[kbaker]

I simulated flow in square cavity its boundaries as : top wall insulated,bottom wall fixed Th, left wall heated with linear temperature profile and the right wall is kept with Tc. The results seem highly reasonable except the stream function it reasonable in shape but the values seems inverse? Fluent shows the maximum values of stream function near the walls and at the center the stream function is zero? this case is differ when compared with the results from other researchers which they show that the maximum stream function occur at the center of vortex and decreases gradually towards the walls? so what the problem? answer is appreciated?


Khalid

[LuckyTran]

Compare the definition of the stream function in Fluent versus the ones you're seeing in your papers. Because stream function is calculated from the velocity gradient, there's an arbitrary sign. It's very common in some fields to flip the sign. The stream function is also not unique for a given velocity field (to within an arbitrary constant). Thus a max in one definition and min in another definition of the stream function could in fact be the exact same velocity field.

Also don't make repeat posts of the same questions because it makes it harder for others to find the same answers.

[kbaker]

Thanks Lucky to the reply. Yes I compared the stream function with that calculated from the papers they define it in dimensionless relation as (epsi/thermal diffusivity) where epsi stream function and I was transform the stream function predicted from Fluent to a similar term by divided it firstly by density as (epsi/rho) to transform it to m2/s then I divided it by thermal diffusivity (k/rho*Cp) then I got it as dimensionless as the paper but it inversely distributed as the last pict.

Originally Posted by LuckyTran

Compare the definition of the stream function in Fluent versus the ones you're seeing in your papers. Because stream function is calculated from the velocity gradient, there's an arbitrary sign. It's very common in some fields to flip the sign. The stream function is also not unique for a given velocity field (to within an arbitrary constant). Thus a max in one definition and min in another definition of the stream function could in fact be the exact same velocity field.

Also don't make repeat posts of the same questions because it makes it harder for others to find the same answers.

[LuckyTran]

But how do they calculate epsi the stream function? Do they take the positive or negative of the velocity gradient?

[kbaker]

They used both of negative and positive velocity gradients. About Fluent is it used just positive velocity gradient so it give positive stream function values?

Khalid

Originally Posted by LuckyTran

But how do they calculate epsi the stream function? Do they take the positive or negative of the velocity gradient?

[kbaker]

The attached two plots are for dimensionless stream function which I got it from several runs. The first plot represents Fluent results while the second one is from a paper which I want to compare with. As it seen from the paper the stream function has negative and positive values while Fluent gives only positive values and consider the negative values as zero? I saw that the distribution of stream function is reasonable at the first vertex the stream function is max at the center and decrease gradually and this fit to the paper just the second vertex which has negative values I do not now why Fluent calculate it as positive values only and give it zero value at the center? I am really have doubt about this result is it acceptable and I can depend it in my work or reject it? any reply please?

Khalid

[LuckyTran]

The stream function is a potential. Positive or negative values of stream function don't mean anything because a particular value of stream function is only meaningful relative to a particular reference. There is no unique stream function for a given velocity field because the stream functions within an arbitrary constant will all give the same velocity field.

You need to compare the stream function definition in your papers with the Fluent definition, they're probably not the same.

[kbaker]

I look through the definition of stream function in Fluent it defined it as
rho*psi (where rho is fluid density,psi the stream function) see the link below please:

https://www.sharcnet.ca/Software/Ans...ug_fvdefs.html

so in order to obtain psi in terms of m2/s , I simply used a custom field function of the form psinew=psi/rho and in this way I got psi similar to that calculated at the paper ? I think this procedure correct?