Calibrate each machine in compliance with an ISO standard, and keep those data as reference. Then periodically recalibrate the machines noting changes in these data until each one fails. An analysis of the data after each failure usually show a marker that portends the failure which you can then use in the future as a diagnostic indicator. This should work because I was once in this sort of business.
I hope that this is helpful
__________________
EditorGBAnalysts - Give a boy fish and he will eat may be for a day, but teach him how to fish and he will eat everyday.