Flow over a composite wall

1
Flow over a composite wall

• Aurélien CUIF
• Marianne SJÖSTRAND

2
Contents

• Introduction/definition
• Flat plate theory
• Results on the Composite wall
• Unsteady simulations
• Conclusions

3
Introduction / definition
Flow over a composite wall (coupled
convection/conduction). How is a house wall
constructed? Where is the most heat transferred
when the wall is subject to different convection
situations. Validate against flat plate theory.
Wind 20 m/s or 10m/s
4
Flat plate theory

• A wall can be considered as a flat plate
• Not so simple to model the flat plate under
  different convection situations

?? Troubles

• For a 20 m/s wind (72 km/h) transition occurs at
  about 1 m from the leading edge
• We want to avoid transition zones but we dont
  want to restrict ourselves to low winds.

5
Fully turbulent flow

• When we want to model strong winds like 20 m/s,
  we will use a turbulent inlet profile to avoid
  transition problems
• We can compare different models with different
  grids

Coarse grid
Fine grid
6

• The coarse grid looks quite the same as the fine
  one
• Close to the leading edge we notice a difference
• The coarse grid Nusselt seems to have a less
  physical behaviour
• The values of the Nusselt number between 31.4 and
  31.9

7

• As expected the temperature and velocity profiles
  has the same thickness

• Almost same profiles predicted by each model

gt We will use the coarse grid (good enough)
8
Laminar flow (Slow wind)
Blasius equation
At different x positions the profiles are
similar!
9
Does the profiles fit well with the Blasius
equation?
Polynomial interpolation gives a rapid way to get
a function and its derivatives
10-3
Fit very well!
10
And fits very well with the correlation of the
profile width
11
Results on the Composite wall
Composition of the wall
Material Properties
12
Grid and boundaries Description
Interior North BC
Interior East BC
Interior East BC
Inlet BC
Outlet BC
Outside South BC
13
Contour plot of temperature for the composite
wall

• It is cold and windy outside but the wall keeps
  the temperature stable inside
• The highest gradients are  absorbed  by the
  glass wool
• The wall seems to be colder close to the inlet
• Nusselt number rises quickly
• Small rise far downstream

Nusselt number along the wall
14
(No Transcript)
15
Contour plot of temperature for the wall without
insulation
Insulation removed!!!

• It is a bit colder inside
• We can see that close to the inlet the wall is
  colder like in the non insulated case
• We see that the nusselt number looks more like it
  should be (flat plate theory)
• As the wall is not insulated the wall is more
  sensible to the flux imposed by the external flow

16
Velocity profile
Room
Wall without isulator
Pylon,
Exterior

• Shape of temperature different from the previous
  case
• Turbulence increased downstream
• Wall colder than the insulated case

Temperature profile
17

• Analogy between the velocity field and the
  Nusselt number
• The velocity decreases in the front of the pylon
  due to stagnation
• It accelerates between the pylon and the wall
• Decelerates downstream (Increase in Turbulence
  intensity

• The Nusselt number decreases in front of the
  pylon
• It increases between the pylon and the wall
• Decreases downstream
• We see a small rise (1,2m) Turbulence?

Nusselt number along the wall
18
Unsteady simulations
We start from the converged solution for the
composite wall
We stop the warming of the house! Let us see what
happens
Wall with slip (heat flux0)
Video of the transversal temperature X1m
19
Wind blowing at 10 m/s at 253 K
Wind blowing at 20 m/s at 253 K
The north temperature is lower when the wind is
blowing faster
20
Attention Starting from the initial conditions
21
Conclusions

• Fluent computes good solutions for the flat plate
  but wa have to be careful about an eventual
  transition zone.
• The composite wall that we used insulate quite
  well the house
• There are problems with the boundary conditions
  for the room Which one to use to model the case
  that we really want
• We saw different variables influencing the
  Nusselt number/heat transfer
• It is possible to run the simulation unsteady
  with Fluent

Improvements

• This work can be improved by a 3D study of the
  domain to see the influence of the Buoyancy effet
  on the temperature profile in the room
• Model the whole house so the boundary conditions
  of the domain become the walls