CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Flow past a sphere (http://www.cfd-online.com/Forums/fluent/47573-flow-past-sphere.html)

Fabio March 11, 2008 09:08

Flow past a sphere
 
Hi all,

I am trying to simulate on a 3D model, a flow past a sphere which is located inside a cylinder, gap width between the sphere face and the cylinder 0.3mm (ball radius 15mm, cylinder radius 15.3mm).

I got 750 PSI (it's true, is really high pressure!) pressure difference between the inlet and outlet.

My goal is to predict the flow rate and the force action on the ball. As I am not interested to catch the turbulences shear stress and so on, then my mesh isn't too fine, no needs to use EWT.

As you see the flow restriction is too small and the pressure is high, which yields huge flow velocities values on the shorter gap area. My point is which viscous model fits better? K-E Standard? Realizable? The solver might be Simple or Couple could be used as well?

Thanks.

Muhammed_iraq November 24, 2009 13:37

energy eqation in flow past a sphere
 
Dear Sir
how are you?
i have problem is flow past a sphere,3-dimension,unsteady state
if the energy equation is applied in flow past a sphere,how the mean value of heat tranfer calculated on surface of the sphere at each time by using FLUENT?,please can you give me the options of FLUENT in order to calculate mean value of heat tranfer calculated on the surface of sphere at each time .
i am waiting your reply quickly
please,please help me
many thanks to you

thecfduser November 27, 2009 15:44

Hi
go to solve--->monitors----> and u will find it

u will choose to monitor the heattransfer over the sphere surface

Muhammed_iraq November 27, 2009 16:37

Quote:

Originally Posted by thecfduser (Post 237987)
Hi
go to solve--->monitors----> and u will find it

u will choose to monitor the heattransfer over the sphere surface

Dear thecfduser

thank you very very much for your help,please Sir ,i will go to solve--->monitors and aftetr that what i will select (force,surface,volume,..or what),please give me the details step by step ,please my problem is waiting your help, for remember my problem is unsteady state,3-dimension,energy equatino,sphere
mant thanks to you

thecfduser November 27, 2009 16:50

Ok
solve--->monitors--->surface
choose 1 surface monitor. Enable plot, print and write

in monitor 1: click on define:
name: put any name ex:heat transfer coeficient

report type: i dont know all report types...Look to Fluent manuel. Me i would put sum,
but then u must divide ure data by the surface of ure sphere

Report of: choose the variable u want to average



Now Fluent will store in a file (because u enabeled write) the sum of the heat transfer coeficient over the sphere surface. U open the file with excel and divide by the sphere surface and u will got the average in function of time


Hope that i could be useful
sorry for my bad english

if i was not clear, rewrite for me and i will try to talk in more details

thecfduser November 27, 2009 16:51

i forget something: in surfaces, u must select the sphere surface

Regards

Muhammed_iraq November 27, 2009 17:10

Quote:

Originally Posted by thecfduser (Post 237995)
i forget something: in surfaces, u must select the sphere surface

Regards

again thank you very much for your helps,your english is not bad
dear in field report time ,what i will select (integral,flow rate,sum,...or what) i want to caculate mean value of heat transfer coefficient on the surface of the sphere at each time for example h at 10sec and h at 20sec,,etc no at each node
in x axis field what i will select (iteration,time step,flow time)
i hope that it is found example for this problem
many thanks to you

thecfduser November 27, 2009 17:19

select integral, but u will get the integral of the heat coefficient over the sphere surface ok? (in function of time)
so u must divide it by the sphere surface to get the the mean value

For the rest of ure quetsion, tell me before how much iteration by time step are u performing and i can respond u then.

If u understand french, it will be easier to me.

Regards

Muhammed_iraq November 27, 2009 17:32

dear my english is best from my french
i understand on your english
your english is very good and no problem
your text"so u must divide it by the sphere surface to get the the mean value"
how can i divide?
my sphere is divided before that in Gambit program
my iteration in tim step is 30 as maximum value
many thanks to you

thecfduser November 27, 2009 17:40

of u are interested in a whole profile, u must divide it on excel or matlab...
if u choose area weighted average instead of integral, i think Fluent will divide it for u...but i am not sure....look to the fluent user's guide.

30 iteration per time step is very lot....
Anyway, in the surface monitors panel, choose WHEN TIME-STEP instead of WHEN ITERATION
In X-AXIS, choose flow-time or time-step....
Fluent will plot for u your variable, and write it in a file too

If u still have a problem, contact me (no need to thank me every time :) )
Good luck and regards comrade

Muhammed_iraq November 27, 2009 17:52

Quote:

Originally Posted by thecfduser (Post 238000)
of u are interested in a whole profile, u must divide it on excel or matlab...
if u choose area weighted average instead of integral, i think Fluent will divide it for u...but i am not sure....look to the fluent user's guide.

30 iteration per time step is very lot....
Anyway, in the surface monitors panel, choose WHEN TIME-STEP instead of WHEN ITERATION
In X-AXIS, choose flow-time or time-step....
Fluent will plot for u your variable, and write it in a file too

If u still have a problem, contact me (no need to thank me every time :) )
Good luck and regards comrade

Dear
by the way and at any time can check me what i will select area weighted average instead of integral ??? because fluent user's guide is not availble in my program.
my name is Muhammed
iraq
regards

Muhammed_iraq November 27, 2009 18:20

Quote:

Originally Posted by thecfduser (Post 238000)
of u are interested in a whole profile, u must divide it on excel or matlab...
if u choose area weighted average instead of integral, i think Fluent will divide it for u...but i am not sure....look to the fluent user's guide.

30 iteration per time step is very lot....
Anyway, in the surface monitors panel, choose WHEN TIME-STEP instead of WHEN ITERATION
In X-AXIS, choose flow-time or time-step....
Fluent will plot for u your variable, and write it in a file too

If u still have a problem, contact me (no need to thank me every time :) )
Good luck and regards comrade

Dear
by the way and at any time can check me what i will select area weighted average instead of integral ??? because fluent user's guide is not availble in my program.
my name is Muhammed
iraq
regards

thecfduser November 27, 2009 18:41

i looked in Fluent user's guide
yes, choose are weighted averag and u will get what u want :)

u can recontact me in case u still have a problem...

thecfduser November 27, 2009 19:11

hi
u are using a steady or unsteady solver ?

Muhammed_iraq November 29, 2009 05:36

Quote:

Originally Posted by thecfduser (Post 238007)
hi
u are using a steady or unsteady solver ?

i am using unsteady state
can you give me relation between Reynold and Nusselt? ,my Re=100,000 in order to check my solution.
i have checked cd with experimental work and the accuracy was very good
can give me example in actual work about flow past a sphere?
regards

CFDchile December 3, 2009 14:00

HI Muhammed. I'm trying to solve a similar problem. I need to know the global heat transfer coefficient of a sphere (in fact, there are two shperes...one into another, separated with a air gap). The physical problema that i'm trying to solve is the drying of walnuts.

But i'm starting with de simulation of the heat trasnfer. Can we maybe get email contact?. I have convergence problems with my model....how about you?.

Which solver and boundary conditions are you using?.

Hopingo to get an answer. Best regards

Matias

Muhammed_iraq December 14, 2009 10:08

Quote:

Originally Posted by thecfduser (Post 238007)
hi
u are using a steady or unsteady solver ?

Dear Karine
i am using unsteady state
can you give me relation between Reynold and Nusselt? ,my Re=100,000 in order to check my solution.
i have checked cd with experimental work and the accuracy was very good
how can i calculate the Strouhal Number(St)by using Fluent?

can give me example in actual work about flow past a sphere?
i am waiting your replr quickly

regards

Muhammed_iraq December 14, 2009 10:10

Quote:

Originally Posted by CFDchile (Post 238630)
HI Muhammed. I'm trying to solve a similar problem. I need to know the global heat transfer coefficient of a sphere (in fact, there are two shperes...one into another, separated with a air gap). The physical problema that i'm trying to solve is the drying of walnuts.

But i'm starting with de simulation of the heat trasnfer. Can we maybe get email contact?. I have convergence problems with my model....how about you?.

Which solver and boundary conditions are you using?.

Hopingo to get an answer. Best regards

Matias

Dear Matais

please explain to me your problem step by step and send it to my e-mail(muham76@yahoo.com)
best regards

thecfduser December 14, 2009 17:06

Hi
For the realtion between Re and Nu, i dont know it, but it is sure that u can find a lot of correlations in bibilography...
For the Strouhal number, it is not easy... U must monitor data in a point that is behind ure sphere, then give this data to Matlab and do a fast fourrier transform to get the frequency of ure eddies....Once u have this frequency, u can calculate the Stouhal number...I dont know if this number is affected by temperature in ure case...
Regards

Muhammed_iraq December 15, 2009 09:54

Quote:

Originally Posted by thecfduser (Post 239992)
Hi
For the realtion between Re and Nu, i dont know it, but it is sure that u can find a lot of correlations in bibilography...
For the Strouhal number, it is not easy... U must monitor data in a point that is behind ure sphere, then give this data to Matlab and do a fast fourrier transform to get the frequency of ure eddies....Once u have this frequency, u can calculate the Stouhal number...I dont know if this number is affected by temperature in ure case...
Regards

Hi

thank you very much
can give me example to caculate Strouhal number and send it to my e-mail(muham76@yahoo.com)?
best regards


All times are GMT -4. The time now is 11:37.