Piled Raft Foundation

[Dirt Guy]

I am working on a project where the geotechnical design engineer has implemented a piled raft foundation for a bridge. The soil conditions in the area consist of relatively weak saturated glacial lacustrine clays (CH). In order to construct the bridge at the required elevation, an excavation on the order of 40 feet is required. Upon completion of the excavation (removal of overburden soils), the designer has recommended placement of a 5-ft thick gravel layer followed by installation of H-piles through the gravel layer into dense glacial tills approximately 30 feet below the soft clays. The H-piles will be spaced 10 to 20 ft O.C. beneath the approximately 75'x300' concrete mat footprint. The cutoff elevation of the piles will be at approximately the mid-point of the gravel layer. The design engineer has incorporated two geogrids layers within the gravel thickness (2.5 feet) above the tops of the piles for support and distribution of the structural loading. Once this is complete, the engineer will pour a reinforced concrete mat with a thickness of 4 feet to form the base of the structure (for support of bridge pier walls/abutments, etc.). The native soils are prone to rebound upon unloading. In this case, when the overburden soils are removed, the stress decrease is estimated to be around 2,600 psf across the footprint of the structure. When the structure is finally constructed, it is supposed the structure will reapply a load of approximately 1,500 psf across the footprint of the mat. My question is this. If you are unloading the site and only partially reapplying the load and all settlement analyses show a net rebound at the site over the life of the structure, what purpose are the H-piles providing if they are not attached to anything (endbear in glacial till and not attached at the top to the concrete structure)? In addition, the H-piles do not have plates welded on top. The engineer claims the piles will reduce the total rebound of the native clays from 3 inches (without piles) to less than 1 inch with piles. Can someone reasonably explain the process of how the piles will benefit the structure?

 

[TigerGuy]

This sounds almost like an inverted approach to stone columns. The H-piles should reduce the rebound (rigid inclusion), as long as the uplift (side friction) from the rebound does not pull them up from the gravel layer. If I was an artist, I would provide a diagram, but I'm not, so I don't.

What does the geotech say when you ask these questions?

 

[Dirt Guy]

The geotech has a difficult time explaining it. They pretty much say this is a modified load transfer platform and that when the structure is constructed, the gravel will arch to the tops of the H-piles, but I have a hard time understanding how the load from the structure gets distributed directly to the piles without plates on top and with the distance between piles being rather significant (10 to 20 feet). Then in another instance they explain that since the top 2.5 feet of the pile is in gravel, the skin friction along the piles from the gravel and load of the structure is enough to reduce rebound of the underlying native clays.

If I look at this from a rebound aspect, a majority of the rebound occurs during excavation prior to gravel and structure construction. This rebound in the clay would be hard to measure since it will occur as the excavation occurs and some of the rebounded material will be removed during excavation. I would think the engineer would have to have a strong handle on the soil's performance to recommend such a system. If your tolerance in movement is 2 inches, what happens if most of the rebound occurs faster than predicted using assumed parameters, say during the excavation stage, then you apply your gravel layer, install H-piles, and pour concrete, and build your structure and no more significant rebound occurs. Potentially, you have installed a million dollars worth of steel for no definitive purpose.

 

[DanielSgeo]

The raft is 40 ft deep and 4 ft thick, I assume the 36 ft above the top of the raft gets backfilled? If so, there would be a net increasing in bearing at the bottom of the raft.

 

[Dirt Guy]

There is no backfill placed above the structure. Just the structure itself. There will be a significant net stress decrease at the bottom of the raft.

 

[DanielSgeo]

Huh, that’s interesting. I couldn’t explain what’s going on based on the info provided.

I have designed a few load transfer platforms and pile supported embankments. 10 to 20 ft spacing for something as small as an H pile is large for the rigid inclusions spacing, based on my experience.

 

[Dirt Guy]

It was designed presuming the weight of the structure would be transferred to the piles through arching in the gravel in combination with the geogrid in the gravel; however, the H-piles are not equipped with plates on top. The pile spacing seemed to be to big for this arching to occur and reasonably apply the load to just the pikes. If the site didn't involve a net stress decrease and rebound of the native clays, I would assume some of the load would be transferred to the soils between the piles at that length of spacing as well.

 

[MTNClimber]

I agree with the concerns about the 20-foot spacing. I would ask the engineer for calculations and project experience with using this method for bridge abutments. We've only used this style of load transfer platform for tank foundations, but we were not concerned with rebound.

 

[EireChch]

The H pile dont need plates on them to take up the load. The gravel between the flanges will likely act like a plug and provide a surface area. Like when we design H piles for end bearing. We design for L*B not the steel surface area of the H profile.

I agree that the pile spacing is too much. I would have thought 10ft (3m) was average, 4 being maximum IMO.

I agree also that the rebound happens during excavation. Most of it will have occured before the piles are in place. I cant see them really doing much if its purely for rebound control. There is no issue of bearing capacity failure since its a net decrease in load.