SAE Fatigue
Design and Evaluation Committee
Component Test
Division
Spring 2002
Meeting,
The
minutes of the division’s progress meeting and planning session held 16-17
October 2001 were approved. A copy
of those minutes is posted at the “fatigue.org” site.
Jeff
Nash (309 578 4151, Nash_Jeff_P(at)cat.com)
is the division’s new vice-chairperson. Welcome
and thank you, Jeff!
Presentation: “Analysis of Creep-Induced Strains and Stresses in Notches
” – Enrique Núñez (presenter), Grzegorz Glinka,
and David Burns
Visco-plastic FEA is an approach that may be used in
calculating creep stress and strain at notches.
The approach is very accurate and provides results in the entire
component domain. However, the
approach is very time-consuming, thus impractical for fatigue calculations.
This presentation proposes an alternative Neuber-based approach meant to
be easily programmed, computationally efficient, and capable of satisfactory
approximations for notch behavior. The
proposed solution ties the Neuber relation to the time-dependent stress-strain
relation found in creep. The
proposal also includes an additional term for cases of non-localized plasticity;
the term compensates for additional deformation at the notch dip due to
far-field constraint relaxation. To
validate the proposal via comparison to FEA, several cases of notch geometry,
such as a hole, a semi-circle, and a V, were analyzed for tension conditions,
then later bending conditions; the FE models were 2D.
Stress and creep strain vs. time were compared for both approaches and
showed good agreement between the Neuber-based solution and FEA, especially for
stress results. Future work will
address cyclic loading, mixed stress state (tension and bending combined), and
3D stress analysis. Please
contact Enrique or Paul Lubinski if interested in further detail, such as the
equations and proposed algorithm of the Neuber-based solution, or plots of
stress and strain vs. time for the analyzed cases.
Presentation: “Effect of Fatigue Damage on Dynamic Response Frequency for Spot-Welded Joints
” – De-Guang Shang (presenter), Mark Barkey, Yi Wang, and Teik Lim
The work presented here was measurement of response frequency during the fatigue process for spot-welded specimens, identification of the relationship between damage and response frequency, and proposal of a non-linear damage model based on response frequency. Each specimen consisted of steel sheets joined by a single spot weld and loaded axially at their ends so that the spot weld was in shear. Fatigue tests were conducted at various positive R-ratios for each of two cases of nominal nugget diameter (5.4 and 8.0 mm). The dynamic response frequency of each specimen was measured at approximately 10% life increments up to ~half the life, then in ~5% life increments from ~half-life until failure. The measurements for the modal test involved mounting the specimen on cushions and exciting the specimen via impact hammer. Plots in the presentation show the decreasing response frequency with respect to increased fatigue damage. The response frequency at imminent failure was generally ~73-90% of the original frequency, with the sharpest decrease in frequency occurring after ~75-80% life. The presentation proposes a relationship for damage as a function of the response frequency in the original (prior to any loading) specimen F0, the frequency at failure Ff, and the frequency at the interval of interest Fn: D = 1 – ((Fn – Ff)/(F0 – Ff)2. Using a damage mechanics formula for damage rate and integrating this rate equation produces a life estimate equation based on the frequency change parameters. Please contact Dr. Shang or Paul Lubinski if interested in further detail, such as the derived equations for the proposed damage model, or plots of test results, or a comparison of damage accumulation in the test results to the proposed model.
Presentation: “Correlation of Fatigue Crack Growth for Different R-Ratios using DK+ and Kmax Parameter
” – Sudip Dinda (presenter) and Daniel Kujawski
The
presentation proposes a crack driving force parameter K* in an alternative
method to the DKeff approach. A
need was identified to address limitations that have been cited regarding the DKeff
crack closure concept, such as the extent to which the “U-ratio” (DKeff/DK)
is affected by R-ratio, specimen thickness, microstructure, load history, and Kmax.
Examples are given for differences in crack growth rate as a function of
DK for different R-ratios (noticeably greater rate for a higher R-ratio at a
given DK). The proposed driving
force parameter is K* = Kmaxa(DK+)(1
– a),
where a
is a material- and environment-dependent parameter.
This generalized form is based on breaking the crack tip zone’s damage
into components of monotonic damage due to Kmax
and cyclic
damage due to DK+,
accounting for the elastic field’s control of damage action, and the existence
of tensile stress in the zone. Two
methods were tried for determining a: one was from Kmax vs. DK for constant da/dn, and the other was from Kmax
vs. (1 – R) for constant da/dn. Both
methods generate very similar results for da/dn vs. K*.
The approach using the K* parameter was then compared to the crack
closure concept for da/dn curves of two grades of aluminum, one grade of
medium-carbon steel, and one titanium alloy.
The new approach was concluded to have “collapsed” the crack growth
data into a narrow band, showing the ability to predict crack growth rate for
many different R-ratios. The
presentation was followed by much discussion about the discussion as well as
clarification about the crack closure (DKeff) approach. Please
contact Sudip or Paul Lubinski if interested in further detail, such as
equations related to the proposed methods, or the published ‘raw’ da/dn
data used in the study, or plots of the method comparisons / ‘collapsing’ of
data from multiple r-ratios.
·
Rig test
data and results need to be compiled and summarized.
VOLUNTEERS?
·
FEDEM and
other parties working with dynamic models of the ATV could use help in
characterizing bushings and other components.
VOLUNTEERS?
·
Bolt
testing: There are extra bolts remaining from the study performed by Caterpillar
Component Technology and presented at the fall meeting.
WHO WOULD LIKE THE BOLTS? Please
contact Dan Lingenfelser or Jeff Nash if interested in more information.
·
Variability
and uncertainty studies: no discussion, no volunteering.
·
Accelerated
testing (new methods, how much acceleration is too much): no discussion, no
volunteering.
·
Thermal
fatigue testing (and acceleration thereof): no discussion, no volunteering.
·
Help for
Structural Analysis division / other ideas: Rivet testing at R = -1, laser weld
studies, weld modeling studies, and rivet modeling studies were among the ideas
discussed. Al Conle gave Al Krause
as a contact to discuss rivet test possibilities, specimen availability, etc.
VOLUNTEERS for other studies?
-----------------------------------------
Respectfully submitted,
Paul Lubinski
Chairperson, Component Testing division
812 341 2362
paul.lubinski(at)arvinmeritor.com