Optimize Solids Cyclone Separation with Simulation

Shawn Wasserman posted 2015-06-10

Phenomena

  • Solid Cyclone Separators

Applications

  • Manufacturing
  • Oil and gas
  • Construction
  • Energy
  • Pharmaceutical
  • Materials and chemical processing
  • Food processing

Main software

  • ANSYS Fluent 

Models

  • Discrete Element Model
  • Dense Discrete Phase Model 
  • Dispersed RSM Turbulence Model 

Analysis Type

  • CFD

Computing Power

  • 2 – 4 hours on eight compute nodes for single simulation

Mesh

  • Block-structured hex mesh
  • 72,000 cells

Findings

  • Pressure drop predicted until 1:1 air/particles mass ratio 
  • Increase solid loading reduces particle rotations 
  • Friction coefficient effects particle flow more than fluid flow

Solid Cyclone Separators in Industry

Solid cyclone separators are typically used in manufacturing to improve air quality or product recovery. The technology relies on the different densities between the solid particles and carrying fluid.  

Cyclone separator geometry. Gas with dust particles enter at the side. Gas will exit at the top while the particles fall to the bottom.


As the carrying gas enters the cyclone, the shape of the unit forces the fluid-particle mixture to swirl into a vortex. The centrifugal force caused by the increased radial acceleration, as well as the varying densities, leads the particles to impinge onto the walls. Once the particles hit the wall, they lose momentum and the solids fall out of the carrying fluid.


Regardless of industry, cyclone separators typically have the same design goals:

  • Improved product recovery
  • Higher separation for cleaner carrying fluid
  • Consistent separation for various operating conditions
  • Optimized fluid flow
  • Reduced pressure drop and pump loads
  • Reduced corrosion/erosion and thermal loading
  • Low maintenance


“CFD simulation can play a crucial role to achieve many of these design goals for a cyclone separator,” said Shitanshu Gohel, technical specialist at ANSYS. “CFD can provide insight into understanding the complex fluid and particulate flow process and can also evaluate various performance parameters.”

Benefits of Simulating Solid Cyclone Separators

The main benefit of simulation is that it reduces trial-and-error and expensive prototyping.


Engineers can use the simulations to explore different operating and geometry conditions in virtual environments on their desktops.  


“Virtual design of experiment can be studied to optimize various parameters of interest, such as pressure drop and separation efficiency,” clarified Gohel. “Simulations can also be used to conduct what-if type studies for geometry and operating condition changes while troubleshooting a sub-optimal performance.”

How to Simulate a Vortex in ANSYS Fluent

Simulation of the particle trajectory. As the particles hit the wall, gravity forces them down. Color based on particle resistance time.  

Gohel said ANSYS has many years of experience modeling multiphase industrial turbulent flows like the ones seen in a cyclone separator.

The flow at the vortex core exhibits  anisotropic turbulence. Therefore, a Reynolds stress model (RSM) is recommended as it will resolve the six Reynolds stress terms by time averaging the Navier-Stokes equation.


“Multiphase turbulence models are the extension of the ones available for single phase flow,” explained Gohel. “Depending on the model, there are various options available for the user to choose from such as mixture, dispersed or per-phase.”


He added, “Dispersed option is generally chosen in cases of relatively low solid loading and the relatively higher particle diameter. In such situations, primary phase turbulence is regarded as the dominant process. Dispersed option solves the transport equations for the continuous phase and predicts turbulence quantities for the dispersed phase based on a proven theory.”

How to Account for Particles in the Cyclone Separator Simulation

Summary of ANSYS CFD Multiphase Models. DEM and DDPM used in cyclone separator simulation.

As for the particle-to-particle and particle-to-wall interactions, Gohel suggested using a discrete element model (DEM). The model is based on a spring and dashpot system between the particles and boundary walls. The deformation and collision forces are determined based on the overlap of the particle spheres.

To speed up your cyclone separator simulation, users may want to select a different time step for the fluid and the particle interactions. The DEM collisions need multiple particle integration steps to resolve each collision. Therefore, the time step is often lower than what is needed for the fluid. 


“ANSYS Fluent provides the flexibility for the fluid flow time step size to be a different value and hence we almost always choose a higher value to speed-up the simulation,” said Gohel.


The dense discrete phase model (Dense DPM) is the final one used to simulate the system. This model is an expansion of the DPM model to handle flows with high particulate concentrations. The Dense DPM will simulate the dampening effects of the particle phase as it blocks the flow of the continuous solid-fluid phase. This is achieved by mapping the particle phase volume fractions onto the Eulerian multiphase mesh and the aforementioned DEM model.


The Dense DPM can use either the DeFelice or Gidaspow drag models to account for the dampening forces of the particles. As the particle flow is influenced more on the collisions than the drag models, both DeFelice and Gidaspow should yield similar results. However, Gohel chose to use DeFelice because “it produced smoother particle flow pattern especially at lower solid loading.”


He added, “All these models [RSM, DEM, and Dispersed DPM] are necessary to be able to accurately predict the pressure drop across the cyclone. Once the appropriate models are chosen, ANSYS Fluent’s robust solution algorithms take over the task of converging the solution.”

Results from Simulating a Cyclone Separator

Simulation validation shows the predictive abilities of pressure drop reduce with larger solid loadings. Pressure drop is a mass weighted average of the sum of dynamic and static pressure.

The resulting model is shown to be able to accurately predict system’s pressure drop across wide range of solid loading. While overall match with experiment is good, a very close comparison with experiments can be seen until the mass particles and air were at a one-to-one ratio.

 “Particulate flow in a cyclone separator is mainly governed by collisions,” Gohel explained. “The importance of correctly modeling collision force increases with increase in solid loading. The present study was carried out using a simplified collision model, which ignores particle rotation and rolling friction. Inclusion of these effects can further improve the pressure drop prediction at higher solid loading.” 


Therefore, for very densely-loaded separators, analysts may want to include the particle rotation and rolling friction in their models.


The results also show that the friction coefficient has a large effect on the particle flow but a small effect on the pressure drop. “Friction coefficient between a pair of particles is an important parameter which determines particle position and velocity after each collision event,” said Gohel. “All these collision events collectively result in a different particle flow pattern with different value of friction coefficient.”


As discussed here, user can optimize a cyclone separator to fit their specific industry needs. The simulations allow the user to reduce the number of prototypes and the design cycle. Using optimization tools to explore the design space will also help with the automation and selecting the ideal cyclone separator for the job.


To learn more about simulating cyclone separators in ANSYS follow this link.

Ansys has sponsored this post. They have no editorial input to this post. All opinions are mine. —Shawn Wasserman