Please use this identifier to cite or link to this item:
Title: Investigation of density variation in triaxial test specimens of cohesionless soil subjected to cyclic and monotonic loading
Authors: Gilbert, P. A.
Keywords: Soils
Soil dynamics
Soil mechanics
Cohesionless soils
Soil density
Dynamic testing
Soil freezing
Publisher: Geotechnical Laboratory (U.S.)
Engineer Research and Development Center (U.S.)
Series/Report no.: Technical report (U.S. Army Engineer Waterways Experiment Station) ; GL-84-10.
Description: Technical report
Summary: The undrained cyclic triaxial test is widely used to evaluate the seismically induced liquefaction potential of saturated cohesionless soil. Concern that internal void ratio changes (redistribution) due to testing characteristics may render the test inappropriate for dynamic analysis prompted research to investigate such changes. This report describes equipment and procedures to: a.) Prepare sand specimens of known average density or void ratio to a high degree of uniformity. b.) Load specimens under undrained (constant volume) conditions to various pore pressure and strain responses. c.) Freeze specimens under test conditions in such manner that the soil skeleton is not disturbed by the freezing process. d.) Dissect frozen specimens to establish spatial density distribution. e.) Conduct control tests involving steps a, c, and d only. This research required specimens with a higher degree of density uniformity than previously demonstrated. A complex trial-and-error laboratory study was conducted to develop equipment and procedures to construct the high-quality specimens. Additionally, special freezing techniques were required to produce an "undisturbed" frozen specimen. This freezing process and the behavior of the triaxial test specimen during freezing are documented. Control specimens with 100 percent water saturation (B factor of 0.96 or more) were tested in the triaxial chamber, frozen under back pressure and confining pressure with the top drainage line open and the cold temperature source at the base of the specimen, and then dissected into 96 elements in a cold room. The density of each segment and, consequently, the density distribution of the specimen were determined from the ice content. Homogeneity, that is, relative density uniformity, was quantified in terms of the average relative density determined for the 96 elements. Because of the requirement for precise density determination, errors caused by sublimation and measurement uncertainty were examined and are discussed in this report. Relative density variation caused by undrained cyclic or monotonic deviatoric loading is documented for initially uniform specimens which were interrupted at various stages of cyclic and monotonic loading, frozen, dissected, and analyzed. Relative density dispersion with increase in strain level is shown at three densities, approximately 40, 60, and 70 percent relative density. The spatial changes in density occurring as a result of cyclic loading are observed and indexed in terms of changes in standard deviation of all dissected elements from the average when compared to control tests described in step e above. Density redistribution as a result of cyclic and monotonic loading is irrefutably demonstrated and quantified in test specimens of Banding sand, which is a specific gradation of Ottawa sand. The study demonstrates clearly that a highly repeatable average density from specimen to specimen is not an indication of a high degree of specimen density uniformity. It was also demonstrated that uniform specimens are stronger and more stable under cyclic loading than nonuniform specimens. It was shown that density redistribution begins at pore pressure responses less than 100 percent, but does not become significant until nearly 100 percent pore pressure response or high peak-to-peak strain levels (greater than 5 percent) are reached. The work is compared with the work of others who have conducted related studies. NOTE: This file is large. Allow your browser several minutes to download the file
Rights: Approved for public release; distribution is unlimited.
Appears in Collections:Technical Report

Files in This Item:
File Description SizeFormat 
TR-GL-84-10.pdf22.68 MBAdobe PDFThumbnail