Crack Propagation in ANSYS 1

[reesh]

I'm going to be working on a problem involving the propagation of a postulated crack in an unpressurized cylinder due to cyclic loading. I'm going to need to calculate the growth rate and the point at which the crack becomes unstable.

I've never performed an analysis such as this in ANSYS and I am currently looking for examples/tutorials that I can use in preparing myself for this project.

Any ideas/places to look? There's not much that I could find in the ANSYS manuals.

Thanks!

 

[prost]

If I were you, I'd start with AFGROW, the crack growth analysis tool, which has models based on finite element solutions for thumbnail (curved that is) cracks in unpressurized cylinders. This program is free, courtesy of the US Air Force, and can be obtained here:

http://www.siresearch.info/projects/afgrow/index.php

Even if you continue with the ANSYS project, AFGROW results will give you a good reference and verification of your ANSYS results.

 

[torpen]

Hello,


You can find some useful information in the text of Bishop and Sherratt (see bibliography).

If the critical location of the crack is known, a stress history at this point can be time-cycle counted.
Then, the crack growth history is predicted using a effective stress intensity (Paris equation) and a cycle by cycle approach.

The whole stress field near to a crack tip can be characterised by a parameter K called stress intensity factor (SIF) :
K = Y * sigma * (PI * a)^1/2
The factor Y is the compliance-function. It is the change in stiffness as the crack grows.
For a central crack in an infinite body, Y = 1. For an edge crack, Y = 1.12.
Closed-form solutions of Y can be found in the litterature (see Rooke & Cartwright) but if not you can generate it by using crack tip elements in your FE model.


Bibliography
Bishop N.W.M., Sherratt F., "Finite Element Based Fatigue Calculations",
The International Association for the Engineering Analysis Community,
NAFEMS, Farnham, 2000.
Rooke, D.P., Cartwright, D.J. (1975) Compendium of Stress Intensity Factors,
H.M. Stationary Office (London).
Test NAFEMS : D.P.Rooke and D. J. Cartwright in Copendium of Intensity Factors HMSO (1976)

Regards,

Torpen

 

[reesh]

Thanks for the suggestions prost and torpen. They'll really come in handy.

Prost, I played around with AFGROW over the weekend and I really think it will help a lot. Thanks!

 

[prost]

I apologize if you know this already.

The quantity "Stress Intensity Factor" which can be computed with most FEA programs (and for that matter, with equations in a few handbooks such as Rooke and Cartwright, Tada and Paris "Stress Analysis of Cracks Handbook," and Murikami "Stress intensity factors handbook") is used to 'drive' the propagation of a crack by using an iterative process with the following three easy steps:
1. Calculate Stress Intensity Factor (SIF) of the crack.
2. Use the Step 1 SIF to look up 'da/dN' in a 'crack growth rate curve' that plots crack growth rate da/dN vs. Delta-SIF (Delta-SIF is the difference SIF(max) minuns SIF (min); that is, the crack SIF at the peak load minus the crack SIF at the minimum load).
3. The increase in crack length 'da' is then computed for one cycle from 'da/dN' computed in Step 2. Increase the crack length in Step 1. by this 'da' -- a(new) = a(Step 1)+da. Go back to Step 1, repeat as necessary until structure failure.

Sounds easy, except for the following complications
1) the crack growth rate curves da/dN are usually proprietary; even if not, they are often considered like gold--precious and therefore expensive. You'll find some da/dN curves in AFGROW, but most people develop their own with testing of their particular material, and are reluctant to give those curves out to anyone outside of their company.
2) 'structural failure' can be a complicated subject. Many people use one of two criteria: net section yield or maximum SIF. Net section yield is essentially when the plastic zone surrounding the crack tip reaches a free boundary, suggesting that section of the structure can't sustain a load. Max SIF criteria uses one of three measures of max. SIF--plane strain fracture toughness, plane stress fracture toughness, or a blend of the two (AFGROW uses a blend). Other types of structural failure criteria: crack length (perhaps an instrument is used to measure crack propagation in a test article, and the test stopped when a certain crack length measured), number of cycles (in case a test is stopped before the structure fails).