An official website of the United States government
Abstract Applications of process‐based models (PBM) for predictions are confounded by multiple uncertainties and computational burdens, resulting in appreciable errors. A novel modeling framework combining a high‐fidelity PBM with surrogate and machine learning (ML) models is developed to tackle these challenges and applied for streamflow prediction. A surrogate model permits high computational efficiency of a PBM solution at a minimum loss of its accuracy. A novel probabilistic ML model partitions the PBM‐surrogate prediction errors into reducible and irreducible types, quantifying their distributions that arise due to both explicitly perceived uncertainties (such as parametric) or those that are entirely hidden to the modeler (not included or unexpected). Using this approach, we demonstrate a substantial improvement of streamflow predictive accuracy for a case study urbanized watershed. Such a framework provides an efficient solution combining the strengths of high‐fidelity and physics‐agnostic models for a wide range of prediction problems in geosciences.
more »
« less
Feature Engineering for Surrogate Models of Consolidation Degree in Additive Manufacturing
Surrogate models (SM) serve as a proxy to the physics- and experiment-based models to significantly lower the cost of prediction while providing high accuracy. Building an SM for additive manufacturing (AM) process suffers from high dimensionality of inputs when part geometry or tool-path is considered in addition to the high cost of generating data from either physics-based models or experiments. This paper engineers features for a surrogate model to predict the consolidation degree in the fused filament fabrication process. Our features are informed by the physics of the underlying thermal processes and capture the characteristics of the part’s geometry and the deposition process. Our model is learned from medium-size data generated using a physics-based thermal model coupled with the polymer healing theory to determine the consolidation degree. Our results demonstrate high accuracy (>90%) of consolidation degree prediction at a low computational cost (four orders of magnitude faster than the numerical model).
more »
« less
- Award ID(s):
- 1910539
- PAR ID:
- 10223904
- Date Published:
- Journal Name:
- Materials
- Volume:
- 14
- Issue:
- 9
- ISSN:
- 1996-1944
- Page Range / eLocation ID:
- 2239
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Two-photon lithography (TPL) is a direct laser writing process that enables the fabrication of cm-scale complex three-dimensional polymeric structures with submicrometer resolution. In contrast to the slow and serial writing scheme of conventional TPL, projection TPL (P-TPL) enables rapid printing of entire layers at once. However, process prediction remains a significant challenge in P-TPL due to the lack of computationally efficient models. In this work, we present machine learning-based surrogate models to predict the outcomes of P-TPL to >98% of the accuracy of a physics-based reaction-diffusion finite element simulation. A classification neural network was trained using data generated from the physics-based simulations. This enabled us to achieve computationally efficient and accurate prediction of whether a set of printing conditions will result in precise and controllable polymerization and the desired printing versus no printing or runaway polymerization. We interrogate this surrogate model to investigate the parameter regimes that are promising for successful printing. We predict combinations of photoresist reaction rate constants that are necessary to print for a given set of processing conditions, thereby generating a set of printability maps. The surrogate models reduced the computational time that is required to generate these maps from more than 10 months to less than a second. Thus, these models can enable rapid and informed selection of photoresists and printing parameters during process control and optimization.more » « less
-
Abstract Mitigating the adverse impacts caused by increasing flood risks in urban coastal communities requires effective flood prediction for prompt action. Typically, physics‐based 1‐D pipe/2‐D overland flow models are used to simulate urban pluvial flooding. Because these models require significant computational resources and have long run times, they are often unsuitable for real‐time flood prediction at a street scale. This study explores the potential of a machine learning method, Random Forest (RF), to serve as a surrogate model for urban flood predictions. The surrogate model was trained to relate topographic and environmental features to hourly water depths simulated by a high‐resolution 1‐D/2‐D physics‐based model at 16,914 road segments in the coastal city of Norfolk, Virginia, USA. Two training scenarios for the RF model were explored: (i) training on only the most flood‐prone street segments in the study area and (ii) training on all 16,914 street segments in the study area. The RF model yielded high predictive skill, especially for the scenario when the model was trained on only the most flood‐prone streets. The results also showed that the surrogate model reduced the computational run time of the physics‐based model by a factor of 3,000, making real‐time decision support more feasible compared to using the full physics‐based model. We concluded that machine learning surrogate models strategically trained on high‐resolution and high‐fidelity physics‐based models have the potential to significantly advance the ability to support decision making in real‐time flood management within urban communities.more » « less
-
When rheological models of polymer blends are used for inverse modeling, they can characterize polymer mixtures from rheological observations. This requires repeated evaluation of potentially expensive rheological models. We explored surrogate models based on Gaussian processes (GP-SM) as a cheaper alternative for describing the rheology of polydisperse binary blends. We used the time-dependent diffusion double reptation (TDD-DR) model as the true model; it takes a 5-dimensional input vector specifying the binary blend as input and yields a function called the relaxation spectrum as output. We used the TDD-DR model to generate training data of different sizes [Formula: see text], via Latin hypercube sampling. The optimal values of the GP-SM hyper-parameters, assuming a separable covariance kernel, were obtained by maximum likelihood estimation. The GP-SM interpolates the training data by design and offers reasonable predictions of relaxation spectra with uncertainty estimates. In general, the accuracy of GP-SMs improves as the size of the training data [Formula: see text] increases, as does the cost for training and prediction. The optimal hyper-parameters were found to be relatively insensitive to [Formula: see text]. Finally, we considered the inverse problem of inferring the structure of the polymer blend from a synthetic dataset generated using the true model. Surprisingly, the solution to the inverse problem obtained using GP-SMs and TDD-DR was qualitatively similar. GP-SMs can be several orders of magnitude cheaper than expensive rheological models, which provides a proof-of-concept validation for using GP-SMs for inverse problems in polymer rheology.more » « less
-
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
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3390/ma14092239
