Evaluation of Machine Learning Models for Enhancing Sustainability in Additive Manufacturing

Additive manufacturing offers an immense potential for sustainability through optimized processes and material usage. This research investigates applications of machine learning models to predict and optimize key sustainability metrics in the energy consumption, part weight, scrap weight, and production time based on additive manufacturing process parameters such as the layer height, infill density, infill pattern, build orientation and number of shells. Four machine learning models, Linear Regression, Decision Trees, Random Forest, and Gradient Boosting, are evaluated with hyperparameter tuning performed using the Limited-memory Broyden-Fletcher-Goldfarb-Shanno with Box constraints optimization algorithm which demonstrates a superior computational efficiency compared to traditional methods like the grid search and random search. Among the models, Random Forest achieves the highest accuracy, and the lowest Mean Squared Error for all target metrics. The results provide actionable insights into optimizing additive manufacturing processes for sustainability, and demonstrate that machine learning can directly link process parameters to environmental and economic impacts. This research bridges a critical gap in sustainable additive manufacturing by offering a computationally efficient and scalable optimization approach.