基于双点加点策略的改进Kriging响应面可靠度计算方法

Improved Kriging Response Surface Reliability Calculation Method Based on Two-point Addition Strategy

针对经典响应面法中多项式模型在拟合高度非线性结构功能函数时能力不足的问题,采用Kriging模型代替真实结构功能函数,提出了一种基于双点加点策略的改进Kriging响应面可靠度计算方法。根据双点加点策略,通过内插公式和评价函数选出最佳样本点,与一次二阶矩法产生的验算点一起加入到每次迭代的样本库中以更新Kriging模型,利用遗传算法对Kriging模型的重要参数进行优化,并用优化后的模型拟合结构功能函数,再结合重要抽样法修正可靠度计算结果。算例分析结果表明,所提方法可以兼顾计算精度和计算效率,验证了所提方法的适用性和高效性。

Aiming at the problems that the polynomial model in the classical response surface method was insufficient in fitting the highly nonlinear structural performance functions, the Kriging model was used to replace the real structural performance functions and an improved Kriging response surface reliability calculation method was proposed based on two-point addition strategy. According to the two-point adding strategy, the best sample points selected according to the interpolation formula and the evaluation function, as well as the checking points generated by first-order second-moment method were added to the sample library together to update the Kriging model . The important parameters of Kriging model were optimized by genetic algorithm, and the optimized model was used to fit the structural performance functions, then the reliability calculation results were corrected with the important sampling method. The results of numerical examples show that the proposed method may take into account both of the calculation accuracy and the calculation efficiency, which verifies the applicability and efficiency of proposed method.