The vertical and horizontal forces and associated stiffnesses on a permanent magnet (PM) above a high-temperature superconductor (HTS) were measured during vertical and horizontal traverses in zero-field cooling (ZFC) and in field cooling (FC). In ZFC, the vertical stiffness was greater in the first descent than in the first ascent and second descent, and the stiffness in the second descent was between those of the first descent and the first ascent. At the FC position, the vertical stiffness was two times greater than the lateral stiffness at each height, to within 1% of the vertical stiffness value. The cross stiffness of vertical force with respect to lateral position was positive for FC, but negative for ZFC. Free-spin-down experiments of a PM levitated above a HTS were also performed. These results showed that the coefficient of friction is double valued at frequencies just below the rotor resonance, a result attributed to cross stiffness in the PM/HTS interaction. A frozen-image model was used to calculate the vertical and horizontal forces and stiffnesses, and reasonable agreement with the data occurred for vertical or horizontal movements of the PM less than several mm from the FC position.

The rotational dynamics of a disc-shaped permanent magnet rotor levitated over a high temperature superconductor was studied experimentally and theoretically. The interaction between the rotor magnet and the superconductor was modelled by assuming the magnet to be a magnetic dipole and the superconductor a diamagnet. In the magnetomechanical analysis of the superconductor part, the frozen image concept was combined with the diamagnetic image, and the damping in the system was neglected. The interaction potential of the system is the combination of magnetic and gravitational potentials. From the dynamical analysis the equations of motion of the permanent magnet were stated as a function of lateral, vertical, tilt, precision and rotating angles. The vibration behaviour and correlation of the vibration of one direction with that of another were determined with a numerical calculation based on the Runge-Kutta method. The various vibrational frequencies identified were vertical, radial, tilt, precession and rotation. The tests performed for experimental verifications were translational and rotational. The permanent magnet was 'spun up' under vacuum conditions to analyse the dynamics of the free 'spin down' behaviour of the permanent magnet.

No abstract

We have constructed a bearing system for an energy storage flywheel. This bearing system uses a combination of permanent magnets and superconductors in an arrangement commonly termed as an Evershed bearing. In an Evershed system there are in fact two bearings which act in concert. In our system we have one bearing constructed entirely out of permanent magnets acting in attraction. This system bears the weight of the flywheel (43.6 kg) but would not, on its own, be stable. Stability is provided by a superconducting bearing which is formed by the interaction between the magnetic field of a permanent magnet sited on the rotor and superconductors on the stator.This overall arrangement is stable over a range of levitation heights and has been tested at rotation speeds of up to around 12 Hz (the maximum speed is dictated by the drive system not the bearing system). There is a sharp resonance peaking at between 2 and 3 Hz and spin down tests indicate that the equivalent coefficient of friction is of the order of 10−5. The rate of change of velocity is, however, not constant so the drag is clearly not solely frictional.The position of the resonance is dictated by the stiffness of the bearing relative to the mass of the flywheel but the amplitude of the resonance is dictated by the variation in magnitude of the magnetic field of the permanent magnets. Large magnets are (at present) fabricated in sections and this leads to a highly inhomogeneous field. The field has been smoothed by using a combination of iron which acts passively and copper which provides magnetic shielding due to the generation of eddy currents and therefore acts as an ‘active’ component. Calculations based on the spin down tests indicate that the resultant variation in field is of the order of 3% and measurements are being carried out to confirm this.

No abstract