How to use 3D prints made of PLA for simulation

baumar

I used to create models of my parts for basic tests. A part of these tests are also stress tests. Now that there is this great possibility of simulation I would like to first calculate the effect of different load scenarios and then compare with the real values measured on the model. 
To do this I need to find 2 things: First I need to find the (generally) most robust way of printing a part from an stl. Then I would like to find the material values needed. I checked in the existing library but did not find any values for PLA, so I guess I need to start from scratch? 
Any input welcome

S1mon

I assume since you mention PLA that you're using FDM. This means layers, which also means that the strength of the parts will be anisotropic (i.e. it will be stronger in some axes than others). I don't think that Onshape Simulation can predict the behavior of anisotropic materials. If you really want to compare a 3D printed part to a CAD model, you're probably better off with SLA or SLS. As far as I know, they are still anisotropic, but less so. 

ChatGPT on this:

"Here's a rough ranking of some common 3D printing technologies based on their anisotropic strength qualities:

1. Selective Laser Sintering (SLS) - SLS typically offers relatively strong parts with less anisotropy because the powder material is fused together using a laser, creating strong bonds in all directions.

2. Digital Light Processing (DLP) / Stereolithography (SLA) - These resin-based printing technologies create parts by curing liquid resin with a light source. While the parts can be strong and exhibit good isotropic properties, the anisotropy in the Z-direction can still be a concern if the layer adhesion is not optimized.

3. Fused Deposition Modeling (FDM) / Fused Filament Fabrication (FFF) - These filament-based printing technologies are the most common and affordable, but they tend to have the most anisotropic strength qualities due to the layer-by-layer deposition process. Strength in the Z-axis can be significantly lower compared to the X and Y axes, which can result in weaker layer adhesion and part strength.

4. Binder Jetting - In binder jetting, a liquid binding agent is selectively deposited to join powder particles. While the process allows for complex geometries, parts often have weaker anisotropic properties due to the reliance on the binder for structural integrity. Strength can be improved through post-processing techniques, such as infiltration or sintering, but may still not reach the levels of other technologies."

billy2

I use SLS and material properties found on the web. I'm using OS analysis with success but I'm careful about what I do. 

FEA uses:
-is this design better than that design?, I never validate a design using FEA, you have to build a prototype to validate
-I use flexural modulus of elasticity for strength to match loading case, I never yank something apart, I never use tensile strength 
-Use the proper Poisson's ratio for the material
-Try to get the mesh correct

Once I get a prototype, I measure everything trying to understand material properties and dial my analysis in to be more predictive.

Linear analysis are:
-only good to 2% strain
-can't handle non-linear loading, materials & geometry (basically the real world)

If you're going to try and use FEA, then get use to analyzing, prototyping and then validating. You'll never want to enter into a design review with your analysis only. There are too many guys like myself that'll question your methods. Just one simple validation will put a smile on some old fart's face when reviewing your design and the approach you use.

There's a lot going against FEA for an engineering tool, but it's better than guessing.

EvanReese

S1mon said:

I assume since you mention PLA that you're using FDM. This means layers, which also means that the strength of the parts will be anisotropic (i.e. it will be stronger in some axes than others). I don't think that Onshape Simulation can predict the behavior of anisotropic materials....

Additionally, you'd have to account for your actual print settings somehow (wall count, infill settings, etc). To make that accurate, you'd need to model the infill in Onshape, which is probably preventatively onerous. There might be a better purpose-built software for this exact use-case though. I know that the Eiger slicer for Markforged recently launched a simulation beta, which I believe takes all of these things into account, but you'd be limited to their materials.

baumar

Thanks all for this great input! Of course I will always use a model to check the predicition as I did too many theoretical calculations in my studies ;-) But since the simualtion possibilities improved a lot in terms of quality and cost, it's worth to include them in the design phase to get an approximation. 

billy2

All the assumptions you have to make when performing an FEA, no wonder the results are never correct. Anisotropic properties of PLA? That's a good one. Does that affect the torsional stiffness of the part or does it affect the tensile strength? Will this not manifest itself during small loads and only occur at higher stress levels normal to the print direction? It's definitely a non-linear characteristic.

Is FEA accurate? Probably not, you just have to get comfortable validating your assumptions and never skip this step.

When I sold analysis tools I'd always ask how accurate is your guess?