Navier-Stokes for engine inlet manifold pressure waves

Numerical methods and mathematical models of Elmer
Post Reply
Dishy
Posts: 2
Joined: 03 Oct 2017, 06:53
Antispam: Yes

Navier-Stokes for engine inlet manifold pressure waves

Post by Dishy »

Hello,
I am trying to model the pressure waves in the inlet manifold of an engine. I have attached a zip of all the files including the case.sif and inlet.stl that I am using to create the mesh.
The basics description of what I am doing is the long tube is the inlet path with the valve as the boundary condition at the bottom and solid walls along the tube. The cuboid at the top is the space inside the air filter with a solid walls top and bottom and constant ambient air pressure (101325 Pa) on the sides for the fresh air coming in through the filter element.

For the engine inlet valve boundary condition I am using
Velocity 3 = Variable Time;Real MATC "if (sin(58*3.14159*tx)>0.44) (-15*sin(58*3.14159*tx)); else 0;"
End
This should give me 29 inlet pulses every second or almost six pulses for the 0.2 second simulation. The reason I am using 15 m/s for the velocity is because the cylinder is 0.375 L and I assume that it fills within 0.013 of a second (I know this is a little rough but it should be ballpark).

I am trying to run the simulation as a Transient with 200 steps of 0.001 sec.

Whenever I try to run the solver I get non-nonsensical velocities of 10^90 m/s.

Any help on how to resolve this would be much appreciated.

Regards

Dishy
Attachments
Navier Inlet.zip
(617.96 KiB) Downloaded 250 times
annier
Posts: 1168
Joined: 27 Aug 2013, 13:51
Antispam: Yes

Re: Navier-Stokes for engine inlet manifold pressure waves

Post by annier »

Dishy,
A. Without looking the detail file, i am telling you from my experience of how the velocity of the solution shoots up.
1. Non-zero boundary conditions for all velocity components at fixed walls or non-zero BCs for one or two components of velocities when the third one is described for boundaries other than fixed walls:
Sometimes when we forget to set up boundary conditions to zero ,

Code: Select all

Boundary Condition n
...
Velocity 1 = 0
Velocity 2 = 0
Velocity 3 = 0
..
End
Sometimes, if we only set Velocity 1 = 1.2345 and do not set velocity 2 and velocity 3; the solutions may deviate from the expected ones.

2. So, setting up complete boundary conditions even at boundaries other than fixed walls, is important for obtaining realistic velocity.


Yours Sincerely,
Anil Kunwar
Anil Kunwar
Faculty of Mechanical Engineering, Silesian University of Technology, Gliwice
annier
Posts: 1168
Joined: 27 Aug 2013, 13:51
Antispam: Yes

Re: Navier-Stokes for engine inlet manifold pressure waves

Post by annier »

Hi Dishy,
B. Going through your file -case.sif; I would point out on pressure boundary conditions:

Code: Select all

Boundary Condition 2
  Target Boundaries(1) = 2 
  Name = "Filter"
  !Pressure 2 = 101325
  !Pressure 1 = 101325
  !Pressure 3 = 101325
  Pressure  = 0.0 !101325 101325 Put 0.0 if you mean atmospheric pressure
  !External Pressure  = -101325 !-101325 0.0
End
Notes:
(i) Pressure 1, Pressure 2 and Pressure 3 are not BCs for scalar pressure (Pa) variable but mean the three components of shear stress (Tau= Force/Area) at the boundary areas.
(ii) Similarly, External pressure means Normal stress (Sigma= force/area) at the boundary areas.
(iii) The word Pressure or Sometimes External Pressure can be used to apply scalar boundary condition for pressure at a boundary.
(iv) Use atmospheric pressure = 0.0 , for the simulation.
(v) Reference link

Yours Sincerely,
Anil Kunwar
Anil Kunwar
Faculty of Mechanical Engineering, Silesian University of Technology, Gliwice
Dishy
Posts: 2
Joined: 03 Oct 2017, 06:53
Antispam: Yes

Re: Navier-Stokes for engine inlet manifold pressure waves

Post by Dishy »

Hello,
Thank you very much for the help. I think I have implemented the changes recommended.

I am able to get the simulation to run now if I use an incompressible fluid. But the results don't look like what I would expect. I can see the pressure drop at the bottom of the inlet when the valve opens and the gas flows, as you would expect. When the valve closes and the flow stops there is a pressure spike as you would expect but that pressure spike doesn't turn in to a wave travelling back up the inlet path. I suspect this is because I am trying to get a pressure wave to travel through an incompressible fluid. But if I change the fluid to ideal gas the solver generates errors and results that NAN.

I have attached a copy of my project again after the changes, any help would be greatly appreciated.

Dishy
Attachments
Navier-inlet.zip
(617.54 KiB) Downloaded 248 times
Post Reply