[itsme_kit]

I wanna put a source term (du/dz=constant) in a cell zone in the u equation to act like a shear stress

However, I have no experience of writing source term (didn't find relevant examples from internet)

Can you please suggest me some hints to write this source term?

Thanks

[A CFD free user]

Hi,
Nothing special makes you worry. The only thing you need to know is velocity gradient in a particular direction. Have a look at the manual and you find how to do that.

[itsme_kit]

Originally Posted by A CFD free user

Hi,
Nothing special makes you worry. The only thing you need to know is velocity gradient in a particular direction. Have a look at the manual and you find how to do that.

Hi Azarafra

Thanks for your reply

There is one example shown in the manual (adding a momentum source to a duct flow)

The link: http://aerojet.engr.ucdavis.edu/flue...df/node232.htm

For this formula: S=-CV, I can understand this is to force velocity into zero, even though I don’t know how the value of constant C is figured out

For my case, I have a relationship: du/dz=constant value at a specific position

However, I have no idea how I can transfer my formula (du/dz=constant) into this kind of form (S=-CV)

[A CFD free user]

Originally Posted by itsme_kit

Hi Azarafza

Thanks for your reply

There is one example shown in the manual (adding a momentum source to a duct flow)

The link: http://aerojet.engr.ucdavis.edu/flue...df/node232.htm

For this formula: S=-CV, I can understand this is to force velocity into zero, even though I don’t know how the value of constant C is figured out

For my case, I have a relationship: du/dz=constant value at a specific position

However, I have no idea how I can transfer my formula (du/dz=constant) into this kind of form (S=-CV)

Well, I had a look at the manual and I think nothing is ambiguous. S, is source term and as it was mentioned in order to enhance stability and convergence, it's recommended that one employes the derivative term which is derivative with respect to the variable. So, in your case you need to use C_U_G(c,t)[2], which actually is dU/dz and G denotes gradient and index [2] means in direction z.
I hope it helps

[itsme_kit]

Originally Posted by A CFD free user

Well, I had a look at the manual and I think nothing is ambiguous. S, is source term and as it was mentioned in order to enhance stability and convergence, it's recommended that one employes the derivative term which is derivative with respect to the variable. So, in your case you need to use C_U_G(c,t)[2], which actually is dU/dz and G denotes gradient and index [2] means in direction z. So, I suppose your source term is something like this:

S=-*C_U_G(c,t)[2]=K (cons)
dS[eqn]=0.0

I hope it helps

Hi Azarafra

I think I got your point

There are two things I'm still in doubt:

a) you said I need to use C_U_G(c,t) which is a gradient vector macro, but I found another macro for cell velocity derivative: C_DUDZ(c,t) in UDF manual

I'm not sure what's the difference between them

b) as the manual said: The solver linearizes source terms in order to enhance the stability and convergence of a solution. To
allow the solver to do this, you need to specify the dependent relationship between the source and
solution variables in your UDF, in the form of derivatives.

I couldn't understand this sentence. why derivative form can enhance stability and convergence

Thanks for your help again

[A CFD free user]

Originally Posted by itsme_kit

Hi Azarafra

I think I got your point

There are two things I'm still in doubt:

a) you said I need to use C_U_G(c,t) which is a gradient vector macro, but I found another macro for cell velocity derivative: C_DUDZ(c,t) in UDF manual

I'm not sure what's the difference between them

b) as the manual said: The solver linearizes source terms in order to enhance the stability and convergence of a solution. To
allow the solver to do this, you need to specify the dependent relationship between the source and
solution variables in your UDF, in the form of derivatives.

I couldn't understand this sentence. why derivative form can enhance stability and convergence

Thanks for your help again

OK, for the first question, I suppose there is no actually any difference between the two macros. Because C_U_G(c,t)[2] means as the same du/dz. For, the second question, I think you'd better have a look at the following book:

"Numerical heat transfer and fluid flow" written by Patankar pages 48 and 143. There's a good explanation there regarding source term.

[itsme_kit]

Originally Posted by A CFD free user

OK, for the first question, I suppose there is no actually any difference between the two macros. Because C_U_G(c,t)[2] means as the same du/dz. For, the second question, I think you'd better have a look at the following book:

"Numerical heat transfer and fluid flow" written by Patankar pages 48 and 143. There's a good explanation there regarding source term.

Really appreciate your useful help

[itsme_kit]

Originally Posted by A CFD free user

OK, for the first question, I suppose there is no actually any difference between the two macros. Because C_U_G(c,t)[2] means as the same du/dz. For, the second question, I think you'd better have a look at the following book:

"Numerical heat transfer and fluid flow" written by Patankar pages 48 and 143. There's a good explanation there regarding source term.

Hi I got another problem

Is the following structure correct?

DEFINE_SOURCE(dissipation_gradient,cell,thread,dS, eqn)
{
real source;
source=C_D_G(cell,thread)=-0.000003;
dS[eqn]=0.;
return source;
}

I got an error 'invalid lvalue in assignment' in line starting with 'source='