Develop a more extensive process model of Friction Stir Welding that identifies background variables and accounts for variability.
While researching the relationship between traverse force and tool pin length, a discontinuity was discovered. The general trend was that the required force increases proportional to pin length. This trend continues until a ¼ inch pin length, where the required force decreases. Replicate experiments produced similar results with large variability in the traverse force response.
In order to investigate or reproduce these findings, it becomes important to derive a scientific explanation for the variability. Once the source of variability is known, the discontinuity and the factors that contribute to it can be better understood.
A new experimental method is proposed which would use a supersaturated design to link statistically significant main effects to previously unobserved factors. Once these factors and their net effects are better understood, the discontinuity can be reproduced with less variability.
A supersaturated screening design will be used to reveal statistically significant factors and discover important background variables, factors that affect the process significantly but have gone largely unnoticed. Once factors with a significant main effect are identified, further experimentation can develop a more detailed process model.
The immediate growth of the Friction Stir industry would benefit greatly from a more detailed process model than is currently available. Such a model would allow for greater optimization of the overall process, improving weld properties while minimizing processing time and required power.
Figure 1. X Force (the traversing force) vs. Pin Length. Note the discontinuity and variability for a pin length of ¼ inch.