关键词: 碳纤维前体, 共聚聚丙烯腈, 自由基聚合, 人工智能, 数字化综合实验

Abstract: As a critical material in national defense, aerospace, and rail transportation sectors, carbon fiber has been included in China’s strategic development plan. The structural composition of its precursors, particularly copolymerized polyacrylonitrile, serves as the key factor of carbon fiber’s structure and performance. However, China currently faces technological bottlenecks in synthesizing and applying carbon fiber copolymer precursors, necessitating the cultivation of interdisciplinary talents with both professional knowledge and innovative capabilities through undergraduate teaching experiments. Presently, polymer chemistry laboratory courses predominantly focus on homopolymer radical polymerization experiments using single-variable controlled, non-exploratory approaches. The incorporation of copolymer synthesis experiments—which hold significant practical applications—into traditional curricula remains challenging due to time constraints, complex monomer selection and ratio determination, and limited instrument availability. The rapid advancement of digital technologies, particularly artificial intelligence (AI), offers promising solutions. This study designs a digital experimental teaching program for copolymer synthesis and application, leveraging open-source databases to train neural networks. Through AI-assisted predictions of various synthesis strategies, students can optimize parameters on a virtual platform to simulate the complete synthesis process and performance testing of polyacrylonitrile-based carbon fibers. These virtual experiments then guide physical laboratory investigations of carbon fiber precursor synthesis. The experimental data generated can be uploaded to the platform for fine-tuning pre-trained models, thereby progressively enhancing the AI’s predictive accuracy. Ultimately, by integrating with relevant virtual simulation experiments, this approach establishes a comprehensive modular experimental system encompassing the entire “synthesis-structure-property-application” workflow of carbon fiber precursors, providing students with a systematic, exploratory, and innovative digital integrated experiment that significantly improves talent development quality.

Key words: Carbon fiber precursor, Polyacrylonitrile copolymer, Free radical polymerization, Artificial intelligence, Digital comprehensive experiment