Stochastic properties of strains of randomly-oriented chopped strand composites under unidirectional loading

Randomly -oriented chopped strand composite (ROS) specimens were tested under unidirectional load, and the results were statistically treated as the material properties of ROS composites varying across specimens. The strain fields were measured using digital image correlation (DIC). To evaluate the elastic properties quantitatively from the experimentally measured in -plane strain data, the probabilistic properties of the strain were obtained and investigated. The average and standard deviation, probabilistic density function and spatial autocorrelation function, were obtained to determine an appropriate test method and data treatment rules to thoroughly understand the elastic properties of the ROS composites. Assuming spatial ergodicity, the statistical quantities of the strains were determined from sample averages and spatial averages simultaneously. Reasonable test conditions, such as the number and sizes of specimens, gauge sections, data measurement, and analysis procedures for appropriately determining the elastic performance, were discussed through statistical quantities of the strain.