類神經網路在南横公路邊坡破壞潛能之預測

An Evaluation of Highway Slope Failure Using Artificial Neural Networks

李宗霖(T. L. Lee) ； 林宏明(H. M. Lin) ； 盧育聘(Y. P. Lu)

《立德學報》 2卷1期 (2004/12) Pp. 97-115

https://doi.org/10.7080/JLU.200412.0097

摘要

近年來，公路邊坡的破壞已陸續引起許多字家學者及工程師廣泛的注意。一般而言，公路邊坡破壞受到許多因子的影響，如坡度、坡高、地質材料、地質構造、土地利用、雨量及地震等等。傳統公路邊坡破壞的評估方法中，如加權評估方法乃根據各因子間簡單之線性關係來進行危險度之評估。然而，每個因子對邊坡崩壞的影響非屬於線性關係，因此較無法反應出實際狀況，其可信度有待進一步之探討。本研究採用之類神經網路，因其具有高度的學習及處理非線性問題的能力，可用來評估邊坡崩壞的可能性。文中並以南橫公路為例，經研究結果顯示，類神經網路可相當準確地預測公路邊坡崩壞潛能的機率。

關鍵字

類神經網路 ； 邊坡破壞

並列摘要

In recent years, highway slope failure has been widely studied by geotechnical engineers. However, Conventional investigations for determining slope failure have focused on the linear relationships among many factors, such as slope angle, slope height, material, construction, rainfall, earthquake and so on. In fact, this problem is still a complex nonlinear relationship. This paper presents an application of an artificial neural network for assessing slope failure using these factors. On site slope failure data for Taiwan’s Highway 20 (South-Cross Highway) and the Highway 18 (A-Li-San Highway) were used to test the performance of the artificial neural network model. The results indicate that the artificial neural network can efficiently estimate slope failure potential using the major factors.

並列關鍵字

artificial neural network ； slope failure