Skip to main content
main-content

Tipp

Weitere Kapitel dieses Buchs durch Wischen aufrufen

2023 | OriginalPaper | Buchkapitel

15. A Physics-Based Reduced Order Model with Machine Learning-Boosted Hyper-Reduction

verfasst von: Konstantinos Vlachas, David Najera-Flores, Carianne Martinez, Adam R. Brink, Eleni Chatzi

Erschienen in: Topics in Modal Analysis & Parameter Identification, Volume 8

Verlag: Springer International Publishing

share
TEILEN

Abstract

Physics-Based Reduced Order Models (ROMs) tend to rely on projection-based reduction. This family of approaches utilizes a series of responses of the full-order model to assemble a suitable basis, subsequently employed to formulate a set of equivalent, low-order equations through projection. However, in a nonlinear setting, physics-based ROMs require an additional approximation to circumvent the bottleneck of projecting and evaluating the nonlinear contributions on the reduced space. This scheme is termed hyper-reduction and enables substantial computational time reduction. The aforementioned hyper-reduction scheme implies a trade-off, relying on a necessary sacrifice on the accuracy of the nonlinear terms’ mapping to achieve rapid or even real-time evaluations of the ROM framework. Since time is essential, especially for digital twins representations in structural health monitoring applications, the hyper-reduction approximation serves as both a blessing and a curse. Our work scrutinizes the possibility of exploiting machine learning (ML) tools in place of hyper-reduction to derive more accurate surrogates of the nonlinear mapping. By retaining the POD-based reduction and introducing the machine learning-boosted surrogate(s) directly on the reduced coordinates, we aim to substitute the projection and update process of the nonlinear terms when integrating forward in time on the low-order dimension. Our approach explores a proof-of-concept case study based on a Nonlinear Auto-regressive neural network with eXogenous Inputs (NARX-NN), trying to potentially derive a superior physics-based ROM in terms of efficiency, suitable for (near) real-time evaluations. The proposed ML-boosted ROM (N3-pROM) is validated in a multi-degree of freedom shear frame under ground motion excitation featuring hysteretic nonlinearities.
Literatur
1.
Zurück zum Zitat Wagg, D., Worden, K., Barthorpe, R., Gardner, P.: Digital twins: state-of-the-art and future directions for modeling and simulation in engineering dynamics applications. ASCE-ASME J. Risk Uncert. Engrg. Sys. Part B Mech. Engrg. 6(3), 030901 (2020) CrossRef Wagg, D., Worden, K., Barthorpe, R., Gardner, P.: Digital twins: state-of-the-art and future directions for modeling and simulation in engineering dynamics applications. ASCE-ASME J. Risk Uncert. Engrg. Sys. Part B Mech. Engrg. 6(3), 030901 (2020) CrossRef
2.
Zurück zum Zitat Peherstorfer, B., Willcox, K.: Dynamic data-driven reduced-order models. Comput. Methods Appl. Mech. Eng. 291, 21–41 (2015) MathSciNetCrossRef Peherstorfer, B., Willcox, K.: Dynamic data-driven reduced-order models. Comput. Methods Appl. Mech. Eng. 291, 21–41 (2015) MathSciNetCrossRef
3.
Zurück zum Zitat Benner, P., Ohlberger, M., Cohen, A., Willcox, K.: Model Reduction and Approximation: Theory and Algorithms. SIAM, Philadelphia (2017) CrossRef Benner, P., Ohlberger, M., Cohen, A., Willcox, K.: Model Reduction and Approximation: Theory and Algorithms. SIAM, Philadelphia (2017) CrossRef
4.
Zurück zum Zitat Agathos, K., Bordas, S.P., Chatzi, E.: Parametrized reduced order modeling for cracked solids. Int. J. Num. Methods Eng. 121(20), 4537–4565 (2020) MathSciNet Agathos, K., Bordas, S.P., Chatzi, E.: Parametrized reduced order modeling for cracked solids. Int. J. Num. Methods Eng. 121(20), 4537–4565 (2020) MathSciNet
5.
Zurück zum Zitat Farhat, C., Avery, P., Chapman, T., Cortial, J.: Dimensional reduction of nonlinear finite element dynamic models with finite rotations and energy-based mesh sampling and weighting for computational efficiency. Int. J. Num. Methods Eng. 98(9), 625–662 (2014) MathSciNetCrossRef Farhat, C., Avery, P., Chapman, T., Cortial, J.: Dimensional reduction of nonlinear finite element dynamic models with finite rotations and energy-based mesh sampling and weighting for computational efficiency. Int. J. Num. Methods Eng. 98(9), 625–662 (2014) MathSciNetCrossRef
6.
Zurück zum Zitat Carlberg, K., Farhat, C., Cortial, J., Amsallem, D.: The gnat method for nonlinear model reduction: effective implementation and application to computational fluid dynamics and turbulent flows. J. Comput. Phys. 242, 623–647 (2013) MathSciNetCrossRef Carlberg, K., Farhat, C., Cortial, J., Amsallem, D.: The gnat method for nonlinear model reduction: effective implementation and application to computational fluid dynamics and turbulent flows. J. Comput. Phys. 242, 623–647 (2013) MathSciNetCrossRef
7.
Zurück zum Zitat Ghavamian, F., Tiso, P., Simone, A.: POD–DEIM model order reduction for strain-softening viscoplasticity. Comput. Methods Appl. Mech. Eng. 317, 458–479 (2017) MathSciNetCrossRef Ghavamian, F., Tiso, P., Simone, A.: POD–DEIM model order reduction for strain-softening viscoplasticity. Comput. Methods Appl. Mech. Eng. 317, 458–479 (2017) MathSciNetCrossRef
8.
Zurück zum Zitat Peherstorfer, B., Butnaru, D., Willcox, K., Bungartz, H.-J.: Localized discrete empirical interpolation method. SIAM J. Sci. Comput. 36(1), A168–A192 (2014) MathSciNetCrossRef Peherstorfer, B., Butnaru, D., Willcox, K., Bungartz, H.-J.: Localized discrete empirical interpolation method. SIAM J. Sci. Comput. 36(1), A168–A192 (2014) MathSciNetCrossRef
9.
Zurück zum Zitat Kapteyn, M.G., Willcox, K.E.: Predictive digital twins: Where dynamic data-driven learning meets physics-based modeling. In: International Conference on Dynamic Data Driven Application Systems, pp. 3–7. Springer, Berlin (2020) Kapteyn, M.G., Willcox, K.E.: Predictive digital twins: Where dynamic data-driven learning meets physics-based modeling. In: International Conference on Dynamic Data Driven Application Systems, pp. 3–7. Springer, Berlin (2020)
10.
Zurück zum Zitat Zhang, Y., Combescure, A., Gravouil, A.: Efficient hyper reduced-order model (HROM) for parametric studies of the 3D thermo-elasto-plastic calculation. Finite Elem. Anal. Des. 102, 37–51 (2015) MathSciNetCrossRef Zhang, Y., Combescure, A., Gravouil, A.: Efficient hyper reduced-order model (HROM) for parametric studies of the 3D thermo-elasto-plastic calculation. Finite Elem. Anal. Des. 102, 37–51 (2015) MathSciNetCrossRef
11.
Zurück zum Zitat Hijazi, S., Stabile, G., Mola, A., Rozza, G.: Data-driven POD-Galerkin reduced order model for turbulent flows. J. Comput. Phys. 416, 109513 (2020) MathSciNetCrossRef Hijazi, S., Stabile, G., Mola, A., Rozza, G.: Data-driven POD-Galerkin reduced order model for turbulent flows. J. Comput. Phys. 416, 109513 (2020) MathSciNetCrossRef
12.
Zurück zum Zitat Simpson, T., Dervilis, N., Chatzi, E.: Machine learning approach to model order reduction of nonlinear systems via autoencoder and LSTM networks. J. Eng. Mech. 147(10), 04021061 (2021) CrossRef Simpson, T., Dervilis, N., Chatzi, E.: Machine learning approach to model order reduction of nonlinear systems via autoencoder and LSTM networks. J. Eng. Mech. 147(10), 04021061 (2021) CrossRef
13.
Zurück zum Zitat Takata Mücke, N., Hjuler Christiansen, L., Karup-Engsig, A.P., Bagterp Jørgensen, J.: Reduced order modeling for nonlinear PDE-constrained optimization using neural networks (2019). arXiv e-prints, pages arXiv–1904 Takata Mücke, N., Hjuler Christiansen, L., Karup-Engsig, A.P., Bagterp Jørgensen, J.: Reduced order modeling for nonlinear PDE-constrained optimization using neural networks (2019). arXiv e-prints, pages arXiv–1904
14.
Zurück zum Zitat Lee, K., Carlberg, K.T.: Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders. J. Comput. Phys. 404, 108973 (2020) MathSciNetCrossRef Lee, K., Carlberg, K.T.: Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders. J. Comput. Phys. 404, 108973 (2020) MathSciNetCrossRef
15.
Zurück zum Zitat Fresca, S., Manzoni, A.: POD-DL-ROM: enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decomposition (2021). Preprint arXiv:2101.11845 Fresca, S., Manzoni, A.: POD-DL-ROM: enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decomposition (2021). Preprint arXiv:2101.11845
16.
Zurück zum Zitat Swischuk, R., Mainini, L., Peherstorfer, B., Willcox, K.: Projection-based model reduction: formulations for physics-based machine learning. Comput. Fluids 179, 704–717 (2019) MathSciNetCrossRef Swischuk, R., Mainini, L., Peherstorfer, B., Willcox, K.: Projection-based model reduction: formulations for physics-based machine learning. Comput. Fluids 179, 704–717 (2019) MathSciNetCrossRef
17.
Zurück zum Zitat Farhat, C., Chapman, T., Avery, P.: Structure-preserving, stability, and accuracy properties of the energy-conserving sampling and weighting method for the hyper reduction of nonlinear finite element dynamic models. Int. J. Num. Methods Eng. 102(5), 1077–1110 (2015) MathSciNetCrossRef Farhat, C., Chapman, T., Avery, P.: Structure-preserving, stability, and accuracy properties of the energy-conserving sampling and weighting method for the hyper reduction of nonlinear finite element dynamic models. Int. J. Num. Methods Eng. 102(5), 1077–1110 (2015) MathSciNetCrossRef
19.
Zurück zum Zitat Vlachas, K., Tatsis, K., Agathos, K., Brink, A.R., Chatzi, E.: A local basis approximation approach for nonlinear parametric model order reduction. J. Sound Vibr. 502, 116055 (2021) CrossRef Vlachas, K., Tatsis, K., Agathos, K., Brink, A.R., Chatzi, E.: A local basis approximation approach for nonlinear parametric model order reduction. J. Sound Vibr. 502, 116055 (2021) CrossRef
20.
Zurück zum Zitat Agathos, K., Tatsis, K.E., Vlachas, K., Chatzi, E.: Parametric reduced order models for output-only vibration-based crack detection in shell structures. Mech. Syst. Signal Process. 162, 108051 (2022) CrossRef Agathos, K., Tatsis, K.E., Vlachas, K., Chatzi, E.: Parametric reduced order models for output-only vibration-based crack detection in shell structures. Mech. Syst. Signal Process. 162, 108051 (2022) CrossRef
21.
Zurück zum Zitat Amsallem, D., Zahr, M.J., Washabaugh, K.: Fast local reduced basis updates for the efficient reduction of nonlinear systems with hyper-reduction. Adv. Comput. Math. 41(5), 1187–1230 (2015) MathSciNetCrossRef Amsallem, D., Zahr, M.J., Washabaugh, K.: Fast local reduced basis updates for the efficient reduction of nonlinear systems with hyper-reduction. Adv. Comput. Math. 41(5), 1187–1230 (2015) MathSciNetCrossRef
22.
Zurück zum Zitat Allemang, R.J.: The modal assurance criterion–twenty years of use and abuse. Sound Vibr. 37(8), 14–23 (2003) Allemang, R.J.: The modal assurance criterion–twenty years of use and abuse. Sound Vibr. 37(8), 14–23 (2003)
23.
Zurück zum Zitat Vlachas, K., Tatsis, K., Agathos, K., Brink, A.R., Quinn, D., Chatzi, E.: On the coupling of reduced order modeling with substructuring of structural systems with component nonlinearities. In: Dynamic Substructures, vol. 4, pp. 35–43. Springer, Berlin (2022) Vlachas, K., Tatsis, K., Agathos, K., Brink, A.R., Quinn, D., Chatzi, E.: On the coupling of reduced order modeling with substructuring of structural systems with component nonlinearities. In: Dynamic Substructures, vol. 4, pp. 35–43. Springer, Berlin (2022)
24.
Zurück zum Zitat Carlberg, K., Bou-Mosleh, C., Farhat, C.: Efficient non-linear model order reduction via a least squares Petrov-Galerkin projection and compressive tensor approximations. Int. J. Numer. Methods Eng. 86(2), 155–181 (2011) CrossRef Carlberg, K., Bou-Mosleh, C., Farhat, C.: Efficient non-linear model order reduction via a least squares Petrov-Galerkin projection and compressive tensor approximations. Int. J. Numer. Methods Eng. 86(2), 155–181 (2011) CrossRef
25.
Zurück zum Zitat Carlberg, K., Farhat, C., Cortial, J., Amsallem, D.: The GNAT method for nonlinear model reduction: effective implementation and application to computational fluid dynamics and turbulent flows. J. Comput. Phys. 242(C), 623–647 (2013) MathSciNetCrossRef Carlberg, K., Farhat, C., Cortial, J., Amsallem, D.: The GNAT method for nonlinear model reduction: effective implementation and application to computational fluid dynamics and turbulent flows. J. Comput. Phys. 242(C), 623–647 (2013) MathSciNetCrossRef
26.
Zurück zum Zitat Spiridonakos, M.D., Chatzi, E.N.: Metamodeling of dynamic nonlinear structural systems through polynomial chaos NARX models. Comput. Struct. 157, 99–113 (2015) CrossRef Spiridonakos, M.D., Chatzi, E.N.: Metamodeling of dynamic nonlinear structural systems through polynomial chaos NARX models. Comput. Struct. 157, 99–113 (2015) CrossRef
Metadaten
Titel
A Physics-Based Reduced Order Model with Machine Learning-Boosted Hyper-Reduction
verfasst von
Konstantinos Vlachas
David Najera-Flores
Carianne Martinez
Adam R. Brink
Eleni Chatzi
Copyright-Jahr
2023
DOI
https://doi.org/10.1007/978-3-031-05445-7_15