Bennie covered it well with the hard start situations of a Transformer under load. Glad to see you re-thought the secondary load problem! That's where the problem will fall under.
If you have secondary loads with a long inrush time, the Transformer will need to deliver that via the Primary windings. This, along with the inherit inrush of a transformer will be seen on the Primary as a high and long duration inrush.
Using circuit breakers with adjustable trips could help this situation, unless you can install a C/B frame with time current curves similar to the ones of an FRN/RK5 fuse.
One thought would be to unload any non-essential loads, especially those with high inrush values-like motors- during transfer over to separate systems, just long enough to allow the transfer and stabilize the circuits/system. If you find transfer equipment that has better time lag trips, that is great.
Any Induction motors that are running when the transfer takes place will do the same as if there is a short circuit - they become Transformers. Loaded motors supply the most KVA, which is near the LRA of the motor for a motor loaded to near it's capacity. [kind of interesting, isn't it??
..]. The transfer acts as if there is a temporary loss of KVA supplying the motor. Since the Induction motor is just a Transformer with a rotating secondary, it will supply KVA through the Inductive Reactance that is present in the primary [windings], along with the XL in the secondaries [rotor]. This continues until either the load stops turning the rotor at a sufficient speed that can create the Transformer action, or the system becomes stable. A locked rotor will not develop KVA outwards, but requires an inrush KVA to get things moving- similar to the way a Transformer needs a high inrush current to get a high flux current flow moving in the core.
After the system becomes energized again, the motors will demand any inrush /LRA according to how much KVA is required to get the loads moving again, up to the speed needed [or slipped below a synchronous speed].
Also, any HID lighting will have the same scenario of starting as it would normally. Some HID ballasts draw a high starting current, which tapers off to the somewhat constant VA load it has while at full operation. CWA [Constant Wattage Autotransformer] Ballasted HID lamps will have the same KVA [or current] load through starting and operating. HPS lamps typically use the CWA ballast, as do most Metal Halides.
Anyhow, hope this was interesting.
My father had to do the breaker cooling trick one time. He was the EC on Playboy Mansion West, back in the late 60's/early 70's. When the place was almost finished [everyone knows that situation with clients!!], Mr. Hefner wanted to throw one of his world famous parties. The load was extremely high on one certain subpanel, due to the fact that things were kind of temped here and there for his party!! To keep the breaker from tripping, they used a fire extinguisher to cool the frame [the CO2 type, I guess]. He told me this story about 15 years ago when I was getting serious about the technical side of the trade [you know - the young wanna-be Engineer!!]. Not too sure if He would admit to doing this now
That was the only trouble with that entire project, and everything is still rock solid to this day.
There is a problem with Utility power quality. The location is still fed via 4160 3 phase utility distribution and on a very small branch. Since it's nearly impossible to upgrade for the Utility Co. [DWP], there is a high occurrence of power loss!!
To remedy this, the on-site crew have installed diesel powered backup systems. They have a nice one now, which from what I have heard is really huge! I think it's rated at like 750 KVA, but not too sure.
Need to go back and visit everyone there. I can get some insight to their UPS / backup protocols and all the extra details! It's been over 7 years since I have last been there, so it will be a nice trip.
See everyone in a couple of weeks! I am off for a "Vacation". details to follow.