Continuous water-assisted flow (CWAF), where a water layer surrounds a viscous oil core, provides low energy, long distance transport of heavy oil and bitumen without requiring heating or solvent addition. In industrial applications of CWAF, the pipe wall is fouled by a thin coating of oil, an effect not considered in many studies of water-lubricated pipe flows. In the present study, a new method to model pressure loss in the water-assisted pipeline flow of heavy oil is introduced. The hydrodynamic effects produced by the wall-fouling layer are incorporated in the model as input parameters for CFD simulations. The most important of these parameters are the thickness of the wall-fouling layer and the equivalent hydrodynamic roughness it produces. The CFD methodology described here was developed on the ANSYS-CFX platform and is able to capture the effects of the wall-fouling layer, the hydrodynamic roughness produced by this layer, and the water hold-up. The new CFD model was validated using previously collected data from tests conducted in two separate pipeline loops (100 and 260 mm in diameter), using a range of oil viscosities, water fractions, and mixture velocities. Compared to existing models, the one presented here provides more accurate predictions and requires significantly fewer computing resources. Because the model was developed using a physics-based approach, it is a useful tool in evaluating the effects of pipe diameter, oil viscosity (or temperature), water cut, and mixture velocity on pressure losses in water-assisted heavy oil pipelines.

In water-lubricated pipeline transportation of heavy oil and bitumen, a thin oil film typically coats the pipe wall. A detailed study of the hydrodynamic effects of this fouling layer is critical to the design and operation of oil-water pipelines, as it can increase the pipeline pressure loss (and pumping power requirements) by 15 times or more. In this study, a parametric investigation of the hydrodynamic effects caused by the wall coating of viscous oil was conducted. A custom-built rectangular flow cell was used. A validated CFD-based procedure was used to determine the hydrodynamic roughness from the measured pressure losses. A similar procedure was followed for a set of pipe loop tests. The effects of the thickness of the oil coating layer, the oil viscosity, and water flow rate on the hydrodynamic roughness were evaluated. Oil viscosities from 3 to 21300 Pa s were tested. The results show that the equivalent hydrodynamic roughness produced by a wall coating layer of viscous oil is dependent on the coating thickness but essentially independent of oil viscosity. A new correlation was developed using these data to predict the hydrodynamic roughness for flow conditions in which a viscous oil coating is produced on the pipe wall.Keywords Pipeline transportation Á Heavy oil Á Wall fouling Á Lubricated pipe flow Á CFD simulation List of symbols

Applications of machine learning algorithms (MLAs) to modeling the adsorption efficiencies of different heavy metals have been limited by the adsorbate–adsorbent pair and the selection of specific MLAs. In the current study, adsorption efficiencies of fourteen heavy metal–adsorbent (HM-AD) pairs were modeled with a variety of ML models such as support vector regression with polynomial and radial basis function kernels, random forest (RF), stochastic gradient boosting, and bayesian additive regression tree (BART). The wet experiment-based actual measurements were supplemented with synthetic data samples. The first batch of dry experiments was performed to model the removal efficiency of an HM with a specific AD. The ML modeling was then implemented on the whole dataset to develop a generalized model. A ten-fold cross-validation method was used for the model selection, while the comparative performance of the MLAs was evaluated with statistical metrics comprising Spearman’s rank correlation coefficient, coefficient of determination (R2), mean absolute error, and root-mean-squared-error. The regression tree methods, BART, and RF demonstrated the most robust and optimum performance with 0.96 ⫹ R2 ⫹ 0.99. The current study provides a generalized methodology to implement ML in modeling the efficiency of not only a specific adsorption process but also a group of comparable processes involving multiple HM-AD pairs.

No abstract