Vitrimeric polymers offer a unique solution by combining the recyclability of thermoplastics with the superior thermo-mechanical properties of thermosets. The study introduces a novel graph variational autoencoder (VAE) model to generate and guide the inverse design of vitrimers based on Tg. By leveraging MD simulations and ML techniques, a diverse dataset of vitrimers is created, allowing for property-guided inverse design. The VAE model employs dual graph encoders and overlapping latent dimensions to represent multi-component vitrimers accurately. Through Bayesian optimization, novel vitrimers with targeted Tg are efficiently discovered beyond the training regime.
To Another Language
from source content
arxiv.org
Thông tin chi tiết chính được chắt lọc từ
by Yiwen Zheng,... lúc arxiv.org 03-14-2024
https://arxiv.org/pdf/2312.03690.pdfYêu cầu sâu hơn