论文部分内容阅读
针对直接蒙特卡罗方法在高维非线性结构混合可靠度模拟中稳定性差和效率低的问题,将马尔可夫链自适应重要性抽样技术引入混合可靠度计算过程.经实验确定了马尔可夫链转移概率分布,可快速获得近似服从最优重要性抽样函数的系列状态点.根据状态点的一阶原点矩和二阶中心矩来确定所采用的重要性抽样函数;对区间变量进行等距划分,并计算以区间变量为自变量的可靠度平均值,可得到混合可靠度值;通过算例对本方法、非自适应重要性抽样法和直接蒙特卡罗法进行了对比,表明了本方法稳定高效的优势,为结构混合可靠度计算提供了新途径.
Aiming at the poor stability and low efficiency of direct Monte Carlo method in the simulation of mixed reliability of high-dimensional nonlinear structures, the Markov chain adaptive importance sampling technique is introduced into the mixed reliability calculation process. Markov chain transfer The probability distribution can quickly obtain a series of state points that approximate to the sampling function of the optimal importance.According to the first-order moment of origin and the second-order central moment of the state point, the importance sampling function is determined, and the interval variables are equally divided, And calculate the reliability of the interval variable as an independent variable to obtain the value of the mixed reliability; comparison of the method, non-adaptive importance sampling method and direct Monte Carlo method shows that the method is stable and efficient The advantages of this approach provide a new way to calculate the mixed reliability of structures.