| 摘要: | 高頻地動模擬為近年來地震工程上重要目標之一，其必須考量震源、路徑及場址效應，目前，隨機式地動模擬技術於台灣已可有效考慮此三項效應並對於區域內之地震進行符合物理假設之頻譜及地表峰值加速度（Peak ground acceleration, PGA）預估（Huang et al., 2017），其結果已可與傳統使用統計模型之地動預估模式（Ground motion prediction equation, GMPE）達到相似預估成效，且較具物理意義。但對於震度較大時之非線性土壤效應，目前仍尚未找到合適的方式進行合理預估，前人研究對於非線性土壤效應曾嘗試定性並定量描述頻率域之行為，包括顯著頻率降低及壓抑放大，但對於實際應用於地動預估而言仍不能夠套用。而大地工程領域常用之模擬非線性土壤效應之方式為等效線性模擬法（SHAKE, Schnabel et al., 1972），其對於已知詳細地層介面速度及材質之淺井構造，套用合適之應力-應變關係曲線後，即可經由解波傳方程式之方式進行線性及非線性之模擬預估，但此方法對於較大構造（太深之構造）或層面過多時之解析力較差。本計畫即嘗試結合此二高頻模擬方法之優點，第一年先運用台北盆地之井下地震儀陣列進行此模擬方法之驗證，第二年則以2016年美濃地震為例，將第一年建置之模擬方法應用於對於非線性強之地動記錄進行預估，以期應用於未來可能發生之大地震事件。 ;High frequency ground motion simulation was quite important for engineering seismology in recent years, including source, path and site effect should be considered to get precisely prediction. Currently stochastic ground motion simulation technique could simultaneously consider abovementioned three effects in Taiwan. Which could predict earthquake spectrum and peak ground acceleration reasonable and all of the parameters used in the technique were from physical assumption (Huang et al., 2017). The simulation could carry physical meaning and it reached similar prediction level with traditional ground motion prediction equation (GMPE). However, nonlinear soil response would inference ground motion a lot when increasing of intensity level and previous studies could not reasonable predict till now. Previous studies once tried to quantitatively and qualitatively describe soil nonlinearity in frequency domain including dominant frequency drop and de-amplification, but application for ground motion prediction was still not suitable. Meanwhile, there’s an equivalent linear simulation technique (SHAKE, Schnabel et al., 1972) could consider soil nonlinearity problem in geotechnical engineering filed. While velocity structure, geological material and suitable stress-strain curve were well investigated and constructed for shallow borehole system, linear and nonlinear ground motion simulations could be done from solving wave propagation equation. However, the equivalent linear technique had some limitations of deeper structure or multiple layers. Thus, in this proposal, the advantages from both simulation techniques will be combined to solve nonlinear soil response for future possible large earthquakes. Downhole records from borehole array in the Taipei basin will be verified for this newly combined technique in first year. Finally, while the ground motion prediction ability of the technique was prove to provide reasonable nonlinear strong ground motion, an example earthquake will be used to check it in highly nonlinearity records during 2016 Meinong, Taiwan earthquake. |
