Rotterdam, Oct. 2013 1
*Scotton P., *Rossi D., **Barberi M., **De Toni S.
*University of Padova, Department of Geosciences
** Barberi Srl, Trento (Italy)
COMSOL CONFERENCE ROTTERDAM 2013
Heat-Accumulation Stoves: Numerical
Simulations of Two Twisted Conduit
Configurations
Rotterdam, Oct. 2013 2
COMSOL CONFERENCE ROTTERDAM 2013
• Description of the Physical Problem;
• Hydrodynamic and Heat Transfer Equations;
• Results of twisted refractory pipe
on a vertical plane
refractory pipe spacially
twisted
INDEX
Rotterdam, Oct. 2013 3
COMSOL CONFERENCE ROTTERDAM 2013
Components of a
heat accumulation
stove
Burning Process of Woody Material
time T
he
rma
l P
ow
er
[KW
]
V, T ↑↑ V, T ↓↓
He
at sto
rag
e
an
d r
ele
ase
P [
W]
time [h]
HEAT STORAGE
in the refractory
HEAT RETURNED
to the environment
• Description of the Physical Problem
Rotterdam, Oct. 2013 4
Examples of heat accumulation stoves
Historical heat accumulation
stove “Sfruz” (Valle di Non,
Trentino, Italy).
COMSOL CONFERENCE ROTTERDAM 2013
Classical stove
Contemporary
stove
Modern stove
• Description of the Physical Problem
Rotterdam, Oct. 2013 5
COMSOL CONFERENCE ROTTERDAM 2013
The driving force acting
on the flue gases
Z = 0
Z = H
ΔHg)P,(TρPP
ΔHg)P,(TρPP
affHB
aaaHA
ΔHgρ-ρΔHg)P,(Tρ-)P,(TρPP faaffaaaBA
air supply
• Description of the Physical Problem
Rotterdam, Oct. 2013 6
Hydrodynamic flow aspects
Sharp Curve – Turbulent motion
IRe = 28400 x/D = 1.4
0.1
0.4
0.7
1.0
1.3
-5 15 35
x/d [-]
E [m
]
A
Apii
curv
e
DE
DP/g
COMSOL CONFERENCE ROTTERDAM 2013
Temporal evolution of Reynolds number
Resis
tance n
um
be
r
[-]
• Description of the Physical Problem
Rotterdam, Oct. 2013 7
• Physics and Equations
Transport Equations, k- model
Reynolds-averaged Navier–Stokes eq.
FuuIpuuuut
u T
''
0)(
u
t
k
k
T Pkkut
k
kCP
kCu
tk
T2
21
Turbulent energy eq.
Turbulent Dissipation energy eq.
where
2kCT
+
COMSOL CONFERENCE ROTTERDAM 2013
Rotterdam, Oct. 2013 8
Wall Functions
Wall dw
06.11
ddd w
ww
u
influence of
choosen mesh
on the results
COMSOL CONFERENCE ROTTERDAM 2013 • Physics and Equations
Rotterdam, Oct. 2013 9
Heat Transfer
Heat transfer is guaranteed by three terms:
i
ix
Tkq
conduction
TTAhq sconvection
4sTAq radiation
Qput
p
T
TSqTu
t
TC
p
p
:
0
v
t
Equation of mass conservation
Equation of heat transfer
heat flux by conduction
.. the conserved property is
the total energy not the heat rquTkHu 0 heat flux by radiation
COMSOL CONFERENCE ROTTERDAM 2013 • Physics and Equations
viscous heating
total energy flux
Rotterdam, Oct. 2013
Straight Steel Pipe
COMSOL CONFERENCE ROTTERDAM 2013
Thermotechnical characteristics
stainless
steel
black
steel
thickness [mm] 0.2 2.0
emissivity [-] 0.1 0.95
conductivity [W/mK] 17 50
C-shaped refractory pipe
Thermotechnical characteristics
refractory calcespan
density [kg/m3] 2550 600
heat cap. [J/kgK] 859 1000
conductivity [W/mK] 3.16 0.15
emiss. [–] 0.95 0.70
• Physical and numerical experiments
Rotterdam, Oct. 2013 11
Twisted Refractory Pipe on a vertical plane: numerical model
Mesh properties
Boundary conditions: 1. Mass flow rate + Temperature at the inlet face;
2. Pressure at the outlet face;
Mass flow rate
and temperature
at the inlet
section.
Inlet
section
3. Convective cooling on the outer surface: ;
4. Surface to Ambient Radiation: ;
COMSOL CONFERENCE ROTTERDAM 2013 • Physical and numerical experiments
Radiation in participating media – Discrete ordinate method: S2 Scattering = 0; absorption k = 1.524 E-3 [1/cm]
Modest,
‘83
Rotterdam, Oct. 2013 12
Twisted Refractory Pipe on a vertical plane: numerical results
t= 76 min.
Residual combustion
activity inside the first
vertical stretch ?
COMSOL CONFERENCE ROTTERDAM 2013 • Physical and numerical experiments
Rotterdam, Oct. 2013 13
Twisted Refractory Pipe on a vertical plane: numerical results
S1 S2 S3
CENTER
St
Sp
Dx Sx
Temperature gauges distribution
inside an instrumented section
COMSOL CONFERENCE ROTTERDAM 2013 • Physical and numerical experiments
Rotterdam, Oct. 2013 14
Twisted Refractory Pipe on a vertical plane: numerical results
MASS BALANCE
ENERGY BALANCE
ERROR 15%
COMSOL CONFERENCE ROTTERDAM 2013 • Physical and numerical experiments
Rotterdam, Oct. 2013 15
Measurement positions
of the temperature at the
section 13.
Temperature and mean velocity
of the combustion air (pipe
diameter = 0.16 m) .
Calculated flue gas mass
discharge at section 01.
Thermotechnical properties of the
materials.
Refractory Pipe spacially twisted: physical model
COMSOL CONFERENCE ROTTERDAM 2013 • Physical and numerical experiments
Rotterdam, Oct. 2013 16
Refractory Pipe spacially twisted: numerical model
Boundary conditions:
3. Convective cooling on the outer surface;
1. Calculated mass flow rate + Measured temperature at the
inlet face;
2. Pressure at the outlet face;
4. Surface to Ambient Radiation;
INLET
OUTLET Tethraedical
meshes
Radiation in participating media – Discrete ordinate method: S2 Scattering = 0; absorption k = 1.524 E-3 [1/cm]
COMSOL CONFERENCE ROTTERDAM 2013 • Physical and numerical experiments
Rotterdam, Oct. 2013 17
Refractory Pipe spacially twisted: numerical results
t= 62 min.
Mesh M2
Mesh M3 Mesh M3
COMSOL CONFERENCE ROTTERDAM 2013 • Physical and numerical experiments
Rotterdam, Oct. 2013 18
Refractory Pipe spacially twisted: numerical results
MASS BALANCE
ENERGY BALANCE
Mesh M3: ERROR 30%
Mesh M2: ERROR 17%
Mesh M3
Mesh M3
COMSOL CONFERENCE ROTTERDAM 2013 • Physical and numerical experiments
Rotterdam, Oct. 2013 19
Conclusions
•The numerical results have been obtained considering the thermal- and the hydrodynamic
equations, including buoyancy forces (Boussinesque approximation). Radiation has been
considered both outside and inside the pipe;
• The software is able to describe correctly the mass balance until the time of occurrence of a
numerical instability;
• The energy balance depends strongly on the mesh refinement. A refined mesh, realized
with the use of a boundary layer, can induce an anticipated instability;
• The pressure variations are not significantly far from the measured ones (this experimental
measure is very difficult);
• The mean velocity values along the pipes are coherent with the mean temperature values;
• The calculated mean temperature values along the pipes are generally much higher than
the measured ones; the measured temperature variation inside the pipes are much larger
than the measured ones; temperature variations inside the pipe, both measured and
calculated, decrease increasing the distance along the pipe;
• The use of a boundary layer is not always the best choice, having an influence on the
reduction of the stability period.
COMSOL CONFERENCE ROTTERDAM 2013
Rotterdam, Oct. 2013 20
What’s more …
• Use of Comsol in the final stove
configuration, where a radiant
covering envelops the twisted pipe;
• Need of a higher computing
power (now Dell T7500 24GB Ram)
to reduce the energy loss and to
hope to give technical importance
in the design process. At the
moment, due to the large
calculation time, of the order of a
week, is almost useless;
• Another road: decoupling the hydrodynamic equations relevant to the pipe from the thermal
equations relevant to the refractory;
COMSOL CONFERENCE ROTTERDAM 2013
Thank you for your
attention
Rotterdam, Oct. 2013 21
• Description of the Physical Problem
General Features of global system
combustion air supply
refractory
flue gas
rad. heat exchanges
rad. heat
exchanges
conv. heat
exchanges
conv. heat exchanges
walls of room
Barberi Srl
COMSOL CONFERENCE ROTTERDAM 2013
Rotterdam, Oct. 2013 22
COMSOL CONFERENCE ROTTERDAM 2013
Software version: Comsol 4.3
Computing machine: Workstation Dell T7500 equipped with
two Intel R Xeon R Processor X5550 and
24GB DDR3 1333MHz ECC-RDIMM.
Rotterdam, Oct. 2013 23
Temperature gauges distribution inside an instrumented section
COMSOL CONFERENCE ROTTERDAM 2013
Rotterdam, Oct. 2013 24
THE MASS FLOW RATE HAS BEEN DETERMINED
BY MEANS OF VELOCITY AND TEMPERATURE
MEASURED UPSTREAM DEL COMBUSTION
CHAMBER.
COMSOL CONFERENCE ROTTERDAM 2013
The technical regulation
prEN15544 has been used.
prEN 15544 – “One off
tiled/kachelofen stoves - Calculation
method”
Rotterdam, Oct. 2013 25
COMSOL CONFERENCE ROTTERDAM 2013
Velocity measurement
Propeller anemometer
Pitot
Hot wire anemometer
Rotterdam, Oct. 2013 26
COMSOL CONFERENCE ROTTERDAM 2013
Analogical piezometer
Furness FC0332
Testo 521
Pressure measurement
Rotterdam, Oct. 2013 27
COMSOL CONFERENCE ROTTERDAM 2013
Thermocouples k;
Pt100;
Pt1000.
Temperature measurement