Evaluation and Design of a Flywheel Energy Storage System

This thesis deals with the energetic evaluation and design of a flywheel energy storage system (FESS). The first purpose is to give a quantitative evaluation of the energetic performance of the systems equipped with flywheels. Two systems are chosen: one 5 kWp (kWp: peak power) household PV system equipped with a 3 kWh flywheel to accumulate the excessive energy generation in daytime, and one tramway power system equipped with a 1.5 kWh onboard flywheel to accumulate the recovered braking energy. The energy saving potential of each system is analyzed based on the given profiles and the modeled FESS, in which various losses are taken into account. The results show that, energy savings can be achieved for both systems by using flywheels: 15.1 % for the PV system and 20.9 % for the tramway system. But the overcall energy efficiency of the flywheel in the PV system is only 40 %, much lower than 75.5 % in the tram. The main reason is that the high self-discharge due to internal losses (7.7 % of the maximum stored energy per hour) causes considerable energy loss for the long term idling operating cycle (for hours) of the flywheel in PV system. As a comparison, in the tramway system, charge/discharge cycle is much shorter (approx. 1 min) so that the self-discharge is less critical, leading to a higher efficiency of the flywheel. The second part of this work is design and prototyping of a flywheel demonstrator in order to verify the energetic evaluation and the implementations of high-speed technologies. The demonstrator has an energy capacity of 0.5 kWh at the maximum operating speed of 24000 min-1 and the power rating is 35 kVA. This thesis introduces the design methodology of the key components. For the flywheel rotor, a constant thickness rotor body with the inertia of 0.57 kg⋅m2 is designed. The mechanical issues regarding the stress caused by the centrifugal force are analyzed. The construction issues, such as balancing solutions and rotor hardening process are discussed. As an energy conversion component, a 4-pole PMSM is designed. Low loss in the rotor is required due to the inefficient rotor cooling in vacuum. Therefore, the surface mounted magnets are segmented so that the losses can be reduced to 28.4 W (approx. 0.1 % of the rated power). The calculated power efficiencies at two defined operating points are both above 96 %. Based on the designed rotor and PMSM, magnetic bearings are selected and the housing is designed. The components are assembled and a complete system is built up, which is validated both by a 3D CAD program and by prototyping. The main issues concerning the component processing and the assemble work are presented. Due to the safety consideration, two outer housings are designed as vacuum and also burst containments in case of the rotor structural failure. The loss analysis is carried out for the PMSM, magnetic bearings, and also the rotor due to air friction. A lumped parameter network of the system is built up for the thermal analysis. The rotor is painted black in order to improve the radiative heat dissipation. The calculated temperature rise on the black-painted rotor and stator is approx. 87.4 K for continuous operation and 73.0 K for operation with the fully-utilized duty cycle. Compared to the temperature limit of 141 °C for the E-machine carbon fiber bandage, the system can operate with the designed duty cycle with a 28 K thermal margin and is sustainable for continuous operation. The flywheel demonstrator design is validated by FEM calculation and the prototype construction so far. Relevant testing of the prototype has been performed in order to verify the performance, including spin testing of the rotor and levitation testing. In the end of the thesis, a conceptual design of high power (150 kW) machine is presented, as an outlook for the application of the flywheel in the railway systems. Additionally, the design criterions of light weight rotor constructed with composite materials, aiming to achieve higher energy density, are presented. The critical considerations are pointed out, as an outlook for further structural optimization in the future