本研究是以類神經網路之一的倒傳遞網路(Back-propagation network),以及模糊控制(Fuzzy Control)的理論為基礎,應用在非線性的流量控制閥中,用來改善一個泵浦的量測系統的控制。其研究方法是在執行量測控制的動作前,先以類神經網路為基礎來做系統的鑑別,之後再利用模糊控制理論來調整泵浦輸出流量控制閥,以快速完成控制目標值。 泵浦的主要輸出特性是流量和揚程,其特性曲線通常都以兩者來表現出來;在不同馬達轉速之下,便有不同的特性曲線。因此在泵浦量測系統的控制過程當中,就必須要考慮揚程、流量和馬達轉速、控制閥開關大小之間的關係,來進行系統鑑別,並進行控制,經實驗結果,可達到理想的控制效果。 本研究以Visual Basic的程式語言來建構控制系統的環境,另外以Matlab軟體進行類神經網路和模糊理論的運算,並針對所得到的結果加以討論和探討。 This research applies the most extensively-used back-propagation neutral network theory and fuzzy control theory to the hydraulic pump measurement system control. The method of this research is to identify the system on the basis of back-propagation network, and then determine which may approximate the target. Afterward the theory of fuzzy control applies to regulate the system in order to approach the target which is required. The output characteristics of pump are flow and head, which are two main properties representative of the characteristic curve. Different motor rotation speed contributes to different characteristic curve. Therefore, flow and head of pump's output should be taken as referent indices during the control process in order to approach the ideal result.