TL;DR
This paper introduces a novel Deadbeat Direct Torque and Flux Control (DB-DTFC) algorithm for Permanent Magnet Synchronous Motors (PMSMs). By shifting to a stator-flux-oriented coordinate system and employing a forward Euler method for voltage calculation, the proposed method simplifies computation and improves accuracy compared to traditional rotor-flux-based DB-DTFC, leading to reduced torque ripple.
What Problem Does This Paper Address?
Permanent Magnet Synchronous Motors (PMSMs) are highly efficient and widely used, but their control is complex due to their multivariable, coupled, and nonlinear nature. Traditional Direct Torque Control (DTC) offers fast response but suffers from significant torque ripple. While Deadbeat Direct Torque and Flux Control (DB-DTFC) aims to improve upon DTC by achieving more precise torque and flux control, the conventional approach relies on a rotor-flux-oriented coordinate system. This method requires solving quadratic equations to determine stator reference voltages, which increases computational burden and can lead to less accurate results. The paper addresses the need for a more computationally efficient and accurate DB-DTFC method.
Methods
The core innovation of this paper lies in a new DB-DTFC algorithm that utilizes a stator-flux-oriented coordinate system instead of the traditional rotor-flux-oriented one. This shift allows for the use of the forward Euler equation to calculate the reference stator voltage, which is computationally simpler and avoids the complex quadratic equation solving of the conventional method. Additionally, the algorithm incorporates a discrete mathematical model to predict future control currents, enabling deadbeat control. The authors also explicitly account for the stator resistance voltage drop in their calculations, which they argue leads to more accurate voltage reference values. The proposed algorithm is then implemented using a MATLAB/Simulink simulation platform.
Datasets
Not discussed in the paper. The paper focuses on simulation-based evaluation using a defined PMSM model.
Key Insights and Ideas
- Stator-Flux Orientation Simplifies Voltage Calculation: Moving from rotor-flux to stator-flux orientation fundamentally changes how the reference voltage is computed, replacing complex quadratic solutions with a simpler forward Euler approach. This is a key insight for reducing computational load.
- Forward Euler for Voltage Calculation: Using the forward Euler method for voltage calculation in the stator-flux frame is presented as a more direct and less computationally intensive alternative to traditional methods.
- Accounting for Stator Resistance: Explicitly including the stator resistance voltage drop in the voltage calculation leads to a more accurate reference voltage, which is crucial for achieving precise deadbeat control.
- Predictive Current Control for Deadbeat Performance: The use of a discrete mathematical model to predict future currents is essential for achieving the "deadbeat" aspect of the control, ensuring the system reaches its target state in the next control period.
Key Results
The simulation results demonstrate that the proposed DB-DTFC algorithm achieves superior control performance in terms of torque and flux compared to both standard DTC and SVM-based DTFC (SVM-DTFC). Specifically:
- Reduced Torque Ripple: The proposed DB-DTFC algorithm exhibits significantly lower torque ripple. In a limited speed range, the torque ripple was reduced by approximately 6% compared to the DTC algorithm.
- Fast and Accurate Response: The PMSM controlled by the proposed algorithm demonstrated quick and accurate tracking of reference speed and torque.
- Ideal Flux Trajectory: The stator flux trajectory was observed to be an ideal circle, indicating precise flux control.
Limitations and Future Work
- Simulation-Based: The study is based on simulations using MATLAB/Simulink. The authors acknowledge that the next step is to test the proposed DB-DTFC algorithm on a real experimental platform to validate its performance in a practical setting.
- Current Prediction Methods: The paper mentions that there are many current prediction methods available. Future work could involve exploring and implementing different current prediction techniques to further enhance the deadbeat control performance.
Summary updated