Probabilistic Site Characterization

Abstract

The special collection on Probabilistic Site Characterization is available in the ASCE Library (https://ascelibrary.org/page/ajrua6 /probabilistic_site_characterization). Site investigation and the interpretation of site conditions, profiles , and data are necessary aspects of sound geotechnical practice. Any geotechnical design methodology, be it reliability-based design or otherwise, should place site investigation as the cornerstone of the methodology. The purpose of this special collection on prob-abilistic site characterization is to advance research on how soil and rock databases and data-driven methods (probabilistic or more general computational intelligence methods such as machine learning) can be effectively deployed to augment existing physics-based methods and engineering judgment in decision making. Some emerging topics that are relevant to probabilistic site characterization , including clarifying its role and value to geotechnical practice, are as follows: 1. In contrast to structural materials such as steel and concrete, naturally occurring geomaterials such as soil and rock are not manufactured to meet prescribed quality specifications. 2. Spatial variability is an intrinsic feature of a site profile. 3. The problem of small sample size is more conspicuous in geotechnical engineering than structural engineering. 4. Site-specific data, correlations, and conditions have to be considered in relation to information from other sites with comparable geology. Site specificity arises because of the natural origin of soil and rock. 5. The characteristic value of a soil or rock property is defined as a cautious estimate of the value affecting the occurrence of a limit state. The limit state manifests itself physically as a critical slip surface or a soil-structure deformation mechanism, which is dependent on spatial variability. The statistics of this mobilized value governing the limit state are complex because of this coupling between spatial variability and mechanics. 6. Update site information and design during construction through the observational method (design and construction less clearly delineated than structural engineering). 7. Standardization in design codes is more challenging because of site-specific conditions and testing methods, design approaches, and local practices that evolved to suit these conditions. 8. The role of global multivariate distributions in capturing data from a basket of laboratory and field tests and how global distributions can be localized to be more relevant for a single site. 9. Statistical characterization of spatial variability and statistical identification of stratification, particularly in the presence of small sample size. In short, there are some characteristics of site investigation data that are relatively unique to geotechnical engineering. These characteristics can be succinctly described as MUSIC: multivariate, uncertain and unique, sparse, and incomplete. Probabilistic site characterization must take account of these realistic data characteristics to be useful to the profession. In particular, although it is commonly accepted that each site is unique to some degree, there is no method of characterizing this uniqueness that can lead to an automatic selection of similar sites. It is evident that generic correlation models that are widely used in the absence of sufficient site-specific data can be refined when the supporting database is drawn from similar sites only. More research is needed to address this site challenge under the constraint of MUSIC. The compilation of large generic databases is an important first step in a broader digitaliza-tion agenda to connect geotechnical engineering to Industry 4.0. It is possible to envisage realizing "precision construction," where characterization of "site-specific" model factors and "site-specific" soil parameters based on both site-specific and generic data can lead to further customization of design to a particular site and and even a particular location in a site. © ASCE 02018002-1 ASCE-ASME J. Risk Uncertainty Eng. Syst., Part A: Civ. Eng.