Flow 2D simulation fails when using air instead of water 回答済み

I used the Flow 2D simulation as a starting point. I changed the flow velocity (30 m/s) and the medium from water to air:

symb vel = 340

matr
  wvsp on
  type elas
  prop air 1.225 340 0
  vdmp air $freqdamp db 0.002 0 1e+06 2 0.01 /* needs changing for air
  hrgl air visc 0.02

However, the simulation appears to crash. I've attached two screenshots from the output mp4 showing the crash. 

Last good frame:

First bad frame:

This cloud of energy floats away in the wind field!

What could be the problem? Should I decrease the step size? Here are the relevant parameters:


c Flow
symbx vflow = 30 /* flow velocity
symb nsc = 20 /* boxes at edge to scale flow

c Frequency and mesh
symb freqint = 0.5e6 /* Center frequency
symb freqdamp = $freqint /* Damping frequency
symb vel = 340 /* Minimum velocity to mesh for
symb wavelen = $vel / $freqint
symb nelem = 15 /* Number of elements per wavelength for mesh (minimum 100, increase for accuracy but at expense of longer computation time)
symb box = $wavelen / $nelem

c Simulation time
symb ncycles = 200 /* Number of cycles
symb nloop = 20 /* Number of times to plot the simulation
symb tfact = 0.6 /* Time stability factor

c Calculations including Meshing
symb simtime = $ncycles / $freqint
symb pipeLen2 = $pipeLen / 2
symb txDist = $pipeRad / cos ( $txAng * 3.14 / 180 ) + 10e-3
symb txTop = $txDist + $txThk
symb ygap = $txTop + 3e-3
symb ybeg = 0 - $pipeRad - $ygap
symb ang = $txAng + 90

 

 

 

1件のコメント

  • Hi Andrew,

    Another user is having a similar issue, I just got some feedback from our VP of engineering and he made these suggestions:

    1. Flow moving too fast for the timestep (36 m/s in air is about 10% of the speed of sound, so the flow would move the ultrasound beyond what the timestep supports)
    2. Fast flow bounding regions with no flow – it would usually be a gradual transition between ‘boundaries’
    3. Flow all the way up to the absorbing boundaries
    4. Poor mesh quality in terms of aspect ratio etc
    5. Just running for too many timesteps

     

    コメントアクション パーマリンク

サインインしてコメントを残してください。

お探しのものを見つけられませんでしたか?

新規投稿