Using machine learning, Mitsubishi Chemical Performance Polymers (MCPP) developed a model to predict the performance of novel bio-based plasticisers with minimal lab work. Trained on legacy data, this tool simulates key properties like glass transition temperature, drastically cutting time, cost, and raw material use. Compared to traditional testing methods, it speeds up product development and in turn the shift from fossil-based to sustainable alternatives. The approach is now being proposed to customers for validation at Tier 1 and OEM levels. What powers MCPP’s innovations: its access to Mitsubishi’s thousands of researchers and labs worldwide.
Awards 2025