使用遞迴子空間識別法實現線上系統識別之試驗研究

地震作為台灣最具代表性的自然災害，每每皆造成災難性的破壞，包括大規模的人員傷亡與財產損失，為了減輕這種災害，結構健康監測技術應在地震事件中提供連續且自動的診斷結果，並憑藉相關系統能提供即時的警報訊息，有鑑於此，本研究發展能夠即時識別結構狀態、追蹤模態參數、提供損傷訊息，並幫助震後評估和復原的線上系統識別技術。考量地震預警技術近年的蓬勃發展，連續且自動的結構健康監測技術可透過融合來自地震預警和系統識別技術的資訊，預期在偵測到地震後利用遞迴方式達成上述目標。藉此，本文提出基於遞迴方式的子空間識別法，透過理論推導與試驗驗證來核實其可行性，並針對線上運行的細節進行研討論證，其中利用三層樓與四層樓實尺寸鋼結構的振動台試驗來驗證本法的即時性與有效性，驗證使用遞迴子空間識別法實現線上系統識別之的可行性。

關鍵字

遞迴子空間識別法；線上系統識別；結構健康監測

並列摘要

Earthquakes, as one of the most catastrophic natural disasters in Taiwan, often caused destructive damages including large-scale loss of human lives, public/private properties, and so on. To mitigate such damages, structural health monitoring (SHM) techniques should be continuous and autonomous during seismic events so that the SHM system can provide timely warnings. In this regard, real-time or online system identification (SI) techniques can be used to identify the state of the structure, track the modal parameters, provide the warning message about damage, and help the post-earthquake reconnaissance and rehabilitation. Considering that the earthquake early warning (EEW) techniques flourishes this decade, a continuous and autonomous SHM technique can be preliminarily achieved by fusing the information from EEW and SI. To be specific, the recursive algorithm is proposed and, in this study, the recursive subspace identification (RSI) algorithm is applied to investigate the time-varying dynamic characteristics after the seismic event is detected. The proposed algorithm is carefully derived and examined; moreover, the suggestion for implementation is described and discussed. For the verification of the RSI-based SHM, two laboratory tests, the shaking table tests using a three-story steel frame and a four-story steel frame, are conducted and the proposed algorithm is applied to identify the modal parameters reclusively.

並列關鍵字

recursive subspace identification ； online system identification ； structural health monitoring

主題瀏覽