CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Troubleshooting probe and average functions with Angular Velocity (http://www.cfd-online.com/Forums/cfx/66921-troubleshooting-probe-average-functions-angular-velocity.html)

siardica July 28, 2009 13:17

Troubleshooting probe and average functions with Angular Velocity
 
LIBRARY:
CEL:
EXPRESSIONS:
radius = 2 [m]
prevAngVel = (probe(Velocity u )@omegaPt)/radius
timestep = .002 [s]
inertia = 40 [kg * m^2]
curAngVel = prevAngVel + (torque()@Blades * timestep / inertia)
END
END
END

omegaPt is defined as a user point located on the outer perimeter of a rotating domain.

The rotating domain is set to rotate at angular velocity curAngVel.
CFX-Pre returns no errors.


Solver returns:
"Error processing expression 'Angular Velocity'.
The expression is invalid because:
Velocity u is not available for use in this term
Error processing expression: Angular Velocity = curAngVel"

I get this error with every combination of variable and probe or average function I can think of. Having a dumb blonde moment... what am I missing?

stumpy July 29, 2009 07:01

I think that Angular velocity is just not allowed to be a function of the fluid velocity. If you look in the VARIABLES file in <install root>/<version #>/CFX/etc/ and search for "Angular Velocity" it will tell you what variable this can be a function of - i think it's called "Dependency List". Now you could try editing this file to comment out the dependency list, and there's a at least a chance it will work, but it may not. The work around is probably to make an algebraic additional variable equal to probe(u)@omegaPt, then make the Angular Velocity a fn of the AV.

siardica July 29, 2009 15:02

You are officially my hero for the day -- just added velocity to the list of dependencies in the rules file and it worked like a charm. Thanks so much!

ghorrocks July 29, 2009 19:13

Hi:

Be very cautious doing this. The variable is probably not in the list of dependencies for good reason! I suspect the influence of variable velocity is not correctly taken into account. I would do a careful validation of this before using it.

It sounds like a great idea, just check it works before you do anything important.

Glenn Horrocks

pingub July 13, 2011 09:26

HI everyone,
I'm performing the same simulation (overrunning of the turbine).
I'm calculating the rotating velocity at the previous time step as:

LIBRARY:
CEL:
&replace EXPRESSIONS:
Gamma = 9806 [N/m^3]
Raggio = 0.0454845 [m]
Torcente = torque_x()@gir_wall
deOmega = omegapunto * dt
dt = 0.002083333333333333333333333333333 [s]
grav = 9.806 [m/s^2]
inerzia = 0.00167941 [kg*m^2]
omegapunto = Torcente/inerzia
omegauno = deOmega + omegazero
omegazero = areaAve(rotvel)@gir_wall / Raggio
rho = Gamma/grav
END
END
END


I also added "rotvel" to the dependancies list in RULES file.

The error I get is:

Details of error:-
----------------
Error detected by routine PEEKL
CDANAM = /FLOW/PHYSICS/ZN2 /LIMSOLID
CRESLT = NONE

Current Directory : /FLOW/CEL/CBCK_CALLS

+================================================= ===================+
| ****** PROBLEM REPORT ****** |
|--------------------------------------------------------------------|
| Subsystem: Input |
| Subroutine name: ErrAction |
| Severity level: Fatal Error |
| Error message number: 001100279 |
|--------------------------------------------------------------------|
| Message: |
| |
| Stopped in routine MEMERR |
| |
| |
| |
| |
| |
+================================================= ===================+


What's wrong?
How can I workaround? I've also tried with a fortran user routine, but CFX returns me the same MEMERR.

pingub July 15, 2011 03:47

Any advice?

ghorrocks July 15, 2011 06:28

Isn't it obvious? rotvel is not available for callbacks (ie areaAve function), and you tried to add it in the VARIABLES files but funnily enough the solver cannot find it.

If you read the second half of stumpy's post he suggests how to work around this.

pingub July 15, 2011 07:41

Thanks for your answer.

I'm sorry, but I also tried to avoid the use of rotvel, I made an attempt with a constant value of omegazero. But I had problem also with the torque_x.

It seems I cannot relate the Angular Velocity to any variable but the time...

I'm trying with the stumpy's workaround, I will acknowledge you!
If you have any other ideas, please post!!!!

pingub July 15, 2011 08:58

Ok, the workaround with algebraic additional variable also does'nt work.
I also tried to use Velocity w instead of rotvel, but the problem is that I cannot relate Angular Velocity to any variables but the time.

Furthermore, if I use rotvel to calculate the INLET condition (or OUTLET, or anything else), it perfectly works. The only problem is with Angular Velocity!!


My only idea is to manage it manually, step by step (PRE-SOLVER-POST-PRE-SOLVER-POST...). I only miss a slave...

I wonder how siardica did...

hos747 May 4, 2013 11:31

Angular velocity depend on Additional variable
 
Hi all,

I have similar error in my simulation:
I need a Angular velocity which is function of an additional variable(myspeed).

As mentioned hear, I tried to change dependency list in Rules file. but there are many Angular velocity in this file. I tried to change each of them and add A.V (myspeed).

but I faced again error:

| ERROR #001100279 has occurred in subroutine ErrAction. |
| Message: |
| Angular velocity can depend only on time for transient simulation- |
| s.


Has anybody got an idea how to edit rules file for angular velocity?

Thanks in adavance.

ghorrocks May 5, 2013 07:47

stumpy's post above describes how to get around this. There is no need to edit the rules file.

hos747 May 5, 2013 08:53

Thank you for your comment

As stumpy mentioned I have to change dependency list for angular velocity.

I have no idea how do that.
since, several times I tried edit rules file, but again crushed.

could you please explain me how I must change dependency list

hos747 May 5, 2013 16:43

no one help me?

ghorrocks May 5, 2013 18:51

Have you read stumpy's post #2? He describes the work around there.

hos747 May 5, 2013 20:01

yes, I have read.
but I dont know how I must edit dependency list in rules file as stumpy mentioned "Now you could try editing this file to comment out the dependency list"

please guide me

hos747

ghorrocks May 5, 2013 20:04

Here is a direct quote of the comment:

Quote:

The work around is probably to make an algebraic additional variable equal to probe(u)@omegaPt, then make the Angular Velocity a fn of the AV.
Have you tried this?

hos747 May 5, 2013 20:29

yes, I have made an additional variable and then made it function of additional variable. but it crushed

error: Angular velocity can depend only on time for transient simulation

w-idiot May 29, 2013 09:08

Quote:

Originally Posted by hos747 (Post 425436)
yes, I have made an additional variable and then made it function of additional variable. but it crushed

error: Angular velocity can depend only on time for transient simulation

I also have similar error in my simulation!!!I know nothing to deal with it

ghorrocks May 29, 2013 19:03

Have you read this thread? It explains what is going on in detail.

w-idiot May 29, 2013 23:53

Quote:

Originally Posted by ghorrocks (Post 430853)
Have you read this thread? It explains what is going on in detail.

It doesn't work! Maybe I did something wrong.Thanks!


All times are GMT -4. The time now is 09:20.