[Hemo]

Sorry Marjan, that's about the limit of my knowledge on the subject. Let us know if you find a solution!

[happy]

Originally Posted by singer1812

No need to use user Fortran. As an example, edit the additional variable in your LIBRARY with the update loop as follows:

ADDITIONAL VARIABLE: OldVel
Option = Definition
Tensor Type = SCALAR
Units = [m s^-1]
Variable Type = Specific
Update Loop = TRANS_LOOP
END

You can then use this to create your derivative.

does any one help me understand how the
Update Loop = TRANS_LOOP]

added to additional variable commands??
need more explaintions please, ASAP

[happy]

Originally Posted by happy

does any one help me understand how the
Update Loop = TRANS_LOOP]

added to additional variable commands??
need more explaintions please, ASAP

No one can help you happy!!!

[singer1812]

What do you want to know?

[happy]

Originally Posted by happy

does any one help me understand how the
Update Loop = TRANS_LOOP]

added to additional variable commands??
need more explaintions please, ASAP

first, how I can add update loop into my AV. I've tried using comand editor, but when I open it again, I found it is gone!!!. After you show me how it is done. How I can use AV to create derivative - is that by cel expression?

Originally Posted by singer1812

What do you want to know?

Hi singer 1812
Actually, I would like to apply derivative w. r. t. distance at my b.Cs for u, v,w,E, and K. So does I can apply that using your advices into this thread? if so, how we can handle this without using user fortran into CFX solver?

Thanks in advance

[singer1812]

You are trying to get d/dx not d/dt?

You dont need this method (update loop) to do that, that data is in already in solution.

[happy]

Originally Posted by singer1812

You are trying to get d/dx not d/dt?

yes, I need d/dx and not d/dt so which way I should follow, plz?
Regards

[happy]

Originally Posted by singer1812

You dont need this method (update loop) to do that, that data is in already in solution.

Hi again
Actually U misunderstand me because you talked about d /dx,y,z into post stage but what I'm looking for is set d../dx, d../dy &d../dz=0 at the boundary consitions (outlets) into set up stage.
any input can help
Regards

[singer1812]

Dont need transloop. Just setup a CEL expression for the derivative and put it into the BC under flow direction.

[singer1812]

Ahh, I didnt read your message carefully enough. You need it on the outlet.

To what purpose are you forcing the derivatives on the outlet?

[happy]

Originally Posted by singer1812

Ahh, I didnt read your message carefully enough. You need it on the outlet.

To what purpose are you forcing the derivatives on the outlet?

I would like to represent the forced plume that its source elevated on the ground into large surroundinf fluid domain. I tried using many BCs but I found 3 journal article used Neumann conditions at the top domain. Actually all the 3 articles did not use ready software solver like CFX or Fluin but they had used their own code. So I want to apply neumann condition for my problem and does not matter if the boundary is outlet or openning. so is this possible through CFX.
Any input can help.

regards

[happy]

Originally Posted by singer1812

Dont need transloop. Just setup a CEL expression for the derivative and put it into the BC under flow direction.

the flow direction is available only with opening boundary condition. However, whan I choose the cartesian components that means provide distance (m) and not dv/dx (s^-1). Is it correct?
Regards

[singer1812]

I dont understand why you are trying to force an outlet to have a certain gradient. Is this purely to match something in liturature?

The opening will only apply those conditions to the flow coming into the domain, not the flow leaving it. The conditions of the flow leaving the domain are a function of what takes place inside.

If you need the gradient to be zero wrt the boundary, use a symmetry BC. If it is non zero, I think you will need to use User Fortran to enforce that.

[happy]

Originally Posted by singer1812

I dont understand why you are trying to force an outlet to have a certain gradient. Is this purely to match something in liturature?

The opening will only apply those conditions to the flow coming into the domain, not the flow leaving it. The conditions of the flow leaving the domain are a function of what takes place inside.

If you need the gradient to be zero wrt the boundary, use a symmetry BC. If it is non zero, I think you will need to use User Fortran to enforce that.

Thanks singer1812 for your input
If I required any further recomments I will reask you

[lihui54312]

Originally Posted by singer1812

You are getting it from a fortran routine that you are calling while CFX is running (User Fortran?)

Or are you getting it from a pre-ran fortran routine, and delta_velocity is already set for the CFX run.

This method is right, frist you must run the fortan in your computer.

[anon_f]

Hello,

I am tring to use the TRANS_LOOP operation to solve the following equation in a CFX Domain:
XX=1-(1/((1-XXold)^-a+B*exp(-C/(D*T[K]))*F*n))^a

I am doing the following steps:

· Create an additional variable XX which is defined by the expression :
XX value: 1-(1/((1-XXold)^-a+B*exp(-C/(D*T[K]))*F*n))^a The variable is function of temperature.

· Create another additional variable XXold which defined by the expression:
XXold value: ave(XXold)@domain+XX()@domain*Time Step Size.

· Initially, I run a simulation with initial values in Additional Variable Models XX=0 and XXold=0. But when I set for each variable its CEL expression I am getting this following message:

"Bad expression value 'XXold val' detected in parameter 'Additional Variable Value' in object '/FLOW:Flow Analysis 1/DOMAIN:domain/SOLID MODELS/ADDITIONAL VARIABLE:XXold'.
CEL error:
The following unrecognised name was referenced: XX."

Is there anyone who can help me to overcome this problem please?

Thank you in advance for your advices and suggestions

[anon_f]

Hello,

I am tring to use the TRANS_LOOP operation to solve the following equation in a CFX Domain:
XX=1-(1/((1-XXold)^-a+B*exp(-C/(D*T[K]))*F*n))^a

I am doing the following steps:

· Create an additional variable XX which is defined by the expression :
XX value: 1-(1/((1-XXold)^-a+B*exp(-C/(D*T[K]))*F*n))^a The variable is function of temperature.

· Create another additional variable XXold which defined by the expression:
XXold value: ave(XXold)@domain+XX()@domain*Time Step Size.

LIBRARY:
&replace ADDITIONAL VARIABLE: XXold
Option = Definition
Tensor Type = SCALAR
Units = [ ]
Update Loop = TRANS_LOOP
Variable Type = Specific
END
END

· Initially, I run a simulation with initial values in Additional Variable Models XX=0 and XXold=0. But when I set for each variable its CEL expression I am getting this following message:

"Bad expression value 'XXold val' detected in parameter 'Additional Variable Value' in object '/FLOW:Flow Analysis 1/DOMAIN:domain/SOLID MODELS/ADDITIONAL VARIABLE:XXold'.
CEL error:
The following unrecognised name was referenced: XX."

Is there anyone who can help me to overcome this problem please?

Thank you in advance for your advices and suggestions

[Hemo]

Hi emotep, I could be wrong, but on first glance, you have ave(XXold) and XX().

I think the issue amy be here.......

[anon_f]

what do you suggest me. I am not using a variable such MASSFLOW or velocity for the variable I want to calculate. It is in fact independent. To calculate this I have already set an additional variable and express a CEL.
Thank you for your answer

[cfdonlineuser]

this is covered here: http://www.edr.no/blogg/ansys_blogge...rations_in_cfx