It can also be a switching transient in the sense of a very brief overload - if the pick up coil in the RCD is not perfectly balanced (and real magnetic devices are good but never perfect ) then consider that even a 1% magnetic imbalance on a circuit carrying 10 amps looks like a  100mA of the kind that ought to fire the RCD  - in fact for some cheaper RCD designs this effect can be a limiting factor for inrush current handling.

I'd hope the switches are break before make if they are jumping motor windings, but you never know - if they manage a brief short then the current for the next half cycle may be much more than the normal load.

I think as the next question will be what to do about it the first thing to do is to look inside and see how they are made. It may be possible to suppress the worst effects either with series NTC (for inrush limiting) or shunt snubber (for voltage spikes) arrangements.. Or it may be that a change of supply polarity or brand of RCD is the way forward.
Mike.

Switching to a different brand of RCD might be a wise decision. I have observed that a low-cost 300 mA RCD tripped prematurely at 195 mA, whereas a more costly one tripped at a closer value of 225 mA.

Switching to a different brand of RCD might be a wise decision. I have observed that a low-cost 300 mA RCD tripped prematurely at 195 mA, whereas a more costly one tripped at a closer value of 225 mA.