The long-standing problem of characterizing failure for fiber composite materials will be reviewed. Emphasis will be given to the lamina level involving nominally aligned fibers in a matrix phase. However, some consideration will also be given to laminate failure using the lamina form as the basic building block along with the concept of progressive damage. The many different lamina level theories will be surveyed along with the commitment necessary to produce critical experimental data. Four particular theories will be reviewed and compared in some detail, these being the Tsai-Wu, Hashin, Puck, and Christensen forms. These four theories are reasonably representative of the great variety of different forms with widely different physical effects that can be encountered; also for comparison, the rudimentary forms of maximum normal stress and maximum normal strain criteria will be given. The controversial problem of how many different individual modes of failure are necessary to describe general failure will receive attention. A specific and detailed methodology for evaluation of all the various theories will be formulated.
IntroductionTheories of failure for anisotropic materials have been propounded for at least the past forty years. The advent of high strength, highly anisotropic fiber composite materials has accelerated the activity and accentuated the importance of the search. The lack of agreement on a single, best theory has not been for lack of activity. If one includes all forms of theoretical failure characterization, there are probably well over one hundred different theoretical forms, sometimes applicable over widely different conditions. To put some scope and limits on the present considerations, only reasonably comprehensive theories (not individual mechanism theories) will be considered, meaning theories applicable to fully three dimensional states of stress and strain. That the consideration and evaluation of such theories of behavior is a difficult proposition should be self evident. Let it just be said that even in the case of isotropic materials, the theoretical characterization of failure is not a settled issue. The corresponding problem for highly anisotropic materials could reasonably be expected to be much more difficult than that for isotropy.A sampling of some fully three dimensional theories of failure are given in Table 1. The sources for these are given in the list of references. By no means are the theories limited to the case of just 10 adjustable parameters. There are theories with 15 or more parameters. This situation immediately raises the issue of the degree of practicality for theories with large numbers of parameters. It would appear that somewhere in the range of 7-9 parameters is pushing the upper limit. There is a corresponding problem at the other end of the scale. What is the fewest number of parameters that can capture the immensely complicated interactive physical effects that occur at the threshold of failure. Additionally, in developing theories of failure one mu...