m8kwr
|
| posted on 24/6/09 at 08:02 AM |
|
|
Solidworks experts
I am looking to do some stress anaylsis on a chassis i am designing, i was under the impression that in Solidworks you could do this.
I am using the 2009 version, does anyone know how to go about this, or give me any tips and tricks of how to get started, and then hopefully i should
be able to figure the rest.
I believed in 2008 there was cosmos.
|
|
|
|
|
andyd
|
| posted on 24/6/09 at 08:08 AM |
|
|
According to this 2009 has it built-in.
Edit to say that link does seem to be about 2009 but the screen shots show 2007 and COSMOSXpress. 
[Edited on 24/6/2009 by andyd]
Andy
|
|
|
m8kwr
|
| posted on 24/6/09 at 08:12 AM |
|
|
I just found that out myself... always the way.
I will now search the internet finding help of how to use it.
Does anyone how much load is put onto a chassis - approximately
|
|
|
balidey
|
| posted on 24/6/09 at 08:28 AM |
|
|
This has been asked many times before, and I often write this type of reply.
Unless you have a LOT of experiense using FEA and you know what you are doing then don't expect your results to be any good.
I have been using SW for over 11 years. I know a brief understanding of FEA (being modest  ) and I know that I only limit myself to analysing small
simple parts. NOT full chassis. To do that then you need to be a bloody good FEA expert.
So, have a play by all means, but don't use your results for anything oter than making nice coloured pictures.
And I'm not trying to be rude, its just that Cosmosexpress makes it so easy to use FEA that it fools people into believing they are doing it
correctly.
And the first rule of FEA (or is it the second) is that FEA is only a tool to help you do less tests. ie if you have a bracket that you could make out
of 3mm, 4mm, 5mm, 6mm, 7mm, 8mm, 9mm or 10mm thick steel, instead of making and testing each one, you use FEA to narrow down your options, ie it can
tell you to ignore 3mm to 6mm as they are not within the FOS, and therefore you can make and test the other thicknesses.
There, that makes FEA sound less appealing now doesn't it?
The salesmen at my work love bringing customers to me anf giving the old BS about FEA, so I set up a quick demo, show them the coloured pics, they
love it. But that all it is. We usually farm out complex work to specialist companies, who often need several days with us to figure out loads and
restraints.
EDIT TO ADD: theres nothing wrong with using balsa wood to make a scale model. The late great Allan Staniforth was a believer in this, and its true,
it does work, costs almost nothing and helps you visualise the frame well.
[Edited on 24/6/09 by balidey]
|
|
|
wescottishmatt
|
| posted on 24/6/09 at 08:42 AM |
|
|
As above but I don't think the COSMOS express works vey well, you need to get the full COSMOS sysytem to get any decent results.
We use Solidworks but we use ANSYS for FEA which is not a cheap option and second the comments about you need an expert to get the best results
although as a comparison tool you should be able to get some positive feedback.
|
|
|
cadebytiger
|
| posted on 24/6/09 at 09:00 AM |
|
|
i like the balsa wood model idea. Might be quite good fun too
|
|
|
m8kwr
|
| posted on 24/6/09 at 09:01 AM |
|
|
I think you are completely right, i was after pretty colours...
And how the hell am i going to know what force are going to but onto a chassis, i am no expert.
I have done a few test, on simple box shapes, and got it to bends the metal depending on the loads, fixing points etc. This may come in useful to
seeing if i removed beams, or added, how much this affect the overall design.
I may put into solidworks the 7 chassis, at least then i have something to compare it against.
|
|
|
JekRankin
|
| posted on 24/6/09 at 09:13 AM |
|
|
If you're looking to compare the torsional stiffness of the two chassis, then I think the load you apply is arbitrary.
I would try fixing the rear suspension points and applying a moment of say 100Nm to the fronts. Torsional stiffness is measured in Nm per degree, so
if you can measure the angle of deflection across the front pickup points, you should be able to work out the torsional stiffness.
Ps - if Solidworks allows you to simplify your analysis using beam elements, it will solve lots quicker!
|
|
|
balidey
|
| posted on 24/6/09 at 09:35 AM |
|
|
The load becomes very important the nearer you get to the yeild point of the material. So putting 100Nm into it may give similar results on two
different chassis, but put 1000Nm in and it may well give very different results. Then it could be that the design of one chassis over the other
handles the load transfer better, so you may not actually be comparing like for like. Example, say you have a 7-esque chassis, and you also have a
super stiff Le Mans GT racer chassis. Putting 100Nm into both could possiblly give similar stress analysis results regarding torsional stiffness. But
put 10x that load in, then you notice the differences.
And what load to put in? What are we really looking for? Suspension forces? If so then its probably close to mass of the complete car, times whatever
vertical G you may get, about 3G? I have no idea. Then you have the engine putting its load into the chassis, again you need to allow for its maximum
load, which I think is when you shut off the throttle (chassis has been laoded with the torque from the engine, then imediatly shut off, but i have no
idea how to figure that out into real numbers). Then you need to allow for it all to be bending due to steering input, again probably a couple of G
side. All at the same time, all interacting with each other. And thats just the loads. The restraints will be just as difficult. I have seen people
use wrong restraint types without realising their errors.
Good luck 
|
|
|
balidey
|
| posted on 24/6/09 at 10:37 AM |
|
|
Just thinking about this a bit more regarding force inputs and restraint locations.
One major force input has got to be the suspension points. This is where major loads enter the chassis.
And to restrain the chassis you would need to select the points that define the location of the chassis and allow it to move, ie the suspension
points.
So, how do you deal with that, the load (or atleast some of them) and the restraint points are one and the same.
|
|
|
m8kwr
|
| posted on 24/6/09 at 10:55 AM |
|
|
I was just thinking that myself...
|
|
|
hellbent345
|
| posted on 24/6/09 at 11:34 AM |
|
|
i was lookingat doingsomething similar with the hcassis i did, and cosmos on solidworks is very poor for this kind of work. Iwent to a cosmos demo day
and they saidthere is no way cosmosexpress could do that in a million years, the full version would possibly be the only way to do it, as you can
change to beam elements. This is because xpress changes the part into a series of triangular shapes, nodes, and to get them small enough for the model
to work would mean having to calculatemillions upon millions of them.
Also its a lot hard to test a chassis than you would think, restraining faces and force loadingfaces arent where you expect them to be etc. Basically,
dont use solidworks for the analysis of chassis, use a dedicated beam loading FEA program (there are free ones) and even then treat your results with
a hefty handful of salt.
Al
|
|
|
JekRankin
|
| posted on 24/6/09 at 12:26 PM |
|
|
Another way of looking at the problem could be to constrain all the suspension points and feed inertial loads into the chassis at the location of all
the large masses? - i.e through the engine mounts, seat mounts etc. How you could do this accurately I have no Idea!
As has been said, I think a simple beam model could be used to compare the stiffness of a couple of similar designs, but anything more than that is
probably very difficult!
|
|
|
luke
|
| posted on 24/6/09 at 04:41 PM |
|
|
Im doing a degree in CAD using solidworks and used it briefly toward the end of last year.
I modelled a trolley jack and tested it by constraining the wheels and the ram and applying a pressure to the top surface.
Couple of tips, think long and hard about the type of force applying. I originally placed a force and got really wierd results. Converted it to a
pressure on a face and got better results.
As people have said take results with a pinch of salt. the jack i modelled was rated to 2 tonne so i placed a two tonne load on it. I had huge
displacement on the main arm and pivot points. It scared me because i used on of these at home!
I then discovered solidworks automatically scales up the results to make them more obvious. so it had multiplied the results by over 10 times. Once
this ratio had been reduced to 1:1 i had a none deformed jack to display!
be prepared to run the model at least 10 times to tweak it and to smack ur head on the desk alot!
|
|
|
