- 學位論文
以最小平方法估測永磁式直流機參數
Parameter Estimation for Permanent Magnet Direct Current Machine by Least Mean Square Method
摘要
在現今的工業上，直流機已成為不可或缺的動力來源及發電來源，為了提升直流機之效能，直流機數據量測及直流機參數估測之研究扮演非常重要的角色。本文利用直流電動機啟動時的暫態數據：電壓、電流及轉速來計算直流機的參數估測。為了更準確的參數估測結果，必須先對原取樣訊號作最佳化處理。利用多項式化方法嵌合，將濾除原取樣訊號之雜訊，保留原曲線特性。直流機之電壓、電流及轉速曲線可表示為電壓方程式，建構出直流機之等效電路模型，以最小平方法估測直流機等效電路參數。該組等效電路參數可以建立動態模型中的輸出轉矩，得一轉矩曲線。獲得轉矩曲線後，使用最小平方法計算出機械系統之轉動慣量及摩擦係數。最後，將估測出的等效電路參數及機械系統參數代入等效電路方程式及轉矩方程式，求出電樞電流曲線及轉速曲線。模擬值與原取樣訊號做比對驗證，證實本文方法的可行性且準確度高。本文採用多項式法對原取樣信號作最佳化處理，使用最小平方法作等效電路參數及機械參數的估測，具有誤差低，效率高及計算方法簡易等優點。 本文另實現一組完整的直流機監控系統來實現直流機信號擷取的目的。監控目標為直流機，作為直流電動機及直流發電機的行為表現。另設有交流伺服馬達，作為直流機的原動機；交流伺服馬達由伺服驅動器所驅動，可以控制馬達的轉速，且提供轉速與轉矩的即時資訊。伺服驅動器的速度命令由電腦提供，電腦的人機介面由LabVIEW撰寫而成，有助於使用者監視及控制整個系統。使用者可以透過人機介面編輯直流機轉速，透過信號擷取卡將轉速命令輸入至8051微處理機，由8051微處理機轉成RS-232格式輸入伺服驅動器，再由伺服驅動器驅動伺服馬達，完成轉速控制的目的。直流機的電壓、電流及轉速信號透過信號轉換電路，轉為適合擷取的電壓準位，再輸入至信號擷取卡中，並記錄直流機在不同行為下所產生的數據。這些資料有助於了解直流機的動態特性，並估測直流機的參數。
並列摘要
Today, DC machine has become indispensable power and generation sources in industry. In order to improve DC machine efficiency, DC machine measurement and parameter estimation play an important role for the research. In this paper, the start-up transient signals of voltage, current and speed in DC machine are used for parameter estimation. In order to obtain more accurate results of parameter estimation, the sampling signal must be optimization. This paper uses polynomial fitting method to filter noise in the sampling signals and keep the original curve characteristics. The voltage, current, speed curve of the DC machine can be expressed as a voltage equation, and then the DC machine equivalent model can be established. After that, this paper estimate DC machine equivalent circuit parameter with least square method. The equivalent circuit parameter can set up output torque in the dynamic model, and obtain a torque curve. By the torque curve of mechanical system, this paper calculates moment inertia and friction coefficient with the least square method. Lastly, the simulated armature current and speed can be found by equivalent circuit parameters and mechanical system parameters. Comparing simulation curve with the original one, the feasibility and accuracy of this method can be proved. In this paper, the original sampling signals are optimized with polymerization, equivalent circuit parameters and mechanical parameters are estimated with least square method. The method has advantages of is low deviation, high efficiency, and simple calculation. This paper also sets up a DC machine monitoring system to get the DC machine signals . The monitoring target is a DC machine, which can perform as a DC motor or a DC generator. An AC servo motor is installed in this system as the prime motor. The AC servo motor driven by AC servo driven can control the motor speed and provides real time information about speed and torque. The speed command of servo driver is provided by the computer. Human–machine interface is written by the LabVIEW, which helps users to monitor and control the entire system. Users can edit DC machine speed by the human–machine interface. User can deliver the speed command to servo driver by DAQ card, and DAQ card delivers the speed command to AT89S51. AT89S51 transforms the command to servo driver by RS-232 form. And then the servo drive servo motor, speed control purposes is completed. DC machine’s voltage, current, speed and torque signal transform to suitable voltage standard by signaling conversion circuit and input to DAQ card. It can save the DC machine data with different actions. The information can help to understand DC machine dynamic characteristic and estimate the DC machines parameter.