A further assessment of the predictive capabilities of current failure theories for composite laminates: comparison with experimental evidence

This paper presents supplementary conclusions from an international exercise to establish the current status of failure prediction theories for polymer composite laminates. In previous stages of the exercise leading theoreticians in the field used 15 different approaches to predict a range of biaxial failure envelopes and stress–strain curves for composite laminates. The theoretical predictions were compared with experimental results and their performance was assessed using a process designed by the organisers Hinton, Soden and Kaddour [Compos Sci Technol 62 (2002) 1725]. During the course of the exercise four additional theories have emerged that add to the overall picture. The performance of the new theories is assessed here and compared with that of the other fifteen theories, using an identical methodology. The theories are ranked according to their abilities to predict the experimental results for failure of a unidirectional fibre reinforced lamina, initial and final failure of multi-directional laminates and large deformation of laminates under biaxial loads. An attempt is made to identify the most accurate approaches for use in a wide range of applications. The predictions of the four most highly ranked theories, which included two of the new approaches, were within ±50% (i.e. a factor of 2) of the experimental results in more than 75% of the test cases.