AI Model Predicts Flame Resistance of Epoxy Resins

Researchers at the IMDEA Materials Institute have developed a machine-learning framework capable of accurately predicting the fire resistance of epoxy resins, a widely used polymer across the automotive, aerospace, electronics, and construction sectors. The methodology, detailed in a recent publication in Polymer Degradation and Stability, addresses the persistent safety challenges posed by the inherent flammability of these materials while eliminating the time-intensive and costly trial-and-error processes that traditionally govern flame-retardant development. Led by Professor De Yi Wang’s High-Performance Polymers and Fire Retardants Research Group, the team trained their algorithm on a dataset comprising 510 epoxy composite samples formulated with phosphorus-based flame retardants. These halogen-free additives represent a safer alternative to conventional chemical treatments. The model evaluates molecular structures, resin formulations, and auxiliary variables to forecast two critical fire-safety metrics: the UL-94 vertical flammability rating and the Limiting Oxygen Index. By simultaneously analyzing combustion duration, afterglow behavior, flame-dripping characteristics, and minimum oxygen concentration requirements for ignition, the system generates a comprehensive profile of real-world fire performance. Beyond raw prediction, the research introduces a standardized evaluation framework that translates complex algorithmic outputs into four discrete semantic tiers: excellent, good, moderate, and poor. This classification system provides engineers and materials scientists with immediate, actionable guidance for selecting and prioritizing polymer formulations. The approach has already demonstrated robust validation through independent case studies, confirming its practical utility in industrial material design. Dr. Qiong Tan, the study’s lead author, emphasized that the integration of artificial intelligence into materials discovery fundamentally accelerates the identification of safer, high-efficiency fire retardants. By reducing dependency on extensive laboratory testing, the platform enables faster iterative cycles and lowers development expenditures. The IMDEA Materials Institute intends to expand the underlying database to encompass alternative polymer classes and novel retardant chemistries, further broadening the technology’s applicability. Commercial deployment of the AI-driven assessment tool carries direct implications for critical infrastructure and consumer technology. Enhanced predictive capabilities are expected to improve the safety profiles of electronic components, electric vehicle battery systems, commercial aviation interiors, and modern construction materials, establishing a new standard for rapid, data-driven polymer engineering.