- Email: [email protected]
Performance of laterally loaded piles in improved coal ash deposit
HOSTED BY
Available online at www.sciencedirect.com
ScienceDirect Soils and Foundations xxx (2017) xxx–xxx www.elsevier.com/locate/sandf
Performance of laterally loaded piles in improved coal ash deposit Jiunn-Shyang Chiou a,⇑, Teng-Ruei You b, Cheng-Chang Tsai a, Jin-Hung Hwang b b
a Department of Civil Engineering, National Taiwan University, No. 1, Sec. 4, Roosevelt Rd., Taipei 10617, Taiwan, ROC Department of Civil Engineering, National Central University, 300, Jhongda Rd., Jhongli City, Taoyuan County 32001, Taiwan, ROC
Received 19 May 2016; received in revised form 3 May 2017; accepted 1 July 2017
Abstract A series of pile load tests was conducted at a fossil fuel power plant in Taiwan to evaluate the performance of piles in a coal ash deposit that were improved using gravel compaction. The data from lateral load tests on two large-sized reinforced concrete piles were analyzed in this study. The original penetration resistance of the coal ash layer was low, but it increased after the ground improvement. Based on the load-defection curves of the test piles, the complete relationships of equivalent subgrade reaction coefficient kh with lateral displacement were derived and compared with those suggested by the Architectural Institute of Japan (AIJ) and the Japan Road Association (JRA). A composite SPT-N value was adopted to include the contribution of the gravel piles. It was found that the JRA equation using the composite SPT-N value fitted well with the relationships. Moreover, the experimental pile response and p-y curves were deduced using the inclinometer slope data. The relationships of the secant kh of the p-y curves vs. the normalized lateral displacement were further compiled and compared with those predicted by the JRA method. It was shown that the predicted curves using the composite SPT-N values also agreed well with the experimental relationships. Through a comparison of the profiles of the soil reaction with some existing lateral limiting soil pressure formulas, it was found that Prasad and Chari’s method was able to yield predictions close to the ultimate soil reaction. Ó 2017 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society. This is an open access article under the CC BYNC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Keywords: Coal ash; Ground improvement; Piles; Lateral loads; Field tests
1. Introduction Fossil fuel is a major source of power in Taiwan. A great deal of coal ash is generated during the combustion of coal; and therefore, the treatment of coal ash is an important issue for fossil fuel power plants. In Taiwan, coal ash is often added to Portland concrete to improve its engineering properties. It is also hydraulically filled into ponds near power plants to create land for constructing generator factories or coal storage warehouses. A hydraulically filled coal ash deposit is normally loose; and therefore, it has to be improved in order to increase its
Peer review under responsibility of The Japanese Geotechnical Society. ⇑ Corresponding author. E-mail address: [email protected] (J.-S. Chiou).
strength before any construction is done on it. The sand/gravel compaction pile method is commonly used for ground improvement. It installs piles made of compacted sands or gravels into the soft ground. The method can increase the liquefaction resistance and the bearing capacity of loose sandy soils (Okamura et al., 2003; Hatanaka et al., 2008) or accelerate the consolidation of soft clayey soils and the drainage of sandy soils (Rao et al., 1997; Yi et al., 2013). Sand/gravel compaction piling has also been introduced in Taiwan and is becoming increasingly popular for improving coal ash deposits (Hwang and Tu, 2002; Lin et al. 2009). The sand compaction pile method was shown to be effective for hydraulically filled coal ash ponds to increase SPT-N and CPT-qc values. In addition, another method, the heavy compaction method, was recently applied to a coal ash ground and found to be effective for improving the ground (Kokusho et al., 2012).
https://doi.org/10.1016/j.sandf.2017.08.019 0038-0806/Ó 2017 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
2
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
Piles are often used as foundations of coal storage warehouses in power plants to support the vertical structure loads. In addition to vertical loading, the piles may also be subjected to lateral loading, such as wind and earthquake loads on the structures. In engineering design practice, the p-y method is commonly used to analyze the nonlinear behavior of laterally loaded piles; it employs sets of nonlinear relationships between the soil reaction (p) and the pile displacement (y) to define the subgrade horizontal stiffness for different depths of soil along the piles. Using the data from pile load tests, a number of p-y models for sands (Reese et al., 1974; O’Neill and Murchinson, 1983) and clays (Matlock, 1970; Reese and Welch, 1975) have been proposed. However, few studies have been done on the lateral behavior of pile foundations in improved coal ash deposits; and therefore, more research is needed on the rational design of lateral piles in this type of ground. In 2013, the construction of ten coal storage silos was planned in an area filled with coal ash near a fossil fuel power plant in Linko, Taiwan. The area was to be improved by gravel compaction piles. Each silo was designed to have an inner diameter of 46 m and a height of 77.3 m, with the capacity to store about 7000 MN of coal. Piles were adopted to be the foundations of the silo structures. The piles were reinforced concrete piles with a diameter of 2 m and a length of 26.5 m. In order to understand the actual performance of the piles in a coal ash area improved by gravel compaction piles, a series of pile loading tests, including a compression test, a tension test, and two lateral tests, were conducted before the above construction.
For a better understanding of the lateral behavior of piles in an improved ash ground, this study analyzed the data from the lateral load tests and retrieved the experimental load-transfer curves (i.e., p-y curves). From the results, the characteristics of subgrade reaction coefficient kh and the p-y curves of the test piles, along with the applicability of general kh formulas in the design codes to improved coal ash grounds, were investigated. 2. Site conditions The soils at the coal silo construction site are composed of coal ash fill from EL. 9.0 m to EL. 9.0 m, and weathered sandstone below EL. 9.0 m. On the west side of the construction site, a region of 32 m 22 m in plan, which had been improved by gravel compaction piles, was chosen for performing the pile load tests, as shown in Fig. 1. The ground level at and around this region was lowered to EL. 4.5 m and the gravel piles were installed from this elevation. The water level was located at about EL. 1.0–1.4 m. As shown in Fig. 1, 16 boreholes (eight outside and eight inside the improved area) were adopted for the SPT and CPT subsoil investigation. Boreholes T5-T8 were located outside of the improved area; the soil properties in this region without the ground improvement are shown in Table 1. The penetration resistance of the coal ash fill was quite low; the average SPT-N value was 5 and the CPT-qc values were about 2500–3500 kN/m2. The gravel piles had a diameter of 1.0 m, a length of 11.3 m (the upmost 1.3 m was null), and a spacing of 3 m (three times the pile diameter). The backfill material for the piles was
N
6m
SPT-T5 CPT-T5
Construction area of coal silos Testing area
SPT-T1 CPT-T1
A1
A2
A3
A4
A5
B1
B2
B3
B4
B5
P1
5m
P6
SPT-T2 CPT-T2
5m
P4
P7
P2
5m
SPT-T8 CPT-T8
P5
5m
SPT-T3 CPT-T3
SPT-T6 CPT-T6
SPT-T4 CPT-T4
Pile cap
: gravel piles for ground improvement (pile diameter=1m) P3
3m
P8
6m
Coal silo (inner diameter= 46m)
: test piles and anchor piles (pile diameter=2m)
3m 3m
6m
5m
5m SPT-T7 CPT-T7
6m
Compression test: P4 (test pile); P1, P2, P6, P7 (anchor piles) Tension test: P5 (test pile); P2, P8 (anchor piles) Lateral test: P3 (test pile); P2, P8 (anchor piles) : SPT and CPT boreholes outside the improved area (T5-T8) : SPT and CPT boreholes in the improved area (T1-T4)
Fig. 1. Construction site of coal silos and test site of pile loading tests. Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
3
Table 1 Soil properties of region without ground improvement. Zone
Elevation (m)
Soil type
SPT-N
ct (kN/m3)
wn (%)
e
Gs
CPT-qc (kN/m2)
I II
4.5 to 9.0 <9
Coal ash Rock
2–9 (5) >50
8.34–16.97 (13.73) 20.70
32.1–98.4 (60.1) –
0.80–2.12 (1.59) –
2.14–2.63 (2.29) –
2500–3500 –
Note: bracketed values are the average values.
SPT-N
100 0
5
10
15
20
4
80
T3
70 2
60
T4 T6
50 0
40 30
Elevation (m)
Percent passing by weight
90
20 10 0 100
10
T7
-2
-4
1
Grain diameter (mm) -6
Fig. 2. Grain-size distribution for backfill material of gravel piles. -8
crushed granite. The grain-size distribution of the backfill material was controlled by suitability number SN, which was proposed by Brown (1977) and is defined as follows: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3 1 1 S N ¼ 1:7 þ þ ð1Þ 2 2 ðD50 Þ ðD20 Þ ðD10 Þ2 where D50, D20, and D10 are the diameters (in mm) through which 50%, 20%, and 10%, respectively, of the material is passing. The SN value in a range of 0–10 is an excellently rated backfill. Fig. 2 displays the grain-size distribution of the gravel pile material, in which D50, D20, and D10 are 18 mm, 14 mm, and 12.5 mm, respectively, which correspond to SN = 0.245. From triaxial compression tests on the samples taken from the gravel piles, the deformation modulus of the gravel piles was about 28.6 MN/m2 under an effective confining pressure of 50 kN/m2; the internal friction angle of the gravel piles was 38.7°. The area replacement ratio of the gravel piling was not high, only about 10%. Boreholes T1-T4 were located inside the improved area; the soil properties in this region after the ground improvement are shown in Table 2. The average increase in the total unit weight of the soil was from
-10
Fig. 3. SPT-N profiles near area of lateral pile load testing before and after improvement.
13.73 kN/m3 to 16.48 kN/m3. The average SPT-N value increased from 5 to 12 by a factor of 2.4. And the CPTqc values increased to about 7500–8500 kN/m2 by a factor of about 3. Fig. 3 further compares the SPT-N profiles with depth before the ground improvement (T6 and T7) with those after the ground improvement (T3 and T4) around the area of the lateral pile load tests. It was found that the effect of the improvement in the SPT-N values was more significant in the deeper soil than in the shallower soil. This result could be expected since ground improvement by gravel piles is generally less effective for shallow soil due to lower confining pressure. The profile after improvement could be roughly divided into two layers: the average SPT-N values of the upper layer (EL. 2 m to 4 m) increased from 4 to 10, while the average SPT-N values of the lower layer (EL. 4 m to 8 m) increased from 6 to 17.5.
Table 2 Soil properties of region after ground improvement. Zone
Elevation (m)
Soil type
SPT-N
ct (kN/m3)
wn (%)
e
Gs
CPT-qc (kN/m2)
I II
2.7 to 9.0 <9
Coal ash Rock
6–19 (12) >50
14.81–17.76 (16.48) 20.70
30.5–56.7 (46.1) –
0.64–1.47 (1.19) –
2.14–2.42 (2.28) –
7500–8500 –
Note: bracketed values are the average values. Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
4
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
3. Lateral load testing The ground level of a region of 12 m 22 m in the improved area was further lowered from EL. 4.5 m to EL. 2.7 m. In this low area, eight reinforced concrete piles were cast in place by the full casing method, and their positions are shown in Fig. 1. P3, P4, and P5 were the test piles for the lateral test, the compression test, and the tension test, respectively. Except for P3, P4, and P5, the piles were anchor piles. After the tension test, the tension load was released and then horizontal loading was applied to P5 for an additional lateral load test. The test piles had a length of 26.5 m and a diameter of 2.0 m. P3 and P5 had the structural details shown in Table 3. The pile length was divided into three regions: L1 for the top region of 6.5 m, L2 for the middle region of 8 m, and L3 for the rest of the region. The anchor piles had a length of 30.5 m and a diameter of 2.0 m. The structural details are shown in Table 4, in which Region ANL1 was for the upper region of 6.5 m and Region ANL2 was for the rest of the region of 24 m. The concrete strength was 30.4 MPa and the steel strength was 412 MPa. According to the material properties of the pile sections, the plastic moments were estimated
through section analyses, as shown in Table 3. Furthermore, assuming that the soil was cohesionless and considering two possible pile failure modes, the ultimate lateral capacities were estimated to be about 44,000 kN for the rigid body rotation due to soil failure and 6631 kN for the plastic hinge formation. It was found that the failure might be governed by the formation of plastic hinging. Fig. 4(a) displays the test setup of P3. P3 was pushed by an actuator supported by two anchor piles (P2 and P8) via a reaction beam. In P3, a load cell and LVDTs (linear variable displacement transducers) were set to measure the applied load and the pile-head displacement, respectively. An inclinometer and rebar stress meters were installed in the pile to measure the profiles of the displacement and the bending stress of the pile with depth. In addition, LVDTs were also set on the anchor piles to simultaneously record their pile-head displacements. The maximum lateral load in the test was 3630 kN. This load was far below the estimated ultimate lateral capacity, implying that plastic hinging had not occurred. During the testing, a gap of about 6–7 cm formed between the back of the pile and the adjacent soil. The pile-head load-deflection curve of P3 is shown in Fig. 5, in which the pile-head displacement
Table 3 Structural details of P3 and P5. Region
Length (m)
Diameter (m)
Cover (m)
Longitudinal reinforcements
Transverse reinforcements
Calculated plastic moment (kN m)
L1 L2 L3
6.5 8 12
2
0.1 0.1 0.1
92-D36 46-D36 32-D36
[email protected] [email protected] [email protected]
27,105 15,505 10,980
Table 4 Structural details of anchor piles. Region
Length (m)
Diameter (m)
Cover (m)
Longitudinal reinforcements
Transverse reinforcements
ANL1 ANL2
6.5 24
2
0.1 0.1
92-D36 46-D36
[email protected] [email protected]
P2
P3
P3 P8
P5 P2
Steel pad
Steel pad
Excavated part Load cell Reaction Actuator beam Steel pad
Anchor piles
P3
Not excavated EL. 2.70 m
EL. 2. 70m Load cell
EL. 2.20m
EL. 1.70 m Reaction Actuator beam Steel pad
Anchor piles
(a) P3
P5
(b) P5 Fig. 4. Test setups of P3 and P5.
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
namely, about 1.2 times that of P3. The larger resistance of P5 might come from the contribution of the unexcavated soil around the pile head.
5000 4500
Lateral load (kN)
4000 3500
4. Preliminary analysis for subgrade reaction coefficient
3000 2500
For a semi-infinitely long pile in Winkler-type soil with constant horizontal subgrade reaction coefficient kh under pile-head lateral loading, the governing equation of the pile response can be expressed as
2000 1500
P3 Anchor piles (P3) P5
1000 500 0 0
10
20
30
40
50
60
70
80
Lateral displacement (mm)
Fig. 5. Load-deflection curves of P3, anchor piles (P3) and P5.
4000 3500 3000
Lateral load (kN)
5
d 4y ¼ p ¼ k h Dy dx4
2000 1500
1000 500 0
0.002
0.004
0.006
0.008
0.01
Lateral rotation (rad)
Fig. 6. Load-rotation curve of P3.
at the maximum load was 68.4 mm (about 3.5% pile diameter). The pile-head load-rotation curve is shown in Fig. 6, in which the pile-head rotation at the maximum load was 0.0087 rad. Fig. 5 also shows the average load-deflection curve of the anchor piles, in which the load represents one-half of the total load applied and the displacement is the average displacement of both anchor piles. The anchor piles showed larger lateral resistance because the soil in front of the upper parts of the piles above the position of the applied loads had not been removed (Fig. 4(a)). Fig. 4(b) displays the test setup of P5, which was similar to that of P3. The main difference in the setups of P5 and P3 is that the soil in front of P5 above the position of the actuator was not excavated. This part of the soil provided some resistance to the lateral loading. In P5, a load cell and LVDTs were set to measure the applied load and the pilehead displacement, respectively. An inclinometer was installed in the pile to measure the profiles of the pile displacement with depth. Fig. 5 also shows the pile-head load-deflection curve of P5. The maximum load was about 4415 kN under which the pile-head lateral displacement reached 73 mm (about 3.7% pile diameter). The maximum load had not reached the estimated ultimate lateral capacity; and therefore, the plastic hinging had not occurred in the pile. Compared to P3, P5 had larger lateral resistance,
ð2Þ
in which EI is the pile flexural rigidity, p is the soil reaction, D is the pile diameter, and kh is the horizontal subgrade reaction coefficient.The closed-form solution of the above equation has been derived (Hetenyi, 1946). Letting qffiffiffiffiffiffi hD b ¼ 4 k4EI , the deflection along pile y can be expressed as y¼
2500
0
EI
H ebx cos bx 2EIb3
ð3Þ
in which b is referred to as the characteristic coefficient of the pile-soil system. Due to the analytical form of the solution, the solution is extensively adopted in Japanese design codes (e.g., Architectural Institute of Japan, 2001; Japan Road Association, 2012). In these design codes, the subgrade reaction coefficient is usually calibrated with the SPT-N value. Two approaches for determining the subgrade reaction coefficient are presented below. (1) Architectural Institute of Japan (AIJ, 2001) The Architectural Institute of Japan suggests 0:5 y k h ðyÞ ¼ k hr yr
ð4Þ
where k h is the subgrade reaction coefficient (kN/m3), y is the lateral displacement (m), and k hr is the subgrade reaction coefficient (kN/m3) at yr (yr = 0.01 m), which is expressed as k hr ¼ 2:53E0 D3=4
ð5Þ
in which D is the pile diameter (m) and E0 is the soil deformation modulus, which can be determined by unconfined or triaxial compression testing or borehole deformation testing, or it can be estimated based on an empirical formula, namely, E0 = 700 N (kN/m2) (N is the SPT-N value). (2) Japan Road Association (JRA, 2012) The Japan Road Association suggests 0:5 y k h ðyÞ ¼ k hr ð6Þ yr where k h is the subgrade reaction coefficient (kN/m3), y is the lateral displacement (m), k hr is the subgrade reaction
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
6
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
coefficient (kN/m3) at yr (yr = 0.01 D (0.05 m), and D is the pile diameter in meters), which is expressed as 3=4 BH ð7Þ k hr ¼ k h0 0:3 k h0 ¼
1 aE0 0:3
ð8Þ
in which BH is the equivalent foundation width (m), E0 is the soil deformation modulus, and a is the modification factor associated with method used to determine E0. E0 can be determined by plate load testing, unconfined or triaxial compression testing, or borehole deformation testing, or it can be estimated using an empirical formula E0 = 2800 N (kN/m2). When E0 is determined using unconfined or triaxial compression testing or borehole deformation testing, a = 4, and when E0 is determined by plate load testing or estimated by the SPT-N value, a = 1. Equivalent foundation width BH is expressed as follows: pffiffiffiffiffiffiffiffiffi BH ¼ D=b ð9Þ where, as mentioned previously, b is the characteristic coefficient of the foundation (m1). Based on the data on the lateral load and the pile-head displacement of each test, Eq. (3) for x = 0 (i.e., the pilehead displacement) was applied to back-figure the equivalent subgrade reaction coefficients with lateral displacement. Fig. 7 shows comparisons of the trends in the equivalent subgrade reaction coefficients with displacement for both tests. The equivalent subgrade reaction coefficients at a small displacement were from the anchor piles of the P3 test, while the equivalent subgrade reaction coefficients at a large displacement were from the test piles. The trends were consistent and the subgrade reaction coefficients decreased with displacement. At the same displacement level, the subgrade reaction coefficient from the P5 test was slightly larger than that from the P3 test. The difference might be due to the unexcavated soil around the pile head of P5. Fig. 7 further compares the equivalent subgrade reaction coefficient data with the AIJ equation (Eq. (4)) and
Subgrade reaction coefficient (kN/m3)
180000 P3 test (P3 pile) P3 test (anchor piles) P5 test (P5 pile) AIJ (N=12) JRA (N=12) JRA (Ncomp=14.9)
160000 140000
120000 100000 80000
60000 40000 20000 0 0
10
20
30
40
50
60
70
80
Displacement (mm)
Fig. 7. Comparisons of experimental horizontal subgrade reaction coefficients with AIJ and JRA equations.
the JRA equation (Eq. (6)). In the figure, it can be seen that when the average field improved SPT-N value (N = 12) was used, the computed kh using the AIJ method was significantly smaller than the equivalent kh, while the computed kh using the JRA method was slightly smaller than the equivalent kh. This comparison also shows that the kh from the AIJ method is much smaller than that from the JRA method. The main reason for the above difference is that the E0 estimated by the JRA method is four times that estimated by the AIJ method. Since the field-improved SPT-N values were measured among the gravel piles, they could only represent the improved properties of the coal ash, although such an improvement is normally evaluated based on the SPT-N values among the gravel piles. Tamura et al. (2012) conducted cyclic lateral loading centrifuge tests, and the results showed that the lateral resistance of a new structural pile, located in a place surrounded by existing structural piles (spacing = 3.5 D), increased slightly. Based on their study, and considering the possible influence of gravel piles, the following simple equation was used to estimate a weighted SPT-N value Ncomp to represent the whole composite improved ground (Japanese Geotechnical Society, 1998): N comp ¼ ð1 As ÞN 1 þ As N p
ð10Þ
where N1 is the SPT-N value among the gravel piles, Np is the SPT-N value at the center of the gravel pile, and As is the replacement ratio. Due to the lack of an SPT-N value at the center of the gravel pile, Np = 41 was estimated based on E0 = 28.57 MN/m2 (from the triaxial compression test) multiplied by the modification factor of a = 4. For As = 10% and N1 = 12, Ncomp = 14.9 was computed based on Eq. (10). Using N = Ncomp (=14.9) instead of 12, the AIJ method still underestimated the equivalent kh; however, the JRA method gave closer kh, as shown in Fig. 7. 5. Pile response analysis and p-y curves As mentioned before, for P3 and P5, the inclinometers were installed to measure the slope profiles of the piles with depth during lateral loading. With the inclinometer slope data, the regressive method proposed by Chiou et al. (2008) was applied to analyze the pile response based on the beam theory. The method uses a composite moment function for regression with the inclinometer slope data. In the analysis, the flexural rigidity of the pile section is required to link the curvature and the moment. Therefore, this study performed section response analyses to construct the nonlinear moment-curvature relationships of the pile sections. The simplified curves for regions L1 and L2 are shown in Fig. 8, on which four points are marked, namely, the cracking moment, the moment at the initial yielding of steel, the effective yielding moment, and the ultimate moment. According to the moments at the different depths calculated from the moment function, the corresponding reduced flexural rigidity of the pile sections would be chan-
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
7
1200
30000
25000
1000
20000
800
p (kN/m)
Moment (kN-m)
(a)
15000
10000
z=0.5m z=1.0m z=2.0m z=3.0m z=4.0m z=5.0m z=6.0m z=7.0m
600
400 L1-simplified
5000
200
L2-simplified
0 0
0.002
0.004
0.006
0.008
0.01
0.012
0
0.014
0
Curvature (1/m)
0.01
0.02
0.03
0.04
0.05
1200
Fig. 8. Moment-curvature curves for pile sections in L1 and L2 regions.
(b)
z=0.5 m z=1.0 m z=2.0m z=3.0m z=4.0m z=5.0m z=6.0m z=7.0m
1000
ged based on Fig. 8, depending on the occurrences of cracks or the yielding of the pile sections. By using the above-mentioned regressive method, the slope profiles and the corresponding regressive moment functions of P3 were obtained as shown in Fig. 9(a) and (b), respectively. Based on the moment functions and the sectional moment-curvature curves of the test pile, it was found that test pile P3 cracked very early, prior to the lateral load of 1962 kN; however, it had not yielded at the maximum load. Fig. 9(b) also compares the regressive moment functions with the moments back-computed based on the rebar stresses measured from the rebar stress meters, of which the trends were similar. Integrating the slope profiles and differentiating the moment functions gave the displacement and soil reaction profiles with depth, as shown in Fig. 9(c) and (d), respectively. By collecting y and p at specific depths for different load levels, the experimental p-y curves could be deduced. Fig. 10(a) shows the p-y curves at depths from 0 to 7 m; the curves were softer for the depths from 0 to 1 m, the stiffness and strength increased with depth, and the soil appeared to reach the ultimate state and have a slight softening response at a larger displacement. For the depths from 2 to 5 m, the trends in the curves were close; for the depths from 6 to 7 m, the (a) -0.01
Depth (m)
-0.015
(b)
Slope (m/m) -0.005
0
0.06
y (m)
0
5000
p (kN/m)
600
400
200
0 0
10000
0.01
0.02
0.03
0.04
0.05
0.06
0.07
y (m)
Fig. 10. Experimental p-y curves: (a) P3 and (b) P5.
curves showed linear trends. For the depths larger than 7 m, because of the influence of the rock stratum, the displacement response was very small and even had reverse displacements in the rock stratum, such that the trends in the associated p-y relationships became random and irregular. As a whole, the subgrade stiffness around this area was large, about 1350–1500 MN/m3, and the corresponding displacement range was about 0.0003 m. The same analysis approach was applied to the P5 test. Fig. 11 displays the profiles of the slope, the moment, the (c)
Moment (kNm)
0.005 -5000
800
15000 -6
(d)
y (cm) -4
-2
0
2
-1500
Soil reaction (kN/m) -1000
-500
0
0
0
0
0
2
2
2
2
4
4
4
4
6
6
6
6
8
8
8
8
10
10
10
10
12
12
12
12
500
1000
981kN 1962kN 2943kN 3630kN 981kN-exp
981kN
1962kN 14
14
2943kN 3630kN
1962kN-exp
981kN (rebar stress meter)
2943kN-exp
1962kN (rebar stress meter)
16
16
2940 kN (rebar stress meter)
3630kN-exp
981KN 1962kN
14
981kN
14
1962kN 2943kN
16 3630kN
2943kN
16
3630kN
3630kN (rebar stress meter) 18
18
18
18
Fig. 9. Response profiles of P3: (a) slope, (b) moment, (c) displacement, and (d) soil reaction. Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
8
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx (a)
Depth (m)
-0.012
(b)
Slope (m/m) -0.007
-0.002
981kN 1962kN 2943kN 3433kN 3924kN 4415kN 981kN-exp 1962kN-exp 2943kN-exp 3433kN-exp 3924kN-exp 4415kN-exp
(c)
Moment (kNm)
0.003 -5000
0
5000
10000
15000 -8
(d)
y (cm) -6
-4
-2
0
2
-1500
Soil reaction (kN/m) -1000
-500
0
0
0
0
0
2
2
2
2
4
4
4
4
6
6
6
6
8
8
8
8
10
10
10
10
12
12
12
500
12 981kN
14
16
18
(a)
981kN 1962kN 2943kN 3433kN 3924kN 4415kN
14
16
981kN 1962kN 2943kN 3630kN 3924kN 4415kN
1962kN
14
14
2943kN 3433kN
16
16
3924kN 4415kN
18
18
(b)
18
(c)
(d)
Fig. 11. Response profiles of P5: (a) slope, (b) moment, (c) displacement, and (d) soil reaction.
displacement, and the soil reaction. Fig. 10(b) shows the experimental p-y curves of P5 for depths from 0 m to 7 m. Compared to Fig. 10(a), it is seen that the trends in the p-y responses for P3 and P5 are similar.
(1) Secant subgrade reaction coefficient of p-y curves With the obtained p-y curves, this section further analyzes the secant slope of the curves which is called the secant subgrade reaction coefficient. Simplifying the soil layer as a homogeneous layer, a displacement of 1% pile diameter (0.01 D) is used as the normalization factor to build the relationship of the secant subgrade reaction coefficient (p/(D y)) and the normalized lateral displacement (y/0.01 D), as shown in Fig. 12. Compared to the JRA equation using Ncomp = 14.9, it was found that the trend agreed averagely well with the JRA equation. As mentioned before, the soil after improvement was not uniform with depth and could be roughly divided into two layers. Therefore, the soil was considered as two layers for P3 in order to build the corresponding trends in the secant subgrade reaction coefficient vs. normalized displacement. Consistent with the soil profile, the kh in the lower layer was larger than that in the upper layer. Furthermore, the JRA method was applied to predict the relationships using the average SPT-N values for the two layers. It can be seen that the predicted curves are in good agreement with the experimental curves, as shown in Fig. 13. The JRA method seems to be applicable for simulating the respective relationships of the subgrade reaction coefficients when composite SPT-N values of different depths are adopted. (2) Lateral limiting soil pressure There are several formulas proposed for lateral limiting soil pressure pu. From the soil reaction profiles of P3 and
P3 P5
kh_sec (kN/m3)
200000
JRA equation (Ncomp=14.9)
150000
100000
50000
0 0
0.5
1
1.5
2
y/0.01D
Fig. 12. Secant subgrade reaction coefficient - normalized displacement curve.
250000 P3 (z=0-4m) JRA equation (Ncomp=13.1 for Nave=10)
200000
P3 (z=4-7m)
kh_sec (kN/m3)
6. Discussion
250000
JRA equation (Ncomp=19.4 for Nave=17.5) 150000
100000
50000
0 0
0.5
1
1.5
2
y/0.01D
Fig. 13. Secant subgrade reaction coefficient - normalized displacement curve for different soil layers.
P5, it was observed that the soil reactions at the shallow depth of soil seemed to have a limit when lateral loading is increased. Therefore, with the obtained soil reaction profiles, the applicability of the existing formulas for the
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
(a) -1500
(b)
Soil reaction (kN/m) -1000
-500
0
500
1000
-1500
1200
Soil reaction (kN/m) -1000
-500
0
0
0
1
1
2
2
3
3
4
4
9
500
5
Depth (m)
1962kN 2943kN
8
3630kN Brom Brinch Hansen
9
Prasad and Chari
6
1962kN 2943kN 3433kN 3924kN 4415kN Brom Brinch Hansen Prasad and Chari
10
600 z=0.5m z=1.0m z=2.0m z=3.0m z=4.0m z=5.0m z=6.0m z=7.0m
400
5
6
7
800
p (kN/m)
Depth (m)
1000
200
7
0
8
0
0.01
0.02
0.03
0.04
0.05
0.06
y (m)
9
Fig. 15. Derived p-y curves for use in numerical simulation.
10
Fig. 14. Comparisons of lateral limiting pressure formulas: (a) P3 and (b) P5.
5000 4500
4000
Lateral load (kN)
limiting pressure were investigated. In this study, three formulas for pu, proposed by Brinch Hasen (1961), Brom (1964), and Prasad and Chari (1999), were adopted for comparison. These formulas were basically derived from the theory of earth pressure to represent the lateral ultimate resistance provided by the surrounding soil at the fully mobilized passive stress state. These three methods for pu are briefly introduced below.
3500 3000 2500 2000
P3
1500
P5
1000
Analysis (P3) Analysis (P5)
500 0
(1) Brinch Hasen method (1961)
0
10
20
30
40
50
60
70
80
Lateral displacement (mm)
pu ¼ czK q
ð11-1Þ
Fig. 16. Comparisons of simulated and experimental load-displacement curves for P3 and P5.
(2) Brom method (1964) pu ¼ 3czK p
ð11-2Þ
(a)
(3) Prasad and Chari method (1999)
(b)
Moment (kNm)
-5000
0
5000
10000
15000
-1500
Soil Reaction (kN/m) -1000
-500
0
0
0
ð11-3Þ
2
1
where c is the unit weight of the soil, z is the depth, Kq is the Brinch Hasen earth pressure coefficient, Kp is the Rankine passive earth pressure (=tan2(45 + //2)), and / is the angle of internal friction of the soil. Fig. 14 plots the profiles of the soil reaction for P3 and P5, on which the limiting pressures by the three methods are shown. It can be seen that the limiting pressure estimated by the Prasad and Chari method could envelop the ultimate soil reactions at shallow depths of soil, but the other two methods significantly underestimated the limiting pressure.
4
2
3 6
4
Depth (m)
Depth (m)
pu ¼ 0:8cz10ð1:3 tan /þ0:3Þ
500
8
5
10
6 12
981kN (rebar stress meter) 1962kN (rebar stress meter) 2940 kN (rebar stress meter)
14
3630kN (rebar stress meter)
981 kN (analysis) 16
1962 kN (analysis) 2943 kN (analysis) 3630 kN (analysis)
18
7 981kN 1962 kN 2943 kN 3630 kN Limiting pressure
8
9
10
Fig. 17. Simulated responses of P3: (a) moment and (b) soil reaction.
7. Verification of JRA’s kh method and Prasad & Chari’s pu model In order to evaluate the applicability of the JRA’s kh method and Prasad and Chari’s pu model, a numerical simulation was performed to analyze the P3 and P5 tests using the Winkler-beam model, in which the pile is
modeled by beam elements and the soil reactions are modeled by p-y curves. The secant slope of the p-y curves was determined by the JRA formula and the maximum soil reaction was limited by Eq. (11-3). Based on the profile of the average SPT-N value of boreholes T3 and T4 and the limiting pressure, the derived p-y curves varied with
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
10
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
depth, as shown in Fig. 15. Fig. 16 compares the numerical load-deflection curves of P3 and P5 with the experimental ones. It can be seen that except for some underestimation of the displacement at the largest load, the overall simulated curves were in good agreement with the test data. Fig. 17(a) further compares the simulated moment profiles of P3 with those observed in the test. It can be seen that the simulated moments were slightly larger than the observed ones, but that their overall trends were consistent. In addition, Fig. 17(b) compares the simulated soil reaction profiles with the set limiting pressure. It was found that the soil shallower than about 2.5 m had reached the limiting value. 8. Conclusions From this study, the following conclusions have been deduced: 1. The SPT-N values were increased by a factor of 2.4 and the CPT-qc values were increased by a factor of about 3 for the coal ash ground improved by gravel compaction piles with an area replacement ratio of about 10%. The effect of the improvement was larger in the deeper soil than in the shallower soil. 2. The piles in the improved coal ash ground had large lateral resistance. At the lateral displacement of 1% of the pile diameter (20 mm), the lateral load was about 2200 kN. The soil right in front of the upper parts of the piles above the applied load increased the lateral resistance by about 20%. 3. Based on the load-defection curves of the piles, the trends in kh with lateral displacement could be well constructed. Using the field-average improved SPT-N value, the AIJ method significantly underestimated kh, while the JRA method slightly underestimated kh. However, when a larger composite SPT-N value was used, considering the possible contribution of the stiffness of the gravel piles, the predicted trends in kh using the JRA method were improved. 4. The experimental p-y curves could be expressed as relationships of the secant subgrade reaction coefficients with lateral displacement. The JRA equation, using composite SPT-N values, fitted well with the relationships. Moreover, the limiting pressure proposed by Prasad and Chari was able to envelop the ultimate soil reactions at shallow depths of the soil. 5. The numerical simulation using the p-y curves whose secant stiffness and ultimate pressure were determined by the JRA method and Prasad and Chari’s model, respectively, yielded predictions that corresponded well with the experimental results.
Acknowledgements The authors would like to thank the Taiwan Power Company, T.Y. LIN International Taiwan Consulting Engineers, Inc., and Kung Sing Engineering Corporation for providing the test data. References Architectural Institute of Japan (AIJ), 2001. Recommendation for the Design of Building Foundations (in Japanese). Brinch Hansen, J., 1961. The ultimate resistance of rigid piles against transversal forces. Danish Geotechnical Institute, Copenhagen, Bulletin No. 12, 5–9. Brom, B.B., 1964. Lateral resistance of piles in cohesionless soils. J. S MFE Divis., ASCE 90, 123–156. Brown, R.E., 1977. Vibrofloation compaction of cohesionless soils. J. Geotech. Eng. Div., ASCE 103 (12), 1437–1451. Chiou, J.S., Chen, C.H., Chen, Y.C., 2008. Deducing pile responses and soil reactions from inclinometer data of a lateral load test. Soils Found. 48 (5), 609–620. Hatanaka, M., Feng, L., Matsumura, N., Yasu, H., 2008. A study on the engineering properties of sand improved by the sand compaction pile method. Soils Found. 48 (1), 73–85. Hetenyi, M., 1946. Beam on Elastic Foundations. University of Michigan Press, Ann Arbor, Mich. Hwang, J.H., Tu, T.Y., 2002. Ground vibration during gravel pile construction. J. Mar. Sci. Technol. 10 (1), 36–46. Japan Road Association (JRA), 2012. Design Specifications for Highway Bridges - IV: Substructures (in Japanese). Kokusho, T., Nakashima, S., Kubo, A., Ikeda, K., 2012. Soil investigation of fly ash deposit improved by heavy compaction method. J. Geotech. Geoenviron. Eng. 138 (6), 738–746. Lin, S.S., Chien, S.J., Liao, J.C., 2009. Ground improvement of a coal ash storage pond using sand compaction piles. Lowland Technol. Int. 10 (2), 21–29. Matlock, H., 1970. Correlations for design of laterally loaded piles in soft clay. In: Proc., 2nd Annual Offshore Technology Conference, Houston. Okamura, M., Ishihara, M., Oshita, T., 2003. Liquefaction resistance of sand deposit improved with sand compaction piles. Soils Found. 43 (5), 175–187. O’Neill, M.W., Murchinson, J.M., 1983. An Evaluation of P-Y Relationships in Sand. Rep. to American Petroleum Institute, Washington, D. C.. Prasad, Y.V.S.N., Chari, T.R., 1999. Lateral capacity of model rigid piles in cohesionless soils. Soils Found. 39 (2), 21–29. Rao, S.N., Reddy, K.M., Kumar, P.H., 1997. Studies on group of stone columns in soft clays. Geotech. Eng., Bangkok 28 (2), 165–182. Reese, L.C., Cox, W.R., and Koop, F.D. 1974. ‘‘Analysis of laterally loaded piles in sand. In: Proc., 6th Annual Offshore Technology Conference, Houston, pp. 672–690. Reese, L.C., Welch, R.C., 1975. Lateral loading of deep foundations in stiff clay. J. Geotech. Eng., ASCE 101 (7), 633–649. Tamura, S., Adachi, K., Sakamoto, T., Hida, T., 2012. Effects of existing piles on lateral resistance of new piles. Soils Found. 52 (3), 381–392. The Japanese Geotechnical Society, 1998. Remedial Measures Against Soil Liquefaction – From Investigation and Design to Implementation. A.A. Balkema, Rotterdam. Yi, J.T., Goh, S.T., Lee, F.H., 2013. Effect of sand compaction pile installation on strength of soft clay. Geotechnique 63 (12), 1029–1041.
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
Available online at www.sciencedirect.com
ScienceDirect Soils and Foundations xxx (2017) xxx–xxx www.elsevier.com/locate/sandf
Performance of laterally loaded piles in improved coal ash deposit Jiunn-Shyang Chiou a,⇑, Teng-Ruei You b, Cheng-Chang Tsai a, Jin-Hung Hwang b b
a Department of Civil Engineering, National Taiwan University, No. 1, Sec. 4, Roosevelt Rd., Taipei 10617, Taiwan, ROC Department of Civil Engineering, National Central University, 300, Jhongda Rd., Jhongli City, Taoyuan County 32001, Taiwan, ROC
Received 19 May 2016; received in revised form 3 May 2017; accepted 1 July 2017
Abstract A series of pile load tests was conducted at a fossil fuel power plant in Taiwan to evaluate the performance of piles in a coal ash deposit that were improved using gravel compaction. The data from lateral load tests on two large-sized reinforced concrete piles were analyzed in this study. The original penetration resistance of the coal ash layer was low, but it increased after the ground improvement. Based on the load-defection curves of the test piles, the complete relationships of equivalent subgrade reaction coefficient kh with lateral displacement were derived and compared with those suggested by the Architectural Institute of Japan (AIJ) and the Japan Road Association (JRA). A composite SPT-N value was adopted to include the contribution of the gravel piles. It was found that the JRA equation using the composite SPT-N value fitted well with the relationships. Moreover, the experimental pile response and p-y curves were deduced using the inclinometer slope data. The relationships of the secant kh of the p-y curves vs. the normalized lateral displacement were further compiled and compared with those predicted by the JRA method. It was shown that the predicted curves using the composite SPT-N values also agreed well with the experimental relationships. Through a comparison of the profiles of the soil reaction with some existing lateral limiting soil pressure formulas, it was found that Prasad and Chari’s method was able to yield predictions close to the ultimate soil reaction. Ó 2017 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society. This is an open access article under the CC BYNC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Keywords: Coal ash; Ground improvement; Piles; Lateral loads; Field tests
1. Introduction Fossil fuel is a major source of power in Taiwan. A great deal of coal ash is generated during the combustion of coal; and therefore, the treatment of coal ash is an important issue for fossil fuel power plants. In Taiwan, coal ash is often added to Portland concrete to improve its engineering properties. It is also hydraulically filled into ponds near power plants to create land for constructing generator factories or coal storage warehouses. A hydraulically filled coal ash deposit is normally loose; and therefore, it has to be improved in order to increase its
Peer review under responsibility of The Japanese Geotechnical Society. ⇑ Corresponding author. E-mail address: [email protected] (J.-S. Chiou).
strength before any construction is done on it. The sand/gravel compaction pile method is commonly used for ground improvement. It installs piles made of compacted sands or gravels into the soft ground. The method can increase the liquefaction resistance and the bearing capacity of loose sandy soils (Okamura et al., 2003; Hatanaka et al., 2008) or accelerate the consolidation of soft clayey soils and the drainage of sandy soils (Rao et al., 1997; Yi et al., 2013). Sand/gravel compaction piling has also been introduced in Taiwan and is becoming increasingly popular for improving coal ash deposits (Hwang and Tu, 2002; Lin et al. 2009). The sand compaction pile method was shown to be effective for hydraulically filled coal ash ponds to increase SPT-N and CPT-qc values. In addition, another method, the heavy compaction method, was recently applied to a coal ash ground and found to be effective for improving the ground (Kokusho et al., 2012).
https://doi.org/10.1016/j.sandf.2017.08.019 0038-0806/Ó 2017 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
2
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
Piles are often used as foundations of coal storage warehouses in power plants to support the vertical structure loads. In addition to vertical loading, the piles may also be subjected to lateral loading, such as wind and earthquake loads on the structures. In engineering design practice, the p-y method is commonly used to analyze the nonlinear behavior of laterally loaded piles; it employs sets of nonlinear relationships between the soil reaction (p) and the pile displacement (y) to define the subgrade horizontal stiffness for different depths of soil along the piles. Using the data from pile load tests, a number of p-y models for sands (Reese et al., 1974; O’Neill and Murchinson, 1983) and clays (Matlock, 1970; Reese and Welch, 1975) have been proposed. However, few studies have been done on the lateral behavior of pile foundations in improved coal ash deposits; and therefore, more research is needed on the rational design of lateral piles in this type of ground. In 2013, the construction of ten coal storage silos was planned in an area filled with coal ash near a fossil fuel power plant in Linko, Taiwan. The area was to be improved by gravel compaction piles. Each silo was designed to have an inner diameter of 46 m and a height of 77.3 m, with the capacity to store about 7000 MN of coal. Piles were adopted to be the foundations of the silo structures. The piles were reinforced concrete piles with a diameter of 2 m and a length of 26.5 m. In order to understand the actual performance of the piles in a coal ash area improved by gravel compaction piles, a series of pile loading tests, including a compression test, a tension test, and two lateral tests, were conducted before the above construction.
For a better understanding of the lateral behavior of piles in an improved ash ground, this study analyzed the data from the lateral load tests and retrieved the experimental load-transfer curves (i.e., p-y curves). From the results, the characteristics of subgrade reaction coefficient kh and the p-y curves of the test piles, along with the applicability of general kh formulas in the design codes to improved coal ash grounds, were investigated. 2. Site conditions The soils at the coal silo construction site are composed of coal ash fill from EL. 9.0 m to EL. 9.0 m, and weathered sandstone below EL. 9.0 m. On the west side of the construction site, a region of 32 m 22 m in plan, which had been improved by gravel compaction piles, was chosen for performing the pile load tests, as shown in Fig. 1. The ground level at and around this region was lowered to EL. 4.5 m and the gravel piles were installed from this elevation. The water level was located at about EL. 1.0–1.4 m. As shown in Fig. 1, 16 boreholes (eight outside and eight inside the improved area) were adopted for the SPT and CPT subsoil investigation. Boreholes T5-T8 were located outside of the improved area; the soil properties in this region without the ground improvement are shown in Table 1. The penetration resistance of the coal ash fill was quite low; the average SPT-N value was 5 and the CPT-qc values were about 2500–3500 kN/m2. The gravel piles had a diameter of 1.0 m, a length of 11.3 m (the upmost 1.3 m was null), and a spacing of 3 m (three times the pile diameter). The backfill material for the piles was
N
6m
SPT-T5 CPT-T5
Construction area of coal silos Testing area
SPT-T1 CPT-T1
A1
A2
A3
A4
A5
B1
B2
B3
B4
B5
P1
5m
P6
SPT-T2 CPT-T2
5m
P4
P7
P2
5m
SPT-T8 CPT-T8
P5
5m
SPT-T3 CPT-T3
SPT-T6 CPT-T6
SPT-T4 CPT-T4
Pile cap
: gravel piles for ground improvement (pile diameter=1m) P3
3m
P8
6m
Coal silo (inner diameter= 46m)
: test piles and anchor piles (pile diameter=2m)
3m 3m
6m
5m
5m SPT-T7 CPT-T7
6m
Compression test: P4 (test pile); P1, P2, P6, P7 (anchor piles) Tension test: P5 (test pile); P2, P8 (anchor piles) Lateral test: P3 (test pile); P2, P8 (anchor piles) : SPT and CPT boreholes outside the improved area (T5-T8) : SPT and CPT boreholes in the improved area (T1-T4)
Fig. 1. Construction site of coal silos and test site of pile loading tests. Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
3
Table 1 Soil properties of region without ground improvement. Zone
Elevation (m)
Soil type
SPT-N
ct (kN/m3)
wn (%)
e
Gs
CPT-qc (kN/m2)
I II
4.5 to 9.0 <9
Coal ash Rock
2–9 (5) >50
8.34–16.97 (13.73) 20.70
32.1–98.4 (60.1) –
0.80–2.12 (1.59) –
2.14–2.63 (2.29) –
2500–3500 –
Note: bracketed values are the average values.
SPT-N
100 0
5
10
15
20
4
80
T3
70 2
60
T4 T6
50 0
40 30
Elevation (m)
Percent passing by weight
90
20 10 0 100
10
T7
-2
-4
1
Grain diameter (mm) -6
Fig. 2. Grain-size distribution for backfill material of gravel piles. -8
crushed granite. The grain-size distribution of the backfill material was controlled by suitability number SN, which was proposed by Brown (1977) and is defined as follows: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3 1 1 S N ¼ 1:7 þ þ ð1Þ 2 2 ðD50 Þ ðD20 Þ ðD10 Þ2 where D50, D20, and D10 are the diameters (in mm) through which 50%, 20%, and 10%, respectively, of the material is passing. The SN value in a range of 0–10 is an excellently rated backfill. Fig. 2 displays the grain-size distribution of the gravel pile material, in which D50, D20, and D10 are 18 mm, 14 mm, and 12.5 mm, respectively, which correspond to SN = 0.245. From triaxial compression tests on the samples taken from the gravel piles, the deformation modulus of the gravel piles was about 28.6 MN/m2 under an effective confining pressure of 50 kN/m2; the internal friction angle of the gravel piles was 38.7°. The area replacement ratio of the gravel piling was not high, only about 10%. Boreholes T1-T4 were located inside the improved area; the soil properties in this region after the ground improvement are shown in Table 2. The average increase in the total unit weight of the soil was from
-10
Fig. 3. SPT-N profiles near area of lateral pile load testing before and after improvement.
13.73 kN/m3 to 16.48 kN/m3. The average SPT-N value increased from 5 to 12 by a factor of 2.4. And the CPTqc values increased to about 7500–8500 kN/m2 by a factor of about 3. Fig. 3 further compares the SPT-N profiles with depth before the ground improvement (T6 and T7) with those after the ground improvement (T3 and T4) around the area of the lateral pile load tests. It was found that the effect of the improvement in the SPT-N values was more significant in the deeper soil than in the shallower soil. This result could be expected since ground improvement by gravel piles is generally less effective for shallow soil due to lower confining pressure. The profile after improvement could be roughly divided into two layers: the average SPT-N values of the upper layer (EL. 2 m to 4 m) increased from 4 to 10, while the average SPT-N values of the lower layer (EL. 4 m to 8 m) increased from 6 to 17.5.
Table 2 Soil properties of region after ground improvement. Zone
Elevation (m)
Soil type
SPT-N
ct (kN/m3)
wn (%)
e
Gs
CPT-qc (kN/m2)
I II
2.7 to 9.0 <9
Coal ash Rock
6–19 (12) >50
14.81–17.76 (16.48) 20.70
30.5–56.7 (46.1) –
0.64–1.47 (1.19) –
2.14–2.42 (2.28) –
7500–8500 –
Note: bracketed values are the average values. Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
4
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
3. Lateral load testing The ground level of a region of 12 m 22 m in the improved area was further lowered from EL. 4.5 m to EL. 2.7 m. In this low area, eight reinforced concrete piles were cast in place by the full casing method, and their positions are shown in Fig. 1. P3, P4, and P5 were the test piles for the lateral test, the compression test, and the tension test, respectively. Except for P3, P4, and P5, the piles were anchor piles. After the tension test, the tension load was released and then horizontal loading was applied to P5 for an additional lateral load test. The test piles had a length of 26.5 m and a diameter of 2.0 m. P3 and P5 had the structural details shown in Table 3. The pile length was divided into three regions: L1 for the top region of 6.5 m, L2 for the middle region of 8 m, and L3 for the rest of the region. The anchor piles had a length of 30.5 m and a diameter of 2.0 m. The structural details are shown in Table 4, in which Region ANL1 was for the upper region of 6.5 m and Region ANL2 was for the rest of the region of 24 m. The concrete strength was 30.4 MPa and the steel strength was 412 MPa. According to the material properties of the pile sections, the plastic moments were estimated
through section analyses, as shown in Table 3. Furthermore, assuming that the soil was cohesionless and considering two possible pile failure modes, the ultimate lateral capacities were estimated to be about 44,000 kN for the rigid body rotation due to soil failure and 6631 kN for the plastic hinge formation. It was found that the failure might be governed by the formation of plastic hinging. Fig. 4(a) displays the test setup of P3. P3 was pushed by an actuator supported by two anchor piles (P2 and P8) via a reaction beam. In P3, a load cell and LVDTs (linear variable displacement transducers) were set to measure the applied load and the pile-head displacement, respectively. An inclinometer and rebar stress meters were installed in the pile to measure the profiles of the displacement and the bending stress of the pile with depth. In addition, LVDTs were also set on the anchor piles to simultaneously record their pile-head displacements. The maximum lateral load in the test was 3630 kN. This load was far below the estimated ultimate lateral capacity, implying that plastic hinging had not occurred. During the testing, a gap of about 6–7 cm formed between the back of the pile and the adjacent soil. The pile-head load-deflection curve of P3 is shown in Fig. 5, in which the pile-head displacement
Table 3 Structural details of P3 and P5. Region
Length (m)
Diameter (m)
Cover (m)
Longitudinal reinforcements
Transverse reinforcements
Calculated plastic moment (kN m)
L1 L2 L3
6.5 8 12
2
0.1 0.1 0.1
92-D36 46-D36 32-D36
[email protected] [email protected] [email protected]
27,105 15,505 10,980
Table 4 Structural details of anchor piles. Region
Length (m)
Diameter (m)
Cover (m)
Longitudinal reinforcements
Transverse reinforcements
ANL1 ANL2
6.5 24
2
0.1 0.1
92-D36 46-D36
[email protected] [email protected]
P2
P3
P3 P8
P5 P2
Steel pad
Steel pad
Excavated part Load cell Reaction Actuator beam Steel pad
Anchor piles
P3
Not excavated EL. 2.70 m
EL. 2. 70m Load cell
EL. 2.20m
EL. 1.70 m Reaction Actuator beam Steel pad
Anchor piles
(a) P3
P5
(b) P5 Fig. 4. Test setups of P3 and P5.
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
namely, about 1.2 times that of P3. The larger resistance of P5 might come from the contribution of the unexcavated soil around the pile head.
5000 4500
Lateral load (kN)
4000 3500
4. Preliminary analysis for subgrade reaction coefficient
3000 2500
For a semi-infinitely long pile in Winkler-type soil with constant horizontal subgrade reaction coefficient kh under pile-head lateral loading, the governing equation of the pile response can be expressed as
2000 1500
P3 Anchor piles (P3) P5
1000 500 0 0
10
20
30
40
50
60
70
80
Lateral displacement (mm)
Fig. 5. Load-deflection curves of P3, anchor piles (P3) and P5.
4000 3500 3000
Lateral load (kN)
5
d 4y ¼ p ¼ k h Dy dx4
2000 1500
1000 500 0
0.002
0.004
0.006
0.008
0.01
Lateral rotation (rad)
Fig. 6. Load-rotation curve of P3.
at the maximum load was 68.4 mm (about 3.5% pile diameter). The pile-head load-rotation curve is shown in Fig. 6, in which the pile-head rotation at the maximum load was 0.0087 rad. Fig. 5 also shows the average load-deflection curve of the anchor piles, in which the load represents one-half of the total load applied and the displacement is the average displacement of both anchor piles. The anchor piles showed larger lateral resistance because the soil in front of the upper parts of the piles above the position of the applied loads had not been removed (Fig. 4(a)). Fig. 4(b) displays the test setup of P5, which was similar to that of P3. The main difference in the setups of P5 and P3 is that the soil in front of P5 above the position of the actuator was not excavated. This part of the soil provided some resistance to the lateral loading. In P5, a load cell and LVDTs were set to measure the applied load and the pilehead displacement, respectively. An inclinometer was installed in the pile to measure the profiles of the pile displacement with depth. Fig. 5 also shows the pile-head load-deflection curve of P5. The maximum load was about 4415 kN under which the pile-head lateral displacement reached 73 mm (about 3.7% pile diameter). The maximum load had not reached the estimated ultimate lateral capacity; and therefore, the plastic hinging had not occurred in the pile. Compared to P3, P5 had larger lateral resistance,
ð2Þ
in which EI is the pile flexural rigidity, p is the soil reaction, D is the pile diameter, and kh is the horizontal subgrade reaction coefficient.The closed-form solution of the above equation has been derived (Hetenyi, 1946). Letting qffiffiffiffiffiffi hD b ¼ 4 k4EI , the deflection along pile y can be expressed as y¼
2500
0
EI
H ebx cos bx 2EIb3
ð3Þ
in which b is referred to as the characteristic coefficient of the pile-soil system. Due to the analytical form of the solution, the solution is extensively adopted in Japanese design codes (e.g., Architectural Institute of Japan, 2001; Japan Road Association, 2012). In these design codes, the subgrade reaction coefficient is usually calibrated with the SPT-N value. Two approaches for determining the subgrade reaction coefficient are presented below. (1) Architectural Institute of Japan (AIJ, 2001) The Architectural Institute of Japan suggests 0:5 y k h ðyÞ ¼ k hr yr
ð4Þ
where k h is the subgrade reaction coefficient (kN/m3), y is the lateral displacement (m), and k hr is the subgrade reaction coefficient (kN/m3) at yr (yr = 0.01 m), which is expressed as k hr ¼ 2:53E0 D3=4
ð5Þ
in which D is the pile diameter (m) and E0 is the soil deformation modulus, which can be determined by unconfined or triaxial compression testing or borehole deformation testing, or it can be estimated based on an empirical formula, namely, E0 = 700 N (kN/m2) (N is the SPT-N value). (2) Japan Road Association (JRA, 2012) The Japan Road Association suggests 0:5 y k h ðyÞ ¼ k hr ð6Þ yr where k h is the subgrade reaction coefficient (kN/m3), y is the lateral displacement (m), k hr is the subgrade reaction
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
6
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
coefficient (kN/m3) at yr (yr = 0.01 D (0.05 m), and D is the pile diameter in meters), which is expressed as 3=4 BH ð7Þ k hr ¼ k h0 0:3 k h0 ¼
1 aE0 0:3
ð8Þ
in which BH is the equivalent foundation width (m), E0 is the soil deformation modulus, and a is the modification factor associated with method used to determine E0. E0 can be determined by plate load testing, unconfined or triaxial compression testing, or borehole deformation testing, or it can be estimated using an empirical formula E0 = 2800 N (kN/m2). When E0 is determined using unconfined or triaxial compression testing or borehole deformation testing, a = 4, and when E0 is determined by plate load testing or estimated by the SPT-N value, a = 1. Equivalent foundation width BH is expressed as follows: pffiffiffiffiffiffiffiffiffi BH ¼ D=b ð9Þ where, as mentioned previously, b is the characteristic coefficient of the foundation (m1). Based on the data on the lateral load and the pile-head displacement of each test, Eq. (3) for x = 0 (i.e., the pilehead displacement) was applied to back-figure the equivalent subgrade reaction coefficients with lateral displacement. Fig. 7 shows comparisons of the trends in the equivalent subgrade reaction coefficients with displacement for both tests. The equivalent subgrade reaction coefficients at a small displacement were from the anchor piles of the P3 test, while the equivalent subgrade reaction coefficients at a large displacement were from the test piles. The trends were consistent and the subgrade reaction coefficients decreased with displacement. At the same displacement level, the subgrade reaction coefficient from the P5 test was slightly larger than that from the P3 test. The difference might be due to the unexcavated soil around the pile head of P5. Fig. 7 further compares the equivalent subgrade reaction coefficient data with the AIJ equation (Eq. (4)) and
Subgrade reaction coefficient (kN/m3)
180000 P3 test (P3 pile) P3 test (anchor piles) P5 test (P5 pile) AIJ (N=12) JRA (N=12) JRA (Ncomp=14.9)
160000 140000
120000 100000 80000
60000 40000 20000 0 0
10
20
30
40
50
60
70
80
Displacement (mm)
Fig. 7. Comparisons of experimental horizontal subgrade reaction coefficients with AIJ and JRA equations.
the JRA equation (Eq. (6)). In the figure, it can be seen that when the average field improved SPT-N value (N = 12) was used, the computed kh using the AIJ method was significantly smaller than the equivalent kh, while the computed kh using the JRA method was slightly smaller than the equivalent kh. This comparison also shows that the kh from the AIJ method is much smaller than that from the JRA method. The main reason for the above difference is that the E0 estimated by the JRA method is four times that estimated by the AIJ method. Since the field-improved SPT-N values were measured among the gravel piles, they could only represent the improved properties of the coal ash, although such an improvement is normally evaluated based on the SPT-N values among the gravel piles. Tamura et al. (2012) conducted cyclic lateral loading centrifuge tests, and the results showed that the lateral resistance of a new structural pile, located in a place surrounded by existing structural piles (spacing = 3.5 D), increased slightly. Based on their study, and considering the possible influence of gravel piles, the following simple equation was used to estimate a weighted SPT-N value Ncomp to represent the whole composite improved ground (Japanese Geotechnical Society, 1998): N comp ¼ ð1 As ÞN 1 þ As N p
ð10Þ
where N1 is the SPT-N value among the gravel piles, Np is the SPT-N value at the center of the gravel pile, and As is the replacement ratio. Due to the lack of an SPT-N value at the center of the gravel pile, Np = 41 was estimated based on E0 = 28.57 MN/m2 (from the triaxial compression test) multiplied by the modification factor of a = 4. For As = 10% and N1 = 12, Ncomp = 14.9 was computed based on Eq. (10). Using N = Ncomp (=14.9) instead of 12, the AIJ method still underestimated the equivalent kh; however, the JRA method gave closer kh, as shown in Fig. 7. 5. Pile response analysis and p-y curves As mentioned before, for P3 and P5, the inclinometers were installed to measure the slope profiles of the piles with depth during lateral loading. With the inclinometer slope data, the regressive method proposed by Chiou et al. (2008) was applied to analyze the pile response based on the beam theory. The method uses a composite moment function for regression with the inclinometer slope data. In the analysis, the flexural rigidity of the pile section is required to link the curvature and the moment. Therefore, this study performed section response analyses to construct the nonlinear moment-curvature relationships of the pile sections. The simplified curves for regions L1 and L2 are shown in Fig. 8, on which four points are marked, namely, the cracking moment, the moment at the initial yielding of steel, the effective yielding moment, and the ultimate moment. According to the moments at the different depths calculated from the moment function, the corresponding reduced flexural rigidity of the pile sections would be chan-
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
7
1200
30000
25000
1000
20000
800
p (kN/m)
Moment (kN-m)
(a)
15000
10000
z=0.5m z=1.0m z=2.0m z=3.0m z=4.0m z=5.0m z=6.0m z=7.0m
600
400 L1-simplified
5000
200
L2-simplified
0 0
0.002
0.004
0.006
0.008
0.01
0.012
0
0.014
0
Curvature (1/m)
0.01
0.02
0.03
0.04
0.05
1200
Fig. 8. Moment-curvature curves for pile sections in L1 and L2 regions.
(b)
z=0.5 m z=1.0 m z=2.0m z=3.0m z=4.0m z=5.0m z=6.0m z=7.0m
1000
ged based on Fig. 8, depending on the occurrences of cracks or the yielding of the pile sections. By using the above-mentioned regressive method, the slope profiles and the corresponding regressive moment functions of P3 were obtained as shown in Fig. 9(a) and (b), respectively. Based on the moment functions and the sectional moment-curvature curves of the test pile, it was found that test pile P3 cracked very early, prior to the lateral load of 1962 kN; however, it had not yielded at the maximum load. Fig. 9(b) also compares the regressive moment functions with the moments back-computed based on the rebar stresses measured from the rebar stress meters, of which the trends were similar. Integrating the slope profiles and differentiating the moment functions gave the displacement and soil reaction profiles with depth, as shown in Fig. 9(c) and (d), respectively. By collecting y and p at specific depths for different load levels, the experimental p-y curves could be deduced. Fig. 10(a) shows the p-y curves at depths from 0 to 7 m; the curves were softer for the depths from 0 to 1 m, the stiffness and strength increased with depth, and the soil appeared to reach the ultimate state and have a slight softening response at a larger displacement. For the depths from 2 to 5 m, the trends in the curves were close; for the depths from 6 to 7 m, the (a) -0.01
Depth (m)
-0.015
(b)
Slope (m/m) -0.005
0
0.06
y (m)
0
5000
p (kN/m)
600
400
200
0 0
10000
0.01
0.02
0.03
0.04
0.05
0.06
0.07
y (m)
Fig. 10. Experimental p-y curves: (a) P3 and (b) P5.
curves showed linear trends. For the depths larger than 7 m, because of the influence of the rock stratum, the displacement response was very small and even had reverse displacements in the rock stratum, such that the trends in the associated p-y relationships became random and irregular. As a whole, the subgrade stiffness around this area was large, about 1350–1500 MN/m3, and the corresponding displacement range was about 0.0003 m. The same analysis approach was applied to the P5 test. Fig. 11 displays the profiles of the slope, the moment, the (c)
Moment (kNm)
0.005 -5000
800
15000 -6
(d)
y (cm) -4
-2
0
2
-1500
Soil reaction (kN/m) -1000
-500
0
0
0
0
0
2
2
2
2
4
4
4
4
6
6
6
6
8
8
8
8
10
10
10
10
12
12
12
12
500
1000
981kN 1962kN 2943kN 3630kN 981kN-exp
981kN
1962kN 14
14
2943kN 3630kN
1962kN-exp
981kN (rebar stress meter)
2943kN-exp
1962kN (rebar stress meter)
16
16
2940 kN (rebar stress meter)
3630kN-exp
981KN 1962kN
14
981kN
14
1962kN 2943kN
16 3630kN
2943kN
16
3630kN
3630kN (rebar stress meter) 18
18
18
18
Fig. 9. Response profiles of P3: (a) slope, (b) moment, (c) displacement, and (d) soil reaction. Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
8
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx (a)
Depth (m)
-0.012
(b)
Slope (m/m) -0.007
-0.002
981kN 1962kN 2943kN 3433kN 3924kN 4415kN 981kN-exp 1962kN-exp 2943kN-exp 3433kN-exp 3924kN-exp 4415kN-exp
(c)
Moment (kNm)
0.003 -5000
0
5000
10000
15000 -8
(d)
y (cm) -6
-4
-2
0
2
-1500
Soil reaction (kN/m) -1000
-500
0
0
0
0
0
2
2
2
2
4
4
4
4
6
6
6
6
8
8
8
8
10
10
10
10
12
12
12
500
12 981kN
14
16
18
(a)
981kN 1962kN 2943kN 3433kN 3924kN 4415kN
14
16
981kN 1962kN 2943kN 3630kN 3924kN 4415kN
1962kN
14
14
2943kN 3433kN
16
16
3924kN 4415kN
18
18
(b)
18
(c)
(d)
Fig. 11. Response profiles of P5: (a) slope, (b) moment, (c) displacement, and (d) soil reaction.
displacement, and the soil reaction. Fig. 10(b) shows the experimental p-y curves of P5 for depths from 0 m to 7 m. Compared to Fig. 10(a), it is seen that the trends in the p-y responses for P3 and P5 are similar.
(1) Secant subgrade reaction coefficient of p-y curves With the obtained p-y curves, this section further analyzes the secant slope of the curves which is called the secant subgrade reaction coefficient. Simplifying the soil layer as a homogeneous layer, a displacement of 1% pile diameter (0.01 D) is used as the normalization factor to build the relationship of the secant subgrade reaction coefficient (p/(D y)) and the normalized lateral displacement (y/0.01 D), as shown in Fig. 12. Compared to the JRA equation using Ncomp = 14.9, it was found that the trend agreed averagely well with the JRA equation. As mentioned before, the soil after improvement was not uniform with depth and could be roughly divided into two layers. Therefore, the soil was considered as two layers for P3 in order to build the corresponding trends in the secant subgrade reaction coefficient vs. normalized displacement. Consistent with the soil profile, the kh in the lower layer was larger than that in the upper layer. Furthermore, the JRA method was applied to predict the relationships using the average SPT-N values for the two layers. It can be seen that the predicted curves are in good agreement with the experimental curves, as shown in Fig. 13. The JRA method seems to be applicable for simulating the respective relationships of the subgrade reaction coefficients when composite SPT-N values of different depths are adopted. (2) Lateral limiting soil pressure There are several formulas proposed for lateral limiting soil pressure pu. From the soil reaction profiles of P3 and
P3 P5
kh_sec (kN/m3)
200000
JRA equation (Ncomp=14.9)
150000
100000
50000
0 0
0.5
1
1.5
2
y/0.01D
Fig. 12. Secant subgrade reaction coefficient - normalized displacement curve.
250000 P3 (z=0-4m) JRA equation (Ncomp=13.1 for Nave=10)
200000
P3 (z=4-7m)
kh_sec (kN/m3)
6. Discussion
250000
JRA equation (Ncomp=19.4 for Nave=17.5) 150000
100000
50000
0 0
0.5
1
1.5
2
y/0.01D
Fig. 13. Secant subgrade reaction coefficient - normalized displacement curve for different soil layers.
P5, it was observed that the soil reactions at the shallow depth of soil seemed to have a limit when lateral loading is increased. Therefore, with the obtained soil reaction profiles, the applicability of the existing formulas for the
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
(a) -1500
(b)
Soil reaction (kN/m) -1000
-500
0
500
1000
-1500
1200
Soil reaction (kN/m) -1000
-500
0
0
0
1
1
2
2
3
3
4
4
9
500
5
Depth (m)
1962kN 2943kN
8
3630kN Brom Brinch Hansen
9
Prasad and Chari
6
1962kN 2943kN 3433kN 3924kN 4415kN Brom Brinch Hansen Prasad and Chari
10
600 z=0.5m z=1.0m z=2.0m z=3.0m z=4.0m z=5.0m z=6.0m z=7.0m
400
5
6
7
800
p (kN/m)
Depth (m)
1000
200
7
0
8
0
0.01
0.02
0.03
0.04
0.05
0.06
y (m)
9
Fig. 15. Derived p-y curves for use in numerical simulation.
10
Fig. 14. Comparisons of lateral limiting pressure formulas: (a) P3 and (b) P5.
5000 4500
4000
Lateral load (kN)
limiting pressure were investigated. In this study, three formulas for pu, proposed by Brinch Hasen (1961), Brom (1964), and Prasad and Chari (1999), were adopted for comparison. These formulas were basically derived from the theory of earth pressure to represent the lateral ultimate resistance provided by the surrounding soil at the fully mobilized passive stress state. These three methods for pu are briefly introduced below.
3500 3000 2500 2000
P3
1500
P5
1000
Analysis (P3) Analysis (P5)
500 0
(1) Brinch Hasen method (1961)
0
10
20
30
40
50
60
70
80
Lateral displacement (mm)
pu ¼ czK q
ð11-1Þ
Fig. 16. Comparisons of simulated and experimental load-displacement curves for P3 and P5.
(2) Brom method (1964) pu ¼ 3czK p
ð11-2Þ
(a)
(3) Prasad and Chari method (1999)
(b)
Moment (kNm)
-5000
0
5000
10000
15000
-1500
Soil Reaction (kN/m) -1000
-500
0
0
0
ð11-3Þ
2
1
where c is the unit weight of the soil, z is the depth, Kq is the Brinch Hasen earth pressure coefficient, Kp is the Rankine passive earth pressure (=tan2(45 + //2)), and / is the angle of internal friction of the soil. Fig. 14 plots the profiles of the soil reaction for P3 and P5, on which the limiting pressures by the three methods are shown. It can be seen that the limiting pressure estimated by the Prasad and Chari method could envelop the ultimate soil reactions at shallow depths of soil, but the other two methods significantly underestimated the limiting pressure.
4
2
3 6
4
Depth (m)
Depth (m)
pu ¼ 0:8cz10ð1:3 tan /þ0:3Þ
500
8
5
10
6 12
981kN (rebar stress meter) 1962kN (rebar stress meter) 2940 kN (rebar stress meter)
14
3630kN (rebar stress meter)
981 kN (analysis) 16
1962 kN (analysis) 2943 kN (analysis) 3630 kN (analysis)
18
7 981kN 1962 kN 2943 kN 3630 kN Limiting pressure
8
9
10
Fig. 17. Simulated responses of P3: (a) moment and (b) soil reaction.
7. Verification of JRA’s kh method and Prasad & Chari’s pu model In order to evaluate the applicability of the JRA’s kh method and Prasad and Chari’s pu model, a numerical simulation was performed to analyze the P3 and P5 tests using the Winkler-beam model, in which the pile is
modeled by beam elements and the soil reactions are modeled by p-y curves. The secant slope of the p-y curves was determined by the JRA formula and the maximum soil reaction was limited by Eq. (11-3). Based on the profile of the average SPT-N value of boreholes T3 and T4 and the limiting pressure, the derived p-y curves varied with
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019
10
J.-S. Chiou et al. / Soils and Foundations xxx (2017) xxx–xxx
depth, as shown in Fig. 15. Fig. 16 compares the numerical load-deflection curves of P3 and P5 with the experimental ones. It can be seen that except for some underestimation of the displacement at the largest load, the overall simulated curves were in good agreement with the test data. Fig. 17(a) further compares the simulated moment profiles of P3 with those observed in the test. It can be seen that the simulated moments were slightly larger than the observed ones, but that their overall trends were consistent. In addition, Fig. 17(b) compares the simulated soil reaction profiles with the set limiting pressure. It was found that the soil shallower than about 2.5 m had reached the limiting value. 8. Conclusions From this study, the following conclusions have been deduced: 1. The SPT-N values were increased by a factor of 2.4 and the CPT-qc values were increased by a factor of about 3 for the coal ash ground improved by gravel compaction piles with an area replacement ratio of about 10%. The effect of the improvement was larger in the deeper soil than in the shallower soil. 2. The piles in the improved coal ash ground had large lateral resistance. At the lateral displacement of 1% of the pile diameter (20 mm), the lateral load was about 2200 kN. The soil right in front of the upper parts of the piles above the applied load increased the lateral resistance by about 20%. 3. Based on the load-defection curves of the piles, the trends in kh with lateral displacement could be well constructed. Using the field-average improved SPT-N value, the AIJ method significantly underestimated kh, while the JRA method slightly underestimated kh. However, when a larger composite SPT-N value was used, considering the possible contribution of the stiffness of the gravel piles, the predicted trends in kh using the JRA method were improved. 4. The experimental p-y curves could be expressed as relationships of the secant subgrade reaction coefficients with lateral displacement. The JRA equation, using composite SPT-N values, fitted well with the relationships. Moreover, the limiting pressure proposed by Prasad and Chari was able to envelop the ultimate soil reactions at shallow depths of the soil. 5. The numerical simulation using the p-y curves whose secant stiffness and ultimate pressure were determined by the JRA method and Prasad and Chari’s model, respectively, yielded predictions that corresponded well with the experimental results.
Acknowledgements The authors would like to thank the Taiwan Power Company, T.Y. LIN International Taiwan Consulting Engineers, Inc., and Kung Sing Engineering Corporation for providing the test data. References Architectural Institute of Japan (AIJ), 2001. Recommendation for the Design of Building Foundations (in Japanese). Brinch Hansen, J., 1961. The ultimate resistance of rigid piles against transversal forces. Danish Geotechnical Institute, Copenhagen, Bulletin No. 12, 5–9. Brom, B.B., 1964. Lateral resistance of piles in cohesionless soils. J. S MFE Divis., ASCE 90, 123–156. Brown, R.E., 1977. Vibrofloation compaction of cohesionless soils. J. Geotech. Eng. Div., ASCE 103 (12), 1437–1451. Chiou, J.S., Chen, C.H., Chen, Y.C., 2008. Deducing pile responses and soil reactions from inclinometer data of a lateral load test. Soils Found. 48 (5), 609–620. Hatanaka, M., Feng, L., Matsumura, N., Yasu, H., 2008. A study on the engineering properties of sand improved by the sand compaction pile method. Soils Found. 48 (1), 73–85. Hetenyi, M., 1946. Beam on Elastic Foundations. University of Michigan Press, Ann Arbor, Mich. Hwang, J.H., Tu, T.Y., 2002. Ground vibration during gravel pile construction. J. Mar. Sci. Technol. 10 (1), 36–46. Japan Road Association (JRA), 2012. Design Specifications for Highway Bridges - IV: Substructures (in Japanese). Kokusho, T., Nakashima, S., Kubo, A., Ikeda, K., 2012. Soil investigation of fly ash deposit improved by heavy compaction method. J. Geotech. Geoenviron. Eng. 138 (6), 738–746. Lin, S.S., Chien, S.J., Liao, J.C., 2009. Ground improvement of a coal ash storage pond using sand compaction piles. Lowland Technol. Int. 10 (2), 21–29. Matlock, H., 1970. Correlations for design of laterally loaded piles in soft clay. In: Proc., 2nd Annual Offshore Technology Conference, Houston. Okamura, M., Ishihara, M., Oshita, T., 2003. Liquefaction resistance of sand deposit improved with sand compaction piles. Soils Found. 43 (5), 175–187. O’Neill, M.W., Murchinson, J.M., 1983. An Evaluation of P-Y Relationships in Sand. Rep. to American Petroleum Institute, Washington, D. C.. Prasad, Y.V.S.N., Chari, T.R., 1999. Lateral capacity of model rigid piles in cohesionless soils. Soils Found. 39 (2), 21–29. Rao, S.N., Reddy, K.M., Kumar, P.H., 1997. Studies on group of stone columns in soft clays. Geotech. Eng., Bangkok 28 (2), 165–182. Reese, L.C., Cox, W.R., and Koop, F.D. 1974. ‘‘Analysis of laterally loaded piles in sand. In: Proc., 6th Annual Offshore Technology Conference, Houston, pp. 672–690. Reese, L.C., Welch, R.C., 1975. Lateral loading of deep foundations in stiff clay. J. Geotech. Eng., ASCE 101 (7), 633–649. Tamura, S., Adachi, K., Sakamoto, T., Hida, T., 2012. Effects of existing piles on lateral resistance of new piles. Soils Found. 52 (3), 381–392. The Japanese Geotechnical Society, 1998. Remedial Measures Against Soil Liquefaction – From Investigation and Design to Implementation. A.A. Balkema, Rotterdam. Yi, J.T., Goh, S.T., Lee, F.H., 2013. Effect of sand compaction pile installation on strength of soft clay. Geotechnique 63 (12), 1029–1041.
Please cite this article in press as: Chiou, J.-S. et al., Performance of laterally loaded piles in improved coal ash deposit, Soils Found. (2017), https://doi. org/10.1016/j.sandf.2017.08.019