Due to the increasing importance of cyclic oxidation as a long term failure mechanism for high temperature components in the energy conversion and transport fields, it is vita1 that reliable predictive models are
developed. Unfortunately current test methods are, in general, neither good simulations of what occurs in industry, nor do they provide a fundamental basis for the development of predictive models. A programme is suggested which addresses these issues, through better experimental procedures, and the use of sophisticated finite element analysis and complimentary stress analysis techniques.