本論文提出一套在線式智慧型太陽能最大功率追蹤系統,系統中 以數位信號控制器dsPIC30F4011 為控制核心,並整合模糊最大功率追蹤法(Fuzzy Maximum Power Point Tracking, FMPPT)、類神經網路(Artificial Neural Network, ANN)及LabVIEW 虛擬儀表軟體等,以實現太陽能之最大功率追蹤控制。FMPPT 已證實能有效追蹤到太陽能最大功率點,但其缺點為追蹤速度稍嫌緩慢;ANN 具有快速響應的優點,但其訓練資料須涵蓋所有範圍才能達到準確的輸出。為達成上述之快速響應與有效訓練的目標,本論文中藉ANN 進行太陽能之最大功率追蹤控制,同時,ANN 的輸出參考電壓隨時與FMPPT 的輸出參考電壓進行比較,當誤差超出預設誤差值時,則將此資料收集並重新訓練網路,此部份則藉LabVIEW-Matlab 介面達成。由實驗結果證實,本論文所提出的方法能快速有效地追蹤到太陽能最大功率點。
This thesis presents a maximum power point tracking (MPPT) technique for solar photovoltaic systems using on-line intelligent method. The proposed system regards the digital signal controller dsPIC30F4011 as the control center. It combines the fuzzy maximum power point tracking (FMPPT) technique, artificial neural network (ANN) approach, and LabVIEW virtual instrument software to achieve the MPPT for solar photovoltaic systems. The FMPPT has been validated to be an effective method to track to the maximum power point. However, disadvantage of this method is that the tracking speed is slow. Although ANN has the feature of fast response, the training data must cover all over the possible range to generate the accurate output. To reach the goals of both fast response and effective training, this thesis employs ANN method to carry on the MPPT for solar photovoltaic systems. In the mean time, the output reference voltage of ANN is compared with the output reference voltage of FMPPT at anytime. When the error between ANN and FMPPT outputs exceeds the setting tolerance, this data will be collected and the network is then re-trained, which is reached by the LabVIEW-Matlab interface. The experimental results have verified the proposed method can track fast and effectively to the maximum power point of solar photovoltaic systems.