蘭陽溪洪水預報模式之研究

A Study on Flood forecasting Model for the Lan-Yang Creek

陳信中(Hsing-Chung Chen)

指導教授 ： 許銘熙

國立臺灣大學/生物資源暨農學院/生物環境系統工程學研究所/碩士(2006年)

https://doi.org/10.6342/NTU.2006.00060

摘要

蘭陽溪下游之蘭陽平原因地勢低窪，夏秋之際常因梅雨季或颱風所帶來的豪雨造成淹水災害，加上自然地形的陡峻、河川短促，經常導致嚴重的洪水災害，造成人民生命及經濟的重創。因此，若能提供沿岸的洪水位剖面線及特定地點的洪水位預報資訊，作為洪水預警之參考，將可有效減低洪水所帶來的災害損失。
對於流域內設置水文站之地點，可利用其水位資料依據類神經網路模式做短時期預報資料，但對於其它地點之洪水位則需利用洪水演算模式作為工具。本文嘗試將類神經網路模式之預報水位納入河川洪水演算模式中，進行演算得到蘭陽溪河系合理的洪水位資訊及沿岸的洪水位剖面線。
本文以最佳化理論所建立的參數修正，配合類神經網路模式於蘭陽大橋水位站之預報水位而獲得最符合流況的曼寧係數值，為驗證模式的效能，本文以兩場颱洪事件進行模式測試。由模擬的結果顯示，本文所建立的河川洪水預報模式，可以在颱洪期間提供合理及準確的河川洪水資訊。

關鍵字

動力波模式 ； 類神經網路 ； 洪水演算 ； 洪水位預報

並列摘要

The downstream Lan-Yang Creek is located in the low-lying Lan-Yang Plain, where the inundation disasters occur frequently in summers and falls because of the torrential rain. The high intensity precipitation combining with short river course usually results in flood inundation, which inflicts disastrous losses of life and economy. If the real-time forecast information, including the water stage at the significant locations and longitudinal profiles along a river is available before flooding, the damages would be effectively mitigated.
The real-time forecasting can be performed by applying ANN model using river stage data at gauged station. But there is no river stage data for the forcasting using the ANN model other than the places with gauged station. The purpose of the study is to develop a flood forecasting model which integrates river stage prediction at gauged station with ANN model and flood routing model. The integrated model makes numerical calculations of water stage at the significant locations and longitudinal profiles along the river.
The model parameter of the flood routing model is updated by using the optimization technique which minimized the differences of stages between the ANN model and river flood routing model. Two typhoon events were simulated to confirm the accuracy of the forecasting model. The present model can provide a satisfactory and reliable river stages forecasting for a short period following a storm.

並列關鍵字

Dynamic routing model ； Artificial Neural Network ； Flood routing ； River stage forecasting

參考文獻

31. 吳建均，2005，河川洪水位演算模式之研究，國立台灣大學生物環境系統工程學研究所碩士論文。

78. 蔡亞欣，2003，模糊範例學習推論系統於水位預測之研究，國立台灣大學生物環境系統工程學研究所碩士論文。

1. Campolo, M., Andreussi, P., and Soldati, A., 1999 , River flood forecasting with a neural network model, Water Resources Research, Volume 35, Issue 4, pp.1191-1197.

2. Chiang, Y. M., Chang, L. C., and Chang, F. J., 2004, Comparison of static-feedforward and dynamic-feedback neural networks for rainfall-runoff modeling, Journal of Hydrology, Volume 290, pp.297-311.

3. Chang, F. J., Chang, L. C., and Chiang, Y. M., 2005, Reply on ‘Comparison of static-feedforward and dynamic-feedback neural networks for rainfall-runoff modeling’ by Y.M. Chiang, L.C. Chang and F.J. Chang, 2004, Journal of Hydrology 290, 297-311., Journal of Hydrology, Volume 314, pp.204-206.