A Machine Learning-Based Model for Multiple Material Density Prediction Developed by Powder Bed Fusion Additive Manufacturing
摘要
In metal-based additive manufacturing, powder bed fusion has been adopted for printing different metals and alloys. However, process parameter selection for printing dense and defect-free components is still challenging. Therefore, in this work, a data-driven machine learning predictive model has been developed for density prediction of different materials. The decision tree-based model predicted the density with an accuracy and RMSE of 99% and 0.21, respectively. A process parameter window was predicted for Al50Si, IN718, and SS316L to develop dense components. A machine learning-backed GUI was also developed to assist the non-expert in experimental design.