This makes me think about another problem that is sort of interesting. A child on a swing. How does the child get the swing going higher and higher when there is nobody pushing him?
You are on the right track, I see it as torque = (moment of inertia) x angular acceleration.
S is force, r is radius, 1/2 m rsq is moment of inertia, and a is linear acceleration, a/r is angular acceleration.