Machine learning guided optimization of engineered wood dust reinforced hybrid polymer composites
Abhilash Purohit, Alok Agrawal, Priyabrat Pradhan, Arvind Kumar, Suresh Palanimuthu, Pankaj Kumar Chauhan
Abstract
Within this study, the modification of engineered wood dust-epoxy (EWD-EP) composite with the addition of Linz-Donawitz (LD) sludge as a filler to improve physical, mechanical properties, and abrasion resistance of composites was investigated. The neat epoxy has a density of 1100 kg/m3 and it decreased with the addition of EWD and LD sludge addition. A clear improvement in tensile, flexural, and impact strengths, along with micro-hardness, is observed as the LD sludge content increases. The tensile strength rises to a high from 48.0 MPa for neat epoxy (B0) to 62.3. MPa for 10 wt. % EWD and 10 wt. % LD sludge addition (B3). Similarly, the flexural strength rose from 22.5 MPa for B0 to 27.9 MPa for 10 wt. % EWD and 15 wt. % LD sludge addition (B4). The impact strength and micro-hardness also improved significantly, increasing from 12.2 kJ/m² and 13.3 Hv for the neat epoxy to 22.7 kJ/m² and 23.9 Hv for B3 composite. Taguchi design showed a significant improvement in abrasion resistance corresponding to the increase in LD sludge content. The results obtained from machine learning demonstrate a clear alignment between the model and the experimental findings.