Modeling of yield-power-law fluid flow in a partially blocked concentric annulus

Horizontal well is one of the most effective ways of developing unconventional oil and natural gas reservoirs. During horizontal well drilling, a stationary cuttings bed forms on the low-side of the well bore. Formation of the bed is undesirable and has detrimental effect on wellbore hydraulic and drilling performance. Hydraulic optimization of the process strongly depends on accurate prediction of bed height and frictional pressure loss. Therefore, it is necessary to investigate hydraulic characteristics of horizontal welibore with cuttings bed, which can be represented by partially blocked concentric annulus when centralizers are used with drilipipe. In this study, flow in a partially blocked concentric annulus is investigated using Computational Fluid Dynamics (CFD) approach. CFD simulation results are validated using established numerical solutions and experimental measurements obtained from concentric annulus. The analysis is performed for laminar flow of yield-power-law (YPL) fluid. Pressure loss and wall shear stress acting on the cuttings bed are obtained from CFD simulation. Results show significant variation in bed shear stress in the lateral direction. Hence, average bed shear stress ((tau) over bar (bed)) is used to analyze the result. Unexpected (tau) over bar (bed) trend is observed when bed height is increased at constant flow rate. As bed height increases, (tau) over bar (bed) bed slightly decreases at low bed heights until it reaches the inner pipe. For bed height greater than annular clearance, (tau) over bar (bed)trend is normal, it increases with bed height. Existing hole cleaning models require accurate prediction of (tau) over bar (bed),and pressure loss in partially blocked concentric annuli. Therefore, combining existing non-circular duct modeling approach and simulation data, simplified models are developed to predict pressure loss and (tau) over bar (bed). Model predictions show reasonable agreement with simulation results. Maximum discrepancy of +/- 10% has been observed. (C) 2016 Elsevier B.V. All rights reserved.