Valorization of Industrial Waste in Polymer Composites: Enhancing Mechanical and Thermal Properties for Insulation Applications Using Machine Learning Analysis

Abstract

This study investigates the incorporation of industrial waste materials into polyurethane-based composites and evaluates their mechanical, thermal, and microstructural properties. The polyurethane matrix was synthesized from methylene diphenyl diisocyanate (MDI) and polyether polyol, into which various waste fillers, including ulexite, colemanite, tincal, and K & imath;rka clay, were introduced in different proportions. Mechanical testing revealed that specific wastes significantly enhance compressive strength, with ulexite- and clay-reinforced composites achieving improvements of 42.19% and 43.54%, respectively, compared to the pure polymer. The ulexite-clay composite exhibited the highest mechanical strength (38.67 kN), whereas tincal-containing samples demonstrated the weakest performance. Shore A hardness values generally decreased with waste incorporation, indicating that filler addition reduces polymer rigidity. Thermal conductivity results showed property variations within +/- 25%, where ulexite increased conductivity while K & imath;rka clay reduced it, thereby improving thermal insulation potential. Microstructural analysis using scanning electron microscopy (SEM) confirmed heterogeneous morphologies with dense filler distribution that intensified with increasing filler ratios. Fourier transform infrared spectroscopy (FTIR) indicated both physical and chemical interactions between the polymer matrix and boron-containing fillers, highlighting the complex interfacial bonding mechanisms. To complement the experimental analyses, machine learning (ML) models were applied to predict composite performance based on waste type and ratio. Among the tested algorithms, Random Forest (RF) demonstrated the highest predictive accuracy (R-2 > 0.90), confirming its suitability for modeling composite properties. The integration of ML provided quantitative insights into the role of individual and combined waste fillers, aligning closely with experimental observations. This research demonstrates that the controlled selection and optimization of waste fillers can enhance the performance of polyurethane composites, promote recycling of industrial byproducts, and support the development of sustainable materials for applications such as thermal insulation and structural components.