| 摘要: | 本研究蒐集集集地震建築物震害資料與強地動測站紀錄，利用「等震度」與「等建築物母體數」的概念，透過GIS軟體的運算與分析，以建立不同強地動參數與建築物倒塌率之易損曲線，並以集集地震為模擬地震，進行建築物倒塌分佈之推估及驗證。 在「等震度」方面，分別以PGA, PGV, CAV, AI四種地動參數進行分析。首先利用Kriging內插方法完成全台強地動參數的空間分佈。在「等建築物母體數」則以(1)100,000 (2)50,000 (3)25,000與(4)12,500戶四種尺度進行探討，完成不同的強地動參數與建築物倒塌之易損曲線。另外，將倒塌建築物區分為土塊厝、磚構造、鋼構造、鋼筋混凝土、其它5類，逐一完成各類建築物之易損曲線。 本研究成果顯示：以「等震度」與「等建築物母體數」所獲得之建築物倒塌率與強地動參數的關係，資料點遠較以「行政區」獲得資料點集中，顯示本文提出以「等震度」與「等建築物母體數」的概念合理且可行。當以較小建築物母體數進行分區時，所獲得之資料點較多，建築物整體倒塌趨勢較為明顯。本文利用Logistic與Lognormal CDF方程式對資料點進行迴歸，顯示以 PGA與AI為強地動參數所建立的建築物易損曲線，不僅 較高且推估建築物倒塌之空間分佈亦較佳，故建議採用。 In this research, we collected the building damage data and strong motion data of the Chi-Chi earthquake and proposed the concept of “Iso-seismic intensity” and “Equal buildings intervals”. We construct fragility curves which were relation between strong ground-motion index and building damage rate by the GIS software. Furthermore, we evaluate the distribution of building damage based on the case of the Chi-Chi earthquake. We used four strong ground motion indices, including PGA, PGV, CAV and AI. The Kriging method is used to obtain spatial distribution of strong ground motion in Taiwan. We used four categories of the Equal buildings intervals to decide the scale of analysis region. Finally, we obtain fragility curves with respect to different strong ground motion indices. Additionally we divided building damage data into the four types, including mud-brick, masonry, reinforced concrete and so on. Then we constructed the fragility curves based on four structural types. We obtained the relation between strong ground motion indices and building damage rate by the concept of “Iso-seismic intensity” and “Equal buildings intervals”. The results of analyses show that data points are more concentrated than data points obtained by means of analysis of fragility curves based on the unit administrative districts. It appears that the concept we proposed is reasonable and feasible. The smaller equal-buildings intervals we used, the more data point we got. Therefore, the relationship of the building damage rates as function of the strong ground-motion index is apparent. In this research, Logistic and Lognormal CDF functions are used to finish the regression fitting of fragility curves. There are higher values of R2 when we used the indices of PGA and AI to analysis the fragility curves. At the same time, we obtained a good result to the estimation of the number and distribution of damaged buildings caused by the Chi-Chi earthquake. Finally, the fragility curves with respect to PGA and AI are suggested to the evaluation of future earthquake events. |
