Finite element model of welded joint with crack-tip in heat affected zone 7

FINITE ELEMENT MODEL OF WELDED JOINT

WITH CRACK-TIP IN THE HEAT AFFECTED ZONE

Dražan Kozak

M.Sc.Mech.Eng. D. Kozak, University of Osijek, Mechanical Engineering Faculty,

Trg I.B.Mažuranić 18, HR-35000 Slavonski Brod, Croatia, E-mail:

Keywords: high strength steels, welded joint, strength mismatching, finite element method, crack, fracture mechanics

ABSTRACT

The possibility of failure of welded structures is depending on fracture mechanics properties of weld joints. The aim of this paper was to determine appropriate numerical model to simulate the experimental evaluation of fracture mechanics specimens. To this purpose, standard BxB single edge notch bend (SENB) three points bend (TPB) specimens made with surface notch tip in the heat affected zone (HAZ) were applied. The X-grooved welded joint is homogenous with the strength higher then base material (overmatching). The comparison between numerical and known experimental results is done by using measured values of fracture behaviour as are crack tip opening displacement (CTOD), load line displacement (LLD) and crack mouth opening displacement (CMOD). The finite elements results shown that using of described numerical procedure for threshold determination of safe load for real welded different structures is possible.

1. INTRODUCTION

The actual approach to welding of high strength low allow (HSLA) steels recommends in some cases preheating and welding by electrodes, which produce higher yield strength of weld metal compared to parent metal. Preheating reduces the cold cracking susceptibility and overmatching has a protective effect, preventing growth of planar cracks or other defects in a weldment (N.Gubeljak, 1999). Anyhow, welding of HSLA steels with yield strength above 700 MPa is questionable, because it is difficult to produce electrode that provides high toughness and protective effect regarding crack initiation and growth in a weldment. According this, it is necessary to investigate fracture behaviour of such produced welded joints, particularly in the case of limited ability of deviation (f.e. constraint is greater in the case of three point bending than in tension mode of loading). The shape of weld gap can also play significant influence on increasing of constraint, so X-grooved welded joint may be a barrier (R.J.Dexter, 1996) to the further plastic flow, because a locally overmatched HAZ may act as a barrier against plastic deformation in the base metal. Therefore, the most critical fracture behaviour of specimens with crack tip in HAZ can be expected (C.Thaulow and M.Toyoda, 1996).

The examinations of standard fracture mechanics specimens are very expensive in the rule and require appropriate equipment and qualified staff. In the other hand, the numerical simulation of the fracture investigation is cheaper and asked just one computer or workstation with installed some finite elements package and expert with experience in modelling of fracture problems. The many of today used numerical conventional programs (e.i. ABAQUS, IDEAS, NISA, ANSYS etc.) are applied to evaluate fracture mechanics parameters. The aim of this paper is to mark difficulties by numerical modelling of strength mismatched welded joints with crack-tip in HAZ and to propose the procedure for the valid determination of fracture behaviour.

2. THE EXPERIMENTAL RESULTS

HSLA steel NIONICRAL 70A, corresponding to grade HT80 produced by Steelworks Jesenice - Slovenia as the base metal (BM) has been chosen. Welding was done on plate samples (500x250x40 mm) by Flux Cored Arc Welding (FCAW) process (CO2 shield gas) producing one overmatched (Weltec B 800 electrode) homogenous multipass X-grooved welded joint.


Fig. 1 True stress-strain diagrams for BM, weld metal and HAZ metal

Round tensile specimens (Ø5mm) extracted from a real welded plate for the determination of mechanical properties of BM and weld metal region was used. It is very important to get true stress-strain curves for both materials because these data have a great influence on the later numerical simulation accuracy. True stress-strains diagrams for BM, weld metal and metal of HAZ on the Fig. 1 are given.

The most difficult to achieve is correct determination of HAZ behaviour. One of the possible approaches is to transform data of measured microhardness to the strength properties by correlated formulae (N. Gubeljak, 1998):



where Rp0.2 [MPa] is yield strength, Rm [MPa] is ultimate strength, HV0.1 is Vickers microhardness and n is strain hardening coefficient related to Dt8/5 [s]:


Assessment of the effect of different microstructures on fracture behaviour was obtained by positioning of the pre-crack in HAZ. Standard BxB (B=W=36 mm) single edge notched bend (SENB) specimens (a0/W=0,29) with surface notch tip completely in the HAZ were fatigue pre-cracked from the surface up to HAZ, Fig. 2. The crack front was located 5 mm distant from the symmetry of welded joint. During the test, the load F, load line displacement LLD, crack-tip opening displacement CTOD (d5) and crack mouth opening displacement (CMOD) were recorded. Measuring points are marked on specimen surface, Fig. 3. Experimentally obtained values will be compared with finite element results assessing accuracy of proposed numerical procedure.

Fig. 2 Location of fatigue pre-crack tip

Fig. 3 Three point bending specimen and measuring points (CMOD, CTOD (d5))

3. THE NUMERICAL MODELLING OF SPECIMEN

Considering the aim of this paper finite element model of evaluated specimen by FEM commercial package ANSYS 5.3 was prepared. The 2-D front of model was build up by “down to top” technique. This means that first key points of three weldment zones (weld metal, 1,5 mm thick HAZ and base metal) must be defined. Further, these key points are connected with lines, which make the geometry of fracture model (Fig. 4).

Fig. 4 Geometry of model drawn from lines

The stress-strain curves introduced in the finite element code as a set of numerical data, are the true stress strain curves drawn from the tensile tests previously presented. For different zones of weldment appropriate set of material data was associated.

The most important in a fracture model is the region around the edge of the crack. Crack-tip located in the HAZ as the stress concentration was generated with eight singular circumferential elements in the first row. For reasonable results, the first row of elements around the crack tip should have a radius of 50 mm (C. Ruggieri et all, 1996). The model was discretisate by so called “free meshing” technique with 8-node second-order isoparametric elements. The finite element model shown on the Fig. 5 consist from 1113 elements and 3402 nodes. Enlarged detail of fine mesh around crack-tip on the Fig. 6 is presented.

Fig. 5 Finite element model of fracture specimen with crack-tip in the HAZ

Fig. 6 Detail of singular collapsed elements around crack-tip

At the surface of specimen nodes are specified, which correspond to measure points for LLD, d5 and CMOD. Elastic plastic finite element calculations have been performed with incremental increasing of loading. As the yielding criteria von Mises equivalent stress was applied.

4. RESULTS AND DISCUSSION

The fracture behaviour of specimen has been calculated with quasistatic increasing of load up to the point of experimental instability. Except calculated displacements at specified points and maximum equivalent stress, it is very useful to analyse stress-strain fields at surface of specimen at the moment of failure instability. Such analysis making possible to find out the reasons for collapse of structure, what is impossible just with experimental work (D. Kozak, 1998).

The acceptable agreement between experimental and numerical values for displacements (LLD, d5 and CMOD) confirms proposed finite element model (Figs. 6 and 7). Actually, with FE results is possible to follow to curves obtained by experiment.

Fig. 6 The comparison between experimental and numerical results for LLD

Fig. 7 The comparison between experimental and numerical results for d5 and CMOD

It is interesting to look at stress-strain condition at the moment of instability corresponds to ones at the moment of cleavage failure initiation on testing specimen. The observations of equivalent von Mises stress field at maximal load (117 kN --à to je netočno, jer je za ovu epruvetu max. sila bila 164,49 kN !!!!!!!!!!!) show that fracture appears although the utility of plastic flow in weld gap is not finished (Fig. 7).

Fig. 7 Von Mises equivalent stress field at the moment of instability failure

5. CONCLUSIONS

Performed finite element analysis of weldment with crack-tip located in the HAZ are exhibiting understand good agreement between FE calculated displacements and experimental measured values. Also the stress-strain condition at the crack tip can be numerical calculated and realistic presented in the moment of failure instability. Considering the previous work is evident that used finite element calculation procedure is appropriate not just for mismatching welded joints with crack-tip in the weld metal but also with crack-tip in the HAZ.

Analysing the results it can be concluded that the shape and width of weld joint are strongly effected by constraint effect in the weld joint. In the case of a locally overmatched HAZ with respect to the weld metal, the HAZ will act a barrier against plastic deformation in the base metal. Because of HAZ, following the shape of weld gap, the initiation of instability began when the contour of plastic flow comes to near fusion line. It seems that unstable fracture occur when the utility of plastic flow in weld gap is finished.

6. ACKNOWLEDGMENT

This investigation was performed under joint Croatian-Slovenian project, conducted by the Mechanical Engineering Faculty, Slavonski Brod and Faculty of Mechanical Engineering, Maribor. The work was supported too by stimulated grants for young scientists of Ministry of Science and Technology of Republic of Croatia Nr. 152-504. The author acknowledges many useful discussions and contributions of Prof. Dr. I. Rak, Prof. Dr. F. Matejiček and Dr. N. Gubeljak.

7. LITERATURE

/1/ Fracture Behaviour of Specimens with Surface Notch Tip in the Heat Affected Zone (HAZ) of Strength Mis-Matched Welded Joints, to be published in the International Journal of Fracture

/2/ Significance of Strength Undermatching of Welds in Structural Behaviour, IIW Doc. X-F-042-96

/3/ Strength Mis-Match Effect on Fracture Behaviour of HAZ, 2nd International Symposium on Mis-Matching of Welds, Reinstorf-Lüneburg

/4/ The Effect of Strength Mis-Match on Welded Joint Fracture Behaviour, Dissertation, University of Maribor

/5/ Numerical Modelling of Ductile Crack Growth in 3-D Using Computational Cell Elements, International Journal of Fracture 82, pp. 67-95

/6/ Modified Numerical Modelling of Fracture Behaviour of Weld Joint in Ductile-to-Brittle Temperature Transition Region, KOVINE ZLITINE TEHNOLOGIJE, Vol. 82, No. 5, pp. 337-342