FRACTURE-MECHANICS - THEORIA OR TEKHNE

Ideally a method for use in fracture mechanics should be a harmonious combination of theoria and techne, of contemplation and engineering skill. The main objective of fracture mechanics is to tell whether fracture is likely to occur in a structure under given conditions. This requires a theory of fracture and access to necessary material data, obtainable through laboratory testing. Existing theories rest on the fundamental assumption of near-edge autonomy, i.e. that the same events occur near the crack edge in one and the same material, irrespective of structure geometry and load configuration. Therefore it is of vital importance to identify autonomy classes, vis., those classes of problems within which an acceptable autonomy prevails. The most important autonomy class should contain the most common problems associated with engineering structures, and laboratory tests should be tailored to fit this class. This, however, is rarely done: tests seem rather to have been designed for convenience in the laboratory. In fact the intersection of an autonomy class that contains the most common engineering situations and any class to which commonly used test specimens might belong, is remarkably narrow and scarcely extends beyond the region of validity of linear fracture mechanics. A major part of nonlinear fracture mechanics has been developed for fracture under monotone loading, but it has been used rather indiscriminately for static loading, without considering implications of non-monotone loading. Experiments show that these implications might be very drastic - a hundred load cycles instead of one can sometimes reduce the load carrying capacity by more than 30 percent. Obviously, theory-building in non-linear fracture mechanics has departed very much from the engineering reality. On the other side of the coin, so called engineering tests are usually designed without any aspiration of connection with theory. The continuing popularity of Charpy tests illustrates this. and so do several wide plate test methods. By taking the simple facts referred to here into account, simple but functional methods for nonlinear fracture mechanics can be envisaged. One such method is suggested.