基于人工神经网络的结构可靠性分析

Structural Reliability Analysis Based on Artificial Neural Network

在结构可靠性分析中,通过人工神经网络拟合极限状态函数,借助神经网络的函数映射关系产生极限状态函数值。采用Monte Carlo法随机抽样的思路,对大范围的数据进行概率分析,得到极限状态函数值的均值和标准差,求得结构系统的可靠性指标,该方法为极限状态函数无法显式表达的结构系统可靠性分析问题提供了一个可行方法。

The limit state function can be fitted by artificial neural network and its value can be generated by function mapping relation of neural netwok for the structure reliability analysis. Adopting the thinking of Monte Carlo random samples, the probability analysis to the data of the big scope can be carried on, the mean value of the limit state function and standard deviation can be got, the reliability index of the structure system can be solved. Under the condition that the limit state function is unable to make explicit expression, a feasible method was provided for reliability analysis question of the structure system.