Hybrid modelling by machine learning corrections of analytical model predictions towards high-fidelity simulation solutions

Research output: Journal contributionsJournal articlesResearchpeer-review

Standard

Hybrid modelling by machine learning corrections of analytical model predictions towards high-fidelity simulation solutions. / Bock, Frederic E.; Keller, Sören; Huber, Norbert et al.
In: Materials, Vol. 14, No. 8, 1883, 10.04.2021.

Research output: Journal contributionsJournal articlesResearchpeer-review

Harvard

APA

Vancouver

Bibtex

@article{02e117368c55439ea1be8f12b5f2129e,
title = "Hybrid modelling by machine learning corrections of analytical model predictions towards high-fidelity simulation solutions",
abstract = "Within the fields of materials mechanics, the consideration of physical laws in machine learning predictions besides the use of data can enable low prediction errors and robustness as opposed to predictions only based on data. On the one hand, exclusive utilization of fundamental physical relationships might show significant deviations in their predictions compared to reality, due to simplifications and assumptions. On the other hand, using only data and neglecting wellestablished physical laws can create the need for unreasonably large data sets that are required to exhibit low bias and are usually expensive to collect. However, fundamental but simplified physics in combination with a corrective model that compensates for possible deviations, e.g., to experimental data, can lead to physics-based predictions with low prediction errors, also despite scarce data. In this article, it is demonstrated that a hybrid model approach consisting of a physics-based model that is corrected via an artificial neural network represents an efficient prediction tool as opposed to a purely data-driven model. In particular, a semi-analytical model serves as an efficient low-fidelity model with noticeable prediction errors outside its calibration domain. An artificial neural network is used to correct the semi-analytical solution towards a desired reference solution provided by highfidelity finite element simulations, while the efficiency of the semi-analytical model is maintained and the applicability range enhanced. We utilize residual stresses that are induced by laser shock peening as a use-case example. In addition, it is shown that non-unique relationships between model inputs and outputs lead to high prediction errors and the identification of salient input features via dimensionality analysis is highly beneficial to achieve low prediction errors. In a generalization task, predictions are also outside the process parameter space of the training region while remaining in the trained range of corrections. The corrective model predictions show substantially smaller errors than purely data-driven model predictions, which illustrates one of the benefits of the hybrid modelling approach. Ultimately, when the amount of samples in the data set is reduced, the generalization of the physics-related corrective model outperforms the purely data-driven model, which also demonstrates efficient applicability of the proposed hybrid modelling approach to problems where data is scarce.",
keywords = "Analyticalmodel, Artificial neural networks;model correction, Data driven, Feature engineering, Finite elementmodel, Laser shock peening, Machine learning, Physics based, Residual stresses, Engineering",
author = "Bock, {Frederic E.} and S{\"o}ren Keller and Norbert Huber and Benjamin Klusemann",
year = "2021",
month = apr,
day = "10",
doi = "10.3390/ma14081883",
language = "English",
volume = "14",
journal = "Materials",
issn = "1996-1944",
publisher = "MDPI AG",
number = "8",

}

RIS

TY - JOUR

T1 - Hybrid modelling by machine learning corrections of analytical model predictions towards high-fidelity simulation solutions

AU - Bock, Frederic E.

AU - Keller, Sören

AU - Huber, Norbert

AU - Klusemann, Benjamin

PY - 2021/4/10

Y1 - 2021/4/10

N2 - Within the fields of materials mechanics, the consideration of physical laws in machine learning predictions besides the use of data can enable low prediction errors and robustness as opposed to predictions only based on data. On the one hand, exclusive utilization of fundamental physical relationships might show significant deviations in their predictions compared to reality, due to simplifications and assumptions. On the other hand, using only data and neglecting wellestablished physical laws can create the need for unreasonably large data sets that are required to exhibit low bias and are usually expensive to collect. However, fundamental but simplified physics in combination with a corrective model that compensates for possible deviations, e.g., to experimental data, can lead to physics-based predictions with low prediction errors, also despite scarce data. In this article, it is demonstrated that a hybrid model approach consisting of a physics-based model that is corrected via an artificial neural network represents an efficient prediction tool as opposed to a purely data-driven model. In particular, a semi-analytical model serves as an efficient low-fidelity model with noticeable prediction errors outside its calibration domain. An artificial neural network is used to correct the semi-analytical solution towards a desired reference solution provided by highfidelity finite element simulations, while the efficiency of the semi-analytical model is maintained and the applicability range enhanced. We utilize residual stresses that are induced by laser shock peening as a use-case example. In addition, it is shown that non-unique relationships between model inputs and outputs lead to high prediction errors and the identification of salient input features via dimensionality analysis is highly beneficial to achieve low prediction errors. In a generalization task, predictions are also outside the process parameter space of the training region while remaining in the trained range of corrections. The corrective model predictions show substantially smaller errors than purely data-driven model predictions, which illustrates one of the benefits of the hybrid modelling approach. Ultimately, when the amount of samples in the data set is reduced, the generalization of the physics-related corrective model outperforms the purely data-driven model, which also demonstrates efficient applicability of the proposed hybrid modelling approach to problems where data is scarce.

AB - Within the fields of materials mechanics, the consideration of physical laws in machine learning predictions besides the use of data can enable low prediction errors and robustness as opposed to predictions only based on data. On the one hand, exclusive utilization of fundamental physical relationships might show significant deviations in their predictions compared to reality, due to simplifications and assumptions. On the other hand, using only data and neglecting wellestablished physical laws can create the need for unreasonably large data sets that are required to exhibit low bias and are usually expensive to collect. However, fundamental but simplified physics in combination with a corrective model that compensates for possible deviations, e.g., to experimental data, can lead to physics-based predictions with low prediction errors, also despite scarce data. In this article, it is demonstrated that a hybrid model approach consisting of a physics-based model that is corrected via an artificial neural network represents an efficient prediction tool as opposed to a purely data-driven model. In particular, a semi-analytical model serves as an efficient low-fidelity model with noticeable prediction errors outside its calibration domain. An artificial neural network is used to correct the semi-analytical solution towards a desired reference solution provided by highfidelity finite element simulations, while the efficiency of the semi-analytical model is maintained and the applicability range enhanced. We utilize residual stresses that are induced by laser shock peening as a use-case example. In addition, it is shown that non-unique relationships between model inputs and outputs lead to high prediction errors and the identification of salient input features via dimensionality analysis is highly beneficial to achieve low prediction errors. In a generalization task, predictions are also outside the process parameter space of the training region while remaining in the trained range of corrections. The corrective model predictions show substantially smaller errors than purely data-driven model predictions, which illustrates one of the benefits of the hybrid modelling approach. Ultimately, when the amount of samples in the data set is reduced, the generalization of the physics-related corrective model outperforms the purely data-driven model, which also demonstrates efficient applicability of the proposed hybrid modelling approach to problems where data is scarce.

KW - Analyticalmodel

KW - Artificial neural networks;model correction

KW - Data driven

KW - Feature engineering

KW - Finite elementmodel

KW - Laser shock peening

KW - Machine learning

KW - Physics based

KW - Residual stresses

KW - Engineering

UR - http://www.scopus.com/inward/record.url?scp=85104473182&partnerID=8YFLogxK

U2 - 10.3390/ma14081883

DO - 10.3390/ma14081883

M3 - Journal articles

C2 - 33920078

AN - SCOPUS:85104473182

VL - 14

JO - Materials

JF - Materials

SN - 1996-1944

IS - 8

M1 - 1883

ER -

Documents

DOI

Recently viewed

Publications

  1. Challenges for postdocs in Germany and beyond:
  2. Practice and carryover effects when using small interaction devices
  3. Detection of oscillations with application in the pantograph control
  4. Offline question answering over linked data using limited resources
  5. A Framework for Applying Natural Language Processing in Digital Health Interventions
  6. Leverage points 2019
  7. Statistical methods for the evaluation of hydrological parameters for landuse planning
  8. An approach for dynamic triangulation using servomotors
  9. Hot tearing behaviour of binary Mg-1Al alloy using a contraction force measuring method
  10. Kontext
  11. Separable models for interconnected production-inventory systems
  12. Programmierung einer DELTA-Roboterzelle nach PackML Standard
  13. An experience-based learning framework
  14. Performance incentives in activity-based management
  15. Rethinking the Spatiality of Spatial Planning
  16. Identification of Parameters and States in PMSMs
  17. PID Controller Application in a Gimbal Construction for Camera Stabilization and Tracking
  18. Eulerian and Lagrangian perspectives on turbulent superstructures in Rayleigh-Bénard convection
  19. Do Linguistic Features Influence Item Difficulty in Physics Assessments?
  20. The erosion of relational values resulting from landscape simplification
  21. Resolving conflicts between people and over time in the transformation toward sustainability
  22. Temporal and thermodynamic irreversibility in production theory
  23. Making the most out of timeseries symptom data
  24. Lyapunov stability analysis to set up a saturating PI controller with anti-windup for a mass flow system
  25. Media-Assisted Foreign Language Learning - Concepts and Functions
  26. Effect of grain size and structure, solid solution elements, precipitates and twinning on nanohardness of Mg-Re alloys
  27. An Exploration of humans‘ ability to recognize emotions displayed by robots
  28. Using Digitalization As An Enabler For Changeability In Production Systems In A Learning Factory Environment
  29. Training in Components of Problem-Solving Competence
  30. Dialogic interactions in higher vocational learning environments in mainland China
  31. Anwendungsprogrammierung mit Embedded-SQL