Computational Prediction Of Particle-Laden Slurry Flow In A Vertical Pipe Using Reynolds Stress Model

Correct interpretation of optical measurements of concentration and composition in slurry flows in pipes depends upon an understanding of the solids distribution across the pipe. The distributions are not necessarily uniform. Computational Fluid Dynamics (CFD) was used to evaluate the deviations from concentration uniformity for slurry flow in a vertical pipe. The slurry was composed of Xylene-2 aminoliquid and 4, 6 dimethyl pyrimidine (ADP) solid. Computations were performed for flow in a 0.0508 m diameter pipe with velocities of 2 and 4 m/s. Solids volume concentrations of 10 to 30% and particle diameters between 50 and 780 pm were studied. The corresponding Reynolds numbers were 1.43E05 and 2.86E05. ANSYS FLUENT 13.0 CFD software was used with the Reynolds stress turbulence model and an enhanced wall treatment. The Euler-Euler multiphase model was used for the slurry. A mesh of 1.8 million cells was chosen using a grid independence study. The, modeling approach was validated by comparing concentration and velocity profiles in the fully developed region to experimental data in the literature. In addition, comparisons were made to computations performed using k-epsilon turbulence models.When the volumetric concentration of the ADP particles was less than 10%, their presence did not have a significant effect on the velocity and concentration profiles of the slurry. For all mean concentrations, particles with diameters less than 160 gm showed negligible effects on the velocity and concentration profiles of the slurry flow. The velocity profiles almost matched the velocity profile of the flow with xylene alone and the concentration profiles were uniform across the pipe with very small effects in the near wall region. For particle diameters above 160 microns and concentrations above 10%, the velocity profile had a tendency to become flatter as the particle size and concentration increased. The concentration profiles showed that particles tended to move towards the center of the pipe and away from the wall as particle size and particle concentration increased. These near wall effects were found to be much smaller using the k-epsilon models, whether ANSYS FLUENT's Standard, Realizable, or RNG versions. It can be concluded that the effects of the wall on particles can extend beyond what is considered the near wall region for velocity profiles in flows without particles.