[highhopes]

Dear friends,

I have been using Icem CFD for a long while now and I have a feeling that it is not supported anymore because of features which are not improved such as the interface, etc. Is there any news about this software that I have maybe missed?

I have tried ANSYS and Fluent Meshing but I was rarely satisfied with the quality of their mesh (especially at the boundary!!!) and the necessity of learning their many tips and tricks. I have spent years on Icem CFD and now I am able to control almost any aspect of a mesh and create very good block-structured meshes.

I would really appreciate your input here. If this topic was already discussed, then I am sorry for posting this and I would also appreciate if you pointed me to those posts.

Kindest regards

[siw]

ANSYS will tell you that they are still working on ICEM. A lot of the functionality is now being put into SpaceClaim's new Interactive Meshing. ICEM has also been omitted from the Student Edition.

[Gert-Jan]

Indeed, ICEM has not improved a lot lately. I think it is more or less end of life time.
In my opinion ANSYS discourages its use, e.g. by no longer allowing the import of .stp and .igs files after normal installation. You have to add the library links yourself manually. Grrrrrrr.

I used ICEM for hex meshing but also a lot for tet meshing, because it is very forgiven for sloppy CAD drawings. There was even a time that I drew everything in ICEM. But after the introduction of Spaceclaim, it is easier to repair sloppy CAD in Spaceclaim, making the route to ANSYS Meshing a more logical one. So, I use ICEM less and less for tet meshing. Also because I notice that it is much slower than ANSYS Meshing.

For hex meshing I still use ICEM. However, ICEM hex meshing features are being introduced (as a beta feature) in Spaceclaim. So, I expect that Spaceclaim will take over. ICEM will no longer be supported in the near future.

[bluebase]

Originally Posted by siw

ICEM has also been omitted from the Student Edition.

On the ANSYS homepage you won't even find any information on ICEM anymore, try using the search.

ICEM's front end and many of the functionality is written in tcl/tk. It even has it's own special framework to be quick enough.
I would speculate, this structure has become difficult to fit into the rest of the tool chain of ANSYS or plainly it become to costly to maintain.

As a general question for the discussion, how many peers do you know using ICEM?

I have seen many occasional users (mostly students), which have tried ICEM, but only very few became advanced users. I'd estimate only a few hundred engineers are there worldwide actively using ICEM, maybe a low 4-digit number. If that would be the case, the size of the user group might have become to small for ANSYS to justify further development?


I also still work with ICEM, but mostly to get structured/hybrid meshes, or certain geometry operations. For unstructured meshing, other tool (chains) become more comfortable and reliable.

For some time now, I try to find suitable alternatives. Are the beta features in SpaceClaim promising? I haven't played with the main line tools for while.
Currently, Pointwise is very appealing to me, though in my current work environment subcribing it has been ruled out for now...

[highhopes]

Thank you very much for all of your answers. Now it is clear to me that ANSYS is not keeping the Icem users up to date.

I know currently 7 colleagues actively using Icem. And I know some research groups strictly using block-structured meshing approach for turbomachinery and building aerodynamics problems.

Thank you also for sharing your alternatives to Icem. So much appreciated!

[siw]

Been trying SpaceClaim Meshing for a little while. It's clearly going to be the replacement for ICEM. Just a shame that ANSYS have released this, and removed ICEM from the Studeht version, before SpaceClaim Meshing has even some of the more basic hexa features and abilities.


As a test I tried ANSYS'c 3 part Hexa of an aerofoil on YouTube using SpaceClaim. No success. Could get SpaceClaim to make an O-Grid with breaking the initial Block associations. Cannot split blocks at prescribed points. Cannot align block vertices - only move them freely. Making O-Grids (now just called Boundary Layers) is less straight forward.


Not impressed with SpaceClaim Meshing as this earlt stage. Particulary, not happy that cannot get ICEM only more in the Student edition.

[Gert-Jan]

If possible using student versions, always keep at least 3 versions of ANSYS on your system (I have 5). That would have assured your access to ICEM.

[siw]

I only put the student version on my personal laptop to try features in my own time to then maybe use at work or maybe try a simulation about something that interests me. So I do not use it as a student at university would. That is why I only have one version, the latest, on my laptop and the older versions with ICEM are not now available for download, I think. I can see that once SplaceClaim Meshing has had somemore development over then next few releases then it should be good - but now I find it a bit undercooked. Some online tutorials, e.g. YouTube, would be very helpful from ANSYS for this new capability. Which is getting off the ICEM discussion.

[Far]

Recently I had the chance to talk with someone at ANSYS (was infact), and you know him very well. He is no more working for them and he was one of the best in ICEMCFD and he promoted ICEMCFD to next level. Before him ICEMCFD was alien for us. Whole team was dismantled.

Secondly their plan to replace every meshing software with ANSYS meshing did not work as can be seen from very sloppy interface of Workbench mesher. It requires lot of inputs from design modeler or spaceclaim. Although meshing process is simple and many things are automatic. But in case, we want some tweaking at somewhere, it is not possible there.

Although when ever I teach some one CFD from beginning, I would not go with ICEMCFD. I would prefer workbench mesher, as it can hardly take few minutes to get mesh for very simple geometries which I use as examples. So question arises, why not ICEMCFD. Because ICEMCFD requires lot of inputs and some times it is hard to understand functions of button which can be used in very different ways for different applications. In short, ICEMCFD requires same time for setting up meshing for simple and very complex geometries. But in ANSYS meshing, you need less time for very simple geometries and for very big and complex models, you may crash it.

So for teaching purpose, ANSYS Meshing, for professional work, I would go with ICEMCFD. Though, now I don't create domain or repair geometry in ICEMCFD, that all is done in Spaceclaim.

So ICEMCFD is not going anywhere, until they successfully incorporate all options from GAMBIT, ICEMCFD, Fluent Meshing (Tgrid), CFX meshing etc into workbench meshing in efficient and effective way. At that time ICEMCFD will die and user wont see any mesher at all in interface and every thing will be done in background. This is the Future of CFD.


As far as spaceclaim meshing capabilities are concerned, I have tried them. But it is still not fully developed and I find many bugs in there.

People will continue to use Ansys workbench mesher and will continue to curse ICEMCFD for its steep learning curve. But if you want high quality CFD results, you have to go with ICEMCFD!

As far as new features are concerned in ICEMCFD, I have seen many new options are introduced in Hexa meshing. In tetra meshing side, you can observe inclusion of advance options for Fluent prism setting (PreInflation) in prisms settings. But you are correct in a way, that interface is still that old fashioned. They even did not care to make it more interactive. For example selecting working directory is still pain in the hand. They need to improve many thing in GUI but as far as functionally is concerned, you wont find these features in any software in the whole industry.