There's a good write up on the commercial forum in the "Running two queen colonies" thread. It's post number 3 by Oldtimer. I'm not sure how to link it so this is the copy.
One place I worked did this as standard practise it worked very well for both swarm control and requeening.
Hives were wintered in double deeps. In spring before swarming time, all the hives with 2 year old queens (yes in those pre varroa days, queens lived 2 years), the queen was found and put in the bottom box. An excluder was put on top and then a honey super, there was no flow yet the super was just incase things got crowded. A division board was put on top of that and the other brood box put on that with basically a nuc in it and this was given a queen cell.
A few 1 year old queen hives were done also because there would not be 100% mating.
Around a month later the hives would be combined. The top box above the division board now (hopefully) with a laying queen, was taken off. The honey super was taken off and the excluder taken off. So the hive was now down to the bottom box with the old queen. 6 Sheets of newspaper were put on and then the excluder. Then the top box with the new queen and a choc was put in each front corner on the excluder to raise the second box to make a second entrance. This stopped the nuc suffocating prior to the paper getting chewed out plus these hives later got strong and needed that extra entrance. 2 sheets of paper were put on the second box plus another excluder and the honey box on top of that.
So the configuration from the bottom, was bottom box with old queen, excluder, second box with new queen, excluder, honey super.
After this we did not open the brood nest again that season, just more supers were added as needed.
In fall, honey supers were removed, then the hives wintered down, which involved pulling the excluders and feeding hives that needed it to get required amount of feed for winter. Come spring, the old queen would be gone just the young one was left.
At the time we gave the queen cells, the hive mats were marked with the strain of the queen and the date. So we could know what queen was in each hive by looking at the mat.
In that area, there was a sharp flow lasting roughly 2 1/2 months. Once this flow started the bees lost all desire to swarm and focussed on honey collection. So the 2 queening was timed so that leading into swarming time, the hives were split with the unit & cell put on top. This effectively stopped the hive swarming. Then the hives were recombined a few weeks before the flow would start. They just had time to sort themselves out & build a very strong hive in time for the flow to start. So we achieved requeening, a strong hive just in time for the flow, and swarm control.
The method was perfectly suited for that area, and is still done by some beekeepers in that area. (Canterbury plains, New Zealand).
However when I moved North around 600 miles closer to the equator, no beekeepers used 2 queening, and I found out why when I tried it. Here winters were warm and the bees came out strong and wanting to swarm. There was some kind of flow for around 7 or more months so the seasons were not well defined. Bees happy to swarm anytime, flow or not. So when I tried 2 queening all I got was super strong hives that were impossible to prevent from swarming.
So the method can work well in some areas, poorly in others. It is a prime example of how locality can affect management methods, what is superb in one area is lousy in another.