本計畫擬以三年為期探討水庫排洪對下游河道洪水位之影響，發展水庫系統排洪最佳即 時操作模式和排洪最佳操作規則。本計畫將利用水文即時及預報資料進行水庫排洪最佳即時 操作，以決定水庫最佳即時及預報排洪量。本子計畫之研究成果與其他子計畫整合後，即可 建立一整套之水庫排洪操作與下游洪水預報模式。 本計畫之工作流程包括三大部份，分別為資料收集、模式和規則建立與測試及模式和規 則應用與檢討，各部分之工作項目可分為三年度進行，第一年為資料收集、水庫排洪即時操 作優選模式建立、水庫排洪最佳即時操作模式驗證與測試；第二年為決策樹資料礦掘法演算 建立、水庫排洪最佳操作規則建立、水庫排洪最佳操作規則驗證與測試以及第三年為水庫排 洪最佳即時操作模式與最佳操作規則之配合應用、整體計畫之整合及整合模組應用。本計畫 將以曾文溪流域為範例，建立曾文水庫最佳排洪即時操作模式和排洪最佳操作規則。 We are proposing a three-year project. The major purpose of the project is to analyze the influence of reservoir releases on stages at the downstream stations during a flood. Additionally, it is proposed to develop the optimal real-time reservoir operation model and rule for flood control. Using the real-time and forecast hydrological observations, the developed operation rules can be used to determine the reservoir releases during the flood. Furthermore, using the results obtained from the proposed project and other related projects will allow us to establish a reservoir flood control operations and downstream river flood forecasting model for a river basin. The tasks to be carried out can be grouped into the following three general categories: (1) data collection, (2) model and rule development and verification, and (3) model and rule application and evaluation. The specific tasks to be carried out in the first year of the project are (1) collection of data, (2) development of real-time optimal operation model for flood-control, and (3) verification of the model. In the second year, the tasks are (1) building of decision tree data mining algorithms, C4.5, (2) development of real-time optimal operation rules for flood-control, and (3) verification of the rules. And the tasks in the second year are (1) implementation of the developed operation model and rule, (2) integration and implementation of results obtained from all related projects, and (3) application of the integration model. The developed methodology will be applied for Tsengwen Reservoir in the Tsengwen river basin. The results obtained from the optimal real-time reservoir operation model and rule will provide the optimal real-time reservoir releases.
Spearman’s Rank correlation coefficient