摘要
为准确计算不同参数影响下供水管道的抗震可靠度,依据地震波作用机理,建立地下管道抗震可靠度的功能函数。首先,汲取BP神经网络较强的非线性映射的结构自组织性和自学习性等特点,建立基于神经网络的一次二阶矩可靠度智能计算方法;然后,针对国内最常用的灰口铸铁供水管道,研究剪切波速、特征周期、管径、管壁厚度、地震波入射角对埋地管道失效概率的影响;最后,以秦皇岛市DN300以上的815段地下管道为例,进行不同地震作用下的震害预测。结果表明:一类场地是对管道抗震最有利的地段,四类场地为最不利地段,地震波入射角对管道失效概率的影响在45°左右达到最大值。随着剪切波速的增大、特征周期的减小,管道失效概率随之降低。在二、三类场地条件下可以适当增加管径和管壁厚度,以提高管道抗震性能。秦皇岛市供水管网系统在小震和中震作用下是比较可靠的,在大震作用下损害会很大。
To calculate accurately the seismic reliability of urban water supply pipeline under influence of different parameters,a seismic reliability calculating functional function of underground pipes was established according to mechanism of seismic wave. Taking into account the BP neural network characteristics of strong nonlinear mapping,self-organization and self-learning,a first-order second-moment method reliability intelligent calculation method based on neural network was worked out. Effects of the shear wave velocity,characteristic period,pipe diameter,pipe wall thickness,and seismic wave incidence angle on seismic reliability of the most commonly used gray cast iron pipe were studied to explore the influence on failure probability of buried pipelines. 815 segments of water supply pipeline in Qinhuangdao were taken as an example,an earthquake damage prediction under various seismic conditions was made. The results show that the pipe network system is reliable under slight and medium earthquake conditions,but under strong earthquake conditions it will be damaged severely.
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2015年第11期118-122,共5页
China Safety Science Journal
基金
国家自然科学基金资助(51208017)
国家科技支撑计划课题(2015BAK14B01)
北京市教委科技一般项目(KM201610005029)
关键词
供水管道
抗震性能
BP神经网络
失效概率
一次二阶矩法
water supply pipeline
seismic performance
BP neural network
failure probability
first order second moment
