I do not agree with the "fail." This is not a situation that involves a "series-rated combination." You are not passing the fault current through two devices in series, with the expectation that, on the basis of some prior testing, the combination will sucessfully trip the circuit before anything gets damaged. Instead, someone calculated that the amount of fault current available at the equipment was 14kA. However, that person should have changed that calculation in light of the specific overcurrent device (note: singular, "device") that was being installed. The presence of that device changes the amount of fault current available at the equipment. If there were a circuit breaker instead of a current-limiting fuse, the system might have be able to inflict 14KA of fault current. But with the fault limiting fuse in the circuit, the amount of fault current available is now just 3kA.