FATIGUE AND FRACTURE OF DISSIMILAR FRICTION STIR WELDS

FATIGUE AND FRACTURE OF DISSIMILAR FRICTION STIR WELDS

N. Manuel1*, C. Silva1, L.P. Borrego1, 2, J. M. Costa1, A. Loureiro1

1CEMMPRE, Mechanical Engineering Department, University of Coimbra, Rua Luís Reis Santos, 3030-788 Coimbra,

Portugal 2Instituto Politécnico de Coimbra, ISEC, Department of Mechanical Engineering, Rua Pedro Nunes, 3030-199

Coimbra, Portugal * J. M. Costa: [email protected]

ABSTRACT Friction stir welding (FSW) of T-joints presents several difficulties related to the reduction in the skin thickness, to the production of skin/stringer fillets and to the formation of weld defects. This study aims to analyse the morphology and mechanical behaviour of dissimilar FSW in T geometry of alloys AA6082-T6 and AA5083-H111. The welds were made with a progressive pyramidal tool developed for this purpose, using both alloys as either skin or stringer and an additional plate to prevent skin thickness reduction. The welding parameters were chosen so as to prevent the formation of weld defects. The adopted procedures allowed to obtain welds without macroscopic defects, without reduction in thickness of the skin and with fillets skin / stringer with excellent surface finish. The tensile tests of the skin of the welds caused the fracture of the test pieces in the heat affected zone of the welds whether the skin is AA6082-T6 or AA5083-H111. The fatigue strength was higher in welds where the skin was the alloy AA5083-H111 compared to those where AA6082-T6 was the skin. Fatigue fractures always occurred in the heat-affected zone on skin AA6082 welds whereas on skin AA5083 welds occurred in either the affected zone or the base material. KEY WORDS: Friction stir welding, dissimilar welds, aluminium alloys, fatigue strength

RESUMEN La soldadura con agitación por fricción (FSW) de las juntas en T presenta varias dificultades relacionadas con la reducción del grosor de la piel, la producción de filetes de piel / larguero y la formación de defectos de soldadura. Este estudio tiene como objetivo analizar la morfología y el comportamiento mecánico de FSW disimilares en la geometría T de las aleaciones AA6082-T6 y AA5083-H111. Las soldaduras se realizaron con una herramienta piramidal progresiva desarrollada para este propósito, utilizando ambas aleaciones como piel o larguero y una placa adicional para prevenir la reducción del espesor de la piel. Los parámetros de soldadura se eligieron de manera que se evitase la formación de defectos de soldadura. Los procedimientos adoptados permitieron obtener soldaduras sin defectos macroscópicos, sin reducción de espesor de la piel y con filetes de piel / larguero con excelente acabado superficial. Los ensayos de tracción de la piel de las soldaduras provocaron la fractura de las piezas de ensayo en la zona afectada por el calor de las soldaduras, sea la piel AA6082-T6 o AA5083-H111. La resistencia a la fatiga fue mayor en soldaduras en las que la piel era la aleación AA5083-H111 en comparación con aquellas en las que se encontraba AA6082-T6. Las fracturas por fatiga siempre ocurrieron en la zona afectada por el calor en las soldaduras de la piel AA6082, mientras que en la piel las soldaduras AA5083 ocurrieron en la zona afectada o en el material base. PALABRAS CLAVE: Soldadura por fricción, soldaduras disimilares, aleaciones de aluminio, resistencia a la fatiga. 1. INTRODUTION T-joint welds of aluminum alloys are of great importance in the transport industries as they have an excellent relationship between strength and density, allowing significant reductions in the weight of the structural components. The alloys AA5083-H111 and AA6082-T6 are alloys of different families widely used today [1]. T joints are widely used in welded structures

because they provide high rigidity and good flexural strength [2]. Conventional fusion welding methods, such as MIG welding, cause problems in the welding of these materials, such as porosity, hot cracking, fatigue and strong reduction of mechanical properties. The difficulties of welding dissimilar joints are even greater. FSW welding, because it is a solid state welding

Anales de Mecánica de la Fractura 34, 2017

308

process, removes or reduces these problems [3]. However, FSW like any other process, may present several difficulties. The reduction in thickness is a defect which currently appears on the skin of T-joints, which affects the mechanical strength of the weld [4]. Thus, in order to improve the resistance of these welds it is necessary to avoid the reduction of skin thickness and the stress concentration in the fillets. Another defect that can be found in dissimilar T-joints is the oxide lines. These are zigzag lines only observable after chemical etching. Song-Woo et al., 2004 [5] suggest that the oxide lines do not correspond to the breaking zone in bending or tensile tests, however, they state that the oxide lines may be the preferred crack initiation zones in fatigue testing, since the mechanical properties of that zone will depend on the size, density and location of the oxides present in the weld. In dissimilar welds there is increasing difficulty in choosing and optimizing the process parameters in order to avoid the formation of defects. Fratini et al., 2008 [6] report that the speed of rotation and feed rate of the tool are the main parameters which contribute to the heat generated in the joint, but the tool geometry also plays an important role in the morphology of T-joint welds. The quality of the weld is significantly affected by the speed of rotation, but high tool rotation speeds cause excessive heat generation producing burrs due to the strong plasticization of the material. The foregoing studies address essentially similar welds in alloys AA5000 or AA6000, in butt or overlap joints. The scientific literature on welds in dissimilar joints is still scarce [7], [8]. This research aims to contribute to solve difficulties mentioned above in order to obtain dissimilar T-Lap joints in alloys AA6082-T6 and AA5083-H111 without significant defects and to analyze the mechanical properties of the joints produced. 2. EXPERIMENTAL PROCEDURE

The dissimilar T-Lap welds were produced using base material plates in the alloys AA5083-H111 and AA6082-T6 of 3 mm thick. Their chemical compositions and mechanical properties are presented in Tables 1 and 2, respectively.

Table 1 – Chemical composition of the alloys (% wt).

Liga Cu Mg Mn Fe Si Cr

AA5083‐H111 0,02 4,5 0,57 0,18 0,09 0,06

AA6082‐T6 0,09 0,7 1 0,44 0,53 0,02

Table 2 – Mechanical Properties of Alloys AA5083-H111 and AA6082-T6.

Mechanical properties AA5083‐H111

AA6082‐T6

Tensile strength [MPa] 340,62 321,2

Yield strength [MPa] 158 288

Elongation [mm] 10,4 8,6

Hardness [ . ] 83,5 116 The T-Lap joint configuration was chosen in this study to solve the problem of the thickness reduction in the skin. An extra 1mm thick plate was placed on top of the skin to provide material to fill the fillets without having to resort to the material of the skin. The aluminum plates were cut into 160x330x3 mm for the skin, 33x330x3 mm for the stringer and 100x330x1 mm for the plate over the skin. Figure 1 schematically illustrates the geometry of the adopted T-Lap joint.

Figure 1 – T-Lap joint configuration

Two series of welds were carried out. The welds of the 65LPP series consist of a stringer in the AA5083 alloy and a skin in the alloy AA6082 and the welds of the series 56LPP the skin plate was in the alloy AA5083 and the stringer in the alloy AA6082. The plate placed on the skin was in the same material as the skin. The tool with progressive pyramidal pin was used to allow a more intense flow of the material and the heat input directly in the zone of the stringer and of the fillets. The tool was made of hot working steel H13, subsequently quenched and tempered to obtain a surface hardness of 50HRC. The tool has a concave shoulder of 18mm diameter. The cylindrical pin part is threaded with a length of 3.7mm and the pyramidal pin of 2.5mm length. Figure 2 illustrates the geometry of the tool.

Figure 2. Pyramidal progressive tool.


309

The welding parameters used in the two series are indicated in table 3. These parameters were chosen based on previous work. The welds were made on a Cincinnati Milacron 207Mk milling machine which has high stiffness required for the process.

Table 3 – Welding parameters used in the series.

Series W

(rpm) V

(mm/min Tool depth

(mm)

65LPP 1140 60 6,8

56LPP 500 30 7 After welding the test pieces for metallographic analysis and hardness measurement were cut approximately 90x19mm size and polished by conventional techniques. In order to reveal the microstructure and grain, the specimens were chemically etched with modified Poulton’s reagent. The samples were analyzed and photographed in an optical microscope Leica MD4000 M LED. The Vickers microhardness analysis was performed by applying a 200g load for 15s using a Struers Duramim 1 machine. The hardness profile along the skin was analyzed at 0.5mm from the lower surface of the skin until the distance of 20mm from the central zone to each side. The test specimens for the skin fatigue and tensile tests were cut transversely to the weld approximately 160x20mm size and the fatigue tests were still machined in the form of dog bone. The tensile tests were performed using an Instron 4206 machine and the values of the strains were acquired using a GOM ARAMIS optical system with digital image correlation. A DARTEC servo-hydraulic machine coupled to an Instron Fast Track 8800 acquisition and control system was used for the fatigue tests and determination of the S / N curves. In all fatigue tests, the stress ratio used was 0.05 and the stress range varied between 100 and 190 MPa with a frequency between 15 and 25 Hz, depending on the maximum load applied. A Carl Zeiss Gemini Scanning Electron Microscope was used to analyze the fracture surfaces of fatigue specimens. 3. RESULTS AND DISCUSSION 3.1. Metallographic analysis The progressive pyramidal tool allowed to obtain welds without significant defects in the 65 series, however, in the preliminary tests, the welding produced with 1140 rpm and 60 mm / min of tool rotation speed and advance rate, respectively, presented superficial defects in the advancing side. These defects diminished until disappearing as the penetration of the tool increased. For a penetration of 6.8 mm the weld showed no defects, as shown in Figure 3a.

a) 65LPP

b) 56LPP

Figure 3 – Welds morphology: a) 65LPP e b) 56LPP.

The welds of the 56LPP series, produced at 600 rpm and 60 mm / min, showed along their length a surface cavity on the advancing side. This defect was decreasing successively for larger tool penetrations, without having closed completely. For minor rotation and feed speeds of 500 rpm and 30 mm / min, defects for tool penetrations of 7 mm were eliminated as shown in Figure 3b. Figure 3 also shows that the plate superimposed on the T-Lap geometry allowed to obtain perfect fillets without any reduction in thickness of the skin. It was verified that the adopted strategy allowed to obtain the desired results. An analysis in greater magnification allowed to observe in more detail the structure of the grain in the zone of the nugget. The average grain size in the nugget is influenced by the grain size of the base material and the heat input in the weld. The mean grain size measured was 7.4 μm in the 56LPP weld series produced with low tool rotation speed while was 11.4 μm for 65LPP welds performed with higher rotation and feed speeds. 3.2. Microhardness Figure 4 shows the hardness profiles measured throughout the skin for both series and their base materials. The hardness of the base materials was of 83.5 HV0.2 and 116 HV0.2 respectively for the alloys AA5083 and AA6082 and are represented by a solid line and a dotted line, respectively. For 65LPP welding it is possible to identify the typical W curve associated with thermally treatable alloys. The loss of hardness observed in the HAZ and TMAZ zone is due to the dissolution and coalescence of hardening precipitates. This result suggests that the alloy present on the stringer or sheet superimposed on the skin does not significantly influence the hardness profile throughout the skin. For the welds of the 56LPP series there is a slight loss of hardness compared to the base material in HAZ and


310

TMAZ, with a small increase in hardness in the nugget. This increase in hardness is due to the plastic deformation induced by the tool.

Figure 4. Profile of microhardness of welds 65LPP and 56LPP.

3.3. Tensile behaviour Figure 5 shows an example of a force / elongation curve obtained for each series of welds.

Figure 5. Load-elongation curves for specimens of the 56LPP9 and 65LPP1 series.

The figure shows that between the two series of welds there is much difference in regard to the elongation suffered, in addition to the series 56LPP welds have higher values of load in relation to the series 65LPP. The maximum load sustained by the 65LPP series test pieces averaged 13.32 kN while for the 56LPP series test was 15.72 kN. The 65LPP series specimens generally broke into the HAZ or TMAZ along a line of oxides while the 56LPP series broke into the HAZ. Dividing the maximum strength by the section of the test pieces resistance losses in relation to the respective base materials in the order of 27% for the 65LPP series while only 18% for the 56LPP series were obtained. These results show that the welding process induced some loss of resistance in any of the series tested. 3.3. Fatigue behaviour The results of the fatigue tests are shown in figure 6. The stress ranges used varied between 100 and 180 MPa for 65LPP series and between 105 and 190 MPa for 56LPP series. Both weld series showed a fatigue

behavior much lower than the respective base materials, with the fatigue strength for the 65LPP series being about 44% lower than the respective skin BM while the 56LPP series being about 38% lower than the BM. The two weld series have a similar fatigue behaviour, although being superior for the 56LPP series, mainly for higher stress ranges. These results are in agreement with those observed in the tensile tests, where the highest resistance was observed for welds of the 56LPP series. The 65LPP series fatigue specimens fractured in the heat-affected zone while the series 56LPP specimens fractured either in the weld, in the ZTA and even in the base material.

Figure 6. S-N curves for welds and base materials.

3.4. Fracture surface analysis In Figure 7 the fracture surfaces of a 65LPP fatigue specimen are illustrated. In image 7 a), on the left side of the image, it is possible to observe at low magnification the fracture surface of the central zone of the specimen. In the right part of the image, in greater magnification, it is possible to identify a zone formed essentially by streaks characteristic of propagation by fatigue. In the Figure 7 b), characteristic plastic deformation dimples are shown, indicating that the fracture had essentially ductile character in the area surrounding the fatigue fracture, in the same specimen. Figure 8 shows the fracture surface of a 56LPP series fatigue specimen. In the illustrated sample the fracture surface showed essentially three types of fracture. In one zone a perfectly ductile fracture surface was observed, with very thin dimples present, as shown in Figure 8 a). These dimples were much thinner than those shown in Figure 7 (b). In addition, it was verified that the crack propagated along a line of oxides, which leads to the conclusion that these oxides may have contributed to the appearance of dimples. Figure 8 b) illustrates the fracture mode in the central zone of the specimen, which occurred by fatigue. Fracture was initiated in the point indicated by the arrow in the cross-section of the specimen, shown above from Figure 8 (a).


311

Figure 6. 65LPP weld fracture surface.

Figure 8 b) illustrates the typical striations of

fatigue crack propagation. The final rupture of the specimen occurred outside zones a and b, marked in the cross section of the specimen, with ductile morphology identical to that shown in Figure 7 b). 4. CONCLUSIONS This study has drawn the following conclusions:

It is possible to carry out dissimilar friction stir welding on the T-Lap joint without significant defects;

The technique of placing an extra thin sheet on the skin effectively resolves the skin thickness reduction defect;

At the T-Lap joint, the material present in the stringer and the extra sheet, showed no influence on either the hardness or the tensile strength of the welds;

The two series of welds showed similar fatigue strength. The fatigue strength of any of the series of welds is lower than the respective base material;

In the tensile tests and fatigue tests the fracture occurred in the HAZ or TMAZ or even in the base material.

Figure 7. 56LPP series fracture surface. ACKNOWLEDJEMENT The authors are indebted to the Portuguese Foundation for the Science and Technology (FCT) for the financial support through the Project UID/EMS/00285/2013. REFERENCES [1] Cui L., Yang, X., Xie Y., Hou., X., “Process Parameter Influence on Defect and Tensil Proprerties of Friction Stir Welded T-Joints on AA6061-T4 Sheets,” Materil and Design, 51, pp. 161-174, 2013. [2] Zhao Yong, Lilong Zhou, Quingzhao Wang, Keng Yan and Jiasheng Zou, “Defects and tensil properties os 6013 aluminum alloy T-joints by friction stir welding.,” Materials and Design, vol. 57, pp. 146-155, 2014


312

[3] Ana C.F. Silva, Daniel F.O. Braga, M. A.V de Figuereido, P.M.G. Moreira, “Friction Stir Welding of T-Joints Optimization,” Material and Design 55, Elsevier Ltd., pp. 120-127, 2014. [4] Yadava, M. K., Mishra R. S. Chen Y. L., Carlson B., Grant G. J., “Estudy of friction stir joining of thin Aluminium sheets in lap joint configuration,” Science and Technology of Welding and Joining, 15, pp. 70-75, 2010. [5] Song-Woo Song, Byung-Chul Kim, Tae-Jin Yoon, Nam-Kyu Kim, In-Bae Kim and Chung-Yun Kang, “Effect of Welding Parameters on Weld Formation and Mechanical Properties in Dissimilar Al Alloy Joints by FSW, Material Tansactions, 51, pp. 1319-1325, 2010. [6] Fratini, L., Buffa, G., Micari, F., Shivpuri. R., “On the material flow in FSW of T-joints: Influence of geometrical and technological parameters,” International Journal of Advanced Manufacturing Technology, 44, pp. 570-578, 2008. [7] S.T. Amancio-Filho, S. Sheikhia, J. F. dos Santosa and C. Bolfarini, “Preliminary study on the microstructure and mechanical properties of dissimilar friction stir welds in aircraft aluminium alloys 2024-T351 and 6056-T4,” Journal of Materials Processing Technology, 206(1-3), pp. 132-142, 2008. [8] H. Jamshidi Aval, “Microstructure and residual stress distributions in friction stir welding of dissimilar aluminium alloys,” Materials & Design, vol. 87, pp. 405-413, 2015. [9] Da Silva, A., Arruti, E., Janeiro, G., Aldanondo, E., Alvarez, P., Echeverria, A., “Material flow and mechanical behavior of dissimilar AA2024-T3 and AA7075-T6 Aluminium alloys friction stir welds,” Materials and Design, 32, pp. 2021-2027, 2011.


313

FATIGUE AND FRACTURE OF DISSIMILAR FRICTION STIR WELDS

Documents