Development of Polyurethane-Based Composites With Salt Clay and Industrial Wastes as Fillers: Corrosion, Mechanical Properties, and Machine Learning Insights

In this study, a polyurethane-based composite is developed by incorporating salt clay, ulexite, colemanite, and various other industrial waste materials. The effects of these fillers on the composite are evaluated and modeled using machine learning techniques. Among the tested models, random forest and neural network demonstrate the highest performance in predicting changes in compressive strength, hardness, and thermal conductivity. The dispersion of salt clay within the polyurethane matrix provides a 300%-500% increase in compressive strength and a 25%-40% improvement in hardness. Ulexite enhances compressive strength by 250%-350% and increases hardness by up to 30%, while colemanite contributes to a 400%-500% rise in compressive strength and a 35%-40% improvement in hardness. The addition of K & imath;rka clay waste and tincal further improves the composite's hardness and overall durability. Fly ash significantly increases compressive strength, although its effect on hardness is limited. The machine learning models effectively capture the relationship between input parameters and composite performance. The random forest model achieves a mean squared error (MSE) of 0.15 for compressive strength and 0.20 for hardness, while the neural network model yields the best results for thermal conductivity prediction with an MSE of 0.12. These findings highlight the potential of the developed composite for industrial applications, particularly in thermal insulation and low-load structural components. Future studies will focus on evaluating its performance under real-world conditions and further assessing its long-term durability.