| OCR Text |
Show (ii) in a drop tube furnace the heating is via radiation from the heated wall with a conduction contribution from the gases if they are preheated. The nature of the heating-up curve is n ow different; with a tube temperature of 1673 K, the heating-up time to half the temperature is approximately 15 m s if by radiant heating only, and approximately 5 m s if both radiation and conduction are present. These heat up times should be compared with the reaction time. If the higher reaction rate is chosen (FG-DVC) the tVi « 0.5 to 1.0 ms. If the lower reaction rate is chosen, for example the flame values, then tt/a « 1 s. Clearly, the mode of rate of heating of the particle affects the measured devolatilisation rate, and thus different apparent rates are measured using different experimental designs. However, the use of these apparent rates in C F D modelling can often be beneficial, since these types of heat transfer mechanisms exist in the burner and some self-correction of data m a y occur. High speed video of coal particle combustion in flames was investigated in this laboratory and illustrate the following general points: (i) Particle size, or particle size distribution, is overwhelmingly important in pf flames. (ii) The devolatilisation rates amongst the suite of power station coals studied are estimated as shown in Figure 2. (iii) Disintegration of the particles during heating, and/or devolatilisation is important, vastly so in some plastic coals such as Pittsburgh 8. Volatile species As discussed previously, "volatiles" are often represented as a single species for which the combustion rate (steps 2 and 3) is controlled by mixing with oxidant. A n improved approach is to include measured or predicted volatile compositions and yields either within the code or as a pre-processor. One experimental approach of this type is pyroprobe-GC in which the coal sample is pyrolysed within the injector port of a gas chromatograph and the products analysed. In this w a y the gaseous and light tar hydrocarbons in the volatiles can be quantified. When combined with a nitrogen-sensitive detector, this technique provides useful data regarding partitioning of fuel-N between volatiles and char, and quantifies major N-volatile species. Typical pyroprobe-GC results show that most of the volatile nitrogen is evolved as hydrogen cyanide and methyl cyanide, although there is still uncertainty in the amounts of N H 3 and N2 evolved. Preliminary studies indicate that the H C N yield is directly related to the pyrrolic nitrogen content of the coal, as shown in Figure 3. In contrast, there is no obvious relationship between pyridinic or quaternary nitrogen in the coal and the H C N or C H 3 C N yields. These results m a y explain certain differences between N O x release from coals of similar ultimate and proximate analyses; fiiel-N released as H C N in the volatiles will have sufficient residence in the fuel rich zone of the burner to be reduced to N2. Thus coals with higher pyrrolic-N should give lower fuel (char) N O x. 5 |
