Creep cavitation is a major degradation mechanism in polycrystalline metals and alloys at high temperature for current and future nuclear electrical power generating plant. In general, creep deformation and failure have been studied intensively since the middle of the 20th century.

In this presentation we will consider and review the need to underwrite the integrity for safe and economic operation of existing and future generations of nuclear power plant where the design lives can be of the order of 60 years. Creep deformation and creep damage are strongly linked with the microstructure of the material, in this case high alloy steels. In general, creep cavities are considered to grow on grain boundaries by stress- directed diffusion at temperatures and stresses encountered in such plant. Examples will be given where cavitation has developed in Advanced Gas Cooled Reactor boiler components. High- resolution microscopy on creep-tested specimens can provide important information about the initiation, early growth and closure of creep cavities – however – in general microscopy is limited to a small region of a specimen, making interpolation to engineering-relevant length- scales challenging.

The talk will explore a novel correlative microscopy approach for characterising creep cavitation in a creep-tested 316H austenitic stainless steel to address this length-scale challenge. Here, backscattered electron images at a resolution to identify cavities are combined into millimetre-length stitched datasets, and deep-learning software is used to process these images to identify the location and frequency of the cavities. Factors leading to cavitation, such as grain boundary misorientation, can then be evaluated statistically.