14
$\begingroup$

Since I'm running a 3D printing facility of an engineering school, students are always wondering how much infill percentage affects the stiffness of the part. I know that it is impossible to get a numerical solution for this question, but maybe there is an option to simulate in software an already sliced model. I haven't seen in any slicer an option to export as .stl or .step or any other format which can be accepted by simulation software. Has anyone seen or thought about something similar?

$\endgroup$
6
$\begingroup$

I don't believe that slicing engines create any sort of solid model that would be useful for CAD simulation. When a slicing engine slices a 3D model, it's goal is to spit out the preferred machine paths in G-Code (of some kind). However, I've read a few articles, done some tests, and heard through the grape vine that anywhere between 10%-35% is good enough for most applications. I once watched a webinar for understanding the new MakerWare interface that explained how they chose such settings. Although I can't find the clip directly, here is the page for all of MakerBot's webinars. I think this webinar was the one I watched explaining a little bit about preferred infill percentages.

From experience, anything over 35% doesn't yield much more strength from infill side of things. Beyond 35% and you're going to want to reconsider how you're orienting the print when you print it and what you're printing to utilize the grain structure for proper strength.

However, infill percentage/patterns are not the only variable for creating strong parts. Infill is really just a way to save time and material. Here are some other ways to potentially increase strength:

  • Increase your shell. Shell is the number of profile patterns per layer. Typically (in FDM/FFF), each shell is about the diameter of your extruder nozzle.
  • Increase your floor/roof. Similar to shell, floor/roof refers to the number of layers that make up the "bottom" and "top" of the part with regard to the build plate.
  • Print orientation. Pay attention to which areas of the part are susceptible to strain along the "grain" of the layers. Try to rotate your part on the build plate in a way that minimizes potential failure both in print and post-print use.
  • Post process. Don't be afraid to do some post-processing to increase the strength. There are some 3D printers on the market that go as far as including Kevlar strands in the printing process to beef up their prints. However, it may be as simple as just coating the part in an epoxy with some basic finishing techniques. It's a bit more work, but it turns weak 3D printed parts into full production quality prints.

Update: Based on some of the comments, it sounds like your best bet might be to find a custom application that can either convert the g-code file into a solid model (try CAM software?), or create a plugin for your CAD software (I know Unigraphics NX and Solidworks allow for this) and essentially recreate your own slicing engine that takes your solid model and generates the same infill pattern dynamically inside.

Perhaps look into the works of Simlab or similar which has a lot of 3D software plugins. I'm not promoting them and I don't work for them, this is just a reference of what to look for.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks for link, I'll check it out. For engineering applications we need precise data, at least it is required by project description. Engineers can relay on some experiments and experience, but it would be much more reliable source of information, to slice your part and then by simulation prove that this parameters will withstand such a load. $\endgroup$ – Anton Osadchy Apr 24 '16 at 21:37
  • $\begingroup$ You could potentially print your own tensile bars and test different infill % in-house. However, it will greatly depend with the source of the material and shell parameters. $\endgroup$ – tbm0115 Apr 24 '16 at 21:41
  • 1
    $\begingroup$ For bending loads, which are the primary failure mode in real printed parts (because they concentrate tension on the surface of the object where the layer lines present plentiful fracture initiation sites), adding perimeters is significantly more impactful than adding infill. This behavior is kind of masked in standard tensile tests, but you'll see it in a four-bar bending test. $\endgroup$ – Ryan Carlyle Apr 24 '16 at 22:30
  • $\begingroup$ Thanks for elaborating on perimeters, I forgot to really emphasise that $\endgroup$ – tbm0115 Apr 24 '16 at 22:34
  • $\begingroup$ Of course with bars it can be done by printing test bars. But problem is that shapes are always different, and there are no possibility to print thousands of test parts for each of projects. So that I though about software which can simulate stiffness of already sliced part. $\endgroup$ – Anton Osadchy Apr 25 '16 at 8:12
3
$\begingroup$

Since I am not able to comment on this question yet, I thought I would provide an answer in addition to the already helpful insight provided.

If the question in general is regarding infill percentage, and the common follow-up regards part stiffness, then it should be explained that choosing infill percentage is much more than just part stiffness.

Printing out tensile bars would be a great thing to have for educational purposes. The bars should not only vary in infill percentages, but in also the various infill patterns. Depending on the type of stresses and loads applied, different patterns may be stronger at lower infill rates, for example.

Also, the infill rate should correlate to the thickness of the bottom, top, and sidewalls. This is especially important with ABS when it comes to shrinkage, warping, and delayering. In order for the part to be universally as strong as possible, as the part cools, it must shrink evenly. This is a well known factor for machinists creating molds for injection molding and casting. Otherwise, you may have unintended additional points of weaknesses arise.

Lastly, if creating tensile bars, make sure to take into account the shrinkage experienced along each axis, for each individual example. I would also suggest cutting each one open, as well as attempting to break a few of the apart (in a very crude way). This could stimulate a lot of thought when designing a part, before printing it.

| improve this answer | |
$\endgroup$
2
$\begingroup$

Anton, I parse the G-code and build a finite element model and a thermal transient event to simulate the printing of the part, followed by a structural simulation to determine the deformation and stress state in the resulting part. This part could then be further analyzed with external loads to determine mechanical characteristics. I use ANSYS software and the element birth & death capability to activate a small volume of material per each thermal transient step. The structural sim is static but also done at the same time steps as the thermal transient.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Wow! That sounds very complicated and is probably a bit too advanced for most of our users. It would be very interesting if you could tell us what you learnt from these simulations (some rules-of-thumb and perhaps some statistics) because not many people will be able to replicate this. $\endgroup$ – Tom van der Zanden Jul 16 '16 at 16:21
  • $\begingroup$ Thanks man! I have an experience with ANSYS, so I can try it. The only question is - How did you imported g-code to Ansys simulation? $\endgroup$ – Anton Osadchy Jul 17 '16 at 9:59
1
$\begingroup$

I could think of a way. But it might require a few softwares to get everything done.

First get the CAD file. Import to magics (Materialise proprietary software) There is a function for structures, you can build your custom internal structure. So add trusses etc. Export stl. (There is one software which allows direct stl to step conversion, I think its called Instep) Or i think you could reduce mesh density and use FreeCAD to convert it back to step and run your analysis.

It would be intersting to see a report on the part. Also there is no good way of simulating a 3d printed part. Maybe the closest would be with composites.

| improve this answer | |
$\endgroup$
0
$\begingroup$

If you're an engineering class, you may very well print out big cubes of different infils and test them.

| improve this answer | |
$\endgroup$
  • $\begingroup$ The OP mentioned in some of the comments that testing is not feasible as new tests would need to be conducted for each new part. The part geometry and requirements vary such that a few tensile bars would not cover the engineering requirements. Perhaps elaborate on the process you mentioned to potentially aid others in the future? $\endgroup$ – tbm0115 May 2 '16 at 14:02
  • $\begingroup$ This should be a comment. $\endgroup$ – Tom van der Zanden Aug 4 '16 at 9:05
0
$\begingroup$

You have to do it by eye. However you need to think about what infill you use. a triangle pattern will be very strong.

Following I never use high infill. Most of my items have a strong outer shell, IE 3 layers. Inside I will do between 7 and 14 %. If I printed a square 200mm^3 I would have no concerns standing on it.

All depends. Really though for your settings I would not thing you would really ever need more than 14%.

At a company I worked at that was testing 3d printed parts. They evaluated material by printing the same test shapes and seeing what their tolerances were. You will need to develop your own method as such.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.