Modelling Combustion in Wood Pellet Stoves.pdf

ANSYS Fluent Guide for Modelling Combustion in Wood Pellet Stoves

1. 1 © 2015 ANSYS, Inc.
Modelling Combustion in Wood Pellet
Stoves
Ren Liu
ANSYS UK Ltd
March 2015

2. 2 © 2015 ANSYS, Inc.
•Introduction
•Guidelines
–Using the Finite-rate/Eddy-Dissipation Model
•Model setup steps
•Material property inputs
•Recommended solution methods & controls
•Solution strategy
•Judging solution convergence
–Biomass Model
•Biomass characterisation
•Volatiles model
•Gaseous reaction mechanisms
–User-Defined Functions
•Simple examples
Index

3. 3 © 2015 ANSYS, Inc.
Introduction
•Main characteristics
–Solid fuel
–Packed bed combustion
–Multiphase flow
–In-Bed combustion/gasification
–Gaseous combustion in freeboard
–Turbulence
–Radiation
–Emissions, NOx, CO
GUNTAMATIC HeiztechnikGmbH
www.guntamatic.com

4. 4 © 2015 ANSYS, Inc.
•Finite-Rate/Eddy-Dissipation Model (FR/EDM)
–EDM for mixing limited reaction –effects of turbulence
–FR for kinetic effects
–Computationally robust and efficient
•Eddy-Dissipation Concept (EDC)
–Allows detailed kinetic mechanism
–More accurate prediction of CO, other intermediate species and
slow forming species such as NOx.
–Computationally expensive
•Can be sped up using acceleration tools
–ISAT
–Chemistry agglomeration (CA)
–Dynamic mechanism reduction (DMR)
Gaseous Combustion

5. 5 © 2015 ANSYS, Inc.
•Discrete Phase Model (DPM)
–Lagrangianapproach
–Particle size distribution is accounted for naturally
–Allows inclusion of various heat & mass transfer processes
•Including heterogeneous reactions
–Great UDF customisation flexibility
–Not valid for packed bed
•Requires custom models
•Eulerian-Granular multiphase model
–Valid for packed bed
–UDF necessary to customise heterogeneous reactions
–Computationally inefficient for polydisperseparticles
•Each size class requires a granular phase
–Computationally expensive overall
Multiphase Flow Modelling

6. 6 © 2015 ANSYS, Inc.
Using Finite-Rate/Eddy-Dissipation Model

7. 7 © 2015 ANSYS, Inc.
•N species mixture
•N-1 transport equations of species mass fraction
??????෪ഥ??????�
�
??????�
+
??????ഥ??????෪��
෪��
??????��
=
??????
??????��
ഥ??????�
���
??????෪�
�
??????��
+෪??????
�
•The mean reaction rate is calculated from the smaller of
–EDM rate
•A, B: model constants
–Kinetic rate: ??????=??????�
??????
�??????�
−�
�
??????
��
�
�
�
�
�??????
�
Transport Equation

8. 8 © 2015 ANSYS, Inc.
•Launch Fluent and read in a mesh file
•Activate Energy Equation
•Activate a turbulence model
•Activate a radiation model
•Activate Finite-rate/Eddy-Dissipation model
•Define the reactions & properties of the mixture and fluid materials
•Define cell and boundary conditions
•Adjust solution methods and controls
•Create monitors
•Initialise the solution
•Solve
Overview of Simulation Procedure

9. 9 © 2015 ANSYS, Inc.
•Typically one of the 2-equation models
–SST k-w, or
–Realizable k-e
•Scalable Wall Function is recommended
Turbulence Model

10. 10 © 2015 ANSYS, Inc.
•Discrete Ordinates (DO) model
–Typical Angular & Pixelationdivisions:4x4x2x2
–WSGGM for mixture absorption coefficient
•UDF option available
Radiation

11. 11 © 2015 ANSYS, Inc.
Finite-Rate/Eddy-Dissipation Model

12. 12 © 2015 ANSYS, Inc.
Mixture Material

13. 13 © 2015 ANSYS, Inc.
Mixture Components

14. 14 © 2015 ANSYS, Inc.
Reactions

15. 15 © 2015 ANSYS, Inc.
Fluid Material

16. 16 © 2015 ANSYS, Inc.
•Mixture Density
–Incompressible-ideal-gas law
•Mixture Cp
–Mixing law
•Fluid Cp
–temperature-dependent
•Polynomials or the like
•Kinetic theory
•UDF
•Thermal conductivity, viscosity, mass diffusivity
–Constant if accurate data unknown
•Acceptable for turbulent flow
–Polynomials
–Kinetic theory
Material Properties

17. 17 © 2015 ANSYS, Inc.
•Pressure-Velocity coupling
–SIMPLE typically used
–Coupled sometimes beneficial
•Pressure
–Second Order: now default
–PRESTO! For hex mesh
–Standard: works for most cases
•DO first-order upwind
–Second order upwind less stable
•Turbulence
–First order by default
–Second order upwind can be used usually
•All other equations
–Second order upwind
Solution Methods

18. 18 © 2015 ANSYS, Inc.
•Pressure: 0.7
•Density: 0.8
•Momentum:0.3
•Turbulence: 0.4 -0.8
•Species: 1
•Energy: ~1
–Initially <1, e.g., 0.9
–Final stage: 1
•DO: ~1
–Initially <1, e.g., 0.9
–Final stage: 1
Solution Controls

19. 19 © 2015 ANSYS, Inc.
•Set up ALLmodels required for the final simulation
•Save the case file
•Deactivate Volumetric Reaction in Species Model dialog box
•Disable solution of all equations except Flow, Turbulence & Energy
•Reduce energy URF to, say, 0.9
•Initialise the flow field as air
•Calculate a non-reacting flow solution
–A fully converged solution is not necessary
–A reasonably established flow field is often adequate
–Typically 100 to 200 iterations are required for the segregated
solver
•Save the case & data
Solution Strategy: Step 1: Non-Reacting Flow Solution

20. 20 © 2015 ANSYS, Inc.
•Re-activate the reaction model setting
•Enable solution of all equations
–Alternatively, keep solution of DO equations disabled
•Ignition
–Mark cells in a region where flame is expected
–Patch, e.g., T=1800K,
•Necessary because kinetic rates are temperature-dependent
–Patch a small value, e.g., 0.01 for mass fractions of reactants and products, except O2
•Otherwise reaction rates will remain zero because of EDM’s way of working
•Set energy & DO URF to ~0.9, perform, say, 5 gas flow iterations, check
the flame is sustained
•Perform, say, ~200 iterations
•Save the case & data
•Restore energy & DO URF to 1
•Iterate to convergence
–If calculation becomes unstable, slightly reduce energy & DO URF, e.g. 0.97
Solution Strategy: Step 2: Reacting Flow Solution

21. 21 © 2015 ANSYS, Inc.
•Convergence should not be judged by the residual values alone.
•The following should be checked:
–Residuals for energy & DO < 1e-6 and < 0.001for all other equations
–Good global mass balance, e.g., < 0.1%
–Good global heat balance, e.g., < 0.1%
–Monitored parameters of interest do not change more than
acceptable values with more iterations
–Distributions of flow variables in the domain are realistic and make
sense.
Judging Convergence

22. 22 © 2015 ANSYS, Inc.
Biomass Chemistry Model

23. 23 © 2015 ANSYS, Inc.
•Proximate analysis
–Volatiles, char, ash, moisture
•Ultimate analysis
–C, H, O, N elemental composition
•Volatiles representation in CFD
–Option 1: single artificial species: C
xH
yO
zN
m
–Option 2: mixture of CH4, CO, CO2, H2, H2O, N2
•Char burnout
–CO, CO2 or a combination of both
Biomass Characterisation

24. 24 © 2015 ANSYS, Inc.
Common Global Reaction Mechanisms
Yang Y.B., et al, Combustion of a single particle of biomass, Energy & Fuels 2008, 22, 306-316.
Yin, C, et al, Co-firing straw with coal in a swirl-stabilized dual-feed burner: Modelling and experimental
validation, BioresourceTechnology 101 (2010), 4169-4178.

25. 25 © 2015 ANSYS, Inc.
User-Defined Functions

26. 26 © 2015 ANSYS, Inc.
Example 1: Customise Boundary Conditions
#include "udf.h"
DEFINE_PROFILE(profil_T,thread,i)
{
real r[ND_ND];
real y;
face_tf;
begin_f_loop(f,thread)
{
F_CENTROID(r,f,thread);
y = r[1] / 0.1;
y = MIN(y, 1.);
F_PROFILE(f,thread,i) = 1000. -200*y*y;
}
end_f_loop(f,thread)
}

27. 27 © 2015 ANSYS, Inc.
Example 2: Customise Volumetric Reaction rate
#include "udf.h"
DEFINE_VR_RATE(user_vr_rate, c, t, r, mw, yi, rr_k, rr_t)
{
real T = C_T(c,t);
real Keq= 1.;
real tau = TRB_TIM_SCAL(c,t);
real tau_min= 1e-5;
if (tau < tau_min)
*rr_t*= (tau/tau_min);
if ( ! STREQ(r->name, "reaction-9") ) return;
if (T < 600.) T = 600.;
if (T > 2000.) T = 2000.;
Keq= 2408.1/T + 1.5350*log10(T) -7.452e-5 * T -6.7753;
Keq= pow(10, Keq);
/* alternative expression */
/* Keq= exp(-4.33 + 4577.8/T); */
*rr_k/= Keq;
}