本研究將土壤液化潛能評估法分成兩個方向來加以探討研究，第一部分是驗証傳統的簡易土壤液化評估法，並發展新的簡易土壤液化評估法，第二部分是以機率與統計的方式發展一套完整的全機率土壤液化評估法，並應用於實際工程之液化潛能風險評估。 本研究蒐集共669組的SPT-N、388組的CPT-qc及250組的震測VS現地液化與非液化案例資料，以預判準確率與至少液化安全係數誤差指標，來驗証現有液化評估法之適用性，並利用這些資料建立一套物理意義較明確的TAI液化評估法。驗証結果顯示，SPT-N法以Seed法、NCEER法及TAI法為較佳之方法；CPT-qc法以NCEER法、Juang法與TAI法為較佳之方法；而震測VS法則以TAI法為較佳之方法。 傳統的液化評估法以安全係數的大小為液化潛能的評估標準，安全係數的要求則由工程師之經驗判斷。本研究利用所蒐集的現有液化案例資料庫，以Seed’85法及一次二階矩法為基礎，將影響土壤液化評估結果之主要參數的變異性量化，建立單一地震事件之土壤液化可靠度分析法。再進一步結合地震危害度與土壤液化可靠度分析法，建立完整考量地震發生與土壤液化強度變異性的全機率土壤液化評估法。最後以價值工程的觀念，建立一套土壤液化風險評估法，期望能為液化防治工程建立一套具有風險價值意義的決策分析方法。 This research examines the existing methods of evaluating soil liquefaction potential and seeks to develop new methods of evaluation. In the first part of the study, it verifies the traditional simplified methods for evaluating soil liquefaction and develops a new simplified method. In the second part of the study, a full probabilistic method of evaluating soil liquefaction is developed and applied to liquefaction risk analysis. In total, there are 669 sets of SPT-N data, 388 sets of CPT-qc data, and 250 sets of shear wave velocity data on liquefaction and non- liquefaction case histories used in this research. Based on this data, a new simplified method for evaluating soil liquefaction called the “TAI method” is established that can use either SPT-N, CPT-qc or shear wave velocity data (Vs). Two indexes - the accuracy of prediction and the index of the at least error of safety factor - are used to verify the traditional simplified methods and compare them with the TAI method. The results show that the choice of the evaluation method used should be dependent on the type of data collected. The Seed method, NCEER method and TAI method work best for SPT data. The NCEER method, Juang method and TAI method are the best predictors for CPT data and the TAI method is the best choice for Vs data. The traditional method for evaluating soil liquefaction assesses the liquefaction potential based on the safety factor it produces. However, criterion of safety factor adopted should depend on engineering’s experience. This study presents a reliability analysis method based on the popular Seed ’85 method and well known "first order, second moment" method. This method quantifies the variance of main factors affecting the result of soil liquefaction assessments from the liquefaction and non-liquefaction case histories. Next, a full probabilistic model of evaluating soil liquefaction is built by combining the reliability analysis of liquefaction method with earthquake hazard analysis. The model also takes into account the variance in the occurrences of earthquakes and the variance of soil resistance. Finally, monetary values are attached to the various outcomes of the model in order to establish a complete soil liquefaction risk analysis and decision making methodology.
