Looking for a cheap, easy to use FEA soft with easy assembly import 2

[GGuillaume]

Hi,
I'm looking for a software that could match for working on ultrasonic transducers.
I used to perform dynamic analysis with Abaqus (displacement vs frequency) using Piezoelectric elements.
I also used Ansys Designspace for modal analysis and it's been great for CAD assemblies, unlike Abaqus (CAE)...
The input was the electrical potential and the output was the displacement, stress, etc...
I heard about Algor, which seems to be not too expensive and with a nice GUI.
Can someone give me some advice on Algor and on some other softwares ?
Thanks a lot

 

[Drej]

If it's opinions you're after then thread727-120567 is a good place to start. Personally I think Workbench sounds ideal for what you need at the right price.


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[teb1]

If you are purchasing this software for yourself or a small company, I don't think you can beat Algor for ease of use, tech support and solid model importation.

Linear static analysis with Algor and Solid Model importation from Alibre has come a long way.

For the money invested you can't beat Algor or Alibre.

My 2 cents worth.

teb1

 

[GGuillaume]

The most important for me is to perform some steady-state dynamic analysis. Defining damping has taken a lot of time (by using Rayleigh damping)to fit result with measurements.
For this reason using some input files I created in the past using Abaqus CAE would be great.
Here's the point: Algor claims that it's possible to import files from Abaqus to Algor...

Has anyone ever tried to do this kind of things (opening a *cae file with Algor or another software)?
Thanks

 

[jcfoley]

GGuillaume,

The best answer for this is it to check it out yourself. Algor will supply you a full trial copy of its software for you to test (as will anyone else). I have found that most companies will tell you that their software can be the cure all but only your evaulation will tell the truth.

I went through and looked at Cosmos, Algor and NE Nastran before finally settling on Algor (but just barely over Ne Nastran).

Chris Foley,
Midland, TX

 

[GBor]

I've moved back and forth between .nas (Nastran) files and Algor. It's usually pretty good. What is generally lost has been on the Nastran side (I don't mean to criticize Nastran here, it was actually FEMAP that redefined the element types from shell to plate).

Don't know what they would do with an Abaqus file, but send them one and ask them to send you a graphic of what the import looks like in Algor. Then ask what properties came through. Algor can print out a pretty detailed report that is fairly easily read of what the model information looks like...ask them to send it to you (it's in html format, so you should be able to pull it up in your web browser). If you're not satisfied, ask them to run it (make it fairly simple, but include some mixed element types).

What kind of "steady-state" dynamic analyses are you running? Is this just a modal or a frequency response?

 

[GGuillaume]

Hi GBor,

I used to run steady state direct dynamic calculation on Abaqus, because I didn't achieve to make it with the "submodal" way.
I used to run dicrete frequency response : calculation with complex displacement calculated on 360° phase for each frequency step.

But modal response is better because a transducer works at a resonance frequency. It's faster compared to a frequency response since I don't need to determine the resonant frequency by increments.

 

[fkmeyers]

I would recommend NEiNastran. I do not believe the Alibre with Algor can handle assemblies or anything past linear statics analysis. Usually, they offer this at a very low price as an introductory package.

Our experiences with NEiNastran have been generally very good. We looked at Algor and they were actually more expensive so I think if you are looking for a lower price option that includes nonlinear and assembly support you may want to consider them. And it is not difficult to use. Certainly easier that some of the other packages you have mentioned which I have also used.

As for support that seemed to be emphasized a lot on theses threads. My recommendation is to evaluate and decide yourself.

 

[dmacx]

I agree with fkmeyers. We also use NE Nastran here at AFRL. We also considered the same products and went with NE Nastran. NE Nastran (like NX and MSC) supports both modal and direct frequency and transient response solutions. It also supports residual vectors which I highly recommend for modal response analysis so that you do not need to request extra modes to get the static responses for rigid type structures. In "my" opinion Nastran has the best support for dynamic analysis.

 

[jcfoley]

Frank,

Just FYI, I have used Algor v17 on assemblies (linear dynamics mostly). I am currently working on a nonlinear MES problem for a small assembly.

Chris

 

[GBor]

Alibre can handle assemblies...not sure I would recommend it as there are MUCH better CAD packages. From an FEA standpoint, however, Algor does just fine. It has joint creators for universal and pin joints that automatically establish the rotational motions that you may want. Alibre basic comes bundled with Algor...my preference is SolidWorks or AutoDesk products for the CAD side.

Both NENastran and Algor handle modal response dynamics well. If Algor is more expensive for this type of package, go with NENastran. Get a quote and get a demo. Run some basic modal response analyses on each of them and see what you get. I think the answers will be very similar.

At that point, you can be confortable going with the low-bidder.

Garland E. Borowski, PE

 

[fkmeyers]

Well put Garland.

Chris, as I understood, the Alibre + Algor package does not support assemblies, only single parts. Of course Algor supports assemblies but the cheap version you get with Alibre does not. Maybe this has changed over the last year.

 

[jcfoley]

Garland,

Well put again, buddy.

Frank,

I may have been a bit confused. Alibre ships with Algor's package. Basically, I have a limited version of Alibre with the full Algor software (Linear, NL, and MES) I forgot that Alibre has a VERY LIMITED version that ships with Alibre. This is the same as the version of Cosmos that ships with SolidWorks (& that version of Cosmos does not do assemblies either).

I have a real problem with people "doing FEA" without having any background in it what so ever. I am sure that you (as well as others) would agree.

GGuillaume,

I hope that we have not confused you. As you may be aware, some CAD packages are now shipping very limited FEA software (Algor & Alibre, Cosmos & SolidWorks, and now from what I hear Inventor & Ansys) with the CAD packages. It seems that some companies have decided that FEA has "become so simple that anyone could do it." This is not the software to which I am referring.

I would reiterate on test the software yourself.

Chris

 

[GGuillaume]

I really agree with you Chris, considering CAD softwares offering crapy useless FEA software. This is marketing...

However, FEA packages can sometimes offer some good in-build CAD (Abaqus does, for sure). I've done several geometry modification with Abaqus allthough it was on an imported STEP file, which means you have to "cheat" to eliminate some chanfers and other geometry details.

But the point is we've got Pro/E so I don't expect to get some good in-build CAD from a FEA soft.

And I will test the softwares myself, that's sure. I wanted to know which one I should evaluate.
I've allready wasted some time evaluating FEMLab (more mathematics oriented in my opinion).

 

[teb1]

GGuillaumne,

I'll stick with my original statement. Algor and Alibre can do just about anything that you might want to accomplish with regard to modeling an object and performing an FEA analysis of it. But you have to be looking at the full versions of both software packages to get the maximum capabilities from both programs.

I am currently running several models with over 250 parts to them. My largest model to date had over 312 parts in it. These models have contact and friction.

Most of my experience (18 years+) has been with Algor so I am a little predjudiced.

Regards,

teb1

 

[vikingfea]

We recently purchased NEiNastran for Windows and have been very impressed with the product overall. The support has been very good from both the dealer and directly from Noran. We considered Algor and Ansys as well. The decision to go with Noran was based on price, reputation, and most important our evaluation of the product itself. In our case we asked them to assist with the building of a complex nonlinear assembly of a worm gear which rotates over 720 degrees. For us the process to generate this model was the easiest in NEiNastran. Ansys also seemed capable and was a close 2nd in our decision process. Algor was not considered seriously for multiple reasons.