FEA result

[melissa.tabandehjoo4YFEJ]

Hi everyone,

I am a new user of Inventor, I modeled a Feed hopper to start FEA. One of my colleagues is a professional user of Solidworks. he modelled the hopper to do a pre-analysis to see if I was on the right track. the result shows a big gap between two software (Inventor is almost twice above Solidworks in displacement values!)

Please find attached the both software reports to see if anyone can guide me as to what is wrong and where this gap comes from.

Regards,

Melissa

[JDMather]

Attach the actual Inventor and SolidWorks files here.

[Frederick_Law]

First they don't even have same mass.

Second, IV has 2 material: A516-70 and 316SS.

SW only A516-70.

Then material spec for A516-70 is not the same in IV and SW.

What's else are different?

If you want to compare, compare the same thing.

[CCarreiras]

---

@Frederick_Law wrote:

First they don't even have same mass.

Second, IV has 2 material: A516-70 and 316SS.

SW only A516-70.

Then material spec for A516-70 is not the same in IV and SW.

What's else are different?

If you want to compare, compare the same thing.

---

The name of the materials are different, but the density, yield, and other properties values are very similar, so i believe that is not a reason to double the values.

I guess it's the kind of boundary conditions regarding the constraint, but it's just a guess, only by comparing both models we can say something valuable.

[melissa.tabandehjoo4YFEJ]

thanks for the link, can I find the reasons through this channel?

[melissa.tabandehjoo4YFEJ]

thanks for the response, but the mechanical values are the same, so we are looking for a reason that could make such a big differences.

[pcrawley]

@melissa.tabandehjoo4YFEJ - the very first reply to your post is the most likely to yield (pun intended) an answer not based on guesswork.

Without the two models to compare, it is impossible to answer the question.

[melissa.tabandehjoo4YFEJ]

hello everyone, thanks for your feedback. Please find attached the two models in Inventor & Solidworks. I really appreciate your time and support.

Cheers,

Melissa

[JDMather]

Unfortunately your version of SolidWorks is later than what I have and therefore I cannot open your SolidWorks file - but it appears to me that your Pressure is applied differently in SolidWorks than in Inventor.

As a side note - without seeing the SolidWorks file I would also say that the Inventor file is poorly modeled with unconstrained sketches and far more work than necessary to create this simple geometry.

I can recreate the Inventor geometry if you would like to see how I would create it.

[melissa.tabandehjoo4YFEJ]

Thanks for sharing your Channel, I watch them. they are awesome videos. as I explained I am new that's why the model is poor. do you think that can be the reason for the big gap? it would be much appreciated if you recreate the Inventor geometry and we do the analysis to see the result and compare with Solidworks. Cheers, Melissa

[JDMather]

In your Sketch1 you are missing a Tangent Constraint.

You are also missing a Coincident Constraint between the center of the arc and the vertical axis of revolution.

Since I can't open your later version *.sldprt file - can you post screenshot of expanded browser so that I can get an idea if it was modeled correctly?  Something like this Inventor shot...

[JDMather]

@melissa.tabandehjoo4YFEJ 

This is how I would model the geometry.

I would expect to see a Fillet along this edge.

I did not take the time to set the Material.

Note only one fully defined sketch (could have used 3 sketches, one for the Revolve, one for the Split and one for the Hole).

Note the simplicity of the operations.

https://www.youtube.com/playlist?list=PLp5izJt_zvN0KUccGh5OrgbNTZ_c4vYxs

[melissa.tabandehjoo4YFEJ]

@JDMather 

much appreciated. the tips are really helpful.
I analyzed your model and the result is the same as mine, which is still different from Solidworks!
besides, we need to apply a Fixed Constraint on the bottom of the hopper as below snipped, so the cylinder should be cut from the bottom and a rolled plate attached to it

I look forward to your comments

[melissa.tabandehjoo4YFEJ]

@JDMather in order to find the reason of differences between Solidworks and Inventor, I modelled a simple Cantilever Beam and applied a force load at the end. Both software result were exactly the same as hand calculation. I also modelled a square hollow section beam, the Inventor result was %0.2 lower than the hand calculation and the Solidworks result was almost %7.5 higher which both are acceptable. Considering the above two samples, it shows us that the result of applying force load is the same in both software. So, we give the probability of error in applying the pressure load, hence we modelled a thin sheet and apply the pressure load. the displacement in both software is the same, but safety factor and stress are different. Please find attached the model of both software to see if anyone can guide me what is wrong and where this gap coming from.

Cheers,

Melissa

[cadman777]

Have you tested this FEA example in NASTRAN or some other FEA dedicated software?

Incidentally, I just looked at your model. If it's a pressure vessel, and if you're making it out of sheetmetal or thin plate, then according to standard pressure vessel handbooks, it's not modeled correctly. For example, no joints are shown (metal connection + weld prep+ weld). Also, the dished head needs a lead-in flange. There are other considerations as well. So if you do an FEA on a tank that isn't properly designed, I wonder how it can represent anything in terms of real world conditions? Just a thought...

UPDATE:

Find attached another way to model this tank. The flat side could've been made like the cylindrical side (stitch-cut on the laser table & roll formed as a full sheet & then full cut along the stitch-cut line, then fit-&-weld). The assumption is the tank head was purchased from a head mfgr. Note: I didn't look-up and add any of the weld preps and welding beads.

I don't have FEA in my version of Inventor, so this model hits my limits in this thread.

[Frederick_Law]

I cannot open you IV file so I created it and compare to your SW (which I can open):

Every little bit of different in model and setting adds up.

Both IV and SW are NAFEMS validated.  So slim chance there is software problem.

So model and sim setup will be the reason for different results.

Different in mesh definitely will give different results.

[Frederick_Law]

[Frederick_Law]

Safety Factor is compared to Yield.

Hopper and Test plate has different material with different Yield and Tensile.

What did you used in Inventor?

[melissa.tabandehjoo4YFEJ]

na, just Inventor and Solidworks. Thanks for the model, I did FEA on your model, it still doesn't match the solidworks result.

[cadman777]

I really can't comment on the apps.

The reason I posted that model was to show you how this would be fabricated by shops around here.

I didn't make any of the weld-prep and weld-beads, which should be added for FEA purposes.

You don't only test the plate, you also test along the joints, nozzle connections and support appertenances.

But the formulas and rules in the PresureVessel design books may help with all of that.

Most pressure vessel books address cylindrical tanks, so the info in them would have to be adapted to your design.