Finite Difference Analyses for Sand Drain Problems

by Roy E. Olson, (F.ASCE), Professor of Civil Engineering; University of Texas, Austin, TX,
David E. Daniel, (A.M.ASCE), Staff Engineer; Woodward-Lundgren, Oakland, CA,
Thomas K. Liu, (F.ASCE), Partner; Haley and Aldrich, Cambridge, MA,

Part of: From Soil Behavior Fundamentals to Innovations in Geotechnical Engineering: Honoring Roy E. Olson


The geotechnical literature contains a number of papers in which sand drain installations are described and field performance is compared with theoretical predictions. Some of the field experiences suggest that the design theories in general use correlate rather well with field measurements (19, 20, 22, 23, 27, 28, 30, 31, 49, 55, 58, 59) whereas others indicate major discrepancies (24, 29, 32, 33, 36, 50, 57, 60). Unfortunately, the theories in general use (5, 6, 9, 35) contain simplifying assumptions not fully satisfied by any of the case histories published so far. In addition, many of the case histories contain important uncertainties regarding actual field conditions and nearly all are deficient in terms of well-defined soil properties. In a few cases where excellent correlations were obtained between field measurements and theoretical predictions, attempts by the authors to re-analyze the results led to considerably poorer correlations than those reported in the original publication. The result of the uncertainties in theory and field case histories is that a designer cannot be certain that his embankment will perform in the manner predicted by his analyses. Studies of the available case histories indicate clearly that improvements are needed in both the methods of analysis and in the techniques used for measurement of soil properties. In particular, the analysis should account for layered systems, two-dimensional water flow, time-dependent loading with the load dependent on settlement, partially saturated soils, soil properties that are functions of effective stress, nonlinear stress-settlement curves, and secondary effects. The major defect in measurement of soil properties is clearly the lack of tests in which the soil is allowed to drain radially and in which the samples are large enough to include representative stratification (17, 37, 38, 39, 40, 48). The authors have developed a finite difference solution that accounts for some of the deficiencies in theories currently in use. The method has been applied to one case history and found to be successful in predicting the rates of settlement in the field. The data for the case history will be presented first in order to clarify the variables that must be included in the analyses. Then the analytical technique will be explained and applied to the case history.


Return to search