FLASHCHAIN has been developed to predict yields and product characteri
stics from any coal for any operating conditions. This evaluation demo
nstrates the model's utility for the usual situation where the ultimat
e analysis is the only sample-specific information available. It also
identifies the key reaction centers in coal as its structural componen
ts called labile bridges. Their elemental compositions are grossly dif
ferent than the analogous whole-coal properties, showing much stronger
rank dependences and a much higher degree of sample-to-sample variabi
lity. In light of these findings, it is inconceiveable that bridge con
version rates are rank-independent. Parameters in the rate law for bri
dge conversion in FLASHCHAIN are now explicitly related to the element
al compositions of bridges. The (O/C)(B) ratios are the best regressio
n variable for the rate constants because oxygen is the most effective
promoter of pyrolytic decompositions. The (O/H)(B) rates are best for
the selectivity coefficient between scission and condensation into ch
ar links because oxygen promotes crosslinking but hydrogen addition to
broken bridge fragments stabilizes them. These extensions are evaluat
ed in comparisons against a database of 27 coals that span all ranks f
rom lignite to anthracite, for heating rates from 5 to 5000 K/s, ultim
ate temperatures to 1300 K, and pressures from vacuum to 70 MPa. In fo
ur out of five cases, predicted total and tar yields are within experi
mental uncertainties. The model is also used to rigorously define nomi
nal devolatilization rates for diverse coal types and broad ranges of
operating conditions. Nominal rates have very low activation energies,
proving that heat and mass transport resistances are not responsible
for the low values because this theory is completely free of these con
siderations. Whereas nominal rates are rather insensitive to coal type
variations and independent of pressure, they vary in proportion to ch
anges in heating rate.