MODELING CHAR OXIDATION AS A FUNCTION OF PRESSURE ...
MODELING CHAR OXIDATION AS A FUNCTION OF PRESSURE ... MODELING CHAR OXIDATION AS A FUNCTION OF PRESSURE ...
effectiveness factor for the Langmuir rate equation with a correction function to improve its accuracy; 3) a pore structure model for calculation of the effective diffusivity, taking into account both Knudsen diffusion and molecular diffusion; and 4) general correlations for Nusselt and Sherwood numbers, which allowed the HP-CBK model to be used for both entrained-flow pulverized char oxidation and large-particle combustion in fixed beds. The HP-CBK model was evaluated by comparison with five sets of high-pressure experimental data: 1) graphite flake oxidation data (Ranish and Walker, 1993); 2) rough sphere combustion data (Banin et al., 1997a); 3) large-particle oxidation data (Mathias, 1996); 4) pulverized-char drop-tube data (Monson, 1992), and 5) TGA and FFB data from this study. Results showed that the HP-CBK model was able to quantitatively explain: 1) the effects of temperature, total gas pressure, oxygen mole fraction, particle size and gas velocity on reaction rates; 2) the change of reaction order with temperature and oxygen partial pressure observed by Ranish and Walker (1993) and by Banin et al. (1997); 3) the reaction order (typically about 0.7) and activation energy (typically 34 kcal/mol) observed in TGA experiments (Reade, 1996); and 4) the apparent reaction order of 0.5 typically observed at high temperatures, implying a true order of zero (Smith, 1982). Therefore, the Langmuir rate equation, when used with the appropriate effectiveness factor, seems to be satisfactory for modeling char oxidation over wide ranges of experimental conditions. In modeling the data by Mathias (1996) and those by Monson (1992), it was found that the Langmuir rate equation reduced to an intrinsic zero-th order equation for both cases. The intrinsic zero-th order equation implies an apparent order of 0.5 in Zone 126
II, in agreement with many observations (Smith, 1982; Mitchell et al., 1992). This suggests that an intrinsic m-th order (m = 0) is adequate for modeling char oxidation rates as a function of total pressure at high temperatures, although a global n-th order rate equation has been shown to be inadequate for that task. However, an intrinsic m-th order rate equation is inadequate for modeling char oxidation over a wide range of particle temperature since the intrinsic reaction order is typically non-zero at low temperatures and may change with temperature (Ranish and Walker, 1993). In the specific cases examined in this study, micropores can be neglected compared to macropores in modeling the effective diffusivity in the porous char matrix, in agreement with the experimental observation (reactivity correlates well with feeder-pore surface area, but not with micro-pore surface area; see Appendix). Principal Conclusions The principal conclusions drawn from this study are listed below: 1) The HP-CBK model, which uses the Langmuir rate equation and a corrected general asymptotic solution of the effectiveness factor and has three to five adjustable parameters depending on the data set, satisfactorily explains char oxidation rates over wide ranges of experimental conditions (including total pressure, temperature, oxygen mole fraction and particle size) for the following data sets: • graphite flake oxidation data (P = 2 to 64 atm; pure oxygen; T p = 733 to 814 K) • entrained flow data (67 μm; P = 1, 5 and 10 atm; T g = 1006 to 1469 K) • large particle data (ca. 6 to 9 mm; P = 1, 3 and 5 atm; T g = 825 to 1250 K) • small particle data (5 μm; P = 8 atm; T p = 1480 to 2850 K) • extrapolation of atmospheric TGA data to high temperature rate data 127
- Page 96 and 97: calculation uses a 7 × 7 × 7 matr
- Page 98 and 99: HP-CBK Model Development The HP-CBK
- Page 100 and 101: Effective Diffusivity The major obs
- Page 102 and 103: where r p1 and r p2 are the average
- Page 104 and 105: where r p1 is the macro-pore radius
- Page 107 and 108: 7. Model Evaluation and Discussion
- Page 109 and 110: experiments are non-porous, the rat
- Page 111 and 112: and 2850 K). For consistency with t
- Page 113 and 114: The value of the roughness factor w
- Page 115 and 116: = S int S ext D e r p a 2 2M C M O2
- Page 117 and 118: Reactor Head Flow Straightener Reac
- Page 119 and 120: the large size of the particle, and
- Page 121 and 122: taking into account convection, rad
- Page 123 and 124: 2.5x10 -4 2 /sec) 2.0 1.5 Rate (g/c
- Page 125 and 126: Table 7.5. The Experimental Conditi
- Page 127 and 128: The burnout and particle velocity d
- Page 129 and 130: The HP-CBK was used to predict the
- Page 131 and 132: TGA and FFB Data-This Study The rea
- Page 133 and 134: This equation can be derived as fol
- Page 135 and 136: q = A 1p e − E 1 p / RT P os 1 +
- Page 137 and 138: m obs = 0 at high temperatures) and
- Page 139 and 140: Currently the correlations between
- Page 141 and 142: 8. Summary and Conclusions The obje
- Page 143 and 144: 0.5 due to the contribution from th
- Page 145: Langmuir rate equation, the reactio
- Page 149 and 150: 9. Recommendations The predictive c
- Page 151 and 152: References Ahmed, S., M. H. Back an
- Page 153 and 154: Essenhigh, R. H., D. Fortsch and H.
- Page 155 and 156: Mehta, B. N. and R. Aris , “Commu
- Page 157 and 158: Szekely, J. and M. Propster, "A Str
- Page 159 and 160: Appendices 139
- Page 161 and 162: Introduction Appendix A: Experiment
- Page 163 and 164: detaching the flame from the burner
- Page 165 and 166: To study the effects of steam, CO w
- Page 167 and 168: times at heights of 1, 2, 4, and 6
- Page 169 and 170: analysis. The char reactivities (in
- Page 171 and 172: Table A.5. Moisture, Ash and ICP Ma
- Page 173 and 174: Table A.9. Elemental Analyses of Fo
- Page 175 and 176: temperature profile of the post-fla
- Page 177 and 178: Apparent densities 1.00 0.75 0.50 0
- Page 179 and 180: This observation is somewhat surpri
- Page 181 and 182: It is interesting to compare Figure
- Page 183 and 184: The N 2 BET surfacea areas and H/C
- Page 185 and 186: collected in the #4 reactor conditi
- Page 187 and 188: Rate (gC /g C remaining /sec) 1.6x1
- Page 189 and 190: close to zero, the accumulated erro
- Page 191: Appendix B: Errors and Standard Dev
effectiveness factor for the Langmuir rate equation with a correction function to improve<br />
its accuracy; 3) a pore structure model for calculation of the effective diffusivity, taking<br />
into account both Knudsen diffusion and molecular diffusion; and 4) general correlations<br />
for Nusselt and Sherwood numbers, which allowed the HP-CBK model to be used for<br />
both entrained-flow pulverized char oxidation and large-particle combustion in fixed beds.<br />
The HP-CBK model was evaluated by comparison with five sets of high-pressure<br />
experimental data: 1) graphite flake oxidation data (Ranish and Walker, 1993); 2) rough<br />
sphere combustion data (Banin et al., 1997a); 3) large-particle oxidation data (Mathias,<br />
1996); 4) pulverized-char drop-tube data (Monson, 1992), and 5) TGA and FFB data<br />
from this study.<br />
Results showed that the HP-CBK model was able to quantitatively explain: 1)<br />
the effects of temperature, total gas pressure, oxygen mole fraction, particle size and gas<br />
velocity on reaction rates; 2) the change of reaction order with temperature and oxygen<br />
partial pressure observed by Ranish and Walker (1993) and by Banin et al. (1997); 3) the<br />
reaction order (typically about 0.7) and activation energy (typically 34 kcal/mol)<br />
observed in TGA experiments (Reade, 1996); and 4) the apparent reaction order of 0.5<br />
typically observed at high temperatures, implying a true order of zero (Smith, 1982).<br />
Therefore, the Langmuir rate equation, when used with the appropriate effectiveness<br />
factor, seems to be satisfactory for modeling char oxidation over wide ranges of<br />
experimental conditions.<br />
In modeling the data by Mathias (1996) and those by Monson (1992), it was<br />
found that the Langmuir rate equation reduced to an intrinsic zero-th order equation for<br />
both cases. The intrinsic zero-th order equation implies an apparent order of 0.5 in Zone<br />
126