Pressure boundary condtion numerical error ?

Questions and remarks about code_saturne usage
Forum rules
Please read the forum usage recommendations before posting.
Post Reply
sbardwell

Pressure boundary condtion numerical error ?

Post by sbardwell »

Hello,

I am currently modeling the flow in hydrodynamic seals with singularities. The boundaries of the model must be pressure based to model the different working ranges of the seal. My initial model is for a pressure difference of 2 bars. The study becomes a steady state after the flow is established.

My first study was for a certain distance of 1.5mm between the two wall boundaries (top & bottom) and i had no apparent numerical problem linked to the pressure boundary conditions. However my next study consists in reducing this distance to simulate a dysfunction of the seal, but now the pressure boundaries are no longer are stable and pressure waves move from the outlet to the inlet (inlet = left; outlet =right) creating an oscillation in the seal. I looked into some books only to find that this is linked to the pressure boundaries and the correction on the flow to keep mass continuity
at the moment this is how i impose my pressure on the boundaries

Code: Select all

call getfbr('inlet', nlelt, lstelt)

do ilelt = 1, nlelt

  ifac = lstelt(ilelt)
 
  itypfb(ifac) = ientre 
  icodcl(ifac,ipr) = 1 
  rcodcl(ifac,ipr,1) = 400000.d0

  icodcl(ifac,iu) = 3 
  icodcl(ifac,iv) = 3
  icodcl(ifac,iw) = 3

  rcodcl(ifac,iu,3) = 0 
  rcodcl(ifac,iv,3) = 0
  rcodcl(ifac,iw,3) = 0

enddo

call getfbr('outlet', nlelt, lstelt)

do ilelt = 1, nlelt
 
  ifac = lstelt(ilelt)

  itypfb(ifac) = isolib
  icodcl(ifac,ipr) = 1 
  rcodcl(ifac,ipr,1) = 200000.d0
enddo
I have tried many suggestions, for example:

- Creating long entrance and exit zones, but this generates a false pressure distribution in the seal because the exit zone near the seal is no longer at the boundary pressure corresponding to the operating pressures.
- Secondly i tried imposing pressure losses near the exit boundary but its hard to estimate the correct head loss value and also this will only work for one configuration of wall distances. The futur aim is to create an ALE model with moving walls.
- Lastly i have also tried Long simulations to monitor an eventual stability of the pressure and velocity in time and also finer meshes. But nothing changes.

An example model with a similar configuration is attached below.

Thank you for any comments
Capture.png
(9.8 KiB) Not downloaded yet
Post Reply