Abstract
This introductory paper describes how thermal stresses arise in service, mainly during start-up and .shut-down sequences. Two classes of thermal transients are distinguished, colloquially known as ‘upshock’ (promoting intergranular damage) and ‘downshock’ (promoting transgranular damage) and by reference to both service examples and laboratory tests, these differences can be detected by careful metallographic examination of the craze cracks. Several industrial applications where thermal shock is likely to occur are considered. Simulative laboratory testing can be reduced to the simplest technique possible so that materials may be ranked according to their resistance to fatigue crack initiation, but much more effective analytical results may be obtained by measuring the temperature transients. In this way the appropriate peak stress intensity profile (which governs crack behaviour) may be deduced and hence it can be assessed whether the structural integrity of the component is compromised. Simple laboratory tests on cylindrical structures show that cracks commonly arrest at some point across the section, but two factors separately or together cause breakthrough (simulating component behaviour). These are an embrittling heat treatment and the application of an external stress, respectively. However, in certain circumstances some relief on penetration depth can be claimed from the ‘multiple crack’ effect, which becomes less damaging than a single crack. Finally it is investigated whether steps can be taken to avoid the onset of craze cracking. These may be adopted at the design stage (choice of material and thickness considerations) and sometimes during manufacture, but it is seen that only at the operational stage can the maximum ameliorating measures be taken.