| 摘要: | 液化引致之沉陷及側向擴張會造成建物、橋梁及地下管線的嚴重損壞。在規模較大的地震中皆可觀察到由液化引致沉陷及側向擴張造成的重大損失。例如，在2010-2011紐西蘭基督城地震中，約有兩萬餘戶房屋受損，液化引起的總損失約為150億美金。精準的預測液化引致沉陷及側向擴張變形對於防範液化災害至為重要。目前預測液化引致地表變形之經驗模式主要有三個缺失。其一，現有經驗模式是根據早期的個案發展而成，並無納入近期地震的案例。其二，過去個案存在大量不確定性，但模式發展過程中鮮少考慮到這些不確定性。其三，過去很少有關模式本身不確定性的研究。基於上述難題，準確的預測液化引致之地表變形仍然是一個困難的挑戰。因此亟需建立將所有不確定性納入液化引致之地表變形資料庫及模式。在本研究中，經驗模式的率定是根據嚴謹的機率理論及貝氏定理，並考慮各項不確定性。研究目的包括：（1）建立一個新的液化引致之地表變形資料庫，並明確考慮隱含在每個案例中的參數不確定性及模式不確定性；（2）利用新的資料庫發展液化引致地表沉陷模式；（3）發展液化引致側向擴張變形模式。為推廣所建立的模式，將提供使用指南及範例，以供工程師參考。 ;The settlement and the lateral spreading are the two major forms of liquefaction-induced ground movements that have caused great damage to infrastructure such as buildings, bridges, and lifelines in previous earthquakes. For example, severe damages to buildings and roadways from liquefaction-induced settlement and lateral spreading were observed in the 1999 Chi-Chi, Taiwan earthquake (Chu et al., 2006), the 2007 Niigata, Japan earthquake (Apostolakis et al. 2007), the 2010–2011 Christchurch, New Zealand earthquakes (Bowen et al., 2012), and the 2011 Tohoku, Japan earthquake (Ishihara et al., 2011; Yasuda et al., 2012). In particular, in 2010-2011 Christchurch, New Zealand earthquakes, also known as the Canterbury earthquake sequence, more than 20,000 residential homes were damaged beyond repair. The total loss due to soil liquefaction in the Canterbury earthquake sequence was estimated at US$15 billion (Cubrinovski et al., 2014).Realistic prediction of liquefaction-induced settlement and lateral spreading is crucial to the mitigation of liquefaction hazard. There are three major limitations regarding the existing empirical methods for prediction of liquefaction-induced ground deformation. Firstly, the models were developed based on relatively old case histories, and the valuable information provided by recent earthquakes are not considered. Secondly, there are substantial amount of uncertainties in the case histories, but such uncertainties are seldom considered in the model development stage. Thirdly, the model uncertainty associated with these models were seldom reported. Thus, new models for prediction of liquefaction induced ground deformation are needed. The objectives of this research are to: (1) compile a state-of-the-art ground deformation database with explicit consideration of uncertainties involved in the database; (2) develop new models for prediction of liquefaction-induced ground settlement using the new database, and (3) develop new models for prediction of liquefaction induced lateral spreading based on the new database.Strong research efforts will be made to accomplish these objectives in three years. The database creation will emphasize the explicit treatment of uncertainties. Model calibration and development will involve formulations and simulations using Bayes’ theorem, maximum likelihood principle, and the Bayesian model comparison method. To facilitate the application of the new models, guideline and illustrative examples for applying the validated models will be prepared. Open-access codes will be developed to automate the entire process. The anticipating impact of this three-year research project is significant, as there is a pressing need to create procedures to estimate the liquefaction-induced settlement and lateral spread that are hazard-consistent and incorporating all the uncertainties. The outcome of the project will also provide a basis/reference for other sub-projects. In many past cases, the degree of damage of the superstructure and lifelines is closely related to the liquefaction induced ground deformation. The results obtained from this project will also provide the basis for assessing the degree of damage of superstructures and lifelines. Overall, the proposed research is a step forward on advancing the resilient, hazard-consistent earthquake design of buildings and infrastructures in Taiwan, leading to safer, more economical and more efficient designs. |
