Evolution of geotechnical risk analysis in North American practice - Risk and Reliability in Geotechnical Engineering - page 467

The interesting and original aspect of this body of work was that statistical thinking
and geometrical probability led to nonintuitive inferences about geological and engineer-
ing properties of the subsurface. Sampling for soil properties had been, and remains,
straightforward. Sampling for geometric properties raised issues of truncation, censor-
ing, orientation bias, nonplanar support of probability distributions, and a host of other
complications.
The purpose behind so much interest in joint systems at the time was analyzing rock
slope stability, especially in surface mines. For economic reasons, these slopes are designed
with low factors of safety; yet, it is undesirable for them to fail prior to the completion of
mining operations. Their stability depends both on the engineering properties of the intact
rock and on the fabric of the rock mass including the joint systems. The combination of eco-
nomic incentive, significant uncertainties about in situ conditions of the rock mass, and the
need to balance costs against safety made this an obvious candidate for statistical decision
theory. McMahon was among the early contributors to this direction of work (McMahon
1975), and the Ministry of Natural Resources, Canada, developed guidance along these
lines (Coates 1977). Einstein et al. (2010) surveyed this early work.
12.5 oFFShore relIabIlItY (1974-1990)
Following the oil crisis of 1973-1974, oil exploration and production became an inter-
national priority. Offshore exploitation of oil had started in California in the late 1800s
and in the Gulf of Mexico before World War II, but accelerated following the war. By the
1970s, offshore production was routine and widespread (National Commission on the BP
Deepwater Horizon Oil Spill and Offshore Drilling, 2010). The industry had experienced
a series of serious failures over the years. In 1969, a blow out at a Union Oil Company
rig in the Santa Barbara Channel produced an 800 square miles oil slick on the coast,
which was followed in the next 2 years by three other blowouts and a major fire on rigs
in American waters. By 1980, these were followed by failures in the Persian Gulf, Niger
Delta, North Sea, and Mexico. In 1988, the Piper Alpha tower in the North Sea exploded
killing 167.
Geotechnical reliability developments in the offshore realm are closely integrated with
those of structural reliability. Since earlier generations of offshore platforms were typically
founded on driven piles, a great deal of research was devoted to piles. Some of this early
work is summarized in Høeg and Tang (1978) and in Wu et al. (1989). Developments on
the reliability of offshore piles are summarized by Lacasse and Goulois ( 1989) and by Tang
et al. (1990); a recent overview is provided by Gilbert et al. (2013). A major contribution to
this body of work was the introduction of pile resistance calibrations against historical data
and the relation of such calibration to codes.
A closely related area of code development driven by offshore applications was the intro-
duction of load and resistance factor design (LRFD) to foundation engineering. Motivated
by the AISC structural steel code, which was an early adopter of LRFD, both the offshore
industry and the Federal Highway Administration (FHWA) became early proponents of
LRFD in the geotechnical design. These efforts were closely integrated with structural reli-
ability. Benchmark contributions include Lloyd and Karsan (1988) and Moses and Larrabee
(1988). In transportation engineering, corresponding contributions include Paikowsky et al.
(2004, 2010).
In more recent years, the offshore industry has moved to updating and calibrating reli-
ability-based methods in light of field observations in a series of serious hurricanes: Andrew
(1992), Ivan (2004), Katrina/Rita (2005), Ike (2008), and others.