Files in this item

FilesDescriptionFormat

application/pdf

application/pdfASMAR-DISSERTATION-2017.pdf (270MB)
(no description provided)PDF

application/zip

application/zipElectronic files-part1.zip (385MB)
(no description provided)ZIP

application/zip

application/zipElectronic files-part2.zip (303MB)
(no description provided)ZIP

Description

Title:Next generation triaxial apparatus using combined computational-experimental testing framework
Author(s):Asmar, Randa Khaleel
Director of Research:Hashash, Youssef M. A
Doctoral Committee Chair(s):Hashash, Youssef M. A
Doctoral Committee Member(s):Ghaboussi, Jamshid; Olson, Scott M.; Rutherford , Cassandra J.
Department / Program:Civil & Environmental Eng
Discipline:Civil Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Modified triaxial device
Digital photogrammetry
Image processing
3D deformations
Inverse analysis
Deep learning
Abstract:Solving complex boundary value problems in geotechnical engineering requires a soil constitutive model that reliably captures the soil behavior under general loading conditions. However, soil behavior is commonly characterized based on laboratory tests with imposed or assumed uniform stress and strain distribution within the soil specimen for convenient data reduction. This uniformity assumption limits each test to a single stress–strain path, and therefore extensive laboratory testing is required to represent real soil behavior such as small strain nonlinearity and anisotropy. The process of development of material constitutive models remains lengthy and requires numerous tests to cover a broad range of loading paths. This limited information generally results in a constitutive model that may not be justifiable to represent loading conditions that differ substantially from the ones in laboratory tests. This study presents the development of a modified triaxial device that can generate multiple stress paths in a single test which can be extracted using Self-learning simulations (SelfSim) inverse analysis – an advanced deep learning computational engine. The new device inherits all features of a conventional triaxial test, and adds lateral restraint clamps, to increase non-uniformity in specimen deformation, combined with a digital photogrammetry system to measure the 3-D deformed shapes of the specimen. Using two high resolution digital cameras mounted in front of the cell, the system is able to capture the specimen’s deformed shape synchronously with measurements of loads, axial displacement, and pore pressures/volume change during the shearing process. The design of the restraint clamps was optimized using numerical simulations which showed that the sheared specimen includes shear modes that cannot currently be mobilized with available testing devices. Evaluation of the proposed device was done by testing soil specimens of Ottawa sand using both conventional and modified triaxial devices. The photogrammetry system was able to successfully capture the specimen deformations and these deformations are highly non-uniform in the new apparatus compared with those using a conventional triaxial device. SelfSim is employed to interpret Ottawa sand shear behavior. The tested sand specimens cover three different relative densities, and were tested under three different confining pressures. A soil-specific material constitutive model can be generated from this information. The constitutive model can then be directly used within a numerical analysis (e.g., finite element (FE) method) of a geotechnical problem.
Issue Date:2017-04-21
Type:Text
URI:http://hdl.handle.net/2142/97439
Rights Information:Copyright 2017 Randa Asmar
Date Available in IDEALS:2017-08-10
Date Deposited:2017-05


This item appears in the following Collection(s)

Item Statistics