FOUNDATION SETTLEMENTS - hcmut.edu.vn

1 CHAPTER. 5. FOUNDATION SETTLEMENTS . 5-1 THE SETTLEMENT PROBLEM. FOUNDATION SETTLEMENTS must be estimated with great care for buildings, bridges, towers, power plants, and similar high-cost structures. For structures such as fills, earth dams, levees, braced sheeting, and retaining walls a greater margin of error in the SETTLEMENTS can usually be tolerated. Except for occasional happy coincidences, soil settlement computations are only best es- timates of the deformation to expect when a load is applied. During settlement the soil tran- sitions from the current body (or self-weight) stress state to a new one under the additional applied load. The stress change Aq from this added load produces a time-dependent accu- mulation of particle rolling, sliding, crushing, and elastic distortions in a limited influence zone beneath the loaded area.

2 The statistical accumulation of movements in the direction of interest is the settlement. In the vertical direction the settlement will be defined as AH. The principal components of AH are particle rolling and sliding, which produce a change in the void ratio, and grain crushing, which alters the material slightly. Only a very small frac- tion of A// is due to elastic deformation of the soil grains. As a consequence, if the applied stress is removed, very little of the settlement AH is recovered. Even though AH has only a very small elastic component, it is convenient to treat the soil as a pseudo-elastic material with "elastic" parameters E59 G', /i,, and ks to estimate SETTLEMENTS .

3 This would appear reason- able because a stress change causes the settlement, and larger stress changes produce larger SETTLEMENTS . Also experience indicates that this methodology provides satisfactory solutions. There are two major problems with soil settlement analyses: 1. Obtaining reliable values of the "elastic" parameters. Problems of recovering "undis- turbed" soil samples mean that laboratory values are often in error by 50 percent or more. There is now a greater tendency to use in situ tests, but a major drawback is they tend to obtain horizontal values. Anisotropy is a common occurrence, making vertical elastic values (usually needed) different from horizontal ones. Often the difference is substantial.

4 Because of these problems, correlations are commonly used, particularly for preliminary design studies. More than one set of elastic parameters must be obtained (or estimated) if there is stratification in the zone of influence H. 2. Obtaining a reliable stress profile from the applied load. We have the problem of com- puting both the correct numerical values and the effective depth H of the influence zone. Theory of Elasticity equations are usually used for the stress computations, with the in- fluence depth H below the loaded area taken from H = 0 to H - o (but more correctly from 0 to about AB or 5B). Since the Theory of Elasticity usually assumes an isotropic, homogeneous soil, agreement between computations and reality is often a happy coinci- dence.

5 The values from these two problem areas are then used in an equation of the general form AH = \ edH. Jo where e = strain = Aq/Es\ but Aq = f(H, load), Es = f(H, soil variation), and H (as pre- viously noted) is the estimated depth of stress change caused by the FOUNDATION load. The principal focus in this chapter will be on obtaining Aq, Es and H. It is not uncommon for the ratio of measured to computed AH to range as < ^ mea * >. 2. Current methodology tends to minimize "estimation" somewhat so that most ratios are in the to range. Note too that a small computed AH of, say, 10 mm, where the measured value is 5 or 20 mm, has a large "error," but most practical structures can tolerate either the predicted or measured value.

6 What we do not want is an estimate of 25 mm and a subsequent settlement of 100 mm. If we err in settlement computations it is preferable to have computed values larger than the actual (or measured) ones but we must be careful that the "large". value is not so conservative that expensive (but unneeded) remedial action is required. SETTLEMENTS are usually classified as follows: 1. Immediate, or those that take place as the load is applied or within a time period of about 7 days. 2. Consolidation, or those that are time-dependent and take months to years to develop. The Leaning Tower of Pisa in Italy has been undergoing consolidation settlement for over 700. years. The lean is caused by the consolidation settlement being greater on one side.

7 This, however, is an extreme case with the principal SETTLEMENTS for most projects occurring in 3 to 10 years. Immediate settlement analyses are used for all fine-grained soils including silts and clays with a degree of saturation S ^ 90 percent and for all coarse-grained soils with a large coef- ficient of permeability [say, above 10~3 m/s (see Table 2-3)]. Consolidation settlement analyses are used for all saturated, or nearly saturated, fine- grained soils where the consolidation theory of Sec. 2-10 applies. For these soils we want estimates of both settlement AH and how long a time it will take for most of the settlement to occur. Both types of settlement analyses are in the form of AH = eH = V - ^ (i = 1 to n) (5-1).

8 Where the reader may note that the left part of this equation is also Eq. (2-43<z). In practice the summation form shown on the right may be used where the soil is subdivided into layers of thickness Ht and stresses and properties of that layer used. The total settlement is the sum obtained from all n layers. The reader should also note that Es used in this equation is the constrained modulus defined from a consolidation test as \/mv or from a triaxial test using Eq. (e) of Sec. 2-14, written as ts {5la). ~ mv~ (1 + M ) ( 1 - 2 / 1 ). where ESyir = triaxial value [also used in Eq. (5-16)]. Note, however, that if the triaxial cell confining pressure 0-3 approximates that developed in situ when the load is applied, the tri- axial Es will approximate l/mv.}

9 In most cases the actual SETTLEMENTS will be somewhere be- tween SETTLEMENTS computed using the equivalent of \/mv as from a consolidation test [see Eq. (5- Ia)] and Es from a triaxial test. Unfortunately the use of Eq. (5- Ia) also requires estimating a value of Poisson's ratio /JL. 5-2 STRESSES IN SOIL MASS DUE. TO FOOTING PRESSURE. As we see from Eq. (5-1), we need an estimate of the pressure increase Ag from the applied load. Several methods can be used to estimate the increased pressure at some depth in the strata below the loaded area. An early method (not much used at present) is to use a 2 : 1. slope as shown in Fig. 5-1. This had a great advantage of simplicity. Others have proposed the slope angle be anywhere from 30 to 45.

10 If the stress zone is defined by a 2 : 1 slope, the Figure 5-1 Approximate methods of obtaining the stress increase qv in the soil at a depth z beneath the footing. pressure increase qv = Ag at a depth z beneath the loaded area due to base load1 Q is A. * - * - (B + ZXL+ Z) <5-2). which simplifies for a square base (B X B) to (5 2a). * = (BT^ - where terms are identified on Fig. 5-1. This 2 : 1 method compares reasonably well with more theoretical methods [see Eq. (5-4)] from z\ = Bio about zi = 4B but should not be used in the depth zone from z = 0 to B. The average stress increase in a stratum (H = Zi ~ Z\) is fZ2 O 1 I O \Zl 4. * -I,<BT3S*-*-i?l-sH ^*. 5-3 THE BOUSSINESQ METHOD FOR qv One of the most common methods for obtaining qv is the Boussinesq (ca.)