you are right, the stress will be gone when the diaphragm rests against the upper housing. as a pressure sensor, the diaphragm will move up and down under applied variable pressure (against a loaded spring on the stop/piston which is not shown). as another requirement, customer wants to apply motion to the stop/piston against the system pressure to verify the operation of the pressure sensor. seems the calculation of the stress is not an easy job (to me), i ran a test, it passed my life cycle target. thanks. to all.
ps, the diaphragm is not welded to the stop/piston.