1 Introduction
In tall buildings, constructing basements increases the floor area, for parking purposes and for retaining the excavation. The most critical structural component of a tall building is its foundation and it’s very important to design for safety and economic reasons. The basement’s contribution to lateral loads is usually not taken into account during foundation design even though, in reality, it can resist lateral loads and increase the lateral stiffness of the foundation.
Many researchers have worked on the settlement behavior and capacity of the foundation of tall buildings to make them more economical, and safer, satisfying the desired lateral and vertical loading conditions. The common types of foundations for tall buildings are raft foundations, pile foundations and piled raft foundations. A raft foundation is used when isolated footings cover 50%–70% of the building area under a superstructure. Such a foundation supports a building of moderate height but for very tall buildings it is not feasible due to large lateral and bending moment loadings that the raft foundation cannot resist. So, for taller buildings, pile foundations are provided to resist lateral and bending moment loadings. Piles also reduce settlement by transferring the load from soft strata to hard strata [
1] but in most cases an increased number of piles result in an uneconomical foundation. Prabha [
2] carried out two case studies and concluded that settlement can be reduced by using an appropriate number of piles. The raft is also effective even though the piled raft foundation is designed based on the pile group concept [
3]. Many experimental and analytical studies have been carried out by researchers [
4–
8] to determine the response of piled raft foundations to vertical loading and have concluded that rafts contribute to resistance to vertical loads. After research studies, raft contribution to vertical load in foundation capacity is considered by the engineers but the raft’s contribution to resistance to lateral load is usually ignored. That’s why in the last two decades, many experimental and analytical studies have been carried out to determine the response of piled raft foundations to lateral loading [
3,
9]. Chanda et al. [
10] proposed simplified expressions for lateral capacity of piled raft based on regression analysis. Recently, Jamil et al. [
11] conducted an experimental study on a small-scale model piled raft foundation under lateral loading and concluded that the raft contributes to resisting the lateral load in a piled raft system by an extent that depends on the vertical contact pressure beneath it. Hamada et al. [
12] proposed a simplified analytical method and theoretical equations for pile rafts under lateral load and validated their results with an experimental approach. However, in tall buildings, the substructure components also include the basements and there is little research on the effect on or contribution of the basement wall to lateral loading response in the case of piled raft foundations.
Research by Chow and Poulos [
13] involved numerical study using PLAXIS 3D, a finite element analysis program to determine the contribution of basements to the overall stiffness of the piled raft foundation system. The researchers performed the numerical study by selecting the different heights of the basement wall in a 16 m ×16 m piled raft with 16 piles. They concluded that basement walls make a meaningful contribution to the piled raft stiffness and capacity under lateral load. Furthermore, they concluded that an increase in height of the basement wall will cause an increase in lateral stiffness of the piled raft.
Taiebat et al. [
14] and Amirzehni et al. [
15] performed a study on seismic design and seismic performance of deep basement walls but their study lacked consideration of the basement’s effect on the overall foundation capacity (i.e., maximum load that a foundation can sustain).
No experimental study has been done on the contribution of the basement wall to overall foundation capacity (i.e., maximum load that a foundation can sustain) involving comparison of piled raft and pile groups with the same pile configuration. So, this study relates the increase in stiffness of the foundation by considering basement walls and their response to lateral loading through experimental small-scale models. It is worth mentioning here that the purpose of this experimental study is not to simulate any prototype foundation or the response of a large foundation. For this study, a total of nine small-scale models were fabricated consisting of 3 piled raft models, 3 pile group models and 3 piled-raft with basement wall models with the same pile configuration.
2 Test setup
A total of nine tests were performed experimentally in a soil box of 0.9 m × 1.2 m × 1.5 m using small-scale models of pile group, piled raft and piled raft with basement wall. Three configurations of piles, i.e., 2 × 2, 2 × 3, and 3 × 3, were tested as a piled raft and pile group, in addition to tests for basement wall with piled raft foundation.
All the experimental works were performed at the laboratory of soil mechanics at the University of Engineering and Technology Peshawar and the UET Jalozai campus.
2.1 Soil testing
In this research study, sandy soil was used throughout the depth of the sandbox. Different laboratory tests were performed to determine the properties of soil used; i.e., specific gravity test for soils according to ASTM-D854, sieve analysis for particle gradation curve according to ASTM–D422, D2487 shown in Fig.1, direct shear test for the angle of internal friction ASTM–D3080 and also classification of the soil according to USCS. The summary of the soil properties is shown in Tab.1.
2.2 Model
The piles were fabricated from galvanized iron (hollow pipes) having internal and external diameters 16.70 mm and 19.05 mm respectively with a length of 0.45 m. Specifically, hollow piles were selected because they are more sensitive, in terms of deformation, to the strains developed at very small loads.
The raft and basement wall were fabricated from aluminum and welded together as shown in Fig.2. The raft had a dimension of 0.3 m × 0.3 m and a thickness of 25.4 mm, while the basement wall had a 152.4 mm height and 10.2 mm thickness. Dimensions of the basement wall were so selected to minimize the bending effect induced due to lateral load and to handle it easily during the test. Spacing between piles was 127 mm for all models. The material properties of raft, piles and basement wall are shown in Tab.2.
2.3 Soil box
The soil box of 1.2 m in length, 0.9 m in width and 1.5 m in depth, made of steel, was used for the model test as shown in Fig.3. The soil box was braced in the lateral and transverse directions to make it rigid and strong enough. Dimensions of the soil box were so selected to reduce the influence of boundary conditions as defined by Narasimha et al. [
16], as shown in Fig.4. Narasimha stated that the soil box dimension must be extended up to 8
d–12
d (where
d is pile’s diameter) in parallel direction and 3
d–4
d in a perpendicular direction to applied lateral load. A dimension of 6
d (see Fig.4) was used, to further reduce the boundary influence [
17].
2.4 Soil sample preparation
To compare the results of different model tests, sand relative density had to be the same for all tests. There are different methods to achieve the same relative density such as tamping, pluviation, etc. The Pluviation technique generally results in uniform sand density [
17]. A newly raining device was designed to consist of a hopper, sieves, shutter and a moveable steel frame. Unit weight of the sand depends on the falling height and pouring rate, i.e., shutter’s size [
18]. Pluviator was calibrated using four different shutters and for different falling heights. From the results, a 13 mm shutter was selected to obtain a relative density of 60% with a constant falling height of 0.45 m from the sand surface as shown in Fig.5. A pluviation method is a tedious job and so a 13 mm shutter was selected to reduce the pouring time. To verify 60% relative density, a small-scale dynamic cone penetrometer test was performed during the calibration process.
2.5 Other arrangements
In this study, a vertical load of 5350 N was applied in increments of 157 N through steel plates. A lateral load of 2634 N was applied with the help of a hydraulic load machine. Its capacity was 5 t with a loading rate of 4.9 N/s. Two different load indicators were used for measuring the applied load, i.e., vertical load cell and lateral load cell with a total capacity of 5 t. Before using for tests, both load cells were properly calibrated and calibration factors for vertical and lateral load cells were measured as 0.016 and 0.018894, respectively.
Using the phenomenon of shear load, a load cells (combination of strain gauges using Wheatstone bridge principle) as shown in Fig.6, were installed at the top of piles to measure the lateral load on the piles. Two strain gauges were installed on the tension side and two strain gauges on the compression side of each pile. After shear load cells installation, each pile was calibrated against known lateral load to find calibration factors for a 9.8 N load. A total of nine piles were calibrated and their calibration factors are shown in Tab.3.
For measuring lateral displacement, very precise transducers were used having a maximum capacity of 25 mm with a least count of 0.001 mm. For data acquisition, a data logger, with a frequency of 4.9 N/s, was used. This has 30 channels to which the strain indicators, vertical and lateral load cells and transducers (displacement indicators) were connected. The data logger was further connected to the laptop as indicated in Fig.7.
2.6 Test procedure
The soil box was filled with sand using the pluviation technique with a fall height of 0.45 m to achieve a relative density of 60%. To acquire a uniform sand surface the hopper was moved constantly in both directions with the help of a moveable steel frame. When a sand bed of 0.9 m height was prepared, the model basement wall with piled raft was placed exactly in the center of the soil box without fixing pile bolts. After that, air pluviation was resumed till the sand level reached a height equal to two-thirds of the pile length. The raft and basement wall were removed such that the pile’s location was not disturbed. This was done to obtain the same relative density below the raft. The remaining sand was poured through air pluviation until the sand reached the pile top. The raft and basement wall were re-installed above the piles, holding each pile in position with the long handle wrench. The purpose of the wrench was to prevent a change in direction of lateral strain gauges and hold the pile firmly while fixing the bolt on the raft top. The pile heads were fixed with the raft. The direction of the lateral load cells placed in the piles was kept the same as the lateral load in all model tests. The sand bed was leveled with a straight edge after removing excess sand. Thereafter, a vertical load of 5350 N was applied incrementally through steel plates. A lateral transducer was attached to the side of the wall to monitor lateral displacement. Lateral load cells (strain gauges arrangement) were connected to the data logger. A lateral load of 2634 N was applied through the hydraulic machine at a rate of 4.9 N/s. A lateral load was applied at the mid of the basement wall while it was applied at the raft top in the case of piled raft and pile group. A total lateral load was selected such that the whole apparatus did not collapse but remained stable. All data from the data logger was stored in a computer-based system (i.e., MATLAB coded file). The arrangement is shown in Fig.8. In the case of piled raft, the raft was in contact with the sand surface while in the pile group, a raft was about 50.8 mm above the soil surface.
2.7 Test summary
A total of 9 tests were performed on a small-scale model under lateral load. One model was tested for each pile configuration in a piled raft, pile group and piled raft with a basement wall.
The experiments consist of small-scale models of piles, raft and basement wall. The variation of piles along with spacing between them is shown in Fig.9. To study the effect of the number of piles on the lateral contribution of the basement wall, three different configurations were used. 3D sketches are shown in Fig.10.
3 Result and discussion
A total of 9 tests were performed on piled raft and basement wall models. Fig.11 shows the load−displacement curves for each pile configuration. From Fig.11(a), which is for 4-piles configuration, it can be observed that displacement was reduced significantly because of the basement wall. For displacement of 0.8 mm, the pile group resisted only 170 N and piled raft resisted about 750 N while the basement wall with piled raft resisted about 2000 N. The increase in load resistance was due to basement wall shear resistance and soil resistance in front of the basement wall. A decrease in displacement was a result of an increase in stiffness. Similarly, for the 6-pile and 9-pile configurations, shown in Fig.11(b) and Fig.11(c) respectively, the same response was observed. The load-resistant capacity of the basement wall with a piled raft was about 3–4 times that of the piled raft and 5–15 times that of the pile group.
By comparing the response of basement wall models with piled raft and pile group in Fig.11, it became clear that piled raft with basement wall models had a very stiff response compared to piled raft and pile group models, which was due to the raft and basement wall contribution to resisting lateral load. Similarly, the piled raft response was stiffer than that of pile group models due to the raft lateral load contribution.
One of the most important outcomes of this study relates to the distribution of load in front and back piles. Fig.12 and Fig.13 show lateral load that was resisted by front and back piles against displacement in 2 × 2, 3 × 2, and 3 × 3 pile configurations for piled raft, pile group and piled raft with basement wall model. In the piled raft, back piles were taking more load than front piles which is the same as results obtained by Refs. [
11,
12]. A plausible reason for this response is an increase in stiffness of soil in front of back piles due to vertical raft contribution. The basement wall model showed the same response as in case of the pile group, i.e., front piles were resisting more load than back piles. This was because of the increased stiffness in the vicinity of the front piles due to the basement wall and applied vertical load. By comparing the pile group and piled raft with basement wall models, it is clear that the back piles of the piled raft with the basement wall model were resisting more load than the back piles of the pile group. Initially, in the 9-pile configuration, the piled raft back piles were resisting less load than the back piles of the piled raft with basement wall model which is because of the lower raft vertical contribution in the case of the 9-piled-raft model.
Tab.4–Tab.6 shows the summary of the percentage contribution of each component (i.e., piles, raft and basement wall) in each pile’s configuration. It was observed that raft and basement wall contributed 80% to the resistance to applied lateral load and 20% of lateral load was transferred to piles in the 4-pile configuration as shown in Tab.4. In the case of the piled raft, 39% of the total applied lateral load was resisted by the raft while the remaining was transferred to piles. Similarly, the component contributions for 6-pile configuration and 9-pile configuration are shown in Tab.5–Tab.6, respectively. It was observed that the load transfer to piles increased as the number of piles increased due to pile-pile interaction and a decrease in raft contact with the soil. Furthermore, it was observed that in the piled raft, 71% of lateral load was transferred to piles while in a basement wall with a piled raft, only 36% of lateral load was transferred to piles in a 6-pile configuration. So, by incorporating basement resistance in foundation design we can reduce the pile’s dimensions which can make the design more economical as well as safer.
Tab.7–Tab.9 shows stiffness and stiffness increase in each pile configuration. The stiffness of an overall foundation system increased in the presence of a basement wall. The stiffness of each type of foundation increased with an increasing number of piles and the stiffness of the 9-pile configuration with the basement wall was the highest. Moreover, it must be noted that stiffness of the piled raft model was greater than that of pile group model due to raft contact with the soil surface. Stiffness increase between piled raft with basement wall and piled raft decreases as the number of piles increased.
4 Conclusions
A program of 9 experimental tests of 3 different pile configurations in a piled raft system with a basement wall embedded in a uniform soil layer was carried out and responses to lateral load were compared with those of pile group and piled raft models.
From the above results and discussion, the following conclusions can be outlined.
The lateral resistance capacity of the basement wall with a piled raft is about 3–4 times more than that of the piled raft and 5–15 times that of a pile group.
Basement wall contribution to resisting lateral load ranges from 50%–80%, so this can be considered a major contribution and should not be ignored in the design stage.
The lateral deflection of the piled raft with basement wall model was much less than those of the piled raft and pile group model which was because of the lateral contribution of the basement wall and piled raft. Similarly, the piled raft stiffness was more than the pile group stiffness which was because of the raft’s lateral contribution.
In the piled raft, back piles were taking more load than front piles, depending on the raft’s vertical contribution. This response was opposite to the conventional pile group response in which the front piles resist more load than the back piles. Basement wall foundations showed the same response as in case of pile group foundations, i.e., front piles were taking more load than back piles.
Load transfer to piles increased as the number of piles increased in both piled raft and basement wall with piled raft due to pile-pile interaction and decrease in raft contact with the soil.
Most tall buildings have 3–4 basement levels and this study suggests that design can be more economical in terms of limiting lateral deflection, limiting lateral demand on piles, and reducing the number of piles and their diameter.
PDF
(3586KB)
