部分感測器量測下辨識剪力構架結構之破壞位置

Abstract

結構物會因材料老化或受外力作用而產生損壞，所以一套不斷監測結構物健康狀況的系統是有其必要性。結構健康監測在土木工程中逐漸受到重視，而破壞位置辨識為結構健康監測重要的一環。大部分破壞位置辨識均需利用模態資訊，因此需要在所有的自由度上安裝感測器。但因為成本與安裝位置限制的關係，在現實中通常只使用少數感測器。此時就必須使用模態擴展技術來得到每個自由度上的模態資訊，舊有的模態擴展技術需要利用結構參數來進行，因為結構參數的改變對擴展結果有所影響。故本研究利用雲線擬合來擴展，達到不需要任何結構參數的擴展方法。利用matlab軟體建造數值模型，模型均為剪力構件。運用少量模態資訊進行雲線擬合，配合虛擬樓層策略與線搜尋來增加模態的精確度，再透過有效質量與雲線擬合精度來決定感測器數目。擴展結果由MAC(Modal Accuracy Criteria)來比較。研究結果顯示當結構參數差越多時，本論文提出的雲線擴展方法在低頻模態下效果較佳，且當結構參數變異性不大時擴展結果越好。再來利用雲線擬合的擴展結果進行破壞位置辨識以達到僅需要部分感測器量測下辨識結構之破壞位置，在感測器誤差在萬分之一的狀況與結構參數變異性不大下，破壞位置辨識具有高可信度。

Structures may be damaged due to ageing of material or by external force. It is necessary to acquire accurate and real-time information on the structure condition. For this purpose, structure health monitoring (SHM) for civil engineering has received considerable attention. Identify damage locations is one important research of structure health monitoring. Many damage detection method has been developed in last decade. Most of damage detection method use complete mode shapes information. In practical cases, we often use sparse sensors because of the high cost of sensor. Mode shape expansion techniques deal with these condition in order to acquire all degrees of freedom’s mode shape information. But traditional expansion techniques can’t expand well due to the change of structure parameter. This study use spline curve-fitting to expand mode shape information without using any structure parameter. Use pseudo story strategy and line search strategy to improve expanded mode shape information accuracy. Then use expanded mode shape information on damage detection. This study use matlab to simulate shear building structure. The result show that the expansion method of this study is better than traditional expansion method in lower mode. Also, this study identify damage locations successfully in the absence of sensor error.