuel component in these combustible mixtures. PETROLEUM COKE DEVOLATILMZATION FLASHCHAIN was recently expanded to describe the thermal decomposition of petroleum cokes, which generally have much higher carbon contents and somewhat lower H/C ratios than low volatility coals. Cokes and coals have the same macromolecular architecture, composed of condensed aromatic structures bound together by various labile bndge structures. The same reaction mechanisms can be applied to cokes and coals, but the database of fuel properties especially proton and carbon aromatiaties. had to De expanded. Since cokes have much less oxygen than coal their volatiles release rates are also slower The predicted devolatilization and fuel-N partitioning of a database of 17 cokes appears in Fig. 3. This database was assembled from reported proximate and ultimate analyses for various forms of petroleum cokes ranging from green cokes to fully calcined electrode cokes. The simulations imposed a nominal heating rate of 104 K/s to 1600 K at atmosphenc pressure. The predicted total weight loss generally decreases for cokes with progressively higher carbon contents, although there is substantial sample-to-sample variability For cokes with less that 93 daf. wt. % carbon, the nominal weight loss is roughly 20 wt. % and the variations tend to be within z 5 wt. %, except for the sample with about 87 % carbon Weight loss falls off sharply for carbon contents between 93 and 96 %, and remains below 2 wt % for the highest carbon contents. The predicted tar yields closely mimic the variations in weight loss values, except that they fall off more gradually for progressively higher carbon contents and become negligible for carbon contents above 94 % The maximum predicted tar yield is only 10 wt. % The predicted partitioning of coke-N is much less sensitive to coke properties that the volatiles yields Between 20 to 30 % of coke-N is retained in the char for carbon contents under 95 %, The release of coke-N decreases for higher carbon contents, but remains significant even for cokes with the highest carbon contents. As with coal, the portion of voiatiie- N released in tar molecules is roughly proportional to the fractional tar yields. Consequently, the tar-N fraction never exceeds 0.10. |