本計畫擬以三年為期探討颱風侵襲期間考量取水濁度限制下水庫各層放水口放水對下游河道洪 水位、取水口濁度、防淤與蓄水功效之影響，發展多階段智慧型水庫排洪與防淤之最佳即時操作模式。 本計畫將利用水文與水質即時及預報資料進行水庫排洪與防淤最佳操作之即時決策，以決定水庫各層 放水口之最佳即時放水量。 本計畫之工作流程包括三大部份，分別為：水文與水質預報模式之建立、操作決策模式之建立、 預報模式與操作模式之整合與即時模擬測試，各部分之工作項目可分為三年度進行，第一年為水文與 水質等相關資料之蒐集、定量降雨預報模式之建立、短時距入庫流量預報模式之建立與總入庫流量預 報模式之建立、WASP(Water Quality Analysis Simulation Program)水庫與下游河道泥沙濃度模擬模式之 建立、水庫排洪與防淤最佳即時操作模式之建立；第二年為應用第一年水庫最佳操作優選模式進行整 場歷史事件多層放水口最佳操作歷線之優選、應用調適性網路模糊推論系統(Adaptive Network-based Fuzzy Inference System, ANFIS)配合最佳操作歷線建構智慧型多階段多層放水操作模式、ANFIS 智慧 型多階段多層水庫放水操作模式之驗證；以及第三年為預報模式與操作模式之整合、實際颱風事件之 即時決策操作、比較各模式之操作成效以評估各模式之優劣、藉由模擬測試建立操作規則以提供給水 庫管理局操作參考。本計畫以大漢系石門水庫流域為研究區域，建立多階段多層智慧型水庫即時操作 模式，以評估所建立之模式應用於水庫考量取水濁度限制下防洪與防淤實務操作之成效。 We are proposing a three-year project. The major purpose of the project is to analyze the influence of multi-layer reservoir releases on stages at the downstream stations, turbidity at the downstream inlet, sedimentation control, and storage efficacy during a typhoon-flood event. Additionally, it is proposed to develop the optimal multi-phase multi-layer intelligent real-time reservoir operation model and rule for flood and sedimentation control. Using the real-time monitored and forecasted hydrological and water quality information, the developed operation rules can be used to determine the multi-layer reservoir releases during the flood. The tasks to be carried out can be grouped into the following three general categories: (1) data collection; (2) development of hydrologic and water quality forecast model; (3) development of operated decision model; and (4) integration of forecasted and operation model and real-time simulation testing. The specific tasks to be carried out in the first year of the project are (1) collection of data; (2) development of quantitative precipitation forecast model; (3) development of short lead-time and total reservoir inflow forecast model; (4) development of TSS (total suspended solids) concentration simulation model among reservoir and downstream channel using WASP (water quality analysis simulation program); and (5) development of optimization model. In the second year, the tasks are (1) optimizing multi-layer optimal reservoir release hydrograph during historical typhoon event; (2) development of real-time intelligent multi-phase multi-layer reservoir operation model using ANFIS (adaptive network-based fuzzy inference system) with optimal operation hydrograph for flood and sedimentation control; and (3) verification of the model. The task in the third year is to process (1) integration of forecasted and operation model; (2) real-time operation testing using historical typhoon events; and (3) assessing the goodness of the developed operation models. The developed methodology will be applied for Shimen Reservoir in the Dahan river basin. The results obtained from the optimal multi-phase multi-layer real-time reservoir operation model and rule will provide the optimal real-time reservoir releases considering turbidity constraints, flood and sedimentation control at the same time.
real-time reservoir operation