Advanced search
1 file | 5.10 MB Add to list

Neural network augmented physics models for systems with partially unknown dynamics : application to slider-crank mechanism

Author
Organization
Abstract
Dynamic models of mechatronic systems are abundantly used in the context of motion control and design of complex servo applications. In practice, these systems are often plagued by unknown interactions, which make the physics-based relations of the system dynamics only partially known. This article presents a neural network augmented physics (NNAP) model as a combination of physics-inspired and neural layers. The neural layers are inserted in the model to compensate for the unmodeled interactions, without requiring direct measurements of these unknown phenomena. In contrast to traditional approaches, both the neural network and physical parameters are simultaneously optimized, solely by using state and control input measurements. The methodology is applied on experimental data of a slider-crank setup, for which the state-dependent load interactions are unknown. The NNAP model proves to be a stable and accurate modeling formalism for dynamic systems that ab initio can only be partially described by physical laws. Moreover, the results show that a recurrent implementation of the NNAP model enables improved robustness and accuracy of the system state predictions, compared to its feedforward counterpart. Besides capturing the system dynamics, the NNAP model provides a means to gain new insights by extracting the neural network from the converged NNAP model. In this way, we discovered accurate representations of the unknown spring force interaction and friction phenomena acting on the slider mechanism.
Keywords
Control and Systems Engineering, Electrical and Electronic Engineering, Computer Science Applications, Mathematical model, Neural networks, Load modeling, Servomotors, Predictive models, Physics, Mechatronics, Explainable artificial intelligence, load identification, neural networks, nonlinear dynamic system modeling, IDENTIFICATION, DESIGN

Downloads

  • (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 5.10 MB

Citation

Please use this url to cite or link to this publication:

MLA
De Groote, Wannes, et al. “Neural Network Augmented Physics Models for Systems with Partially Unknown Dynamics : Application to Slider-Crank Mechanism.” IEEE-ASME TRANSACTIONS ON MECHATRONICS, vol. 27, no. 1, 2022, pp. 103–14, doi:10.1109/tmech.2021.3058536.
APA
De Groote, W., Kikken, E., Hostens, E., Van Hoecke, S., & Crevecoeur, G. (2022). Neural network augmented physics models for systems with partially unknown dynamics : application to slider-crank mechanism. IEEE-ASME TRANSACTIONS ON MECHATRONICS, 27(1), 103–114. https://doi.org/10.1109/tmech.2021.3058536
Chicago author-date
De Groote, Wannes, Edward Kikken, Erik Hostens, Sofie Van Hoecke, and Guillaume Crevecoeur. 2022. “Neural Network Augmented Physics Models for Systems with Partially Unknown Dynamics : Application to Slider-Crank Mechanism.” IEEE-ASME TRANSACTIONS ON MECHATRONICS 27 (1): 103–14. https://doi.org/10.1109/tmech.2021.3058536.
Chicago author-date (all authors)
De Groote, Wannes, Edward Kikken, Erik Hostens, Sofie Van Hoecke, and Guillaume Crevecoeur. 2022. “Neural Network Augmented Physics Models for Systems with Partially Unknown Dynamics : Application to Slider-Crank Mechanism.” IEEE-ASME TRANSACTIONS ON MECHATRONICS 27 (1): 103–114. doi:10.1109/tmech.2021.3058536.
Vancouver
1.
De Groote W, Kikken E, Hostens E, Van Hoecke S, Crevecoeur G. Neural network augmented physics models for systems with partially unknown dynamics : application to slider-crank mechanism. IEEE-ASME TRANSACTIONS ON MECHATRONICS. 2022;27(1):103–14.
IEEE
[1]
W. De Groote, E. Kikken, E. Hostens, S. Van Hoecke, and G. Crevecoeur, “Neural network augmented physics models for systems with partially unknown dynamics : application to slider-crank mechanism,” IEEE-ASME TRANSACTIONS ON MECHATRONICS, vol. 27, no. 1, pp. 103–114, 2022.
@article{8710449,
  abstract     = {{Dynamic models of mechatronic systems are abundantly used in the context of motion control and design of complex servo applications. In practice, these systems are often plagued by unknown interactions, which make the physics-based relations of the system dynamics only partially known. This article presents a neural network augmented physics (NNAP) model as a combination of physics-inspired and neural layers. The neural layers are inserted in the model to compensate for the unmodeled interactions, without requiring direct measurements of these unknown phenomena. In contrast to traditional approaches, both the neural network and physical parameters are simultaneously optimized, solely by using state and control input measurements. The methodology is applied on experimental data of a slider-crank setup, for which the state-dependent load interactions are unknown. The NNAP model proves to be a stable and accurate modeling formalism for dynamic systems that ab initio can only be partially described by physical laws. Moreover, the results show that a recurrent implementation of the NNAP model enables improved robustness and accuracy of the system state predictions, compared to its feedforward counterpart. Besides capturing the system dynamics, the NNAP model provides a means to gain new insights by extracting the neural network from the converged NNAP model. In this way, we discovered accurate representations of the unknown spring force interaction and friction phenomena acting on the slider mechanism.}},
  author       = {{De Groote, Wannes and Kikken, Edward and Hostens, Erik and Van Hoecke, Sofie and Crevecoeur, Guillaume}},
  issn         = {{1083-4435}},
  journal      = {{IEEE-ASME TRANSACTIONS ON MECHATRONICS}},
  keywords     = {{Control and Systems Engineering,Electrical and Electronic Engineering,Computer Science Applications,Mathematical model,Neural networks,Load modeling,Servomotors,Predictive models,Physics,Mechatronics,Explainable artificial intelligence,load identification,neural networks,nonlinear dynamic system modeling,IDENTIFICATION,DESIGN}},
  language     = {{eng}},
  number       = {{1}},
  pages        = {{103--114}},
  title        = {{Neural network augmented physics models for systems with partially unknown dynamics : application to slider-crank mechanism}},
  url          = {{http://doi.org/10.1109/tmech.2021.3058536}},
  volume       = {{27}},
  year         = {{2022}},
}

Altmetric
View in Altmetric
Web of Science
Times cited: