So, I have a new problem with your model, that has to do with a double sided Earth having any thickness; you posit that what we recognize as hemispheres are the two sides of the flat earth, with the equator on the rim. Here's the problem: that means there must be a ton of land on the rim, and thus a large area both on and off the rim where you would expect to see the earth curve (relatively) sharply downward, with no horizon. You estimate the earth to have a thickness of about 1/4 the flat earth's circumference. Turns out, I didn't need that estimate, because no estimate works. If the Earth is thin, there isn't be enough iron in the earth's core to create the magnetic field, plus the view on the rim would be incredibly peculiar, with ground sharply dropping off before the horizon, even on the equator. If the earth is thick, then much more land is on the rim.
That's a problem for one big reasons: measurements. I know that this chapter of FE doesn't trust all the measurements we get for the world, especially at sea, but this would take it to a whole new level. The thinner the rim is, the less land is on it, so it works. But as previously mentioned, this would lead to perspective issues. But if it's thicker, than the circumference of the earth shrinks, because the contiguous landmasses over the equator mean that we can't fudge it: there's less surface area on the disk part. If you DID fudge it, and you would have to fudge it pretty heavily for this to work, than measurements on land, which can be verified much more easily, would be so incredibly far off it would be impossible to keep under wraps. So, the circumference logically must shrink.
But if the circumference shrinks, that brings up all kinds of new and exciting issues with the landmasses near the equator. See, the circumference of the cylinder would be in line with measurements for things near the edge of the rim. As you get farther though, things start shrinking fast. Think about it; the landmasses near the equator would have to be much smaller, because they now have less latitude to work with than in both the globe model and most FE models. Once again, though measurements can be fudged, this would take a LOT of fudging.
And isn't the whole reason FE doesn't have one conclusive model because none of them line up with measurements?
Oh, and the good ol' Coriolis effect has come back once more! See, this effect just wouldn't work on the rim. like, at all.