My guess is that, doing it the way you have chosen, you will need to trim your long edges so all the longitudinal seams have progressively bigger gaps as you move towards the (bow and stern) ends of each length of curved plate, in comparison with the ideal (small) gap at the midsection.
You should start by tacking each of these semi-cylindrical plate runs to the frames only near their mid-length, then start welding the seams from the middle towards the larger gaps at each end, rather than the usual skip welding. I would still keep the beads short, and alternate between welding towards the bow and towards the stern.
is that this will induce a compound curve in each piece, by shrinking the weld zones differentially. You will certainly find that the more open seams at the end will close as you weld towards them. As (say) the one nearest the keel
closes significantly, switch to the other side of the same plate and run some weld along the seam nearest the gunwhale. Try to keep the gaps even, IOW.
The key would be not to do ANY welding near the ends until the plate has pulled down (at the ends) to where you want it: ie kissing the stringers and/or frames.
However if you judge that it's pulling down TOO rapidly when you're still some distance from the ends, leave some gaps, and come back to them later.
The reason welding will induce a compound curve is that the edges of the semi-cylinder will shorten due to the weld metal bonding when hot, then shrinking as it cools. If you were to lie your semi-cylinder on the bench, so it can rock about, then run weld beads along the uppermost edges, they will be shortened in comparison with the 'silhouette edge' touching the bench. Consequently both ends of that silhouette edge will lift
off the bench, at which point you have your compound curve.
Getting back to your boat: If you do any welds whatsoever to adjacent plates or to framing near the ends before the desired shrinkage has taken place, you'll have 'locked in' the cylindrical shape, and you'll have a fight you won't be able to win.
This method would not
be a good idea with aluminium (for the benefit of others reading this) because it's important to prep alu butt welds with the optimum gap. I'm guessing this is why alu radius chine amateur builds tend to use shorter, full width pieces, rather than longer, half-width pieces, as (if I understand correctly) you are doing.
Steel OTOH is extremely forgiving when it comes to filling large gaps: the trouble amateurs most often seemed to encounter (certainly in the 'bad old days' before the www) was not providing enough
gap, so that when they ground off the weld, there was nothing left holding the edges together, and they'd get that lovely 'ping' sound, and slivers of daylight. Luckily, in these days of flexible, 1mm thin cutting disks, that situation is easily recovered from.
If you find you didn't provide enough gap towards the ends, just slit the weld you've just laid back a bit with the thin disc , then use it in both directions al gusto
; this will let you have another crack without having to start all over again. And it will enable you to err on the conservative side in how much gap increase to allow towards the ends, at least until you find out if the method will work
Ideally I would also recommend staggering the ends of these half-width radius chine sections, but that would make the advice above harder to follow. Welds in steel are so strong that it's probably a counsel of perfection: better on balance to do it in such a way that the welds are not excessively tortured, whatever that takes.
(ON EDIT) Apologies if this is what you were already doing, and I'm probably just stating the bleeding obvious, but someone else might benefit from considering this proposal...