Herein, an image‐processing‐based random field modeling of microstructure in an actual specimen of a unidirectional fiber‐reinforced composite material and multiscale stochastic stress analysis with considering a random fiber arrangement having the random field identified from an actual specimen is described. Multiscale stochastic stress analysis will be effective for evaluation of uncertainty propagation through the different scales in heterogeneous materials. It will be important when a microscopic randomness has a significant influence on apparent material properties. Some mechanical properties of heterogeneous materials take an influence of the microscopic randomness, and such multiscale stochastic analysis has been attracted. In this article, a random field modeling technique for a digital image of microstructure of a unidirectional fiber‐reinforced plastic (FRP) is presented, and the multiscale stochastic stress analysis with considering a random fiber arrangement in FRP based on the random field model identified by the presented method is attempted. For this purpose, a method for generating a set of realizations according to the identified random field is also developed. From the results, effectiveness of the presented approach is discussed.

No abstract