Artificial intelligence and machine learning in additive manufacturing

Abstract

The adoption of artificial intelligence (AI) and machine learning (ML) in part with additive manufacturing (AM) is radically changing the functionality of advanced manufacturing technologies. The chapter includes a thorough analysis of the application of AI and ML approaches to some of the most important issues in AM, such as material development, process optimization, quality control, and defect detection. The review is the synthesis of current developments in several AM modes, such as powder bed fusion, directed energy deposition, and extrusion-based processing. Some of the main ML methods that have been discussed are convolutional neural networks to monitor objects visually, reinforcement learning to optimize parameters, and physics-informed surrogate models to predict material properties. This chapter also outlines a comprehensive discussion on AI-based material discovery models that combine high-throughput thermodynamic simulations with multiobjective optimization algorithms that show accelerations of alloy development cycles by orders of magnitude. The use of quality control applications is discussed in terms of real-time monitoring systems that achieve defect detection rates of more than 93% and reduce scan time by more than 90% with the help of AI-based computed tomography reconstruction. The closed-loop AI-AM model, consisting of design, process, monitoring, and control processes, is scheduled in a schematic diagram, and the overall table is used to find particular ML algorithms and AM applications with the performance metrics assigned to them. Nevertheless, in spite of the great progress, the issues of data sparseness, generalization of models, and real-time integration remain. Future trends are federated learning to achieve data privacy, digital twins to manage lifecycle, and autonomous manufacturing systems. The review offers the researcher and practitioners a systematic overview of the present state of affairs and future directions of the AI and AM intersection.