Model predictive control of induction motor drive with constraints

Gan, L 2014, Model predictive control of induction motor drive with constraints, Doctor of Philosophy (PhD), Electrical and Computer Engineering, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Gan.pdf Thesis application/pdf 20.13MB
Title Model predictive control of induction motor drive with constraints
Author(s) Gan, L
Year 2014
Abstract Induction motor drives play a significant role in various industrial applications. It is an essential machine that converts the electrical energy to mechanical motion. Meanwhile, Model Predictive Control (MPC) is an optimal control algorithm developed for constrained control of Multi-Input-Multi-Output (MIMO) systems. In general, there are two categories of MPC methods which are applied to motor drives: the traditional MPC and the finite control set MPC. The former type of MPC has modulation-based implementation, which generally requires sufficient computational time and a Linear Time Invariant (LTI) model. Complementarily, the Finite Control Set (FCS) predictive control method is proposed based on the benefit of finite switching states of the inverter in the induction motor drive. In this thesis, both MPC methods will be investigated and discussed along with own contributions in both theory and applications to induction motor control. For the modulation-based MPC design, the continuous-time model based predictive control scheme is selected due to its advantages. In this thesis, there are two different approaches investigated for this type MPC: centralized and cascaded control structures. On one hand, the centralized MPC is proposed to achieve the induction motor speed control by using a single model predictive controller, the major challenge is found due to the non-linearity of the full order model of the induction motor. Thus, the Gain-Scheduling (GS) technique is proposed for MPC by pre-defining the operating conditions according to different equilibrium points of the system operations. By employing the Direct Field Oriented Control (FOC) concept, the Gain-Scheduled MPC is developed and validated for the Variable Speed Drive (VSD) of induction motor. On the other hand, the cascaded MPC control is proposed to further develop the continuous-time Model Predictive Control of the induction motor drive. Based on the Indirect FOC strategy, two MPC controllers are designed according to inner-loop electrical model and outer-loop mechanical model, respectively. Furthermore, position control is also investigated for servo drive application by using the cascaded MPC technique. Another major contribution of this thesis is on the advance of theory and applications of FCS-MPC to induction motor control. The new analytical results have been obtained in terms of closed-loop feedback control gain and the closed-loop poles of the FCS-MPC system, based on which closed-loop stability is established for the system without constraints. More importantly, integrator has been incorporated in the FCS-MPC system to overcome the steady-state errors existed in the original system. The proposed approach not only maintains the simplicity of the original algorithm, but also improves its robust performance in the presence of parameter uncertainty. In the alpha-beta reference frame, a resonant controller is derived to overcome steady-state errors of the original FCS-MPC system by following the same thought process as introduced in the dq reference frame. All of proposed control strategies are compared with the conventional PI-based FOC method, the control performances of the experimental results are studied and analysed. Moreover, the robustness analysis of the proposed control methods is investigated by comparing the experimental results based on mismatched model parameter values.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Electrical and Computer Engineering
Keyword(s) Induction motor drive
Control system
Model Predictive Control
Finite Control Set
Gain Scheduling
Non-linear Control
PI control
Version Filter Type
Access Statistics: 799 Abstract Views, 4229 File Downloads  -  Detailed Statistics
Created: Fri, 12 Sep 2014, 14:51:15 EST by Maria Lombardo
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us