Résumé - Abstract
Extraction and processing of oil sands generates extensive volumes of tailings slurry that is treated and deposited into tailings ponds. The deposited material settles down over time due to self weight consolidation. Several uncertain aspects are involved in tailings settlements due to the variable properties of the tailings and hence, differential settlement is expected to occur in different areas of the ponds. Estimating the future surface of a consolidated pond and the differential settlement is an essential input in reclamation phase and closure landform design. This paper presents an integrated methodology that simulates the one-dimensional (1D) settlement of a column of tailings based on a large strain consolidation model in presence of uncertainty, assigns the probabilistic interface heights to several points on the pond surface, generates a quasi-three-dimensional (3D) surface of the pond using inverse distance weighting (IDW) and kriging, and finally calculates the differential settlements over the consolidated pond surface. The 1D large strain consolidation and the surface contours are developed in GoldSim using Monte-Carlo techniques. A series of scenarios are performed to study the sensitivity of the model to different parameters, investigate the performance of kriging versus IDW in calculating the consolidated surface, and the performance of the experimental tables for void ratio and hydraulic conductivity against the analytical constitutive relationships in simulating the height interface.