Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/14866
標題: 邊坡破壞發生之預測方法研究-以梨山地滑地為例
Study on time prediction of slope failure -An illustration of lisan landslide
作者: 郭力行
shen, kuo li
關鍵字: time prediction;邊坡破壞;landslide;預測
出版社: 土木工程學系
摘要: 
梨山地區為一古老崩塌地,自民國79年四月中旬因連日豪雨,造成台七甲線73K+500處路基崩踏而告交通中斷。有鑒於地滑地整治規劃至今,對於邊坡滑動的管理基準值方面一直採用日本高速道路調查會所提出的管理基準值,而未有一專為梨山現地情況進行分析而提出的管理基準值,在適用性及安全性上恐有諸多疑慮,因此本文嘗試以邊坡滑動的速率對梨山歷年來所得到的監測資料進行分析,進而訂定出一個能符合梨山現地情況的邊坡滑動管理基準值來,並配合類神經網路對地滑地邊坡破壞的發生進行預測。
研究中以齋藤迪孝之理論分析梨山歷年來地表位移之應變速率,並以福囿輝旗之理論分析梨山歷年之地表位移速度,將兩者所得的結果進行分析討論,訂定出一符合梨山現地情況的地滑管理基準值,於地表位移應變速率達到 10-6(%/分)為警戒,以福囿法預測破壞時刻與最後觀測時間小於5小時為疏散的標準值,並運用類神經網路配合相關度迴歸分析,學習滑動體內部因地下水升高所引致的地表位移間之內在對應關係,對邊坡破壞的發生進行預測。

Lisan is an old landslide area,the grand-scale landslide which occurred in April 1990 caused road service interrupted .Because the standards of danger of slop is from japan highway study orgnization,therefore it's accuracy for lisan area is doubted.So this search try to analyze the data getting from auto monitor station with slop sliding rate,and setting a standards of danger to fit the situation of Lisan,and use the neural network model to predict the time when landslide is happened.
This search try to analyze the strain rate of landslide with the method of Saito and the rate of landslide with the method of Fuzokono,through these analyzing to set a standards of danger to fit lisan situation,and have a conclusion that if the strain rate of landslide up to 10-6(%/min) this is a alert situation ,and when we use the method of Fuzokono to predict the failure time Tr<5hr we have to evacuate resident,and combine neural network and correlation regression to learn the correlation between groundwater and movement of groundserface to predict the time of failure.
URI: http://hdl.handle.net/11455/14866
Appears in Collections:土木工程學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.