how to find crack initiation in gear tooth root

[LMGK]

Hi all,
I am doing gear tooth failure analysis for that i could know the following in fracture mechanics:
1. How to find crack initiation and crack propagation ?. (i found some articles about crack propagation but still i dunno about crack initiation, i mean to find a, where it initiates and what r the parameters associated to crack initiate)

I hope someone have a solution or ideas. thanks in advance.

regards,
lmgk

 

[EricZhao]

Maybe I am wrong, but I don’t think there is a crack initiation theory in crack analysis. Usually it is assumed the crack is initiated by a small imperfection on the surface. You have to assumed the worst case - crack at where the max stress intensity factor is. Or if you got some gear already failed, examine the crack location of your gear.

 

[GBor]

I would suggest that you send this question to the metallurgical forum (forum330)

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.

http://www.borowskiengineering.com

 

[LMGK]

Thanks Eric and gbor. sorry guys if i am in wrong forum. I will post my query in the mentioned forum too. my main aim is to find the partial damage of the tooth root as function of torque and speed in planetary gear. still i dont have an exact idea to do it, but before i would know all the parameters which related to fracture mechanics. am i right ?. thanks

lmgk

 

[GBor]

Is this a fatigue issue, or an overtorque? If fatigue, you will need to know the S-n (or Goodman/modified Goodman) curve which predicts cycles to failure for different metals under certain cyclic loadings. Generally, the prediction comes from the stress state induced by the peak cyclic loading compared with the number of cycles anticipated to result in a crack. This data is generally empiracle and is often provided by a metals handbook.

If it is due to an overtorque, you need to know the post yield behaviour up to ultimate strength and rupture. This would require a pretty good knowledge of the entire stress-strain curve, including the strain-hardening portion.

This is really a pretty complex problem...more complicated than anyone should try to fully explain on a forum. You may want to find a metallurgist in your local area and have a long talk with him.

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.

http://www.borowskiengineering.com

 

[gwolf]

You can tackle this in two ways:

1) "Safe life" in which you use an S-N curve, appropriately factored from test to design the gears to survive a certain number of load cycles.

2) "Damage Tolerant" in which you assume that there is an initial crack (corner quarter or semi-circular). For the initial crack size you choose the largest crack which would NOT be detectable by the inspection methods used for the component. You then use fracture mechanics to see how many cycles it takes to grow the crack to the critical crack length.

In either case you need to know the stresses at the tooth root and this is difficult. You can either use an FE model or resort to empirical gear tooth design software. I personally prefer the empirical approach in this case because it is based on a lot more real world data than the FE. With FE you run into problems in determining the stresses around tooth-tooth contact points.

You will also have difficulty in determining the real loading on the tooth because gearboxes flex (well, the lightweight aerospace ones do) and will often bias the load and contact towards one end of the tooth. This is a significant effect if the gearbox/gear ensemble has not been designed to generate even contact at the most commonly used load condition.

You are into specialist analysis here - consult experts better than me locally as GBor suggests.

 

[feajob]

EricZhao,
There are two important theories for fatigue studies:
1) Stress Life Theory (S-N) approach estimates total life without distinguishing crack initiation from crack propagation.
2) Strain Life Theory (E-N) approach is called as the crack initiation method.

AAY

 

[LMGK]

thanks to all, actually my ultimate aim is to correct the tooth profile to reduce the tooth root damage as a funtion of applied torque and speed. For that i have to know the parameters such as fatigue life, durability, crack initation, crack length, etc.
I have tooth root stress values which are taken from MSC Nastran linear static analysis. I like to know at what torque and speed (normally load distribution along the face width may be constant or variable) , the crack initiates in the root and minize it by profile correction which is my ultimate aim as i said before.
I hope the theories and ideas which u people suggested could help me. Let me go thro it and see what is says. thank u guys.
We dont have msc.fatigue module which is easier to find the fatigue life. Instead of that i am looking alternative method to do ie. MSC Nastran/Ansys/or ..(i dunno)

lmgk

 

[rb1957]

LMGK,

i'd be surprised if the crack growth portion of your damage is significant, compared with the crack initiation life. if this is what you're interested in, (the crack growth portion of the crack life), then your DT would start at a non-detectable flaw size (something like 0.05" is tyipcal in aerospace to represent manufacturing (or in-service) damage, or 0.005" for typical material imperfections). I'd think that you could do your fatigue life as a hand calc, you have a simple spectrum (i think, 1 stress cycle per gear revolution). maybe something like ncode could analyze your entire model ...

anyways, i'd also be surprised if linear FEM can give you an accurate stress concentration stress, i'd expect plasticity at the root of the gear tooth.

modifying the gear tooth profile is a way to increase fatigue life, remember sometimes taking material away can improve the fatigue life (ah, the magic of fatigue !).

consider too, surface treatments (maybe you can shot-peen a non-contact part of the tooth), surface hardening, cold working (again in a non-contact area) are also ways to improve fatigue life. also, maybe you can modify how the gear is used; possibly changing the time-profile of how the gear is loaded can help (if the load is rapidly applied, maybe this'll increase the stresses on the tooth).

good luck