[Melvins]

It says:
------------------------------------------------------------------+
| *** INSUFFICIENT CATALOGUE SIZE *** |
| |
| ACTION REQUIRED : Increase the file catalogue size. |
| |
| If the situation persists please contact the CFX Customer Helpline |
| giving the following details:- |
| Current catalogue size : 28960 |
+--------------------------------------------------------------------+

My element number is 34 million, so I wonder if it is the huge element number that leads to the memory size problem, but it is OK in steady simulation.

Please help, thanks a lot...

[Martin_Sz]

There are many reasons of this error.
Check your mesh quality. Make these mesh more coarse.
Second check your BCs

[Opaque]

In the ANSYS Solver Manager, click on the Show Advanced Controls, go to the Solver tab and click on Detailed Memory Overrides and use a multiplier for the Catalogue Size.

If using the command line, cfx5solve, please use cfx5solve -help and read the details to modify the catalogue size (-smms ?)

The software failed to selecting the proper catalogue size for its memory system.

Hope the above helps,

[NielsB]

I've encountered this problem as well, increasing the catalog size will solve the problem. But I'm wondering, what's actually the 'meaning' of the catalog size? Is it related to memory somehow?

Kind regards,
Niels

[jjpbuaa]

Originally Posted by mactech001

Hi Glenn, thanks for your response.
i wanted to try the approach above to increase it, but my problem is, CFX prompted me that i can't add parameters anymore with the following message:

No New parameters can be added to the /FLOW:Flow Analysis 1/SOLVER CONTROL section of the CFX Command File.

Is it a setup problem or installation problem would you think?

regards,

hi, I have met totally the same question with yours. Have you solve this problem? Have you figured out why this problem occurred?

regards.

Originally Posted by Melvins

It says:
------------------------------------------------------------------+
| *** INSUFFICIENT CATALOGUE SIZE *** |
| |
| ACTION REQUIRED : Increase the file catalogue size. |
| |
| If the situation persists please contact the CFX Customer Helpline |
| giving the following details:- |
| Current catalogue size : 28960 |
+--------------------------------------------------------------------+

My element number is 34 million, so I wonder if it is the huge element number that leads to the memory size problem, but it is OK in steady simulation.

Please help, thanks a lot...

I have met the same problem. How did you solve this?

Originally Posted by Bruyère
;84069

I have tried with catalogue size multiplier = 1.5 and it is working. So thank you very much. I keep your mail in mind. Regards. Bruyère.

hi, I have met the same problem. When i tried to add parameter "catalogue size multiplier", it indicated "No New parameters can be added to the /FLOW:Flow Analysis 1/SOLVER CONTROL section of the CFX Command File." Also, the try with "cfx5solve -size-mms 3" failed. I'd like to fingure out why this happens?

regards.

[ghorrocks]

When CFX starts a simulation it estimates the amount of memory it will require to run and grabs that chunk of memory. For most simulations the estimation is pretty good and this causes no problems. The solver solver requires additional memory as the solution progresses and if this rises above the amount initially estimated the solver will stop with a catalog size error.

In these cases the first thing to do is to increase the catalog size, x1.2 is usually enough but if you have the memory to spare x2.0 is good. In a small number of cases this does not work - this indicates a more fundamental problem with the simulation, where function requiring memory has proven far more complex and it should be, and this invariably means a problem in that function and the problem should be fixed (rather than just allocating more memory and hoping).

[jjpbuaa]

Originally Posted by ghorrocks

When CFX starts a simulation it estimates the amount of memory it will require to run and grabs that chunk of memory. For most simulations the estimation is pretty good and this causes no problems. The solver solver requires additional memory as the solution progresses and if this rises above the amount initially estimated the solver will stop with a catalog size error.

In these cases the first thing to do is to increase the catalog size, x1.2 is usually enough but if you have the memory to spare x2.0 is good. In a small number of cases this does not work - this indicates a more fundamental problem with the simulation, where function requiring memory has proven far more complex and it should be, and this invariably means a problem in that function and the problem should be fixed (rather than just allocating more memory and hoping).

Thank you very much for your answer. The error only occurred in the transient simulation when i used the steady simulation results as the initially value. I have the model with 3 million elements. Maybe the interpolation got error since the huge calculation or the complex interface? Can i solve this problem by using distributed parallel simulation?

[ghorrocks]

The large mesh will not cause the error - the memory required for meshes is easy to calculate.

But complex interfaces can cause it. If the interface is more complex than normal the built in estimation can be wrong and lead to this error. But it is also a sign that the interface might not be well formed. Check the mesh for folds, bad quality elements and other badness near the interface.

[jjpbuaa]

Originally Posted by ghorrocks

The large mesh will not cause the error - the memory required for meshes is easy to calculate.

But complex interfaces can cause it. If the interface is more complex than normal the built in estimation can be wrong and lead to this error. But it is also a sign that the interface might not be well formed. Check the mesh for folds, bad quality elements and other badness near the interface.

The simulation I made is for a patial admission turbine. The model was meshed using IGG/AutoGrid, and I think the quality is not bad. However, eace side of the interface containing more than 50 faces in CFX. Can this be the reason for the error?

[ghorrocks]

It is hard to be sure but that is a strong suspect. Try make the interface consist of a smaller number of separate faces if possible.

[Milsey]

The error isn't caused by mesh at all. Its caused by the solver detirmining how much space/ memory is required to complete all the programmed iterations. By changing the catalogue allocation/ iteration size you can allow the computer to take up more space, or reduce the amount of space required.

[ghorrocks]

Posts #26 and #28 explain the source of this problem. If the error in memory estimate is small you can fix it with the catalog allocation parameter, but if that is enough you have to look at where the error in the estimation comes from. In my experience it is usually from large numbers of surfaces, interfaces or bodies; or very complex interfaces. It can also come from huge amounts of CEL expressions. To fix it you need to simpilfy the mesh faces/bodies or reduce the CEL.

[Rikki-Tikki-Tavi]

Originally Posted by Marvin

Just to mention that this error can also occur at the interpolation phase of the simulation. Even, for example, when starting the simulation from a current solution onto identical mesh. In this case, increase the catalogue size on the "Interpolator" tab of the solver manager.

Thank you for making me feel stupid by stating the obvious and solving my problem with it.

[TGS]

Thank you very much for your answer. Also, need to use the suffix x with the multiplier.