Hi,
I am getting very high excess pore pressures at particular nodes beneath the pile cap, adjacent to the embedded beams as shown below,. This is causing very slow convergence and excess pwp is not changing after particular point. (fig 1 - phase before, fig 2 - interested phase)
The pile cap is a volume element with Linear elastic material (non-porous)
The permeability of clayey soil beneath the pile cap volume is 2.0 E-11 m/s. (all directions)
Around pile cap there is granular fill with K = 1.0 E-06 m/s and clay with K = 2.0 E-11 m/s (all directions)
If I increase permeability of clay by 100 times I can run the model properly, but we are expecting very low permeability as stated.
What could be the reason of such spurious excess pwp at such location?
Thanks for any help in advance.
Dear Omkar,
1) As I already mentioned: reduce the ratio. It really doesn't matter if a material is 100 times less permeable or 10,000 times less permeable, in both cases the material is practically impermeable.The art of geotechnical engineering is not to just input the numbers given and consider it a problem if it doesn't work .... the art is to recognize the problem and the solution and then wonder if the solution is really giving a significantly different answer from engineering point of view or not.
2) I just realized this has become superfluous because since Plaxis V22 the software will always transform the input parameters to SI units and then calculate. So whether you fill in permeabilities in m/day or in mm/sec makes no difference.In the past PLAXIS just calculated with the numbers as given. But what I was trying to say is that when solving the matrix when it contains very small values the rounding off error increases when the numbers are smaller. So then indeed, having a permeability 2e-11 will generate more rounding off error than a permeability of 1.726e6. And more rounding off error means that more iterations are required to solve it. Therefore changing to m/days could help, but only for PLAXIS versions prior to V22...
With kind regards,
Dennis Waterman
Answer Verified By: Omkar Sugade
Thanks Dennis,
I have few queries on your points as below:
1) How to possibly deal with the high permeability ratio numerically, where the mitigation measures (such as highly permeable granular fill) over the very stiff low permeable clays are proposed?
2) In general I use 'seconds' to input the permeabilities and then switch back to 'days' to input consolidation time. Does defining parameters in particular units makes difference to the calculations in Plaxis? I could not understand your second point properly, please can you give some more insights on this?
Kind regards,
Omkar
As you already found out, the problem is in the very low value of the permeability. There can be 2 things going on:
1) it's the ratio between permeabilities that is a problem. Right now you have a ratio of 50,000 between your permeabilities, which is very high. If the ratio is 5000 or even 500 the results with be the same because where the clay has a permeability that is 500 times lower or 50,000 times lower really doesn't make a difference that in terms of the uncertainties in geotech can be justified.
2) the absolute value of 2e-11 is the problem as it's very small. In that case you could consider not using "seconds" as your unit of time, but for instance "days" because then all values will become 86,400 times larger and the permeability of the clay will be 1.726e6 m/day. Usually the problem is the ratio, which is not just for permeabilities. If one would use to soil materials with a stiffness ratio of 50,000 there may also be convergence problems.