Code_Aster forums

Welcome on the forums. Please post in English or French.

You are not logged in. You won't be able to post until you register and log in using the top-right bar.

#1 2011-08-20 18:18:23

JMB365
Member
Registered: 2008-01-19
Posts: 778

[Solved] Fatigue life prediction

Hello,

I would like to perform a fatigue life prediction on a steel component in CA_11+, where I would like see the predicted life represented by number of cyclical loads (represented by a color plot) and an accompanying legend.  Some commercial codes are capable of this.

I did read and run the example provided here: http://caelinux.com/CMS/index.php?optio … mp;catid=5

But I do not understand what the resulting "Fatigue___________" field with the "4.94061e-319,-" is supposed to mean.  Is this even the right example for what I want to do?  I have attached a screen shot to assist in showing what I am looking at.

Since I am new to this field and to this aspect of CodeAster's capability can someone provide some guidance and insight, please?  Thank you for any help!

Regards, JMB

Last edited by JMB365 (2011-08-24 16:51:23)


Attachments:
Post-Pro-fatigue.png, Size: 84,916 bytes, Downloads: 151

STA10.3/NEW11 Ubuntu 64 bit on multiple PCs

Offline

 

#2 2011-08-21 20:06:48

JMB365
Member
Registered: 2008-01-19
Posts: 778

Re: [Solved] Fatigue life prediction

Hello,

I am re-running the example fatigue problem posted by CAELINUX admin Joel Cugnoni from:

<... www.caelinux.com seems to be down, so I will post the URL later...>

which is for a simple piece of right angle steel, but with values for the Wohler S-N Curve taken from sslv135a.comm for Steel.

I have attached a screenshot of the VonMises stress on the left and the Fatigue on the right. [See fatig2-VonMises-and-Fatigue-PostPro.png].  What does not make sense to me is, for a region (elements) where the stress is between 450~600 MPa one would expect a life cycle of less than 2000 cycles based on the table below which is the supplied S-N data.  However, the Fatigue presentation shows values between 4.2e-5 ~ 5.6e-5 (which is 24K ~ 18K cycles).  The fatigue damage prediction is lesser (or life cycles prediction is greater) by at least an order of magnitude!   Why?!

Based on this and the defined S-N Curve:

Code:

# From sslv135a.comm
SNcurve3=DEFI_FONCTION(NOM_PARA='SIGM',
                       INTERPOL='LOG',
                       PROL_DROITE='LINEAIRE',
                       PROL_GAUCHE='LINEAIRE', 
                       VALE=(  138.,    1000000.,
                               152.,     500000.,
                               165.,     200000.,
                               180.,     100000.,
                               200.,      50000.,
                               250.,      20000.,
                               295.,      12000.,
                               305.,      10000.,
                               340.,       5000.,
                               430.,       2000.,
                               540.,       1000.,
                               690.,        500.,
                               930.,        200.,
                              1210.,        100.,
                              1590.,         50.,
                              2210.,         20.,
                              2900.,         10.,),);

All the neccessary files for running this in Ver11.0+ (or Ver10.3- with minor alterations) are attached.  Can somebody please explain or correct what I am doing or interpreting incorrectly?  Thank you!

Regards, JMB


Attachments:
fatig2.zip, Size: 441,552 bytes, Downloads: 143

STA10.3/NEW11 Ubuntu 64 bit on multiple PCs

Offline

 

#3 2011-08-21 20:10:11

JMB365
Member
Registered: 2008-01-19
Posts: 778

Re: [Solved] Fatigue life prediction

The screenshot [fatig2-VonMises-and-Fatigue-PostPro.png] is attached below for ease of looking at it, without having to extract it from the fatig2.zip attached earlier.

JMB


Attachments:
fatig2-VonMises-and-Fatigue-PostPro.png, Size: 200,072 bytes, Downloads: 160

STA10.3/NEW11 Ubuntu 64 bit on multiple PCs

Offline

 

#4 2011-08-24 10:49:18

XuanSpring
Member
Registered: 2011-07-18
Posts: 2

Re: [Solved] Fatigue life prediction

Hi,

Thanks for your post.

1. For your question: "What does not make sense to me is, for a region (elements) where the stress is between 450~600 MPa one would expect a life cycle of less than 2000 cycles based on the table below which is the supplied S-N data.  However, the Fatigue presentation shows values between 4.2e-5 ~ 5.6e-5 (which is 24K ~ 18K cycles).  The fatigue damage prediction is lesser (or life cycles prediction is greater) by at least an order of magnitude!   Why?!"

Actually, in your command file, you gave the load-history (via your parameter "Modulate") between 0 and sigma_max (which you obtained about 540 MPa in term of Von_Mises stress). You only applied a quarter of a full cycle (if the load ratio = -1) or a half of cylce (if the load ratio = 0).
In Code_Aster, as mentioned in page 9:32 of document U4.83.02, the value of sigma_alternated ( = 1/2(sigma_max - sigma_min) to extract the number of cycle from the Wohler curve. So, with your loads between 0 and sigma_max, sigma_alternated = 1/2 sigma_max ~  540/2 = 270 MPa in term of Von_Mises stress. That's why you got the values between 4.2e-5 ~ 5.6e-5.
If you change you loads between  - sigma_max and + sigmax_max, you will get the damage value of about 1e-3, associated well with the Wohler curve for sigma_laternated = 540 Mpa.

. For thsi question "But I do not understand what the resulting "Fatigue___________" field with the "4.94061e-319,-" is supposed to mean.  Is this even the right example for what I want to do?  I have attached a screen shot to assist in showing what I am looking at."
The value of "4.94061e-319,-"  is a bug of visualation. Please disregard.

Hope that my responses are clear.

Best regards,

Xuan

Offline

 

#5 2011-08-24 16:50:56

JMB365
Member
Registered: 2008-01-19
Posts: 778

Re: [Solved] Fatigue life prediction

XuanSpring wrote:

Actually ,...You only applied a quarter of a full cycle (if the load ratio = -1) or a half of cylce (if the load ratio = 0). Hope that my responses are clear.
Xuan

Hello XuanSpring,

Thank you very much for the reply and the excellent and clear explanation!  You have solved my problem...

Regards, JMB

Last edited by JMB365 (2011-08-24 16:53:07)


STA10.3/NEW11 Ubuntu 64 bit on multiple PCs

Offline

 

Board footer

Powered by PunBB
© Copyright 2002–2005 Rickard Andersson