頁籤選單縮合
題 名 | 以類神經網路評估鋼筋混凝土結構之損壞=Damage Assessment of Reinforced Concrete Structures by Artificial Neural Network |
---|---|
作 者 | 蔡中暉; 徐德修; | 書刊名 | 中國土木水利工程學刊 |
卷 期 | 10:1 1998.03[民87.03] |
頁 次 | 頁31-38 |
分類號 | 441.557 |
關鍵詞 | 倒傳遞神經網路; 鋼筋混凝土結構; 損壞評估; 自然頻率; Back-propagation network; Reinforce concrete structure; Damage assessment; Natural frequency; |
語 文 | 中文(Chinese) |
中文摘要 | 本文嘗試引用倒傳遞神經網路應用在"診斷"及"識別"方面之理論,來建立評估 鋼筋混凝土結構損壞程度及損壞位置之模式。在網路架構內,文中將隱藏層處理單元之集 成函數加設一適量之"放大係數",建立識別雛型應用在所舉三個實例上均能使網路加速收 斂並分別獲得良好之識別效果。因此配合計算機平行運算之技術及模糊集理論之應用,將 此評估模式推展應用到大型鋼筋混凝土結構,將是結構物損壞評估一個極具研究潛力的方 向。 |
英文摘要 | A method for assessing the reinforced concrete structural damage is developed using the back-propagation network, which can be applied to the diagnosis and identification problems. A magnification factor is introduced to the summation function of the hidden layer unit to speed up the convergence. A prototype of identification is developed and applied to damaged reinforced concrete cantilever structures. Results show that the proposed method can speed up the convergence of the network in searching for the satisfied identification goal. By combining the technique of parallel computation and fuzzy theorem, the proposed algorithm of reinforced concrete structural damage assessment has the potential of being applied to real structures for damage assessment. |
本系統中英文摘要資訊取自各篇刊載內容。