Attempts on representing sink strengths with machine learning formulations and the long-term role of crystalline interfaces in the development of irradiation-induced bubbles

The present article addresses an early-stage attempt on replacing the analyticity-based sink strength terms in rate equations by means of a surrogate representation. Here we emphasise, context multiscale modelling, combinative use machine learning with scale analysis, through which set fine-resolution problems partial differential describing local behaviour, can be asymptotically sorted out from mean-field kinetics. Hence training formulation is restrictively oriented, that is, to express and already identified, but analytically unavailable nonlinear functional relationships between strengths other continuum field quantities. Besides, normally faster diffusion mechanisms crystalline interfaces are particularly modelled locally planar equations, their linkages for bulk formulated imposing irregular boundary condition where point defect fluxes discontinuous across interfaces. Thus distinctive role as sinks quick channels investigated. Methodologicalwise, treatment also applicable studying more complicated situation long-term behaviour observed irradiated materials.