Modeling, Analysis and Control of Switched Reluctance Generator

Abstract

Switched reluctance machines (SRM) are permanent magnet-free and have a simple rotor with no current-carrying conductors. SRM can be operated as a generator and driven by renewable energy resources such as wind, solar-thermal, and ocean tide. A switched reluctance generator (SRG) is also suitable for operating at high speeds with direct coupling to the high-speed turbine (without any gearbox). This ultimately reduces the space requirement and improves the efficiency of the overall generation system. However, modeling and controlling the SRG system is challenging because phase inductance and induced EMF of SRG depend on phase current and rotor position. This thesis focuses on the analysis, modeling, and control of an SRG system. This thesis proposes a novel flux-tube-based approach to derive closed-form analytical expressions for the effective air-region permeance at the aligned and unaligned rotor positions in terms of machine dimensions. Unlike existing approaches, the proposed method does not require preliminary FEA and includes the effect of rotor pole height on the effective air-region permeance, which is particularly useful in the design of high-speed SRG. The predictions by the proposed method are compared with the results of FEA and three existing analytical methods, considering seven different SRM designs and different sets of machine dimensions. The results of the proposed method are found to be accurate and comparable to those of FEA in all cases. Further, based on magnetic material characteristics and machine dimensions, the static flux-linkage characteristics at the aligned and unaligned rotor positions are predicted. These analytical characteristics are then used to develop a design tool. Generally, for power generation, SRG is operated close to or above its base speed. At these speeds, single-pulse mode (SPM) and modified single-pulse mode (MSPM) are the obvious choices. The optimal choice of control angles for a given average DC-bus current in any of these operating modes is challenging. This thesis develops simplified simulation models and optimization algorithms to choose optimal control angles for both of these operating modes. The optimal control angles are validated on a 4-phase, 8/6-pole, 4-kW, 1500-rpm SRG. With optimal control angles, a very significant reduction in RMS phase current, RMS capacitor current, and power loss in SRG drive over the entire output power range is found. The thesis also proposes a large-signal model where average DC-bus current, mean-square DC-bus current, and mean-square phase current are related to control angles, DC-bus voltage, and prime-mover speed of SRG. A small-signal model is also reported where the rate of change of average DC-bus current, mean-square DC-bus current, and mean-square phase current with control angles are related to the control angles, DC-bus voltage, and prime-mover speed of SRG. These models are developed for both SPM and MSPM. All these models are validated through simulations and experiments over a wide operating range of the SRG. Real-time optimal predictive average DC-bus current controllers for SPM and MSPM are also developed in this thesis. For SPM, the proposed controllers can optimize the control angles in real time to minimize the mean-square DC-bus current while tracking the commanded average DC-bus current reference. For MSPM, the proposed controllers can optimize the control angles in real time to minimize both mean-square DC-bus current and mean-square phase current while tracking the commanded average DC-bus current reference. The most important advantage of the proposed optimal controllers is that they do not require prior measurement of flux-linkage characteristics, extensive simulation of the SRG system, and offline optimization. The proposed optimal controllers can also operate without a rotor position sensor. However, speed sensing is still required. This thesis also proposes a predictive average DC-bus voltage controller (i.e., an outer loop) to regulate the average DC-bus voltage as desired. The discrete-time model of the average DC-bus current controlled SRG feeding a load (i.e., an inner loop) is determined. Then, the discrete-time model is used to calculate the average DC-bus current required to track the desired average DC-bus voltage. The proposed controller is validated at multiple values of reference voltage, speed, and load resistance using a 4-phase, 8/6-pole, 4-kW, 1500 rpm SRG. The performance of the proposed voltage controller is also compared with that of a PI voltage controller. The experimental results confirm the superior performance of the proposed voltage controller over a PI voltage controller.