OFR-161(1)-82 Probabilistic Modeling Of Tailings Embankment Designs - Volume I - Model Development And Verification

Traditionally, an evaluation of the safety of slopes has been based on computing a safety factor against failure. In computing the safety factor, the geometry of the slope, the soil parameters and the pore pressures are treated as deterministic quantities even though they are known to be random variables. A pseudo-three-dimensional probabilistic slope stability model has been developed that treats shear strength as a random variable. The model uses the probability of a slope failure as an assessment of slope reliability. The model can accommodate zoned embankments of soil in which the strength is described by the Mohr-Coulomb strength envelope. Autocorrelation functions are used to describe the spatial correlation of the strength parameters, c and tan ø. Model error can also be included in the analysis. The model has been applied to an existing tailings dam and to several hypothetical examples. These examples are presented to illustrate the influence of the choice of the soil strength variability parameters on the probability of failure. The results show that the critical failure surface based on the minimum safety factor is not necessarily the failure surface that will yield the maximum probability of failure. The analysis of the tailings dam shows that variance reduction can be large and that the model error is the dominant source of uncertainty for the embankment analyzed.