A Deep MLP-CNN Model Using Metadata to Predict Melt Pool Morphology in LPBF: Effect of Geometrical Factors

Abstract

Abstract The laser powder bed fusion (LPBF) process enables complex geometries to be created in bespoke parts. The LPBF process reduces lead times and costs for high-value parts by reducing material waste, unlike traditional subtractive manufacturing techniques. However, the fabrication of complex geometries using the LPBF process for industrial applications remains challenging in terms of achieving consistent mechanical properties. Experimental investigations show that geometrical factors such as shape and size impact the thermal history, microstructure, defect structure, and thus mechanical and fatigue properties. Consequently, the partially known effect of design parameters on part quality causes significant challenges in the standardization, qualification, and certification of additively manufactured (AM-ed) parts. In this paper, a deep Multilayer Perceptron and Convolutional Neural Network (MLP-CNN) is proposed to predict the melt pool depth for various geometries during LPBF of stainless steel 316 L using metadata (i.e., numerical and image data). Quite often the melt pool depth generated during the AM processes is employed as a surrogate for thermal, defect- and micro-structural signatures. To predict melt pool depth during the LPBF process of complex geometries, a digital twin environment using a finite element model (FEM) is developed and validated using experimental melt pool measurements for different geometries within multi-tracks and layers. The FEM was in agreement with experimentation with an error of less than 15%. Through the process simulation, a large dataset of melt pool depths was obtained for various part geometries at different locations and process parameters. Next, the MLP-CNN framework was established to identify the impact of part design and process parameters on melt pool depths. To enhance the performance and robustness of this model, data augmentation has been implemented by rotating and transferring geometries to artificially expand the dataset. Data augmentation helps to mitigate overfitting and promote better generalization, especially in the context of limited training data. After training, the proposed model is found to give accurate melt pool prediction even for new geometries not considered during training. The fusion of CNN and MLP has led to melt pool predictions for unseen geometries with an accuracy of 95%.