Experimental and ANN-based mechanical analysis of jute fiber-reinforced rHDPE sustainable composites
Küçük Resim Yok
Tarih
2026
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Taylor & Francis Ltd
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
High-density polyethylene (HDPE) waste poses a severe environmental threat, with only 9% recycled globally. This study develops sustainable composites by reinforcing recycled HDPE (rHDPE) with RUCOSTAR EEE6-treated jute fabric at 6 and 12 wt.% loadings. Two treatment concentrations (50 g/L and 70 g/L) were applied via dip-pad-dry-cure to enhance hydrophobicity and interfacial adhesion. Six composite variants were fabricated through compression molding, and their physical, mechanical, and hygroscopic properties were evaluated. The 70 g/L treatment significantly reduced water absorption (E70F12 < 3% vs. 9% for untreated RF12), while increasing density confirmed improved impregnation and reduced void content. Tensile tests revealed enhanced ductility (up to 34% higher strain at break) with modest strength trade-offs at 12 wt.% loading. A feed-forward artificial neural network (ANN) was developed to predict the stress-strain constitutive relationship sigma = f(epsilon), with hyperparameters selected via 5-fold cross-validation and stability assessed over ten reruns. The optimized model achieved median R >= 0.97 and MAE below 0.50 MPa on the held-out test set, outperforming polynomial fitting, SVR, and random forest. This hybrid experimental - computational approach demonstrates that EEE6-treated jute/rHDPE composites offer moisture-resistant, high-ductility green materials for structural applications.
Açıklama
Anahtar Kelimeler
Jute Composites, Recycled Polymer Composites, Hygroscopic And Mechanical Properties, Biocomposites, Ann Modeling
Kaynak
Composite Interfaces
WoS Q Değeri
Q3
Scopus Q Değeri
Q1
