A rigorous possibility approach for the geotechnical reliability assessment supported by external database and local experience

Tombari, Alessandro; Dobbs, Marcus; Holland, Liam M.J.; Stefanini, Luciano. 2024 A rigorous possibility approach for the geotechnical reliability assessment supported by external database and local experience. Computers and Geotechnics, 166, 105967.

Before downloading, please read NORA policies.
Text (Open Access Paper)
1-s2.0-S0266352X23007243-main.pdf - Published Version
Available under License Creative Commons Attribution 4.0.

Download (3MB) | Preview


Reliability analyses based on probability theory are widely applied in geotechnical engineering, and several analytical or numerical methods have been built upon the concept of failure occurrence. Nevertheless, common geotechnical engineering real-world problems deal with scarce or sparse information where experimental data are not always available to a sufficient extent and quality to infer a reliable probability distribution function. This paper rigorously combines Fuzzy Clustering and Possibility Theory for deriving a data-driven, quantitative, reliability approach, in addition to fully probability-oriented assessments, when useful but heterogeneous sources of information are available. The proposed non-probabilistic approach is mathematically consistent with the failure probability, when ideal random data are considered. Additionally, it provides a robust tool to account for epistemic uncertainties when data are uncertain, scarce, and sparse. The Average Cumulative Function transformation is used to obtain possibility distributions inferred from the fuzzy clustering of an indirect database. Target Reliability Index Values, consistent with the prescribed values provided by Eurocode 0, are established. Moreover, a Degree of Understanding tier system based on the practitioner’s local experience is also proposed. The proposed methodology is detailed and discussed for two numerical examples using national-scale databases, highlighting the potential benefits compared to traditional probabilistic approaches.

Item Type: Publication - Article
Digital Object Identifier (DOI):
ISSN: 0266352X
Date made live: 24 Jan 2024 13:38 +0 (UTC)

Actions (login required)

View Item View Item

Document Downloads

Downloads for past 30 days

Downloads per month over past year

More statistics for this item...