seep/W sand boiling

[marioooo16]

Hi all,

I have a question regarding analysis of so called 'sand boiling' under the embankment. On existing embankment there is an issue where water appears on the bottom of the slope due to internal erosion of sand in upper layers, below is gravel. Any ideas how to model this in seep (geostudio)? which BC to use to simulate this. I've tried to do something and I can obtain hydraulic gradient which can be considered as cause of sand boling, but when I model an interface which should simulate cutoff, these gradients are even larger. Any ideas?

Thanks!

 

[GeoPaveTraffic]

I think you are confusing different conditions.

Sand boils occur when the water is confined by a surficial layer causing increased pressure below the surficial layer. When the water finds a weak spot in the surficial layer and escapes, a boil can form.

You describe a wet area on the slope and indicate that it is due to flow through the embankment. This is a seepage issue, not a sand boil issue.

I recommend that you get with someone who has more experience with type of analysis so they can help you understand the differences.

good luck.

Mike Lambert

 

[oldestguy]

How's about using the hydraulic conductivity values for each soil layer and draw your own flow net. That will then develop the hydraulic gradient at the outlet. It is not that difficult. Probably takes less time than setting up all the details on a program.

 

[marioooo16]

Water appears on surface, next to the toe of embankment (some 2 - 5m), it is not on the slope. So I thnik it is not seepage problem, but sand boiling. The subsoil strata is gravel and above (some 1 m thick) is sand with some silt. In my opinion, the main issue is internal erosion of sand due to high gradients as consequence of water flow through gravel...

 

[oldestguy]

Sand does not "boil" by itself. That location is sufficiently close to embankment to be part of the flow coming from the reservoir. The flow net analysis of the area will answer the question. Don't try to model it with a computer program, because your situation sounds more complicated than a simple soil situation. If need be get someone else to do the flow net if that is strange for you. Water pressure and gradients in a lower gravel layer is likely part of the overall groundwater flow system, influenced by the reservoir. In the unlikely possibility of artesian water, you should check area for artesian wells. Being that close to an embankment makes the possibility of artesian pressures much less likely.

 

[oldestguy]

Take a look at this link and watch the video. Treatment of sand boils involves lengthening the flow path, generally by building a circular dike around it. This video ia about the Mississippi levee system but the principle applies many places of impounded water.

http://deltafarmpress.com/government/corps-engineers-levee-boards-watch-sand-boils-levee-breaks

 

[oldestguy]

Here are some references. I see some flow nets for simple situations, but not for yours. However, the references deal with a situation like yours figuring only the flow involving that porous lower zone and the mechanics of flow mathematically. They deserve some study..

The last reference deals with a way to correct the sand boil problem by wells, other than using the sand bag surround technique.

http://etd.lsu.edu/docs/available/etd-0403103-111134/unrestricted/Ozkan_dis.pdf

http://web.stanford.edu/~meehan/flood/mech.html

http://ussdams.com/proceedings/2012Proc/659.pdf

http://www.dailyjournal.net/view/story/86549eb022dd493d9813974cee901b5b/MS--Delta-Relief-Wells

 

[oldestguy]

I took a look at the WEEP W instructions. I'd bet you have not defined the limits of the problem sufficiently wide and deep.
To help those who use the program, why don't you copy the cross section of your program set up diagram and post it for them to comment on.

My cursory viewing of the WEEP W instructions would lead me to believe it would be simpler for you to forget it and go the old fashoned way and draw that flow net.

You will note that the referenced documents have checked their results against WEEP W, yet they go about a more straight forward approach.

 

[marioooo16]

Sorry for not replying before, I was not at home... Here is setup..

 

[marioooo16]

conturs shown are y gradients....

 

[GeoPaveTraffic]

What is the purpose of this embankment?

I can't tell enough about the gradients from the posted images to offer any advice on how the gradients are changing from before and after. I would caution you against using the gradients calculated by the software and instead manually calculate the gradients across the sand layer by hand using the constant head contours.

Based on these images, the easiest way to control the exit gradients and seepage would be to install either a trench drain or a series of relief wells just off the landside toe of the embankment. Either would allow you to collect and transport the seepage a way from the structure and would provide a safe exist for the excess seepage head.

Mike Lambert

 

[fattdad]

thd drawing of a flow net is independent on hydraulic conductivity.

Boiling is the result of exceeding the critical gradient.

f-d

ípapß gordo ainÆt no madre flaca!

 

[oldestguy]

Missing on the diagrams are the lines of equal potential. You need them to see the gradients at any location, specifically at the discharge sand boil area. Hand drawing of the flow net is needed. For some reason that program likes to show what happens with rain on the downstream slopes maybe. That is an insignificant part of the problem solving. Why the line from toe of cut-off going up stream?

 

[marioooo16]

i'll try with flow net, even though it is complex due to two layers of subsoil... When I look at the lines of equal potential, whole potential is decreasing in top (low permeable) layer and therefore vertical gradients are large (up to 2.5)... If I put onlya one layer (high permitivity), vertical layers are OK (up to 0.3). But what concerns me is the 'installation' of barrier which does not reduce gradients but enlarges them.... Situation is completely different if I finish barrier in non-permeable layer where exit vertical gradients are reduced significantly. But the reality is that I have gravel to large depths...

 

[GeoPaveTraffic]

The barrier you are including should have very little impact on the gradients.

I stand by my advice above.

Mike Lambert