# Boundary condition confusion. Engine exhaust manifold ventilation.

 Register Blogs Members List Search Today's Posts Mark Forums Read

July 8, 2016, 19:34
Boundary condition confusion. Engine exhaust manifold ventilation.
#1
New Member

Kramer
Join Date: Aug 2013
Posts: 11
Rep Power: 12
Hello all,

First off apologies ahead of time for the large wall of text!

I am hoping someone will be able to give me some guidance on my problem. Currently I have spend weeks and many hours of trial+error, searching the forums, researching the openFOAM user handbook, etc. but alas I have not confidently solved these very important CFD simulations.

A little background:
I am using OpenFOAM 2.4.0 with Helyx-OS 2.3.1 GUI on Linux Ubuntu 16.04.

The Case Model:
A 2D representation of the model may be seen in the "Boundary_Conditions.jpg" attachment. The case involves a manifold with a constant temperature, an inlet, an outlet, and the surrounding engine cover. Both inlet and outlet can be approximated as being atmospheric pressure.
Simulation 1:
This is the easier of the two simulations. The purpose is to find the outlet velocity and temperature by natural convection of the manifold and inlet temperature alone.

Patches:
Engine Manifold Cover:Patch="wall (no slip)", Thermal = "zero gradient"
Outlet: Patch= Patch, Velocity= "Inlet Outlet Velocity" inlet value = [0,0,0], Pressure= "Fixed value" 0 m^2/s^2, Thermal= "zero gradient"
Inlet: Patch="Patch", Velocity="Pressure Inlet Velocity", Pressure= "zero gradient", Thermal = "Fixed Temperature"
Manifold: Patch = "wall (no slip)", Thermal = "Fixed Temperature"

Simulation 2:
The purpose of this simulation is to find the internal temperature of the engine cover and outlet temperature when a known velocity is applied to the outlet. I understand typically a velocity is applied to the inlet, but for this case a fan will be applied at the outlet to create a vacuum, thus I cannot confidently estimate the inlet velocity (besides possibly approximating using the energy equation/bernoulli's equation but I would rather apply an outlet velocity).

Patches:
Engine Manifold Cover:Patch="wall (no slip)", Thermal = "zero gradient"
Outlet: Patch= "Patch", Velocity= "Fixed Value", Pressure= fixed value 0 m^2/s^2, Thermal= zero gradient
Inlet: Patch=Patch, Velocity="Pressure Inlet Velocity", Pressure=zero gradient, Thermal = "Fixed Temperature"
Manifold: Patch = "wall (no slip)", Thermal = Fixed Temperature

Model
The model for both simulations are buoyantboussinesqsimplefoam as I have the energy equation and gravity tuned on. The simulation is also at steady state.

Turbulence
Typically I run these cases on Laminar, but have had more success using k-W SST and setting at turbulence to zero gradient (as I have no idea how to work with/apply/approximate turbulence intensity and mixing)

Mesh
From what I can gather on the forums and research, the mesh is very important to a good simulation. Typically I try to keep the mesh as coarse as possible without compromising the integrity (e.g. holes in the mesh, unrealistic shapes etc.) due to other posts I have read about regarding compounding error with a finer mesh and difficulties converging the solution. I do run the check-mesh command to ensure that there are no other issues such as overly skewed cells. I have spent quite a bit of time trying to learn different functions of SnappyHexMesh that is included with Helyx-os.

Residuals:
Honestly, most of the residuals have been pretty garbage. The density will not regularly converge below 1e-2. I have many examples if you would like to see any. I will adjust the relaxation factors in order to try to converge the residuals quicker, but I have learned that it will not help long term with converging solutions.

Convergence:
As pointed out by the residuals, typically the CFD will not converge. Not only through residuals but by monitoring the variables (variable changes each time-step).

What else I have tried:
There has been a lot of trial and error with this problem. I have tried adjusting the boundary conditions to zero gradient, pressure inlet outlet, pressure directed inlet outlet, inlet outlet velocity, freestream, etc.

I have tried looking up what each boundary condition does, but the documentation is limited or vague so I am not exactly sure which boundary condition is suitable for which scenario.

Anyways, thank you for reading and any assistance would be greatly appreciated. I am more than willing to give any information that I may have missed or help explain my simulation in greater detail. It has been frustrating, but a great learning experience.

Cheers,
UPengineer
Attached Images
 Boundary_Conditions.JPG (23.2 KB, 29 views)

 July 12, 2016, 17:53 #2 New Member   Kramer Join Date: Aug 2013 Posts: 11 Rep Power: 12 Bump, please.

 Tags boundary conditions, buoyantboussinesqsimple, helyx-os, openfoam, thermal