Location

Arlington, Virginia

Session Start Date

8-11-2008

Session End Date

8-16-2008

Abstract

In this paper it is shown how the knowledge embedded in case histories can be used to explicate some of the uncertainties contributing to the gap between theory and practice. With the help of computational intelligence techniques, collections of case histories in data-bases, as a type of collective memory of the geotechnical profession may be explored to turn this memory into collective brains in geo-technics: a GeoBrain. Regarding the scarcity of soil investigation data and the translation of the available data into a model, the ‘schematization factor’ has been introduced as a partial safety factor to account for the influence of data availability and the role of human expertise. Using a database of increasing size on the feasibility of installing sheet pile walls, the determination of optimal parameter values for prediction models is illustrated. It is shown that computational intelligence techniques like Bayesian Belief Networks and Genetic Algorithms can be very helpful to improve predictions of what is likely to happen in geotechnical practice.

Department(s)

Civil, Architectural and Environmental Engineering

Appears In

International Conference on Case Histories in Geotechnical Engineering

Meeting Name

Sixth Conference

Publisher

Missouri University of Science and Technology

Publication Date

8-11-2008

Document Version

Final Version

Rights

© 2008 Missouri University of Science and Technology, All rights reserved.

Document Type

Article - Conference proceedings

File Type

text

Language

English

Share

 
COinS
 
Aug 11th, 12:00 AM Aug 16th, 12:00 AM

Geobrain: Dutch Feasibility Database for Installing Sheet Pile Walls

Arlington, Virginia

In this paper it is shown how the knowledge embedded in case histories can be used to explicate some of the uncertainties contributing to the gap between theory and practice. With the help of computational intelligence techniques, collections of case histories in data-bases, as a type of collective memory of the geotechnical profession may be explored to turn this memory into collective brains in geo-technics: a GeoBrain. Regarding the scarcity of soil investigation data and the translation of the available data into a model, the ‘schematization factor’ has been introduced as a partial safety factor to account for the influence of data availability and the role of human expertise. Using a database of increasing size on the feasibility of installing sheet pile walls, the determination of optimal parameter values for prediction models is illustrated. It is shown that computational intelligence techniques like Bayesian Belief Networks and Genetic Algorithms can be very helpful to improve predictions of what is likely to happen in geotechnical practice.