Load transfer on single pile near new tunneling in sandy ground

Transcription

1 Load transfer on single pile near new tunneling in sandy ground Chung-Jung Lee, Kuo-Hui Chiang 2 Associate Professor, Department of Civil Engineering, National Central University, Chungli, TAIWAN 2 Graduate student, Department of Civil Engineering, National Central University, Chungli, TAIWAN ABSTRACTS: A series of centrifuge model tests were performed to assess the tunneling-induced ground deformations and their detrimental effects on the neighboring single pile in the saturated sandy ground. The relative elevations between the pile tip and the buried depth of new driven tunnel dominated the distributions of bending moment and axial force along the depths. A load transfer mechanism is proposed to provide constructors making an assessment and taking effective measures for preventing piles from bending failure and suffering excess foundation settlements.. INTRODUCTION Geotechnical centrifuge modeling has been frequently used to study the tunneling-induced problems. Lee et al. (999) studied tunneling-induced ground movements (surface and subsurface settlement troughs in clays). Bezuijen & Schrier (99) presented the influence of a bored tunnel on pile foundations. Similarly, Loganathan et al. (2) assessed tunnelinginduced ground deformation in clay and their adverse effects to adjacent pile foundations. Jacobsz et al. (2) also investigated the effects of tunneling on nearby single piles in dry sand. In the present study, a series of centrifuge model tests were performed to assess tunneling-induced ground deformation and their detrimental effects on the adjacent pile foundations in saturated sandy ground. The mechanisms of load transfer (bending moment and axial force) on a single pile with and without axial load anile head deformation during tail voids closure were discussed. Due to increasing the traffic congestion in urban areas, the transportation authority in Taiwan has been planning to construct the public mass rapid transit system to slow the growth of traffic flow down. Shield tunneling in the soft ground has become more and more widely used to subway construction for the benefit to reducing interference with the congested traffic already in streets. However, the subway authorities have also been forced to make use of a limiteublic area (i.e., beneath existing streets, in the vicinity of or even passing under existing buildings) to build tunnels because of the high density of development and difficulty on land expropriation in the crowed cities. When a tunnel is excavated, the ground around the tunnel inevitably moves towards the tunnel opening. The difference of linings and excavated boundaries is defined as tail voids. Consequently, a tunnel driven close to or beneath a building that founded on pile foundations in soft ground may cause damages of buildings. Lee et al. (99) reported a case that the tunnel was driven between the pile foundations supporting a seven-story building with a two-story basement in London. The structure is founded on piles taken through 28 m of London clay to end bear on the hard bed. They found the piles deformed in the same manner as the surrounding soils. However, the impact of tunneling on the pile foundations and how to evaluate building s safety were not discussed. The load transfer mechanism of a pile embedded in the moving soils resulted from tunnel construction is considerably complex and needs to study in detail. Several researchers developed numerical methods to evaluate the detrimental effects of tunneling-induced soil movements on adjacent existing piles (Loganathan et al. 2, and Chen et al. 999). Both finite element method and boundary element method were used to study pile response due to excavation-induced lateral soil movement by Poulos and Chen (997). They presented the design chart for estimating bending moments and deflections on pile. 2. CENTRIFUGE MODELING AND TEST PROCEDURES This experimental work was undertaken in the geotechnical centrifuge at the National Central University (NCU). The NCU geotechnical centrifuge, with a nominal radius of m, is capable of accelerating a -ton model package to g and. ton to 2 g. All the tests reported in the paper were conducted at an acceleration of g. The layout and dimensions of the models are described in model units; however, the test results are presented in terms of prototype units. 2. Model tunnel and instrumented model piles The model tunnel was a.-mm-thick cylindrical rubber bag on which a sheet of filament tape of. mm thickness was pasted to enhance tunnel stiffness in an attempt to ensure that the tunnel maintained a diameter of 6 mm during increasing the supporting pressure in the rubber bag, representing a 6-m diameter tunnel at prototype scale. Four deformation gages, made of four thin sliced strain-gauged cantilevers, were placed inside the rubber bag. The free end of each cantilever

2 was bonded to the appropriate positions on the inner surface of the bag, make it possible to measure the deformations of the crown, invert, and two sidewalls of the tunnel during the tests. Two instrumented model piles (axial pile and bending pile) were constructed using aluminum tubes of 9.6 mm outer diameter and 8.6 mm internal diameter (Figure. ). Embedment depth of model pile was mm. The model piles were designed to replicate.96 m diameter circular concrete pile at prototype scale. Strain gauges were placed externally at seven locations to measure lateral response on the bending pile and at six locations to measure axial response on the axial pile. Both the axial pile and the bending pile were calibrated prior to each test. The corresponding strain gauge outputs were then related to the axial force and bending moment, respectively. =.96 unitcm =.96 unitcm row of marked spaghetti was vertically implanted on the surface along the central line of the model at a regular interval and at predetermineositions as shown in Figure.2. After the spaghetti absorbs water it softens and will deform together with surrounding soil. The colored sand layers (2 mm in thickness) used together with the vertical marked spaghetti were good indicators of the deformations caused by the imposed test conditions. Table. Characteristics of Quartz sand G s D, in C u γ max mm g/cm γ min g/cm Quartz sand The maximum and minimum densities of sand were measured in the dry state, according to the method (JSF T 6-99) specified by the Japanese Geotechnical Society. g =.6. Strain Gage (Half-bridge) Expoxy coating. Bending Pile Axial Pile Figure. Configuration of bending pile and axial pile. =.6 Strain Gage (Full-bridge) Expoxy coating 8cm LVDT C D Figure 2. Test setup. axial pile Tunnel Vertical spaghetti cm bending pile 82cm Colored sand layer strain gage cm Deformation gauge Loading device 2.2 Preparation of saturated sand bed The model tunnel, the sand around it and the model piles were contained in an aluminum alloy strong box. The test setup is illustrated in Figure. 2. The internal dimensions of strongbox are 82 mm in length, 22 mm in width and 8 mm in height. In an effort to reduce wall friction, the wall surfaces of the strongbox were thoroughly greased and thin latex membranes were attached. The axial pile and the bending pile were installed at either side of the tunnel and at the locations having equal distance from the tunnel axis in the tests. After the model piles and the model tunnel were set on the predetermineositions, Quartz sand was pluviated into the strongbox with a regular path from a hopper at a falling height of 7 cm and at a constant flow rate. The characteristics of Quartz sand are summarized in Table. A fairly uniform sand bed (D r =6±2%) was achieved and its friction angle is around 8 o. The raining process was interrupted as needed to rain a layer of colored sand at specified elevations. Finally a After completion of model preparation, the package was mounted on the centrifuge platform. An acrylic plate was tightly covered on the strongbox. Simultaneously and continuously vacuumed the air both in the strongbox and in the model tunnel out and at the same time de-air water was carefully dripped into the strongbox to saturate the sand bed until the water level rose to 2 mm above the sand surface. It took 7 hours to do the saturation process. After removing the acrylic plate, a total five LVDTs (linear variable displacement transformers) were attached to a mounting unit to measure the surface settlements. In addition, two LVDTs mounted at the different elevations were used to measure the lateral displacement and rotation angle of the pile and one LVDT fixed on the top of the pile to measure the vertical settlement. The configuration of the LVDTs is shown in Figure. 2. Two series of tests were 2

3 conducted; one was without axial load on the pile head, the other with axial load on it. For the pile with axial load on it, loading devices were housed on both bending and axial pile heads as shown in Figure Test Procedures The model tunnel was connected with an air pressure line and the centrifuge was then accelerated in g steps until it reached g. The air pressure was cautiously regulated to balance the overburden pressure so that no crown settlement allowed during the acceleration of the centrifuge. Once the pore water pressure reached equilibrium in the acceleration of g, the tunnel diameter was reduced by gradually lowering the air pressure to zero at an decrement of kpa each seconds. The deformation of tunnel, the surface settlement and the deflection of pile, and the axial forces and the bending moments on the piles at various depths were continuously measured. Two series of tests were conducted in the study. In Test series A, one instrumenteile (axial pile) was used to perform a pile load test to investigate the load transfer mechanism and the ultimate bearing capacity for a pile embedded in ground where no tunnel near by. In Test series B, the piles were near a new driven tunnel. Two instrumented model piles in a penetration depth of m were located at either tunnel side and at a distance of. m from a tunnel embedded in depths of various cover-to-diameter ratios (C/D=, 2, and, and C = thickness of cover = a distance between ground surface and tunnel crown, and D = tunnel diameter). In Test series B, axial load was applied to the pile head to simulate a pile that structures rested on. All the tests in Test series B were used to investigate the mechanism of load transfer (including bending moment, and axial force on piles) and the pile head deformation during the tail voids closure of tunnel. Details of test arrangements are given in Table 2.. TEST RESULTS AND DISCUSSION. Ultimate bearing capacity of pile The load transfer mechanism from a pile to the soil is complicated. Figure displays the axial load and settlement curve in the pile load test (Ltest). The ultimate bearing capacity of pile is defined as the load capacity corresponding to the pile head settlement δp =.Dp (D p =pile diameter), it was around 66kN in the study. The relation of the ratio of point bearing load to total load (Q p /Q t ) with settlement is shown in Figure. The arrows in both the figures point the ultimate load. The unit skin friction mobilized on the pile with depths is shown in Figure. The unit skin frictions at the depths deeper than about 2 m didn t increase anymore. Settlement, (m) Axial force,(kn) Figure. Test result of pile load test. Settlement, (m).6 Figure. Ratio of point bearing load to total load. Depth, (m) Ltest Ratio of point bearing load to total load, Q p /Q t,(%) Ltest Ltest kN kn 8kN 67kN 87kN 9kN Figure. Variations of unit skin friction versus depths. Table 2. Test arrangements and test conditions Test Series Test No. C/D Tunnel depth, z o (m) Embedment Distance depth of x (m) pile, L p (m) Axial load, (kn) A Ltest Pile load test B Atest Atest2 Atest Atest Atest Atest

4 .2 Bending moment on pile due to tunneling about the same values of positive and negative bending moments on pile (Figure 7-a and 8-a), while deeper tunneling (the pile tip located above the tunnel invert) produced the big negative bending moments but the small positive from the tunnel buried in the depth of C/D= experienced the larger bending moments. However, no matter how large the axial load on the pile was, no obvious difference in the bending moment for the pile near the tunnel buried in the depth of C/D=2 was observed from a comparison of Figure 7-a with Figure 8-a. Tunneling adjacent to an existing pile can induce extra bending moments and axial forces and cause settlements and lateral deformations on the pile. As is listed in Table 2, all the tests were conducted using the pile, having the same embedment depth of m and located at a distance of. m from a new driven tunnel embedded at various depths, to discuss the impact on the pile resulted from tunneling. Figure 6 shows the notations on bending moment and axial force and test conditions. Figures 7 and 8 show the measured bending moments along the depths versus the various ground loss, (, %), that is defined as the ratio of the soil volume flowing into the tunnel during tunneling divided by the volume of the tunnel per unit tunnel length and expressed as a percentage, for the piles with various axial loads (8 kn and 2 kn) and located at the distance of. m from the tunnels axis and embedded in the cover-to-diameter ratios of 2 and. It can be observed from the Figures 7 and 8 that the bending moments at different depths increased with an increase of the ground loss until tunnel ( =-.%). Shallow tunneling near a long pile (the pile tip located below the tunnel invert) produced Figure 6. Notations on bending moment and on axial force. Bending moment (kn-m) Collapse =.% =2.% =.% Pi=.kPa Bending moment(kn-m) (C/D=,X=.m) Axial load=8kn - =% V - L =% Pi=kPa (C/D=2,X=.m) -2-2 Axial load=8kn - - Figure 7. Bending moments along depths for piles with axial load=8kn Depth(m) (C/D=2,X=.m) Axial load=2kn Bending moment (kn-m) (C/D=,X=.m) Axial load=2kn Figure 8. Bending moments along depths for piles with axial load=2kn.. Axial force on pile due to tunneling Figures 9 and display the distributions of the measured axial force along the depths versus the ground loss for the piles with axial load = 8 kn and 2 kn, respectively, those which were installed at the distance of. m from the tunnel embedded in the case of C/D =2 and. Nearly the same load transfer curves for the pile with the different axial forces but installed near the new driving tunnel having the same C/D were observed from a comparison of Figures 9 and. On the other hand, there were totally different load transfer curves for the piles installed near the tunnel having different cover-to-diameter ratios (e.g. C/D= 2 and ). Figures and 2 display the variations of the unit skin friction mobilized on the pile along the depths. The load transfer mechanisms on a pile when tunneling nearby are investigated in detail in the study. The magnitudes of unit skin friction mobilized on the pile body at various depths were dependent on both the normal stresses on the pile body and the relative displacements between the soil and the pile. The distributions of the unit skin friction on the long pile ( m in the study) near to shallow tunneling (e.g., C/D=2) are totally different from that near to deeper tunneling (e.g., C/D=). For the long pile near shallow tunneling, the unit skin frictions on the pile body between the elevation of D above and /2 D below the tunnel axis

5 (C/D=2,X=.m) Axial load=8kn =% =% Pi=kPa (C/D=,X=.m) Axial load=8kn - =% =% Pi=kPa Figure 9. Load transfer curves for a pile with axial force=8kn near to a new driving tunnel U nit skin friction, (kn/m 2 ) (C/D=2,X=.m) Axial load=8kn =% =% -2-2 Figure. Variations of unit skin friction along depths with various ground loss (axial load =8kN) (C/D=,X=.m) Axial load=8kn =% =% epth(m) D (C/D=2,X=.m) Axial load=2kn =% - =% Pi=kPa (C/D=,X=.m) Axial load=2kn =% =% Pi=kPa Figure. Load transfer curves for a pile with axial force=2kn near to a new driving tunnel (C/D=2,X=.m) Axial load=2kn =% =% (C/D=,X=.m) Axial load=2kn =% =% Figure 2. Variations of unit skin friction along depths with various ground loss (axial load =2kN). rapidly decreased with an increase of ground loss and reduced to zero or even to negative values. More loads transferred to the pile tip as shown in Figures 9-a and -a (in the case of C/D=2). The settlement of pile head would increase because of losing the partial resistance derived from skin friction due to nearby tunneling. Consequently, more axial load must carry by the point bearing. The more ground loss caused by tunneling, the less skin resistance can be mobilized, and the more axial load on the pile must be carried by the point bearing and the pile will eventually experiences the larger pile settlement. In addition, the more load already carried on the pile, it will suffer the more pile settlement after new driving tunnel, as shown in Figure. For the pile near to deep tunneling and its pile tip locating above the tunnel axis (e.g., in the cases of C/D=), the settlement of pile increased rapidly with the result that a large amount of point bearing resistance disappeared due to the ground loss caused by new tunneling. The downward movement of pile was larger than the settlement of the soil adjacent to the pile, therefore, the pile surface still mobilized the positive unit skin friction, which was slightly larger than that before new tunneling, to carry the pile load. This fact can also be seen clearly in Figure. The maximum surface settlements caused by new tunneling at tunnel are display with the hollow triangle symbols. For the pile near shallow tunneling(c/d=2 and ), the soils adjacent to the pile settled more than the pile head did, therefore, the negative skin friction mobilized as shown in Figures -a and 2-a in the Pile settlement or surface settleme nt, (c m) 2 2 without axial load with axial load = 8 kn with axial load = 2 kn surface settlement 2 Cover-to-diameter ratio, C/D Figure. Settlement of pile head with different axial loads and maximum surface settlement versus C/D.

6 Figure. Load transfer mechanism for the pile near new shallower tunneling (C/D=2). Figure. Load transfer mechanism for the pile near new deeper tunneling (C/D=). case of C/D=2. On the other hand, the settlement of pile head settled more than the soils did for the pile near to deeper tunneling. No negative skin friction force would develop on the pile. These load transfer mechanisms can be summarized in Figures and in the cases of C/D=2 and. A comparison of the distributions of unit skin friction on the piles indicates that the load transfer mechanism is considerably different when tunneling in various depths.. SUMMARY AND CONCLUSIONS A series of centrifuge model tests were performed to assess the tunneling-induced ground deformations and their detrimental effects on the neighboring single pile in the saturated sandy ground. The relative elevations between the pile tip and the burial depth of tunnel dominated the distributions of bending moment and axial force along the pile depths. The maximum bending moment and axial force on the pile increased rapidly until ground loss reached -.%. Shallow tunneling near a long pile (the pile tip located below the tunnel invert) produced about the same values of positive and negative bending moments, while deeper tunneling (the pile tip located above the tunnel invert) produced the big negative bending moments but the small positive bending moment. Shallow tunneling Unit skin friction Unit skin friction near the long pile, the unit skin frictions on the pile between the elevation of D above and /2 D below the tunnel axis rapidly decreased with an increase of ground loss and reduced to zero or even negative values. More loads transferred to the pile tip and caused the pile settlement. Deep tunneling near the pile the pile settlement increased rapidly with the result that a large amount of point bearing resistance disappeared due to the ground loss. The downward movement of pile was larger than the settlement of the soil adjacent to the pile, therefore, the pile surface still mobilized the positive unit skin friction, which was slightly larger than that before new tunneling, to carry the pile load. Hence the constructor must take the measures preventing the damage from both bending failure and foundation settlements into consideration during tunneling adjacent to the existing pile.. ACKNOWLEDGMENTS Financial support provided by the National Science Council of the Republic of China under the Grant NSC E-8- is gratefully acknowledgement. REFERENCE Bezuijen, A. and Schrier Van der, J. (99): The influence of a bored tunnel on pile foundations, Proceedings of the International Conference Centrifuge 9, (eds. Leung, Lee and Tan), pp Chen, L.T., Poulos, H.G., and Loganathan, N. (999): Pile responses caused by tunneling, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, Vol. 2, No., pp Jacobsz, S.W., Standing, J.R., Mair, R.J., Soga, K., Hagiwara, T., and Sugiyama, T. (2): The effects of tunneling near single driven piles in dry sand, Proceedings of Asian Regional Conference on Geotechnical Aspects of Underground Construction in Soft Ground, Shanghai, pp.29-. Lee, R.G., Turner, A.J., and Whitworth, L.J. (99): Deformation caused by tunneling beneath a piled structure, Proceedings of the XIII International Conference on Soil Mechanics and Foundation Engineering, New Delhi, pp Lee, C.J., Wu, B.R, and Chiou, S.Y. (999): Soil movements around a tunnel in soft soils, Proceedings of the National Science Council, Part A: Physical Science and Engineering, Vol.2, No.2, pp Loganathan, N., Poulos, H.G., and Stewart, D.P. (2): Centrifuge model testing of tunneling-induced ground and pile deformations, Geotechnique, Vol., No., pp Loganathan, N., Poulos, H.G., and Xu, K.J. (2): Ground and pile-group responses due to tunneling, Soils and Foundations, Vol., No., pp Poulos, H.G. and Chen, L.T. (997): Pile response due to excavation-induced lateral soil movement, Journal of Geotechnical and geoenviromental Engineering, ASCE, Vol. 2, No.2, pp