一种结构可靠性的模拟数值计算方法

Simulating the Numerical Computation Approach of Structure Reliability

常用结构可靠性计算的方法包括有蒙特卡洛模拟法,响应面法等。大多数方法在计算时必须采用有限元理论,并进行基于Monte carlo抽样仿真,以研究多种不确定因素的影响下结构的可靠性。因此文中提出了一种基于kriging理论建模和遗传算法求解的模拟数值可靠性分析方法,通过采用重要抽样方法以减少抽样仿真次数,并利用kriging模型的高效预测建立结构的状态功能函数,在此基础上,应用遗传算法计算结构可靠性指标。将此方法用于典型框架结构可靠度分析的实例,表明了该方法的有效性和较高的计算效率。

The reliability analysis methods include Monte Carlo simulation, first order second moment method and so on. Most of the methods need the finite element theory and Monte carlo sampling-based simulation. In this paper,we build the Kriging model and solve the reliability by genetic algorithm. With sampling the importance method, the sam pling simulation frequency can he reduced, and the kriging model is used to build the state function. On this basis, the application of genetic algorithms can help us to solve the reliability. Examples of the kinds of frame structure are provided to show the efficiency of the method.