HEC-6模式判別與敏感度分析研究

Abstract

本文旨在以50組合成資料對美國陸軍工兵團水文工程中心所發展的 HEC-6 輸砂模式進行模擬,並由模擬結果選取欲分析的模式輸出，對寬深比、起 始河道坡度、中值粒徑、砂石幾何標準偏差、福祿數、雷諾數六基本河道 參數，建立一敏感度分析架構，分析基本參數對模式輸出的敏感度，在模 擬過程中對於異常沖刷的現象亦加以探究，並做判別分析，以區分造成異 常沖刷的參數群。對於模擬結果合理性之探討，主要是以兩種不同的時間 間隔，模擬 50組合成資料，依其模擬結果之底床變化有無振盪現象或異 常刷深現象加以判定，並試由判定之結果找出較重要之影響參數。此外並 依50組資料之模擬結果建立一套無因次迴歸模式且進行六基本河道參數對 於模式輸出的敏感度分析。 The purpose of this study is to perform the validation and sensitivity analysis for HEC-6 model(Thomas,1991) by using 50 sets of synthesized data. A framework of sensitivity analysis has been established to examine the effect of six selected key fluvial parameters including width-depth ratio、channel slope、 medium sediment size、sediment geometric standard deviation、 Froude number、Reynoids number. The anomalous excessive scouring phenomenon which was observed in simulation process on the model's outputs has also been extensively studied. Discriminant analysis is used classify the six key fluvial parameters. The simulations have been performed for two different time intervals on the basis of 50 sets of synthesized data generated. The rationality of simulation results was defined by examining whether the oscillation of bed level change and/or the excessive scouring phenomenon appeared or not. From the discriminant analysis one can classify the influence level of the key parameters to the model outputs. Furthermore, based on the simulation results of 50 sets of synthesized data a nondimensional regression model was established and the model's sensitivity analysis was also performed.