BTW: I worked for a company that required heat treating of some small steel covers not for increased hardness, but for the change in the metals magnetic properties. So verifying the hardness would not have validated the process at all.
Validation can only be performed against the specification requirements - if this was hardness, grainsize or magnetic properties - these would be the factors that are validated against.
A special process is any production or service delivery process that generates outputs that cannot be measured, monitored, or verified until it's too late.
I would disagree with this definition - A special process is one where the output cannot
effectively inspected, tested or verified.
An example - I worked on a blasting process that had to achieve a specific surface finish (between 6 and 20 Ra). Obviously we can measure surface finish, but only at discrete points, measuring the surface finish at every point would be completely excessive, take far too much time and just be ridiculous. Hence it was treated as a special process and was validated.
This has a parallel in heat treatment - hardness can only be measured at discrete points on the surface on the product, grainsize at sections where it is cut up mounted & polished etc etc
How do you know the heat treatment process gives even properties across the component? Different component geometries, variations in the component before heat treatment, variations in temperature uniformity across the furnace or heating process and many other factors can affect these properties - hence without validation of the process at the worst case limits of the variables, or combinations of variables, hardness testing each output from the process is insufficient.
Heat Treatment is a different process, and basically in order to be effective all that is required is that temperature and time be controlled.
What about cooling rates, atmosphere, pressure etc etc? Are there not plenty of other variables that would need to be considered?
So what you need is to be able to show that each parts batch was exposed to the correct temperature and time. It also follows that you have to periodically calibrate (verify) that the indicated temperature is indeed correct
Process records like these do not validate the process. To validate the process you would need to show the process, even at the worst case limits of each variable or combination of variables still produces conforming product. The process records would then be of use in showing that the manufacturing process conforms to this validated process.