2006-05-042024-05-14https://scholars.lib.ntu.edu.tw/handle/123456789/659578摘要：本研究將選用模糊推論之類神經網路，建立低流量時期水庫操作模式，利用模糊推論規則庫能隱含專家知識與操作者經驗之優點，期能提供水庫管理單位於操作運轉之參考。研究將除收集石門水庫基本水文、氣象與下游用水資料，並針對低流量時期河川流量狀況進行相關統計分析，以其建置低流量時期智慧型水庫經營策略。預期以模糊推論之類神經網路架構，建立水庫操作之規則庫，檢討模式對於水庫操作之結果，推求未來水庫之最佳營運策略，並與水庫既定規線(M-5)運轉方式之操作結果進行比較，以研議智慧型水庫操作模式為最終目標。<br> Abstract: In order to ensure the sustainable use and management of water resource, we intend to use the intelligent control methodology to devise a suitable reservoir operation model in this project. A fuzzy inference neural network would be used to construct the reservoir operation model in drought period. In the advantage of the rule base, obtained from expert knowledges and the operator's experiences, we could make the optimizing strategy for reservoir operation in land fallow during shortage period. We would collect the inflow data of the Shihmen reservoir and the water demand amount in downstream, and analyze the discharge in drought period. The knowledge acquired and implemented and the inference system are first investigated, and then a fuzzy inference neural network will be constructed to search the intelligent reservoir operation policy. For the purpose of comparison, a simulation of the current used M-5 operating rule curve is also performed. The applicability and practicability of the developed intelligent reservoir operation model could then be obtained.水庫操作類神經網路模糊推論系統reservoir operationartificial neural networkfuzzy inference system