Solidworks simulation

[Dylan]

Hey guys, I’m having some trouble in solidworks simulation. I’m trying to test the rigidity of the chassis. The process that I have applied is as follows:

Firstly, I created offset surfaces for each member according to the shell thickness because it kept giving me an error when I tried make offset surfaces all at once. Therefore, each the offset surfaces are grouped according the thickness of the member eg 1.2mm members are all together, 1.6mm members all together and so on.

Second step is I defined each shell according to the member thickness.

I then applied fixed geometry to the rear hard points and a remote load the front hardpoints.

It meshes fine but when it tries to solve, it gives me an error saying: “Preconditioner-fails- to-factor may not have sufficient restraints”. I looked up the error and it apparently means one of my members is ‘floating’ and it’s not attached to the chassis.
Is the process I have outlined fairly accurate? Does anyone know what I’ve done wrong?

Any help would be much appreciated.

Thanks

[CameronBeaton]

e in solidworks simulation. I’m trying to test the rigidity of the chassis. The process that I have applied is as follows:

Firstly, I created offset surfaces for each member according to the shell thickness because it kept giving me an error when I tried make offset surfaces all at once. Therefore, each the offset surfaces are grouped according the thickness of the member eg 1.2mm members are all together, 1.6mm members all together and so on.

Second step is I defined each shell according to the member thickness.

I then applied fixed geometry to the rear hard points and a remote load the front hardpoints.

It meshes fine but when it tries to solve,

My guess would be (never used solidworks, but I have done simular things in other packages) that the the surfaces need sewing together, this would need to be done after step 1 but would probably make step 2 difficult so you would probably have to apply the thickness to sections of the mesh.

Alternativly heres probably a function that removes duplicate nodes that you could use at the end but you probably have to mess around with the mesh alot.

[wagemd]

You might want to look at the SW tutorial on creating weldments. You can create a 1.2 and a 1.6mm profile and it does (most) of the work for you.

Using weldments also puts simulation into beam mode automatically, which is a good thing for chassis analysis.

SW also has a Youtube channel, and one of the parts they demo is an FSAE chassis. You can get a pretty good idea of how to use weldments with that.

[Dylan]

Thanks for the reply guys.


surfaces need sewing together, this would need to be done after step 1 but would probably make step 2 difficult so you would probably have to apply the thickness to sections of the mesh.

How did you sew the surfaces together in the packages you have used?


Using weldments also puts simulation into beam mode automatically, which is a good thing for chassis analysis.

Yeah the chassis is built using weldments, the problem is, a remote load (which I'm using to apply a moment and the suspension hardpoints) can't be attached to beams, which is why I have had to shell the whole chassis.

[CameronBeaton]

Originally posted by Dylan:
Thanks for the reply guys.

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> surfaces need sewing together, this would need to be done after step 1 but would probably make step 2 difficult so you would probably have to apply the thickness to sections of the mesh.

How did you sew the surfaces together in the packages you have used?

</div></BLOCKQUOTE>

There was a sew tool, you just selected it and then selected the the shells. I have just opened solidwork and in the surfaces tabs there is something called (in spainish) 'coser superficie' which is what you what I think

[Dylan]

There was a sew tool, you just selected it and then selected the the shells. I have just opened solidwork and in the surfaces tabs there is something called (in spainish) 'coser superficie' which is what you what I think

Ah yep found it. Would it cause issues if I sew to surfaces together that have different shell thickness?

[Joe_Little]

Thinking back, it seems that during the beam mesh creation there was a process where I "reviewed" the beam mesh and made sure all of the beam are connected properly.

Based mostly in the KISS philosophy, I would try to make it work with beams, but that's just my opinion. I feel like a shell analysis would lead to a lot of unrealistic stresses at tube junctions, but perhaps could give reasonable hardpoint deflections if that's all your after.

I also did chassis analysis where I wrote in notepad enough xyz beam points from solidworks to define the chassis and created them as "points" in a dedicated FEA program then connected the dots with beam elements. It turned out to be a more work but a lot less frustration, probably better results too.

[Dylan]

Based mostly in the KISS philosophy, I would try to make it work with beams, but that's just my opinion. I feel like a shell analysis would lead to a lot of unrealistic stresses at tube junctions, but perhaps could give reasonable hardpoint deflections if that's all your after.

I also did chassis analysis where I wrote in notepad enough xyz beam points from solidworks to define the chassis and created them as "points" in a dedicated FEA program then connected the dots with beam elements. It turned out to be a more work but a lot less frustration, probably better results too.



Thanks for your reply mate. Yeah I understand the shells will cause some problems at the junctions, but as you said, hopefully it will give fairly accurate hardpoint results.

I like that idea of recording the xyz beam points and then using and fea program to connect the dots with beam elements. Which program did you use to do that?

[Joe_Little]

Abaqus. (Hopefully no one from MSC checks these forums!) This should be straightforward on any of the big commercial FEA packages.

If you don't have one readily available, you can download Student Edition Nastran from MSC's website, which can handle this fine. I'm sure other companies offer Student Editions as well, or if you're like me, your school is a Abaqus powerhouse where every 3rd guy is a Abaqus expert and Abaqus is on every computer, or wherever other FEA program.

But... if you go with Nastran I could justify helping you make the model as this would make a good write-up to put in my collection of MSC/FSAE materials. I might even make an Adams flex body from it. Email me at joseph.little[at]mscsoftware.com if you're interested.

[Paul Becker]

It's just a matter of bonding each of the shells together.

[Dylan]

It's just a matter of bonding each of the shells together.

Hey mate, thanks for the reply. How did you go about bonding the shells together in solidworks?

[RobbyObby]

Originally posted by Dylan:

Yeah the chassis is built using weldments, the problem is, a remote load (which I'm using to apply a moment and the suspension hardpoints) can't be attached to beams, which is why I have had to shell the whole chassis.

You don't need to apply a "Remote Load". Simply apply a "Force/Torque" as you would if you were applying it to a solid face, but select the "Joint" input and enter the magnitude and direction of each load, applying it to whatever node you are analyzing from (whether it be the inboard mounts, or outboard suspension nodes). I think they explain this method in the YouTube tutorials, as Daniel mentioned.