Kinetic Energy Storage and Magnetic Bearings : for Vehicular Applications

One of the main challenges in order to make electric cars competitive with gas-powered cars is in the improvement of the electric power system. Although many of the energy sources currently used in electric vehicles have sufficientlyhigh specific energy, their applicability is limited due to low specific power. It would therefore be advantageous to create a driveline with the main energy storage separated from a smaller energy buffer, designed to have high power capabilities and to withstand frequent and deep discharge cycles. It has been found that rotating kinetic energy storage in flywheels is very well suited for this type of application. A composite shell, comprising an inner part made of glassfiber and an outer part made of carbonfiber, was analyzed analytically and numerically, designed, and constructed. The shell was fitted onto a metallic rotor using shrinkfitting. The cost of the shell, and the complexity of assembly, was reduced by winding the glass- and carbonfiber consecutively on a mandrel, and curing the complete assembly simultaneously. Thereby, the shell obtained an internal segmentation, without the need for fitting several concentric parts onto each other. The radial stress inside the composite shell was kept compressive thanks to a novel approach of using the permanent magnets of the integrated electric machine to provide radial mechanical load during rotation. Two thrust bearing units (one upper and one lower) comprising one segmented unit with the permanent magnets in a cylindrical Halbach configuration and one non-segmented unit in a up/down configuration were optimized, constructed and tested. Each thrust bearing unit generated 1040 N of repelling force, and a positive axial stiffness of 169 N/mm at the nominal airgap of 5 mm. Two radial active magnetic bearings (one upper and one lower) were optimized, constructed and tested. By parameterizing the shape of the actuators, a numerical optimization of force over resistive loss from the bias cu