[ardismo]

Hi;

Im trying to simulate gas-solid cyclone separator used in cement industry. I will mention about boundary conditions and operating conditions. I need some ideas to solve more accurate , quicker solution with same accuracy or to correct my mistakes. These will helpful for me and for others who has need help on this subject.

operating conditions ;

- It is a multilevel system consists of 5 cyclone levels.
- very high dust loading particle.( 0.2 kg particle/ m^3 air)
- 12 to 20 m/s inlet velocity
- nearly 700 m^3 control volume

Boundary Conditions;

1 - velocity inlet
2 - pressure outlet or outflow

****normally outlet is not open to atmosphere. Gas always goes to upper cyclone level. So there must not be reverse flow.

3 - wall ,of cyclone body
4 - wall of dust outlet

Setup settings;

1 - Mesh count : 160K , 700K , 3M, 8M , 30M
2 - Eularian multiphase model with dense DPM option.
3 - Reynols stress model with linear pressure strain option
4 - pressure-velocity coupling scheme;
----Simple : there is no need to set a CFL but long computational time.
----Coupled : I have enough memory. But in Fluent theory manual, it says below 10 is not recommended for Courant number. And if you can't choose wright CFL in coupled , solution monitors fluctuates a lot.
5 - Spatial Discretization ;
---- Gradient :Green Gauss node based
---- Pressure : Presto
---- others : second or third order upwind---- solution monitors fluctuates a lot. So it says you to switch time-dependent simulation.
6 - Under Relaxation Factors : 0.5 for all parameters
7 - Initialization : Normal + FMG with 7 levels


QUESTIONS

1 - I will use Multi-Objective algorithm to get functions of separation efficiency and pressure loss depending on geometric and flow parameters. I could not get efficiency directly . I use 17 injections for 17 different diameters, I need efficiency for all diameters like output parameters.

2 - Is it necessary to set a boundary layer mesh while using RSM ?

3 - Which model would be more suitable for particle drag force calculations ?

Thank you in advance

[chen_xmu]

I am alos interested in cyclone separator simulation.
The reference below may help your problem.

[1] S.K. Shukla, P. Shukla, P. Ghosh, Evaluation of numerical schemes using different simulation methods for the continuous phase modeling of cyclone separators, Adv. Powder Technol. 22 (2011) 209–219. doi:10.1016/j.apt.2010.11.009.

[ardismo]

Thank you so much for your response. But I can solve primary phase flow accurately. My main problem is to solve DPM physics. As I mentioned above I have 16 diameters that ı modeled as 16 injections. I want to see seperation efficiencies of each diameter.

[chen_xmu]

Originally Posted by ardismo

Thank you so much for your response. But I can solve primary phase flow accurately. My main problem is to solve DPM physics. As I mentioned above I have 16 diameters that ı modeled as 16 injections. I want to see seperation efficiencies of each diameter.

This problem troubled me too. If you have solved , Please share the solution with me , Thanks.

[Tal Eluk]

Hi- here is my suggestion:
In case you are not reaching high levels of solid volume fraction (let's say below 10%) then you might be able to say that the particle don't affect each other. in that case there is no need to follow all the 17 injections at ones. you can run each diameter on it's own and get explicit efficiency for each one. this will require 17 DPM runs, but if you do it "one way" after the fluid is already converged you can be done with all 17 runs after 1/2 houre.

Good luck
Tal