Comparison of Predicted and Measured Settlement of a Test Embankment Over Soft Soil

Chang, Kuo-Hsia; Kovacs, William D.; Wu, Ming-Jiun; ,

Comparison of Predicted and Measured Settlement of a Test Embankment Over Soft Soil

by Kuo-Hsia Chang, (M.ASCE), Shannon & Wilson, Inc, Seattle, United States,
William D. Kovacs, (M.ASCE), Shannon & Wilson, Inc, Seattle, United States,
Ming-Jiun Wu, (M.ASCE), Shannon & Wilson, Inc, Seattle, United States,

Document Type: Proceeding Paper

Part of: Vertical and Horizontal Deformations of Foundations and Embankments

Abstract:

A test embankment was instrumented with six (6) settlement plates, three (3) vertical extensometers, two (2) inclinometers, and two (2) horizontal profilers in order to verify settlement predictions for a highway project in Tacoma, WA. The subsoil consists of about 13 meters of soft lacustrine clayey silt and loose silty sand underlain by over 43 meters of alternate layers of medium and dense alluvial silty sand resting on Till. The data obtained offered an unique opportunity to compare different models for their capabilities to predict both the maximum and differential settlements in very heterogeneous soils. Under such conditions, the cone-penetration test seemed to be the most valuable basis for predicting maximum settlements when there was sufficient local experience to establish empirical relationships between the tip resistance and modulus (Schmertmann 1970), regardless of which model was used. After comparing the three most commonly used models in predicting differential settlements for soil-structure interaction analyses, it was found that in deep alluvial deposits, a modified semi-infinite elastic model assuming that Boussinesq stresses was still valid in a layered soil with variable Young's moduli and limited depth seemed to be better than either the Winkler or the Terzaghi-Taylor model.

Return to search