I have yet to see a tandem motor situation (without a VFD) that does share the load totally equally! The difference is usually not much, but not equal either. The best I have seen is 1% delta between the two motors, the worst is around 4%, most are around 2% (running amps). I saw one where the delta was in excess of 60% even though the motors were identical, but I traced it to poor connections in one motor.
Gunnar is correct from a theoretical sense with regards to drooping characteristics making them share the load, but in practice I don't think that truly "identical" slip values exist. My guess would be that natural differences in materials and construction probably make motor response slightly different, even if the tested slip appeared the same in an unloaded condition. Still, 2% differential is nothing to get worried about.
I did a project least year with several VFDs on 800HP crushers with tandem motors (2 x 400HP on each). We had the VFDs for their intended purpose (changing finished product size), but instead of one x 800HP VFD we used 2 x 400HP, with one as a speed follower from the control system and the other in torque follower to the first. We managed to keep them within 1% delta at any speed; worked great. The side benefit we discovered later is that we were able to accomplish that with dissimilar motors as well! That proved to be a big benefit to the end user because they could get back up and running without having to find matched motors.
Twin VFDs for 2MW MV motors however would likely be hard to justify from a financial standpoint if variable speed was unimportant. The only kind of crusher I have seen be successful with speed modulation is Vertical Shaft Impactors (VSI). Jaws, HSIs, Rollers, Hammer Mills and Cones (Gyratory) do not benefit from changing speed on the fly.
Never heard of "phase alignment", sounds like a red herring to me. Belt tension is critical, motor lead connection is critical (as I mentioned above), bearing wear is critical, mechanical alignment is critical (I always recommend laser alignment), and making sure that the location of each motor is such that they are balanced in the structure, otherwise torque on the frame can pull them out of alignment. The nature of the beast is that even slight differences in those things can set up a load imbalance that can escalate quickly. On the ones I've done without VFDs, I put current sensing relays on each motor and look at the differential. If it climbs too much above what it is at commissioning, I would set up an alarm to have them go look at all of those issues above.
