of Croatian Society of Mechanics
September, 18-20, 2003
Bizovac, Croatia /
EXPERIMENTAL AND NUMERICAL ANALYSIS
OF CRACK TIP POSITION EFFECT LOCATED IN HAZ
Nenad Gubeljak, Dražan Kozak, Franjo Matejiček, Maks Oblak
Keywords: / Fracture toughness testing, heat affected zone, finite element analysis
1. Introduction
The heat affected zone (HAZ) is in many cases considered to be the preferred location for the initiation and propagation of weld joint cracks. In order to evaluate the fracture toughness and fracture behaviour, the stress-strain field at HAZ cracks must be well understood.
In the last decade the extensive fracture toughness testing of weld heat-affected zones in structural steels was performed [1, 2]. Many investigation of fracture behaviour shows that unpredictable fracture behaviour of specimen with crack tip located in HAZ can occur. The reason for such behaviour may lie in the HAZ microstructure heterogeneity and existing of local brittle zones (coarse-grained heat affected zone - CGHAZ). Very important role plays also the yield stress ratio between base metal and weld metal (so-called mismatch factor M=Rp0.2,WM/Rp0.2BM). It is expected that strength overmatch should ensure protection effect from the yielding in the region of weld metal, what extent the load carrying capacity [3]. However, this effect may lead to brittle fracture in the case when the crack path deviated to brittle microstructure of welded joint. The situation becomes more complex in the case when the crack tip is located in the region between weld metal with higher yield strength and softer base metal. Here the crack tip may be shifted from the interface between weld metal and CGHAZ, between CGHAZ and soft (fine-grained heat affected zone) FGHAZ and finally between FGHAZ and base metal. The spreading of stress and strain fields in the vicinity of crack tip, which is located in the HAZ, may be analysed very effective by finite element calculation [4-6].
The aim of this paper is to determine the effect of different crack tip position in HAZ on crack driving force and consequently on fracture behaviour of strength mismatch fracture toughness specimen.
2. The fracture toughness specimens and testingHigh strength low alloyed HSLA steel (with thickness of 40 mm), corresponding to the grade HT80, was used as the base metal (BM) in a quenched and tempered condition (Q+T). Three different overmatched X-grooved multi-pass welded joints are studied. The first was pure homogeneous with 21% higher yield strength related to BM (Fig. 1a). The second and the third joint were heterogeneous with two root passes welded by wire, which ensures 13% overmatch with the root height of 9 mm (Fig. 1b) and 14 mm (Fig. 1c), respectively. The rest passes were filled also by electrode with 21% overmatch [3]. Mechanical properties were determined according to DIN 50125, using standard tensile specimens with section diameter of 5 mm, made from the root and top region of the weld metal along the welding direction. The tensile tests were performed at room temperature. The average values of tensile properties are given in Tab. 1.
Fig. 1 Homogeneous 1 a), inhomogeneous 1 b) and 1 c) strength overmatch X-welded
Table 1. Mechanical properties of the base metal and weld metal
Material / Temp. / E / RP0.2 / Rm / o / n / At / vE+ / M°C / GPa / MPa / MPa / MPa / - / % / J
BM / 20 / 201 / 711 / 838 / 679 / 0.091 / 19.6 / 54-40C / -
WMfill / 20 / 205 / 861 / 951 / 833 / 0.074 / 11.7 / 56-10C / 1.21
WMroot / 20 / 221 / 807 / 905 / 780 / 0.075 / 15.3 / 61-10C / 1.13
Reference stress o and strain hardening exponent n given in the Tab. 1 obtained by fitting of true stress-strain plot. They are needed for the determination of Ramberg-Osgood relation between true stress and true strain.
The single edge notched bend (SENB) specimens with a machined surface notch tip completely in the HAZ were extracted from the welded plates (Fig. 2). The thickness of the BxB specimens was 36 mm. These specimens were used for the estimation of the specified microstructures with regard to fracture behaviour.
Fracture toughness specimens (a0/W ≈ 0,3) were fatigue pre-cracked in accordance with BS 7448 5 from the surface to a distinct welded joint microstructure, as shown in Fig. 3. The aim of producing the root layer with different heights by heterogeneous welded joint is to provide that crack tip location will be changed from high strength filler metal to the root metal with medium strength. The single specimen method was used. The DC potential drop technique was applied for stable crack growth monitoring. The CTOD values were directly measured with a 5 clip gauge, developed by GKSS 8. The measuring points for CTOD (5) and CMOD (crack mouth opening displacement) are marked on the specimen surface, Fig. 3.
Figure 4. shows the records of loading as plots of load (F) vs. CMOD. The load saturates at a different levels of CMOD values, depending on the crack depth (a/W). In all specimens, except those made from BM, an unstable fracture appeared after some amount of stable crack growth, as shown in the F - CMOD plots.
Figure 2. Location of fatigue pre-crack tip
Figure 3. SENB specimen with heterogeneous welded joint cracked in the HAZ
The initiation of cleavage fracture is probably caused by crack tip position in the coarse-grained microstructure of HAZ. The mechanism of cleavage fracture is related to existence of local brittle zones (LBZ). The inclination of linear part of the F - CMOD diagram depends on the a/W ratio and it does not depend on the interaction of single materials in the joint. On the Fig. 4 it is obviously very different behaviour of the specimens with the almost the same crack length, but with the crack tip, which is moving through the HAZ width. This is the proof of great scattering of the results when the fracture toughness specimen is cracked in the HAZ.
Figure 4. Experimentally obtained curves F vs. CMOD for BxB specimens with approximately the same crack depth, but with the different position of the crack tip
3. Numerical analysisFinite element analyses were performed to calculate the magnitude of maximum principal stress, what may serve as a global parameter by assessment of material resistance with respect to brittle fracture [5]. Distribution of stresses and strains could not be considered as a fracture toughness parameter, but it may help by establishing a fact about the reasons and mechanisms of fracture. However, in the case where crack tip is positioned in the HAZ, the accuracy of FE results is strongly influenced by defining of proper material yielding law and correct modeling of geometry as well. Some of input data is difficult to determinate i.e. shape and width of the HAZ, true mechanical properties of all HAZ regions, effective crack length etc. HAZ in this paper was modeled as bimetal region with total width of 1,5 mm. Near the base metal a softer part of HAZ was supposed, while harder material was assigned to the coarse-grained HAZ in the vicinity of the overmatched weld metal. Yield strength as well as ultimate strength needed for the HAZ yielding law is determined from empirical relations based on the microhardness [9]:
/ (1)/ (2)
where HV0,1 presents Vickers microhardness and n is material hardening coefficient.
It has been noticed that during fatigue of the specimens, a significant plastification appears in the vicinity of the crack. Because of that additional yielding, it was necessary to correct initial crack length. Here plastic zone was calculated by using of known expression from Schwalbe:
/ (3)where aeff is effective crack length, a0 is initial crack length, rp is plastic zone radius, K is stress intensity factor by maximal fatigue load and σyis yielding strength of the HAZ material. Adding of plastic zone to initial crack length in the FE model move the crack tip from the fusion line to CGHAZ region.
Figure 5 depicts coarse automatically generated finite element mesh consisted from 413 8-node isoparametric plane strain elements with 1278 nodes. The first row of singular elements around crack tip has the size of 50 μm, what is nearly the grain size. The crack line is almost 6 mm displaced from the welded joint symmetry line in the case when the crack tip is at the interface between BM and FGHAZ.
Validity of the FE model has been proved comparing measured load line displacement LLD on the SENB specimen with homogeneous welded joint with the same calculated by finite element method (Fig. 6). It has to be mentioned that the best agreement shown the model, which allows the movement of the both specimen supports in the horizontal direction, what is adequate to the rollers sliding. The ordinate in aforementioned diagram presents the ratio between the applied force F and yielding force Fy. The yielding force for the homogeneous SENB specimen is defined as [10]:
/ (4)where a, B, W and S are determined by specimen geometry, y is usually equal to yielding stress Rp0,2 and the parameter A can be approximately determined from:
/ (5)Detail A:
/
Figure 5. Finite element mesh for the specimen with crack on the interface FGHAZ-BM
Figure 6. Variation of FE results due to different crack tip position for the fracture toughness specimen with homogeneous weld (M=1,21)
Due to dissimilarity of the materials in the interaction, the different values for local CTOD (5) displacement left and right for 2,5 mm from the crack tip are present. This asymmetry is obviously from the Fig. 7, where the 5 value in weld metal is greater than the 5 value in base metal for the crack tip located at the middle of the HAZ.
Figure 7. Asymmetry of 5 values for the position of the crack tip in the middle of HAZ
The main aim of numerical analysis was to get an insight into stress and strain distribution in the vicinity of moving HAZ crack tip. Both, the position of the maximal stress and spreading direction of the plastic zones should help by assessing of the influence of crack tip position on the fracture behaviour of the specimen.
Figure 8 shows principal stress x, equivalent stress eq (von Mises) and equivalent plastic strain eq corresponding to different crack tip position at HAZ in homogeneous and inhomogeneous weld metals. Analyses show that the highest stress are occurred in high strength material where yielding process deviated from the crack tip to low strength material due by mis-match effect. Figure 8, also shows that x and eq for homogenous specimens do not show significant difference regarding to crack tip position. A contrary, one can observe that in homogenous weld joint crack tip position play important role to elastic-plastic strain eq behaviour at the vicinity of crack tip. Therefore only a small difference at the crack tip position has a strong effect in crack behaviour in elastic and plastic behaviour of specimen. This can explain experimentally obtained different behaviour of specimens cut out from same weld plate and with similar crack depth in Figure 8.
Inhomogeneous weld joint with same crack tip position but different size of overmatch metal (M=1.13) do not show any significant different in stress-strain behaviour of weld metals at the crack tip. Hereby, the slightly over match root passes have no effect on fracture of specimens.
4. ConclusionIn this paper the different fracture behavior of specimens cut out same welded with crack tip in different regions of heat-affected-zone (HAZ) were analyses. The different fracture behavior is explained by numerical analysis using finite elements modeling, Numerical analysis shows that changing of crack tip position has a significant effect on fracture behavior in elastic and plastic loading range. Significantly low effect on fracture behaviour has changing of amount of overmatching weld metal in root layer. Obtained results give principles for idealization of finite elements modeling of fracture behavior of welded joints.
x stress / eq stress / pl, eq strainWeld metal-CGHAZ / / /
CGHAZ-FGHAZ interface / / /
FGHAZ-BM interface / / /
hroot=9 mm / / /
hroot=14 mm / / /
Figure 8. Stress and strain fields in the vicinity of the moving crack tip
5. References[1]Fairchild, D.P., "Fracture Toughness Testing of Weld Heat-Affected Zones in Structural Steel, Fatigue and Fracture Testing of Weldments", ASTM STP 1058, H. I. McHenry and J. M. Potter, Eds., American Society for Testing and Materials, Philadelphia, 1990, 117-142.
[2]Fairchild, D.P., Theisen, J.D. and Royer, C.P., "Philosophy and Technique for Assessing HAZ Toughness of Structural Steels Prior to Steel Production", Paper OMAE-88-910, Seventh International Conference on Offshore Mechanics and Arctic Engineering, Houston, TX, February 1988
[3]Gubeljak, N.,"Fracture behaviour of specimens with surface notch tip in the heat affected zone (HAZ) of strength mis-matched welded joints", International Journal of Fracture 100, 1999, pp 155-167
[4]Thaulow, C., Toyoda, M., "Strength mis-match effect on fracture behaviour of HAZ", IIW Doc. X – F – 033 -96, Reinstorf – Lüneburg, 1997
[5]Thaulow, C., Ranestad, Ø., Hauge, M., Zhang, Z., Toyoda, M. and Minami, F., "FE calculations of stress fields from cracks located at the fusion line of weldments", Engineering Fracture Mechanics, Vol. 57, No. 6, 1997, pp. 637-651
[6]Matejicek, F., Gubeljak, N., Kozak, D. and Koçak. M., "Stress-Strain State at the Vicinity of the Crack Tip in Strength Mis-match Welded Joint", 13th European Conference on Fracture, San Sebastian, Proceedings on the CD-Rom, 1U.6, 2000
[7]BS 7448, "Fracture mechanics toughness test, Part 2. Method for determination of KIC, critical CTOD and critical J-values of welds in metallic materials", TWI Abingdon Hall Cambridge, 1997
[8]GKSS, Displacement Gauge System for Applications in Fracture Mechanics, Patent Publication, Geesthacht, 1992
[9]Akselsen, O.M. and Rorvik, G., "Tensile properties of heat affected zone of medium strength low carbon C – Mn and 2.25Cr – 1Mo steels, Materials Science and Technology, Vol. 6, 1990, pp. 383-390
[10]Schwalbe, K.-H., "The Prediction of Failure Situations Using the CTOD Concept Based on the Engineering Treatment Model (ETM)", The Crack Tip Opening Displacement in Elastic-Plastic Fracture Mechanics Workshop on CTOD Methodology, Geesthacht, April 23-35, 1985
1