I'll put it this way:
The chances that it is something on the LINE side causing ONLY THIS BREAKER to trip are far far slimmer than something on the LOAD side. If you had for example a grid glitch that caused a brief outage and if it came back on almost immediately, and if you had some big spinning loads, then if their motors had not had time to collapse the magnetic fields, the regen from those big spinning loads that can occur because the restoration of power is at a different frequency than the regen can cause a severe transient that can trip the magnetic trips of the breaker. Notice there are a lot of "ifs" in there and by the way, you could STILL consider this a load problem, because all you have to do to fix it is to not allow loads to stay on after an outage.
Or how about this: the likelihood that the breaker is DOING ITS JOB and preventing a fire because of some yet undiscovered but perfectly plausible issue on the load side are far greater than it being defective or getting fooled by some "phantom" even on the line side.
Before wasting a lot more time, I'd find out if a coordination study was ever done. If it was, then immediately check ALL other protective devices down stream of this breaker to make sure they are EXACTLY as depicted in that study. The one that is not the same or the one that is new and not included is the first place to look for undiscovered problems. If there was NO coordination study done (as I suspect), then that may be problem #1. You may have a down stream protective device that is set to trip at a level higher than the setting of this breaker. That's exactly what a "coordination" study prevents, it's intended to make sure that protective device settings are coordinated so that faults are cleared at the lowest (closest) possible level.
