Simulation-Based Inference for the Amortized Determination of Physical Model Parameters of a Mechanical System for Diagnostics and Prognostics
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
Abstract
Many diagnostic and prognostic applications rely on complex measurements, including time series and high-dimensional
data. In a common approach, one extracts key features that capture the system degradation while ignoring nuisance effects. Physical system properties are of interest due to their (direct) relation and relevance to degradation and failure modes, often allowing for superior interpretability compared to purely data-driven approaches. However, determining them
from the observed data is difficult due to the inherent non-identifiability in already rather simple models. Recently developed simulation-based inference (SBI) approaches, based on neural posterior estimates (NPE) and conditional invertible neural networks (cINN), allow the incorporation of domain knowledge in the form of simulation capabilities to extract model parameters as physics-based features for diagnostics and prognostics. This is demonstrated in the case of a mechanical actuator that operates medium voltage breakers. Simulations of a simplified multi-body model are used as input for the training of a cINN that not only provides a set of physical parameter values but also their respective uncertainties. By using simulated data as synthetic measurements and
conducting a number of statistical checks, the performance of the trained cINN is confirmed. We demonstrate that accurate multi-body parameter estimation is possible for some parameters, whereas those that cannot be identified from either the
opening or closing motion remain distributed according to the prior distribution, without significantly affecting the others.
We further show that the opening and closing motions are sensitive to different parameters, complementing each other
in this respect. Data from a real mechanical endurance test are used to demonstrate the method’s effectiveness in a real-
world application. Its integration into a diagnostics or prognostics framework is discussed as an outlook.
How to Cite
##plugins.themes.bootstrap3.article.details##
simulation-based inference, physics-based diagnostics, multi-body mechanical model, parameter estimation, Bayesian inference, neural-posterior estimate, Bayesian workflow
Baniasadi, A. A., Darijani, H., & Matin, M. (2025). Analytical multi-body approach for component selection and timing analysis in spring-actuated vacuum circuit breakers. Scientific Reports, 15(1), 42478. doi: 10.1038/s41598-025-26609-6
BayesFlow Authors. (2023–2026). BayesFlow web page. Retrieved from https://www.bayesflow.org
Cranmer, K., Brehmer, J., & Louppe, G. (2020). The frontier of simulation-based inference. Proceedings of the National Academy of Sciences, 117(48), 30055–30062. doi: 10.1073/pnas.1912789117
Gelman, A., Vehtari, A., Simpson, D., Margossian, C. C., Carpenter, B., Yao, Y., ... Modrák, M. (2020). Bayesian workflow. arXiv preprint arXiv:2011.01808.
Hencken, K., Borrelli, E.-M., Ceccarelli, D., & Krivda, A. (2022). Approximate Bayesian computation as a new tool for partial discharge analysis of partial discharge data. PHM Society European Conference, 7, 182–192.
Hoffmann, M. W., Wildermuth, S., Gitzel, R., Boyaci, A., Gebhardt, J., Kaul, H., ... Tornede, T. (2020). Integration of novel sensors and machine learning for predictive maintenance in medium voltage switchgear to enable the energy and mobility revolutions. Sensors, 20(7), 2099. doi: 10.3390/s20072099
Janssen, A., Makareinis, D., & Sölver, C.-E. (2014). International surveys on circuit-breaker reliability data for substation and system studies. IEEE Transactions on Power Delivery, 29(2), 808–814. doi: 10.1109/TPWRD.2013.2274750
Kobyzev, I., Prince, S. J., & Brubaker, M. A. (2020). Normalizing flows: An introduction and review of current methods. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(11), 3964–3979.
Li, Y., Lin, G., Lau, T., & Zeng, R. (2019). A review of changepoint detection models. arXiv preprint arXiv:1908.07136.
Lipman, Y., Chen, R. T. Q., Ben-Hamu, H., Nickel, M., & Le, M. (2023). Flow matching for generative modeling. In Proceedings of the 11th International Conference on Learning Representations (ICLR 2023).
Mukin, R., Hencken, K., & Cilibrasi, M. (2026). Application of the Bayesian online change-point detection algorithm to the degradation of an electrical device. In Proceedings of the PHM Europe Conference.
Papamakarios, G., Nalisnick, E., Rezende, D. J., Mohamed, S., & Lakshminarayanan, B. (2021). Normalizing flows for probabilistic modeling and inference. Journal of Machine Learning Research, 22, 1–64.
Radev, S. T., Mertens, U. K., Voss, A., Ardizzone, L., & Köthe, U. (2020). BayesFlow: Learning complex stochastic models with invertible neural networks. arXiv preprint arXiv:2003.06281.
Radev, S. T., Schmitt, M., Pratz, V., Picchini, U., Köthe, U., & Bürkner, P.-C. (2023). JANA: Jointly amortized neural approximation of complex Bayesian models. In Proceedings of the 39th Conference on Uncertainty in Artificial Intelligence (Vol. 216, pp. 1695–1706).
Radev, S. T., Schmitt, M., Schumacher, L., Elsemüller, L., Pratz, V., Schälte, Y., ... Bürkner, P.-C. (2023). BayesFlow: Amortized Bayesian workflows with neural networks. Journal of Open Source Software, 8(89), 5702.
Sisson, S. A., Fan, Y., & Beaumont, M. (2018). Handbook of approximate Bayesian computation. CRC Press.
Stanek, L., & Fröhlich, K. (2000). Model-aided diagnosis: A new method for online condition assessment of high-voltage circuit breakers. IEEE Transactions on Power Delivery, 15(2), 585–591.
Stanek, M., Morari, M., & Fröhlich, K. (2001). Model-aided diagnosis: An inexpensive combination of model-based and case-based condition assessment. IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews, 31(2), 137–145.
Talts, S., Betancourt, M., Simpson, D., Vehtari, A., & Gelman, A. (2020). Validating Bayesian inference algorithms with simulation-based calibration. arXiv preprint arXiv:1804.06788.
Wieland, F.-G., Hauber, A. L., Rosenblatt, M., Tönsing, C., & Timmer, J. (2021). On structural and practical identifiability. Current Opinion in Systems Biology, 25, 60–69. doi: 10.1016/j.coisb.2021.03.005
Wildberger, J., Dax, M., Buchholz, S., Green, S., Macke, J. H., & Schölkopf, B. (2023). Flow matching for scalable simulation-based inference. Advances in Neural Information Processing Systems, 36, 16837–16864.
Xie, Y., & Cheng, X. (2025). Flow-based generative models as iterative algorithms in probability space. arXiv preprint arXiv:2502.13394.
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.