The snag with all these conversions is getting enough flux in the gap. Perfectly curved magnets about 1/4" thick would most likely be fine. With the gaps between round magnets and the inability to follow the curvature correctly I suspect that you would do better with at least 3/8" thick. you are unlikely to saturate the rotor core  even if you reduce diameter to use thicker magnets. You will hit saturation in the teeth first so I wouldn't worry too much about the rotor iron unless you are planning to use the original cage rotor and you get down into the part where you have more aluminium cage than steel core.

If you can get N50 cost effectively the go for it, but it isn't normally cost effective and using 3/8 thick N40/42 would be better than 1/4" N50.

I suspect that the 230v connection will be slow for a 24v machine so you may well have to dive inside to split the pole groups. If you can break the star point and use Jerry connection it may be better. I find all this infinitely more hassle than building a dual rotor but everyone to his own choice.