Prediction of Compression and Permeability Characteristics of Mine Tailings Using Natural Computation and Large-Strain Consolidation Framework

Prediction of consolidation behavior of soft soil deposits, such as mine tailings and hydraulic fills, is an essential and routine exercise in mining and dredging industries using large-strain consolidation theory. The values of compressibility and hydraulic conductivity deduced from settlement vs. time response are influenced by the selected form of e – log σ' and e – k relationship. In this paper, we propose a method of back-analysis to deduce compressibility and hydraulic conductivity of soft soils from its settlement vs. time response. The curve-fitting parameters for the nonlinear relations between e – log σ' and e – k are inferred using the settlement response and the initial conditions of the soil for the cases of self-weight consolidation of oil sands mature fine tailings (MFT). A piecewise-linear model of large-strain consolidation is used. The inverse analysis is carried out using natural computation algorithms. It is demonstrated that accurate deduction of compressibility and hydraulic conductivity is possible using natural computation algorithms and that these algorithms are much more stable ad accurately predicts the parameters.