Seepage analysis

[g8star]

Is it possible to conduct a steady-state seepage analysis in STKO? QuadUP elements are capable of analyzing excess pore pressure diffusion during shaking. Could they be used to calculate the steady state piezometric head within a slope or embankment? This is often a necessary initial step in the dynamic analysis of slopes (for example https://opensees.berkeley.edu/wiki/inde ... s_of_Slope).

I've developed a preliminary model with fixed phreatic pressure on either ends of the model. The model converges and appears to output a reasonable head at depth, but the pressure becomes negative near the top of the slope. This would cause cavitation, air-entry, etc.

Is there a way to implement a pressure cut-off? Or iteratively fix the pressure to zero when air entry becomes likely? I've attached a screenshot of the results for reference.

head.png (270.9 KiB) Viewed 6020 times

[STKO Team]

OpenSees and STKO can be very flexible due to the possibility of using custom scripts.
To speed it up, could you please make a bullet list of all the operations you would like to perform? I will then explain to you how to do them

[g8star]

Sure. I mainly want to calculate the piezometric head, but I don't know the groundwater elevation a priori. For subsequent slope stability analyses, I just want to fix nodes above the groundwater table with negative pressure to zero. I think the easiest way would be:

• Fix the pressure (head) where the groundwater surface is known at wells and ponded water locations

• Run the analysis using elastic material properties with UP elements (this is where I am now)

• Fix the pressure (head) at nodes with pressure < 0. (this what I don't know how to do)

• Rerun the analysis with the updated conditions

Thanks

[STKO Team]

Ok I see.
Well it can be done with an iterative approach using some customCommand (Tcl scripts).
If you share your file set up to this points:

Fix the pressure (head) where the groundwater surface is known at wells and ponded water locations
Run the analysis using elastic material properties with UP elements (this is where I am now)

I will show you how to do the remaining 2 points

[g8star]

Sounds great. The STKO file is zipped and attached.

Section A - 2H1V seepage.zip

(180.13 KiB) Downloaded 324 times

[STKO Team]

Here it is:

Section A - 2H1V seepage.zip

(171.35 KiB) Downloaded 333 times

Have a look at the 2 custom commands I used before and after your gravity analysis and let me know if it's clear how it works or if you need some further explanation.

Notes:

• I changed the duration of your pseudo-static analysis from 1 sec. to 1.0e6 sec. With 1 sec you were having very high accelerations (reaching 1 g) due to the fact that the self-weight in the element property is applied instantaneously. To make sure to avoid inertia effects in transient analyses used for pseudo-static simulations, use large time steps
• The scripts I gave you run multiple times the gravity analysis you defined, resetting the results to 0 if some nodes have negative pressure. At the end of the iteration loop, it keeps the last converged static analysis, so there is no need to re-run it.
• The script I gave you works in sequential analyses. If you plan to use them in parallel (OpenSeesMP) you need to do some minor changes, making sure the convergence is achieved in each processor. If and when you will need it, we will think about it.
• In this specific example, it converges in 2 iterations, so in the MPCO recorder, you will see 2 model stages. If you check the pressure in the first one, you will have negative pressures, while in the second one the min pressure will be 0

[g8star]

You guys are amazing! This looks great. I'm really impressed by the STKO teams responsiveness and support for challenging modeling problems. Thanks so much!

Here is a screenshot of the results, showing a minimum of zero pore pressure above the phreatic surface.

head update.JPG (118.79 KiB) Viewed 5974 times

[STKO Team]

You're welcome!

If and when you will need this script for parallel analysis let us know because it will need a few minor changes to deal with communicating the convergence of the seepage analysis through the different partitions in the different processors.

Then, when you are done with testing it, and if you think this is a useful feature in geotech analysis, we can think of including it as a built-in analysis step in STKO, so that you won't need a custom script anymore for this.

[yuweihwang]

Dear STKO team,

This is a very useful function for the geotechnical group.
I think it will be great if we can make this become a built-in command.
In my case, we need this script for parallel analysis. Can you show us how to do this for parallel analysis?


Best,
Yu-Wei Hwang

[STKO Team]

I think it will be great if we can make this become a built-in command.
In my case, we need this script for parallel analysis. Can you show us how to do this for parallel analysis?

Sure, we put it on our to-do list, and hopefully it will be available in the next version.