fatigue for simple-to-make-but-hard-to-get-right structure

[PerKr]

For some time now I have been testing a new door stopper design for our products (refrigerators) and I figured I would see if an FE analysis would have helped us do this faster. However, I have one problem: the impact force is not known and neither is the impact deformation of the door stopper and the hinge (the door stopper is attached to the door and stops against the hinge). There is no equipment available here with which to measure this and I don't have access to any software which could help in obtaining this information either.
I figured I could just use a force and do a simple comparison of designs, but what if I wanted to do this 'for real'? I have a couple of designs which failed after half the test and one which passed the test with flying colors (but which is way too expensive), should I select a force based on this?

 

[rb1957]

you mention a test ... presumably someone has defined the service requirements, or are you guessing a number.

if it's the former (i doubt it tho') then there's no problem (hence my doubts), so i assume you're guessing your test loads. why not test an existing design and interpret the result based on the known performance of the test piece ?

i think too that you're trying to simplfy the testing. FE will allow you to compare designs under a nominal load.

back to testing, i think you should test several specimens (at least 3) at a couple of load levels (hi, medium, and lo).

good luck

 

[PerKr]

you are right, the requirement is known and includes the impact energy (although it is defined as angular speed and door load) and number of impacts (30 000).

So I should start with a design I know failed the test and modify the force, right, but what force do I set, considering I can only perform a static analysis? Would it be fair to set it high enough to reach the yield strength of the material for the failing door stops? Or should I try to find the Wöhler curve for the materials we are testing and find the stress value for ~15000 cycles and then dimension the force to reach this stress value?

 

[rb1957]

i would start with a design i Know something about , either it's not strong enough or it's too strong/expensive, and go from there. maybe you can redesign a small component (the one that failed), maybe change the machining of a detail to reduce a stress concentration.

what's your constraint ... can you do lots of FE modelling to reduce the amount of testing, or lots of testing 'cause you're limited on the structural analysis side ? ... no dig intended, but it'll help with suggestions.

btw, it looks like your load spectrum is a safe life (10 openings per day over 10 years, in round numbers). so you should get a test result of about 100,000 cycles to meet the requirement.

 

[Mech151]

I would recommend determining the peak load using hand calculations. Then perform a static FE analysis on the part to determine peak stress. Just make sure this stress is less than the endurance limit of the material, or at least under the sn curve based on your cycle count. Since you have performed test to failure, you should also be able to calibrate/check your hand calculations and overall method by performing an FEA on the part and showing stresses in excess of yield (post failure analysis).

 

[GregLocock]

You could use a dynamics program to estimate your impact loads - there is a demo version of Working Model 3D that would handle this quite nicely (good luck finding it). You'd need to know how stiff the stopper is, and the inertia of the door.

Then feed that peak load into your FE model.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[GBor]

There is a product on the market called AMPS (Advanced Multi-Physical Simulation) and you can download a demo. Not sure how large your model would be, but if it stays under their demo model size, the software looks pretty impressive.

http://www.ampstech.com

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Lower Alabama SolidWorks Users Group

 

[PerKr]

I've been working on the analysis today. I started out with using soft virtual parts (SVP) replacing the structures to which the door-stop is attached (it's sitting in a slot in the top plastic part of the refrigerator door, fastened with a screw). Then I started messing around and realized that soft and rigid virtual parts (RVP) seem to stick to the part I'm analysing and that a contact virtual part seems to work better (but I'm still using an SVP for the screw)

anyway, how about this: if I managed to measure the deflection of one door stop, I could use this to find the proper force. Of course, I will have to find something to measure with. Either a real instrument or just something which deforms easily (and permanently)...

 

[GregLocock]

Modelling clay

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[rb1957]

you could use a dial gauge on your test (to measure deflection.

to measure the you could load up your detail thru a load cell, and apply load untill you get your measured deflection

 

[PerKr]

I actually used a digital caliper for this. of course, I couldn't do a proper measurement but at least I have an approximate deflection at the maximum angle which the door will reach during testing. Think I will have to make some sort of attachment for the caliper so that measurements can be performed during the actual testing.

 

[btrueblood]

Don't forget to vary the amount of crap that is sitting on the refer door shelves...

 

[PerKr]

So I used my values in the GPS, but encountered a problem. The maximum stress value is above my Rm! So, obviously, my restraints are faulty. What boundary conditions would you use? The door stop is L-shaped with a relatively large radius in the corner. The horizontal part is attached to the door with a screw.

Something I didn't take into account is the fact that the part to which the stop is fastened actually deforms (although the measurement was taken between the screw and the vertical part of the doorstop).