CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CD-adapco (http://www.cfd-online.com/Forums/cd-adapco/)
-   -   samm vs. ICEM (http://www.cfd-online.com/Forums/cd-adapco/52460-samm-vs-icem.html)

Lars Ola Liavåg October 6, 2000 09:17

samm vs. ICEM
 
ProAm from adapco appears to offer many advantages to STAR-CD users. I currently use ICEM Tetra/Prism in conjunction with the direct CAD interface to I-DEAS, and to be honest, I am relatively satisfied with this combination. However, all to frequently, I get frustrated over the many difficulties with generating prisms on my tetmeshes. If this step could be left out, I would probably cut my meshing time with two thirds - at least.

If anyone has experience with both ICEM and ProAm/samm, I would really appreciate to hear your opinions on the strong and weak points of the two softwares, however subjective they may be.

Thanks in advance.

Lars Ola

Bart Prast October 9, 2000 10:34

Re: samm vs. ICEM
 
We are using ICEM Tetra/Prism also. We had lots of problems generating prism layers. We are now switching to SAMM (pro-am). I think the problems will remain the same. If you have troubles generating a prism layer in ICEM, it will also be difficult in SAMM. The advantage of SAMM (as I saw it) is that you have a structured/hybrid/tet mesher in one package, with a user friendly interface. I would like to have a hex mesh where ever I can and minimise the number of tet cells (this is more efficient in number of cells, more accurate and more robust). ICEM is supposed to be somewhat faster so it is a matter of taste and priority. Another advantage of SAMM is is the connection with prostar. From the beginning you are working on a 'cell level' not with a discription of the geometry. During each step in SAMM (proAm) you see the resulting grid in prostar

Lars Ola Liavåg October 10, 2000 02:05

Re: samm vs. ICEM
 
I reckon there will always be difficulties in prism generation, but there ought to be differences enough between the softwares/approaches to support at least a subjective preference. That's what I was after, so thanks for your contribution. Another thing, besides the mere generation of the prisms, is the possibilities for efficient y+-control. What do you think about ProAm in this respect?

The advantages of hexas are well known, but with a single cell size parameter (and local discrete refinement, of course), do you think you have enough control over the mesh density? How about the trimmed cell methodology. I've heard that the cells near the boundaries can be of rather questionable quality. Have you got any experiences yourself?

The opportunity to throw away PROSTAR entirely and replace it with ProAm makes for a tidy preprocessing environment, and I really like the idea. On the other hand, with a switch to ProAm, we would lose the direct CAD interface between I-DEAS and ICEM, which is a really practical feature. How are the surface preparation capabilities of ProAm? Do you believe the loss of the interface could be outwheighed by the other qualities of ProAm?

Regs.

Lars Ola

David Minns October 11, 2000 16:27

Re: samm vs. ICEM
 
Hello,

If you have a particular model that you are having problems generating prism then please speak to your local ICEM office. What version of ICEM are you using ? ICEM CFD version 4.1.1 has many new prism features to tackle complex CAD.

Another good method is to generate some prisms and perform the smoothoing, then use the diagnostic tools to see the location of the bad prisms. You can then vary the prism parameters in this area.

From my experience prism meshing is something that is not a trival task. You may feel my view is bias as I work for ICEM CFD but I feel it is one of the best solutions on the market. I would strongly advice you check with your support office that you are using the best approach.

Regards

David

Lars Ola Liavåg October 12, 2000 05:24

Re: samm vs. ICEM
 
Hello David,

As I mentioned in my first posting, I am basically quite satisfied with the STAR-CD/ICEM bundle that we have, and after using ICEM Tetra and Prism for more than a year (I am currently using v4.0.4), I have of course realised that it is a potent tool. The method you recommend for prism meshing has been the basis for my work for several months, and in November I will attend to an advanced training at CD's locations in London, so I am not exactly giving up here.

However, I don't have much experience with other meshing softwares, and ProAm seems to have many attractive features that ICEM cannot offer. That is why I started this thread: to find out the good AND bad sides of ProAm compared to ICEM Tetra/Prism.

Regs.

Lars Ola

John C. Chien October 13, 2000 21:38

Re: samm vs. ICEM
 
(1). I don't want to say that a commercial code is written in such a way that the only efficient way to find the right solution is through a friend in the support engineer group. (2). In reality, you can not separate the support eningeers from the codes they are supporting. It is impossible to put that personal feeling and experience into a general commercial code. (3). So, as a token to the forum readers, I would say that: The first step to use a commercial code is to get to know your support engineers. Know his name, his secretary, his phone number, his e-mail, and his FAX number. and Do Not Feel Shy to Get Help From the Support Eningeers. You can never figure out the missing part which is required to solve your problem in using a code. (4). It is useless to get help from the forum readers for the code and problem specific questions.

Lars Ola Liavåg October 17, 2000 06:05

Re: samm vs. ICEM
 
John,

Thanks for your comments, but I think perhaps you are misjudging my objectives a bit. I merely wanted to hear other people's opinions on samm/ProAm vs. ICEM Tetra/Prism, be they objective or subjective, ICEM or adapco support engineers, or just others. No more, no less.

If, as it may seem, further responses to my original inquiry are not posted, I suppose the forum readers have made a clear statement entirely on their own: either, they don't have experience with both softwares, or they don't feel inclined to respond for some other cause. Regardless of their reasons, this thread is then closed as far as I concern.

Lars Ola

John C. Chien October 17, 2000 10:04

Re: samm vs. ICEM
 
(1). Just want to add one personal example to the answer. (2). I have been working on a highly complex 3-D geometry recently, and I was very frustrated with the ICEM/Hexa ,because I could not get the good mesh generated. (3). But when I sent the files to the support engineers, they were able to solve the problem quickly and sent the file back through e-mail. Since then, I have been following the same topology and good meshes have been created right away. The time saving is many times, not just a factor of two. (4). In my case, I am dealing with real things, and I don't have the time to become an expert in the code like the support engineers. In the resultant geometry , I have 80 blocks and each edge curve is critical and must be optimized to derive the best efficiency. (5). The message is just a general guideline based on my personal experience. (people normally don't want to tell you a real good story, because they earned it the hard way. A good code is a code which can solve my own problem, not someelse's problem. Since the code does not solve my problem by itself, the support engineer's experience becomes very critical. Even with this story, it is still not possible to tell you exactly how my problem is solved,because then you need to know my problem in the first place. ) (6). I like your question. The point I am trying to make is that: To make the code work for you, you need the experience of the support engineers. This is because the road map is missing from the code. You see the road, but not the map to your destination. Support engineer is just like the AAA office where you can get the free map and route before traveling by car.


All times are GMT -4. The time now is 05:49.