An official website of the United States government
Abstract Melt pool dynamics in metal additive manufacturing (AM) is critical to process stability, microstructure formation, and final properties of the printed materials. Physics-based simulation, including computational fluid dynamics (CFD), is the dominant approach to predict melt pool dynamics. However, the physics-based simulation approaches suffer from the inherent issue of very high computational cost. This paper provides a physics-informed machine learning method by integrating the conventional neural networks with the governing physical laws to predict the melt pool dynamics, such as temperature, velocity, and pressure, without using any training data on velocity and pressure. This approach avoids solving the nonlinear Navier–Stokes equation numerically, which significantly reduces the computational cost (if including the cost of velocity data generation). The difficult-to-determine parameters' values of the governing equations can also be inferred through data-driven discovery. In addition, the physics-informed neural network (PINN) architecture has been optimized for efficient model training. The data-efficient PINN model is attributed to the extra penalty by incorporating governing PDEs, initial conditions, and boundary conditions in the PINN model.
more »
« less
Accelerating Thermal Simulations in Additive Manufacturing by Training Physics-Informed Neural Networks With Randomly Synthesized Data
Abstract The temperature history of an additively manufactured part plays a critical role in determining process–structure–property relationships in fusion-based additive manufacturing (AM) processes. Therefore, fast thermal simulation methods are needed for a variety of AM tasks, from temperature history prediction for part design and process planning to in situ temperature monitoring and control during manufacturing. However, conventional numerical simulation methods fall short in satisfying the strict requirements of time efficiency in these applications due to the large space and time scales of the required multiscale simulation. While data-driven surrogate models are of interest for their rapid computation capabilities, the performance of these models relies on the size and quality of the training data, which is often prohibitively expensive to create. Physics-informed neural networks (PINNs) mitigate the need for large datasets by imposing physical principles during the training process. This work investigates the use of a PINN to predict the time-varying temperature distribution in a part during manufacturing with laser powder bed fusion (L-PBF). Notably, the use of the PINN in this study enables the model to be trained solely on randomly synthesized data. These training data are both inexpensive to obtain, and the presence of stochasticity in the dataset improves the generalizability of the trained model. Results show that the PINN model achieves higher accuracy than a comparable artificial neural network trained on labeled data. Further, the PINN model trained in this work maintains high accuracy in predicting temperature for laser path scanning strategies unseen in the training data.
more »
« less
- Award ID(s):
- 2309250
- PAR ID:
- 10439731
- Date Published:
- Journal Name:
- Journal of Computing and Information Science in Engineering
- Volume:
- 24
- Issue:
- 1
- ISSN:
- 1530-9827
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Current metal additive manufacturing (AM) systems suffer from limitations on the minimum feature sizes they can produce during part formation. The microscale selective laser sintering (μ-SLS) system addresses this drawback by enabling the production of parts with minimum feature resolutions of the order of a single micrometer. However, the production of microscale parts is challenging due to unwanted heat conduction within the nanoparticle powder bed. As a result, finite element (FE) thermal models have been developed to predict the evolution of temperature within the particle bed during laser sintering. These thermal models are not only computationally expensive but also must be integrated into an iterative model-based control framework to optimize the digital mask used to control the distribution of laser power. These limitations necessitate the development of a machine learning (ML) surrogate model to quickly and accurately predict the temperature evolution within the μ-SLS particle bed using minimal training data. The regression model presented in this work uses an “Element-by-Element” approach, where models are trained on individual finite elements to learn the relationship between thermal conditions experienced by each element at a given time-step and the element's temperature at the next time-step. An existing bed-scale FE thermal model of the μ-SLS system is used to generate element-by-element tabular training data for the ML model. A data-efficient artificial neural network (NN) is then trained to predict the temperature evolution of a 2D powder-bed over a 2 s sintering window with high accuracy.more » « less
-
Capturing Local Temperature Evolution During Additive Manufacturing Through Fourier Neural OperatorsHigh-fidelity, data-driven models that can quickly simulate thermal behavior during additive manufacturing (AM) are crucial for improving the performance of AM technologies in multiple areas, such as part design, process planning, monitoring, and control. However, the complexities of part geometries make it challenging for current models to maintain high accuracy across a wide range of geometries. Additionally, many models report a low mean square error (MSE) across the entire domain (part). However, in each time step, most areas of the domain do not experience significant changes in temperature, except for the heat-affected zones near recent depositions. Therefore, the MSE-based fidelity measurement of the models may be overestimated. This paper presents a data-driven model that uses Fourier Neural Operator to capture the local temperature evolution during the additive manufacturing process. In addition, the authors propose to evaluate the model using the R2 metric, which provides a relative measure of the model’s performance compared to using mean temperature as a prediction. The model was tested on numerical simulations based on the Discontinuous Galerkin Finite Element Method for the Direct Energy Deposition process, and the results demonstrate that the model achieves high fidelity as measured by R2 and maintains generalizability to geometries that were not included in the training process.more » « less
-
Abstract Machine learning can be used to automate common or time-consuming engineering tasks for which sufficient data already exist. For instance, design repositories can be used to train deep learning algorithms to assess component manufacturability; however, methods to determine the suitability of a design repository for use with machine learning do not exist. We provide an initial investigation toward identifying such a method using “artificial” design repositories to experimentally test the extent to which altering properties of the dataset impacts the assessment precision and generalizability of neural networks trained on the data. For this experiment, we use a 3D convolutional neural network to estimate quantitative manufacturing metrics directly from voxel-based component geometries. Additive manufacturing (AM) is used as a case study because of the recent growth of AM-focused design repositories such as GrabCAD and Thingiverse that are readily accessible online. In this study, we focus only on material extrusion, the dominant consumer AM process, and investigate three AM build metrics: (1) part mass, (2) support material mass, and (3) build time. Additionally, we compare the convolutional neural network accuracy to that of a baseline multiple linear regression model. Our results suggest that training on design repositories with less standardized orientation and position resulted in more accurate trained neural networks and that orientation-dependent metrics were harder to estimate than orientation-independent metrics. Furthermore, the convolutional neural network was more accurate than the baseline linear regression model for all build metrics.more » « less
-
Powder Bed Fusion (PBF) faces ongoing challenges in the areas of process monitoring and control. Standard methods for alleviating these issues rely on machine learning, which requires costly and time-consuming training data. Expense is compounded by the perceived necessity of using sensors with extremely high resolutions. This research avoids this cost by employing an Ensemble Kalman Filter (EnKF), which uses measured data to correct physics-based model predictions of the process, to monitor part internal temperature fields during building. This work tests EnKF performance, in simulation, for two model architectures, using simulated cameras of varying resolution as our measuring instruments. Crucially, we show that increasing camera resolution produces diminishing returns in EnKF accuracy, relative to the model predictions, with up to 81% error reduction. This result shows that current AM quality control practices with expensive sensors may be inefficient; with appropriate algorithms, cheaper setups may be used with little additional error.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1115/1.4062852
