Sand Boils Observed During Excavation For Segmental Wall

[MRM]

Here's the scenario; there is a site with a 10 foot +/- tall earth embankment. Although the embankment has been stable for the last 30 years, numerous seeps were observed a few years ago. Soil borings were completed in the area at the time the seeps were observed. The geology in the area is complex, with alternating silty clay and water-bearing sandy strata to depth. It is believed that several discrete aquifers exist within these soils. Some of these aquifers are artesian, while others represent the regional groundwater flow towards the bay about 800 feet away.

Now, to provide room for a new development, the embankment will be replaced with a new reinforced segmental retaining wall.

A contractor came in and installed a sheet pile wall (believed to be about 20 feet in length), along with 2 inch PVC well points (believed to be about 20 feet long, screened on the bottom 5 feet) on the side of the sheet wall opposite of the proposed excavation for the new segmental wall. The location of the sheet wall was near the crest of the existing embankment at the top. The wells were installed by jetting in under high water pressure on approximate 3 to 4 feet centers. Prior to the pump being turned on, some artesian flow was observed in about 50% of the wells. This was not surprising based on the original soil boring information.

After the pump had been started and was running for a few days, the contractor excavated the area adjacent to the sheet pile wall down approximately 8 to 10 feet. They noticed that even with the pumps running, there were sand boils at some areas near the wall, and the area of the proposed retaining wall was generally wet.

We were called out to the site to help them understand why their dewatering system was not working. First of all, it was obvious to me that the well points were not doing the job for which they were intended. I believe the well points are probably pulling water from deeper aquifers that probably do not need to be dewatered in order to dry the site up. There are more shallow aquifers making it through and up and around the sheet wall that are not affected by the well points.

My opinion about the sand boils is this; while they may be result of natural discontinuities in the soil strata that are now exposed to air for the first time, I believe that the pressure of the jets may have aggravated the problem by weakening certain strata and ultimately causing more sand boils to form during excavation in that area than would have existed otherwise. What do you think of that? In your experience, have you seen the act of jetting well points in cause damage in a highly stratified soil system? This is an important question to answer since long term dewatering and water management of the segmental wall will be dependent in part on how much water is rising upwards from depth after construction.

My concern is that if the contractor installs additional well points (at varying tip depths) in an attempt to dry the site up for construction, additional damage to the soil strata may occur and more upward flows will be the end result. My thought at this point is that the area should be dewatered using a more gentle approach; namely shallow sump pits along that area to avoid breaching these deeper artesian layers.

What do you think about the dewatering and sand boil theory? Thanks for your comments about this!

P.S. There are other concerns that we have been addressing on the fly for this site too; namely a cantilever sheet pile wall that was installed by the contractor with no design or analysis...

 

[GeoPaveTraffic]

I don't have any experience that matches exactly with your situation, however, your conclusions appear to be correct. I would also not want more well points jetted in, however, you/the contractor could drill some more points in.

I'm not sure I completly understand what is being constructed. It seams that the land side of the existing embankment/dike is being replaced by a segmental wall and that the dike has had seepage problems in the past. I would question the wisdom of this design. Has anyone looked at the new seepage path through the dike since the path will now be shorter? If long term seepage will be a problem, it may make since to install a permanent cut off wall along the crest of the dike. Given the soil conditions the wall would likely need to be a slurry wall with a depth dependent on the specific geometry and soils.

 

[BigH]

Hush! But I had a site once that had very fine silt to silt some clay at the surface. The site was dry - in a small creek valley. We put a boring down for a bridge investigation - punched into a very local artesian deposit and, to make it short, we tried and tried to seal off the borehole - with no success. I figured there was always some seepage but that it was slow enough through the silty surficial layer that evaporation, etc. keep the site dry. You might have something like this. Punching through different strata and hence connecting them.
You may need to install a french drain along the base of the wall - and direct the artesian flow away from the site - hopefully you have some ground contour. I had something like this happen too in Vancouver. No rain for months and the french drain was flowing like a river.
And, you would be quite right to know who designed what and who authorized the method of support (sheet-pile walls) etc. Was a specialist dewatering contractor used? . . . or just a makeshift attempt by a general contractor?

 

[DRC1]

If I understand the situation correctly, you have a highly statified system with aquifer layers and aqualudes. To perform an excavation, the contractor has driven 20 ft steel sheet piles to form a wall for a 10 ft excavation. On the retained side they have installed 20 ft deep well points. On the excavated side they have encountered sandboils durring excavation.

My guess is that the sheet piling has cut off the flow of at least one of the aquifers. With an impermiable layer above and below and an impermiable wall downstream, head pressure at the wall will rise, causing the boils as excavation relies overburden pressure.
Although good in sand, vaccum well points don't function well in low permiability soils. They also will have little effect through the sheetpile, so if boils are occuring in the inside of the wall, ponts should be installed on the inside to the depth of the layer that has the boils. I suspect you are correct that the wellpoints go too deep and do not drain the layer of concern. Generally jetting produces only localized effects which quickly dissapate. I would not worry about a jetting operation, as long as it performed with proper care.
It seems that the sheeting and dewatering systems were installed with little or no engineering. This should be remidied.
Note that OSHA requires a PE stamp the sheeting drawings for the excavation, and if you have sand boils the soil below the dredge line may be liquifing, which will dangerously reduce the capacity of the sheeting. This should be examined before anyone renters the excavation.

 

[MRM]

Thanks for your time in responding. You're comments are very useful and appreciated, as always. You're right in inferring that this is a situation where the contractor sort of went off on their own in an attempt to "get-er-done!" No design pertaining to the sheet wall or the dewatering system was undertaken.

Based on a soil exploration completed about 5 years ago, we knew that there would be a lot of water to contend with after construction. Our site plan has a fairly robust drainage system for the new wall as well as an underdrain system for the new building to be constructed. Although we anticipated a large amount of water approaching the wall horizontally (regional groundwater, etc.) and vertically (via artesian pressure, communication between horizontal strata, etc.), if the act of installing well points using high pressure damages those confining layers, we may have more water than ever expected. That's my theory, at least.

GeoPaveTraffic, you seemed to agree that the idea of jetting wells in and damaging the natural stratification allowing for greater communication between those strata is plausible. DRC, you seemed to think that it would not be cause for any unreasonable concern. This is good as it gives me both sides of the story. I would be interested to hear further why each tends to lean one way or the other on the well point jetting issue.

We're also in the process of looking further at the stability of the wall itself. One thing the contractor did do before calling us, is stop excavation near the wall when the boils were discovered!

 

[DRC1]

I have used well points in artsian sand aquifers overlyn by clay, although not with multiple layers. The jetted wellpoint uses a very small amount of water, that after driving may exhibit exess pore pressures, however,they dissipate quickly. Any exess hydrostatic pressure either flows out the pipe or along the pipe. Well points use high pressure but not neccesarily high volume of jet water. The amount of water added to the confined aquifer is not huge relative to the volume of water in the aqufer. that is why I would expect only temporary localizied problems to occur.

 

[BigH]

?? - did/has the boiling decreased with time? Sometimes acquifers are in fact large pockets that eventually drain out if not connected with other layers. Just another thought. Still think that a "deep" french drain may be needed in the long run. Maybe not based on original condition but certainly after the contractor did his thing!

 

[MRM]

Good point...another engineer from the office has been on the site since then and that would be a good question to ask him. However, when I observed the boil, it had been at least a half day or so since the contractor first noticed it. At that time, they mentioned that it hadn't appeared to have lost any "steam."

 

[BigH]

I had some that took several days to a week once - but once drained, no more water.

 

[MRM]

That would certainly be the best scenario for long term drainage!

BigH, when you say "french drain," a near-surface drainage structure made of stone and geotextile in some configuration comes to mind. When you say "deep french drain," I think of what we may have already designed. We've got a system that includes clean sand backfill behind the wall with a horizontal drain near the base. This drain is near bottom of wall level and contains angular, uniformly graded stone. This horizontal drainage layer is intended to intercept upward groundwater flow (the wild card, if you will), and to help lower groundwater approaching the wall from higher elevations within the backfill area of the wall. Then there is a drainage layer immediately behind the wall, as many walls of this type have, to help keep water away from the immediate back-of-wall area. The water is collected in a perforated 6 inch line near the base of the footing where it travels to daylight, away from the wall.

Is this what you have in mind?

 

[BigH]

I suppose - what I meant by "deep" is that it might be necessary to install the french drain to a greater depth than, say 1m or so below the ground. The one we used in the Vancouver Skytrain project went as deep as 4m. It was a open graded stone (usually, now, wrapped in geotextile) with a 4 to 6 inch perforated pipe near the bottom. You seem to put this well forward of your sheet pile with the finger drains - good idea (actually an old tailings dam concept).

 

[MRM]

Thanks again for your comments, everyone. Very helpful, indeed.