本研究首先介紹結合類神經網路(ANN)及一階可靠度法(FORM)或蒙地卡羅模擬法(MCS)之可靠度分析技術(ANN-basedFORM、ANN-basedMCS)。接著以臺北市一軟弱地盤深開挖設計的可靠度分析為例，完整說明整個可靠度分析流程。此結合ANN、FORM或MCS技術之可靠度分析模式，在系統反應之模擬、計算效率之提昇，以及分析精度之改善上，相較於大地工程慣用之簡化法（如一階二次矩法，FOSM）具有非常明顯之優勢。透過案例研究與探討，本文具體提供了深開挖工程之可靠度設計（或性能設計）一個可資參考，且簡易、可行的落實方法，以及具風險觀念的安全量化評估指標。 In this study, a reliability evaluation method, integrated with artificial neural network (ANN) and first-order reliability method (FORM) or Monte-Carlo simulation (MCS), is explored. By performing a case study on the reliability of deep excavation within soft ground, an analysis procedure for reliability analysis is proposed. The evaluation model of ANN-based FORM or ANN-based MCS is superior to traditional reliability method, such as first-order second-moment method (FOSM), in view of many aspects, such as system modeling, computational efficiency, and analysis precision. Based on these methods, the reliability of different serviceability performance or limit states for braced excavation problems can be assessed easily, efficiently and accurately. The effectiveness of ground modification measures, once needed, against excavation failure can also be evaluated quantitatively. Hence, it will supply the deep excavations engineering with a powerful tool for safety evaluation when using the established reliability-based risk analysis and design method in this research. And also, excavations safety in metropolis will be guaranteed definitely.
Relation:
中國土木水利工程學刊=Journal of the Chinese Institute of Civil & Hydraulic Engineering 19(4)，頁541-555
