Analysis on Mechanical Behavior of Additively Manufactured PLA/Eggshell Composites Using Machine Learning Algorithms

Machine learning techniques have become a powerful tool in material research to correlate the input parameters and their corresponding outputs especially in process optimization and material development. The influence of the input parameters on the strength of polylactic acid/eggshell composites fabricated by fused deposition modeling is analyzed in this study. The input parameters considered are the layer thickness, percentage of eggshell, nozzle temperature and printing speed. Increase in nozzle temperature and percentage of eggshell augmented the strength, while higher printing speed and layer thickness declined the strength of composites. Artificial neural network, decision tree, random forest, K-nearest neighbors, support vector machine and extreme gradient boosting are the machine learning algorithms used in this study to predict the strength of composites. The significance of the machine learning models was analyzed using the regression parameters, namely coefficient of correlation, mean-squared error, mean absolute error and mean absolute percentage error. Artificial neural network produced better results followed by support vector machine algorithms in predicting the strength of the composites. Maximum correlation exists between the flexural strength and tensile strength (0.943) followed by impact strength and flexural strength (0.782). The highest positive correlation between the input and output variables exists between the tensile strength and nozzle temperature (0.540). The least integration is between the layer thickness and printing speed (0.079).