基于DRHOSVM的复杂结构瞬态可靠性分析

Transient Reliability Analysis of Complex Structures Based on DRHOSVM

为提高结构瞬态可靠性分析的效率和精度,基于降维策略和参数优化思想,以支持向量机为建模基础,提出基于降维策略的参数优化支持向量机算法(DRHOSVM)。以航空发动机涡轮叶盘的疲劳寿命数据集为算法验证对象进行瞬态可靠性分析,通过对比直接模拟、传统支持向量机、Kriging,以验证其在建模特性和仿真性能方面的有效性和适用性。结果表明:DRHOSVM在高维数据规模下具有良好的算法精度和效率,适用于复杂结构的可靠性分析。

In order to improve the efficiency and accuracy of structural transient reliability analysis,based on dimension reduc⁃tion(DR)and parameter optimization strategy,and support vector machine(SVM)was taken as modeling basis,a hyperparameters optimization support vector machine-based on dimensionality reduction(DRHOSVM)method was proposed.Taking the fatigue life da⁃ta set of aero-engine turbine disk as the algorithm verification object,the transient reliability analysis was carried out.By comparing with the direct simulation,traditional support vector machine and Kriging,the effectiveness and applicability of DRHOSVM in model⁃ing characteristics and simulation performance were verified.The results show that DRHOSVM has good accuracy and efficiency in highdimensional data scale,and it is suitable for reliability analysis of complex structures.