Computational fluid dynamics simulations of fully developed turbulent flows with Newtonian and non-Newtonian drag reduction fluids in the cylindrical geometry

Abstract

While turbulent flows with the dynamic and sinuous flow pattern have a wide variety of engineering applications in enhancing fluid mixing and heat/mass transfer, they unfavorably lead to dramatic increases in friction drags that need to be eased. The polymer induced drag reduction, first discovered by Toms in 1948, is one of the most efficient techniques to achieve so, which has the capability of reducing the friction drag in turbulent flows by up to 80%. As a result, it has been broadly used for anticorrosion in oil and gas conduits, energy loss prevention in irrigation, water heating/cooling systems and sewer systems, improving hydraulic fracturing in reservoir engineering, and enhancing cutting transport for extended-reach well in drilling engineering. Despite being studied for over 70 years, many problems in this research area are still waiting to be solved. In particular, to accurately predict drag reduction numerically remains challenging, especially in engineering problems where flows with large Reynolds numbers in complex geometries widely exist. Consequently, in this study, a reliable numerical approach has been proposed in the computationally inexpensive Reynolds-averaged-Navier-Stokes (RANS) framework to estimate polymer induced drag reduction in the cylindrical geometry, with the rheological behavior of polymer solutions represented by the finitely-extensible nonlinear-elastic model with Peterlin’s function (FENE-P model) and the turbulent flow field characterized by the k-e-v²-f model, both coded into the commercial software of FLUENT using the User-Defined-Function (UDF). ANSYS 14.0 is used to complete all the simulations, CFX for the base case Newtonian flow and FLUENT for the drag reduction flow. By using the correlations suggested in this study to determine the essential rheological parameters, simulation results are validated successfully against experimental studies, showing the robust performance of the proposed model to predict accurate drag reduction in the cylindrical geometry for solutions with both rigid and non-rigid flexible polymers. By comparing behaviors of the non-Newtonian drag reduction and the Newtonian fluids in the concentric annulus, where the geometrical transverse curvature effect plays a vital role in determining the flow field, polymers are found to behave differently close to the inner and outer walls of the annulus, leading to more decay of turbulence at the outer wall than the inner. Such a phenomenon has been explained by relating the elongation of the polymer chains in the solution to the intensity of the flow field they experience, which is also found in this study to be highly dependent on the inherent rheological properties of themselves. This thesis provides a benchmark study about how the polymer induced turbulent drag reduction in the cylindrical geometry can be numerically estimated using RANS modelling. Substantial progress has also been made in understanding the geometry-dependent behavior of polymers in the turbulent flow, which could inspire research on broader applications of the polymer induced drag reduction in more complex geometries.