摘要
基于BP网络的人工神经元模型和可靠度理论,建立极限状态的抗液化阻力比函数和液化概率函数。沿用原抗震规范中液化标准贯入锤击数基准值概念,建立了简化的液化判别概率方法。该法以液化标准贯入锤击数作为估计液化势的基本依据。基准值是给定地面加速度、土层埋深、地下水位的液化临界锤数,也与震级大小和液化概率有关。为了对不同震级和土层中任一点进行液化判别,引入土层埋深水位以及震级大小对基准值的修正系数。为了方便工程应用,也给出了按地震分组的液化判别方法。
Based on ANN model and reliability theory the limit state function of liquefaction and the probability function of liquefaction are established.In terms of the reference value of SPT in China code a simplified probabilistic procedure for liquefaction evaluation is proposed.The assessment of soil liquefaction potential is based on the liquefied reference value of SPT which is the liquefaction critical value of SPT for given ground acceleration,the depth of soil layers and ground water levels,and is also related to earthquake magnitude and liquefaction probability.The two corrections factors for earthquake magnitudes,and the depth of soil layers and ground water levels,are respectively given for estimating liquefaction potential at any soil layer and various earthquake magnitudes.In addition,based on the procedure another liquefaction criterion according to the design earthquake groups is also given to be convenient to engineering practice in China.
出处
《岩土工程学报》
EI
CAS
CSCD
北大核心
2011年第1期112-116,共5页
Chinese Journal of Geotechnical Engineering
关键词
神经网络模型
液化
可靠度
标准贯入锤击数基准值
artificial neural network（ANN）
liquefaction
reliability index
reference value of SPT
