Verification of geotechnical reliability using load tests and integrity tests - Risk and Reliability in Geotechnical Engineering - page 523

Table 14.1 Factor of safety for pile capacity

Minimum factor of safety
Compression

Method of determining
capacity Loading condition Tension

Theoretical or empirical
prediction to be verified
by pile load test

Usual
Unusual
Extreme

2.0
1.5
1.15

2.0
1.5
1.15

Theoretical or empirical
prediction to be verified
by pile driving analyzer

Usual
Unusual
Extreme

2.5
1.9
1.4

3.0
2.25
1.7

Theoretical or empirical
prediction not verified
by load test

3.0
2.25
1.7
Source: After US Army Corps of Engineers (USACE). 1993. Design of Pile Foundations, New York:
ASCE Press.

Usual
Unusual
Extreme

3.0
2.25
1.7

In a reliability-based design (RBD), the value of pile tests can be further maximized. In
an RBD incorporating the Bayesian approach (e.g., Ang and Tang 2007), the same load
test results reveal more information. Predicted pile capacity using theoretical or empirical
methods can be very uncertain. Results from load tests not only suggest a more realistic pile
capacity value, but also greatly reduce the uncertainty of the pile capacity since the error
associated with load-test measurements is much smaller than that associated with predic-
tions. In other words, the reliability of a design for a test site can be updated by synthesizing
existing knowledge of pile design and site-specific information from pile tests.
This chapter presents systematic methods to incorporate results of quantitative proof load
tests and qualitative integrity tests into foundation design. Two examples are worked out
to evaluate the reliability of piles after verification by proof load tests and integrity tests.
Reliability and FOS are used together in this chapter since they provide complementary
measures of an acceptable design (Duncan 2000) as well as a better understanding of the
effect of pile tests on the costs and reliability of pile foundations. This chapter will focus
only on ultimate limit states.

14.2 WIthIn-SIte VarIabIlItY oF PIle CaPaCItY

In addition to uncertainties with site investigation, laboratory testing, and design models,
the capacity values of supposedly identical piles within one site also vary. Suppose several
“identical” test piles are constructed at a seemingly uniform site and are load tested fol-
lowing an “identical” procedure. The measured values of the ultimate capacity of the piles
would usually be different due to the so-called “within-site” variability following Baecher
and Rackwitz (1982). For example, Evangelista et al. (1977) tested 22 “identical” bored
piles in a sand-gravel site. The piles were “all” 0.8 m in diameter and 20 m in length, and
construction of the piles was assisted with bentonite slurry. The load tests revealed that the
coefficient of variation (COV) of the settlement of these “identical” piles at the intended
working load was 0.21, and the COV of the applied loads at the mean settlement corre-
sponding to the intended load was 0.13.
Evangelista et al. (1977) and Zhang et al. (2004) described several sources of the within-
site variability of pile capacity: inherent variability of properties of the soil in the influence
zone of each pile, construction effects, variability of pile geometry (length and diameter),
variability of properties of the pile concrete, and soil disturbance caused by pile driving