In a recent conversation, someone asked me about the benefits of friction stir welding to rocket design. After having initially raised the subject, I then embarrassed myself by not being able to answer the question coherently, all my knowledge on the subject having suddenly dropped down the memory hole.
Naturally, it all came back to me later, when it did me no good. Hmpf. But like lemons to lemonade, my embarrassment becomes your enlightenment — here’s a brief intro to FSW: what it is and what it’s good for.
Friction stir welding (FSW) is not a “weld” process in the typical sense, as it does not involve the melting of metal in the parts to be joined. It is instead a form of localized forging operation, in which the metal on both sides of a joint is plasticized via friction and then interleaved between the moving tool face and a fixed reaction anvil.
A generic FSW butt-weld operation, for example, would start with two metal plates butted together and clamped to a sturdy, fixed structure. This structure, the “anvil”, is needed to react the forces imparted on the workpieces by the FSW tool. The tool resembles a bolt with a short grip-length and large-diameter head. The welder (basically, a standard multi-axis milling machine) drives the FSW tool’s “pin” into the joint, and moves it slowly along the weld line. The spinning pin plasticizes the mating surfaces of the parts via friction heating, and the softened material then swirls around to the trailing side of the pin. There, the material from the two parts is continuously interleaved, and squeezed between the tool’s shoulder (the “bolt head” above the pin) and the anvil. As the tool passes it leaves behind a flat surface, with only a faint swirl pattern the width of the tool shoulder to mark the join.
It’s straightforward, but when you see it done it looks like magic.
And in a sense it is — it’s a joining technique which “magically” eliminates many of the problems associated with fusion welding. The biggest direct benefit from FSW is that the weld has nearly the same mechanical properties as the parent material. Fusion welding melts the material to join the parts together, resulting in a bead of material with dramatically-reduced properties. The welded assembly can be subsequently heat-treated to restore some portion of the original properties at the joints, but this is an expensive additional step, and one which may not be possible with large or oddly-proportioned structures. The typical remedy in those cases is to make the weld lands thicker, in proportion to the knockdown in properties due to the weld — the zone with reduced mechanical properties is given a larger cross-sectional area to compensate.
A major drawback to this with regards to launch vehicles is that they are weight-critical structures with many total feet of welds — the increased thickness at the weld lands adds up quickly. Compounding this problem is the need to make the weld lands wide as well as thick, to permit the use of clamping tools and to shunt away heat from the weld and thereby limit the extent of the heat-affected zone where the reduction in properties occurs. When the weld land is significantly thicker than the acreage of the mated parts, the transition between these thicknesses must be affected through a taper or a series of steps, further adding to the weight of the joined parts (not to mention the design complexity and machining costs of the individual parts). By allowing the designer to use weld lands which are thinner and narrower, FSW can reduce the overall weight of the launch vehicle.
FSW increases design flexibility, as well. Alloys with desirable mechanical properties but poor weldability can be joined, as can different alloys and even dissimilar metals which cannot be joined with typical fusion welding techniques. FSW lends itself to the welding of odd geometries, extreme thicknesses, and even different and variable thicknesses.
Beyond the design benefits, there are the positive effects FSW has on produceability. It allows most welds to be performed in a single pass, rather than the multiple passes required for certain fusion welds, and requires no filler material (weld wire). Weld defects are almost nonexistent, and those which do occur are of a nature which permits easy detection through ordinary NDE methods. When a defect does occur, it can be eliminated with a second pass of the weld tool (albeit with a tiny additional decrease in the joint’s mechanical properties). Step and gap tolerances at the weld line can be relaxed in comparison to traditional fusion weld processes, weld quality is much less sensitive to surface preparation, and for aluminum alloys the need for an inert gas purge is eliminated. Finally, post-weld surface preparation (“bead shaving”) is eliminated, since there is no longer a raised weld bead and associated defects to shave off. These benefits reduce manufacturing costs by reducing the amount of touch labor and consumables required for the production of a given assembly.
Note that my description of the process only covers the most primitive form of the technology. Many variations and applications have been developed in the past decade, improving produceability, simplifying design of manufacturing tools, and even substituting for certain kinds of mechanical joints.
(This post was mostly off the top of my head, but I checked my facts where possible against the TWI website.)