4.3 TEST RESULTS ON END-BEARING SINGLE PILE
4.3.4 Stage 4: NSF Due to Water Drawdown
The increase in soil effective stress due to ground water drawdown is another major cause of NSF on piles. In an attempt to simulate the effect of water drawdown on the development of NSF on the pile, at the end of the soil re-consolidation after pile driving, the ground water level was drawn down by about 2 m to the elevation of the top sand-clay interface. This was achieved by lowering the end of a hose connecting to a small opening located at the interface of the top sand layer and the underlying soft clay layer to an elevation lower than that of the opening by an mechanically controlled method to drain the water out of the soil container, as explained in detail in Chapter 3. The drawdown of the water head was recorded by the 1-bar PPT-6 embedded at the bottom of the top sand layer (see Fig. 4.1). As shown in Fig. 4.14, the water head was drawn down quickly by about 1.5 m within 4 days in prototype scale. Subsequently, the rate of water drawdown reduced due to the reduction of water head difference inside the soil container and the water outlet.
The water head was drawn down by about 2 m after about 20 days as reflected by the readings of PPT-6 shown in Fig. 4.14. The lowering of hydraulic head at the elevations of the PPTs embedded in the clay since the start of the water drawdown is also shown in Fig.
4.14. It can be seen that the readings of PPTs embedded in the clay generally lagged behind the actual variation of the ground water head since it took time for the clay to consolidate and reached the final hydrostatic state after about 180 days since the commencement of water drawdown, as shown in Fig. 4.14. The pore pressure profiles with depth in the soil at some selected time during the process are shown in Fig. 4.15. The distribution of hydrostatic pore pressure before water drawdown was essentially linear with depth and the projected ground water surface was slightly above the top sand layer.
108
After the water drawdown, all the PPTs began to converge towards the new hydrostatic state. After about 180 days, the pore pressure distribution in the soft clay stabilized and showed a linear distribution along the depth with the projected ground water level located near the interface of the top sand layer and the underlying soft clay as intended.
The variation of dragload during water drawdown is shown in the right portion of Fig.
4.11. Surprisingly, the dragload was observed to reduce abruptly upon the occurrence of water drawdown before it slowly picked up again after about 4 days. This was quite unexpected as it was the common wisdom that ground water drawdown would increase the vertical effective stress in the soil which would further induce NSF on the pile.
However, such puzzle dissolves the moment we begin to examine the soil settlement data plotted in Fig. 4.10. It can be seen that the soil settled during the soil re-consolidation stage after pile driving. However, upon water drawdown, an appreciable amount of ground heave in the magnitude of about 1~2 mm was observed accompanying the process of water lowering, as clearly indicated in Fig. 4.10. Obviously, it is this temporary ground heave that partially relieved the downdrag load on the pile. It is believed that the temporary ground heave was actually due to the sudden reduction of total and effective overburden stresses in the short term upon water drawdown, as illustrated in Fig. 4.16.
Before water drawdown, the stress on a soil element in the clay can be represented as:
sat clay sat
sand total
v1, h1γ , h2γ ,
σ = + (4.4)
u1 =h1γw+h2γw (4.5)
' 2 ' 1
1' sand clay
v hγ h γ
σ = + (4.6)
109
where γsand,sat and γ’sand are the saturated and effective unit weight of the top sand layer, respectively; γclay,sat and γ’clay are the saturated and effective unit weight of the clay, respectively; γw is the unit weight of water.
Upon water drawdown to the sand-clay interface, the hydraulic head at the soil element lies between the initial hydrostatic level, h1+h2, and the final hydrostatic level of h2 in the long term. Thus, the new stress state of the soil element after water drawdown can be expressed as:
sat clay wet
sand total
v2, h1γ , h2γ ,
σ = + (4.7)
h w
u2 = γ (h2 ≤h≤h1+h2) (4.8)
w sat clay wet
sand
v hγ h γ hγ
σ 2'= 1 , + 2 , − (4.9)
where γsand,wet is the unit weight of the unsaturated top sand layer after water drawdown.
Thus, following water drawdown, the increment of stresses on the soil element is the difference between the above two stress states as follows:
, 0
1 , 1
, = − <
Δσvtotal hγsandwet hγsandsat (4.10)
0 )
( − 1− 2 ≤
=
Δu h h h γw (4.11)
w sat
sand wet
sand
v hγ hγ h h h γ
σ '= 1 , − 1 , −( − 1− 2)
Δ (4.12)
Thus, the reduction of total vertical stress = h1γsand,wet −h1γsand,sat
110
≈ ×2 (14.1- 18.2)
≈ −8 kPa
In the short term under almost undrained condition, h may be close to h1+h2, thus,
Δσv'=h1γsand,wet −h1γsand,sat −(h−h1−h2)γw
≈h1(γsand wet, −γsand sat, )≈ −8kPa
which means that in the short term, the total as well as effective vertical stress will be reduced by about 8 kPa, inevitably leading to the immediate elastic rebound of the clay as verified by the test data in Fig. 4.10, and the corresponding relief of NSF along the pile shaft as shown in Fig. 4.11.
However, in the long run with the pore pressure within the clay converged to the hydrostatic state with reduced phreatic level, h = h2, and the increase in effective stress becomes:
) ) (
' 1 sand,wet 1 sand,sat 1 2 w
v hγ hγ h h h γ
σ = − − − −
Δ
=h1(γsand wet, −γsand sat, +γw)=h1(γsand wet, −γ' )sand
=h1(γsand wet, −γ' )sand (4.13)
which is approximately equal to 2 (14.1- (18.2-9.81))× ≈11 kPa.
It is this gradual increase of vertical effective stress that causes the further soil consolidation settlement with time after initial soil rebound as demonstrated in Fig. 4.10.
111
Accordingly, NSF along the pile shaft recovered with time upon the initial relief as shown in the right portion of Fig. 4.11. The NSF on the pile essentially stabilized 180 days after water drawdown, or 337 days after pile installation.
The reduced downdrag load along the pile 4 days after the commencement of water drawdown is shown by the leftmost curve in Fig. 4.17. With further soil consolidation, there was a certain increase in dragload along the pile compared to the dragload profile before water drawdown as shown in Fig. 4.17. The α and β curves are once again found to fit the test data reasonably well except the obvious over-estimation around the neutral point. The derived β value based on the increased vertical effective stress after water drawn down is 0.24, exactly the same as that derived at the soil re-consolidation stage.
Based on the undrained shear strength before water drawdown as presented in Fig. 4.2(a), the derived α value is 1.1, in excess of unity. The consolidation after water drawdown should have marginally increased the undrained shear strength of the clay beyond that before the water drawdown, which is not available for the present study. This explains why the α value is more than unity.