姓名 |
蔡元耀(Yuan-Yao
Tsai) |
電子郵件信箱 |
E-mail 資料不公開 |
畢業系所 |
營建工程系碩士班(Department
and Graduate Institute of Constrction Engineering) |
畢業學位 |
碩士(Master) |
畢業時期 |
93學年第2學期 |
論文名稱(中) |
類神經網路應用於現地深開挖水平地盤反力參數推估之研究 |
論文名稱(英) |
Application of
Artificial Neural Network on Estimating Horizontal Subgrade Reaction
Coefficients for Deep Foundation Excavations |
檔案 |
本電子全文僅授權使用者為學術研究之目的,進行個人非營利性質之檢索、閱讀、列印。 請遵守中華民國著作權法之相關規定,切勿任意重製、散佈、改作、轉貼、播送,以免觸法。
|
論文使用權限 |
校內外完全公開 |
論文語文/頁數 |
中文/109 |
摘要(中) |
本研究利用類神經網路倒傳遞模式,以案例的現地土壤鑽探資料及土壤參數,進行TORSA程式之正算分析,並依此分析結果建立類神經反算模式,輸入現地監測之壁體變位反算,推估符合現地監測值之土壤設計參數,以預估下一階開挖施工造成擋土壁體變形量。 研究結果顯示利用TORSA程式進行壁體位移分析,使用類神經網路反算之Kh分析時,壁體變位相對誤差小於30%評比為優者佔驗證案例之63.8%,優於使用經驗公式建議之Kh分析的24.6%。經由本分析模式反算Kh,進行壁體位移分析能有不錯之結果,實務上使用應增加工程案例驗證,提升類神經反算模式之可靠性及普遍性,對整體深開挖工程有較佳幫助,進而提供施工單位作為施工安全參考,降低工程失敗的機率。 |
摘要(英) |
This research mainly
bases on the Back-Propagation Network model of artificial neural
network. The data of in-situ exploration results soil and engineering
properties of soils are used for processing the analysis of TORSA. On
the basis of the analysis conclusion, the model of back analysis of
artificial neural can be built. After obtaining the design coefficients
of soil which match the field monitoring data resulting from
back-analysis, the prediction of displacement of construction excavation
retaining wall can be made for the next excavation stage. The
results of this research show that when using TORSA to process the
analysis of displacement, the relative deviation between displacement
analysis and monitoring result less than 30% is only 24.6% of examined
cases. In addition, the Kh(horizontal subgrade reaction)obtained from
back-analysis using artificial neural network with relative deviation
less than 30% is 63.8% for all examined cases. A useful result can be
generated by using the analysis model developed in this study to predict
the displacement of retaining wall at different excavation stages. In
general the result of this study can be used as a valuable reference and
can provide a safety guideline for future construction excavations to
reduce the risk of excavation failures. |
關鍵字(中) |
回饋分析
TORSA
類神經網路
深開挖 |
關鍵字(英) |
artificial neural network
deep excavation
back-analysis
TORSA |
指導教授 |
張子修
|