Recent pilots on emerging oil sands tailings technologies have confirmed that deposit thickness is an important parameter controlling tailings dewatering, and a key parameter governing cost. Controlling or managing deposit thickness continues to be an important challenge for full-scale implementation of tailings technologies that exhibit a yield stress. From the perspective of deterministic modelling of such deposits, there are many challenges, including measurement of the relevant rheological parameters, and how to handle time-variant rheology when modelling.
This research characterizes the rheology of a mineral slurry with relatively high clay content, which is treated with a high molecular weight anionic polymer to induce flocculation. Rheometry results showed that while flocs break down under high shear, flocs reform at lower shear rates. Breakdown and recovery of flocs was confirmed by measuring the shear modulus under dynamic loading and a set of microstructural analysis. Moreover, it was shown that the tailings manifest viscosity bifurcation behaviour similar to pure clay, including shear history dependent apparent yield stress. The measured rheology was then modeled using a previously published viscosity bifurcation model that accounts for hysteresis in the apparent yield stress.
The rheology results are used semi-quantitatively to explain deposition rate dependent behaviour seen in flume tests. The geometry of tailings in flume tests with relatively slow deposition is affected by the behaviour of earliest deposited tailings, which appear to have recovered structure sufficiently to manifest a large yield stress. This yield stress is much larger than the yield stress exhibited by tailings when they initially come to rest. This full recovery of the yield stress seems to be particularly important to managing surface deposition, as zones of tailings that have stopped moving substantially steepen the slope of deposits near the deposition point.
Finally, and using the rheological models obtained, an attempt was made to model such flows at bench and pilot scales using 3D/2D numerical simulations. The flume test and field deposition conducted were simulated using CFX. It is found that using the lower limit yield stress value and by conducting simulation in several stages (to account for ageing behaviour), more realistic results could be obtained.