Assessing the total settlement from the settlement profile

[slam00000]

Your advanced analysis shows that all points of the soil under your foundation will settle and non uniform settlement curve is observed ; say look like a half or quarter circular shape (depending on the load of the foundation).
How the total settlement usually assessed:
Do you take an average value or you take settlement of the point having maximum settlement along this curve and consider it as the total settlement.
Please advise

P.S I did not expect to get stuck at a simple matter like this, this is probably due to lack of practical experience

 

[fattdad]

I'd need an explanation of a "non-uniform settlement curve" to address your specific question.

Generally speaking, I look at the distribution of stresses beneath the center of a footing and calculate the anticipated settlement. I typically calculate settlement using elastic properties rather than recompression index (i.e., 1-dimensional consolidation characteristics). I typically base the modulus profile on a representative soil profile derived from the overall field exploration. I typically consider one-half the total calculated settlement as the differential settlement.

Some items to consider:

The distribution of stresses beneat the center of a footing will be greater than the distribution of stresses beneath the edge or corner of the footing. As such, you could calculate that the center is more prone to settlment then the edge or corner. Won't happen like that though as the footing is too rigid for that to occur. Instead, the settlement at the center is the upper bound of what is likely to happen - just as the settlement at the corner is the lower bound of what is to happen.

Correlation between N-value and soil modulus is qualitative and developing a "typical" soil profile is subjective. Additionally, interpretation between borings and the sample frequency (i.e., what you actually see in the split spoon) is loaded with data gaps. Try not to get a pencil too sharp on these things, use reason and be conservative (to a degree).

It all needs to make sense though. . . .

f-d

¡papá gordo ain’t no madre flaca!

 

[namron2]

f-d

just to jump in on this thread and get some advice on something you touched on.

You mentioned that you generally use elastic parameters for footing settlements rather than Cr values from oedometer testing. Can I get some clarification on this for own edification:
1. Can I assume that you do this in overconsolidated soils only where Pc is not exceeded.
2. Does the elastic solution generally give comparable settlements to Cr values <Pc (i.e. is it naive to say that on the oedometer curve the Cr portion in the overconsolidated range is representative of elastic conditions?)
3. Where one was to derive elastic modulus values from other tests (triaxial, SPT, CPT, etc) where would you use undrained moduli vs drained moduli (for cohesive soils). I note that alot of engineers take the undrained case (short term), would this change for large heavy structures (e.g. tanks), or is this a case better handles with Cr (oedometer) type settlement analysis.
4. In deriving elastic moduli for settlement analysis I generally take tangential modulus values (e.g. from stress strain plots). Although I have little idea on some of the empirical correlations to field tests (most simply state and E value) is it fair to assume that they are also tangential values?
5. Are there any good easy to read papers on the use and abuse of elastic parameters in geotechnical engineering I can brush up on?

Thanks for you patience and sorry to highjack the thread

 

[himat42]

namron2
In response to your point # 5, please read paper published ASCE Geotechnical Journal Dec. 2007, (paper by: J.Brian Anderson et al. (pp.1494-1502). You may draw your own conclusion.

 

[BigH]

Might think about checking out the following for point#5:

http://www-civ.eng.cam.ac.uk/geotech_new/people/bolton/mdb_pub/148_ICSFF_Singapore_2004.pdf

http://www-civ.eng.cam.ac.uk/geotech_new/people/bolton/mdb_pub/143_IS_OSAKA_June_2004_93_98.pdf

http://ceprofs.tamu.edu/briaud/buchanan web/Lectures/Eighth Buchanan Lecture.pdf

For some of your other points - remember that "elastic" settlement is what, for clay, is called, generally, immediate settlement (basically using Recompression Ratio) - see Bolton's papers. I would use the E value representing the average of E values in the range of loads you are talking about. You might try to wade through a discussion in Terzaghi Peck and Mesri on modulus values.

Consolidation settlement is the long term settlement happening due, in simple explanation, to the squeezing out of the water (initially water takes the load being incompressible); as it is "forced" out the load is taken by the clay skeleton - which compresses under the load . . . etc. There is also secondary consolidation which is not really critical unles you have organics or certain types of clay minerals.

 

[fattdad]

1. Can I assume that you do this in overconsolidated soils only where Pc is not exceeded.

Yes. If you are exceeding Pc you should look at 1-d consolidation parameters.

2. Does the elastic solution generally give comparable settlements to Cr values <Pc (i.e. is it naive to say that on the oedometer curve the Cr portion in the overconsolidated range is representative of elastic conditions?)

Don't know whether naive or not. . . For overconsolidated soils, you often have the difficulty getting representative samples, interpreting the recompression ratio and finding a budget to look at consolidation parameters. If you know it's overconsolidated (i.e., LL>Wn) and you know the local soils correlation to Es is often easier/reliable. That said, I've never taken lab testing and looked for comparability - my judgement says they'd be very similar. I'd like to think that there'd be a relationship between Es and Cr - just don't know what it is.

3. Where one was to derive elastic modulus values from other tests (triaxial, SPT, CPT, etc) where would you use undrained moduli vs drained moduli (for cohesive soils). I note that alot of engineers take the undrained case (short term), would this change for large heavy structures (e.g. tanks), or is this a case better handles with Cr (oedometer) type settlement analysis.

I'm not sure how to respond to this. I appreciate the distinction between drained and undrained testing (i.e., in the triaxial cell) and I appreciate the distinction between drained (long-term) and undrained (short-term) loading. It's just that for overconsolidated soils, there is not alot of change in void ratio for the give loading condition. As such, I'd believe the distinction to be moot. That said, I'm mostly looking at the behavioral response in soils above the water table, where the distinction is truely moot.

4. In deriving elastic moduli for settlement analysis I generally take tangential modulus values (e.g. from stress strain plots). Although I have little idea on some of the empirical correlations to field tests (most simply state and E value) is it fair to assume that they are also tangential values?

Really not sure on this either. I mostly use correlation to SPT N-value to approximate the Es value. I have some references that I use for this - just not right now at home and all. . . . That said, I'd think tangential modulus value would be appropriate. For large areal loads, I've actually used hyperbolic parameters to depict the increase in modulus value with varying confining stress. That said, I had to do some guessing to do this and, well, a few assumptions - ha.

5. Are there any good easy to read papers on the use and abuse of elastic parameters in geotechnical engineering I can brush up on?

Schmertman (sp) has correlations for cone data. Ray Martin has correlation between SPT and Es for residual soils (Piedmont of Virginia). Martin published in ASCE GT journal.

Good questions and all, just not sure that I've fully thought through all the answers. I use a method that I worked up in graduate school and Prof. Duncan seemed to agree with (well at the time and all. . .)

Good luck.

f-d

¡papá gordo ain’t no madre flaca!