| 摘要: | 近??，受到溫室效應以及能源短缺之影響，節能減碳觀?逐漸受到重視，因此使用傳統內燃機引擎之各種交通運輸工具市場需求受到嚴重衝擊。此外，考慮未來汽?使用者習慣的調整與大眾運輸系統的發展，具有高能源效?與?污染排放優點之輕型電動?(Light Electric Vehicle, LEV)被視為未?取代傳統內燃機引擎?輛之最佳選擇。另一方面，具錯誤容忍控制之馬達驅動系統，可於系統發生故障時避免馬達失去正常運轉能力，適合應用於工具機、航太工程、汽車工業、冷氣壓縮機、機械手臂與機器人、電動載具及電動機需持續運轉之特殊應用場合。有鑑於此，本論文之目標即為發展以數位訊號處理器(Digital Signal Processor, DSP)為基礎之智慧型錯誤容忍控制(Fault Tolerant Control)六相永磁同步馬達(Permanent Magnet Synchronous Motor, PMSM)驅動系統，並應用於輕型電動車中輪內馬達(In-Wheel Motor)驅動系統上，以滿足輪內馬達運轉上之安全性與穩定性需求。
本論文首先發展以數位訊號處理器TMS320F28335為基礎之控制系統，並詳述六相永磁同步馬達驅動系統之架構，再進行六相永磁同步馬達的分析與推導其動態模型。此外，六相永磁同步馬達驅動系統為高度非線性之系統，且對於系統參數變化和外來干擾相當敏感，尤其是發生馬達繞組斷線或是反流器故障時，不平衡電流將使馬達轉矩抖動，導致馬達無法平順運轉，造成系統毀損，因此發展錯誤容忍控制成為六相永磁同步馬達驅動控制系統重要的議題。故本論文提出錯誤偵測與運轉決策判斷方法(Fault Detection and Operating Decision Method)，以達到錯誤容忍控制之成效。接下來進行輕型電動車與輪內馬達驅動系統之動態模型分析與推導。而在控制法則上則提出了具非對稱歸屬函數之TSK型模糊類神經網路 (Takagi-Sugeno-Kang Type Fuzzy Neural Network with Asymmetric Membership Function, TSKFNN-AMF)控制器，以及結合互補式滑動模式控制(Complementary Sliding Mode Control, CSMC)與非對稱歸屬函數之TSK型模糊類神經網路之智慧型互補式滑動模式控制器(Intelligent Complementary Sliding Mode Control, ICSMC)，以改善控制性能且達到錯誤容忍控制六相永磁同步馬達驅動系統之穩定性需求。此外，本論文亦提出機率模糊類神經網路(Probabilistic Fuzzy Neural Network, PFNN)控制器，並將上述錯誤容忍控制六相永磁同步馬達驅動系統應用於輕型電動車之輪內馬達驅動系統上，發展利用機率模糊類神經網路之錯誤容忍控制輪內馬達驅動系統，以達到輕型電動車應用所需之高控制性能，以及維持故障發生時輪內馬達驅動系統之穩定度，使車輛在加減速時提供更好的加減速控制響應，和駕駛者與乘客在車輛行進時更加舒適與安全。最後，由實驗結果可驗證本論文所發展之智慧型錯誤容忍控制六相永磁同步馬達驅動系統，確實具備優異之控制性能與錯誤容忍能力，且可有效應用於輕型電動車之輪內馬達驅動系統上。
In recent years, due to the greenhouse effect and fossil energy shortage, the concept of carbon reduction and energy saving has been valued highly. Therefore, the market of the internal combustion engine vehicles will be impacted seriously due to the increasing price of fossil fuel. Moreover, to consider the possible change of the habit of driver and the development of the public transportation in the future, light electric vehicles (LEVs) with high energy efficiency and low emissions are believed to be the best choice of transportation in the future. On the other hand, fault tolerant control for a motor drive system enables a motor to continue operating properly in the event of the failure. The fault tolerant control motor drive system is suitable for industrial applications such as mechanical tools, aerospace technology, vehicle technology, compressors, robotic arms, robots, electric vehicles (EVs) and some specific applications. For the above reasons, the purpose of this dissertation is to develop a digital signal processor (DSP)-based intelligent fault tolerant control of six-phase permanent magnet synchronous motor (PMSM) drive system. Furthermore, the developed intelligent fault tolerant control motor drive system will be applied to an in-wheel motor drive system in LEV to fulfill the requirements of the safety and system stability in LEV applications.
In this dissertation, the DSP-based control system using a 32-bit floating-point DSP, TMS320F28335, and the six-phase PMSM drive system are presented in detail first. Then, the dynamics of the six-phase PMSM is analyzed and derived. Moreover, the six-phase PMSM drive system is a highly nonlinear system and is very sensitive to the parameter variations and external disturbance. When the motor winding or respective inverter is broken, the unbalanced current will cause torque fluctuation so that the motor may be operated under a non-smooth situation and lead to a serious damage. Thus, the fault tolerant control for the six-phase PMSM drive system should be considered. Hence, the fault detection and operating decision method is proposed in this dissertation. Furthermore, the dynamics of LEV and in-wheel motor drive system are described in detail. In addition, two control approaches including Takagi-Sugeno-Kang type fuzzy neural network with asymmetric membership function (TSKFNN-AMF) control and intelligent complementary sliding mode control (ICSMC), which combines the merits of complementary sliding mode control (CSMC) and TSKFNN-AMF, are proposed to improve the control performance and to maintain the stability of the fault tolerant control six-phase PMSM drive system under faulty condition. Additionally, a probabilistic fuzzy neural network (PFNN) control is proposed to control the fault tolerant control six-phase PMSM drive system for an in-wheel motor drive system in LEV, and to achieve high control performance of the LEV and to maintain the stability of the in-wheel motor drive system under faulty condition. The fault tolerant control with PFNN of in-wheel motor drive system can provide better acceleration and deceleration control performance of the vehicle when the vehicle is accelerated or decelerated promptly and make the driving more comfortable with safety for the driver and the passengers. Finally, according to the experimental results, the developed intelligent fault tolerant control of six-phase PMSM drive system possesses good control performance and fault tolerance ability, and can apply to the in-wheel motor drive system in LEV effectively. |
