Crack in Pile Foundation

[brdgbldr]

I just received a report from our contractor about a pile lateral load test. The report states that during testing, one of the piles made a loud pop at about 30k of load. The soils engineer states that he believes this is from some of the concrete cracking. The pile is to be part of a larger group of 20 piles. The design load for the pile system (for a ped bridge) shows that if all piles are loaded evenly in this system, that the max resultant lateral load to go to the pile will be close to 50k. Our contractor is having problems with this soils company and the designers of the foundation and we are trying to check their work.

My question is (as this is my first foundation design) will this cracking lead to future problems with rebar and pile degradation? I am sure we can expect more cracking in other piles as the design load is 50% higher than the tested load. The water level fluctuates from roughly 5 to 10 ft below grade. I am worried about the pile capacities being compromised. We have already recommended redesign, but the design firm is resistant, even though previous reviews show some major concerns with their design. Any suggestions? Thx

 

[Sean2]

brdgbldr,

I often feel the need for further information when I read about an interesting problem, and this case is no exception.

Your piles should ALL be tested for integrity. For example, the sonic echo test can be used and is reliable to depths which depend on the aspect ratio (ratio of length to diameter) of your piles. This should have been done before any load testing. Such a test would show up any discontinuities (voids or cracks) or sudden changes in cross-section in the pile.

In any event, the pile which seems to have cracked must be tested now.

You would expect that the pile's lateral displacement would have been measured during the lateral load test. If the pile DID crack, the curve of load versus displacement should change significantly at that point.

The question of whether the pile has failed the prescribed load test should not be a question of opinion: it can be established as a FACT. If there is any doubt, test the same pile again, making sure that the direction of loading is the same as before.

I strongly recommend that you and your employer obtain the advice of a specialist to do the following, at least:

1) Review the design calculations for the piles, including the DESIGN BRIEF and the design criteria.

2) Review the pile installation method, the test method, and the results of the tests aw well as any conclusions drawn from the tests.

3) Advise on how to proceed.

If the design firm has not employed a competent specialist, then you need one. I suggest that you ask the design firm to explain why they are resisting a re-design. If they can't put it to you in clear terms which are easy to understand, they probably don't understand the problem themselves.

Whether you engage a specialist of your own or not, the project will be best served if both companies (your employer and the design company) try to resolve the problem TOGETHER, as you both (should) have some common goals, as well as probably some goals which differ.

By the way, a single pile group of 20 piles for a pedestrian bridge sounds like an unusually large number. Also, you haven't mentioned the units of force you are using; I don't know what you mean by "k".

Good luck with this, and learn from it!

 

[Qshake]

The k is obviously for kips (1 kip = 1000 pounds).

I don't understand you part in this. You state that this is your "first foundation design" but the "design firm" is resistent to redesign. This implies that you didn't design the foundation and your not part of the design firm that did design the footing. So are you the owner or an inspector acting on behalf of the owner? You might even be retained by the owner as a consultant to this specific problem. The reason I ask is due to the different approaches that should be taken.

With many clients the contractor shoulders the burden of proving this foundation is compromised will lie with the contractor. In the event that the foundation is compromised it will be the contractor who will ultimately propose what is to be done to correct this situation.

Frankly, I am amazed that the contractor hasn't already agreed to drive an additional pile(s) near the location of the allegedly damage pile. Since the equipment is there on site and since testing is expensive and the contractor risks finding out what is already presumably known, most would just drive another pile and be done with it.

In reference to that comment about the twenty piles for a pedestrian bridge - many reasons for that including poor soil conditions (in fact water was already mentioned) and the lack of hardpan at some reasonable depth. So it sounds like this is friction pile.

 

[DRC1]

First, If there is a problem with the design, the contractor is not "just going to drive another pile" He will wait to recieve explicit instructions on how to proceede. If the piles are deficient, driving an additional pile will not help. By the way, What type and size are the piles. If the design firm is reluctant to redesign, I would ask the owner to hire a third party to review the design and make recomendations. I would consider retesting the pile to see if the is any loss in capacity. Latreal load capacity is difficult to compte accurately. If the piles can not devlop sufficent latral capacity, it may be necessary to drive batter piles to resist lateral loads.

 

[Qshake]

DCR1,

I'm didn't mean to imply that the contractor would just drive the extra pile, but rather would have made that suggestion and waited for some concurrence.

As I stated, in most cases the means of correction is up to the contractor with the review and approval (usually for meeting general specifications) up to the owner. And since the burden falls on the contractor to make a proposal, he is not going to wait for "explicit instructions on how to proceed". Time is money to contractors, and with crew and equipment sitting around doing nothing until some direction is found doesn't sit well with most contractors.

 

[brdgbldr]

Maybe I wasnt clear in our part of this project. The footing design was done by a local firm and the superstructure design and all of the construction is being performed by a branch of my company. The ped bridge consists of an arched structure in a high seismic zone and is very intolerant to lateral displacement of the footings. When the superstructure was being designed, the footing designer was resistant in releasing results such as displacement and the process of their design, which sent up red flags to our contractor and superstructure designer. We were asked to check the footing design to ensure it would be sufficient (we are the third party reviewer in other words). We found the design was questionably sufficient (we could get it to work in some cases, but not in others, among other problems we found) but the owner is hesitant to change designers and the current footing designers are sticking to their design.

Chances are, we will replace the pile. Our fear is that since the design loads are so much higher than the tested loads, there is still a chance for the design to fail. Like I said, the superstructure is VERY intolerable to footing movement. We have suggested a redesign, but the footing design firm is willing to go all the way with their design, and the owner seems ok with it. At this point, we are trying to place as much liability away from our company as possible.

I guess my original question was, I know that LPile uses P-Y curves in its tests, which I know are conservative. So far, the lateral test showed deflection much less than what LPile predicted. However, the load tested was at a much lower magnitude than the design loads and one pile experienced cracking. When reloaded, this pile's deflection only increases about 1% so the crack is not severe structurally. However, will these cracks pose problems with water getting to the rebar and having degradation in the long term? I'm not sure if we should expect long term problems seeing as one pile cracked under the tested load. Thanks for all your help.

 

[SoilRocks]

This is a very involved situation, and someone needs to throw up the red flag and call time out. The first obligation on everyone's part is to build a safe bridge - so, don't let a design engineer who won't discuss, and can't explain, his design just ignore the serious concerns you have. Unfortunately, about 50 percent or more of geotechnical firms out there shouldn't be working on this type of project.

Hat's off to Sean2 for an excellent response above.

My feeling is you shouldn't be talking about re-design (yet). It's all about testing the piles. That IS your re-design.

1. Usually, deformation controls your lateral test, or at least you test it to 200% of design. A structural failure of the pile, which it sounds like you had, shouldn't occur at 150% of design. It sounds like you either exceeded the moment capacity of the pile, or had a pile defect. Further testing on other piles, which is needed, should tell you that.

2. Lpile is not "conservative," only the judgment of the user. Conservative is not always a dirty word, either. That's why we have testing programs.

3. A lot of important information is also not noted, such as, is the design lateral load seismic or static? I assume seismic, since you mentioned you are in a high seismic area. Then, it seems you should have less concerns about corrosion due to cracking. The group does seem very big for a pedestrian bridge - must be a biggy or a long span. Other information such as the pile size and maximum allowable deflection would be interesting to know.

At this point, I would consider refining the testing program based on the information to date. You know, I would be concerned about the vertical capacities too, based on your dealings to date with the geotechnical design firm.

 

[Focht3]

Sean2 and SoilRocks have given you excellent advice. Without a lot of additional information, there isn't much more guidance that I can offer to you. I can comment - generally - on the design and testing of laterally loaded piles and piers. I received my Master's from the University of Texas - Lymon Reese was my second reader. (Lyman's firm supports and markets LPILE1.) My master's report addressed the group effects on laterally loaded pier behavior. And I've been around the drilled shaft business all of my professional career. (San Antonio, my home, was the birthplace of equipment to mechanically construct drilled piers.) You're dealing with my home turf, so to speak.

LPILE1 is a tool, nothing more. A very useful tool, but still a tool. Garbage in, garbage out. The results from the program are not gospel; they're computer output subject to the assumptions of the person who chose the input parameters.

First things first: is the 50k load a typical working load, or the maximum design load? I assume this represents earthquake loading (i.e. maximum design load.) Do not load your production piers beyond their design load unless required by code. Lateral behavior is not a "factor of safety" design; it's a performance design. How well do the designer's predictions of load-deflection match the load test?

Several things could have gone wrong with either the design or construction of the pier in question. Did the designer use the gross uncracked section to calculate EI? What was used for the concrete tensile stress to select a maximum allowable pier stress? Did the concrete achieve the design tensile strength? Is the as-built shaft the correct size? Is the steel in the right place? How well do you know the applied loads?

Finally, I question the use of a 50k "average" load for a 20 pile group. What is the pile layout? Were group effects considered?

Please give us more details - this is a very interesting problem!

 

[brdgbldr]

You are correct in assuming the 50k load was for the max design load. As for your questions, Ill try to answer them as brief as possible. The layout is 4 rows of five columns. Also, the pile was designed correctly but we are still asking them to remove the cracked pile and determine the type and severity of the crack to see if there is a flaw in the design. Easier now than when the bridge is built.

As for some Lpile questions, since you have a good background, our analysis using the 3-d LPile Group Analysis, predicted deflections in the range of 3 to 12 inches, depending on the load case. Analyizing in the the 2-d mode, however, shows deflections in the range of 1 to 3 inches. The bridge is an arched structure (three main arches support the superstructure) and so there is a lot of torsion and moments in both horizontal (x and y) directions as well as moment in the vertical (z) direction. How sensitive is Lpile to these 3-d loadings as the deflections increase so dramatically even when loading in a third direction is not large compared to the basic 2-d loadings.

Second, when analyzing a single pile, we predicted a deflection of 1.5in for a fixed head case using only lateral load. The lateral field test showed a deflection of 0.1in for a free head test. This seemed unusually small as the area is near the coastline (within several hundred feet), groundwater is high (less than 10ft from surface), and a segment of the strata is an old fill (from about 20 to 30 feet below surface). We used soil data recommended by the soil engineer and were greatly surprised by the actual load test. Is there any reason offhand you may know that would give this great of a difference? Our original model (with 1.5in deflection) was almost exact to how the soil engineer himself set up his model. Yet he didn't even flinch. The owner trusts him though as he is familiar with the area. THanks again for all your help.

 

[Focht3]

Thanks, brdgbldr, for the additional details.

Here is my initial impression of what you have told me so far. Please keep in mind that I am primarily focusing on evaluating a load test, not on developing a rational design for your circumstance. Evaluating a load test requires a different mind set than developing a good design.

For me, the key piece of information you provided was the relatively small measured groundline deflection compared to the predicted groundline deflection. I will assume that your load test setup was carefully evaluated by all of the design professionals, and that all agree the results appear correct. (This may not be a correct assumption...) I will also assume that the subsurface stratigraphy has been properly defined and does not vary spatially within five pier diameters of the test pile.

I suspect that you have several factors affecting your prediction; it is often hard to determine which controls the majority of behavior, but I'll hazard a guess at that at the end of my comments.

The Story of k
The first suspect is the value chosen for k. I'll bet that you used the k values included in the LPILE manual; you may have even reduced them. This is a fairly common choice that Lymon Reese has promoted (for good reason.) The input value of k is used to limit the initial slope of calculated p-y curves; the majority of p-y curve equations use a parabolic curve equation such that the slope of the curve approaches infinity as the deflection approaches zero. This was a serious computational problem in the early days of p-y analyses (the 1950' through 1970's.) Computers couldn't handle very small - or very large - numbers.

So Lymon Reese, Hudson Matlock (and perhaps others) devised the secant modulus E_S value to limit the initial slope of the p-y curve. But they needed a rational way to choose this parameter. In some of their field research on relatively slender (L/d > 20) piles, they back-calculated E_S with depth and discovered - for this test site, pile and range of applied loads - that assuming E_S = k * x (where x is depth) reasonably matched their data points. The value of E₀ (secant modulus at the groundline) was small and was ignored.

And this approach works quite well in most cases. The computer code does not generate large initial slopes that cause computer code instability, and the resulting designs have some built-in conservatism (because lower bound, conservative k values are chosen resulting in larger shafts and greater required pile / pier penetrations.)

The Impact of Pile Length and/or Stiffness
Pile length is also an important factor for the obvious reason that, for a long enough pile, the lateral load at the surface is not (significantly) felt at the pile tip - and the tip does not deflect. This pile would be considered as relatively flexible and would likely have a L/d of more than 20. Conversely, a short pile would have significant tip deflection and a much smaller L/d ratio (perhaps less than 10) and would be considered stiff - even rigid.

I'll bet that your piers are relatively stiff - with L/d between 15 and 20. These piers can act in a way that makes the assumption that the effective E₀ = 0 incorrect; and if you have granular materials present in the shallow subsurface, the recommended k values may be entirely too small for your problem - evaluating the results of a load test. They may be entirely reasonable and rational for design.

Recommendations
1. Re-run your 2d single pile and 3d pile group analyses using k values for all strata that are at least 10 times greater than your current analyses used. Be sure and use the same load increments that you used in your field load test.

2. Plot load-deflection curves for all three (field, single pile, pile group) and compare them. If you observed an improvement in the prediction, but the single pile analysis still isn't close to the field test results, continue to increase the k values until they don't seem to affect the results. Choose the smallest k value that gives the stiffest response (i.e. smallest groundline deflection.)

3. Increase the pier modulus until you get a reasonable match on the load-deflection behavior. Modifying the other soil parameters usually won't cause much change (unless you change the e₅₀ values.)

4. Post your load-deflection results in this forum (in tabular form) and let me know whether you think this has been helpful -

 

[02111933]

Looks like the problem is the designer not the pile .. By the way is the bridge still intact? I am certain you will have long term deflection increase in the case of lateral loads on the piles .. esp given the poor quality (as explained) of the soil in which the pile was founded..

Interesting to know the results.. What rebar have you used ... epoxy caated or plain?