The interplay between field measurements and soil behavior for learning supported excavation response

In urban areas, estimation of ground movements due to excavation is critically important. In this thesis after a short review of currently used methods in practice for estimating excavation induced ground movements, a novel inverse analysis approach, self-learning in engineering simulation (SelfSim), is presented. SelfSim is applied to deep excavations in order to extract underlying soil behavior. The performance of the SelfSim inverse analysis is compared to inverse analysis based on a genetic algorithm. In the SelfSim approach, soil behavior is extracted from in situ measurements without a pre-defined constitutive model. In the genetic algorithm approach, soil parameters of an existing constitutive model are identified using field measurements. The performance of both techniques in capturing soil displacements and in predicting of soil behavior associated with the Lurie Center excavation in Chicago is presented. In order to demonstrate SelfSim’s capabilities in learning soil behavior using different instrument measurements, a simulated deep excavation is analyzed. The quality of the extracted behavior is examined by deploying different instrument configurations. The instruments required to provide sufficient information for SelfSim to extract soil behavior are identified. Then, some of the findings are further demonstrated in a case study of an excavation in Taipei soft clays. To illustrate that it is possible to learn from local experience and predict excavation performance in similar soil stratigraphy, case studies in Texas, Shanghai and Taipei are analyzed. The difficulties associated with the use of measured excavation response that is incompatible with recorded construction activity and the importance of engineering judgment in preparing measurement data for inverse analysis are highlighted. Finally, it is shown that the 2D extracted soil behavior of excavation in Chicago clays can not provide a reasonable excavation performance for an elevated ground surface excavation in Chicago suburbs within similar soil stratigraphy. It is demonstrated that the 3D effects of excavation are captured via 3D modeling using SelfSim. At the end, the extracted soil behavior from 3D analysis is discussed and compared to extracted soil behavior from 2D analysis.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Civil Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2010

Author dc:creator

• Osouli, Abdolreza

Contributors dc:contributor

• Long, James H.
• Olson, Scott M.
• Ghaboussi, Jamshid
• Hashash, Youssef M.

• Excavation
• soil behavior
• Numerical modeling
• Neuralnetwork
• constitutive models
• in situ measurements

Statement dc:rights

• Copyright 2010 Abdolreza Osouli

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/15589

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F15589