結構動力歷時分析之地震強度調整方法研究

Abstract

由於在建築耐震設計與分析程序中，動力歷時分析可供進行更詳細的結構耐震性能評估，本研究首先介紹幾種常見用於動力歷時分析調整地震加速度歷時的方法，並且探討不同地震歷時調整方法的優缺點，綜合不同地震加速度歷時調整方法的優缺點，而發展出多振態調整法。此法建立在考量主要之前數個振態效應，使用誤差最小平方法調整考量的主要振態在自然地震與設計反應譜間的誤差，因此可以有效改善調整倍率過大或過小的缺點、使自然地震反應譜與設計加速度反應譜在主要振譜處更接近。本文先使用美國SAC Project中位於洛杉磯的三棟鋼構造抗彎構架，探討多振態調整法用於不同高度結構的結構物受震反應評估效應，發現使用多振態調整法於低、中高樓層，皆可有效降低結構物受震反應的變異性，尤其是使用在高層建築中更為有效。本研究亦探討使用不同振態數目對多振態調整法的影響，以一棟位於高雄的34層鋼構造超高層大樓為例，使用不同振態數目調整地震加速度歷時後，顯示只要多考慮一個振態，便可以有效降低結構物受震反應變異性，且當考量振態數目越多，可以有效的降低調整倍率。 本研究並探討多振態調整法與其他地震加速度歷時調整法在結構物可信度導向耐震性能評估程序中的適用性，以前述34層超高層大樓作為評估案例，使用FEMA-350, FEMA-355F, FEMA-351等耐震性能設計規範所建議之評估程序，採2500年迴歸期設計反應譜進行評估，結果顯示由於多振態調整法可以有效降低結構物受震反應，而使評估結果更加合理，該34樓案例方能符合對應於2500年設反應譜的95％可信度。

Non-linear dynamic time-history analyses conducted as part of a performance-based seismic design approach often require that the ground motion records be scaled to a specified level of seismic intensity. Recent research has demonstrated that certain ground motion scaling methods can introduce a large scatter in the estimated seismic demands. The resulting seismic demands may be biased, leading to designs with significant uncertainty and unknown margins of safety. For high-rise buildings, high-mode effects can be rather pronounced. If a ground acceleration record is scaled without properly incorporating the design spectral accelerations at the significant periods of a building, it could seriously overestimate or underestimate the seismic demands. This study proposes a multi-mode ground motion scaling (MMS) method. This method takes into account the structural modal characteristics and aims to minimize the difference between the spectral responses of a given ground motion and the smoothed design response spectrum at the first few modes. In this research, various ground motion scaling methods are investigated using a set of nine model buildings─ referred to as the SAC buildings, each subjected to 20 specific ground motions. This study compares the effectiveness of the MMS method with the ground motion scaling procedures prescribed by current seismic codes or Sa(T) method proposed by Shome and Cornell in 1998. It is illustrated that the MMS method is effective in reducing the scatter in peak seismic demands computed from both the response spectrum analysis (RSA) and nonlinear response history analysis (NLRHA) procedures. It is demonstrated that historical ground acceleration records can be conveniently scaled with only a few more vibration modes considered for high-rise buildings. It is found that the MMS method is effective in minimizing the scatter in peak seismic demand estimates for tall buidlings. In addition, the effects of different scaling methods for a new confidence-based demand and resistance factor seismic evaluation procedure are investigated, respectively. Again, it is found that the MMS method could reduce the scatter of seismic demands and make the evaluation more effective.