fig8

Discovering polyimides and their composites with targeted mechanical properties through explainable machine learning

Figure 8. Prediction and evaluation of 800,000 virtual PI structures to identify high-performance candidates with enhanced mechanical properties for PI materials. Six PI structures were selected, in which three based on their outstanding individual mechanical properties and three chosen for their overall balanced performance. (A) 3D scatter plots showing 10,000 random PI data points (red color) and the six high-performance PI structures, each marked with distinct symbols: triangles for the structure with high ε, circles for high σ, squares for high Et, and stars for structures with balanced mechanical properties. Each high-performance structure is projected onto three 2D planes, shown in color, for better visualization and comparison with the general distribution (gray shadow) of random PI data; (B) Structures of six PI candidates and the corresponding mechanical properties for both PI materials and PI composites. PI: Polyimide.

Journal of Materials Informatics
ISSN 2770-372X (Online)
Follow Us

Portico

All published articles are preserved here permanently:

https://www.portico.org/publishers/oae/

Portico

All published articles are preserved here permanently:

https://www.portico.org/publishers/oae/