Application of Cokriging method to structural reliability problems

Approximation techniques are widely used in structural reliability analysis to build computationally inexpensive surrogate models. Recently, the kriging method which is a semi-parameter interpolation technique that can be used for deterministic optimization and structural reliability has gained popularity. However, to fully exploit the kriging method, especially in large-scale systems, a large number of sample points should be generated to fill the design space and this can be very expensive and even impractical in practical engineering analysis. Therefore, in this paper, a new method-the cokriging method, which is an extension of kriging, is proposed to calculate the structural reliability. This paper presents four different database augmenting schemes to explore the use of the cokriging method for structural reliability problems by comparing it with the kriging method based on numerical examples. Also, designs of experiments, which are usually used in the cokriging models, are discussed by comparisons. Finally, this paper suggests the points for which the cokriging model could be improved to get better results for structural reliability problems.