1) there is nothing repairable on a PF4 or PF40, they are all surface mount single board (for the most part) and the power devices are IPMs, meaning all of the diodes and transistors are in a single potted unit. Some people attempt to do repairs on surface mount boards, they generally spend more fixing them than they do replacing them.
2) The only possible thing that can change as drives get old is that the caps fail, in which case the drive would stop functioning, or at best MAYBE trip on UNDER voltage, not OVER voltage. So that's not it, and to replace them without finding the problem is pointless.
3) The first thing I would start looking at is what has changed? If it NEVER did this in the past, but started doing it recently, what's different? If on the other hand it has always done this, then it's a design flaw from the outset. Being that it is getting attributed to age, I'd think it was working fine for a long time, so again, what's changed? Even something you might think is innocuous can have effects nobody considered. For example, lets say that in the past, the conveyors were used less and never tripped while loaded, so never had to re-start while loaded. It's then likely that when you re-start your conveyors, inertia in the load ON the conveyors is causing the motors to regenerate as soon as they start to move. I've seen this before; the infeed system that is putting load ON the conveyor has been not doing so when off, so when re-started, that end of the conveyor is empty, but the other end, possibly going down hill, is full of product. So as soon as you re-start, the lack of load at one end means the downhill portion accelerates beyond what the drive is telling it to and turns the motors into generators, which trips the drives on over voltage.
Again, that's just an example to make you think through of what MIGHT have changed when it started happening.