I am guessing the recommendations are more numerous for the Ansys, Abaqus, and Comsol because of different criteria. And I do know that Ansys and Abaqus are regarded more-so as "high-end" products esp. in the nonlinear realm (rubber/hyperelastic deformation, plasticity, and so on). They are reputed to have very "robust" solvers for nonlinear. Also, another criterion might be that SolidWorks Simulation will take-away a license of SolidWorks CAD (in addition to the SolidWorks Simulation license). The other products are known for running "stand-alone" which is an advantage if you have other users that need to use the SolidWorks CAD independently of Simulation. On the other hand, I have never seen a product as well-integrated as SolidWorks Simulation. The Simulation add-in uses similar management tree, picking entities, and wonderful "help" files. I have experienced ANSYS 11 and I can say the "help" is not nearly as user-friendly as SolidWorks Simulation. Another criterion is the target market: SolidWorks Simulation is targeted at Designers and Engineers in what SolidWorks Corp. calls the "mainstream" crowd and does handle basic linear, thermal, and dynamics analysis quite well; whereas, the ANSYS, ABAQUS products are targeted directly at dedicated analysts and do quite well at handling nonlinearities. The retail price reflects this targeting to some extent. You should find that ANSYS can run upwards of $70K (and perhaps ABAQUS). To compare apples-to-apples in terms of capability (i.e., types of analysis) you do need to make sure the comparison is fair in terms of what you need to do in the analysis realm. I know ANSYS has a slew of element types (upwards of 20, I believe), while SolidWorks Simulation is somewhere around 5. Well, that's my two cents. -Tony
Good summation of a complex subject for Hijazi. For the record, I believe you can get a usable seat of ANSYS for around $20K, as that is what I use for most of my analysis. But I'm sure you can put in the $70K that you referred to, as we are just doing structural analysis. No thermal, no explicit solver, no lots of other bells and whistles. And, in fact, now that I think about it, we got a particularly good deal because some of the marketing literature mistakenly claimed some capability for our package that it didn't actually have, so we ended up getting an added module that we didn't pay for.
Oh, and I believe ANSYS has more like 200 element types.
For what it is worth, ANSYS has a good reputation for general structural analysis. Abaqus has a good reputation for non-linear analysis. Comsol is known for multi-physics.
Message was edited by: Jerry Steiger
I recommend it for simple analysis such as basic stress analysis. Here is my experience with it.
- I ran stress analysis. No issues but I did find that there is limit to the amount of elements you can run.
- I ran vibration analysis, a lot of bugs. The set up was easy and straight forward (this is what you pay for, and it’s very nice when everything works). This is not the case with complex models. Your biggest struggle will be going around/through all the bugs. In one case it would not solve for me because it just did not like a mesh of a single part, I had no clue which one it was so here is what I had to do: suppress most of my parts and run a simpler model (less components) until it would solve and it did. And then work my way up a part at a time and solve it. This took me days to accomplish but it worked. It does not take much to learn how to use the software but it will take time to learn how to go around the bugs.
- I ran Shock analysis. Similar to vibrations.
- I ran thermal analysis (conduction). Same as the vibrations, it had many bugs with it. There is a system bug for couple years now, there is a button to pick all surfaces to apply a condition that you want. Well it worked great for heat sinks. Well now it does not work. I have to pick each surface one by one (I was applying convection). Other things I learned is that you better freeze the model when doing heat transfer analysis or else your file will become corrupt and you will have to start all over again. Another thing, I was not able to open older thermal models (previous versions), files were corrupt. I was able to view last results but could not work on it. On the current models you have to save it as a different files name and only work of thermal part of it, if you start changing the geometry it will most likely become corrupt. Such as you will not be able to mesh the model normally (this is the 2012 version).
I have not tried their Simulation Flow. Looked at the tutorials and it looks great, set up look straight forward and easy, as expected. But again, solid works is great in creating user friendly and intuitive controls. But I am afraid to purchase the package because If it s as good as other simulation products then it will have bugs that I will have to learn how to go around and there is no tutorial book for that.
Just to give some pro’s about it. The learning curve is by far day and night difference. You will not be able to master any FEM or CFD software quicker than Solidworks Simulation. It’s integrated into SolidWorks, and it just very intuitive. It took me a couple of day of seminars to understand the functionality. We had another guy just pick it up and work it, I think he ran into single problem. As far as time goes it will take a person at least twice as less time to learn SolidWorks Simulation than any other package out there. (only comparing ANSYS, COMSOL and SolidWorks Simulation)
Brings me to price tags. I am gathering quotes for multiple CFD software packages right now that we are considering: Comsol, Solidworks Simulation Flow, NEi Nastran and ANSYS. So far Comsol came out the least expensive solution. I am still waiting for ANSYS. But I am sure it will be more than COMSOL. But in the end they are all around the same price range.
I will be running trial version of all of them. I already ran COMSOL trial version. But that’s for a different topic.
In the end it’s not a very strong FEA package but it’s very easy to use and intuitive. Would I buy it again? I would only consider it when most of the bugs are worked out, so far I have not seen it. It’s like working with a pre-released software package.
The CFD package is brilliant. It is robust, fast, and the easiest and most straight forward CFD tool I have ever used and I have used more than a few. It is essentially the FloEFD program that mentor graphics sells, just rebranded and distributed by SWX. For the physics it handles (which is not everything) the stuff in my experience produces the same results in quicker time than any unstrucutred code you will find. The b'layer handling is first rate and frees the user from hanving to deal with the b'layer. It is the best product in the whole SWX line up.
for the money the SIm products are a pretty outstanding deal. Yeah they have some issues as was mentioned. However, you can avoid a significant number of them once you get a handle on the secret decoder ring aspects. It isn't that tough in my humble opinion, though I do sympathize with some of the frustrations - they happen. The big issues is managing your own expectations as it was all so mind blowingly easy to put the thing together and then you have some issues and not a lot of help to sort them out. I call the SWX sim product a geometric based FEA pre processor as compared to the classically nodal based one. They are evidently not the easiest things to pull off. The other major vendors are all working on them.
got no time for more.