Adaptive backstepping based online loss minimization control of an induction motor drive

The efficiency of an induction motor (IM) can be improved by the optimum selection of a flux level in the motor. Among the numerous loss minimization algorithms (LMA), a loss-model-based approach offers a fast response and no torque pulsation. However, it requires the accurate loss model and the knowledge of the motor parameters to find the optimum flux level. Therefore, a technical difficulty in deriving the loss-model-based LMA lies in the complexity of the full loss model and the on-line parameter adaptation for the precise motor parameters.; In an effort to overcome the drawbacks of on-line loss model controllers (LMC), this thesis presents a new loss-model-based LMA for inverter-fed IM drives aiming at both high efficiency and high dynamic performance.; A new LMC is proposed for the loss minimization of vector-controlled IM drives. An IM model in d-q coordinates is referenced to the rotor magnetizing current and then an iron loss resistance is added in parallel to the magnetizing inductance. This transformation leads to no leakage inductance on the rotor side by incorporating it into other parameters. This decomposition feature into d-q components makes the derivation of the motor loss model and LMC simpler while keeping the effect of leakage inductances.; In order to achieve high dynamic performance, an adaptive backstepping based nonlinear controller (ABNC) is designed incorporating iron loss under the parameter uncertainties of rotor resistance and load torque. In proposed IM equations, no additional state variables are added while iron loss is considered. Thus, an ABNC incorporating iron loss can be designed without much more complexity compared to the one with neglected iron loss. ABNC achieves desirable motor dynamics at any operating point while the flux level is varied by the LMC in order to reduce the input power.; Adaptive backstepping technique provides a tool to design the controller avoiding wasteful cancellations of certain nonlinearities. Another important feature of an adaptive backstepping technique is that it can derive parameter update laws simultaneously with control laws from the error dynamics. With an extra gain introduced in adaptation laws design, we take advantage of this feature by combining the ABNC with LMC, thus an on-line parameter adaptation of LMC can be obtained with no extra effort.; The complete closed loop control of the proposed LMC based IM drive is implemented in real-time using digital signal processor board DS 1104 for a laboratory 1/3 hp motor. The dynamic performance of the proposed controller and parameter adaptation features are examined. The effectiveness of the proposed loss minimization scheme through a wide range of speed regions including the field weakening region is demonstrated through computer simulation and experimental results.