# Combustion Combustion simulation simulation in in laboratory laboratory scale scale starved starved air air incinerator incinerator

Faculty of Mechanical Engineering
Institute for power, process and environmental engineering
Laboratory for combustion and environmental enrgineering
Combustion simulation in
Combustion simulation in
laboratory scale starved air
laboratory scale starved air
incinerator
incinerator
F F.. K Ko ok ka alljj
Vransko, September 2007Outline
• Governing equations of turbulent reacting flow
• The closure problem of turbulent combustion
• Turbulent combustion models
• Practical case of turbulent combustion simulation
applying CFX programme in second stage of pilot
scale starved air incinerator
• Numerical control of “3T” combustion conditions
(Temperature, Turbulence and Time)
• ConclusionsSumary of RANS equations for reacting flow
Continuity:
∂ρ ∂
+ (ρυ )= 0
j
∂t ∂x
j
Momentum:
∂ ∂ ∂p ∂
′ ′
(ρυ )+ (ρυυ )=− + f − (τ +ρυυ )
j j i υi ij j i
∂t ∂x ∂x ∂x
j i j
Bussinesq approximation of Reynolds streeeses
 
∂υ
 
2 ∂υ ∂υ
j
k i
 
′ ′  
ρυ h = δ ρk+η −η +
j ij t t
 
 
3 ∂x ∂x ∂x
 k  j i
 
where
2
k
η = ρC
t η
εSumary of RANS equations for reacting flow
Transport equations for k andε :
   
∂υ
 
∂ ∂ ∂ η ∂k ∂υ ∂υ
j
t i i
 
( )  
(ρk)+ υ k − η+ =η + −ρε
 
j t
 
 
σ
∂t ∂x ∂x ∂x ∂x ∂x ∂x
 
j j  k  j j i j
   
2
 
   
∂υ
 
∂ ∂ ∂ η ∂ε ε ∂υ ∂υ ε
j
t i i
 
( )  
(ρε)+ υ ε − η+ = C η + −C ρ
 
 
j 1 t 2
 
 
σ
∂t ∂x ∂x ∂x k ∂x ∂x ∂x k
j j  ε j   j i j 
 
   
 
Energy:
∂ ∂ ∂p ∂
′ ′
(ρh)+ (ρυ h)− + (q +ρυ h)=I
j j j T
∂t ∂x ∂t ∂x
j j
Turbulent heat flux:
η ∂T
t
′ ′
ρυ h =− c
j p
Pr ∂x
t jSumary of RANS equations for reacting flow
Mass species transport equation:
 
∂ ∂ ∂ ∂ξ
k
 
(ρξ )+ (ρυ ξ )− Γ =I
k j k k ,eff ξ
k
 
∂t ∂x ∂x ∂x
j j j
 
Effective diffusivity:
η η
t t
Γ =ρD + =Γ +
k,eff km k
Sc Sc
t t
Equation of state:
N
ξ
k
p= ρR T
m ∑
M
k=1
kSource terms
• Source term for energy equation:
N
0
I =− ΔH ω
T ∑ f ,k k
k=1
• Sorce term for mass species transport equation:
I =M ω
ξ k k
k
• Formation/consumption rate of reacting flow components
Combustion
d[X ]
k
′′ ′
ω = =(ν −ν )R
k k k k
modells
dt
• General form of chemical reaction
N N
k
f
→
ν X ν X
′ ′′
∑ ∑
k k ← k k
k
b
k=1 k=1Closure problem of turbulent
combustion
( )
R ≠ R T ,ξ k = 1,...,N
k k k
Mean reaction rate term is not equal to the term based on
mean temperature and mean species concentration.
R
k
TTurbulent combustion models
• Kinetic rates by modified Arrhenius expressions or Finite Chemical
Reaction rates – FCR
• Eddy Break-up modell – EBUM
• Eddy Dissipation Combustion Modell – EDCM
• PDF Mixture Fraction based models
• Multi step reaction models based on EDCM or EBUM enables the
calculation of PIC (CO, H , NO, ...) applying additional reactions
2
• Combined models (ie. EDCM/FCR), etc. Finite Chemical Reaction rates – FCR
N N
′ ′′
ν ν
k k
R =k [X ] −k [X ]
k f∏ k b∏ k
k=1. k=1
Modified Arrhenius equation:

k =k +k
f ,b f ,b f ,b
2
 
  ′
E E T
a a
 
k =k (T) 1+ −1 + L
 
f ,b f ,b
2
 
2R T R T T
m m
   
 
2
2 2 3

′   ′ ′    ′
E T E E T −T E E T
a a a a a
′     
k = k (T) + −1 + −1 +L .

f ,b f ,b
2 3
    
R T T 2R T R T T 6R T R T T

 m  m  m  m  m 

depends only on temperature and activation energy.Eddy Break-up Modell – EBUM, 1
Chemical reaction rate prpotional to mixing frequency
1
R ≈
k
t
t
Mixing is controled by turbulence
t
t
1 ε
=
t k
t
Eddy life time, or time for an eddy to break-up
Reaction rate proportional to fluctuation of species
2

R ≈ξ
k kEddy Break-up Modell – EBUM, 2
Final form of EBUM reaction rate
ε
2

R =−C ρ ξ
k EBU k
k
C is an empirical constant which has to be fitted to the experimental data
EBU
• Model for fast chemistry
• Derived for single step chemistry (Fuel + oxidizer → Products)
• Provides reasonable results for heat release and main species
• Simplified chemistry required
• Empirical constant changes from case to caseEddy Dissipation Combustion Modell - EDCM
Chemical reaction rate proportional to minimum
concentration of participating species
 
[X ]
k
 
R ≈ min
k
 
ν '
 k 
Different expressions are used for reaction rate of reactants
(fuel and oxidizer) consumptionn
 
ε [X ]
k
 
R = A min
k
 
k ν '
k
 
and products formation
 
[X ]M
∑ k k
 
ε
P
R = AB
 
k
k ν '' M

k k
 
 P Eddy Dissipation Combustion Modell – EDCM
• Model for fast chemistry.
• Derived for single step chemistry (Fuel + oxidizer → Products).
• Can be extended to multi-step chemistry either 2 or 5 steps
enabling, the claculation of some PIC’s.
• Simplified chemisty required.
• Two constants which must be fitted by experimental data from case
to case.
• Applicable to non-premixed, partially premixed and premixed
combustion.
• Not very sensitive for non-premixed combustion.
• Provides reasonable results for heat release and main species.
• Can be linked to multiphase reacting flows.Practical case of combustion simulation in pilot scale
starved air incinerator applying CFX 5
• (CH + CO)/air diffusion flame
4
• 3-D bounded geometry
• Reduced number of species considering single step
reaction (Fuel + oxidizer → Products).
• Different combustion models offered by CFX 5 have
been tested to find the most suitable one for our case
– MIB (involved just in CFX 4)
– EBUM
– EDCM (1,2,5 step)
– Combined model EDCM/FCR (1,2,5 step)
• Numerical control of “3T” combustion parameters
applying EDCMPilot scale two stage starved air icinerator
750
1500Secondary chamber -
thermoreactorBoundary conditions
Parameters that were changed:
Volatile gases from
• secondary and tertiary air
primary chamber:
amount and their inlet direction
- amount
- composition
Secondary air inlets:
• speed and composition of
- temperature
- amount
combustible gases
- direction
- temperature
Parameters that were not changed
but may influence the combustion:
Tertiary air inlets:
- amount
• geometry
- direction
- temperature
• capacity
Wall: dQ/dt = 0
• additional air inlets
• additional fuel
Combustion
• insulation
products outlet
• feedingThermoreactor discretization
• Main geometry: 2.4 m x 0.75 m
• Inlets and outlet
• Computational grid with 275.674 control
volumesAverage velocity comparison Average temperature comparison
2100
EBUM
MIB
EDCM (1 step)
1900
EDCM (2 step)
EDCM (5 step)
1700
EDCM/FCR (1 step)
EDCM/FCR (2 step)
EDCM/FCR (5 step)
1500
X
X
1300
X
X
1100
900
0 -0,5 -1 -1,5 -2 -2,5
Longitudinal axle of thermoreactor [m]
Temperature [K]Experimental and numerical data comparison-EDCM
1280 K 1360 K
measured
temperature
in secondary
1230 K 1310 K
chamberAverage fuel mass fraction comparison
0,6
EBUM
MIB
EDCM (1 step)
0,5
EDCM (2 step)
EDCM (5 step)
0,4
EDCM/FCR (1 step)
EDCM/FCR (2 step)
EDCM/FCR (5 step)
0,3
0,2
0,1
0
0 -0,5 -1 -1,5 -2 -2,5
Longitudinal axle of thermoreactor [m]
Fuel mass fraction [CH4]Carbon monoxide mass fraction comparison
0,02
EDCM (2 step)
EDCM (5 step)
0,018
EDCM/FCR (2 step)
EDCM/FCR (5 step)
0,016
0,014
0,012
0,01
0,008
0,006
0,004
0,002
0
0 -0,5 -1 -1,5 -2 -2,5
Longitudinal axle of thermoreactor [m]
Carbon monoxide mass fraction [CO](“3T”) combustion conditions control
(using EDCM)
• Temperature: set by legislation, influences on
the speed of destruction of organic pollutants,
depends of calorific value of gas
• Residence time: set by legislation, influences
on the speed of destruction of organic pollutants
together with temperature, depends on amount
of gases (at the same geometry)
• Turbulence: ensures proper mixing of volatile
gases with air throughout the whole
thermoreactorInfluence of temperature
on combustion quality
Calorific value - temperature of combustion:
35 % CH 55 % CH 75 % CH
4 4 4Influence of residence time
on combustion quality - 1
Amount of volatile gases - residence time:
v =0,25 m/s v =0,38 m/s v =0,5 m/s
1 1 1Influence of residence time
on combustion quality - 2
Amount of volatile gases - residence time:
v =0,25 m/s v =0,38 m/s v =0,5 m/s
1 1 1Influence of turbulent mixing
on combustion quality
Mixing - turbulence:
secondary air tertiary air
• velocity: projected 6 m/s projected 4 m/s
4 m/s 6 m/s
4 m/s 8 m/s
• direction: projected projected
o o
projected + 10 projected + 10
o o
projected + 15 projected + 15o o
Projected direction Projected + 10 Projected + 15
Projected
velocity:
s: 6 m/s
t: 4 m/s
Velocity
Velocity
Changed
velocity:
s: 4 m/s
t: 6 m/s
field
field
Changed
velocity:
s: 4 m/s
t: 8 m/s
.o o
Projected direction Projected + 10 Projected + 15
Projected
velocity:
s: 6 m/s
t: 4 m/s
Temperature
Temperature
Changed
velocity:
s: 4 m/s
t: 6 m/s
field
field
Changed
velocity:
s: 4 m/s
t: 8 m/s
.Conclusion
• EDCM and EBUM are well established and simple to
implement
• Single step combustion models (EDCM, EBUM) are
convenient just for combustion macro-parameters
calculation (temperature, velocity and combustion products
field)
• Require empirical input which has to be fitted from case to
case
• They can be used for fast reactions only
• For slow chemistry reaction rates are not mixing controled
• For some PIC’s concentration calculation (i.e. CO, NO, etc.)
multi step combustion models should be applied but there
are still limitations for other minor species (OH, HC, soot,...)
or complex chemical processes
• It is possible to improve EDCM extending it by chemical
kinetics