Difference between potential centrifugal energy and rotational/kinetic energy

It is not possible to conserve the energy in the inertial frame of reference, for someone must be adding (or extracting) energy to keep the object rotating at the same angular velocity. In fact, if you try to use conservation of energy in the inertial frame to solve the problem, you won't even get a solution.

The energy in the non-inertial rotating reference frame, on the other hand, is conserved; and it should lead you to the correct answer. Are you sure that your calculations are correct?

When the problem is solved in the rotating frame, you can use the conservation of the energy, as the work done on the mass is just due to the spring, the gravity and the centrifugal apparent force. These works are captured in the corresponding potentials.

In an inertial frame the energy of the system is no more conserved, because in order to maintain the rod in rotation with constant angular frequency work must be done on the system.

If you take this into account, you will get the same result in both frames, as I will show explicitly.

In the rotating frame, equating the initial energy to the energy in the maximum elongation position you get

$$ U_{grav}[(ell-delta)cosalpha]+U_{spring}(delta) +U_{centr}[(ell-delta)sinalpha] = U_{grav}[(ell+x)cosalpha]+U_{spring}(x)+U_{centr}[(ell+x)sinalpha] $$ where $$ U_{grav}(h)=mgh $$ is the gravitational potential $$ U_{spring}(delta)=frac{1}{2}kdelta^2 $$ is the spring potential and $$ U_{centr}(r) = -frac{1}{2}momega^2 r^2 $$ is the centrifugal potential.

Explicitly we get a second degree equation for $x$ which can be easily solved.

In the inertial frame, the energy is given by $$ E=frac{1}{2} m left[left(frac{dx}{dt} right)^2 +omega^2 (ell+x)^2 sin^2 alpha right] + U_{spring}(x) + U_{grav}[(ell+x)cosalpha] = frac{1}{2} m left(frac{dx}{dt} right)^2 -U_{centr}[(ell+x)sinalpha] + U_{spring}(x) + U_{grav}[(ell+x)cosalpha] $$ where $x$ is the elongation of the spring. In the initial and final configuration the velocity of the mass along the rod is zero. But to obtain the final energy we must add the work done on the system to the initial energy. The work can be written as $$ {cal L} = int vec{M}cdotvec{omega} dt $$ where $vec{M}$ is the torque, which is also equal to the derivative of the angular momentum. So $$ {cal L} = int frac{dvec{L}}{dt} cdotvec{omega} dt = Delta vec{L}cdotvec{omega} $$ because $omega$ is constant. Now we can write $$ E_{final} = E_{initial}+Delta vec{L}cdotvec{omega} = E_{initial}+vec{L}_{final}cdotvec{omega}-vec{L}_{initial}cdotvec{omega} $$ which means that $$ E -vec{L}cdotvec{omega} $$ is conserved. But $$ vec{L}cdotvec{omega} = m(ell+x)^2 omega^2 sin^2alpha = -2 U_{centr}[(ell+x)sinalpha] $$ so $E-vec{L}cdotvec{omega}$ is just the energy in the rotating frame. We obviously obtain the same result.