Summary
The rapid development of electrical systems in transportation, energy storage, and defense demands dielectric polymers with excellent electrical strength, thermal stability, appropriate dielectric constant, and low loss. However, the mutual constraints among these properties pose a major challenge, and the conventional trial-and-verification paradigm is inefficient for exploring the vast polymer chemical space. Combining artificial intelligence with dielectric material design offers a promising route to accelerate multi-property optimization. This paper introduces an AI-assisted design methodology that integrates domain expertise, curated solid dielectric datasets, target characterization, and advanced machine learning. A polymer database with multi-dimensional structural and performance information is first constructed.
Read more Read lessA cross-scale fingerprint system is then developed to encode complex polymer features for model training. Multiple machine learning models are deployed for high-throughput screening of unreported structures, identifying candidates with breakthrough performance for synthesis and experimental verification. Taking the design of electrostatic dielectrics as an example, this work establishes a closed-loop system coupling information-driven screening with experimental feedback, demonstrating the potential of AI in multi-objective material optimization.
Additional informations
| Publication type | Session Materials |
|---|---|
| Reference | D1_11568_2026 |
| Publication year | |
| Publisher | CIGRE |
| Country | China, People's Republic of |
| Study committees | |
| File size | 835 KB |
| Price for non member | 30 € |
| Price for member | 30 € |
Authors
DENG Jingyu - Tsinghua University; WANG Qian - Tsinghua University; ZHOU Ying - Tsinghua University; ZHAN Xingyu - Tsinghua University; ZUO Zhou - Tsinghua University; WU Wenjin - Tsinghua University; LIANG Xidong - Tsinghua University; WU Chao - Tsinghua University
Keywords
Artificial Intelligence, Solid Dielectric, Target Design, High-throughput Screening, Polymer
