1 Session 10 – Fretting Fatigue Experiments and Case Study and Case Study Prof. D. Nowell University of Oxford, UK UTMIS Autumn Course 2011 1 Types of fretting fatigue test • ‘Materials ranking’ tests – e.g. bridge type pads P σ A B Q Q P P σ 0 • Controlled geometry tests – e.g. Hertzian pads • Component geometry tests
20
Embed
Session 10 – Fretting Fatigue Experiments and Case ...utmis.org.loopiadns.com/media/2016/06/Session10.pdf · Fan Blade root failure – January 2001. 5 Blade root failure – January
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Session 10 – Fretting Fatigue Experiments
and Case Studyand Case Study
Prof. D. Nowell
University of Oxford, UK
UTMIS Autumn Course 2011 1
Types of fretting fatigue test• ‘Materials ranking’ tests
– e.g. bridge type pads
P
σA B
Q
Q
P
P
σ0
• Controlled geometry tests– e.g. Hertzian pads
• Component geometry tests
2
Materials ranking testsP
σA B
– Quick and easy to perform– Give qualitative information about materials performance
• But– Lack of control – slip and traction difficult to control
independently– Contact tractions difficult to estimate and susceptible to
alignment errors– Difficult to measure quantities needed for analysis– Analysis not straightforward (FE?)
Q
P
σ ωsin t
Fretting Pad
a a
R
Controlled Geometry tests (E.g. Hertz))
ωt
Q
P
σ ω0sin t
Specimen
-a a
Hertzian contact – repeatable and easy to analyseSalient variables can be controlled and measured
ωt
Salient variables can be controlled and measured (Q,P,σ, f, a)Experiments usually conducted in partial slipBut more difficult to performMay need specialised 2-actuator machine
3
Hertzian fretting tests
Implementation on single actuator machine
Two-actuator alternativesmall actuator
fretting Padsspecimen
hydraulic jaws
contact loads
fret
ting
mot
ion
hydraulic jaws
large actuatorThe Oxford ‘in-line’ rig
4
Tests on ‘realistic’ geometry
• Designed to simulate actual geometry and l diloading
• Useful for component lifing and model validation
• May be expensive and time-consuming to set upup
• Example – dovetail fretting
Fan Blade root failure– January 2001
5
Blade root failure – January 2001
Blade root failure – January 2001
6
Dovetail loads
Disk expansion
Biaxial Dovetail Fatigue Rig
7
Biaxial Rig
Fatigue Failures – biaxial rig
8
Dovetail fretting scar
Rig test results
9
Finite element analysisDifficult to resolve stress field with FEA
Experiments include a range of different coatingsShot-peened residual stress field
12
Real geometry and simplified tests
Real geometry Simplified tests
3D view: realcomponent
2D Biaxial Test fixture
23
2D “slice”“In-line”F&R contact
Experimental and analytical studies may be carried Experimental and analytical studies may be carried out with an approximate geometrical model.out with an approximate geometrical model.
Dovetail fretting approximation
T
Di
M
V
P
Q
M
B
Flat and rounded
24Dovetail joint
Blade
DiscPQ
Dovetail
13
Q
P
Possible “in-line” test configurations
P
Q
Hertzian contactR
2b2a
y
B B x
25
R
x
y
2bB B
‘Flat and rounded’ approximation
PQ
Experimental work – In-line rig
A
PLoads
A
Tensile specimen
R
Q
Tensile specimen
σo
Bulk
t
BA
B
A B
Q
Bulk
t
B
26
Fretting Pad
PQ
Fretting PadQQ
Bulk max
QQ
Bulkmin
Q=±Q max
P PP P
crack crack
14
Shear actuator “In –line” fretting rig
1. Set of experiments carried out on the
fretting padsspecimen
P
f pOxford “in-line” fretting rig using flat and rounded contact pads. Specimens and pads Ti-6Al-4V
2. Tests designed in order to determine fretting performance for various loading conditions and geometries
27 Tension actuator
3. 7 series of tests carried out: 3 geometries, 3 different values of peak pressure and tangential/normal load ratio. Bulk load varied to obtain S-N curves
Empirical scaling of x-axis may be used to collapse curves
1
Failurefl
amp
σσ _0
Predictions for current experiments on Ti 6-4Flat and rounded geometry, half-plane theory
0.1
p0=500, Q/fP=0.25,0.5,0.7,0.9,1
Non-FailureHighly dependent
on the loading conditions
f
36
0.010.1 1 10 100
Q/fP=0.5, p0=500,600,700,800,900
2
0
0⎟⎟⎠
⎞⎜⎜⎝
⎛−
fl
pa
abσ
Empirical scaling less successful here
19
1σ 0_ampl /σ fl
Point method prediction p0=665 [MPa] Q/fP=0 2p /σ =0 723
Comparison with experimental results (and with critical distance approach)
Point method prediction p0=665 [MPa], Q/fP=0.2Short crack prediction p0=665 [MPa], Q/fP=0.2Series 1: experimental dataSeries 3: experimental dataSeries 6: experimental data
test 6
test 9
tests 11-17, 65
test 27
test 28
66
test 60
test 55
test 54
test 63
p0/σy=0.723 ,p0/σy=0.723 ,
37
0.11 10 100(b-a)/a 0 (p 0 /σ fl ) 2
Run out 10 7Broken test 19
test 20
test 29
test 30
tests 67-68test 66
tests 59,61
test 58
test 57 test 64
tests 71,72
test 73
1
σ 0_ampl /σ fl
P i t th d di ti 0 478 [MP ] Q/fP 0 2/ 0 52
Comparison with experimental results (and with critical distance approach)
Point method prediction p0=478 [MPa], Q/fP=0.2Short crack prediction p0=478 [MPa], Q/fP=0.2Series 2: experimental data
test 23
test 26
test 24
p0/σy=0.52
p0/σy=0.52 ,
38
0.11 10 100(b-a)/a 0 (p 0 / σ fl ) 2
Run out 10 7Broken test 25
test 22
20
Summary - Session 10• A range of experimental approaches are available, ranging from
materials ranking tests up to simulation of realistic geometry and loads
• A practical industrial problem was investigated using a geometrically representative test and using a ‘flat and rounded’ geometry to mimic dovetail traction distribution
• Fretting fatigue exhibits ‘notch-like’ behaviour – there is a size effect• Short crack arrest approach using fracture mechanics gives good
predictions of fretting thresholds, based on ΔK0 and σfl as independently measured material parametersindependently measured material parameters
• It is possible to incorporate residual stress fields such as those due to LSP or conventional shot peening
• Friction coefficient is an important parameter. It may be difficult to determine and can vary spatially and temporally
• The short crack approach is used for validation of current designs39