Modulating Waveform Generation for Split-Phase Induction Motor Drives and Discrete Modelling of Switched Reluctance Motor Drives

Abstract

The thesis focuses on the areas for improvement in non-permanent magnet-based electric drives, such as the asymmetrical six-phase (or split-phase) induction motor and the switched reluctance motor drives.

In the first work, a modulating waveform generation for carrier-based space vector pulse width modulation (PWM) control for a split-phase induction motor drive is proposed for the first time. The modulating waveform synthesis for the carrier-based dodecagonal space vector PWM (DSV-PWM) is based only on the sampled reference phase amplitude. The present approach for the DSV-PWM generation is based on vector time computation and accessing the output vectors from the look-up table. The proposed work synthesizes a modulating waveform compared to a triangular carrier waveform to generate the DSV-PWM for the 2-level conventional split-phase induction motor inverters. Like conventional hexagonal space vector modulating waveform, here, dodecagonal space vector modulating waveform is generated by adding the zero sequence signals of triplen, 6n±1; (n=1,3,5,...) order harmonics to the fundamental or modulating signal in two stages. A simple approach, like in a hexagonal space vector scheme, is proposed with less mathematical computation, such that it can be easily implemented in the digital platform. The proposed scheme is explained and verified with experimental results.

The second work proposes a modulating and carrier-based dodecagonal space vector pulse width modulation (DSV-PWM) technique for split-phase open-end induction motor (OEIM). The advantage of the DSV-PWM technique for the split-phase OEIM is that the fifth and seventh order 6n±1; (n=1,3,5,...) harmonics are eliminated from its phase voltages. The modulating waveform is generated from the sinusoidal reference signals by adding the triplen, and 6n±1; (n=1,3,5,...) order harmonics in a single stage. This modified reference waveform is compared to the triangular carrier waveform to generate the PWM gating signals for inverter-1 so that the vector addition from two inverters results in a 12-sided polygonal space vector structure (SVS) for the PWM control. The PWM signals for the 2-level inverter-2 connected to the other end of the OEIM are produced by performing a small logical operation using the PWM signals of inverter-1. The DC voltage relation between inverter-1 and inverter-2 is 1:0.366. The timing duration for the inverter-1 and inverter-2 is the same to generate the dodecagonal space vector structure (DSVS). It can be easily implemented, as in the case of a conventional sine-triangle PWM technique in a 3-phase drive. So, mapping the space vectors of the DSVS of conventional split-phase to that of the OEIM drive is simple. This can be achieved using a logical operator like XOR. Thus, the computation burden of the digital signal processors to produce the DSV-PWM signals is reduced with the proposed algorithm, which is 3.45 times the conventional space vector-based implementation using sector identification, vector time computation, and storing the switching vectors. A detailed analysis of the modulating waveform and DSV-PWM signal generation with the experimental verification of the proposed drive is presented in the second work.

The third work proposes a new method to produce dodecagonal space vector pulse width modulation (DSV-PWM) signals for a split-phase open-end induction motor (OEIM) drive comprising 3-level neutral point clamped (NPC) inverters. The phase voltages of the split-phase OEIM drive cancel the 6n±1; (n=1,3,5,...) order harmonic components produced from the inverters connected at both ends of the split-phase OEIM using the DSV-PWM technique. This work discusses the generation of modulating signals with minimal trigonometric operations. It compares them to the level-shifted triangular carrier signals to produce DSV-PWM signals for the 3-level inverters (inverter-1A and 1B) with DC link voltage of VDC, connected to one end of the OEIM. The generation of DSV-PWM signals for the 3-level inverters (inverter-2A and 2B) with DC voltage of 0.366VDC, connected to the other end of the split-phase OEIM, is done from the logical operation of DSV-PWM signals of inverter-1A and 1B. The approach of minimizing the trigonometric operations while generating the modulating signals and adopting the logical method of generating the DSVPWM signals for 3-level inverter-2A and 2B reduces the overall processing time and is 4.64 times faster than the conventional method of producing the DSV-PWM signals through sector identification, vector timing computations, and storing the timing values. This is experimentally verified, and a detailed analysis is presented in this work.

The fourth, fifth, and sixth works focus on the dynamic modeling of switched reluctance motors (SRM) in a script-based simulation tool like MATLAB/Octave/Spreadsheet (Excel sheet) and finite element method (FEM) based simulation tools like FEMM and JMAG. The dynamic modeling of SRM becomes relevant due to the lack of literature on mathematical modeling covering electrical, magnetic, and mechanical domain equations, to the best of the author's knowledge. The dynamic model of SRM helps in state/parameter estimation studies, digital twin studies, etc.

The fourth work discusses the tuning of PI controllers by deriving the optimized plant transfer function of non-linear plant model SRM and integrating non-linear techniques like an anti-windup scheme for an improved and stable response. The mathematical modeling of 8/6 pole SRM is presented. The implementation of 8/6 pole SRM has been carried out using the look tables of torque vs phase currents, rotor angle, phase inductance vs phase currents, and rotor angle in MATLAB. The look-table data is derived with Finite Element Analysis (FEA) tools like FEMM and JMAG. The detailed steady state and transient response of 8/6 pole SRM with a modified PI controller with anti-windup logic are presented.

The fifth work discusses the discrete modeling of SRM, which can be used for state and parameter estimation studies. It also helps to emulate the discrete models of SRM in digital platforms and study linear and non-linear characteristics of the SRM. Hence, the control techniques can be tested on this discrete model with multiple study options. This work uses state derivative models and the magnetic characteristics derived in the finite element method magnetics tool to realize the SRM.

The sixth work proposes an electromagnet rotor-based switched reluctance machine (ERSRM) for electric vehicle (EV) applications. The proposed machine is realized with the winding on the rotor. Because of these rotor winding, higher torque density (Nm/kg), reduced input current, and wider torque-speed characteristics are achieved. This is done by establishing rotor excitation through inductive power transfer from stator to rotor. In the first approach of rotor excitation through a wireless power transfer (WPT) system, the stator has a transmitting coil fed by a resonant converter, and the rotor has a receiving coil that has a high pass filter to allow high-frequency signals necessary for rotor excitation. A 50kW ERSRM has been modeled and validated in FEM (Finite Element Method) simulation tools like FEMM and JMAG.