Studies on the Effect of Process Aspects on Material Mixing and Defect Formation in Friction Stir Welding

Abstract

Friction Stir Welding (FSW) is a rapidly growing solid state welding process and has been a proven method for welding high strength aluminium alloys which were formerly not recommended for joining by conventional fusion welding methods. Based on the information acquired from previous studies, to obtain a defect free Friction Stir (FS) weld with suitable strength, three basic requirements need to be fulfilled (i) Filling of the cavity created behind the tool pin during its traverse and ensuring satisfactory contact of filled material with newly generated surface (on advancing side trailing edge of the pin) (ii) Disrupting and distributing the oxide layer at the initial weld interface (iii) Adequate level of mixing of both side material (Adjacent and Retreating side) in similar welding. In the case of dissimilar welding mixing is desired in controlled amount (to prevent or curtail formation of intermetallics) depending on material combination. Failure to achieve the first precondition results in void. Second and third precondition are interconnected for similar FSW as adequate mixing in weld helps in disruption and distribution of oxide layer at initial weld interface. Failure to achieve this, results in Joint Line Remnant (JLR). Metal to metal contact cannot be established due to the presence of JLR (aligned oxide particles) and subsequently initial interface is left unwelded which deteriorate the static and dynamic strength of friction stir welds. The problem aggravates while friction stir welding materials with tenacious contaminant layer. Therefore, appropriate stirring (which entails large deformation and mixing) of initial weld interface is essential for successful FS welds. Hence, process aspects assisting mixing of adjacent (Advancing and Retreating side) materials need to identified and studied, which are missing in former studies. Experiments are conducted with classical FS tool (possessing frustum shaped/tapered circular pin) to analyse the effect of welding parameters (tool rotation speed, traverse speed, plunge depth, tool tilt and tool position w.r.t initial interface) and tool runout by changing these parameters over a range. Tool rotation speed, traverse speed, plunge depth and tool position with initial interface are changed continuously and tool tilt and tool runout are changed in discrete steps. Tool geometry is considered to be a prime parameter controlling the magnitude of mixing, as interaction of rotating tool with initial abutting base metal interface makes the process mechanism complex, unlike other solid state welding process, namely forge welding, diffusion welding, friction welding, explosive welding, ultrasonic welding and roll bonding. Furthermore, due to asymmetric nature of material flow in FSW process, the material located in different locations with respect to the tool is subjected to different levels of deformation. For this purpose experiments have been carried out to analyse the effect of different tool geometrical aspects on level of mixing and material flow. On the other hand, visualizing flow and mixing in metals is debatable as insertion of marker material in the weld line can alter the nature of material flow in the weld due to different material flow characteristics of the base and marker materials and introduction of additional interfaces. Further, using dissimilar materials for flow studies cannot be considered for comparison with similar friction stir welds as their flow properties are different. Therefore, an alternate experimental strategy is devised in these studies using physical modelling approach which is effective and helps in identifying and quantifying mixing observed under different tooling and process conditions. In the present investigation, plasticine of primary colours is adopted and the hue attribute of colour is used to study and quantify intermixing. Yellow and Blue plasticine are placed on advancing and retreating sides respectively. The degree of mixing is indicated by the intensity of generated green. Digital images of the cross section in weld nugget region are taken. To obtain hue component of these digital images the RGB color-maps are converted to HSV color-maps. Overall, these studies help in formulating the guidelines which are useful during tool design, and administering the process to obtain a defect free well mixed welds. Based on the experimental results following conclusions are derived. 1. Following process aspects: tool geometry, interface offset, tool rotation and tool runout demonstrate a significant impact on material mixing and breaking and dispersion of initial interface in weld nugget. Tool tilt, plunge depth, tool traverse exhibit negligible effect on degree of mixing. 2. Increase in tool rotation speed (with other parameters fixed) improves mixing substantially but can be increased to a certain limit after which voids emerge due to loss of weld nugget material in the form of flash. 3. Reducing the weld pitch (i.e. increasing tool rotation speed for a given tool traverse speed) reduces the size of the weld nugget and vice versa. Tool traverse speed largely affects advancing side material and rotation speed affects retreating side material. Therefore, for higher weld pitch advancing side material (yellow plasticine) dominates the weld nugget, whereas for lower weld pitch retreating side material (blue plasticine) dominates the weld nugget. 4. The extended macro-structural feature commonly observed in FS welds occurs under influence of plunge depth. Consequently, this macro-structural feature serves as the demarcation point between shoulder affected and pin induced material flow in FS weld. 5. The degree of mixing and subsequent elimination of JLR, improves, when original interface is offset on the advancing side w.r.t tool axis for all the tools investigated in the present study. Triangular and square pin generate larger pin induced mixing which intensifies further with interface on advancing side, indicating tools with such profiles to possess larger safe zone with better mixing characteristics 6. At zero interface offset with all the process parameters fixed, tapered triangular and square pin profile tools produce welds with maximum mixing. For pins with faces, material is transported in lumps around the pin. The size of lump increases with lesser number of faces on pin. Material in the vicinity of the pin experiences spinning/whirling movement. The volume of material experiencing spinning in a single tool revolution depends on (a) weld-pitch (lesser volume of material for smaller weld pitch and vice versa) and (b) number of faces on the pin (lesser volume of material for greater number of faces and vice versa). Therefore, circular pin which can be considered to be made of infinite faces, spinning of material occurs at micro level for relatively smaller weld pitch. 7. For classical FS tool (tapered circular/frustum shape), there exists an optimum ratio (shoulder diameter/pin diameter) situated between 2.7 to 3.6 to produce void free well mixed welds. Tools with ratio of 2.7 and below possess a tendency to produce welds with void but with better mixing in weld region. Tools with ratio of 3.6 and above possess a tendency to produce void free welds but with poor mixing in weld region. Voids appear and grow under following circumstances (a) with increase in pin diameter (for a fixed shoulder diameter), (b) with decrease in shoulder diameter (for a fixed pin diameter), (c) with decrease in pin taper (for a fixed shoulder diameter and top diameter of pin). Pin length has no effect on void formation. However, it is obvious, length of root defect increases with decrease in pin length. The tooling guidelines established in this study through plasticine work can be extended to metallic friction stir welds of various thickness plates by proportionately increasing or decreasing the tool dimensions as long as they fall in the recommended range. 8. Smaller pin diameter tools exhibit higher optimum weld pitch (but with lower degree of mixing) when compared to larger pins (but with higher degree of mixing). Optimum weld pitch represents weld pitch resulting in void free welds. Consequently, tools with higher optimum weld pitch help in welding at a better rate. 9. Tool runout is replicated through tools with eccentric pins. It is interesting to note that, all the tools with pin eccentricities do not assist in mixing but tools with only certain eccentricities (0.3 and 0.6mm assisted in mixing in the present investigation). It implies that tool runout of certain values facilitate mixing in weld. On the other hand density of void increases with eccentricity of pin/tool runout. 10. In dissimilar FSW investigated with plasticine A, B, C and D possessing different flow stresses (flow stresses ascending in the order of A, B, C and D) and strain rate sensitivity of 0.24, 0.22, 0.19 and 0.18 respectively, following inferences are drawn (i) For combination A and B, weldability improves when plasticine B is on Advancing Side (AS) and A is on Retreating Side (RS). The level of mixing also improves when interface is on AS (w.r.t tool axis) for this handedness. On the contrary, severe discontinuities emerge when plasticine B is on RS and A is on AS, especially when interface is closer to the tool pin axis. (ii) For combination A and C, weldability improves when C is on AS and A is on RS. The level of mixing also enhances when interface is on AS (w.r.t tool axis) for this handedness. (iii) For combination A and D, joining is poor for both the handedness. However, nature of defect is different in both the combinations. Cracks are observed when A is located on AS and voids emerge when D is located on AS. On the other hand, placing A on AS results in weld thinning. (iv) For combination B and C, there is no appreciable change in terms of weldability and level of mixing. Both the handedness in this combination yielded fairly similar results. (v) For combination B and D, though discontinuities do not emerge with change in handedness, mixing in weld improves when B is on AS unlike to its location on RS. (vi) For combination C and D, there is no appreciable change in terms of defect formation and level of mixing with change in handedness. Both the handedness in this combination yield fairly similar results.