含非正态随机参数结构可靠度的虚拟变量算法

PSEUDO VARIABLES ALGORITHM FOR STRUCTURAL RELIABILITY EVALUATED NON-NORMAL PROPERTIES

虚拟变量法是一种把随机问题转化成确定性问题求解的方法。基于与失效函数有关的随机结构变量的均值、标准差和概率分布等统计参数的确定性特征 ,本文提出了按一次二阶矩理论准确计算含非正态随机参数结构可靠度的一种虚拟变量算法。由于无需人工计算失效函数的偏导数 ,计算形式较为简单 ;当失效函数形式复杂、与失效函数有关的随机变量较多时 ,可获得较高的计算效率。文中给出一个含裂纹缺陷高压气瓶抗断裂可靠度的计算实例。

The pseudo variables method was an efficient and accurate algorithm which converts the random problems into the certainty one. Based on the certainty statistical characteristics of the probability descriptions whichever the mean value, the deviation and the probabilistic distributions that of the random parameters related to the structural failure function, a pseudo variables method is proposed in this paper, for accurately to compute the structural reliability in accordance with the first order second moment reliability theory. This method is an iterative approach, it is composed of the formulae that to converts non normal distributive variables into normal one by the R F method, to construct the structural pseudo variables, to compute the failure function and structural reliability. This method is simple and its accuracy is equivalent to that the traditional simulation method. In addition, the effects of properties of non normal variables of structures can be easily evaluated. Because the manual calculation of the partial differences of the structural failure function has been eliminated, the efficacy of this method is superior to the traditional method. An example of applied this method to simulate the reliability of a high pressure gas vessel with cracks that the actual model founded by the probabilistic fracture mechanical theory is also illustrated in this paper.