Uncertainty Analysis of Concrete Structures Using Modified Latin Hypercube Sampling Method

This paper proposes a modified method of Latin Hypercube sampling to reduce the variance of statistical parameters in uncertainty analysis of concrete structures. The proposed method is a modification of Latin Hypercube sampling method. This analysis method uses specifically modified tables of random permutations of ranked numbers. In addition, the Spearman coefficient is used to make modified tables. Numerical analysis is carried out to predict the uncertainty of axial shortening in prestressed concrete bridge. Statistical parameters obtained from modified Latin Hypercube sampling method and conventional Latin Hypercube sampling method are compared and evaluated by a numeric analysis. The results show that the proposed method results in a decrease in the variance of statistical parameters. This indicates the method is efficient and effective in the uncertainty analysis of complex structural system such as prestressed concrete bridges.