The effect of residual stresses and sample preparation on progressive debonding during the fiber push-out test

Iterative finite-element analyses have been conducted to determine the debond length as a function of force during progressive debonding in fiber push-out tests. This procedure was applied to data from push-out tests on a steel/epoxy model composite. During push-out testing this composite debonded from the bottom (support) face of the sample. The modeling allowed debonding from the bottom and included pre-existing debonds caused by residual stresses and/or sample preparatiion. The predicted debond lengths were within 10% of the measured debond lengths. Evidence is also shown that cutting and polishing can cause debonds that are longer than one third of the sample thickness in samples with a relatively small diameter steel fiber.