New Life in Fatigue KIVI NIRIA HOUSTON, WE HAVE A PROBLEM...

Size: px

Start display at page:

Download "New Life in Fatigue KIVI NIRIA HOUSTON, WE HAVE A PROBLEM..."

Brent Gray
5 years ago
Views:

New Life in Fatigue 21-11-2011 KIVI NIRIA HOUSTON, WE HAVE A PROBLEM.

1 New Life in Fatigue KIVI NIRIA HOUSTON, WE HAVE A PROBLEM... Ship structure a collection of a large number and variety of fatigue prone locations, hot spots. Delft University of Technology Challenge the future

2 Fatigue in High-Speed Aluminium Ships a total stress concept and joint SN curve formulation Delft University of Technology Challenge the future

3 Introduction For many years, aluminium alloys have become the standard for high speed ships. L pp = [m] F n ~ 1.0 [-] Primary joining method for ship structural components: ARC-WELDING fracture toughness: aluminium << steel 3 48

(stress concentrations): + Dynamic loading = WELD

4 Introduction Arc-welding: Reduces fatigue strength Fillet welds and butt welds introduce NOTCHES (stress concentrations): + Dynamic loading = WELD LINES fatigue governing parts of ship structure Y Z X 4 48

5 Introduction Some typical structural details: weld line + ends Fm t c t b F m t c F m t b F m F m t c Fm t b Fm t c t b 5 48

aluminium vessels, a practical, efficient methodology to predict

6 Introduction Maritime Innovation Project (MIP): VOMAS Consortium: Damen Shipyards, TUD, Marin, TNO, BV, LR, ABS, USCG For high-speed aluminium vessels, a practical, efficient methodology to predict (impact induced) fatigue in an early design stage does not exist. 6 48

7 Introduction fatigue design method hydrodynamic structural fatigue loads response assessment hydroelasticity fatigue resistance 7 48

Motivation Fatigue life prediction concepts 1 SN curve fictitious notch radius transforms nominal stress fatigue strength systems in a uniform local stress value fine solid mesh Battelle SS approach:

8 Motivation Fatigue life prediction concepts 1 SN curve fictitious notch radius transforms nominal stress fatigue strength systems in a uniform local stress value fine solid mesh Battelle SS approach: master SN curve mesh insensitive design environment structural (hot spot) stress concept (fictitious) notch stress concept HSS approach: extrapolation method sensitive fine mesh 1 or 2 SN-curves nominal stress concept SN-curves / detail cat. x SN-curves + SCF design environment 8 48

single SN curve with stress based through-thickness criterion Shell FE environment: no weld modelling WELD LINES / basic joints

9 Motivation Some modelling considerations: Ship structure = shell plated structure Governing parameter: t b << (a,b) HOUSTON, WE HAVE A PROBLEM... Ship structure a collection of a large number and variety of fatigue prone locations, hot spots. single SN curve with stress based through-thickness criterion Shell FE environment: no weld modelling WELD LINES / basic joints / failure locations Include characteristic NOTCH crack propagation behaviour Relation to LEFM approach: welding process induced flaws exist 9 48

10 Idea Combine and extend the advantages of the different fatigue life prediction concepts for applicability in a design environment. structural (hot spot) stress concept (fictitious) notch stress concept A good scientist is a person with original ideas. nominal stress concept A good engineer is a person who makes a design(method) that works with as few original ideas as possible. TOTAL STRESS CONCEPT 10 48

11 Total Stress Concept Joint SN Curve Formulation total stress concept, T σ = 1 : total stress concept, T 10 3 σ = 1 : S R T 10 1 T-joint Butt joint Cruciform joint Longitudinal stiffener joint Gusset plate / bracket PP Butt joint LC Cruciform joint S R T 10 1 bulb st - web fr conn with bracket: type a T st - web fr conn: type c web fr - web fr conn with soft toe bracket: type a web fr - web fr conn with nested bracket: type c st - web fr conn: type b N N small scale specimen based large scale specimen fit 11 48

12 Total Stress Concept Fatigue governing parameters moving dislocations crack nucleation micro-crack propagation macro-crack propagation final fracture crack initiation crack propagation SCF SIF Stress Concentration Factor Stress Intensity Factor Welding process induced flaws already exist fracture mechanics approach (LEFM, loading mode I) Include crack initiation period modify SIF using SCF related notch stress distribution 12 48

13 Total Stress Concept Weld ends (HS type a and b) b a c c Basic welded joint formulations (HS type c): toe root DS / SS butt joint: 4x / 2x t c t b Fm t c t b Fm 2x / 1x t c DS / SS cover plate: 4x / 2x NC / NC t b t c Fm Mb t c t b Fm Mb Fm Fm t c t c DS / SS T-joint: 4x / 2x 2x / 1x t b Fm t b Fm Fm Fm t c t c DS / SS cruciform joint: 8x / 4x 4x / 2x t b Fm t b Fm Fm Fm 13 48

14 Total Stress Concept Notch stress formulation (force {F,M} induced, geometry and far field stress determined parameter): equilibrium equivalent part far field stress + self-equilibrating part weld geometry induced bending + Williams asymptotic solution mr + br = r t b F m sr

15 Total Stress Concept Notch stress formulation, force {F,M} induced: m + br Notch angle α π(fillet weld config): weld toe r t b = sr F m = + Non-symmetry w.r.t. (t b / 2), e.g. DS T-joint m + b = Symmetry w.r.t. (t b / 2), e.g. DS cruciform joint r t b s F m = + Notch angle α =π(crack config): weld root / crack growth specimen Non-symmetry, e.g. DS LC cruciform joint b b t ' b r r ' t b F m Symmetry, e.g. DEN specimen Mb 15 48

16 Total Stress Concept QUESTION 1: effects of self-equilibrating stress part Consider weld toe failure of a DS T-joint and DS cruciform joint. Will there be any difference in fatigue life and if so, which joint has a longer fatigue life? The far field stress is the same for both joints (membrane state). Please note that the nominal-, hot spot- and fict. notch stress is the same. Regarding the DS cruciform joint, consider fatigue failure from 1 side. A. Fatigue life is the same. B. DS T-joint has longer fatigue life. C. DS cruciform joint has longer fatigue life. D. There is not enough information to answer this question. r t b + r t b

17 Total Stress Concept Residual stress formulation (welding induced): Non-symmetry w.r.t. (t b /2): equilibrium equivalent + self-equilibrating part DS T-joint SS butt joint FE analysis, schematic Symmetry w.r.t. (t b /2): DS butt joint: DS long. stiffener: FE analysis, schematic measured [Pagel, 2011] 17 48

18 Total Stress Concept Arc-welded joint: notch stress + residual stress formulation Equilibrium equivalent part + self-equilibrating part Equilibrium equivalent part: force induced + welding induced Force induced: {F m, } Welding induced: membrane + bending (angular distortion) measured / mean stress / fatigue strength C Ss = C N 1 m measured / design value / fatigue strength C Self-equilibrating part: hypothesis: distribution is similar for notch stress and residual stress! 18 48

19 Total Stress Concept Notch stress formulation: Analytical / characteristic (near) singularity / all geometry parameters Related to linear far field stress distribution (FEM) fm m σ s = 6 t t b b 2 b R s 6 mb = 6 m t f b b m t c l w h w Example 1, (αα π), non-symmetry: DS T-joint t b F m 0.0 t c /t b = 1.0; h w /l w = 1.2; l w /t b = 1.0; ρ/t b = t c /t b = 1.0; h w /l w = 1.0; l w /t b = 1.0; ρ/t b = r/t b [ - ] r/t b [ - ] weld notch stress σ w /σ s 0.8 weld notch stress σ w /σ s 0.9 far field stress σ f /σ s FE result 0.9 far field stress σ f /σ s FE result σ w /σ s, σ f /σ s [ - ] σ w /σ s, σ f /σ s [ - ] 19 48

20 Total Stress Concept Notch stress formulation: Example 2, (α π), symmetry: DS cruciform joint t c l w h w t b F m 0.0 t c /t b = 1.0; h w /l w = 1.0; l w /t b = 1.0; ρ/t b = t c /t b = 1.0; h w /l w = 1.0; l w /t b = 0.4; ρ/t b = r/t b [ - ] weld notch stress σ w /σ s far field stress σ f /σ s FE result r/t b [ - ] weld notch stress σ w /σ s far field stress σ f /σ s FE result σ w /σ s, σ f /σ s [ - ] σ w /σ s, σ f /σ s [ - ] 20 48

Total Stress Concept Notch stress formulation: Example 3, (α =π), non-symmetry: DS cruciform joint, SS butt joint Assumed crack path: weld leg section (cr.

21 Total Stress Concept Notch stress formulation: Example 3, (α =π), non-symmetry: DS cruciform joint, SS butt joint Assumed crack path: weld leg section (cr. joint), weld throat (butt joint) Stress state mainly characterised by membrane and bending as well 0.0 t c /t b = 10/10; l w /h w = 10/20; l w /t b = 10/10; a g /t b = 3/10; ρ/t b = 0/ l w /t b = 10/8; h w /t b = 2/8; l w /h w = 10/2; a g /t b = 3/8; ρ/t b = 0/ r 1 r r '/t b ' [ - ] weld root notch stress σ wr /σ sr weld root linear stress σ lr /σ sr r 3 r '/t b ' [ - ] weld root notch stress σ wr /σ sr weld root linear stress σ lr /σ sr 1 r 2 r 3 {σ s,r s } still defines far field stress 0.9 FEM 0.9 FEM σ wr /σ sr, σ lr /σ sr [ - ] σ wr /σ sr, σ lr /σ sr [ - ] 21 48

22 Total Stress Concept Notch stress formulation: motivation Fatigue life effects: equilibrium equivalent stress part {σ s,r s } fatigue life: relative bending stress effects, far field stress only edge crack formulation centre crack formulation 0.25 s [ - ] R s N/N Rs=0 [ - ] Pure bending shows better performance compared to pure membrane: dσ dr f Non-monotonic behaviour means bad news 22 48

23 Total Stress Concept Notch stress formulation: motivation Fatigue life effects: self-equilibrating part 1.00 fatigue life: relative bending stress effects, far field stress only 1.00 fatigue life: relative bending stress effects edge crack formulation centre crack formulation non-symmetric notch behaviour symmetric notch behaviour edge crack formulation R s [ - ] 0.00 R s [ - ] answer Question 1: C N/N Rs=0 [ - ] N/N Rs(M)=0 [ - ] Non-symmetry: weld geometry induced bending part amplifies effect Symmetry: 3 criteria, bending induced anti-symmetry require shift + scaling (no far field bending stress projection) Far field bending effect counter-acted by self-equilibrating stress (notch effect) 23 48

24 Crack Propagation Model Stress Intensity Factor formulation K: generalised formulation not available for basic welded joints K I (α π), K = M Y σ π (α =π) = Y Y σ π a a m g I k g Equilibrium equivalent stress part: consistent with far field stress / solutions for simple geometries Geometry factor: Y g Self-equilibrating unit stress part: notch effect / applied as crack face pressure Magnification factor: Y m 24 48

25 Crack Propagation Model Geometry factor Y g : cracked geometry parameter Covers macro-crack effects: finite thickness (and free surface) Superposition principle: membrane + bending Y g [ sgn( F ) Y + R { sgn( F ) Y + ( M ) Y }] = sgn m gm m gm b gb Handbook solutions Weld root geometry correction M k : cracked geometry, FEM Magnification factor Y m : uncracked geometry parameter 2 π a σ ( r) Y m = dr σ ( ) 0 a 2 r 2 1 r r = σ w 2 R σ s tb r t b 25 48

26 Crack Propagation Model Stress Intensity Factor example Notch angle α π : Y m dominates (a/t b ) 0.2 Williams asympt. sol. dominates (a/t b ) 0.1 C gb dominates 0.1 < (a/t b ) 0.2 Y g dominates (a/t b ) > t c /t b = 1.0; h w /l w = 1.0; l w /t b = 1.0; ρ/t b = t c /t b = 1.0; h w /l w = 1.0; l w /t b = 1.0; ρ/t b = Y m r/t b [ - ] Y g, Y m, Y m Y g [ - ] Y g Y m Y g FEM 0.8 weld notch stress σ w /σ s far field stress σ f /σ s 0.9 FEM σ w /σ s, σ f /σ s [ - ] micro- macro-crack crack a/t b [ - ] 26 48

27 Crack Propagation Model Stress Intensity Factor example Notch angle α =π : geometry with gap a g SIF is M k Y g determined; square root notch behaviour included in K by definition M k Y g < 1, b b < t b Y m (notch effect) affects crack propagation! t c t c /t b = 10/10; l w /h w = 10/20; l w /t b = 10/10; a g /t b = 3/10; ρ/t b = 0/10 M k Y g Y m b b t ' b a g r', a r l w sr R sr s, R s h w t b F m M k Y g, Y m, ( 1/Y m ) [ - ] ( 1/Y m ) FEM a/t b ' [ - ] 27 48

28 Total Stress Concept QUESTION 2: fatigue life comparison for weld toe and weld root failure Consider a DS cruciform joint. Will there be any difference in fatigue life for the weld toe and weld root failure case and if so, which one will be longer? 5.0 t c /t b = 1.0; h w /l w = 1.0; l w /t b = 1.0; ρ/t b = t c /t b = 10/10; l w /h w = 10/10; l w /t b = 10/10; a g /t b = 5/10; ρ/t b = 0/ Y m 4.0 Y m r t b m Y g [ - ] Y m, Y g, Y m Y g Y m Y g FEM r 2a i t b M k Y g, Y m, ( 1/Y m ) [ - ] M k Y g ( 1/Y m ) FEM a/t b [ - ] a/t b ' [ - ] A. Fatigue life is the same. B. Weld root has longer fatigue life. C. Weld toe has longer fatigue life. D. There is not enough information to answer this question

29 Crack Propagation Model Characteristic da/dn- K curve (long crack growth) Paris-based two-stage model: ignore region III da dn K = C 1 K th γ ( K ) m 10-4 III K = crack driving force da/d dn 10-5 Similitude hypothesis: da/dn depends only on ( K, K th, K max, K c ) II 10-8 I 10 0 K th K K c 29 48

30 Crack Propagation Model Crack growth at NOTCHES: elasticity/plasticity 10-3 short (notch) crack growth 10-3 short (notch) crack growth (da/dn) [ mm/cycle ] long crack growth plastic short (notch) crack growth elastic short (notch) crack growth (da/dn) [ mm/cycle ] long crack growth plastic short (notch) crack growth elastic short (notch) crack growth K [ MPa mm ] K [ MPa mm ] Frost-Dugdale model: da/dn = C. a Generalised Frost-Dugdale model (Jones et. al, 2008) - non-similitude behaviour - elastic crack growth at notch da dn = C a m 1 2 ( K ) m 30 48

31 Crack Propagation Model Crack growth at NOTCHES: elasticity/plasticity P. Dong: plasticity dominated micro-crack behaviour da dn = 2 ( ) m M k K g Elasticity/plasticity criterion: r r p, notch p, crack 2 : plasticity < 2 : elasticity Model: work in progress because of slope dependency da dn ( ) m m n 2 n = 1, 2 g = C Y m K 31 48

32 Crack Propagation Model Crack growth at NOTCHES: elasticity/plasticity Example: elastic behaviour 10-2 Al 5083, T-L config Al 5083, T-L config. (da/dn) [ mm/cycle cle ] rolling direction base material; R s = -1.0 base material; R s = 0.0 (da/dn). ( Y l / Y n ) [ mm/cycle ] base material; R s = -1.0 base material; R s = K [ MPa mm ] K. ( Y l / Y n ) [ MPa mm ] 32 48

33 Crack Propagation Model Crack growth at NOTCHES: elasticity/plasticity Example: plastic behaviour incl. notch radius effects 10-2 R = 0.05 [-] 10-2 R = 0.05 [-] (da/dn) [ mm/cycle ] ρ = 0.0 [mm] ρ = 0.4 [mm] ρ = 0.7 [mm] ρ = 1.4 [mm] (da/dn) / Y 2 m [ mm/cycle cle ] ρ = 0.0 [mm] ρ = 0.4 [mm] ρ = 0.7 [mm] ρ = 1.4 [mm] K [ MPa mm ] K. ( 1/Y m ) [ MPa mm ] 33 48

34 Crack Propagation Model Crack growth at NOTCHES: Notch / micro-crack fatigue life effects and notch radius influence 0.31 micro-crack propagation fatigue life contribution 10-3 short (notch) crack growth a i [ mm ] (da/dn) [ mm/cycle ] N i /N tot [ - ] K [ MPa mm ] 34 48

35 Total Stress Concept QUESTION 3: effect of final crack length Consider a DS T-joint with SIF behaviour as shown. At what (dimensionless) crack length do you expect that > 90 [%] of the fatigue life is consumed? 5.0 t c /t b = 1.0; h w /l w = 1.0; l w /t b = 1.0; ρ/t b = short (notch) crack growth 4.5 Y m A. 0.1 (a f /t b ) B. 0.3 (a f /t b ) C. 0.5 (a f /t b ) D. 0.7 (a f /t b ) Y g, Y m, Y m Y g [ - ] 4.0 Y g Y m Y g 3.5 FEM a/t b [ - ] (da/dn) [ mm/cycle ] long crack growth plastic short (notch) crack growth elastic short (notch) crack growth K [ MPa mm ] 35 48

36 Total Stress Concept QUESTION 3: effect of final crack length Consider a DS T-joint with SIF behaviour as shown. At what (dimensionless) crack length do you expect that > 90 [%] of the fatigue life is consumed? 1.0 fatigue life: effect of final crack length N/N af=tb [ - ] A. 0.1 (a f /t b ) B. 0.3 (a f /t b ) C. 0.5 (a f /t b ) D. 0.7 (a f /t b ) a f / t b [ - ] 36 48

37 Crack Propagation Model Fatigue test results suggest a non-linear mean stress dependency. Exponential relation (Kwofie, 2001): σ σ R 1 = e σ R α σ us First order approximation: generalization of several empirical models σ σ R = 1 α σ R 1 σ us Goodman, α = 1 high-cycle fatigue: low σ / small σ R High-cycle fatigue: low σ / large σ R Walker, α = {σ us /(γ. σ R )}. ln{(1-r)/2} σ 1 = σ R 1 2 R γ 37 48

38 Crack Propagation Model σ us Walker s mean stress model using σ eff = σ e : σ σ R α σ eff = 1 superior curve fitting results: γ 0.5 for R 0 ( 1 R ) γ s γ = 0.0 for R < 0 Another definition of σ eff (Kim, Dong, 2006): σ eff = σ max σ + = + σ min mean R 0: σ eff = γ = 0.5 ( 1 R) 1 t 2 σ max R < 0: σ eff = mean γ = 0.0 ( 1 R) t Walker s model: loading dependent mean stress max min

39 Crack Propagation Model Welded joints: alternating material zones 10-2 Al 7075-T651; T-L config. Weld (filler) material Heat Affected Zone (HAZ) Base (parent) material weld HAZ base material (da/dn) [ mm/cycle ] P ~ 6 P R = R = 0.10 R = 0.50 R = 0.75 Base material, loading induced mean stress: micro- and macro crack prop. region affected according to Y m : K [ MPa mm ] 1 2 da Ym Ym = C K 1 γ 1 γ g dn ( 1 R) ( 1 R) Al 5083, T-L config Al 5083, T-L config Al 7075-T651; T-L config. m (da/dn) [ mm/cycle ] rolling direction K [ MPa mm ] R = -1.0 R = 0.0 (da/dn). ( Y l / Y n ). (1-R) 1- γ [ mm/cycle ] R = -1.0 R = K. ( Y l / Y n ) / (1-R) 1- γ [ MPa mm ] (da/dn). ( Y l / Y n ). (1-R) 1- γ [ mm/cycle ] R = R = 0.10 R = 0.50 R = K. ( Y l / Y n ) / (1-R) 1- γ [ MPa mm ] 39 48

40 Joint SN Curve Formulation Basquin type of equation: S T = C N 1 m with s and Initial crack size effect: σ ST = 2 m 1 t 2 m b I ( Rs )m I a t f a b ( R ) s = d m ai tb t b Y n m m 2 1 Y m g a t b total stress concept, T = 1 : 1.53 σ 10 3 total stress concept, T = 1 : 1.18 σ 60 initial crack size (Weibull) distribution, m = 3, µ = S R S R p [ - ] T-joint Butt joint Cruciform joint Longitudinal stiffener joint 10 1 T-joint Butt joint Cruciform joint Longitudinal stiffener joint N N a i /t b [ - ] 40 48

41 Joint SN Curve Formulation Mean stress effects: The fatigue strength of welded joints is mean stress independent. Welding induced residual stress acts as high-tensile mean stress. Micro-crack region: HAZ welding induced high-tensile residual stress dominates loading induced part Macro-crack region: base material (weld toe) / weld material (weld root) σ s ST = 1 γ 2 m 1 ( 1 R) 2 t 2 m I ( R )m b s loading induced part only in macro-crack region Weld performance improvement (technique dependent) factor R i 2 ( ) ( ) ( 1 γ ) i m ( 1 γ ) R = 1 1 R 1 R eq i micro-crack region correction and loading induced part in macro-crack region 41 48

42 Joint SN Curve Formulation Mean stress effects: 10 3 master curve formulation; T σ = 1 : master curve formulation; T σ = 1 : S s S R T-joint; R = 0.0 T-joint; R = -1.0 Butt joint; R = 0.0 Butt joint; R = -1.0 Long. stiff. joint; R = 0.1 Long. stiff. joint; R = 0.5 Long. stiff. joint; UIT N T-joint; R = 0.0 T-joint; R = -1.0 Butt joint; R = 0.0 Butt joint; R = -1.0 Long. stiff. joint; R = 0.1 Long. stiff. joint; R = 0.5 Long. stiff. joint; UIT N 42 48

43 Joint SN Curve Formulation QUESTION 4: stress component for fatigue life calculation Consider the structural response of the stiffened panel, loaded with space averaged pressure (sea side). The welding induced residual stress is assumed to be tensile. What stress value needs to be used at Pos. 2 for fatigue life calc.? 1 ELEMENTS 1 NODAL SOLUTION 1 MN NODAL SOLUTION 1 NODAL SOLUTION REAL NUM STEP=1 STEP=1 STEP=1 SUB =1 SUB =1 SUB =1 TIME=1 TIME=1 TIME=1 2 SX (AVG) RSYS=0 DMX = SMN = SMX =37.5 SZ (AVG) RSYS=0 DMX = SMN = SMX = SEQV (AVG) DMX = SMN = SMX = Y Y MX Y Y Z X Z X Z X Z X LS specimen LS specimen LS specimen LS specimen A. σ x (mode I stress component) B. σ VM (Von Mises stress component) C. σ x (absolute value of mode I stress component) D. σ x / 2 (absolute value of half mode I stress component) 43 48

44 Joint SN Curve Formulation QUESTION 5: effect of tensile/compressive loading Assuming that the welding induced residual stress is tensile, what will be the relative mean stress ratio R w for compressive loading relative to tensile loading? σ min R = σ max s s r tension s a s max compression s min A. R w = -1 B. R w = -2 C. R w = -3 D. R w = -4 t 44 48

45 Joint SN Curve Formulation Compressive cyclic loading: SN curves data: tensile residual stress + TENSILE cyclic loading Walker model, welding induced tensile residual stress: σ σ eff = 1 ( R ) γ 1 w R γ = 0.5 Predominantly, the micro-crack propagation region is affected! Total stress parameter: region I correction answer Question 4: D σ s ST = 1 γ 2 m 1 ( 1 R) 2 t 2 m I ( R )m b s σ s σ 2 for σ < 0 for 0 ST = 1 γ 2 m 1 ( 1 R) 2 t 2 m I ( R )m b s 45 48

46 Total Stress Concept Joint SN Curve Formulation total stress concept, T σ = 1 : total stress concept, T 10 3 σ = 1 : S R T 10 1 T-joint Butt joint Cruciform joint Longitudinal stiffener joint Gusset plate / bracket PP Butt joint LC Cruciform joint S R T 10 1 bulb st - web fr conn with bracket: type a T st - web fr conn: type c web fr - web fr conn with soft toe bracket: type a web fr - web fr conn with nested bracket: type c st - web fr conn: type b N N small scale specimen based large scale specimen fit 46 48

47 Total Stress Concept QUESTION 6: effect of compressive stresses on fatigue life Consider the structural response of the stiffened panel, loaded with space averaged pressure (sea side). At what weld location is the first crack expected? The welding induced residual stress is assumed to be tensile. 1 ELEMENTS 1 NODAL SOLUTION 1 MN NODAL SOLUTION 1 NODAL SOLUTION REAL NUM STEP=1 STEP=1 STEP=1 SUB =1 SUB =1 SUB =1 TIME=1 TIME=1 TIME=1 2 SX (AVG) RSYS=0 DMX = SMN = SMX =37.5 SZ (AVG) RSYS=0 DMX = SMN = SMX = SEQV (AVG) DMX = SMN = SMX = Y Y MX Y Y Z X Z X Z X Z X LS specimen LS specimen LS specimen LS specimen A. Position 1. B. Position 2. C. Position 3. D. Position 1 and 3 Pos 1. σ z = [MPa] Pos 2. σ x = [MPa] Pos 3. σ x = 15.5 [MPa] 47 48

48 Thank you for your attention! Henk, thank you for your presentation, but... N=C S -m QUESTIONS? 48 48

V Predicted Weldment Fatigue Behavior AM 11/03 1

V Predicted Weldment Fatigue Behavior AM 11/03 1 Outline Heavy and Light Industry weldments The IP model Some predictions of the IP model AM 11/03 2 Heavy industry AM 11/03 3 Heavy industry AM 11/03 4