Motivation
The goal of this project was to design a reluctance motor for use in high-speed flywheel energy applications. As renewable energy market penetration is only increasing, so will energy storage technologies play a larger role. Flywheel energy storage is a technology with a number of advantages, including high cycle counts (potentially up to the order of millions of cycles), and very high power density. While the main disadvantages are high-cost, energy density, and short storage time, the latter two are not particularly important in a number of applications, and the cost for such systems are decreasing as the technology improves. For this project I will investigate a design that has potential for very high-efficiency and low cost.
Why a Reluctance Motor?
A reluctance motor has some very key advantages, particularly regarding the construction of the rotor. Unlike an induction rotor, which requires a cast, bar, or wound squirrel cage, or a permanent magnet BLDC motor, which requires bonding magnets to the rotor, a reluctance motor requires a simple steel-only rotor. Reluctance motor rotor geometries vary from simple salient pole designs, to more complex flux barrier designs used in synchronous machines. For this design, I wanted to explore the possibility of a solid rotor (without flux barriers) in a synchronous configuration. The reason behind this decision is that flux-barrier rotors cannot be used in high-speed applications, as the flux-barrier cutouts cause high stresses from centripetal acceleration. A solid rotor, however, can operate at very high speeds. One the the disadvantages of the solid rotor is that it has a low power-factor compared to the flux-barrier designs. For this project, I studied the effects of rotor geometry on both efficiency and power-factor.
Reluctance Motor Basics
- Change in angle of rotor changes the Reluctance of the flux path
- Can relate reluctance to inductance (L = N^2/R), and thus torque using basic conservation of power:
- P = IV = τω
=> τ (dϴ/dt) = I^2 dL/dt) <- Assuming I is constant:
=> τ (dϴ/dt) = I^2 (dL/dϴ) (dϴ/dt)
=> τ = I^2 dL/dϴ
- Equation doesn’t really hold due to B-H nonlinearities, but it shows the basic idea.