The story of statistics in geotechnical engineering

Abstract

The story of statistics in geotechnical engineering can be traced to Lumb's classical Canadian Geotechnical Journal paper on "The Variability of Natural Soils" published in 1966. In parallel, the story of risk management in geotechnical engineering has progressed from design by prescriptive measures that do not require site-specific data, to more refined estimation of site-specific response using limited data from site investigation as inputs to physical models, to quantitative risk assessment (QRA) requiring considerable data at regional/national scales. In an era where data is recognised as the "new oil", it makes sense for us to lean towards decision making strategies that are more responsive to data, particularly if we have zettabytes coming our way. In fact, we already have a lot of data, but the vast majority is shelved after a project is completed ("dark data"). It does not make sense to reduce one zettabyte to a few bytes describing a single cautious value. It does not make sense to expect big data to be precise and to fit a particular favourite physical model as demanded by the classical deterministic world view. This paper advocates the position that there is value in data of any kind (good or not so good quality, or right or wrong fit to a physical model) and the challenge is for the new generation of researchers to uncover this value by hearing what data have to say for themselves, be it using probabilistic, machine learning, or other data-driven methods including those informed by physics and human experience, and to re-imagine the role of the geotechnical engineer in an immersive environment likely to be imbued by machine intelligence.

What is a characteristic value for soils?

“…This is known as the site challenge (Phoon 2018) or site recognition challenge (Phoon, Ching, and Shuku 2021). Recent research shows promising results using data-driven methods (Phoon 2020;Ching et al 2020a;Ching, Wu, and Phoon 2021;.…”

“…The key point is that it is possible to account for uncertainty in an objective way based on measured data without having a precise estimate of the coefficient of the variation. The need to strengthen the link between decisions and data has become more pressing with the rapid growth in digital technologies (Phoon 2020 While EN 1990(CEN 2002 generally recommends the use of mean values as characteristic values for stiffness parameters, EN 1997-1 indicates that cautious estimates should be used as characteristic values also in SLS. In the 2021 draft version of prEN 1997-1 (CEN 2020) the design values are determined using the same principles for ULS and SLS analyses, applying partial factors which, for SLS analyses, the recommended values are 1.0, to representative values, which in turn are based either on cautious estimates or statistics.…”

Deep Foundation Pit Excavations Adjacent to Disconnected Piled Rafts: A Review on Risk Control Practice

“…They indicated that countries have already been shifting the national geotechnical design codes of standards towards reliability-based design concepts where separate "partial" factor of safety is employed unlike that of comprehensive one in the traditional method. In this regard, it is prudent to carefully examine what existing (monitoring) data have to implicate in decision making; especially as data is increasingly being considered as "new oil" to the point whereby data is used to support sensitive decisions apart from project compliance purposes [125] [126]. For instance, if one excavation failure phenomenon among the 15 failure categories seen in china [127] is suspected to possibly occur, then the reliability of the critical event can be checked along with observations made as excavation progresses and remedial measures can be deployed to control it early from happening.…”

Spotlight article “The story of statistics in geotechnical engineering”

“…The aim and scope of Georisk was revised in 2018 to cater to the rapid pace of research in data analytics and data-driven methods: "Many opportunities to leverage on the rapid advancement in Bayesian analysis, machine learning, artificial intelligence, and other data-driven methods also exist, which can greatly enhance our decision-making abilities". In an era in which data is recognised as the "new oil", it makes sense for us to lean towards decision making strategies that are more responsive to data (Phoon 2020). Georisk is publishing an increasing number of papers on compilation and novel applications of databases to better support decision making (e.g.…”