You are on page 1of 7

Deep Foundations on Bored and Auger Piles, Van Impe (ed.

) 2003 Millpress, Rotterdam, ISBN 90-77017-76-3

Piled raft foundation of a railway bridge in tertiary clay

Y. El-Mossallamy
Ain Shams University, Cairo, Egypt, c/o Arcadis Consult, Darmstadt, Germany

H. Schmidt
Deutsche Bahn, Germany

E. Gndling
Walter Bau, Germany

J. Lschner
Arcadis Consult, Darmstadt, Germany

Keywords: piled raft foundation, post shaft grouting, monitoring program, bored piles, organic clay ABSTRACT: The piled raft foundation has proved in the last two decades to be a very economic geotechnical foundation type carrying the structural loads partly by piles and partly by raft/soil contact pressures. The structural serviceability requirements can be fulfilled with relatively fewer piles in comparison with a pure piled foundation. An innovative application of the piled raft is its special adjustment to cases of foundations of highway railway bridges with high requirements regarding the allowable settlement, differential settlement and tilting. For the first time in Germany the piled raft foundation is applied for the foundation of a 4-span railway bridge with total length of 144 m. The paper describes the concept of piled raft for bridges in detail demonstrating the design requirements, design process and results of conducted pile load tests. The developed numerical model and design concept for the complex interaction between different foundation elements is presented. As no possibility does exist to take into account in an adequate manner- the soil disturbance caused due to pile installation by theoretical means, the numerical model as well soil parameters applied to design the piled raft foundation of the bridge were first calibrated and adjusted to the results of single pile load tests. The bridge foundations were then designed using the same numerical model and the determined soil and pile/soil parameters. A summary of the pile load test results including the pile/soil load transfer mechanism and the effect of post shaft grouting will be summarized in this paper. The results of the developed calculation model will be then presented and compared with those of the load tests in a short overview. The design of the piled raft foundation considering the pile load test results will be demonstrated and discussed. 2 CONCEPT OF PILED RAFT


The foundation of the 4-span railway bridge of E Flieden in Germany is planned applying the concept of piled raft foundation. The bridge consists of 4 spans with a total length of 144 m (Figure 1). To ascertain the adequacy of the piles and determine appropriate design values, pile load tests were first conducted on large diameter bored piles with and without post shaft grouting. Vertical, horizontal and cyclic loading have been applied to investigate the pile behavior under various load conditions.

Figure 1. Layout and geological conditions

In the conventional design of pile foundations, it is assumed that the total applied load has to be carried only by the piles with a certain factor of safety against bearing failure. Taking into account the contribution of the raft in the design can lead to a more economic foundation in comparison with a pure pile foundation. Moreover the piles can be designed in such cases to carry loads close to their ultimate ca-


pacity (DIBT, 2000). In such cases the whole system (raft/piles) should have an adequate bearing capacity against failure. The main function of the piles is to reduce the maximum and differential settlements to an acceptable level (serviceability requirements). Figure 2 demonstrates the principles of piled raft foundations and illustrates the settlement reduction Ds of a piled raft in comparison with the corresponding unpiled one as a function of the pile load share to the total load DL.

Table 1 summarizes the results of the conducted laboratory tests of both clay and lignite layers.
Table 1: Geotechnical parameters. layer 3a Liquid limit WL [%] 60 120 Plastic index IP [%] 30 90 1.01 Consistency index Ic Wet unit weight J [kN/m] 19.8 Undrained cohesion 147 Layer 3b 92 124 58 85 0.95 17.7 133

These results show no significant difference between these two types of clay regarding their stiffness and shear parameters. Nevertheless, it was expected that the lignite layer is sensitive to mechanical disturbance due to its fracture structure. 4 PILE LOAD TEST

Figure 2. Principals of piled raft (El-Mossallamy 1996, DI BT 2000)


As shown in figure 1, the subground consists of quaternary silty sandy clay down to about 5 to 7 m followed mainly by tertiary formations of clay with lenses of lignite coal (clay with brown coal). The lignite coal lenses were found in the upper 10 to 15 m of the tertiary layers. The groundwater level is mostly few meters just above the tertiary clay surface. Regarding the conducted in-situ and laboratory tests (Figure 3), the tertiary clay can be classified as stiff plastic clay.

The pile load tests were carried on two piles with a pile diameter of 0.9 m and a total length of 12.0 m. The above 1.5 m was excavated to prevent direct load transfer to the filling required to have a suitable level for the construction of piles. The tested pile lengths were located completely in the tertiary clay formations. The two tested piles (with and without post shaft grouting) were constructed as bored piles using temporary steel casing. The loading system consisted of two hydraulic jacks working against a reaction beam. The reaction beam was supported by two bored piles with diameter of 0.9 m and length of 15 m acting as tension piles. The reaction piles were installed at a distant of 3.6 m (= 4 pile diameter) from the tested pile to minimize the effect of the mutual interaction between the tested pile and the reaction system (Figure 4).

Figure 3. Results of laboratory and in-situ tests

Figure 4. Pile load test arrangement


El-Mossallamy, Y., Schmidt, H., Gndling, E. & Lschner, J.

With this arrangement, 5 piles were installed in a row: two test piles (pile A and B with and without shaft grouting) and three reaction piles (R1, R2 and R3). Vertical loading tests were carried out on piles A and B and then horizontal loading tests were followed on the reaction piles (R2 and R3). The loads were applied in increments and maintained constant till the settlement rate was negligible ( 0.1 mm/ 5 min, according to the german recommendation for pile load test). Cyclic loads were also applied under vertical and horizontal loads to investigate the pile behavior under static and dynamic loads. Both the applied loads and the corresponding displacements at the tested pile head were measured. Additionally the pile displacements in different depths were measured using multipoint extensiometer (Figure 4), so that the skin friction in different layers could be determined. The load carried by pile base was measured directly using a hydraulic load cell attached to the pile reinforcement. The horizontal deformation with depth of the tested piles under horizontal loads were measured using inclinometers. The inclinometers were installed in steel tubes that were welded to the pile reinforcement of reaction pile R2 and R3. 4.1 Results of the vertical pile load tests Figure 5 shows a comparison between the two tested piles with and without shaft grouting.

about 80 kPa. The organic silty clay layer (3 c) has an ultimate skin friction of about 40 kPa. In case of shaft grouting the pile skin frictions of the lignite coal and silty clay increases to about 80 kPa and 72 kPa respectively. In the other hand the increase of the pile skin friction in the stiff plastic clay (3 a) is almost negligible. These results are conform with the expected mechanical sensitivity of the organic silty clay and lignite coal. It can be concluded that the installation of the piles cause a disturbance in the sensitive clay in a very narrow zone around the pile shaft. This disturbed zone causes a high reduction of the pile skin friction. The disturbed soil in this zone consolidates under the pressure of the post shaft grouting and regain its shear strength. Therefore the skin friction in this case depends on the undisturbed shear parameters of the soil outer the disturbed zone. Regarding these results, it was decided to install all foundation piles applying post shaft grouting.

Figure 6. Development of pile skin friction

The results of the 30 cyclic loads are shown in Figure 7. The pile/soil stiffness under short time applied loads (dynamic loads) is about 3 times higher that that under static loads.

Figure 5. Load-settlement behavior of piles with and without post shaft grouting

It can be recognized that the shaft grouting increases the bearing capacity significantly. To evaluate this effect, the skin friction mobilization for both piles are determined using the measurements of the multipoint extensiometers (Figure 6). In case without shaft grouting, the ultimate skin friction of the lignite coal (3 b) is 25 kPa. This value is too smaller than that of the stiff plastic clay (3 a) which reaches

Figure 7. Results of vertical cyclic loads

Piled raft foundation of a railway bridge in tertiary clay


4.2 Results of the horizontal pile load tests The horizontal load deformation relationship is shown in Figure 8. The piles R2 and R3 are installed without post shaft grouting. The extrapolation of the results of the horizontal load test shows an ultimate horizontal load about 800 to 850 kN. This results support the assumption that the mechanical disturbance of the lignite clay is concentrated in a narrow zone around the pile shaft. This illustrates the effectiveness of the post shaft grouting to increase the vertical pile capacity.

Figure 10. Comparison between observed and calculated pile horizontal deformation


Figure 8. Horizontal load - deformation behavior

Figure 9 shows the pile behavior under horizontal cyclic loading, which is needed to investigate the foundation behavior under brake and wind loads. The pile/soil stiffness under cyclic loads is about 3 times higher that the stiffness under static loads.

The bridge piers are consisted of two pillars, each founded on a separate foundation. The design of the piled raft foundation of the bridge is based on the results of the carried load tests. Figure 11 demonstrates schematically the design concept that was developed.

Figure 9. Results of horizontal cyclic loads

The horizontal deformation of the pile with depth is shown in Figure 10 for two load levels in comparison with back calculated values applying theory of subgrade reaction. The pile stiffness was changed depending on the load level to consider the effect of tension cracks. This comparison demonstrates that the soil subgrade reaction modulus is not a soil constant but it depends mainly on the problem dimension as well on the applied load level. The back calculated values of the soil subgrade reaction modulus and its distribution with depth were used then by the design of the whole foundation under horizontal loads. The group action was taken into account according to the german standards 4014.

Figure 11. The proposed flow chart for the design of piled raft

As no possibility does exist to take into account in an adequate manner- the soil disturbance caused due to pile installation by theoretical means, the numerical model as well soil parameters applied to design the piled raft foundation of the bridge were first calibrated and adjusted to the results of single pile load tests. The bridge foundations were then de-


El-Mossallamy, Y., Schmidt, H., Gndling, E. & Lschner, J.

signed using the same numerical model and the determined soil and pile/soil parameters. 5.1 Numerical model Calculation procedures to model the behavior of such complex three-dimensional problems have been developed since the 1970s (e.g. by Butterfield /Banerjee 1970, Poulos 1980 and Randolph 1983). But some important requirements concerning for example the raft stiffness, the nonlinear pile response and the slip developing along the pile shafts even under working loads were not sufficiently considered in these analyses. In the recent years, three dimensional finite element method with enhanced constitutive laws were applied to analyze the behavior of piled rafts (Katzenbach et. al. 2000 and Maybaum/Mhlmann 2003). Nevertheless, such analysis needs a huge memory capacity, that restricts the application of the 3D finite element method. For these reasons an improved numerical model based on a coupled technique of the Finite Element Method (FEM) and the Boundary Element Method (BEM) was developed taking into account all abovementioned effects. The raft is modeled by FEM as plate-in-bending (eventually taking into account the superstructure stiffness) acted on by the superstructure loads and supported by nonlinear elastic springs at each node of the finite element mesh. These springs partly represent piles and partly the soil support beneath each raft/soil element. The stiffness of the springs depends on the load and on the settlement at each node. The pile groups are modeled by BEM taking the raft contact pressures into account. The boundary element mesh contains only elements at the raft/soil and at the pile/soil interfaces. With the developed computer program the compatibility of the deformations of the raft and of the pile group is achieved by an incremental type of nonlinear calculation with iterative adaptation of FEM (applied to the raft) and BEM (applied to the pile group) in each increment. The considered nonlinear pile response (El-Mossallamy 1999) is due to: slip between the pile shaft and the adjacent soil, nonlinear behavior of the soil adjacent to the pile shaft and nonlinear soil behavior at the pile base. The different stress/strain behavior of the soil in reloading and primary loading is also considered. For more details about the method of analysis, reference is made to El-Mossallamy 1996 and El-Mossallamy/ Franke 1997. This technique with the corresponding software was used for the design of piled raft foundations of many high rise buildings in Frankfurt. A recent example is the high-rise building Gallileo (ElMossallamy/Stahlmann 2002).

5.2 Numerical simulation of the behavior of the bridge foundation The first stage is to calibrate the chosen pile and soil parameters regarding the results of the pile load test. Figure 12 shows the observed load settlement behavior of the tested pile with post shaft grouting in comparison with the calculated values. In this analysis, the ultimate pile skin friction is chosen equal to 70 kPa for all soil layers to include the effect of post shaft grouting. The applied soil deformation modulus is taken equal to 45 MPa (conservative value). The results show a good agreement between observed and calculated behavior.

Figure 12. Comparison between observed and calculated load settlement behavior of tested pile B

Figure 13. Layout of the foundation

Piled raft foundation of a railway bridge in tertiary clay


The bridge piers are consisted of two pillars, each founded on a separate group of 6 piles. The pile arrangement is shown in Figure 13. The piles were designed with a diameter of 1.2 m and a length of 18 m regarding the results of the pile load test and the results of the conducted numerical analyses of the whole foundation. To obtain a better understanding of the pile group action, the load settlement behavior of single pile with pile length of 18 m and pile diameter of 1.2 m was detected using the same soil parameters. Figure 14 compares the behavior of the single pile with the average behavior of the pile group of the piled raft. The pile group action defined as the ratio of the pile group settlement to the settlement of the single pile under the same average load can be determined in this case to be in the order of about 3.2. This knowledge shows the importance of considering the pile group action by the design of piled foundation specially in case of mainly skin friction piles.

incrementally applied till the working loads to determine the deformation under working loads (serviceability limit state). Another aim of the analysis under working loads is to determine the pile/soil stiffness and the subgrade reaction distribution beneath the raft, which are necessary for the structural design of the foundation. Furthermore, the loads are increased till twice the working load to investigate the global stability of the whole foundation system (ultimate limit state).

Figure 16. Load settlement behavior of individual piles among the pile group

Figure 14. Pile group action

Figure 16 demonstrates the load settlement behavior of individual piles among the pile group depending on pile position. The edge piles carry a higher load than the middle piles. Therefore, the pile/soil stiffness is not a constant value but it depends on the number of piles, on the pile position and on the applied load level.

Figure 15. Load-settlement behavior of the piled raft foundation

Figure 15 demonstrates the load settlement behavior of the piled raft. Due to the nonlinear response of the foundation system, the loads should be

Figure 17. Bearing capacity of piled raft


El-Mossallamy, Y., Schmidt, H., Gndling, E. & Lschner, J.

It is also appropriate to comment on the safety of the piled raft against the ultimate limit state caused by yielding of the soil. It is suggested here to apply one of the two models shown in Figure 17 (Franke et. al. 2000). In this particular case, both models deliver safety factor higher than 2. 6 SUMMARY AND CONCLUSION

An extensive monitoring program should be considered as a part of the pile design (specially in cases where there is no enough existing experience) to ensure a realistic design concept. Post shaft grouting can be helpful to regain soil shear strength in the close vicinity around the pile shaft due to installation disturbance in sensitive clay. The piled raft foundation provides a new geotechnical concept for the design of foundations for bridges, which in addition to stability and serviceability, allows a high-grade technical and economical optimization of the construction. The fact that any kind of load combination can be safely founded on soft to stiff clay (also on medium to dense sand) by appropriate location of piles, pile length and through suitable pile diameters means almost unlimited possibilities for the construction of bridge piers. Each project has, however, its own individual boundary conditions demanding an optimized adaptation of the foundation through comparison of alternatives and through modeling. Understanding of the effects of the interaction between construction and subsoil based on the appropriate theoretical knowledge and on experienced application of measurement techniques and numerical modeling together with tested and proven implementation design methods are the necessary qualification for a safe and economic design for piled raft foundations. REFERENCES
Butterfield, R., & Banerjee, P.K. (1971). The problem of pile group-pile cap interaction. Gotechnique, Vol. 21, No. 2, pp. 135-142. Deutsches Institut fr Bautechnik (2000) Richtlinie fr den Entwurf, die Bemessung und den Bau von Kombinierten Pfahl-Plattengrndungen (KPP), u.a. erschienen in Bautechnik Heft 9, Zeitschrift fr den gesamten Ingenieurbau, 77. Jahrgang, September 2000 El-Mossallamy, Y. (1996). Ein Berechnungsmodell zum Tragverhalten der kombinierten Pfahl-Plattengrndung. Dissertation, Fachbereich Bauingenieurwesen der Technischen Hochschule Darmstadt El-Mossallamy, Y. & Franke, E. (1997). Piled Rafts Numerical Modeling to Simulate the Behavior of Piled Raft Foundations. Published by the Authors, Darmstadt El-Mossallamy, Y (1999). Load-settlement behavior of large diameter bored piles in over-consolidated clay. Proceeding of the 7 th. International Symposium on Numerical Models in Geotechnical Engineering, Graz, Austria.

El-Mossallamy, Y. & Stahlmann, J. 2002. Economic Application of Piled Raft Foundations: Case Histories. Felsbau, 6/2002 Franke, E., El-Mossallamy, Y. & Wittmann, P. 2000. Calculation methods for raft foundation in Germany. Design applications of raft foundation, edited by Hemsley, Thomas Telford, 283-322 Katzenbach, R., Arslan, U. & Moormann, C. 2000. Piled raft foundation projects in Germany. Design applications of raft foundation, edited by Hemsley, Thomas Telford, 323-391 Maybaum, G. & Mhlmann, J. 2003. Numerische Berechnung zur Kombinierten Pfahl-Plattengrndung des Investment Banking Center, Frankfurt. Pfahlsymposium 2003 an der TU Braunschweig. Poulos, H.G., & Davis, E.H. 1980. Pile Foundation Analysis and Design. Wiley, New York. Randolph, M.F. 1983. Design of piled raft foundations. CUED/D-Soils TR. 143. Related material Burland, J.B., Broms, B.B. & de Mello, V. 1977. Behaviour of foundations and structures. State-of-the- art report, Proc. 9th ICSMFE Tokyo, vol. 2, 495 - 546 Chow, Y.K. 1991. Pile-soil-pile interaction considering weakened zone of soil around piles. Computer and Geotechnics, vol. 12, 163 -174 Cooke, R.W., Bryden-Smith, D.W. & Goach, M.N. 1970. Some observations of the foundation loading and settlement of a multi-storey building on a piled raft in London clay. Proc. Inst. Civ. Engrs., part I, 1981, 433 - 460 El-Mossallamy, Y. & Franke, E. 1997. Pfahl-PlattenGrndungen: Theorie und Anwendung. Bautechnik 74, Heft 11 Hansbo, S. 1984. Foundations on friction creep piles in soft clay, Proc. Intern. Conf. on Case Histories in Geotechnical Engineering, ,St. Louis 2, 913-922. ONeill, M.W. 1983. Side load transfer in driven and drilled piles. Proc. ASCE (GT), vol. 109, 1259 - 1266 Randolph, M.F. 1977. A theoretical study of the performance of piles. PhD thesis, Cambridge University, England Randolph, M.F. & Clancy, P. 1993. Efficient design of piled rafts, Proc. 2nd Int. Seminar Deep Foundation, Ghent: 119130. Sommer, H., Wittmann, P. & Ripper, P. 1984. Zum Tragverhalten von Pfhlen im steif-plastischen Tertirton. Vortrge der Baugrundtagung in Dsseldorf, 501 -531, Deutsche Ges. f. Erd- u. Grundbau e.V., Essen

Piled raft foundation of a railway bridge in tertiary clay