Settlement

[UP830]

Hello, i am first time poster.

I have a site where I need to calculate settlement of a pad (17' x 12') on clay layer over a sand layer.

Clay extends to 15' with sand to 40' plus.

I am calculating settlement down to 24' which is 2B. I have used boussinesq stress distribution to calculate delta P.

I have consolidation data and calculated the settlement of the clay layer using Mv of 0.046 (calculated from consolidation testing). I have calculated settlement in the sand layer using the following formulae based on elastic theory - Layer thickness * Delta P / modulas of elasticity of sand for each layer.

Is the above approach reasonable. Any geotech book i find only details settlement of a clay layer or settlement of sand layer. Never a combination of the two.

I welcome your advice. Thank you very much

 

[Okiryu]

I would just sum the elastic settlement of your sand plus the consolidation settlement of your clay for the total amount of expected settlement. You may also need to check the time for your calculated consolidation settlement to occur.

 

[oldestguy]

A common rule for situations like this is extend the depth of calculation to the point where the added pressure is 10% of existing pressure.

 

[fattdad]

the total extent of settlement is informed by soil modulus and the consolidation index (Cc) or consolidation ratio (Cec). If the clay is overconsolidated, then it'd be Cr, or Cer.

Mv relates to permeability and informs the time frame for the total consolidation to occur. Not sure how Mv would be related to the ultimate amount of consolidation.

I agree that the stresses should go to the point where delta sigma v is less than 10 percent of the overburden stress.

f-d

ípapß gordo ainÆt no madre flaca!

 

[LRJ]

M[sub]v[/sub] is the constrained modulus; m[sub]v[/sub] is the inverse of this. Both of these parameters vary with applied stress, so any settlement calculation should reflect this.

Personally I prefer to use the bilinear (C[sub]c[/sub] and C[sub]r[/sub]) approach (these parameters relate to void ratio, whereas m[sub]v[/sub] relates to strain), but if applied correctly, both approaches will give you the same result.

c[sub]v[/sub] is the parameter which defines time to end of primary consolidation.

Total settlement is equal to: elastic (undrained) settlement + primary consolidation (drained) + secondary consolidation (creep). I've seen several people state that primary consolidation is included in the undrained elastic settlement, though I've never actually seen a definitive reference explaining this.

As noted by others above, you should calculate the settlement of each layer separately (e.g. divide your soil profile into ~0.5m layers). It seems you already has a grasp of how to model varying stress distribution versus depth. You can typically neglect layers where the stress is less than 10% of the applied stress at surface (again, I haven't got a reference for this, but it is standard practice). After calculating these settlement components for each layer, sum them to obtain the total settlement.

 

[UP830]

Thanks you for responding.

I have a reasonable grasp on calculating settlement. I was thought that the Mv settlement method is generally suitable for soft to firm clay. For very soft cohesive soils it is better to use the Cc method.

Do you think the method of calculating settlement of granular using the change in stress x layer thickness / E is acceptable. I know there are other methods likes schemertmann for granular but i dont think these can be applied to mulitlayered soil?

 

[LRJ]

The M[sub]v[/sub] (or m[sub]v[/sub]) approach is no better or worse than the C[sub]c[/sub]/C[sub]r[/sub] approach for different clay soils providing that it is implemented correctly. M[sub]v[/sub] is derived from oedometer stress-strain plots whereas C[sub]c[/sub] and C[sub]r[/sub] are derived from oedometer stress-void ratio plots; strain and void ratio are related, so whether you prefer to calculate settlement in terms of strain or change in void ratio, you should come out with the same answer whatever approach you use. Whenever I have seen a difference in results it has always been due to an assumption that M[sub]v[/sub] is constant with stress/strain, which is totally incorrect. At least with the C[sub]c[/sub]/C[sub]r[/sub] approach the engineer has explicitly considered pre-yield and post-yield behaviour separately.

In sand soils, the M[sub]v[/sub] approach is likely the better option - purely because oedometer laboratory tests on sand will likely be recompacted, so C[sub]c[/sub] and C[sub]r[/sub] may not be representative of in situ conditions (unless you are compacting sand on site, I guess?). However, once again, changes in M[sub]v[/sub] with stress/strain need to be considered explicitly to produce meaningful results.

The M[sub]v[/sub] approach is consistent with your proposed approach for granular soils ('E' being oedometer Young's modulus (E[sub]oed[/sub]), which is the same thing as constrained modulus (M[sub]v[/sub])).