無位置量測三相四開關直流無刷馬達之速度控制研究

Abstract

本論文驗證了低成本之無位置感測器三相四開關驅動器其結構的可行性，並提出一新型的模糊PI控制法則，以處理變取樣回授現象，使整個系統效率最佳化與強健化。除理論推導外，論文並提供完整實驗結果加以驗證。 自動化是未來工業發展的必然趨勢，馬達驅動系統在整個自動化的過程中扮演了相當關鍵的角色，而降低馬達驅動器的製造成本也成為了未來的趨勢。本論文應用了新型的無轉子位置量測器的換相控制技術與三相四開關的直流無刷馬達驅動結構，藉以降低成本：首先，無感測控制是偵測兩相浮動電壓訊號的交越點，以求出適當的換相時機，此無感測控制只需要兩個電壓訊號，不需任何轉子位置感測器，不僅減少感測器的使用，還可增加系統的可靠度；接著，針對三相四開關的直流無刷馬達驅動器本研究提出應用非對稱脈寬調變技術和新型的無位置感測控制，不僅減少了電力開關的數量，且不需要額外的轉子位置感測器，實現了一個相當經濟的直流無刷馬達驅動器。 不論在傳統裝有霍爾感測器的結構，或無位置感測器方式，直流無刷馬達運轉時，其位置回授訊號的頻率，會隨著轉速改變而有所不同，而本論文提出考慮變頻取樣現象的模糊PI控制器，經由模糊邏輯推論和三個不同的PI控制器所組成，根據轉速的不同，觸發定義的PI控制器，最後將命令值權重加總，下達命令給變頻器以控制脈波寬度調變(PWM)的責任周期狀態。如此當馬達轉速改變時，模糊PI控制器可將轉速誤差根據速度狀態的不同計算出適當的命令值，落實了自由度更高的控制器設計及更強健的速度控制。而本文所設計的模糊PI控制器，以馬達轉速為參考，減少模糊邏輯設計的複雜度，使整個理論更容易實踐。最後，使用DSP來撰寫整個控制器的演算法，並針對各種取樣方式之控制器設計的情況下，進行無載及有載的實驗，並且做出相關的比較與討論，以驗證硬體結構與控制器之可行性與效率。

This thesis proves the practicability of the low cost structure using a four-switch three-phase inverter without position sensors, and proposes a novel fuzzy PI control theory. Besides theory computation, the thesis also presents complete experimental results in proof of the theory. Decreasing the cost of the actuators is a tendency in the future. This thesis applies a novel asymmetric pulse width modulation (PWM) scheme and uses the structure of four-switch three-phase (FSTP) brushless dc (BLDC) motor drives to reduce the cost. This PWM scheme not only achieves the BLDC motor using FSTP inverter, but also does not require any current sensor and complex algorithm, so the lowest manufacture cost of the inverter could be achieved. First, this thesis applies a novel commutation control for the FSTP BLDC motor drive without using position sensors. This sensorless control detects the crossing point of two floating voltages, and estimates the appropriate commutation timing. It only requires two voltage sensors to reduce the cost of position sensors and increases the reliability of whole system. Furthermore, this technique reduces the numbers of power switches and doesn’t require any position sensor. When the brushless dc motor runs with or without Hall sensors, the frequency of the position signals depends on the rotor speed. Considering this phenomenon, the fuzzy PI controller including three fuzzy logics and three PI controllers is designed in this thesis. If the speed of BLDC motor varies, the fuzzy PI controller will trigger the defined PI controllers depending on the rotor speed and sum the weighted output of each PI controllers. Then it delivers the command to the inverter in order to modulate the duty cycle. This kind of fuzzy PI controller provides higher flexibility of design, and a more robust goal is accomplished. The fuzzy logic of the fuzzy PI controller designed in this thesis is based on the rotor speed, and therefore the controller reduces the complex of fuzzy logics and is easy to be implemented. Finally, the whole algorithm is implemented in a DSP. This thesis focuses on distinct sampling methods, and the experimental results with distinct conditions are compared and discussed.