This is one of the best gravity wheel ideas I've seen so far.
(many stupid exist)
I played with your model a little. The acceleration and speed graphs look not bad.
But, I have a bad thing to say...
I think it's impossible to balance the spring constant with the weight mass and the movement of the wheel, because It's changing speed, (and also the weight doesn't have a constatnt track) and these properties react to each other continuously, when the wheel turn.
That is the problem. It need some external devices (and / or constant movement track for the weight) to keep the things synchronized. And do this with the less energy, and best method as it possible.
I have also simulated some stupid devices in wm2d, similar to your model (but yours is much better), which makes more than one revolution with friction included, but they aren't work.
I found when a Wheel makes more than one, or two, or lets say 6 revolution... It still doesn't mean it works.
My stupid devices had balance points appeared and disappeared chaotically, but finally, the device reach them, and stopped.
Also, I think springs weren't essential parts of Bessler's wheels. Probably his first two wheels didn't contains springs. The eyewitnesses didn't mention it, just scraching and loud noise. They only mention at the two bigger wheels. I think Bessler used the springs as a part of a mechanical device or mechanism to control the movement of the weights, or to hold them in a position, but not to store the potentional energy in them, I guess. If the weight can compress the spring, the spring can't push enough on the weight to do something useful, and vice versa. (and the machine must starts by itself to be any use)
So, I think the springs can be used with only some mechanical force produced by the wheel itself, and this is not the best solution. But, I also have a weak idea of a spring type wheel. If somebody interested in it, I will post certainly after some modification. (it's a new-born now)
Gregory