Stochastic nonlinear p-y analysis of laterally loaded piles

B.K. Low and C.I. Teh
School of Civil and Structural Engineering, Nanyang Technological University, Singapore
Wilson H. Tang
School of Civil Engineering, Hong Kong University of Science and Technology, Hong Kong, China

The analysis of a laterally loaded pile based on the p-y curve concept is commonly done using the finite element or the finite difference method. The problem is akin to the classical problem of the beam on elastic foundation, except that in a laterally loaded pile the springs which model the resistance offered by the soil medium typically follow a nonlinear resistance-deflection behavior (nonlinear p-y curves). The nonlinear p-y curves are also likely to vary with depth. Hence the deterministic soil-structure interaction analysis is intricate, and traditionally requires iterative numerical methods. Extending the deterministic analysis to a reliability analysis is hitherto a complicated task, by virtue of the implicit, numerical, and iterative nature of the performance function, and the need to account for the spatial correlation of the soil spring properties.

An alternative practical numerical method based on constrained optimization has been presented for nonlinear strain-softening p-y analysis of laterally loaded single piles. Only a minimal amount of programming in the spreadsheet environment is required. The deterministic numerical procedure was then extended into reliability analysis, in which the soil springs were modeled stochastically to reflect spatial variation of soil resistance. An intuitive expanding ellipsoid perspective in the original space of the random variables (Low and Tang, 1997) greatly simplifies the computation of the reliability index as defined by Hasofer-Lind, such that the transition from the numerical soil-pile interaction deterministic analysis to a stochastic nonlinear and strain-sftening p-y analysis requires little additional effort, despite the highly implicit nature of the performance function. The procedure for cross-correlated and autocorrelated variables is the same as for uncorrelated variables, and does not involve orthogonal transformation of the correlation matrix.

Multicriteria reliability-based design was illustrated for a steel tubular pile that forms part of a pile group in a breasting dolphin. Both ultimate limit state and serviceability limit state were considered. Advantages of reliability-based design over design based on partial factors were discussed.

References

Ang, H. S., and Tang, W. H. (1984). Probability concepts in engineering planning and design, vol.
2 - Decision, risk, and reliability. John Wiley, New York.

Haldar, A., and Mahadevan, S. (1999). Probability, reliability and statistical methods in engineering design. John Wiley, New York.
Hasofer, A. M., and Lind, N. C. (1974). “Exact and invariant second-moment code format.” J. of Engineering Mechanics, ASCE, New York, 100(1), 111-121.
Low, B. K., and Tang, W. H. (1997a). “Efficient reliability evaluation using spreadsheet.” J. of Engineering Mechanics, ASCE, New York, 123(7), 749-752.
Low, B. K., and Tang, W.H. (1997b). “Reliability analysis of reinforced embankments on soft ground.” Canadian Geotechnical Journal, Ottawa, Canada, 34(5), 672-685.
Low, B. K., and Teh, C. I. (1999). “Probabilistic Analysis of Pile Deflection Under Lateral Loads.” International Conference on Applications of Statistics and Probability (ICASP8), A. A. Balkema, Rotterdam, The Netherlands, Vol. 1, pp. 407-414.
Tomlinson, M. J. (1994). Pile design and construction practice, 4th Ed., E & FN Spon, London.

Get this paper: B.K. Low