|
In this study, the typological building fragility curves are developed base on the complete building damage records collected after the 1999 Chi-Chi Earthquake in Taiwan. These fragility curves are further applied in selecting routes and planning network of urban emergency earthquake. An aspect of the building fragility curves, a grid-based clustering maximum likelihood estimate (grid-based method), in combination with a grid-based cluster analysis procedure and with a novel maximum likelihood estimate, is proposed to derive fragility curves for 16 building typologies in Taiwan. This new grid-based method generates lower-deviation vulnerability data for reducing the dispersion of datasets than does the traditional district-based method. The proposed grid-based method has three analysis models, including binomial distribution, multinomial distribution (Method 1), and multinomial distribution with a common log-standard deviation (Method 2). The results of the grid-based method show that: (1) the fragility curves are more stable, unsusceptible, and convergent than those from the district-based method; (2) the fragility curves can reasonably expressed vulnerability of buildings thus applicable to the development of building fragility curves for wide-regional damage records; (3) Method 2 provides a more reasonable vulnerability of building thus the common log-standard deviation is a better choice to derive the empirical fragility curves; (4) the fragility curves have acceptable prediction performance even though only two levels of damage in the 1999 Chi-Chi Earthquake. Above results demonstrate that the developed fragility curves can reasonably be implemented for estimating earthquake loss and assessing seismic risk in the future. An aspect of the emergency earthquake routes, a road seismic vulnerability curves analysis is proposed to express the exceeding probability of the road-section block as a function of a specific earthquake intensity measure. The road seismic vulnerability curves are further using in analysis of low disruption risk for emergency earthquake routes (Low-DREER), in combination with road-section block risk analysis and with network analysis. The results of the road seismic vulnerability curves show that: (1) the road-sections with different buildings have different vulnerabilities of road-section block which related the composition of the number of buildings and its typologies; (2) the more number of buildings induced higher and rapidly increasing vulnerabilities of road-section block. An aspect of Low-DREER, the proposed accumulative route risk values can appropriately describe routes disruption risk, which caused by probable road-section block risk induced from the earthquake-induced building collapse. The results of Low-DREER show that a disaster prevention region should have a Primary Low-DREER which has a lowest accumulative route risk values and is able to communicate outside the access road intersections. Finally, the former metropolitan area in Taichung as a case study, the study proposes a method of bi-stage selection for earthquake emergency network. For urban earthquake emergency network in future, the method can be used in selecting primary earthquake emergency routes and applied in suitability assessment of the various types of disaster prevention facility.
|