A Dynamic Evolution Scheme for Structures with Interval Uncertainties by Using Bidirectional Sequential Kriging Method

Yisi Liu,Xiaojun Wang,Lei Wang

DOI: https://doi.org/10.1016/j.cma.2019.01.041

IF: 6.588

2019-01-01

Computer Methods in Applied Mechanics and Engineering

Abstract:This paper proposes a bidirectional sequential Kriging (BSK) method for nonlinear interval uncertainty quantification of dynamic systems. Different from existing surrogate based methods, which build surrogate model at preselected samples, the proposed method constructs surrogate models sequentially. Small amount of initial samples are generated by Latin hypercube sampling (LHS) and a crude Kriging surrogate model (KSM) is constructed. Then, two index functions, named lower confidence bound (LCB) and upper confidence bound (UCB) respectively, are defined for sample collocation. The index functions guide the search for the optima of structural response. After iterations, a sophisticated surrogate model will be obtained and the interval bounds of dynamic response can be calculated quickly with some auxiliary algorithms such as genetic algorithm (GA). Numerical examples and engineering application are studied to verify the effectiveness of BSK. The results show that compared with existing methods, the proposed method can achieve higher accuracy with good efficiency in nonlinear interval analysis.