- 學位論文
永磁同步馬達之參數鑑別與參考模型適應性控制
Parameter Identification and Model Reference Adaptive Control of Permanent Magnet Synchronous Motor
摘要
近年來，在向量控制的發展、變頻器控制技術的提升及具有快速運算能力的數位訊號處理器(DSP)價格下降等有利條件的配合下，已使得交流伺服馬達的控制相對容易，且其伺服性能優於直流馬達，而有逐漸取代直流馬達的趨勢，其中永磁同步馬達(PMSM)已被廣泛地使用在高性能的伺服驅動系統上。 永磁同步馬達(PMSM)的系統參數對控制器的設計有極大影響，因此鑑別之準確性是十分重要的工作，所以本文提出參數鑑別法則來精確地估測出受控體的系統參數包含電氣參數d-q軸電感、電阻與機械參數之轉動慣量與黏滯係數。利用參數鑑別之結果，本文進一步利用極點配置法設計出來的控制器能同時控制d-q軸電流和速度狀態。本文提出極點配置法，相較於傳統串級式控制架構(電流迴路的頻寬為速度迴路頻寬的六到十倍)而言，在不降低控制性能的前提下，電流控制之取樣率可與速度控制一致，降低微處理器運算負荷。但實際上馬達的電氣及機械參數會隨著運轉溫度、磁飽和及外加負載而改變，故本文進一步提出一參考模型適應性補償的方法，此方法同時考量d-q軸電氣和機械模型，可將永磁同步馬達(PMSM)實際模型逼近到參考模型。 本文採用德州儀器公司TMS320C6711 DSP與Xilinx公司Spartan-II XC2S50 FPGA，將DSP規劃成嵌入式系統運算核心，並將其拆解成數個工作模組包含磁極定位、向量控制、伺服控制器設計和參考模型適應性補償等，FPGA則處理外部I/O之橋接界面、PWM輸出等。最後，透過模擬與實驗結果，驗證本文所提出之控制器穩定性佳並且控制性能良好，確實完成在軟硬體架構下之永磁同步馬達控制。
並列摘要
Permanent Magnet Synchronous Motor (PMSM) drives are today gradually replacing classic dc drives in a large number of industrial applications, taking full advantage of key features of PM motors, such as high-power density, efficiency, robustness, reliability, compactness to working environment, and reduced maintenance and service requirements. For PMSM control, the PMSM’s electromagnetic and mechanical parameters must be adequately known and the parameters of controller have to be adjusted accordingly. Firstly, this thesis presents a novel algorithm to identify the resistance, inductance on the rotating reference d-q frame, and inertia, damping constant of PMSM. After, this thesis designs the controller which uses pole placement to let the responses of d-q frame currents and velocity satisfied our design specifications. Comparing to traditional cascade control of current and velocity loop, the sampling rate of current and velocity control are consistent in our proposed control scheme. Hence, the computation load of DSP could be decreased. However, the PMSM’s electromagnetic and mechanical parameters exist some variation in practical operation. This thesis proposes the Model Reference Adaptive Control (MRAC) to force the real model approximates the nominal model which is estimated from the parameter identification procedure. Our proposed algorithms are implemented in the platform of TI TMS320C6711 DSP, and XILINX Spartan-II XC2S50 FPGA. The DSP handles the main tasks, such as DC orientation, field-oriented control, pole placement control and MRAC. The FPGA interfaces the DSP and PMSM amplifier such as encoder decoding and A/D interface. Finally, the simulated and experimental results verify the performance of our proposed algorithms even if the plant parameters are varied and disturbance is added.