Digital twin temperature field prediction of laser powder bed fusion through proper orthogonal decomposition with radial basis function

Abstract

Digital Twin framework, integrated with in-situ sensing and physical simulation, can significantly improve process productivity and product quality of Laser Powder Bed Fusion (LPBF). Online simulation is the key step for implement of Digital Twin of LPBF. For constructing an online simulation model of LPBF, a swift and accurate Reducing Order Modeling, Proper Orthogonal Decomposition with Radial Basis Function (POD-RBF) is designed and applied to the online simulation of single-track and single-layer temperature field of LPBF. The results show that the relative error of temperature between the POD-RBF and the FEM is about 2%, the training time is less than 8 min. These results mean the lower relative error, the less training time, when compared with the typical non-intrusive model reduction method, Proper Orthogonal Decomposition with Gaussian Process Regression (POD-GPR). The proposed POD-RBF method can accurately realize the rapid prediction of the powder temperature field under the designed process parameters, which meets the requirements of online simulation of Digital Twin model of LPBF and makes it possible to provide opportune feedback for correcting incipient anomalies.