Optimization and predictive modelling of chemically crosslinked recycled PE/PP blends for enhanced mechanical properties

• Chemically crosslinked recycled PP/PE blends achieved superior performance. • Optimal blend: 25% PE, 3% DCP, and 10% AcAm via Taguchi–GRA optimization. • ANN accurately predicted mechanical and physical blend properties. • Integrated experimental-statistical approach enhances sustainable polymer design. Recycling plastic has become a prevalent practice for reducing waste in the environment, but producing high-performance recycled polymer blends is still a major technical problem. In order to improve the mechanical performance of recycled High-Density Polyethylene (PE) and Polypropylene (PP) blends, this work offers an integrated experimental and predictive modeling method. In order to enhance the interfacial adhesion of polymer blends through chemical crosslinking, maleic anhydride (MAH) and Acrylamide (AcAm) were employed as coupling agents. Optimization of the of PE content, Dicumyl Peroxide (DCP) as an imitator, and AcAm was accomplished using the Taguchi experimental design, while the Analytic Hierarchy Process (AHP) and Grey Relational Analysis (GRA) were employed to assign response weights and identify the optimal formulation. The blend containing 25 wt.% PE, 3 wt.% DCP, and 10 wt.% AcAm demonstrated the best overall performance, achieving 60% higher tensile strength and 40% higher flexural strength compared to the untreated blend. It is found that the addition of coupling agents also increased thermal stability, but viscosity changes were negligible. Additionally, a strong correlation with experimental results was found when mechanical and physical properties were accurately predicted by an Artificial Neural Network (ANN) model (R = 0.952). A viable and sustainable framework for creating high-performance recycled polymer blends appropriate for a variety of industrial applications is offered by the combination of statistical optimization and ANN prediction.