Investigation of diamagnetic bearings and electrical machine materials for flywheel energy storage applications

Abstract

Recent trends in energy production have led to a renewed interest in improving grid level energy storage solutions. Flywheel energy storage is an attractive option for grid level storage, however, it suffers from high parasitic loss. This study investigates the extent to which passive diamagnetic bearings, a form of electromagnetic bearing, can help reduce this parasitic loss. Such bearings require three main components: a weight compensation mechanism (lifter-floater), a stabilizing mechanism and an electrical machine. This study makes use of a new radial modification of an existing linear multi-plattered diamagnetic bearing. Here a prototype is built and analytical expressions derived for each of the three main components. These expressions provide a method of estimating displacements, fields, forces, energy and stiffness in the radial diamagnetic bearing. The built prototype solution is found to lift a 30 [g] mass using six diamagnetic platters for stabilization (between ring magnets) with a disc lifter and spherical floater for weight compensation. The relationship between mass and number of platters was found to be linear, suggesting that, up to a point, increases in mass are likely possible and indicating that significant potential exists for these bearings where high stiffness is not needed – for instance in flywheel energy storage. The study examines methods of reducing bearing (parasitic) losses and demonstrates that losses occur in three main forms during idling: air-friction losses, electrical machine losses, stabilizing machine losses. Low speed (158 [rpm]) air-friction losses are found to be the dominant loss at 0.1 [W/m3]. The focus of this study, however, is on loss contributions resulting from the bearing’s electrical machine and stabilizing machine. Stabilizing machine losses are found to be very low at: 1 × 10−6 [W/m3] – this leaves electrical machine losses as the dominant loss. Such electrical machine losses are analysed and divided into eddy current loss and hysteresis loss. Two components of hysteresis loss are remanent field related cogging loss and remagetization loss. Eddy current losses in silicon steel laminations in an electrical machine are quite high, especially at high speeds, with losses in the order of 1 × 105 [W/m3]. Noting the further high cost of producing single unit quantities of custom lamination-based electrical machine prototypes, this high loss prompts a look at potentially lower cost ferrite materials for building these machines. A commercial sample of soft magnetite ferrite is shown to have equivalent eddy current losses of roughly 1 × 10−13 [W/m3]. The study notes that micro-structured magnetite has significant hysteresis loss. Such loss is in the order of 1 × 10−3 [W/m3] when referring to both remanence related cogging and remagnetization. This study, thus, extends its examination of loss to nano-structured magnetite. Magnetite nano-particles have shown superparamagnetic (no hysteresis) behaviour that promises the elimination of hysteresis losses. A co-precipitation route to the synthesis of these nano-particles is examined. A detailed examination involving a series of 31 experiments is shown to demonstrate only two pathways providing close-to-superparamagnetic behaviour. After characterization by Scanning Electron Microscope (SEM), X-Ray Diffractometer (XRD), Superconducting Quantum Interference Device (SQUID) and crude colorimetry, the lowest coercivity and remanence found in any given sample falls at −0.17 [Oe] (below error) and 0.00165 [emu/g] respectively. These critical points can be used to estimate hysteresis related power loss, however, to produce bulk ferrite a method of sintering or bonding synthesized powder is needed. A microwave sintering solution promises to preserve nano-structure when taking synthesized powders to bulk material. A set of proof-of-concept experiments provide the ground work for proposing a future microwave sintering approach to such bulk material production. The study uses critical points measured by way of SEM, XRD, SQUID characterization (e.g. remanence and coercivity) to implement a modified Jiles-Atherton model for hysteresis curve fitting. The critical points and curve fitting model allow estimation of power loss resulting from remanent related cogging and remagnetization effects in nano-structured magnetite. Such nano-structured magnetite is shown to exhibit hysteresis losses in the order of 1 × 10−4 [W/m3] from remagnetization and 1 × 10−7[W/m3] from remanence related cogging drag. These losses are lower than those of micro-structured samples, suggesting that nano-structured materials have a significant positive effect in reducing electrical machine losses for the proposed radial multi-plattered diamagnetic bearing solution. The lower parasitic loss in these bearings suggests excellent compatibility with flywheel energy storage applications.