An efficient hybrid uncertainty propagation method for structures based on polynomial chaos kriging

In complex mechanical systems, intricate uncertainty factors often exist, which can significantly impact the key performance and reliability of the system. Conducting propagation analysis of these complex uncertainties is a crucial approach for assessing the stability and safety of structural performance. However, traditional uncertainty propagation analysis methods may encounter a series of challenges, such as difficulties in model construction and time-consuming solutions, particularly when addressing the analysis of complex mechanical systems that involve various hybrid uncertainty factors. In response to these challenges, this study proposes a propagation analysis method for mechanical structures with hybrid uncertainties based on Polynomial Chaos Kriging (PCK). This method constructs a fundamental model for the black-box problems of complex equipment using PCK and incorporates Gegenbauer polynomials during the modeling process to enhance the adaptability of PCK in the presence of complex uncertain variables. Furthermore, this research introduces targeted sampling techniques for PCK modeling under hybrid uncertainty to maximize the efficiency of sample utilization. The analytical efficiency and solution accuracy of the proposed algorithm are validated through one numerical example and one engineering example of a drive axle shell analysis.