sliding grids
 

Hi,

I have a problem when I try to use sliding grids; by the way, the simulation of the centrifugal pump produce high residual mass. this cause the end of the simulation.

have you an idea ?

Thank you very much!

Giuseppe

Re: sliding grids
 

It is difficult to understand your problem at this stage. Would you be more generous with your information. When does this occur, immediately or...?

What are your: -speed -mesh -time step -underelaxation factors -descritization schemes (spatio-, temporal-) etc..??

regards Nyatoto

Re: sliding grids
 

Hi, Nyatoto

The simulation diverge immedyately, after 7-8 iterations.

I have 1400 rpm , my model of rotor has structured mesh with 4 elemet for each arc with an angle of ten degrees.

the fluid was water

I used :

>>FIXED TIME STEPPING

TIME STEPS #NSTEP * #DT

with

#CALC

PI=3.1415926;

DEG=PI/180.0;

#ENDCALC /* ... ASTEP IS GRID MOVEMENT ANGLE PER TIME STEP ... */

#CALC

NSTEP = 144;

MAXIT = 50;

CRIT = 1000;

ASTEP = 2.5*DEG;

#ENDCALC /* ... WORKING VALUES ... */

#CALC

RPM = 100;

OMEGA = RPM*2.0*PI/60.0;

REVOLT = 2.0*PI/OMEGA;

DT = REVOLT*ASTEP/(360.0*DEG);

#ENDCALC

i used default under relaxation factors and the options:

THREE DIMENSIONS

BODY FITTED GRID

UNMATCHED GRIDS

CARTESIAN COORDINATES

TURBULENT FLOW

ISOTHERMAL FLOW

INCOMPRESSIBLE FLOW

TRANSIENT FLOW

TRANSIENT GRID

furthermore i have:

>>FALSE TIMESTEPS

U VELOCITY 0.8

V VELOCITY 0.8

W VELOCITY 0.8
:>REDUCTION FACTORS

K 0.55

EPSILON 0.55
:>DEFERRED CORRECTION

K START 0

K END #MAXIT

EPSILON START 0

EPSILON END #MAXIT

I ran two simulations with the following boudaries conditions:

1) inlet + outlet

with inlet velocity of 3,5 m/s

2) simulation all circuit

No inlet No outlet

I hope for some help

Thanks in advance! Giuseppe

Re: sliding grids
 

Hi Giuseppe,

Is your RPM 1400 or 100? Please check your FC file. Let me first assume that everything is right with your geo file (BC, grid,etc).

Your false time steps appear to be too high, it is generally a good practise to start with low false time steps (if you need them). In your case I would like to know what your under relaxation factors are! These too can initially be set very low, sometimes the default ones may be too high, expecially if your flow is highly turbulent. I assume you are using k-e model, which will INITIALY give you huge errors if you have a flow with high swirl (I suspect your model has this) and the solver ambitiously tries to get fast solution with high under relaxation factors.

Another thing is the discretization schemes: You may start with, say upwind or Hybrid. For time scheme I prefer fully implicit.

I suggest you try under relaxation factors in the range of u,v,w=0.1-0.3 or lower, P can be as high as 0.5 or 0.7, k,e about 0.3-0.5. You can increase these later as you wish. similarlt try smaller times steps to begin with.

Lastly, select your out put monitor position wisely.

Hope this helps Nyatoto