Mechanical Characterization, Biodegradability Evaluation, and Machine Learning-based Prediction of Bio-composites Fabricated from Polybutylene Succinate (PBS) and Cow Dung Powder

The objective of present research was to devise a biocomposite material by analyzing different existing biocomposites enhancing the machining performance focusing on wrenching using a machine-learning methods. In this study, polylactide/cow dung powder biocomposites with 0-20 wt% CDP and ball milling were used to study the physical, mechanical and biodegrade properties of the polylactide biopolymer. The maximum density and mechanical strength observed was in biocomposite having 20% CDP powder. Ten weight percent CDPs contained biocomposite performance ranges critical tensile strength as 22.2 MPa for a particle size of 250 μm and 21.1 MPa for 500 μm. The melt flow index (MFI) increased as the CDP content raised to 5% but decreased with higher CDP percentages within a range of 34.6 to 55.7 g/10 min for 250 μm particles and 30 to 55.7 g/10 min for the 500 μm particles. CDP powder in biocomposite at 5% had the highest melt flow index while at 20% CDP showed the lowest MFI (melt flow index). And for those biocomposites performance ranges for crash there was a decrease in the impact strength value with increasing CDP content, with a range of 4.2 to 2.1 kJ/m2 for the 250 μm particles and 4.2 to 2.5 kJ/m2 for the coarse 500 μm particles. The more the CDP content, the greater the water absorption and it was observed to be the highest at 3.85% for the use of 20 wt% CDP with size 500 μm. The biodegradation studies showed that the weight loss of biocomposites was directly proportional to the CDP content, and the maximum weight loss of 5.6 % for 20 wt% CDP biocomposite was observed after 50 days in soil. It was also shown that the addition of CDP to PBS increases the biodegradability of the materials, though it has an impact on mechanical properties and processability, which can be beneficial for biocomposite applications. The research conducts performance prediction and optimization of a PBS/CDP biocomposite material. The efficiency of different ML methods wherein Extra Trees, Decision Tree and AdaBoost are among the best. In present study the prediction of the properties like density, impact energy and bio degradability of the composite were analysis using ML methods and the best R2 for the density were found in DT, XT, random forest and ada boost ML method. The best R2 for the impact energy was found in DT, XT, adaboost and random forest ML method. The Best R2 for the biodegradability properties was found in XT machine learning method for the present study.