Slurry flow simulation with 3D computational fluid dynamics (CFD)

This paper presents a methodology to simulate the slurry flow in a feed box that is designed to reduce the maintenance associated with the slurry flow, based on a 3D CFD model using ANSYS CFX. In this configuration, the slurry enters the feed box through the inlet pipe, it then fills the first depositing area and, when full, will overflow to the second depositing area before exiting through the outlet pipe (see Figure 1). The main focus of the simulation is to predict the wall surface where the slurry flow will impact and cause material erosion. Then, lining pads are installed at those specific locations to prevent the feed box from material loss, which reduces maintenance costs and downtime.

• Model geometry created in SolidWorks and imported to ANSYS
• Multi-zone meshing technology (see Figure 2) is applied, which generates high-quality hex-dominant mesh
• Average skewness was kept below 0.1 (0 represents the perfect condition) and average orthogonal quality was maintained > 0.9 (1 represents the perfect condition) to improve convergence

Model assumptions and methodology

This model is a three-phase Eulerian-Eulerian model with solid particles, water and air defined individually with different velocity fields:

• Continuous water phase
• Continuous air phase
• Disperse solid phase
• In-homogeneous flow and turbulence is set for three phases
• Each phase is coupled with two other phases through interfacial drag
• A generic yield stress and power-law model [1] is used for the slurry flow rheology

The variable slurry mixture viscosity is defined as follows:

• Set phasic viscosities separately so the mixture viscosity is given by the chosen correlation: μ₃r₃ + μ_Lr_L = μ_M
• Liquid-phase viscosity (μ_L) is set to be of water
• The solid-phase (sand) viscosity (μ₃) is a variable that will change based on its volume of fraction

This model uses a velocity inlet (gravity fed) and pressure out combination for boundary conditions for smoother convergence.

Model results

Model validation

For general model validation methodology, please refer to article “Optimizing the Design of Penstock Manifolds with 3D Computational Fluid Dynamics (CFD) Simulation”^[2]

In addition, some specific sensitivity cases are created for this application:

• 60% solid volume case (see Figure 7)
• Two-phase (air and slurry mix) modelling with equivalent slurry mix viscosity

Conclusion

Figure 8 shows the feed box condition after 3-month’s operation. From the initial observation, the erosion condition seems to align very well with the model prediction. A detailed inspection can be performed at a later date (maybe after a year of operation) to obtain more information to validate model outputs.

The above reveals a CFD modelling technique that simulates the slurry flow. This methodology can provide reliable and detailed information regarding flow characteristics and performance. It can be of great help for engineers when working on an innovative design for a complex problem.

References

[1] Mueller, S., et al, “The Rheology of Suspensions of Solid Particles”, Proc. R. Soc. A, 466, pp.1201-1228, 2010

[2]. Kerem Karakoc, “Optimizing the Design of Penstock Manifolds with 3D Computational Fluid Dynamics (CFD) Simulation”, BBA Publication, BBA, 2019