Correlation of the bridging model predictions for triaxial failure strengths of composites with experiments

The bridging model was incorporated with a pseudo three-dimensional laminate theory to predict the strengths and deformation curves of a set of typical fiber-reinforced composites under triaxial loads. The prediction was achieved using the constituent properties and geometric parameters of the composites based on the information provided by the WWFE-II's organizers. In the present paper, the 'blind' predictions shown in Part A of this work are compared with the experimental results. The overall correlations between the predictions and the experiments are reasonable but somewhat conservative, especially when shear failures or large hydrostatic pressures are involved. Satisfactory improvement has been achieved when the modified maximum compressive stress failure criterion proposed previously is applied to detect fiber failure as well. Even better predictions for a number of the problems have been obtained if the constituent material properties directly taken from the organizers, rather than retrieved from the laminae data, are employed.