DATA-DRIVEN IDENTIFICATION METHODOLOGY FOR GEOTECHNICAL STRUCTURES USING LONG-TERM FIELD DATAS

摘要：

A non-parametric, data-driven methodology for monitoring geotechnical structures subject to long-term environmental variations is presented. The non-parametric monitoring methodology, unlike conventional parametric approach, does not require physical assumptions and excessive simplification in modeling, even with a very limited amount of sensor data for inverse analysis. Using three non-parametric data processing techniques, Principal Component Analysis (PCA), Empirical Mode Decomposition (EMD), and Hilbert-Huang Transforms (HHT), a full-scale retaining wall was studied to demonstrate the validity of the non-parametric methodology. Three tilt gauges were positioned at top, middle and bottom of the wall, and the wall behaviors were monitored for three years. The analysis results indicate that important performance-related information, such as the wall drainage system, can be obtained even with very limited data that were significantly influenced with various environmental effects.

作者: Hae-Bum Yun Ganesh Sundaresan Yoonhwak Kim Lakshmi N.Reddi

作者单位: Department of Civil,Environmental and Construction Engineering,University of Central Florida

会议类型: 国际会议

会议名称: The International Symposium on Innovation & Sustainability of Structures in Civil Engineering(2011年土木工程结构创新与可持续发展国际会议)

会议地点: 厦门

会议语种:英文

页码: 405-411

在线出版日期: 2011-10-28（万方平台首次上网日期，不代表论文的发表时间）