yade-dev team mailing list archive

[Christian Jakob <1362090@xxxxxxxxxxxxxxxxxx>]

> Try imposing velocities to the boundaries of the undrained sample and see if the 
> results are consistent in terms of pressure changes.

I rewrote the script. It is now taking 8 minutes for preparing and 2
minutes for final calc. The script saves the model after preparation. So
you can set load = True at the beginning of the script to load prepared
sample.

changes:
Triax engine is deactivated after preparation steps and velocity of top boundary is imposed, so that the sample is compressed. In the undrained case I get the following result:

Running script pfv-bug-cell-pressure-too-low.py
set fluid bulk modulus to: 2.000000e+09
Starting calculation, number of steps: 100000
p: -0.0 /poro: 48.147175
p: 77.3254702635 /poro: 47.5619234375
p: 128.265866485 /poro: 46.971240625
p: 141.818936358 /poro: 46.3730234375
p: 190.076801101 /poro: 45.7726046875
p: 256.087922324 /poro: 45.163634375
p: 286.638303292 /poro: 44.5447875
p: 336.686906526 /poro: 43.9271546875
p: 370.445026634 /poro: 43.3041359375
p: -246.733682291 /poro: 42.6770140625

If I uncomment line 145: flow.boundaryUseMaxMin=[0,0,0,0,0,0]
triangulation is not working correctly...

Running script pfv-bug-cell-pressure-too-low.py
set fluid bulk modulus to: 2.000000e+09
Starting calculation, number of steps: 100000
6 : Vh==NULL!! id=6 Point=0.000512399 0.00124487 0.00125213 rad=3.15073e-05
7 : Vh==NULL!! id=7 Point=0.000587947 0.000954619 0.000983527 rad=0.000124813
8 : Vh==NULL!! id=8 Point=0.00146642 0.000434536 0.00102062 rad=7.286e-05
9 : Vh==NULL!! id=9 Point=0.00151061 0.0011991 0.000915917 rad=9.27561e-05
...
...
279 : Vh==NULL!! id=279 Point=0.000526466 0.000269253 0.00137925 rad=5.56929e-05
280 : Vh==NULL!! id=280 Point=0.000373563 0.000806535 0.000942971 rad=8.59516e-05
281 : Vh==NULL!! id=281 Point=0.000918819 0.00104285 0.00064448 rad=9.2951e-05
282 : Vh==NULL!! id=282 Point=0.000928043 0.0006173 0.000905637 rad=0.000119065
AREA <= 0!!
AREA <= 0!!
AREA <= 0!!
AREA <= 0!!
AREA <= 0!!
AREA <= 0!!
AREA <= 0!!
AREA <= 0!!
p: nan /poro: 48.147175
...

It seems, that all cells get zero volume with each triangulation step.

c

** Attachment added: "reproduce bug with imposed velocity (takes around 10 minutes)"
   https://bugs.launchpad.net/yade/+bug/1362090/+attachment/4194076/+files/pfv-bug-cell-pressure-too-low.py

-- 
You received this bug notification because you are a member of Yade
developers, which is subscribed to Yade.
https://bugs.launchpad.net/bugs/1362090

Title:
  problem with pressure calculation in PFV cells

Status in Yet Another Dynamic Engine:
  New

Bug description:
  Hi,

  I want to get the pore pressure of a saturated soil sample in an
  undrained triaxial test when the sample volume is decreasing.

  ...
  TriaxialStressController(thickness=0,stressMask=7,internalCompaction=False,label='triax')
  ...

  	triax.goal1 = 35000
  	triax.goal2 = 35000
  	triax.goal3 = 70000
  	triax.wall_back_activated = False
  ...
  	flow.viscosity = 1.3e-3
  	flow.fluidBulkModulus = 2e9
  	flow.bndCondValue=[0,0,0,0,0,0]
  	flow.bndCondIsPressure=[0,0,0,0,0,0]
  ...

  For one of my simulations a decrease of ca. 1,8 % in porosity (=-dV/V)
  is obtained. When I calculate corresponding pressure increase I get

  dp = K*(-dV/V) = 2e9*0.018 = 36000000 Pa

  But the output from PFV shows maximum average pressures around of 16
  Pa (not MPa!). As long as I use SI units the output of
  flow.averagePressure() should be Pa, right?

  So is there a problem in the code or did I miss something in the
  compressibility option of PFV?

  regards,

  christian

  
  ###additional info:

  I double-checked my script and did not see anything wrong. So I think
  it is on C++ side. So I was going on testing PFV method:

  I also tested a drained situation with incompressible fluid. The porosity is decreasing due to a triggered collapse of the grain structure, but the pore pressure gets negative (see attachment).
  So I think it is a bug in pressure calculation of PFV method.

To manage notifications about this bug go to:
https://bugs.launchpad.net/yade/+bug/1362090/+subscriptions