IV Modeling Weldment Fatigue Behavior AM 11/03

Transcription

1 IV Modeling Weldment Fatigue Behavior AM 11/03 before after

2 Outline Modeling difficulties posed Basic Information Possible models AM 11/03 2

3 Applied stresses are always uncertain! 100 Ideal 10 2x Nominal 10x 10x Fatigue Life, N (cycles) A factor of 2 uncertainty in the applied load causes the life to vary by a factor of 10! AM 11/03 3

4 Distortions cause secondary stresses! S B = 3 2 α S A β = L t 3 S A 4E L t tanh β β r Induced Bending Stress Applied stress ²S α t ²S θ L AM 11/03 4

5 Weldment fatigue behavior is dependent on the manner of loading! Fatigue resistance depends upon loading conditions AM 11/03 5

6 So many geometries! There is an infinite number of weldment geometries. AM 11/03 6

7 A simple weld may have many failure modes! While the weldment may be simple, many different failure scenarios may exist. AM 11/03 7

8 Weldment geometry may actually be undefined! Example: T-joint with a variable fit-up Tight fit-up: K t - = 3.6 Normal fit-up: 3.6 < K t < 6.4 Loose fit-up? K t = 6.4 AM 11/03 8

9 Weld shape may vary! The geometry of a weldment may vary with location. AM 11/03 9

10 Weld quality variable! Undercut, Slag Entrapment and/or other Discontinuities Weld Metal 0.1 in Weld Metal HAZ Base Metal HAZ Base Metal "Nominal" Weldment Fatigue Cracks "Ideal" Weldment AM 11/03 10

11 Mean stresses alter fatigue life! Restraint σ r ²Sb ²Sa ²M ²S T r Sfab ²M ²S θ Remote Weld Done Last Applied mean stresses, welding residual stresses, and fabrication residual stresses AM 11/03 11

12 Weldment size affects fatigue life!? M? S t r 180 -θ? S? M? M r Longer Fatigue Life 180 -θ Region of high stress? M Same initial size, crack grows to 2x initial length? S t? S Shorter Fatigue life AM 11/03 12

13 Material properties generally unknown! Base Metal (BM) Weld Metal (WM) Fatigue cracks begin in WM or HAZ not in BM! Heat Affected Zone (HAZ) Good news: Material properties don t matter too much! AM 11/03 13

14 Summary The variables influencing weldment fatigue life can be thought of as being only two: the magnitude of the notch root stresses. the properties of the notch root material. In this sense, the applied stresses, the degree of bending, the welding residual stresses, the fabrication residual stresses, the applied mean stresses, the weldment geometry, the notch root weld defects, and the weldment size all influence the magnitude of the notch root stresses. AM 11/03 14

15 Summary The fatigue behavior of a weldment is controlled by the local (notch root, hot-spot) stress-strain history. For structural steel weldments: material properties are of minor importance except (as we shall see) to the degree that they determine and limit the value of the residual stresses. AM 11/03 15

16 Outline Modeling difficulties posed Basic Information Possible models AM 11/03 16

17 Fatigue of a component The fatigue life of an engineering component consists of two main life periods: Initiation or nucleation of a fatigue crack (NI) And Its growth to failure (NP) A smooth specimen AM 11/03 17

18 Basic Information - crack initiation.1.01 ε ' f c ²e total strain ampltude elastic strain ampltude plastic strain ampltude Fatigue test data from strain-controlled tests on smooth specimens..001 σ' f E b Elastic component of strain,?e e ε t = ε e ε p = σ' f E Reversals (2Nf) ( 2N f ) b ε' f 2N f ( ) c AM 11/ Plastic component of strain,?e p Smooth specimen behavior

19 Log Crack Growth Rate, da/dn (m/cycle) C Basic Information - crack growth Paris Power Law da dn? K th I II III Long cracks m = C (? K) Log Range in Stress Intensity Factor,?K (MPavm) m I Sensitive to microstructure and environment. Dominated by crack closure effects. II Paris power Law. III Approaching fracture when Kmax K IC. AM 11/03 19 K C Near threshold

20 Outline Modeling difficulties posed Basic Information Possible models AM 11/03 20

21 Modeling options a =? i Total Fatigue Life Crack Nucleation N N N P1 N P2 a smooth specimen i Actual Short Crack Growth N I N P2 IP Model Long Crack Growth N N a = 0.01-in. (0.25mm) i N F Modified PICC-RICC Model N I N F AM 11/03 21

22 Models N I only - Strain controlled fatigue N P2 - Current TWI models N F only - Dong et al. N N N F - PICC-RICC model N I N P2 - The I-P model AM 11/03 22

23 Relative importance of NI and NP Percent Total Life (%) 100% 90% 80% 70% 60% 50% 40% 30% 20% Shear Crack Growth Tensile Crack Growth Crack Nucleation The relative importance of fatigue crack initiation and propagation in smooth specimens of SAE 1045 steel. (after Socie) 10% 0% 1E02 1E03 1E04 1E05 1E06 Total Life, Nt (cycles) AM 11/03 23

24 IP Model details N T =N I N P2 Laboratory fatigue tests on smooth specimens Mechanics analysis FEA K K n n K t ε f σ f b c Set-up cycle σ ε σ m σ m K fmax Estimation of N I σ m time Estimation of N T Kfmax N T = N I N P M k Fatigue notch size effect K t -> K fmax Laboratory tests on precracked specimens C n U a i =? Estimation of N P AM 11/03 24

25 CCN model 1. Initiation? a th 2. Growth? N T N F a 3. Life? ²K( a) Crack closure at a notch model. Information about U(a) becomes the sticking point. U( a) 1 ²K ( )a eff ²K th 0 a a a N F = a f a i da dn 1 da da dn da dn a i a f a ²K ( a) AM 11/03 25 eff

26 Model performance Test data [1] A' - LEFM, straight-front crack B' - LEFM, crack-shape development C' - Crack-closure model, linear Paris law D' - Crack-closure model, bilinear Paris law E - I-P model CRACK PROPAGATION LIFE (CYCLES) AM 11/03 26

27 Unresolved modeling difficulties IP - What is the size of the crack at the end of the N I stage? 0.01-in? a th? IP - What is the meaning of K f? CCN - What is the value of U as a function of crack length? What is the value of U? All - What are the residual stresses, and how do they vary with location and change during the service life of the weldment? AM 11/03 27