Odometer on a soil sample (clay) 1

[darragh91]

Maybe I'm thinking too much about this but in regard to the odometer test I have a few quires:

1) If you had a surcharge in reality and the effective stress falls outside the stress range you tested your soil sample in the lab, obviously you can't use these oedometer results and must preform a test with stress similar to your real life problem?

2) Why are we using void ratio e, and not height when plotting our results vs stress? Is this due to convention of is it due to the fact the void ratio is only a ratio and can be applied to any soil thickness? Where as your height of soil sample in the lab verus out say on a site could be 20mm verus 5m and this isn't much good but the void ratio provides the link as its only a ratio? But then again when calculating settlement the more layers you have the more accurate it is (smaller layers correspond to your sample?)

3) Would you take a few soil samples in reality for a soil type and get an average of results

4) What happens if the soil fails in test? How do you link oedometer results to shear strenght? Do you need to perform a triaxial test?

5) Why is the stress on a log scale?

 

[Erdbau]

Holtz and Kovacs has a great discussion on consolidation. There are other great texts as well. As for your questions:

1. If you reach virgin compression (past the preconsolidation pressure) in the test, you would probably be OK from a magnitude of settlement perspective, maybe less so on the time-rate. In a perfect world, you envelope your working stresses when assigning pressures.

2. Yes, you are using the 25-mm-thick sample to model behavior of a much thicker layer. In doing a settlement calculation, you essentially do is find the change in void ratio due to the load, which tells you how much strain occurs in the layer, and then you multiply the strain by the height to get settlement. You split the clay layer into several sublayers for different reasons (the stress increase is usually not uniform down the layer, some of the layer may pass the preconsolidation pressure, while some of it may not, etc.).

3. Depends on the size of the surcharge, your confidence in the quality of the samples collected, your degree of confidence in the testing lab, the amount of inherent variability of your particular clay, etc.)

4. Soil cannot fail in shear in an oedometer because it is not allowed to deform laterally. You cannot derive shear strength parameters from consolidation testing unless a correlation between shear strength and preconsolidation pressure has been developed via extensive previous lab testing of your particular clay.

5. When stress is plotted on a log scale, the recompression and virgin compression lines happen to fall on straight(ish) lines. If they were plotted on an arithmetic scale, they lines would not be straight, and would be concave up. It is just a convention based on observation of the actual soil behavior that makes it easier to compute settlement. You will notice that the settlement equations are based on log changes in stress.

 

[darragh91]

Ok thank you for this it has helped me greatly, but I still have a few questions?

1. Ok I'm still a bit unsure on this. Say for example I've performed a oedometer test for a soil sample. I'm now using these results to model settlemeant from a surcharge but this falls outside my results in the lad? Do I have to re perform the test stressing the sample to a range similar to the stresses my surcharge will induced when on soil in real world? You mention envelope your results can you explain this in more detail?

2. I'm still unsure why you can't just use the height vs. stress in test and have to use void ratio could you be more specific in your ''yes'' as replied earlier?

5. I get your first point on why you use the log scale. But I recently read in a book that if you plot void ratio vs stress on a semi log scale and look at an equal change in stress at lower stress vs higher stress, you'l find the void ratio is alot higher at lower stress then at higher stress for the same stress increase. This is models transfering a load deep on the ground? My question is if you didnt plot on log scale you wouldn't pick this behaviour up?

 

[Erdbau]

1. Hard to say without looking at the curve and the beginning and ending loads. If you are unsure, you have the budget, and it's critical enough, then yes, you may want to consider a retest. What I mean by envelope is to bound your stresses. If you are loading to X kPa, you want to load it at least that far out, and maybe a bit further.

2. You need to have some way to scale up the results. Let's say that in the oedometer (say 20 mm thick), you got 4 mm of settlement when you increased the load from 100 kPa to 1000 kPa. The settlement of a 20-meter-thick clay layer when you load it from 100 kPa to 1000 kPa isn't going to be 4 mm, it's going to be 4000 mm. Plotting void ratio versus load allows you to scale the results up to a thicker layer. In one-dimensional consolidation, the change in void ratio is directly proportional to strain: Change in strain = Change in void ratio / (1 + initial void ratio). To get settlement from strain, settlement = displacement = change in strain X initial height.

5. I don't quite understand what you are saying. A typical consolidation curve on the semi-log (arithmetic void ratio versus log effective stress) plot will have 2 regions: at low stresses, the curve is flatter, and then it gets steeper at a certain point. This point is the preconsolidation pressure. It's the highest load the clay has seen at any point in geologic time. Clay has "memory" and settles relatively small amounts at loads it has already "seen". Once you load past this point, the settlement is much greater for a given change in log stress. You may see the two regions on an arithmetic void ratio, arithmetic stress plot, but it's not going to be nearly as clear.

 

[darragh91]

1. Ya now I understand i think so each test you perform will be relivent for a different type of loading. Say your surcharge is 400kPa. You start your test at 50kPa and load in even increments to say maybe 600kPa. But if ur surcharge was 1000kPa you might start your test at 800kPa and finish at 1200kPa?

2. Ah yes because the void ratio is a ratio of volumes so like if u take a sample of soil and ratio of volumes is 200/400 the void ratio is 0.5 same as if you take a sample 1000/2000. The void ratio is like a property for the soil, its a ratio of volumes. This is the key for predicting settlement in large thick layers which you obviously can't for practically purposes load a 2000mm thick soil layer?

5. Ah yes that makes sense thank you. Could you vertify 1. + 2. ?


Overall your help was great. My name is Darragh I am a graduate civil engineer just trying to expand my knowledge of soils. I work as a site engineer for a tunnelling contractor.

 

[Erdbau]

1. Typically you will load in increments that double the previous load, and unload in increments of 4, so a typical load sequence would be 10, 20, 40, 80, 160, 320, 640, 1280, 320, 80, 20, 10 kPa. At a minimum you would want to get at least 2 loads past the preconsolidation pressure so you really get an idea of the slope of the virgin compression line. You may want to go out further if your surcharge loads are higher.

2. It sounds like you have the right idea. You can load a 2000mm-thick soil layer. It just may settle a lot.

Good for you for getting the theory down while you're in the field.

 

[darragh91]

Can I ask a couple of questions in regard to the Shear strenght of a soil?

1. Whats the main difference between the Consolidated Undrained test and Unconsolidated Undrained test? Why do you get a series of cirles of the same radii for Unconsolidated undrained test with no shear angle of failure? In what particular situation would you use both tests?

 

[hokie66]

I thought this was the Automotive forum for a minute there.

 

[molerat2210]

The UU vs CU question gets asked frequently. Please search the previous threads.

 

[BigH]

Regarding the "virgin" line of the consolidation curve: Some soils do not have, actually, a "straight" line when past the preconsolidation pressure (as when the line does become "straight") but it actually curves where at higher stresses it is a bit flatter than at the lower stresses in the virgin range. See Terzaghi Peck and Mesri (1995) Figure 16.1(b). Ran into this a lot in Northern Ontario soft clays.