A Study on Runoff Forecasting Model Incorporating Fuzzy Theory in Chou-Shui Creek

• Main Authors: Feng-Lien ChangChien, 張簡鳳蓮
• Other Authors: Pao-Shan Yu
• Format: Others
• Language: zh-TW
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/h354ae

碩士 === 國立成功大學 === 水利及海洋工程學系碩博士班 === 90 === Flood disaster is one of natural hazards in Taiwan. Recently, non-structural methods have been recognized as important alternative methods for flood mitigation. Real-time flood forecasting is one of such methods and may include rainfall forecasting, runoff forecasting and river routing. Only real-time runoff forecasting is studied in this work. The runoff forecasting model is developed on the basis of Grey System Theory. Validation from historical events in Chou-Shui Creek basin concludes that the proposed runoff forecasting model has good performance on 1 and 2 hours ahead forecasting. However, the accuracy of model performance will be reduced when lead time is increased. An apparent time lag and error in peak discharge exist in the runoff forecasting hydrographs. Therefore, the study further proposes an error prediction model to update the output of the runoff forecasting model. The grey error prediction model and the fuzzy error prediction model are applied based on the Grey System Theory and Fuzzy Set Theory, respectively. Comparing these two error prediction models, it is shown that the fuzzy error prediction model is more robust and efficient than the grey error prediction model. The study finally combines the Grey Rainfall-runoff Model with a rainfall forecasting model and the fuzzy error prediction model for real-time runoff forecasting on three subbasins in Chou-Shui Creek. It is found that the proposed model has reasonable performance for one to three hours ahead runoff forecasting.