臺灣博碩士論文加值系統

台灣屬狹長型島嶼。河川眾多，因此跨河結構物成為台灣重要的交通樞紐。然而，結構物會阻滯河道造成水流流況改變，加上台灣地勢陡峭、河短流急、降雨量時空上分布不均，易形成河床局部沖淤現象，嚴重者將會造成跨河結構物(如橋墩)的基礎嚴重掏空進而破壞，終致落橋。我國的橋墩結構物種類繁多，其中以沉箱基礎與群樁基礎為主要的橋墩結構類型，為了評估能夠有效瞭解目前擁有的預測方式對於各種橋墩情況的適用性，蒐集不同橋墩類型，水深，以及覆土深的實驗數據輔以本研究的實驗資料，分析出適用於各種情況下的預測方式與合適性。
本研究採用兩個主要的沖刷預測方式，一為Melville & Coleman於2000年時所著作的”Bridge Scour”中所提出的公式，以即美國聯邦高速公路署的Hydraulic Engineering Circulars NO.18(簡稱HEC-18)，HEC-18光是近幾年共有五個版本，本研究採用第五版(2012)，與傳統的Bridge Scour一同進行分析比較，探究這兩者方法中對於沖刷機制影響的因素，使用最小平方支持向量機(LS-SVM)計算兩公式參數迴歸的可靠度來比較彼此的差異性，並以此為基礎，重新最佳化其中的公式，目的為使用簡易的沖刷預測公式的同時，不失去其準確度。最終結果以誤差評估MAPE的標準來看，Melville & Coleman(2000)與HEC-18的MAPE分別為102.7564和57.4965，而最佳化的公式結果，MAPE=28.96542，有效同時達到準確以及精簡的預其效果。雖然所分析的結構物皆為簡易型的非均勻橋墩，並無加入河流沖刷時的沖刷歷程考量，但本研究所建立的分析演算結果，仍對國內水工結構物之分析過程具有重要的參考價值。

Rivers in Taiwan often have steep slope resulting in a rapid flow condition. The situation is more serious during the typhoon or storm periods. As a result, severe pier scours and exposures of bridge foundations occur frequently, which may endanger the safety of the bridges.
In this study, local scour around complex piers under steady clear-water condition was studied experimentally for a variety of configuration, including different sizes and shapes of complex piers. A total of 176 experiments were collected. The predictions of the scour depths has been improved in which the methods of Melville & Coleman (2000) and Hydraulic Engineering Circular No. 18. Least Squares Support Vector Machines (LSSVM) are used as the basis. Two methods have been proposed to improve the accuracy of the scour prediction. We first used LS-SVM to build a surrogate model in which all the important factors described in HEC-18 and Melville & Coleman’s approach are included. The second proposed method utilized the existing optimization tool box to find the optimal coefficients in Melville & Coleman’s approach based on the collected data. Both proposed methods show a better accuracy of scour depth prediction compared to the previous approaches. The proposed methods, is therefore useful for the bridge safety evaluation, especially, if uncertainty is considered.

誌謝 Ⅰ
摘要 II
Abstract III
目錄 IV
圖目錄 VIII
表目錄 XI
第一章 緒論 1
1.1研究背景 1
1.2 研究動機及目的 2
1.3研究方法 2
1.4論文架構 3

第二章 文獻回顧 5
2.1 沖刷原因與分類 5
2.1.1一般沖刷(general scour) 5
2.1.2束縮沖刷(contraction scour) 6
2.1.3局部沖刷(local scour) 6
2.2 影響橋墩沖刷因子分類與探討 8
2.2.1 橋墩幾何性質 9
2.2.2 水流特性 11
2.2.3 底床質特性 13
2.2.4 時間影響因子 13

第三章 研究方法 14
3.1 Melville & Coleman(2000) ”Bridge Scour” 橋墩沖刷深度公式 14
3.1.1 等效迎水面寬度 be 15
3.1.2 Kyb(水深－橋梁形狀影響因子) 18
3.1.3 KI(水流強度修正係數) 19
3.1.4 Kd(底床質特性修正係數) 21
3.1.5 Ks(橋墩形狀修正因子) 22
3.1.6 Kθ(水流攻角修正係數) 24
3.1.7 Kt(時間因子修正係數) 25
3.2 Hydraulic Engineering Circular No.18(HEC-18)橋墩沖刷深度公式 27
3.2.1 HEC-18 沖刷基本公式：(簡易墩柱沖刷) 27
3.2.1.1 寬柱在淺水狀態中沖刷的矯正因子 Kw 29
3.2.2非均勻橋墩沖刷深度計算流程 29
3.2.2.1 判定橋墩結構受影響之元素 31
3.2.2.2 決定ys pc(樁帽部分的沖刷深度) 34
3.2.2.3 決定ys pg 群樁部分的沖刷深度 37
3.2.2.4 HEC-18修正(Revised)公式 43
3.3 預測沖刷公式誤差探討與改良建議 44
3.3.1 試驗設置 45
3.4 反應曲面法(Response Surface Method，RSM) 49
3.4.1 支持向量機(Support Vector Machine, SVM) 49
3.4.2 最小平方支持向量機(Least Square-Support Vector Machine, LS-SVM) 53
3.4.3 評估指標 54

第四章 案例分析與結果 58
4.1 局部沖刷深度預測分析 59
4.1.1 不同橋墩種類與沖刷狀況比較 62
4.2誤差因子探討 65
4.2.1 Melville& Coleman(2000)預測計算部分 65
4.2.2 局部沖刷深度(ys)與覆土深度(Y)之關係 66
4.2.3不同覆土深度(Y)比較 71
4.3.3 HEC-18預測計算部分 74
4.3.4 機器學習 (Machine Learing) 75
4.3.4.1參數設定 76
4.3.4.2 LSSVM預測結果 77
4.4 be修正(Revised)公式 80
4.5最佳化預測公式 84

第五章 結論與建議 94
5.1結論 94
5.2建議 95

[1] L.A. Arneson, L.W. Zevenbergen, P.F. Lagasse, P.E. Clopper (2012).” Evaluating scour at bridges” Publication No. FHWA HIF 12-003, Federal Highway Administration, Washington,D.C. (2012).
[2] Melville, B. W., and Coleman, S. E. (2000). Bridge scour, Water Re- sources Publications, Littleton, Colo.
[3] Melville, B. W., and Raudkivi, A. J. (1996). “Effects of foundation ge- ometry on bridge pier scour.” J. Hydraul. Eng., 122(4), 203–209.
[4] 林呈，(1998)，「本省西部重要河川橋梁橋基災害分析與橋基保護工法資料庫系統之建立」，交通部運輸研究所專題研究計畫成果報告
[5] Imamoto H., & Ohtoshi K., (1987). "Local Scour around a Non-uniform Circular Pier", Proceeding, IAHR Congress, Lausanne, Switzerland, 304-309.
[6] Raudkivi, A. J., (1986). "Functional trends of scourat bridge piers", Journal of Hydraulic Engineering,ASCE, 112(1): 1-13.
[7] Melville, B. W. and A. J. Raudkivi, (1977). "Flow characteristics in local scour at bridge piers," Journal of Hydraulic Research, 15(4): 373-380.
[8] Raudkivi, A. J. and R. Ettema, (1983). "Clear-water scour at cylindrical piers", Journal of Hydraulic Engineering, ASCE, 111(4): 713-731.
[9] Jain, S. C.(1981) ”Maximun Clear-Water Scour around Circular Piers” J.Hydr. Engrg., ASCE Vol. 107.No. 5, 611~626
[10] Melville, B. W.(1997),”Pier and Abutment Scour：Integrated Approach “J.Hydr. Engrg., ASCE Vol. 123, No. 2, 125~136
[11] Melville, B. W., and Sutherland, A. J. (1988). “Deign methods for local scour at bridge piers.” J. Hydraul. Eng., 114(10), 1210–1226.
[11] Ettema, R., Melville, B. W., and Barkdoll, B. (1998). “Scale effect in pier-scour experiments.” J. Hydraul. Eng., 124(6), 639–642.
[12] Melville, B. W., and Chiew, Y. M., (1999). “Time Scale for Local Scour at Bridge Piers” J. Hydr. Engrg. ASCE Vol. 125, No. 1, 59~65,
[13] Salim, M. and Jones, J.S.( 1996).” Scour around exposed pile foundations” North American Water and Environment Congress, A.S.C.E., Anaheim, U.S.A., June.
[14] Smith, W.L., (1999), "Local Structure-Induced Sediment Scour at Pile Groups," M.S. Thesis, University of Florida, Gainesville, FL.
[15] B. Ataie-Ashtiani, Z. Baratian-Ghorghi, and A. A. Beheshti,. (2010)“Experimental Investigation of Clear-Water Local Scour of Compound Piers.” J. Hydraul. Eng., 136(6): 343-351.
[16]Support Vector Machine WIKIAPEDIA
http://en.wikipedia.org/wiki/Support_vector_machine
[17] J.A.K. Suykens, J. Vandewalle, 1998, ”Least Squares Support Vector Machine Classifiers”.
[18] 林冠中，2005，「漸進式支持向量機於人臉辨識之應用」，國立成功大學資訊工程學系碩士論文
[19] Amaury Lendasse, Yongnan Ji, Nima Reyhani, and Michel Verleysen, “LS-SVM Hyperparameter Selection with a Nonparametric Noise Estimator”, Neural Network Research Centre, Helsinki University of Technology.
[20] J. Kennedy and R. C. Eberhart, “Particle swarm optimization, (1995).” in: Proc. IEEE Int. Conf. on Neural Networks, Perth, Australia, vol. 4, 1942-1948.
[21]Parola, A. C., Mahavadi, S. K., Brown, B. M., and El-Khoury, A. (1996). “Effects of rectangular foundation geometry on local pier scour.” J. Hydraul. Eng., 122(1), 35–40.
[22] Coleman, S. E. (2005). “Clearwater local scour at complex piers.” J. Hydraul. Eng., 131(4), 330–334.
[23]彭明正，2008，「變量流作用下非均勻橋墩之局部沖刷量測與模擬」，中興大學土木工程學系碩士論文。
[24]Sheppard, D. M., and Glasser, T. (2004). “Sediment scour at piers with complex geometries.” Proc., 2004 2nd Int. Conf. on Scour and Ero- sion, World Scientific, Singapore.
[25]Sheppard, D. M., and Renna, R. (2005). Florida bridge scour manual,Florida DOT, Tallahassee, Fla.
[26] T. Chai1, and R. R. Draxler, (2014). Root mean square error (RMSE) or mean absolute error (MAE)? –Arguments against avoiding RMSE in the literature.