Dense slurry flow is a feature of the oil sands operation. Modelling of erosion in dense slurry flow under oil sands process conditions is challenging. Although several erosion models are currently used for upstream produced sand application under very dilute sand conditions, extension of those models to dense slurry flow such as conditions relevant to oil sands is highly uncertain. The objective of this study is to develop predictive wear model for dense slurry flow to narrow the gap.
An integrated approach was developed to model the wear in oil sands Coarse Tailings (CT) slurry pipeline. Three techniques, including pilot-scale flow loop experiments, computational fluid dynamics (CFD) simulations and field trial, were jointly used to aid the development of a reliable predictive tool. By collaborating with vendors, a high-resolution, non-intrusive, erosion monitoring system based on ultrasonic technology (UT) was developed and implemented in the flow loop experiments. A data mining analysis based on random forest algorithm was applied to the field trial data to develop a predictive wear model for CT pipelines. Both the Eulerian-Granular and Eulerian-Lagrangian methods were explored in CFD simulations for dense slurry flow in long and large horizontal pipes. The CFD erosion model was calibrated based on the field trial data and validated by the flow loop tests. The effects of critical variables affecting the wear were investigated, and a predictive tool was developed. The modelling tool is capable of predicting erosion rates due to changes in the piping design and operating conditions. The model can help the operator adjust process conditions to minimize wear and optimize inspection and maintenance schedule. This paper summarizes the findings from the various techniques adopted in this study and their limitations.
Key words: wear modelling, oil sands, dense slurry flow, erosion, CFD simulation, data mining analysis