| 摘要: | 本計畫之目標為研製以數位訊號處理器為基礎之高效能同步磁阻馬達驅動系統，並將其結合智慧型控制使系統具有更好的控制效能。本計畫首先於第一年度進行以高效能同步磁阻馬達為基礎之模擬系統設計、分析及推導其動態模型，並發展以數位訊號處理器為基礎之驅動控制系統，用以驅動同步磁阻馬達系統。同步磁阻馬達驅動系統本身是個高度非線性的時變系統，對於系統參數變化和外來干擾相當敏感，尤其是系統所產生的轉矩是被時變的電感與電流所影響。因此，提升同步磁阻馬達的效率以及性能是目前最重要的發展議題。故本計畫第二年度將著眼於設計一個以智慧型步階迴歸控制器為主的控制系統，並使用在高效能同步磁阻馬達驅動系統的定位控制。相較於一般傳統步階迴歸控制，此智慧型控制系統將採用具有線上學習與快速收斂特性的遞迴式特徵選擇模糊類神經網路作為主控制器，來近似理想的步階迴歸控制器以減緩使用切換函數所引起的抖動現象。此外，並利用適應性補償控制器來改善遞迴式特徵選擇模糊類神經網路近似系統所造成的近似誤差。因此智慧型步階迴歸控制器綜合上述，以強化軌跡追隨的控制效能與穩定性。另一方面，因為同步磁阻馬達具有嚴重的磁場飽和現象，同步磁阻馬達的傳統型每安培最大轉矩控制在實際應用上是很難達到最佳的效能，因此本計畫將在第三年度進行設計以適應性步階迴歸控制結合拉格朗日乘數法之每安培最大轉矩控制。適應性步階迴歸控制拉格朗日乘數法之每安培最大轉矩控制是利用拉格朗日乘數法得出在每安培最大轉矩時直軸與交軸的電流命令，且考慮電感變化並使用適應性步階迴歸控制來估測系統中的電感值，並且減緩磁飽和所造成的影響。除此之外，利用李亞普諾夫穩定性理論推導出適應律並且即時更新控制參數，以改善控制性能且達到高效能同步磁阻馬達驅動系統之穩定性要求。 本計畫採用德州儀器之TMS320F28075 DSP為控制核心，以C語言實現所推導之各種控制法則，並利用DSP周邊擴充電路之編碼器介面電路與類比數位轉換電路讀取位置感測器所量測之轉子位置與馬達轉速，以達成以數位訊號處理器為基礎之高效能同步磁阻馬達驅動系統。 ;The purpose of this project is to develop a digital signal processor (DSP)-based high-performance synchronous reluctance motor (SynRM) drive system. In order to increase the control performance of the SynRM, the intelligent control systems are developed in this project. In the first year of this project, the simulation system of the high-performance SynRM is designed, and the dynamic model of high-performance SynRM system is also analyzed and derived. Moreover, the DSP-based motor drive and control system is developed to actuate the high-performance SynRM drive system. The SynRM drive system is highly nonlinear and very sensitive to parameter variations and external disturbance, especially the torque of the SynRM is nonlinear and time-varying and very sensitive to the variations of the inductance and current. Therefore, to enhance the efficiency and performance of the SynRM drive system is the most important issue in the whole development of the system. Thus, in the second year of this project, an intelligent backstepping control (BSC) system is proposed for the position control of the high-performance SynRM drive system. Comparing with the traditional BSC, the recurrent feature selection fuzzy neural network (RFSFNN) is the main controller and used to approximate the idea BSC with online learning and fast convergence capabilities in order to alleviate the chattering phenomena owing to the sign function. In addition, to compensate the possible approximated error of the RFSFNN, an adaptive compensator is designed and augmented. As a result, the proposed intelligent BSC system will be able to have a satisfactory tracking performance with guaranteed stability. Furthermore, an adaptive backstepping control (ABSC) based Lagrange multiplier (LM) maximum torque per ampere (MTPA) control will be proposed and developed to improve the control performance of the SynRM drive system in the third year of this project. Owing to the saturation effect, the optimal performance of the traditional MTPA control is very difficult to obtain in practical applications. Thus, an ABSC based LM MTPA control using the LM to obtain the current command of the direct and quadrature axis is proposed to achieve the optimal MTPA of the high-performance SynRM drive system. Then, the adaptive law can estimate the required inductance to alleviate the saturation effect. Additionally, an adaptive law is developed to improve the control performance and to achieve the requirement of stability for the high-performance SynRM drive system. The adaptive law is derived using the Lyapunov stability theorem to update the control parameters in real-time. A TMS320F28075 DSP made by Texas Instruments is the core of the proposed control system. Moreover, the proposed control algorithms are realized in the DSP using the “C” language. Furthermore, a DSP extension board reads the rotor position, motor speed and phase currents from the sensors using the encoder interface circuit and analog to digital converters. Therefore, the DSP-based of the high-performance SynRM drive system can be achieved. |
