Role of Machine Learning in Additive Manufacturing of Titanium Alloys—A Review

Abstract

Due to their exceptional properties, titanium alloys are ideal for many technologically demanding applications, including aerospace, automotive, marine, military, sports, and biomedical. The increasing demand for these metals to be formed into intricate shapes necessitates the use of additive manufacturing. Additive manufacturing technology has the potential to offer extremely low-cost manufacturing with a high material utilization ratio, particularly for titanium components with more complicated shapes. The complexity and high-quality manufacture of titanium alloys, which demands a wide range of design concepts, mechanical properties, standardization, and quality control, pose significant problems for additive manufacturing. These are difficult to investigate and evaluate using statistical techniques. Machine learning can greatly improve the accuracy of modelling nonlinearities and the evaluation of the effect of various input parameters on material performance. Further, machine learning has been recognized as a reliable prediction tool for data-driven multi-physical modelling, capable of producing accurate results and examining system parameters beyond the scope of conventional computational and experimental investigation. Numerous studies have reported the use of machine learning algorithms, such as artificial neural network (ANN), support vector machine (SVM), convolutional neural network (CNN), decision tree (DT), k-nearest neighbor (KNN), k-means clustering, random forest (RF), Bayesian networks, self-organizing maps (SOM), and Gaussian process regression (GPR) in the design, fabrication, development, and quality control of titanium components via additive manufacturing. This review study consolidates the relevant literature and illustrates the applicability of machine learning approaches in modelling of titanium alloy additive manufacturing. Based on this literature review, a few recommendations for analyzing machine learning methods for modelling various additive manufacturing process parameters are presented, along with some insightful thoughts on prospective future research. © 2023, The Author(s) under exclusive licence to International Center for Numerical Methods in Engineering (CIMNE).