Multivariate probabilistic back analysis of triaxial tests of Copenhagen clay till

Efthymios Panagiotis*, Irene Rocchi, Varvara Zania

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference abstract in proceedingsResearchpeer-review

27 Downloads (Pure)


When multiple field and/or laboratory tests are carried out at one site in close proximity (e.g., triaxial tests, oedometer tests, classification tests) geotechnical data can be considered multivariate. However, the concept of the multivariate distributions has been limitedly applied in geotechnical engineering due to the finite available geotechnical data owing to budget constraints or the incompleteness and sparsity of geotechnical databases [1,3]. Some studies have attempted to cope with this challenge by using data from global databases to construct generic probability density functions [1,2], although it is widely accepted that similar soil types might behave differently, even though they were formed under similar geological conditions. Recently, two multivariate analyses at a regional level were performed ([5,4]) based on compiled databases of geotechnical properties of Shanghai and Finnish clays respectively. The objective of this study is to construct a regional multivariate distribution function in the greater Copenhagen area, consisting of eight parameters derived by 33 triaxial (20 undrained and 13 drained) and 142 soil classification tests for Copenhagen’s Quaternary upper clay till deposits. The constructed distribution will then be used for reliability design of deep excavations. Therefore, the parameters selected for this distribution are those required for the validation of the Hardening-small strain stiffness model, which is the chosen constitutive model. Finally, triaxial tests are back analysed using the generated samples of the multivariate model.
Original languageEnglish
Title of host publication4th International Symposium on Machine Learning & Big Data in Geoscience : Abstract Book
Number of pages3
PublisherInternational Society for Soil Mechanics and Geotechnical Engineering
Publication date2023
Publication statusPublished - 2023
Event4th International Symposium on Machine Learning and Big Data in Geoscience - University College Cork, Cork, Ireland
Duration: 29 Aug 20231 Sept 2023
Conference number: 4


Conference4th International Symposium on Machine Learning and Big Data in Geoscience
LocationUniversity College Cork


  • Multivariate distribution function
  • Nataf transformation
  • Clay till


Dive into the research topics of 'Multivariate probabilistic back analysis of triaxial tests of Copenhagen clay till'. Together they form a unique fingerprint.

Cite this