Advanced control of induction motors

The current industrial standard for the control of the induction motor is the so called vector control (VC) or field-orientated control (FOC) which transforms and controls the induction motor as a direct current (DC) motor. Besides its many advantages, such as fast and decoupled dynamics of speed and flux, it is well known that VC depends on the detailed system model and is very sensitive to parameter uncertainties and external disturbance (load torque). To clarify further the VC is a only a partial feedback linearising control which can achieve the decoupling of speed and flux asymptotically. The coupling still exists when flux is not kept in constant, i.e. when flux is weakened in order to operate the motor at a higher speed and keep the input voltage within saturation limits, or when flux is adjusted to maximize power efficiency of the motor with light load. The thesis will summarise research of advanced control approaches of induction motors in Chapter One. The Chapter Two starts on building a fifth-order nonlinear dynamic model of an induction motor and then recalls the principal of traditional VC of induction motors. The differential-geometric technique based nonlinear control has developed for induction motors, which can convert some intractable nonlinear problems into simpler problems by familiar linear system methods. The partial decoupled dynamic of the conventional VC has been investigated via feedback linearisation control (FLC) at first. Then input-output linearisation control is applied to design a fully decoupled control of the dynamics of speed and flux. To remove the weak robustness and the requirement of an accurate model of the VC and FLC, a novel nonlinear adaptive control of induction motor is designed based on feedback linearisation control and perturbation estimation. The induction motor will be represented as a two coupled interconnected subsystems: rotor speed subsystem and rotor flux subsystem, respectively. System perturbation terms are defined to include the lumped term of system nonlinearities, uncertainties, and interactions between subsystems and are represented as a fictitious state in the state equations. Then perturbations are estimated by designing perturbation observers and the estimated perturbations are employed to cancel the real system perturbations, assumed all internal states are measured. The designed nonlinear adaptive control doesn’t require the accurate model of the induction motor and has a simpler algorithm. It can fully decouple the regulation of rotor speed and rotor flux and handle time-varying uncertainties. The parameter estimations based on nonlinear adaptive controls can only deal with unknown constant parameters and are not suitable for handling fast time-varying and functional uncertainties. Nonlinear adaptive control based on output measurements is addressed in Chapter Five, assuming that the rotor speed and the stator volatge/currents are measurable. A sliding mode rotor flux observer has been designed based on the stator voltage and current. Moreover, two third-order state and perturbation observers are designed to estimate the unmeasured states and perturbation, based on the rotor speed and the estimated rotor flux. Simulation studies have been carried out for verifying the effectiveness of the proposed advanced controllers and compared with the conventional VC and model based FLC