Improved Model Predictive Control Design for Current-Source and Voltage-Source Motor Drives

• Author / Creator

  Xue, Cheng

• The linear control structure shows limited bandwidth for high-performance ac drive applications and also it is difficult to deal with complex systems with multi-objective optimization demand and nonlinear, high-order characteristics. With the development of more powerful digital platforms and the emergence of new switching devices, the control algorithm for industrial drives application is constantly evolving and significant research efforts have been carried out over the past several years. Particularly, model predictive control (MPC) emerges as an attractive solution to power converters and it has become a research hotspot recently. The time-domain modeling process makes MPC naturally suit the multiple-input multiple-output (MIMO) systems and it shows obvious advantages in the field of power electronics such as the intuitive concept, faster dynamic response, multi-variable control capability, nonlinear objective handling ability, etc. The research presented in this thesis thus exploits the benefits provided by MPC to address the prominent control challenges existing in industrial motors driven by current-source converters (CSCs) and voltage source converters (VSCs). Particularly, the medium voltage (MV) CSCs are applied with the megawatt power range such as in the wind tunnel drive system, fans and compressors, etc. While the VSCs can be more frequently seen in the relatively lower power occasion, such as electric vehicles, where the switching frequency can be higher to hundreds of kilohertz. The control scheme of both CSCs and VSCs should meet the increased demand and requirements imposed by industrial drive applications because these are directly related to whether the motor can work normally, and also the system stability, reliability, efficiency, robustness, etc. Considering the limitation of the conventional linear approach and the lack of MPC design for the high-order system and back-to-back(BTB) drive-level optimization, the main work in this thesis can be divided into the following three aspects.
  Firstly, the flexible cost-function design mode provided by MPC is exploited to coordinately control the BTB structure, achieving system-level optimization compared to the conventional linear approach. For the BTB CSCs-fed drives, the proposed cost function with direct multi-variable feedback is developed to solve the filter resonance with the second-order characteristic. Besides, the robustness is enhanced by incorporating the disturbance observer, and the long horizon prediction case with the benefits of only line current feedback is presented for such a second-order system. Furthermore, the selective harmonic magnitude penalization is included in the cost function to reject the grid-side harmonic. According to the operated speed region of the motor, the peak-to-peak value of the common mode voltage (CMV) and its third-order harmonic are respectively suppressed to address the common mode (CM) resonance for the transformerless operation. On the other hand, regarding the VSCs-fed drives for relatively low power applications, the nonlinear control objective in terms of intermediate dc-link capacitor current ripple reduction is developed in the BTB structure, which contributes to the lifetime extension or the size reduction of the dc-link capacitor.
  Secondly, the potential of MPC is fully developed to control multiple state variables in the third-order LCL-filtered system, avoiding the nested multi-loop control structure in the linear approach. This benefit can be seen in both the LCL-filtered grid-connected application and the long cable-fed motor side with an output LC filter. The proposed MPC schemes show significant benefits in terms of the intuitive design concept, faster dynamic response, and simplified tuning procedure compared to the conventional multi-loop approach.
  Finally, when applying the MPC to the power converters with high switching frequency operation, the high sampling/interrupt frequency demand becomes a great challenge because of the big computational difficulty. Therefore, the multi-rate technique is designed into the MPC formulation, where the switching frequency can be significantly increased and a low sampling/interrupt frequency can be adopted. In this case, the quite long interrupt period allows the easy realization of the control algorithm and an abundant computational margin can remain in the digital platform. The proposed multi-rate MPC by enabling longer interrupt duration is more efficient than the conventional way focusing on how to reduce the execution time of the MPC algorithm itself. The multi-rate MPC also stands for a general formulation of MPC in the field of power electronics.

• Subjects / Keywords

  • Power electronics
  • Motor drives
  • Model predictive control
  • Current-Source Converters
  • Voltage-Source Converters

• Graduation date

  Spring 2023

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-1trk-5j36

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.