4.5. fletcher - tar+gas during pyrolysis of bl and biomasswhitty/blackliquor/colloquium... · 2003....
TRANSCRIPT
Prediction of Tar and Light Gas During Pyrolysis of Black Liquor and Biomass
Harland R. PondThomas H. Fletcher
Larry L. BaxterChemical Engineering Department
Brigham Young UniversityAdvanced Combustion and Engineering Research
Center (ACERC)
Wanted: Pyrolysis ModelModel to describe pyrolysis of biomass and black liquor fuelsModel based on individual components
CelluloseHemicelluloseLignin
Model suitable for inclusion in FLUENT and PCGC–3 CFD packages
Chemical Percolation Devolatilization (CPD) Model Approach
For coal, CPD looks at a lattice of aromatic rings (Fletcher et al.)Bridges between rings thermally rupture (kinetic scheme)Lattice statistics determine the fraction of free fragmentsVapor pressure of free fragments determines tar yieldFunction of coal type, heating rate, temperature and pressure
CPD LatticeStructural Parameters required
Number of attachmentsper cluster (σ + 1)Fraction of attachments that are initially bridges (po)Molecular weight per cluster (MCL)MW per side chain (Mδ)
CPD Parameters
Structural parametersMd,, po, MCL, σ + 1, elemental analysisFraction of char links (co)
Kinetic ParametersKinetics for two competing reactions
Model GoalsEach biomass fuel consists of cellulose, lignin and hemi-cellulose (hard- and softwood)Black liquor consists of Kraft lignin and residual carboxylic acidsDevelop kinetic and structural parameters for each componentUse mass-weighted distribution of CPD predictions for individual components
CPD Application to BiomassLinear structure of cellulose is modeled with coand poBridge weights for cellulose type components changed to model ether linkages
Lignin Chemical ModelLignin is similar to low rank coal with lattice structureConiferyl, sinapyland p-coumarylacids base cluster in the CPD model
Setting ParametersStructural parameters (po, MCL, Mδ and σ+1) set with 13C NMR measurements and known structuresKinetic parameters initially taken from previous research (Sricharoenchaikul, Fletcher, Serio, Azevedo)3 parameters need to be fitted to volatile yield data (co, Ec, ρ)
13C NMR Analysis on Black Liquor
Structural parameters for black liquor composite obtained from 13C NMR analysisParent samples sent to Pugmire and Solum at the University of Utah for analysisReacted samples from flat flame burner experiments were also sent for analysis (Webster, BYU, 2002)
O
*
O
O
*
OCH3
OH
R
**
R
OH
OCH3
*
?
OCH3 CH3
OCH2
H
*
OCH CH2CH
ss
Structural ParametersParameters obtained from theory for hardwood and softwood lignin and Kraft ligninCNMR parameters used for composite Black Liquor parameters
Parameter Lignin Lignin Kraft Lignin Black LiquorStructural Parameters Hardwood Softwood Composite
MWCl 207.5 186 195 296.5Md 39 34 21.5 206.43po 0.710 0.710 0.330 0.330σ+1 3.50 3.50 3.50 3.60
Pyrolysis DataVolatile yield data required for fitting the modelNeed data with heating rates and yields at different temperaturesData obtained
Lignin: Nunn, et. al. heated grid dataBlack Liquor: BYU Furnace data
CPD Prediction of Lignin Volatile Yields1000K/s 30s hold, data from
Nunn, T.R., et. al. Ind. Eng. Chem. Process Des. Dev. , 24:844 (1985).
0
10
20
30
40
50
60
500 700 900 1100 1300
Temp /K
% Y
ield
Gas Yield Tar Yield
CPD Model in Previous ResearchKinetic parameters for biomass and black liquor identified from original and previous research (Serio)Some structural parameters in previous applications of the CPD model not set correctly – misunderstood
Sheng & Azevedo (2000) – Biomass Sricharoenchaikul (2002) – Black Liquor
Sricharoenchaikul fit model only to tar yieldsMeasured tar yields were too low (not all tar accounted and carbon balance not closed)
Optimization and Data FittingOptimization required for three parameters (Ec, ρ, co)CPD model used with optimization program to find global optimized parameter values
Values of Ec = 0 gave best model resultsValue of ρ set to 3.9
Model fit to data through fine tuning kinetic parameters and co
CPD Component ParametersParameter Lignin Lignin Kraft Lignin
Structural Parameters Hardwood SoftwoodMW1 207.5 186 195Mδ 39 34 21.5Kinetic ParametersEb, kcal/mol 54.00 54.00 55.60Ab, s-1 2.60E+15 2.60E+15 2.60E+15
σb, kcal/mol 3.972 3.972 3.972Eg, kcal/mol 66.00 66.00 66.00Ag, s-1 3.00E+15 3.00E+15 3.00E+15
σg, kcal/mol 4.776 4.776 4.776ρ 3.9 3.9 3.9Ec, kcal/mol 0 0 0Ecross, kcal/mol 55.68 55.68 55.68Across, s-1 3.00E+15 3.00E+15 3.00E+15co 0.10 0.10 0.28po 0.710 0.710 0.330σ+1 3.50 3.50 3.60fst 0.60 0.60 0.60
Results
Lignin
Yields fit heated grid data well, secondary tar reactions not modeled
CPD Prediction of Lignin Volatile Yields1000K/s 30s hold, data from
Nunn, T.R., et. al. Ind. Eng. Chem. Process Des. Dev. , 24:844 (1985).
0
10
20
30
40
50
60
500 700 900 1100 1300Temp /K
Yiel
d %
Tar Yield Light Gas Yield CPD Tar Prediction CPD Light Gas Prediction
Secondary tar reactions to form light gas
Black Liquor
Yields fit well with furnace data, sodium – carbon reactions not modeled
Black Liquor Volatile Prediction with CPD100K/s with 10 min hold
data from BYU furnace data
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
650 750 800 850 900 950 1000 1100 1200 1300
Temp /K
Yiel
d
Total Volatility Data CPD+Model
Na2CO3 + 2C 2Na + 3CO
Na2CO3 + C 2Na + CO + CO2
Future Work
More volatile data needed for black liquor, cellulose and hemi-celluloseIntegration of a program calculating component composition from elemental analysisPredict volatile yield from weighted average of biomass components
Summary & ConclusionCPD modified slightly to describe biomassPrevious attempts to use CPD model were incompleteDeveloped chemical structural and kinetic parameters for Kraft lignin and lignin based on theory, literature review and curve-fitting
Summary & Conclusion cont’dTwo cases described accurately
Lignin (heated grid)Black liquor (furnace)
More data needed to see if parameters have general applicabilityMore study needed for cellulose, hemi-cellulose and carboxylic acid parametersNeed better understanding and model of secondary reactions