Research on reliability design method of agricultural infrastructures for earthquakes.

本研究分为以下两部分。砂质材料防液化路堤可靠性设计。本文讨论了砂土地基液化概率的计算方法和液化可靠性的设计方法。通常，用经验方程简单地由有效垂向应力sigma_v'、spt吹风数N、中位粒径D_<50>和细粒含量F_c确定液化阻力。后三个参数通常具有较大的空间变异性，在液化分析中需要建立这些参数的统计模型。这些统计模型是通过最小化AIC来确定的。采用确定性设计最大加速度或年最大加速度统计模型确定地震作用下的动荷载。统计模型是根据历史地震资料确定的。在后一种情况下，最大加速度是一个概率参数。采用蒙特卡罗方法对所有的概率参数进行了求解，并对砂土地基液化的概率进行了计算。最后，讨论了地基的最优加固方法和最优液化概率，特别选取了砂压桩法作为地基加固方法。浅滩岛小古肯-南部地震造成的农场池塘灾害调查。我们课题组调查了农场池塘的灾害情况。根据调查结果进行多元回归分析。研究结果明确了水塘失效率与震源距离、堤防纵向、工期、材料特性、施工现场地质等因素的关系。特别地，定量理论II被用于这里的分析。

This research is divided into following two parts.1. Reliability design of embankment constructed by sandy material for prevention of liquefaction.In this research, the calculation method for the probability of liquefaction in a sandy ground and the reliability design method for liquefaciton are discussed.Usually, the liquefaction resistance is simply determined from the effective vertical stree sigma_v', theSPT blow count N,the median grain size D_<50> and the fines content F_c with the empirical equation. The last three parameters generally have large spatial variability and the statistical models of these parameters are required for the liquefaction analysis. These statistical models were determined by minimizing AIC here.The dynamic load by the earthquake was determined based on 1) the deterministic design maximum acceleration, or 2) the statistical model of the annual maximum acceleration. The statistical model was decided from the historical earthquake data records. The maximum acceleration is probabilistic parameter in the latter case.The Monte Carlo method was applied for all probabilitic parameters and probability of liquefaction in a sandy ground was evaluated. Finally, the optimum ground improvement method and the optimum probabiity of liquefaction are discussed, particularly the sand compaction pile method (SCP) was picked up as a ground improvement method.2. Investigation for disaster of farm ponds caused by Hyogoken-Nanbu Earthquake in Awaji Island.We research group investigated the disaster of farm ponds. Based on the investigation results, the multiple regression analysis was conducted. As a result, the relationship of the failure rate of the ponds with the distance from the earthquake source, longitudinal directions of the embankments, construction period, characteristics of the material, and geology in construction sites, was clarified. Particularly, the quantification theory II was used in the analysis here.