Sobol序列
多项式混沌
材料科学
刚度
不确定度量化
灵敏度(控制系统)
结构工程
有限元法
计算机科学
搭接接头
接头(建筑物)
蒙特卡罗方法
工程类
数学
统计
机器学习
电子工程
作者
Hanyu Zhang,Lei Zhang,Can Xu,Zhao Liu,Ping Zhu
标识
DOI:10.1016/j.compstruct.2021.114841
摘要
Though characterized with extraordinary mechanical performance, the application of hybrid joint (HJ) is restricted due to its numerous parameters and complex nonlinearities in design process. Sensitivity analysis can effectively simplify the HJ design by identifying the dominant parameters in the initial design stage. In this work, uncertainty quantification (UQ) tool was adopted to comprehensively analyze the global sensitivity of the mechanical properties in hybrid (bonded/riveted) single lap aluminum-CFRP joints. A three-dimensional finite element model considering material failure was established and verified by experiments. Besides multiple experimentally quantified uncertainty sources arising from material, geometry and assembly were introduced in the mechanical performance simulation. Meanwhile, the relative contribution of different joint parameters to load sharing, stiffness and peak load was quantified by a high-efficient stochastic sensitivity analysis strategy based on polynomial chaos expansion (PCE) and Sobol decomposition. Then the key parameters affecting the stochastic mechanical properties were identified. Finally, the physical meaning of the sensitivity analysis result was explained. Results show that the proposed method can efficiently and accurately determine the important parameters in the HJ design, which can be applied to improve the design efficiency and quality of HJs.
科研通智能强力驱动
Strongly Powered by AbleSci AI