Surrogate Modeling of Melt Pool Thermal Field using Deep Learning

TL;DR

This work creates three datasets of single-trail processes using Flow-3D and uses them to train a convolutional neural network capable of predicting the behavior of the three-dimensional thermal field of the melt pool solely by taking three parameters as input: laser power, laser velocity, and time step.

Abstract

Powder-based additive manufacturing has transformed the manufacturing industry over the last decade. In Laser Powder Bed Fusion, a specific part is built in an iterative manner in which two-dimensional cross-sections are formed on top of each other by melting and fusing the proper areas of the powder bed. In this process, the behavior of the melt pool and its thermal field has a very important role in predicting the quality of the manufactured part and its possible defects. However, the simulation of such a complex phenomenon is usually very time-consuming and requires huge computational resources. Flow-3D is one of the software packages capable of executing such simulations using iterative numerical solvers. In this work, we create three datasets of single-trail processes using Flow-3D and use them to train a convolutional neural network capable of predicting the behavior of the three-dimensional thermal field of the melt pool solely by taking three parameters as input: laser power, laser velocity, and time step. The CNN achieves a relative Root Mean Squared Error of 2% to 3% for the temperature field and an average Intersection over Union score of 80% to 90% in predicting the melt pool area. Moreover, since time is included as one of the inputs of the model, the thermal field can be instantly obtained for any arbitrary time step without the need to iterate and compute all the steps

Figures (11)

• Figure 1: a) The overall schematic of the architecture of T-CNN. b) The convolutional upsampling layers and the feature maps through the network. The number of channels and the dimensions of each feature map are noted above and below it respectively. c) The operators used in the network.
• Figure 2: A simple illustration of the role of the auxiliary M-CNN. a) The initial T-CNN model's functionality. b) How M-CNN and TM-CNN are trained. c) The MT-CNN model augmented with M-CNN
• Figure 3: Performance of MT-CNN on some validation samples from the Ti64-5m dataset. Note that the original data is 3D, and the cross-section of the thermal field is shown in these plots.
• Figure 4: The summarized results for Ti64-5m dataset. a) The average performance of the model over the data distribution. The test samples are marked by a black edge. b) The performance of the model over time. The main line indicates the median value for the metric, and the shaded area indicates the interquartile range.
• Figure B.5: Performance of MT-CNN on some validation samples from the Ti64-10m dataset.