SRM motors - possible overunity, the principle

Hello,
I was experimenting a bit out of boredom and came up with this experiment:
I disassembled an old EI transformer (220V/12V). Using the E-shaped parts and the 220V winding, I made an electromagnet. I connected a starting electrolytic capacitor to it to create an LC circuit with a resonant frequency of about 10 Hz. I pulsed the LC circuit using PWM at its resonant frequency. From the I-shaped parts of the transformer, I made a block, which I brought close to the pulsing electromagnet. The I-block oscillated strongly in my hand, and when I moved it closer to the electromagnet, it stuck to it and wouldn’t come off until I turned off the pulsing (8V, 250mA, duty cycle 45%).
Using a multimeter and oscilloscope, I found that the LC circuit consumed roughly the same amount of power, whether the oscillating I-block was near the electromagnet (oscillating strongly) or stuck to it completely, or even when it was absent—any difference in power consumption was negligible, despite me applying a significant braking force with my hand.
Now, here’s my question:
Imagine we have a switched reluctance motor (SRM) where the rotation of the rotor does not influence the magnetic field of the excitation coils with an opposing magnetic field (unlike standard motors). Instead, it operates on the same principle as my vibrating I-block. If we connected the phase coils of the SRM motor into LC circuits and pulsed them at the resonant frequency of the LC circuits, what efficiency could such an SRM motor achieve?
Does anyone have experience with SRM motors?
 
I want to thoroughly test the energy efficiency of SRM motors.
I understand the physical principle of generating mechanical energy (the motion of a conductor in a magnetic field when current flows through it, or the motion of a magnet in a magnetic field induced by the conductor). By moving the conductor’s loop, its magnetic field affects the magnetic field of the other element, thereby causing increased electrical energy consumption.
But in the case of SRM motors, the situation is different. There is only one electromagnet, and the rotor is magnetically neutral. Therefore, the motion of the rotor does not influence the magnetic field of the excitation coil.
I have verified this through an experiment with an electromagnet in an LC circuit, where I allowed a metallic object to vibrate. The energy consumption without the vibrating object was the same as the energy consumption with the vibrating object. In both cases, the consumption was only due to Joule losses, and the mechanical energy produced appeared to be "extra."
That’s why I want to try connecting a two-phase stepper SRM motor to an LC circuit, where one part of the oscillation would pass rectified through one coil and the other part through the second coil (with an initial mechanical impulse provided by hand to start rotation). I want to monitor the energy consumption of the supplied pulses in the state without the rotor, with the rotor stationary, and during rotor rotation. Then I plan to connect a generator (e.g., a cheap BLDC motor), rectify the output, and measure the generated energy.
Equations describing the motion of a conductor in a magnetic field talk about the force acting on the conductor in the field but say nothing about a magnetically neutral ferromagnetic object affected by the induced magnetic field.
The result of my experiment with the electromagnet in an LC circuit suggests that if we were to use a superconductor and a lossless magnetic core in the LC circuit, it would oscillate indefinitely, and by placing an object made of lossless magnetic material near the electromagnet in the LC circuit, it would generate mechanical energy seemingly "out of nothing."
AI claims that SRM motors can operate with an efficiency of 95%—but that calculation considers only mechanical efficiency. Joule losses are not accounted for. I want to verify whether SRM motors could operate analogously to a heat pump, generating mechanical energy from something we currently don’t understand.