OCR Text |
Show 11 carbon in an ordered phase, is presented in Figure 5 [Davis et al., 1994]. This parameter has been normalized by the value determined for the most highly ordered sample, the residual carbon extract. For comparison, Fig. 5 also presents global char reactivities at various conversions in the form of the preexponential factor, A, extracted from the optical data using a model of gas-particle transport processes [Hurt and Mitchell, 1992; Hurt and Davis, 1994]. These data confrrm that reactivity does indeed correlate well with structural ordering for illinois No. 6 coal char. PARTICLE-TO-PARTICLE VARIATION OF CHAR REACTIVITY Another consideration involving char reactivity that has not been addressed in most attempts to predict unburned carbon levels is particle-to-particle heterogeneity. Optical measurements on single-particles indicate that there is a wide variation in burning rates and particle temperatures from one particle to the next. This effect can dramatically increase the time required for complete burnout due to the persistence of the least reactive and/or most dense particles in the sample. This effect is demonstrated in the simulation results presented in Figure 6. Starting with a distribution of reactivities determined for Pocahontas #3 coal [Hurt et al., 1994b], the combustion of a collection of 100 J..lm diameter char particles was computationally simulated in an isothermal gas environment of 1600 K containing 12 mole-% oxygen. For comparison, a burnout profile is shown for a uniform sample in which each particle has the mean reactivity. The uniform and heterogeneous samples burn at comparable rates up to about 70% carbon conversion, after which they diverge sharply. To reach 95% carbon conversion, the heterogeneous sample requires twice the residence time of the uniform sample. To reach 98% carbon conversion (corresponding to an LOI of 20% for a coal with 10% ash) the heterogeneous sample requires a factor of 3.4 greater residence time than the uniform sample. Although the coal and conditions chosen for this analysis were selected as a worst case, clearly, for the very high carbon burnout required for economic operation of commercial boilers, particle-to-particle variation in char reactivity could play an important role. CONCLUSIONS This paper demonstrates that single-particle laboratory experiments using well controlled combustion conditions and modern optical diagnostics can yield useful insights into the origin and causes of unburned carbon in the flyash of commercial pulverized coal-fued boilers. Our experimental results indicate that conversion dependence and, possibly, particle-to-particle variation of reactivity during char combustion can have a significant impact upon the time required |