Towards a surrogate spatiotemporal model of additive manufacturing for digital twin-based process control
Alexios Papacharalampopoulos, Dionysios Christopoulos, Panagiotis Stavropoulos
Abstract
The fourth industrial revolution has rendered Digital Twins (DT) a critical technology for enhancing manufacturing processes. For the case of digital twins of manufacturing processes in particular, real-time analysis, monitoring and optimization could be achieved. However, there is need for surrogate models, allowing fast solution, correspondence to physics and adaptivity. The objective of the current work is to introduce a concept of a process control-oriented spatiotemporal Digital Twin on a manufacturing process (namely Laser Powder Bed Fusion). To achieve this, a simplified Finite Element Method (FEM) model has been utilized to simulate the heating phase of a Laser Powder Bed Fusion (LPBF) AM process. Then, based on the results of the FEM model, a surrogate model (SM) has been created. This surrogate model is based on difference equations and it can approximate the output of the original FEM model to a desired extent. A spatiotemporal input-output model is elaborated, describing the distribution of temperature in space and its evolution in time, simultaneously. Based on the results, process control is then attempted, indicating the usability of such models.