Algor vs Ansys 1

[newtofea]

Hello guys,

Our company management is seriously considering to purchase a FEA package for our design engineers. We currently use Autodesk Inventor and have evaluated ANSYS/DesignSpace, COSMOS DesignStar and Algor product for Inventor.

I primarily like ANSYS as my background is more in ANSYS. Have never used Algor. But, that has nothing to do with this purchase decision. Cosmos design star is out of the race.

My question:
What are the pros and cons of DesignSpace vs Algor?
Any input will be appreciated as we will be making a decision pretty soon. (In a month's time!)

My thoughts:
Both are pretty easy to use. Both have Good CAD associativity with Inventor. Algor is lesser in cost that DS. Don't know much about meshing capabilities.

Assuming price is not an issue (I mean between DS and Algor), can any one of you guys tell me which is a better value product considering our FEA needs are as follows:
1. Linear stress analysis
2. Easy to use
3. CAD associativity
4. Linear thermal.
5. Modal analysis.

I will sincerely appreciate your advise. Both the sales guys from ANSYS and Algor have been impressive, but want to hear from the real users. Thank you in advance.

Sincerely,

Rob.

 

[Drej]

Your questions:

Assuming price is not an issue (I mean between DS and Algor), can any one of you guys tell me which is a better value product considering our FEA needs are as follows:

1. Linear stress analysis
2. Easy to use
3. CAD associativity
4. Linear thermal.
5. Modal analysis.

The ANSYS Workbench Environment (as you call it DesignSpace) has been developed entirely for people like yourself. The meshing capabilities are outstanding (better than ANSYS itself - it uses a different mesher) and the solver is very good. The 5 types you mention above are very much "bread and butter" to AWE, and will be seamless in terms of CAD integration and ease of use. This software is a pleasure to use in terms of its GUI, and it's very intuitive meaning a short learning curve also. There are, however, cons as well as pros regarding this software. If you intend to use it, get the latest release (v8), as it is significantly more stable than previous versions. Also, although the solver is very robust, it can enforce fairly non-conservative tolerances for analyses (to aid in achieving full convergence). As long as you've been made aware of the limitations (as with any software) there will be no problem.

-- drej --

 

[newtofea]

Thank you drej.

I understand that ANSYS is probably the better solutions, but all that ANSYS claims to do, so does Algor at a cheaper price! So, how do I convince my management? Any tips?

Rob.

 

[Drej]

Apologies if this sounds biased, but I have been using ANSYS for a few years now, and I respect the code.

In my humble opinion, ANSYS is a very well established, respected and supported code, owing to many years of development. In the past I have used many FE codes, including ABAQUS and Nastran, amongst some others which were well below par. The amount of time I have spent in the past chasing the root of "strange" answers has been phenomenal, which doesn't please the client or the management, and hence a need to minimise this. If I were to justify a new FE code, I would do so on the basis that, of course, it has to be cost effective but also that it doesn't have indirect or secondary cost penalties attached (for instance learning time required, solver robustness, CAD integration, IT burdens, ease of post-processing, poor technical support etc.). Going on personal experience, I would always pay a little extra for the satisfaction of knowing that you have the above.

All in my humble opinion, of course.

-- drej --

 

[JGalliera]

I have heard that Algor reports do be able to do a lot, but in the end, they come up short on a lot of capabilites by just not being able to do them well. They will lower their price at the drop of a hat to try to rope you in, and if you do decide to purchase it, you will shortly be wishing that you hadn't.

Just out of curiosity, why did you eliminate COSMOS DesignSTAR from the race? I have used it very successfully for awhile now (1.5 years) and have been overall very happy with it. DesignSTAR will probably fit into your management's budget as well.

 

[mohr]

Hi Rob !!

I am using Ansys, MechanicalToolbar and Design Space, in a mechanical analisis, fatige analysis,etc
I had problems with geometry import in Ansys and MT. But
with DS, I do not how, I entered in the magic world. No
problems at all, you can do a mesh configuration with a special feature named relevance. You can design a very
efficient mesh. The solution time is short an the accuracy is excelent due to own solver.

I worked with Cosmos as add-in with Solid Works. Then exported the mesh done in Cosmos to Ansys. Ansys detected
a lot of elements with negative Jacobians. It was impossible
to work with Cosmos mesh.

It is my little and humble experience

Regards

 

[craigj]

Depends heavily on what kind of buisness you are in. At my job we use a lot of imported industrial design type surfaces, which can be a serious pain in the ass is you dont have a robust mesher. If you analyze more machined type "analytical" objects in the linear regiem, then one of the cheaper codes like Algor or Cosmos might work well.

That said...if you want the best code use Ansys. I have used all of the softwared mentioned above (Algor, Cosmos, Ansys WB) an have concluded that Ansys makes a better overall package. Plus they have been around a long time which means they have the stability and cash flow to keep improving and supporting their software. With Ansys you get the peice of mind that if there ever comes a time when you need to add more capibility the sky is pretty much the limit...I mean you can even get into explicit dynamic stuff in Ansys using the LSDyna add-on.

If for nothing else Ansys mesher is worth the extra couple thousand dollars...I cant tell you how many countless hrs I spend trying to get a good mesh in Algor or Cosmos.

OK you know where I stand. Good luck, Craig

 

[craigj]

PS check out the thread in:
Finite Element Analysis engineering Forum
topic: ansys classical vs workbench

some good info

 

[JGalliera]

CraigJ wrote: "if you want the best code use Ansys"

I am sorry Craig, but that is your opinion. I have also used all the codes that you listed and I have found COSMOS to be a very good solution for me. I did my research on SRAC, the company who develops COSMOS, before investing in them and I found they have been around since 1982, so they have a lot of experience in FEA software as well. Now that they are owned by SolidWorks, they have a lot more backing to innovate and develop their software. I am looking forward to future advances coming out of COSMOS.


CraigJ wrote: "With Ansys you get the peice of mind that if there ever comes a time when you need to add more capibility the sky is pretty much the limit"

COSMOS also has a high-end code called COSMOSM, which is very similar to Ansys Professional, that I use in addition to DesignSTAR. I can do dynamic analysis with this software as well. The database that I create in DesignSTAR can be recreated in COSMOSM by a simple export, where I don't think Ansys can do that with their DesignSpace to Professional codes. I can easily set up the problem in DesignSTAR and then recreate it in COSMOSM and then make the few necessary changes before running the analysis. One reason to use COSMOSDesignSTAR is that I can do non-linear analysis in this interface, whereas Ansys DesignSpace cannot.

"If for nothing else Ansys mesher is worth the extra couple thousand dollars"
I do not concur. There is no mesher that is a no-brainer point-and-click. As a conscientious analyst, I must interrogate my mesh before I am comfortable with the results it produces, no matter what mesher I use. COSMOS gives me the tools I need to do this interrogation of aspect ratio, Jacobian, element size and compatibility. Ansys I believe has similar tools, so you cannot really say one is better than the other. Rather it depends on how the user works with the tools available in the software. I would argue that perhaps you didn't investigate more deeply into the meshing tools in COSMOS. The auto-mesher in COSMOS makes it seem too simple, I agree, but that's not all that is there- it goes deeper. In my experience, I found it also to be much faster and more robust than the mesher in Ansys.

So to summarize, I feel that both Ansys and COSMOS are both very good FEA codes. I found that I can do everything I need (from static linear to non-linear, to dynamic, post-dynamic, fatigue, thermal, thermal stress, contact, electro-magnetics, fluid flow, and more), I can do with COSMOS products, plus at a much more affordable price than Ansys. Without COSMOS, I wouldn't be able to use a FEA code because I cannot afford Ansys, and I don't feel I lose anything by using COSMOS instead of Ansys. This is just my opinion and I think they everyone has a right to their opinion, but I feel that it should be created themselves. Obviously, the person who started this thread concluded that COSMOS was not for him, and my original question was along the lines of why he came to this opinion. It probably was because he has a background with Ansys already, but I wanted see what he had to say about it in more detail the reasons why "COSMOS is out of the race."

As far as Algor, I used it many years ago (about 1997-8) and I didn't like it at all. Whatever I hear about it now comes from another source, and usually it is not pretty what they are saying. Last thing I heard was the other day that their tech support is really bad. The response from Algor that this person got was: Didn't you read the manual?

 

[Drej]

Just a few comments re. JGalliera:

"The database that I create in DesignSTAR can be recreated in COSMOSM by a simple export, where I don't think Ansys can do that with their DesignSpace to Professional codes. I can easily set up the problem in DesignSTAR and then recreate it in COSMOSM and then make the few necessary changes before running the analysis. One reason to use COSMOSDesignSTAR is that I can do non-linear analysis in this interface, whereas Ansys DesignSpace cannot."

Firstly, if you only have a DesignSpace licence, there is no need to create an ANSYS db file since you can't run ANSYS with this licence anyway. When an analysis is run in Workbench (as it is now called) an ANSYS database file - as well as the corresponding input deck in the common database format - can be created on the fly as necessary. Secondly, ANSYS Workbench has all the standard non-linear capabilities, namely material, geometry and contact within the Professional licence remit. Reason being of course is that if you're a company dealing entirely with linear analyses, the DesignSpace licence gives you a tailored and affordable option with a great deal of functionality.

Finally, as craigj pointed out, the mesher in ANSYS is outstanding. This is as close as it gets to point and click. However, as you state, it is of course entirely correct for you to question and to interrogate your mesh. But the difference between an average mesher and an outstanding mesher is that the outstanding mesher (1) minimises massively the time required to produce exactly what the analyst pictures in his/her mind but also (2) gives one a sense of confidence in the solution, both of which lead to greater cost effectiveness and increased productivity: the holy grail in today's market.

All in my humble opinion, of course.

-- drej --

 

[JGalliera]

Drej:

Thank you for informing me about the latest on the Ansys Workbench interface. It's been about 2 years since I last used Ansys Professional, and they were still using the older interface.

Using you definition of an outstanding mesher, IMHO I believe that both Ansys and COSMOS has one.

 

[szhang]

Today I read the discussion. I’d like input my opinion (I am using Algor now).

1) For comprehensive FEA software (not specify on only some areas, such as stress, thermal, fluid, etc.), Ansys is No 1 and Algor is No 2 in popularity (how many companies are using it)
2) There are many detailed differences for Ansys and Algor. But overall, both of them have same (or similar) functions.
3) Algor is much cheaper.

Shu

 

[newtofea]

The main reason why we eliminated Cosmos was its ineffectiveness with Autodesk Inventor series of products. So...

By the way... as the world being so unpredictable..... we haven't yet made any purchase decisions and we are still at the same place as we were when I started this thread!

Still waiting for budget approval.

But I greatly appreciate your inputs and still looking forward for more answers if there are any!

Rob.

 

[JGalliera]

I use all different CAD tools, such as Autodesk Inventor 8.0, SolidEdge ver. 15 and SolidWorks 2004. They all work great with the latest DesignStar 2004 (2004/035) release.

There was a small change that you had to make with in Inventor to work properly with previous versions of DesignSTAR, but the new update works with no problems. The small change was to change in Inventor the version of ACIS(SAT) file that it saved to version 6, so if you had made this change by going to their knowledge base, then you would have had no problem.

Link 1 shows you how to change the Inventor SAT file out to version 6 so that associativity between DSTAR and Inventor works:

http://www.cosmosm.com/cgi-bin/test/knowledgebase.pl/Show?_id=349

Link 2 has the release notes for the latest version of DSTAR and how it now supports Inventor 8:

http://swdownload.solidworks.com/pdownload/cosmos/COSMOSDesignSTAR/version40/DS40Release.doc

You should not have eliminated COSMOS so quickly and I think that you should re-examine your choices and include COSMOS in your decision.

*********
On another subject:

Recently I showed COSMOS DesignSTAR to someone who has used Algor for the past 10 years, and he was amazed at the power and ease-of-use of DesignSTAR. One point he made was about how it is much more difficult to apply boundary conditions in Algor as compared to DesignSTAR. The way he explained how you have to do it in Algor, it doesn't sound like it changed from when I used it many years ago. The speed of the FFEPlus solver in DSTAR knocked his socks off! He also mentioned that Algor's tech support is regional now, and that the person who supports your region has other responsibilties other than support so you don't always get the person on the phone right away. Also I thought that if you only have to rely on one person, maybe that person doesn't know how to do what is your bread-and-butter analysis.

 

[newtofea]

Dear JGalliera,

Can you explain more about "One point he made was about how it is much more difficult to apply boundary conditions in Algor as compared to DesignSTAR."

Why is it difficult to apply boundary conditions in ALGOR? I saw that DesignSpace and DesignStar, both were extremely friendly in applying boundary conditions.

Another reason why we eliminated COSMOS:
We have huge Assemblies, and Cosmosworks was more cumbersome as it had no "automatic contact detection" like that of DesignSpace. (They say they have it now! Not sure!)
Also, they failed to mesh our part on a first go. We had to do lot of tweaking with Cosmos to make it finally work. DesignSpace on the other hand meshed the whole assembly with automatic contacts in less than 30 mins. That really was impressive for me. Cosmos requires "matching nodes" which inturn made the mesher fail. This was profound in the modal analysis we ran.

Also we were impressed to see that DesignSpace has an option of Fatigue module. We are not purchasing right now, but I do see fatigue as a problem for our products on a long run. Would be nice to have that in the future.

So, there were several things that we observed. There are not too many things that will justify us to buy Cosmos over designspace. A couple of thousand is not what we are concerned about. Its the overall feel of the product and I think DesignSpace really did that. I personally did not get a chance to play too much with Algor, so cant make similar comments about Algor and is still in the race as our other user used Algor in his previous life.

Waiting to hear your comments.

Rob.

 

[newtofea]

After much discussions and ado... I am glad to update you folks that we have indeed purchased Designspace and our experience so far is "excellent". Especially the local VARs are really helpful. Tech support is really great.

I sincerely appreciate all your help. Every CAE software in the market, I am sure has its own goods and bad. But I think for our application, ANSYS serves our needs the best.

Another big drawback that I found with Cosmos is our local VAR does not have any one with sound FEA background in their TECS. They are more a CAD house than FEA. The ANSYS VAR on the contrary seem to have tons of experience even as consultants!

Thanks to one and all who replied to this post.

Rob.