FEA - Affordable Finite Element Analysis for Rhino

unhandled

#1

Anyone in this group using easy to use (and affordable) FEA analysis tools with Rhino?

I need to examine some physical capabilities of some designs. I’ve been looking around a bit and software do exist out there (…), but most of the more capable FEA packages seems to be way more expensive than my (private) budget would allow.

Of course my needs determines more exactly what software would be useful, so here’s what I think I need to simulate (during rotation). Think of a ten (10) ton heavy rotating steel clutch :

  • Rotate with high torque (say, 8 MW, 100+kNm at 700 rpm).
  • Friction from alternative bearing types: Plain bearings/surfaces, ball bearings (both axial & radial types), or needle bearings.
  • Estimation of heat generation from friction (see bearings above).
  • Heat convection and dissipation,
  • Tolerances (deformation allowance) for parts sliding against
    each other under radial torque as well as axial load.
  • Frequency/resonance (impact from unbalanced parts) etc.

Apart from optimizing the part design I need to arrive at reasonable “design parameters” for engineers, given input requirements (example above), and in some cases axial loads. With “reasonable” I mean useful for prototype dimensions.

Of major concern is how different types of bearing solutions (plain, ball, needle) impacts performance and design parameters, and thermal effects on tolerances.

I am not a mechanical engineer and no former experience of FEA analysis, which is why good learning material matters.

Mesh generation and CAD integration (iterative adapting of the design) seems to be an issue as well. Apart from pricing.

Anyone?


#2

Have you looked at Scan and Solve?

I know nothing about it other than it exists. We do all of our FEA in Solidworks.

Dan


#3

Thanks for the reply. Yes, I have looked at a number of different software, but many of them takes hours, even days, to examine, and prices are not always shown (well, then they’re most likely too expensive for me anyway…).

WP has lists of FEA software (and I’ve done my google home work before asking) but complex software takes a lot of time to evaluate… and for this reason I would benefit from summaries based on experience from others, “saying all the right things” given the needs I presented above ( :slight_smile: ), which could help me narrow in on which software to evaluate more closely.

// Rolf


#4

How about this:

Virtually stress test your design


#5

The analysis you can do with Scan-and-Solve is strictly linear stress analysis, so that’s really nothing at all to do with what you want, there’s no “low-cost alternative” for that kind of thing.

And it’s not just the software but the Engineer you need to actually get useful results out of it. I’ve seen what happens when some Engineering student using Solidworks does a simple stress analysis on a really basic aluminum part and doesn’t carefully set it up to make sure the result is going to be valid, and know how much work it was in Rhino to set up the model so that Scan-and-Solve would give a result that reasonably corresponded to how it failed under physical testing. Now apply that to trying to test friction, heat dissipation, resonance? Do NOT try this at home unless you can do a ton of actual physical testing to verify.


#6

Yes, only linear analysis isn’t going to bring me very far. Nor is it an alternative to upload the design to the cloud (@Willem). For novel non patented design it is recommended to not even have the computer attached to the internet…

For us who’s not being scared of learning the physics there’s some competent free open source software out there, mostly for Linux for obvious reasons (but I’ve found some for Windows too, see below). A common problem with the most competent open source software is the lack of good UI’s. Not that it doesn’t exist, but they are not very pedagogical & thus helpful (for an exception, see FreeCAD/Calculix below).

Examples:

  • CAELinux - It hasn’t been updated for a while, but the core software is available for Windows in these days, which is why they stopped updating the package.
  • Calculix (cross platform) - “The solver is able to do linear and non-linear calculations. Static, dynamic and thermal solutions are available”
  • Salome - Front end. Generic platform for Pre- and Post-Processing for numerical simulation. Cross-platform solution, open-source (source & executables). A standalone application for generation of CAD model, preparation for numerical calculations and post-processing of the calculation results. Platform for integration of external third-party tools.
  • MBDyn - Multi Body Dynamics simulation. Used with Blender or SciLab as the front end.

Rhino6 + Calculix (bConverged) = Rhino7?

Anyway, now back to learning FEA… :sunglasses:

// Rolf


#7

I hired scientist to do this work. Big companies use ANSYS simulation software. It is very expensive. Most FEA programs need 0.5 hour manual preparation of the model before it can be used. Scan-and-Solve is easy to use because there is no manual preparation. On the other hand, Scan-and-Solve cannot simulate extreme stress and vibration because it is linear. I have not used Scan-and-Solve, but I believe that it is useful for early stage of mechanical design. If you want perfect design suitable for large-scale manufacturing, you have to spend thousands of dollars for analysis of just one part.


#8

Well, the nuclear industry in France uses Salome/Code_Aster which is for free. It’s not like its not capable, but open source software is not always very easy to set up, not always the most friendly UI either, and it’s not integrated with the CAD tools you’d like to use. And one would have to import the CAD files and redo the pre-processing after every change.

Also, any FEA analysis needs pre-processing of the models as to simplify and/or cleanup detailed features like holes, fillets etc. Some software does that automatically, and others yet even record what was removed so as to be able to repeat the exact same cleaning process in the next round if the original model needs change, and so on.

But if those advanced preprocessing features is worth like 10.000 USD+ for most people is questionable. I will happily spend an hour, or ten, cleaning a model before analyzing it, no problem (well, probably like 90% or more of the world’s engineers - outside Europe and the USA and the rest of the West - actually do so, as we speak)

I know full well that basic commercial FEA packages (non-linear) cost like $10.000 and more depending on modules needed (Creo Simulation, for example). Frequency and Thermo analysis is usually separate modules, but the price says nothing about it’s capabilities compared to, well, other free software out there. It seems like its the integration of pre- and post process and user friendliness which is the biggest difference. And integration with the CAD software of preference, and in some cases windows versions are not available.

Also, some basic FEA analysis, if repeated by an amateur with combinations of set ups, would still indicate trends and which direction to go for more analysis, and so give basic hints about weak points in the design that can easily be reduced. After that one can let experts do final tests.

There’s a reason for why some CAD packages integrates simple FEA analyses (like Fusion 360) despite the level of skills need for high accuracy results.

Don’t scare people away from even trying (I’ve done way “too complex” things given my formal or even expected competence) . :slight_smile:

// Rolf


#9

WELSIM is an ease-of-use simulation software program, check it out at https://welsim.com.


#10

Hi @goeasyon ,
Do you know anything about pricing? The page with prices requires registration.

// Rolf


(Rodrigo Bárcena) #11

@RIL
Maybe https://www.simscale.com/ would work for you?


#13

Hi @RIL, WelSim is $9.98/year for now. I am not sure how much will be in the future, but it should be always very affordable.


#14

Hey,
Take a look at this table, which I just found.
http://feacompare.com

Mecway seems to be very interesting. It uses Calculix solver and costs 350$ for perpetual commercial use.
Some look at couple of features:


#15

The most important FEA feature listed in the feacompare website is nonlinear - large displacements. The Mecway program has this feature.


#16

I noticed from the compare list that Code_Aster (free) is a fairly complete package as well. It includes non-linear large displacements. Quality of analysis also matters. Code_Aster was developed for the French nuclear industry and has been under development for 25+ years. Support for Python scripts.

This software may be of special interest also for Architects. Code_Aster includes thermal simulations and geological analysis, concrete and even anisotropic materials etc.

Salome-Meca is used as the front end ( http://www.salome-platform.org/ ) which has tools for pre and post processing (meshes).

// Rolf


#17

RIL,

Thanks for posting this question about affordable, capable FEA. I think many of us would like to be able to do analysis on evening projects, but the cost of most commercially available software is a non-starter for those of us who have a wife and a family budget.
So it looks like the low-cost choices are Welsim for $10/year, Simscale (which is a wrapper for Code_Aster) for an unknown amount, or going directly to Code_Aster for free.

What did you decide to do, and if you have already started - what has been been your experience with the product so far?


#18

I’ve done some FEA experiments in Coder Aster / Solome. Code Aster can be quite powerful, but the problem was, at least when I used it, was the error and feedback is in French, for instance where mallage = mesh. Solome is translated pretty well, and runs pretty good.

AFAIK, Code Aster is used by the French nuclear consortium EDF, for doing stress analysis on their containment structures, as well as heat modeling on their nuclear fuel assemblies.

If you get into FEM, make sure your computer has enough memory, as it takes a lot of memory to create fine 3D meshes, but the solving is pretty memory-frugal.

The spring was fun, as I had heard that a coil sprint is more of a torsion spring than a cantilever spring-yet it’s another thing to see it, with the stress in the outside. The grandiant was caused from me not terminating the ends too smooth.

In these simplified coat hooks, the one on the right is lighter than either of the others, and doesn’t have the stress concentrator.

I wish that I could afford a copy of Scan and Solve, but Code Aster can FEM.