TY - CONF A1 - Andrés Arcones, Daniel A1 - Weiser, M. A1 - Koutsourelakis, P-S. A1 - Unger, Jörg F. T1 - Quantifying the uncertainty of predictive simulations in digital twins through the identification of model bias N2 - This work presents a novel approach to quantifying uncertainty in digital twin simulations by addressing model bias through embedded parameter distributions. Traditional Bayesian methods often underestimate uncertainty due to assumptions of model correctness. We propose a hierarchical Bayesian framework combined with Polynomial Chaos Expansion to better capture and propagate uncertainty. The methodology is validated on an analytical example and a real-world case involving thermal deformation predictions of the Nibelungen Bridge, demonstrating improved predictive accuracy and reliability. T2 - fib Symposium 2025 CY - Antibes, France DA - 16.06.2025 KW - Digital Twins KW - Model Bias KW - Predictive simulations KW - Quantifying the uncertainty PY - 2025 SP - 2867 EP - 2873 PB - The fib, Fédération international du béton CY - Antibes, France AN - OPUS4-63629 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Andrés Arcones, Daniel A1 - Weiser, M. A1 - Koutsourelakis, P.-S. A1 - Unger, Jörg F. T1 - Model bias identification for Bayesian calibration of stochastic digital twins of bridges N2 - Simulation-based digital twins must provide accurate, robust, and reliable digital representations of their physical counterparts. Therefore, quantifying the uncertainty in their predictions plays a key role in making better-informed decisions that impact the actual system. The update of the simulation model based on data must then be carefully implemented. When applied to complex structures such as bridges, discrepancies between the computational model and the real system appear as model bias, which hinders the trustworthiness of the digital twin and increases its uncertainty. Classical Bayesian updating approaches aimed at inferring the model parameters often fail to compensate for such model bias, leading to overconfident and unreliable predictions. In this paper, two alternative model bias identification approaches are evaluated in the context of their applicability to digital twins of bridges. A modularized version of Kennedy and O'Hagan's approach and another one based on Orthogonal Gaussian Processes are compared with the classical Bayesian inference framework in a set of representative benchmarks. Additionally, two novel extensions are proposed for these models: the inclusion of noise-aware kernels and the introduction of additional variables not present in the computational model through the bias term. The integration of these approaches into the digital twin corrects the predictions, quantifies their uncertainty, estimates noise from unknown physical sources of error, and provides further insight into the system by including additional pre-existing information without modifying the computational model. KW - Gaussian process KW - KOH KW - Bayesian updating KW - Digital twins KW - Uncertainty quantification PY - 2024 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-615519 DO - https://doi.org/10.1002/asmb.2897 SN - 1526-4025 N1 - This work was supported by “C07 - Data driven model adaptation for identifying stochastic digital twins of bridges” from the Priority Program (SPP) 2388/1 “Hundred plus” of the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) - Project number 501811638. VL - 417 IS - 3 SP - 1 EP - 26 PB - Wiley CY - Chichester AN - OPUS4-61551 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Andrés Arcones, Daniel A1 - Weiser, M. A1 - Koutsourelakis, F.-S. A1 - Unger, Jörg F. T1 - Evaluation of Model Bias Identification Approaches Based on Bayesian Inference and Applications to Digital Twins N2 - In recent years, the use of simulation-based digital twins for monitoring and assessment of complex mechanical systems has greatly expanded. Their potential to increase the information obtained from limited data makes them an invaluable tool for a broad range of real-world applications. Nonetheless, there usually exists a discrepancy between the predicted response and the measurements of the system once built. One of the main contributors to this difference in addition to miscalibrated model parameters is the model error. Quantifying this socalled model bias (as well as proper values for the model parameters) is critical for the reliable performance of digital twins. Model bias identification is ultimately an inverse problem where information from measurements is used to update the original model. Bayesian formulations can tackle this task. Including the model bias as a parameter to be inferred enables the use of a Bayesian framework to obtain a probability distribution that represents the uncertainty between the measurements and the model. Simultaneously, this procedure can be combined with a classic parameter updating scheme to account for the trainable parameters in the original model. This study evaluates the effectiveness of different model bias identification approaches based on Bayesian inference methods. This includes more classical approaches such as direct parameter estimation using MCMC in a Bayesian setup, as well as more recent proposals such as stat-FEM or orthogonal Gaussian Processes. Their potential use in digital twins, generalization capabilities, and computational cost is extensively analyzed. T2 - 5th ECCOMAS Thematic Conference on Uncertainty Quantificationin Computational Sciences and Engineering CY - Athen, Greece DA - 12.06.2023 KW - Model bias KW - Bayesian Uncertainty Quantification KW - Digital Twins KW - Gaussian Processes KW - Statistical Finite Element Method PY - 2023 UR - https://2023.uncecomp.org/ SP - 1 EP - 15 AN - OPUS4-58227 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - INPR A1 - Andrés Arcones, Daniel A1 - Weiser, M. A1 - Koutsourelakis, P-S. A1 - Unger, Jörg F. T1 - Embedded Model Bias Quantification with Measurement Noise for Bayesian Model Calibration N2 - A key factor in ensuring the accuracy of computer simulations that model physical systems is the proper calibration of their parameters based on real-world observations or experimental data. Inevitably, uncertainties arise, and Bayesian methods provide a robust framework for quantifying and propagating these uncertainties to model predictions. Nevertheless, Bayesian methods paired with inexact models usually produce predictions unable to represent the observed datapoints. Additionally, the quantified uncertainties of these overconfident models cannot be propagated to other Quantities of Interest (QoIs) reliably. A promising solution involves embedding a model inadequacy term in the inference parameters, allowing the quantified model form uncertainty to influence non-observed QoIs. This paper introduces a more interpretable framework for embedding the model inadequacy compared to existing methods. To overcome the limitations of current approaches, we adapt the existing likelihood models to properly account for noise in the measurements and propose two new formulations designed to address their shortcomings. Moreover, we evaluate the performance of this inadequacy-embedding approach in the presence of discrepancies between measurements and model predictions, including noise and outliers. Particular attention is given to how the uncertainty associated with the model inadequacy term propagates to the QoIs, enabling a more comprehensive statistical analysis of prediction’s reliability. Finally, the proposed approach is applied to estimate the uncertainty in the predicted heat flux from a transient thermal simulation using temperature observations. KW - Model bias KW - Bayesian inference KW - Noise KW - Model updating KW - Quantity of Interest PY - 2025 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-652720 DO - https://doi.org/10.48550/arXiv.2410.12037 SP - 1 EP - 37 PB - arXiv.org AN - OPUS4-65272 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Anker, Andy S. A1 - Aspuru-Guzik, Alán A1 - Bechtel, Tim A1 - Bigi, Filippo A1 - Briling, Ksenia R. A1 - Das, Basita A1 - David, Nicholas A1 - Day, Graeme M. A1 - Deringer, Volker L. A1 - Dyer, Matthew A1 - Eardley-Brunt, Annabel A1 - Evans, Matthew L. A1 - Evans, Rob A1 - Franklin, Barnabas A. A1 - Ganose, Alex M. A1 - George, Janine A1 - Goulding, Mark A1 - Hickey, Niamh A1 - James, Gillian A1 - Kalikadien, Adarsh V. A1 - Kapil, Venkat A1 - Kulik, Heather J. A1 - Kumar, Vishank A1 - Kuttner, Christian A1 - Lam, Erwin A1 - Lederbauer, Magdalena A1 - Lou, Yuchen A1 - Martin, Jennie A1 - Marulanda Bran, Andres A1 - Mathea, Miriam A1 - Pickard, Chris J. A1 - Ruscic, Branko A1 - Ryder, Matthew R. A1 - Sabanza Gil, Victor A1 - Schwaller, Philippe A1 - Segler, Marwin H. S. A1 - Sun, Wenhao A1 - Tanovic, Sara A1 - Treyde, Wojtek A1 - Walsh, Aron A1 - Wu, Ruiqi T1 - Discovering synthesis targets: General discussion N2 - This article is a discussion of the paper "Analysis of uncertainty of neural fingerprint-based models" by Christian W. Feldmann, Jochen Sieg and Miriam Mathea (Faraday discussions, 2025, DOI: 10.1039/D4FD00095A). KW - Automation KW - Materials acceleration platforms KW - Machine learning KW - Materials design KW - Materials discovery KW - Density functional theory KW - Ab initio PY - 2025 DO - https://doi.org/10.1039/D4FD90064B SN - 1359-6640 SN - 1364-5498 VL - 256 IS - Themed collection: Data-driven discovery in the chemical sciences SP - 639 EP - 663 PB - Royal Society of Chemistry (RSC) CY - Cambridge AN - OPUS4-62317 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Andres, S. A1 - Hansen, Ulf A1 - Niemann, B. A1 - Palavinskas, R. A1 - Lampen, A. T1 - Determination of the isoflavone composition and estrogenic activity of commercial dietary supplements based on soy or red clover N2 - Dietary supplements high in isolated isoflavones are commercially available for human consumption primarily to alleviate menopausal symptoms in women. The isoflavone composition, quantity and importantly their estrogenic potency are poorly standardised and can vary considerably between different products. The aim of this study was to analyse the isoflavone composition of 11 dietary supplements based on soy or red clover using the HPLC/MS/MS technique. Furthermore, we investigated the transactivational potential of the supplements on the estrogen receptors (ER), ERα and ERβ, performing luciferase reporter gene assays. As expected, we found that the isoflavone composition varies between different products. The measured total isoflavone contents in various supplements were mostly comparable to those claimed by the manufacturers in their product information. However expressing the isoflavone content as isoflavone aglycone equivalents, soy-based supplements had a clearly lower quantity compared to the manufacturer information. All supplements transactivated more or less ERα and ERβ with a preference for ERβ. The transactivational efficiency exceeded partly the maximal 17β-estradiol induced ER activation. While the different soy-based supplements revealed similar transactivation potential to both ERs, red clover-based supplements differed considerably. We conclude that different commercial dietary supplements based on soy or red clover vary in their isoflavone composition and quantity. They are estrogenically active, although especially the red clover-based supplements show considerable differences in their estrogenic potential to ERα and ERβ. Thus, different isoflavone-rich products cannot be necessarily compared regarding possible biological effects. KW - estrogenic activity KW - HPLC/MS/MS technique KW - Estrogen receptor (ER) KW - nutritional supplements PY - 2015 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-357093 DO - https://doi.org/10.1039/C5FO00308C IS - 6 SP - 2017 EP - 2025 PB - The Royal Soc. of Chemistry CY - Cambridge AN - OPUS4-35709 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Andrés Arcones, Daniel A1 - Weise, M. A1 - Koutsourelakis, P-S. A1 - Unger, Jörg F. T1 - Bias Identification Approaches for Model Updating of Simulation-based Digital Twins of Bridges N2 - Simulation-based digital twins of bridges have the potential not only to serve as monitoring devices of the current state of the structure but also to generate new knowledge through physical predictions that allow for better-informed decision-making. For an accurate representation of the bridge, the underlying models must be tuned to reproduce the real system. Nevertheless, the necessary assumptions and simplifications in these models irremediably introduce discrepancies between measurements and model response. We will show that quantifying the extent of the uncertainties introduced through the models that lead to such discrepancies provides a better understanding of the real system, enhances the model updating process, and creates more robust and trustworthy digital twins. The inclusion of an explicit bias term will be applied to a representative demonstrator case based on the thermal response of the Nibelungenbrücke of Worms. The findings from this work are englobed in the initiative SPP 100+, whose main aim is the extension of the service life of structures, especially through the implementation of digital twins. T2 - EWSHM 2024 11th European Workshop on Structural Health Monitoring CY - Potsdam, Germany DA - 10.06.2024 KW - Digital Twins KW - Model Bias KW - SPP100+ KW - Bridge Monitoring PY - 2024 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-622522 DO - https://doi.org/10.58286/30524 SN - 2941-4989 IS - 12 SP - 1 EP - 10 PB - NDT.net GmbH & Co. KG CY - Mayen, Germany AN - OPUS4-62252 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -