.png){kind=icon}
.png){kind=icon}
Return
Decoding microstructural hierarchy for autonomous alloy discovery
.png){kind=icon}
PRE
AI
DOI：10.1016/j.pnsc.2026.06.013.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Accelerated alloy discovery is often hindered by the lack of mathematically rigorous descriptors for complex microstructural hierarchies. This study introduces the MicroStructural Hierarchy Descriptor μSHD, a physics-informed framework for autonomous discovery. Utilizing the Mallat wavelet scattering transform, we extract Lipschitz-continuous feature vectors that resolve morphological energy across eight dyadic scales, spanning from sub-10 nm interfaces to macroscopic dendritic envelopes. Validated across Ti-6Al-4V and Ni-based superalloys, μSHD captures “Hierarchical Bifurcation”—the simultaneous coarsening of macroscopic features and refinement of residual interfaces—mirroring industrial ultrasonic backscatter and Spatially Resolved Acoustic Spectroscopy (SRAS) data with high fidelity. By aligning descriptors with Renormalization Group logic, the framework provides a robust digital twin for structural energy density. When integrated with Gaussian Process Regression within a Self-driving lab, μSHD signatures enable high-fidelity property prediction and uncertainty quantification. This approach significantly reduces experimental iterations, offering a scalable, physically interpretable pathway to bridge laboratory design and industrial manufacturing. The convergence of μSHD with artificial intelligence platforms accelerates the realization of the Materials Genome initiative.
Journal
P
IF:
7.1
Papers:
13
Citations:
0