Abstract
Some of the simplifying assumptions fundamental to shear lag theory-based models of the fiber pullout test ore compared with results from a detailed finite-element (FE) model to examine their validity. A range of material and geometric parameters representative of glass-and carbon fiber-reinforced plastic specimens are examined. The FE model strongly contradicts shear lag theory's assumptions that the axial stress gradient in the matrix can be neglected in the equilibrium equation, and that the radial and hoop stresses can be neglected in the constituitive relations of the matrix. Shear lag's assumptions that the shear strain in the matrix is dominated by the variation of axial displacement, and that the fiber end transmits no load appear to be in better agreement with the FE results. It is speculated that the questionable validity of these assumptions leads to inaccurate predictions for the energy release rate, G.