Effects of fluid viscosity in liquid-solid flow pattern transition in horizontal pipelines

Abstract

Cuttings removal is a problem of significant concern in drilling operations since poor hole cleaning may lead to several operational problems, such as reducing bits life and stuck pipe. Different configurations may be observed during liquid-solid flows, known as flow patterns, and it is essential to understand the conditions in which each configuration is formed, aiming to avoid those operational problems. An experimental rig was designed to develop a flow pattern map and was be assembled in CERNN/UTFPR. Three working fluids were employed: water, water-glycerin 40%wt, and water-glycerin 60%wt mixtures allowing a viscosity variation over the experiments. Urea particles of the average size of 1.44 mm and a specific mass of 1480 kg/m³ were used as the discreet phase. The evaluation of the flow patterns was developed through pressure drop measurement and flow visualization. Based on the flow pattern map, it was stated that the viscosity indeed affects the flow pattern transition. The general statement was that higher viscosity fluids required smaller mean velocity for the transition, which happened at lower pressure drop values. The effect described is more significant for the transition between the moving bed and heterogeneous flows than for stationary bed to moving bed transition. The influence is intensified for the first case due to the bed erosion of flat particles. The water could not suspend particles through impact erosion due to the small cross-section area of the flatshaped particles, which is the most relevant bed erosion cause for low viscosity fluids. On the other hand, the high viscosity intensifies the shear erosion, which is not greatly affected by the flat-shaped particles. Finally, a correlation was developed, allowing the pressure drop determination as a function of the Reynolds number and the fluids’ kinematic viscosity. The results of the correlation were in satisfactory agreement with experimental data.