植基於粒子群最佳演算法之感應馬達直接轉矩控制器設計

Design of Induction Motor Direct Torque Controller Based on Particle Swarm Optimization Algorithm

黃益坤(Wong Yit Kuan)

指導教授 ： 王順源

國立臺北科技大學/電資學院/電機工程系研究所/碩士(2014年)

摘要

本研究使用直接轉矩控制(direct torque control, DTC)理論作為感應馬達之主控制架構，相較於磁場導向控制理論此架構無需解耦合、其動態響應較快、系統架構比較簡單並且容易計算。另外，為了避免破壞感應馬達結構與降低成本，本研究使用轉速估測器來實現無轉速量測器控制。
本研究將比例積分控制器(PI)結合粒子群最佳演算法(particle swarm optimization, PSO)來設計粒子群最佳演算法比例積分控制器(PSOPIC)作為轉速控制器。此轉速控制器能線上調整PI參數值，除了可提升系統性能與動態特性外，還可以改善傳統PI控制器參數固定的缺點。此外，馬達會因為溫升效應而導致定子電阻變動影響動態性能，因此，本研究設計PSO定子電阻估測器，能即時調適定子電阻值，以精準估測磁通量。另外，爲產生電壓向量命令，分別設計模糊轉矩控制器與模糊磁通控制器，以改善磁通響應與降低轉矩漣波並提升系統之動態性能。
經模擬及實驗結果證明，將PSOPIC轉速控制器、PSO定子電阻估測器、模糊轉矩控制器與模糊磁通控制器植入感應馬達直接轉矩控制系統中，在馬達負載轉矩為8 Nm，轉速控制範圍在36 rpm至1800 rpm時，本研究所提出方法皆具有優異的控制性能，且轉速動態響應也有良好的改善。

關鍵字

直接轉矩控制 ； 粒子群聚演算法比例積分控制器 ； 模糊控制理論

並列摘要

In this study, the direct torque control (DTC) is used as a main control structure of induction motor drive. Comparing to field-oriented control, this structure has no decoupling with fast system responses, simple structure and low computational complexity. This study is using speed estimator to achieve speed sensor-less control in order to prevent structure damaging of induction motor drive as well as to reduce cost. The particle swarm optimization PI controller (PSOPIC) is designed as speed controller by using proportional-integral controller with particle swarm optimization (PSO) algorithm. The speed controller can fine-tune the PI parameter online. Despite of system ability improvement and dynamic performance enhancement, the design resolves drawbacks of traditional fixed-parameter PI controller. In addition, dynamic performance of motor is affected by temperature effect of stator resistance. A PSO stator resistance estimator is able to fine-tune stator resistance in real-time to acquire precise flux linkage. To generate the reference of stator voltage vector, Fuzzy torque controller (FTC) and fuzzy flux controller (FFC) are designed to improve the flux response and reduce the torque ripple for better system dynamic performance.
With PSOPIC speed controller, PSOSRE, FTC and FFC are integrated into DTC system of induction motor within the range of motor speed varying from 36 rpm to 1800 rpm with 8 Nm load, the simulation and experimental results concluded the proposed solution has excellent control capability and significant improvement in speed dynamic response.

並列關鍵字

Direct Torque Control ； particle swarm optimization PI Controller ； Fuzzy control theory

參考文獻

[30] 王雅賢，粒子群最佳化演算法改良之研究，碩士論文，私立中原大學資訊管理學系，桃園，2006

[33] 黃郁授，自適應慣性權重改良粒子群演算法之研究，碩士論文，私立中原大學資訊管理學系，桃園，2006。

[31] Cheng-Jian Lin, Cheng-Hung Chen and Chi-Yung Lee, "Efficient immune-based particle swarm optimization earning for neuro-fuzzy networks design," Journal of Information Science and Engineering 24, 1505-1520, 2008, pp. 1505-1520.

[35] 郭天祐，切換式PD-PI 控制器之最佳化設計，碩士論文，私立淡江大學電機工程學系，台北，2007。

[32] R. C. Eberhart and Y. Shi, "Particle swarm optimization: development, applications and resources," Evolutionary Computation, vol. 1, 2001, pp. 81-86.