Why don't you stick to the energy growth rate (instead of the 4.72% velocity growth rate) and then use the relativistic formula of the kinetic energy (in which the relationship between energy and velocity is not quadratic — the quadratic approximation is valid only for small velocities)?