cyclic loading beyond yield strength - spinning disk 2

[nanobot29]

Here's the problem: a disk made Inconel 625 (no heat treatment) was installed with an interference fit on a rotor. Since the disk was not heat treated, the hoop stress will reach the yield strength while its spinning. The yield strength is about 65 ksi and the hoop stress is expected to reach 70 ksi (von mises stress will reach about 80 ksi).

My assumption is that the disk will become permanently deformed at the inner surface and as a result, will loose some of its initial interference pressure when the rotor returns to a stand still.

my question is: what are the long term and short term effects of cycling a material to about 10% above its yield strength? does the disk continue to expand non linearly after every cycle? or is the loss of preload the only major consequence?

any insight would be greatly appreciated.

 

[Cockroach]

Suppose the disk fails during operation. What would it cost the company now? If during the investigation your doubt and second guessing is called into question by some shrud lawyer, what would you say, how would you react? Given you have a specific engineering doubt, how is it possible to suggest or at least explain it away as being "small"?

Take another read of the Challenger loss and o-ring performance at a temperature just below acceptable limits. It's only a piece of rubber, they'll warm during operation and make the necessary seal.

I wouldn't want to be in your shoes, brother.

Regards,
Cockroach

 

[nanobot29]

dheng - the analysis actually was done before the parts were ordered. i looked throw the analysis and the approach looks accurate. The analysis calls out for heat treated inconel 718 to a yield strength of 120 ksi min. The problem was that the designer forgot to add the heat treat spec to the drawing and the checker and engineer failed to catch the mistake. so the parts were ordered and the machine was put together.

The application is for an electric motor. about 2 Megawatts of power. It goes through one to two start/stop cycles per day.

we do have a few senior engineers here and they agree with my stress analysis. however, they too werent sure what the long term effects would be. I appreciate your offer in looking over my analysis but I'm confident that that my results are good and show consistency with FEA results. but i really do appreciate the offer. i took your advice and asked for more time to triple check my numbers and study up on plasticity

 

[nanobot29]

Thought some of you would be interested in an update on the situation....

I took the advice of many of you and demanded more time to do a proper in-depth analysis. I even showed my bosses this thread and it freaked them out too see so many comments of how risky this application is. Thanks for all the comments and suggestions.

Since i dont have any test data to go off of, i took two different approaches. FEA and hand calc. For FEA, i ran a non-linear elastic/plastic analysis which allows me to input an actual stress-strain curve. I simulated 1000 start/stop cycles to 25,000 rpm. The results show that about 15% of the disk undergoes plastic deformation near the inner diameters. It also shows a loss in interference pressure of about 10% (which is acceptable in the design). Most of the pressure loss occurs after 3 full cycles. Pressure loss continues after every cycle but quickly decays and stabilizes within a few dozen cycles. The residual stresses are low enough that cracking will not occur.

I used an analytical approach as well that i found in "Advanced Strength of Materials" by JP Den Hartog and the calculated results were in agreement with FEA results.

Regarding fatigue, i found S-N curves for the exact material in the annealed condition that contains data for continuously stressing well beyond yield. The expected life is > 10^7 for the stresses I'm expecting to see assuming an R ration of 0.5.

One of my graduate professors took a look at my analysis and agreed with my approach and results. He teaches a course in plasticity.

i presented my analysis and results... managers decided to move forward based on an accelerated test were we will shrink fit an annealed disk to the same interference fit and cycle the part for a few hundred cycles, then remove the part and measure the deformation and inspect for damage.

Thanks again everyone for the advice and suggestions

 

[Compositepro]

I haven't been following this thread in much detail. It seems to me that if a spinning disc approaches its yield stress due to centrifugal force, it will instantly fail. Force will increase with yield which will cause greater yield, etc.

 

[GregLocock]

OP - I think you've done a good job on this one, and you should be able to sleep again, at least til you get the results from the accelerated test!

Compositepro - I think you are being overcautious, plastic deformation at the rotor to shaft fit can be made to work, so long as the outer part of the rotor remains sufficiently elastic.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?