Phase Current Reconstruction and Peak Prediction for Switched Reluctance Generators

Abstract

The threat of rising temperatures and increasing sea levels due to the high carbon emissions and greenhouse gases, has prompted a worldwide push towards the production of clean energy through renewable sources. Renewable technologies gained significant traction, especially after the 2015 United Nation’s Paris Agreement, which united countries to preserve the planet for a better future. Electric machines play an important role in the renewable energy sector, as they are widely used in wind turbines and electric and hybrid electric vehicles. However, the commonly used machines use a large amount of permanent magnets, whose prices have been increasing due to the limited supplies of rare earth elements used in their production and also because of their increasing demand. A switched reluctance machine (SRM) is a suitable candidate in the renewable energy sector as it does not use permanent magnets and is extremely versatile and robust. Due to the lack of permanent magnets in a switched reluctance generator (SRG), the machine suffers from low power densities compared to its competition. The research presented in this dissertation pushes the envelope of an SRG’s produced charge, making it more competitive in the renewable energy sector. Maximizing the output charge of an SRG involves its operation at high speed in single pulse mode, wherein the motional back EMF is allowed to build up and reach a substantial value. This results in an SRG’s phase currents entering into a state of positive feedback, wherein even switching off the phase does not bring down their value. Operating the machine in this scenario results in higher output charge; however, it makes the phase currents uncontrollable. The phase currents peak and begin to fall only after the motional back EMF reduces, which occurs as the rotor approaches its unaligned position. Protection of the drive circuit is imperative; however, if the unknown peak values of the phase current exceed the current ratings of the diodes, the drive circuit will be damaged. As a result, either an SRG is not operated in the positive feedback mode (thereby losing out on the additional charge produced) or the drive’s power converter is over engineered for a high current rating, in order to sustain the unknown current levels. Since the system is extremely nonlinear, it poses significant modeling and control challenges in order to safely operate an SRG in single pulse positive feedback. Due to the time delay associated with numerical methods of integration, an iterative approach to predict the phase current is impractical. The presented research reconstructs the phase current of an SRG and predicts its current peak by detecting an optimal turn-off angle, which leads to a more controllable machine with reduced drive constraints and higher output charge, all while maintaining the same size. The research also analyzes the effect of a freewheeling phase in the high speed mode of operation of an SRG.