Destructive MSA
If you think about MSA it makes perfect sense on destructive tests. Let's use a tensile test as an example. You can test both low an high strength samples so it seems to me that one would benefit from knowing if the tensile machine were exhibiting any deviations from linearity over its working range. Is a tensile test subject to seaonal variability in the electronics due to temperature and humidity or are any of the mechanical components subject to wear? In our case the answer is yes, so stability studies are most definitely appropriate. Are you always testing at the same strength level, if so linearity is not needed but bias(accuracy) must be known. Finally, there are most definitely methods for doing destructive R&R's. The only challenge to destructive MSA is getting working standards. These can be obtained by participating in an ASTM round robin test or the like. In our case, we have taken samples and annealed them to reduce variation in strength as much as possible. We pull one of these"master" samples at the beginning of every shift and plot them on an SPC chart. Using the standard SPC methodology it becomes immediately apparent if there is a potential problem with the tensile machine. Sure the load cells and the tensile machines are calibrated annually but that does not go far enough. You need confidence that these machines are working properly every day when you are certifying material. BTW, MSA is a hot topic at audit times these days so be warned!