3.5. Estimating heating values by calculation from proximate analysis
Cordero et al. (2001) studied prediction of heating values on dry
basis for lignocellulosic and carbonaceous materials based on proximate
analysis. Correlation equations used are based on multiple
linear regression analysis with a least square fitting method. The
correlated equations were used to find out higher heating values
as [11]:
HHV ¼ 354:3FC þ 170:8VM ð1Þ
VM þ FC þ ASH ¼ 100 ð2Þ
where HHV represents the higher heating value (kJ/kg), Dry basis),
and FC represents fixed carbon and VM represents the volatile matter
content respectively, both in weight percent on dry basis:
VM þ FC þ ASH ¼ 100 ð3Þ
where ASH represents ash content in weight percent on a dry basis,
the HHV can also be estimated in terms of the measured parameters
as follows:
HHV ¼ 35430 183:5VM 354:3ASH ð4Þ
Using Eq. (1), heating value has been calculated and compared
with the experimental heating values (Fig. 5) Mean absolute error
between predicted and experimental values has been calculated.
Mean absolute error (MAE):
MAE ¼
HHVpredicted HHVexperimental
HHVexperimental
ð5Þ
The measured and estimated heating values for biomass species
were compared and found to be deviated by 1.7% to 0.85% for
proximate analysis as given in Table 7 and graphical representation
is shown in Fig. 5. The difference between the measured and the
estimated heating values is less than 0.5–2 MJ/kg for all the fuels
except for the soybean waste biomass fuel.