High angular displacement torsion bar failure analysis 1

[JacobP12345]

I'm testing two different diameter (1/2'' and 9/16'') torsion bars made of tempered 4140 (tensile strength of 200-220 ksi, tempered at roughly 725 degrees Fahrenheit) for the purpose of acting as a lid assist for heavy vertical opening concrete lids on a subterranean enclosure. The expected maximum shear stress expected to be present in both bars generated from torsion is roughly 120 ksi for the 1/2'' bar and 140 ksi for the 9/16'' bar. Both are 36'' long (this is the maximum length the heat treating facility we outsource to can treat for this treatment process). I am trying to approximate the number of cycles until permanent plastic deformation is experienced and the torsion bar begins to yield. The angular displacement per cycle is from 0 - 90 degrees. I have attempted a rudimentary analysis myself but cannot find tables that include values necessary for a strain-life analysis or stress-life analysis when considering tempered 4140. Currently i have approximated it to roughly 950 cycles for the 1/2'' bar for reference, but I am not confident about this.
Additionally, the time between cycles is variant. During testing the time between cycles can be kept for uniform when opening and closing the lid, but in the field it is expected for the torsion bar to be subjected to maximum stress while the lid remains closed for up to a year until it is opened again, and then within a 24 hour period closed again. In terms of this cyclic rate i have no idea how to approach the analysis. Thank you for any help and suggestions in advance.

 

[blacksmith37]

The application sounds like it could be wet; The high strength steel would be prone to hydrogen stress cracking (also known by other names) if there is any corrosion .

 

[JacobP12345]

Thanks for the help, that is one of our concerns but currently we are remedying this via coating the material. So far it has proven effective. We are currently concerned mainly about mechanical fatigue of the material over time.

 

[rickfischer51]

If Im looking at this right, you're at 45 RC with a yield strength of about 195 ksi in tension and half that in shear. So a shear stress of 120 ksi is gonna yield it for sure, and I think stress life is not valid when yielding has occurred. You say you want to estimate the number of cycles until permanent plastic deformation occurs. I think one cycle is the answer.

The link below gives strain life data for 4140 at 40 RC. The stress ratio R=-1. Nominally you have R=0, meaning you have a non-zero mean stress. For high cycle fatigue you would have to correct for this, but low cycle fatigue is typical insensitive to mean stress, so you're ok.

http://barsteelfatigue.autosteel.org/Bar/Docs/Iteration_69_Final_Report_rev0505_4766209.pdf

Rick Fischer
Principal Engineer
Argonne National Laboratory