Speaker
Description
Tungsten and tungsten-based alloy coatings are key candidates for plasma-facing components in fusion reactors due to their high melting point and favorable thermo-mechanical properties. However, their thermal performance is strongly influenced by microstructure, interfacial integrity, and processing-induced heterogeneities. Reliable, spatially resolved characterization of thermal properties and defect detection are essential for coating qualification.
This work presents a hybrid experimental-computational framework based on active infrared thermography (IRT) for tungsten and W-alloy coating characterization. Flash thermography with pulsed excitation and high-speed infrared imaging has been adapted to tungsten's high thermal conductivity regime. Large-area specimens with thickness gradients and compositional variations are being investigated.
A comprehensive dataset for machine learning training is being generated through three complementary approaches: (i) analytical heat transfer models for rapid generation of parametric defect libraries, (ii) finite element modeling (FEM) for realistic defect morphologies and complex geometries, and (iii) experimental thermographic measurements. AI models trained on transient thermal descriptors from these combined sources enable automated defect classification and thermal property mapping.
Initial results demonstrate detectability of interfacial debonding and distributed porosity in as-deposited coatings. This methodology establishes a scalable platform for plasma-exposed material assessment, providing quantitative thermal diffusivity maps and AI-assisted diagnostics for fusion coating development.
