I think that rather in the way there are fatigue test rigs that keep ahead of the aircraft in fatigue life to identify weak spots,
It is however quite usual to find cracks on the in-service aircraft which have not yet occured (or even will not occur during the whole test), simply because the test conditions do perfectly match the design requirements, but unfortunately not the real operational life... There have always been surprises, there will allways be surprises. That´s why we design and test with safety margins and second lines of defense.
So unless we find a production quality issue to be the cause of the failure(s), clearly the tests were not simulating real life. This can happen for new technology. So analogue to the fatigue test: add temperature cycling, environmental exposure etc. to the test, and you may obtain better results. Experience however shows that we can ignore these effects. Unfortunately for batteries we still do not know what we have to simulate (Temperature, Pressure, Vibrations...) to have an effective test. If we test the wrong conditions, it does not help if we are far ahead of the real fleet with respect to simulated flight hours. It is easy to add costs to a test by adding new aspects. It is hard to add relevance to a test...