TY - CONF A1 - Andres Arcones, Daniel A1 - Weiser, M. A1 - Koutsoureladkis, D.-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 - Athens, Greece DA - 12.06.2023 KW - Digital Twin KW - Simulation Models KW - Uncertainty Quantification KW - Model Bias PY - 2023 UR - https://2023.uncecomp.org/ AN - OPUS4-58226 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Agasty, Amit A1 - Costard, René A1 - Rosenbusch, Sjard Mathis A1 - Kadoke, Daniel A1 - Kind, Thomas A1 - Hicke, Konstantin A1 - Hüsken, Götz T1 - Structural behavior and damage assessment of a reinforced concrete wall by various NDT methods and embedded sensors under blast-loading N2 - A safety or security related assessment of explosions, accidental and intentional scenarios alike, often necessitate performance of replication-tests. Such test results are necessary to clarify the causes within the scope of forensic investigations. To gain important insights into the behavior of structures and materials under such loading, field tests may also be performed in accordance with different test standards. To determine the resistance of building-structures after explosions, estimation of the residual load-bearing capacity in addition to the assessment of dynamic structural response and damage to the building components is important. In most cases an evaluation of structural integrity is based only on the visual damage, resulting in an overestimation of the residual capacity. The Bundesanstalt für Materialforschung und -prüfung (BAM) operates the Test site for Technical Safety (TTS) on an area measuring about 12 km2 in the Federal State of Brandenburg for execution of true-to-scale explosion tests. At the TTS, building component testing was performed to assess the suitability of different non-destructive testing methods to characterize the dynamic structural response and damage resulting from the detonation of high explosives. Different blast-loading scenarios were realized by varying the net explosive mass and the standoff distance with all scenarios representing a near-field detonation. The test object was a reinforced concrete wall 2 m high, 2.5 m wide and 20 cm thick, fixed at both vertical edges. The dynamic loading of the wall was characterized with 8 piezoelectric pressure sensors flush-mounted on the front surface, thus measuring the reflected pressures from the shock wave. The tests were conducted with the aim of characterizing the global behavior of the wall under dynamic shock loading and the resulting local damage pattern, respectively. High speed digital image correlation was implemented in combination with multiple acceleration sensors to observe the rear surface of the wall to chart the dynamic deflection during the loading and to determine the residual deformation after the loading had ceased. In addition, one test specimen was instrumented with fiber optic sensor cables, both fixed to the rebars and embedded in the concrete-matrix, respectively. Firstly, these sensors were interrogated during the blast test by a distributed acoustic sensing (DAS) device using a particularly high sampling rate to measure the shock-induced vibrations in the structure with high temporal resolution. This delivers information on dynamics of compression and tension cycles from within the structure. Secondly, the local damage-pattern emerging during the series of blasts was determined via distributed fiber optic strain sensing (DSS) by interrogating the embedded fiber optic sensors with a high spatial resolution DSS device after each blast. This enabled the characterization of non-visual damage to the structure, in particular with regard to the formation of localized cracks in the concrete matrix. The DSS was further complimented by a structure-scanner based on ultrasonic measurements. Our contribution describes this new test approach in detail. Results of the three datasets, namely dynamic shock loading, global behavior of the test object and the local damage pattern will be presented. The suitability of the implemented measurement methods will be discussed in combination with the challenges in their application for technical safety evaluation of building components under explosive loading. T2 - 26th International Symposium on Military Aspects of Blast and Shock (MABS26) CY - Wollongong, Australia DA - 03.12.2023 KW - Blast tests KW - Reflected pressure KW - Embedded sensors KW - Distributed fiber optic sensors KW - Acceleration sensors KW - Digital image correlation KW - Ultrasonic structure-scanner PY - 2023 AN - OPUS4-58927 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Agasty, Amit A1 - Rosenbusch, Sjard Mathis A1 - Costard, René T1 - Numerical analysis of structures under blast loading N2 - Blast tests are indispensable for investigations of accidental or intentional explosions and to evaluate the level of protection to people and equipment within critical infrastructure. Current capabilities for detailed blast effects assessment are limited to performing full-scale field testing, which, for complex scenarios, are highly resource intensive. In this regard, reliable numerical simulations are an effective alternative option. A discussion of the scope and challenges of using numerical tools for a technical-safety assessment of reinforced concrete structures under blast loading is presented. Different coupling possibilities between shock wave simulations and structural simulations with the help of practical examples is given. An outlook on the development of new methods for structural simulations currently being researched at BAM concludes the presentation. T2 - Exerter Final Conference, 2023 CY - Soest, The Netherlands DA - 24.04.2023 KW - Coupling KW - Numerical Simulation KW - Blast KW - Structural Analysis KW - FSI PY - 2023 AN - OPUS4-57458 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -