Fluent giving different results for the same analysis
 

Hi,
I'm analysing a 2d model of a centrifugal pump using steady k-epsilon turbulence model. I iterate the fluent 5000 times. I use 2d mode, not 2ddp. Residual convergence criteria is 1e-5 for k,epsilon, x velocity, y velocity, continuity.

I get different results each time I run the model. What could be the reason? Should I increase the number of iterations, or decrease the convergence criteria?

Thanks in advance.

I'm changing the option to 2ddp to see if it solves my problem.

Edit: It didn't solve my problem.

What do you mean with "different results". Slightly different or completely different?

Pressure values are 10% or so different for the same setup(same .msh file).

I use a velocity inlet, a pressure outlet, and wall boundary condition at the blades.

I want to use the result for an optimization work.

The equations you solve for are deterministic. For a given set of initial conditions you will always get the same result. Obviously, you do something (unintentionally) different each case you run it.

I think I spotted the problematic portion of the journal file:

-----------------------------------------------------------

yes
"C:\Users\bpehlivan\Desktop\2dpump\ga17\mass_flow_ rate_v2.flp"
re/si
a-w-a
2
6

total-pressure
yes
"C:\Users\bpehlivan\Desktop\2dpump\ga17\total_pres sure_v2.srp"
re/si
q
exit yes
-----------------------------------------------------------



I'm getting the following message on fluent when I run the fluent journal file:

-------.-.-.--.-.------.--..-.--.-.---.-.------.--.-.-.-
> re/fl/mf
all boundary/interior zones [yes] no

Zone(1) [()] 5
Zone(2) [()] 9
Zone(3) [()]
Write to File? [no] yes
filename (*.flp) [""] "C:\Users\bpehlivan\Desktop\2dpump\direct1\mass_fl ow_rate_v2.flp"

> re/si

/report/surface-integrals> a-w-a
()
surface id(1) [()] 2
surface id(2) [()] 6
surface id(3) [()]

area-avg of> total-pressure
Write to File? [no] yes
filename (*.srp) [""] "C:\Users\bpehlivan\Desktop\2dpump\direct1\total_p ressure_v2.srp"

/report/surface-integrals> re/sinvalid command [re]
i

/report/surface-integrals>
area flow-rate sum
area-weighted-avg integral vertex-avg
facet-avg mass-flow-rate vertex-max
facet-max mass-weighted-avg vertex-min
facet-min standard-deviation volume-flow-rate

/report/surface-integrals>

-------.-.-.--.-.------.--..-.--.-.---.-.------.--.-.-.-


So, I think there is a problem when writing the pressure to the file.

Change it to "/re/si". I think that should work for you.