摘要
通过引入ANN(artificial neural network,人工神经网络)计算斜拉桥复合地震易损性,显著减少在分析结构能力随机性过程中的计算量,获得结构能力概率分布.采用IDA(incremental dynamic analysis,增量时程分析)非线性时程分析法获得结构需求概率分布.通过蒙特卡洛抽样得到各PGA(peak ground acceleration,地面峰值加速度)下结构失效概率,进而获得易损性曲线.分析结果说明,使用ANN模拟结构能力分布可以显著减少计算量;在考虑结构能力随机性的前提下,斜拉桥地震易损性有所增加.
Computation work decreases dramatically when resorting to artificial neural network （ANN） in analyzing the randomness of structure capacity. The structure demand probability distribution is obtained by using incremental dynamic analysis （IDA） method. Based on the probability distribution of structure capacity and demand, the failure probability under each peak ground acceleration （PGA） is calculated through Monte-Carlo sampling method. And the fragility curves are drawn out. Analysis results indicate that the computation work decreases dramatically due to the employ of ANN to simulate the capacity distribution; and the seismic fragility of cable-stayed bridge increases by taking the randomness of structural capacity into consideration.
出处
《同济大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第7期970-976,共7页
Journal of Tongji University:Natural Science
基金
国家自然科学基金(50978194
90915011
51278376)
交通运输部西部交通建设科技项目(20113185191410)
关键词
斜拉桥
易损性
人工神经网络
正交设计
蒙特
卡洛抽样
cable-stayed bridge
fragility
artificial neuralnetwork
orthogonal design method
Monte-Carlo samplingmethod
