Comparison of Liquefied Shear Strength of Tailings Estimated using Standard Empirical Methods and Post-cyclic Direct Simple Shear Tests

摘要

There are a large number of upstream raised tailings dams across Canada and worldwide. While many of these are historical and closed, there are still several operating tailings dams which are being raised by upstream construction method mainly due to its lower cost and as continuation of previous dam raising. Static and seismic liquefaction of foundation tailings is a serious concern in these dams. The tailings in these facilities are generally deposited by means of spigotting, and thus are generally in a loose to very loose state, and highly susceptible to liquefaction. The post-liquefaction stability and remediation design of these dams almost entirely depend on the estimated liquefied shear strength (Sr) of the tailings. The Sr of tailings is generally estimated using empirical methods utilizing in-situ penetration test data such as SPT and CPT data. These correlations have been developed based on back analyses of case histories involving mostly natural soils. It is routine practice to use these same methods to estimate Sr of tailings as well. While some practitioners use clean sand equivalent penetration resistance data with these charts to correct for the fines content, others caution that using fines content correction could provide non-conservative Sr. This paper presents data and results from two different tailings dam sites and compares Sr estimated using empirical methods with clean sand equivalent penetration resistance, (ii) empirical methods without correction for fines content, and (iii) post-cyclic direct simple shear (DSS) testing of Shelby tube piston samples and reconstituted samples. A relatively new procedure, involving unload-reload loops, is introduced to study the post-cyclic shear response at large strain and to assess Sr values in laboratory DSS tests. It has been demonstrated that unlike the routine post-cyclic DSS test data, which invariably provides much higher Sr values compared to in-situ test based empirical correlations, the Sr values determined in post-cyclic DSS tests using this procedure provide Sr comparable with the empirical methods. It is recommended to exercise caution on relying on Sr values from routine post-cyclic DSS tests results and to explore the post-cyclic response and Sr using the proposed post-cyclic DSS test procedure. More research and data is needed before adopting the proposed procedure as a method for design.