以複合型模式分析區域淹水潛勢

台灣在夏秋之際常受到颱風及豪雨侵襲，當颱風暴雨來臨時各地區無法及時排水造成中下游平原地區積淹水災情嚴重，需仰賴淹水潛勢圖做為淹水影響範圍及深度之評估資訊，達到事前防災的效果。傳統淹水潛勢圖模擬過程，需要大量的輸入資訊及經過繁複演算方可推估區域淹水潛勢，不僅造成電腦系統運算負擔亦無法達到防災工作所需的即時效果，且只能提供特定降雨量下之淹水情況查詢。本研究提出以複合型模式建立小區域即時淹水災害範圍推估，以獲得即時洪災資訊及災害影響範圍。由於造成淹水的成因相當多且複雜，希望瞭解區域淹水預測模式與可能影響因子間的關係，先以相關性分析及因素分析探討淹水影響因子間的線性關係，再利用K-means聚類分析依淹水特性進行分類，劃分不同淹水特性之資料點群集。並提出以複合型模式建置區域淹水預測系統：控制點之單點BPNN預測模式、線性點之線性迴歸預測模式、非線性點之BPNN多點淹水預測模式。本文選定高雄縣鳳山市與彰化縣員林鎮為研究案例，前者提供24小時一日累積降雨淹水資料；後者則輸入1-24小時降雨與淹水深之時序資料，此兩類不同的淹水資料型態，先由鳳山市的K-means分析結果證明聚類分析能有效地區分出局部地區之淹水特性；更將此方法應用於員林鎮，並以複合型模式建置區域淹水預測系統，其結果驗證(1)聚類分析能有效區分淹水資料間的線性-非線性關係，有助於區域淹水資料的分類；(2)複合型區域淹水預測模式可有效地掌握二維淹水模式模擬淹水之趨勢。

關鍵字

倒傳遞類神經網路、區域淹水； K-Means聚類分析；線性迴歸模式；淹水推估

並列摘要

The typhoon events usually cause downstream flooding in Taiwan. Estimation the flood depths and extent may provide the flood inundation information that will be helpful to deal with some contingencies, then alleviate the risk and losses of the flood disasters. The conventional simulations of flood inundation extent need a huge amount of data and computing time by using a series of numerical models. The study proposes the hybrid models to build the regional flood inundation estimation model. In order to figure out the causes of the flood inundation, the correlation analysis and factor analysis are used to explore the relationship between flood inundation influence factors; then K-means clustering is used to categorize the data points of the different flooding characteristics and to find the control point in each flooding group. The hybrid models are composed of three types of models in each flooding group: a single back-propagation neural network (BPNN) for each control point, the linear regression models for the linear grids and a multi-grid BPNN for the nonlinear grids. Two study areas, Fonshang city, Kaohsiung County, and Yuanlin township, Changhua County, are tested for evaluating the practicability and accuracy of the proposed approach. The results show that the proposed hybrid models can accurately and timely estimate the simulated flood inundation extents and depths.

並列關鍵字

Back-Propagation Neural Network ； Regional Flood Inundation ； K-means Clustering Analysis ； Linear Regression Mode l ； Flood Inundation Estimation

參考文獻

22. 張國強、林杰熙，2007，類神經網路於降雨-淹水預警之應用，第十一屆海峽兩岸水利科技交流研討會。

1. Ahmad S. and Simonovic SP. 2005. An artificial neural network model for generating hydrograph from hydro-meteorological parameters. Journal of Hydrologic Engineering 315：236-251.

2. Aknbi A.A. and Katopodes N.D. 1988. Model for Flood Propagation on Initially Dry Land. Journal of Hydraulics Engineering ASCE 114:689-705.

3. Balloffer A., Scheffler M.L. 1982. Numerical Analysis of the Teton Dam Failure Flood. Journal of Hydrologic Research 20:317-428.

4. Bates P.D., Stewart M.D., Siggers G.B., Smith C.N., Hervouet J.M., Sellin R.H.J. 1998. Internal and external validation of a two dimensional finite element model for river flood simulation. Proceedings of the Institution of Civil Engineers, Water Maritime and Energy. 130：127-141.

國際替代計量

以複合型模式分析區域淹水潛勢

全文下載

主題瀏覽