Machine learning-driven design and mechanical performance evaluation of sustainable bamboo wood/LD sludge reinforced epoxy composites

This research presents a report on synergetic effect of hybrid fillers, i.e. bamboo wood dust (BWD) and Linz-Donawitz sludge (LDS) on the mechanical behavior of epoxy-based composites. The neat epoxy has a density of 1.1 g/cc and it decreases with the incorporation of BWD and LDS. The increase in BWD content to 6 wt.% clearly causes an improvement in microhardness, flexural, tensile and impact strengths. The tensile strength increases from 48.25 MPa for neat epoxy (EP) to 59.18 MPa for epoxy with 10 wt.% LDS and 6 wt.% BWD addition (EP-LD-6BWD). In the same way, flexural strength increases from 22.62 MPa for EP to 27.27 MPa for EP-LD-6BWD composite. The strength of the composite under impact load and the micro-hardness also increase significantly, from 12.32 kJ/m ² and 13.31 Hv for EP to 23.55 kJ/m ² and 22.23 Hv for EP-LD-6BWD and EP-LD-8BWD composites, respectively. The three machine learning models, decision tree, random forest and support vector regression (SVR), are also used to analyse and compare the density, tensile, flexural, impact and micro-hardness properties of the epoxy-based composites. Of these, the SVR model best predicted tensile strength and flexural strength with R ² values of up to 0.9238 and 0.9404, respectively, and prediction errors <1.5 MPa.