Resonant Column And Cyclic Triaxial Testing Of Tailing Dam Material

Abstract

A series of resonant column and cyclic triaxial tests has been conducted in the frame of the analysis of tailing dam stability during earthquakes. The investigation program for a silty sand from uranium tailings is presented. The paper describes the testing procedures and presents all significant results of these experiments. Single-stage and multi-stage resonant column tests were performed in order to determine the dependence of shear modulus and damping ratio on confining pressure and shear strain amplitude. Stress-controlled cyclic triaxial tests on the same material were conducted to determine the increase in pore pressure and the amount of accumulated strain in dependence of the cycles number and the load configuration. Approximate formulas are given for stiffness and damping to be used in seismic response analyses. INTRODUCTION When applying finite element methods for the seismic response analyses of dams and embankments, e.g. Seed*, the nonlinear soil behaviour can be taken into account by use of equivalent elastic constants which vary as a function of stress and strain level. These parameters are obtained either from empirical relations, e.g. Hardin & Drnevich^, or from laboratory tests on the particular soil material. Much of the published data deal with soils such as poorly graded sands, principal of alluvial origin. However, for soils that are dirty, gravelly, well graded or composed of relatively angular particles only a limited amount of test data is available in the Transactions on the Built Environment vol 3, © 1993 WIT Press, www.witpress.com, ISSN 1743-3509 292 Soil Dynamics and Earthquake Engineering literature. In the frame of the stability analysis of tailing dams located in a moderate seismic risk region in the south-east part of Germany an extensive dynamic in-situ and laboratory investigation testing program was conducted. Since the tailings sand was placed hydraulically with little compaction the strength reduction due to pore pressure increase had also to be considered. In the following, results of resonant column and cyclic triaxial tests are presented which were performed on samples of sand and silty sand from uranium tailings. MATERIAL TESTED The material tested was obtained from borings drilled in the crest or slopes of a major tailings dam. It consisted of sub-angular silty sand and had the following properties: i) grain size distribution: D$Q = 0.2 mm, Dso = 0.08-0.12 mm and £>go = 0-5 mm, ii) particle density 2.77 Mg/nr* and iii) effective strength parameters <^' = 31°, d = 5 kN/m^. The dry density pd varied between 1.59 and 1.82 Mg/m^ corresponding to void ratios e from 0.74 to 0.52. All specimens, 50 mm in diameter and 100 mm in height, were formed in an unsaturated state corresponding to the in situ water content and compacted to the required density. RESONANT COLUMN TESTS The tests were conducted on a Drnevich type apparatus, cf. Drnevich et al according to the specification of the ASTM D4015 87 Standard. All specimens were isotropically consolidated using standard techniques. Multi-stage tests were conducted to determine the shear modulus and damping ratio, Gmax and Dmin, respectively, at very low shear strain amplitudes 7. The test was started at the lowest confining pressure and repeated for intermediate pressures up to a maximum confining pressure of 500 kPa. At the highest confining pressure under consideration the amplitude of excitation was gradually increased to study the shear modulus G as a function of shear strain amplitude. Damping was determined by the steady-state vibration of the specimen. Spot checks were made by the free vibration decay method. In addition, single-stage tests were performed at selected levels of confining pressure by stepwise increasing the excitation force up to a maximum shear strain amplitude of approximately 10~~*%. Densities and confining pressures for each test are summarized in Table 1, where the confining pressure <JQ is normalized with respect to the atmospheric pressure pa = 100 kPa. The measured shear moduli Gmax at small strains of approximately 10~* % are depicted in Figure 1 as a function of confining pressure <JQ. Shear modulus reduction curves at two confining pressures are presented in Figure 2 versus shear strain. Transactions on the Built Environment vol 3, © 1993 WIT Press, www.witpress.com, ISSN 1743-3509