7.7 Modellierung und Simulation
Filtern
Dokumenttyp
- Vortrag (60)
- Zeitschriftenartikel (19)
- Beitrag zu einem Tagungsband (12)
- Forschungsdatensatz (7)
- Posterpräsentation (4)
- Dissertation (2)
- Preprint (1)
Schlagworte
- Model order reduction (8)
- Concrete (6)
- Model calibration (6)
- Multiscale methods (6)
- Variational multiscale method (6)
- Digital Twin (5)
- Digital twin (5)
- Model Bias (5)
- Proper generalized decomposition (5)
- Simulation Models (5)
Organisationseinheit der BAM
- 7 Bauwerkssicherheit (105)
- 7.7 Modellierung und Simulation (105)
- 5 Werkstofftechnik (5)
- 5.2 Metallische Hochtemperaturwerkstoffe (5)
- 5.5 Materialmodellierung (4)
- 7.1 Baustoffe (4)
- 7.4 Baustofftechnologie (4)
- 8 Zerstörungsfreie Prüfung (4)
- 9 Komponentensicherheit (4)
- 9.3 Schweißtechnische Fertigungsverfahren (4)
Paper des Monats
- ja (3)
Digitale Zwillinge bieten wertvolle Einblicke in das Verhalten von Bauwerken und ermöglichen eine fundierte Entscheidungsfindung. Durch den Einsatz von Simulationen, die auf physikalischen Gesetzen beruhen, ist es möglich, Vorhersagen über die Struktur auf der Grundlage zukünftiger oder hypothetischer Situationen zu treffen. Die Verwendung solcher Simulationen impliziert jedoch eine Reihe von Annahmen und Vereinfachungen, die unvermeidbare Fehler in die Vorhersagen einbringen. Die Quantifizierung dieser Unsicherheiten ist der Schlüssel für den Einsatz zuverlässiger digitaler Zwillinge von Brücken.
Der 3D Betondruck ermöglicht die effiziente und ressourcenschonendere Herstellung maßgeschneiderter Betonstrukturen. Trotz des großen Potentials gibt es bisher keine einheitlichen Normen und Standards für den sicheren und robusten 3D Betondruck. Der Beitrag zeigt das Potential numerischer Methoden und Modelle zur effizienten und ressourcenschonenden Design- und Prozessoptimierung des 3D Betondrucks. Dazu werden die aktuellen Forschungsarbeiten im Bereich der Entwicklung robuster und zuverlässiger numerischer Zwillinge, geeigneter Optimierungstools sowie digitaler Workflows diskutiert.
Die Entwicklung eines optimierten Designs für Bauwerke erfordert die Berücksichtigung sowohl des Materials als auch des Strukturdesigns. Ziel des Beitrages ist die Vorstellung eines Designprinzips basierend auf automatisierten Workflows, das am Beispiel eines Biegebalkendesign zur Reduzierung des Treibhauspotenzials vorgestellt wird. Eine ganzheitliche Optimierung berücksichtigt Material- und Strukturdesign. Es werden physikalische Modelle mit Ansätzen aus dem maschinellen Lernen kombiniert, die mit experimentellen Daten kalibriert bzw. trainiert werden. Eine besondere Bedeutung hat dabei die Berücksichtigung von Unsicherheiten.
Bias Identification Approaches for Model Updating of Simulation-based Digital Twins of Bridges
(2024)
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. The updated model can be eventually used for the extension of the service life of the bridge based on an accurate description of the structure. Nevertheless, the necessary assumptions and simplifications in these models irremediably introduce discrepancies between measurements and model response. We will prove that quantifying the extent of the uncertainties generated by said discrepancies provides a better understanding of the real system, enhances the model updating process, and creates more robust and trustworthy digital twins. Among others, we identify that the inclusion of the explicit bias term through a Bayesian inference framework corrects the tuned parameters, allows the identification of non-prescribed noise sources and enables the introduction of additional information in the system without modifying the simulation model. The performance of selected model bias identification approaches will be compared in the context of digital twins of bridges. The different methods will be applied to a representative demonstrator case based on 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 through monitorization and digitalization, especially through the implementation of digital twins.
The creation and use of Digital Twins of existing structures, such as bridges, implies precise digital replicas that accurately mirror their physical counterparts. Ensuring the trustworthiness of Digital Twins and facilitating informed decision-making necessitates a robust approach to Uncertainty Quantification (UQ). A suitable model-updating scheme is key in preserving the quality and robustness of simulation-based Digital Twins. Model bias, stemming from discrepancies between computational models and real-world systems, poses a significant challenge in achieving this goal. This study delves into the challenges posed by model bias within Bayesian updating of Digital Twins of bridges. Two alternative model bias identification methods —a modularized version of Kennedy and O’Hagan’s approach and another one based on Orthogonal Gaussian Processes — are evaluated in comparison with the classical Bayesian inference framework. A key innovation lies in the modification of the aforementioned approaches to incorporate additional information into the Digital Twin framework via the bias term. This enables the extension of the model non-intrusively, leveraging large pools of data inherent in Digital Twins. The study showcases the potential of this approach to correct predictions, quantify uncertainties, and enhance the system with previously untapped information. This underscores the importance of everaging available data within Digital Twins to identify deficiencies and guide potential future model improvements.
The RILEM TC 304-ADC has set up a large interlaboratory study on the mechanical properties of 3D printed concrete (ILS-mech). The study was prepared in 2022 by a preparation group leading to a Study Plan which the TC approved on 29 November 2022 (https://doi.org/10.14459/2023mp1705940). The ILS-mech was performed in 2023. The data was collected using a pre-prepared spreadsheet template. For data management, a database was derived and set-up in openBIS. The underlying Postgres database of openBIS was exported to the here-published SQLite database for sharing without maintaining a server. The structure of the database is described in (doi). The results are discussed in three associated papers focusing on the overall outcomes and evaluation of the procedures (doi), the compressive test results (doi), and the tensile test results (doi).
Optimization Framework for Powder Bed 3D Printed Concrete Structures under Various Constraints
(2024)
The fusion of 3D printing with concrete has transformed large scale construction, accelerating creative design possibilities. In recent years, multiple printing techniques have been developed and are still under investigation. The powder bed technique enables overhang constructions, thereby increasing design flexibility. To maximize the advantages of those printing techniques, a holistic approach is required taking topology optimization through finite element analysis into account. The challenge is to consider a range of constraints, from material to geometrical to manufacturing.
The current research focuses on the development of a holistic approach to large-scale 3D powder bed printed concrete structure optimization, navigating the in-tricacies of constraints, leveraging the advantages of powder-based printers, and integrating essential information from material tests. A workflow starting from geometry input and ending with printing instructions is derived, minimizing user interaction, and offering fast and reliable results for different structures. Starting from a geometric model of the global design with the required load cases, the geometry is optimized with respect to reduce the mass and thus CO2 emissions. Thereby, the topology optimization includes various constraints: stress con-straints (mitigating tension and limit compressive stress using Drucker-Prager based models), geometric constraints (preserving specific surfaces from optimi-zation), and manufacturing constraints (controlling member size and avoiding trapped powder). Different objective functions, like minimizing compliance or mass are considered in this study. The process receives input from upstream ma-terial tests, offering insights into strength, stiffness and possible anisotropy. This integration of material testing enhances the accuracy and reliability of the opti-mization, aligning the design with real-world behavior of 3D printed concrete. An important aspect of the workflow is the segmentation of the optimized global structure into substructures, aligning with the size limitations of the 3D powder-bed printer. As final output of the developed workflow, printer specific fabrica-tion instructions are generated for each substructure. The developed optimization framework is demonstrated in the design of a multiperson shading element.
The integration of advanced simulation models into industrial applications, especially in civil engineering, is a challenge that requires to develop new concepts. In particular, this is related to the validation of models using open access and FAIR data as well as objective methods for identifying model parameters and model comparison with the integration of both the models and their calibration into reproducible workflow systems.
The challenges are illustratively discussed for a variety of applications ranging from bridge monitoring over the structural behavior under blast loading up to the additive manufacturing of concrete. An important basis for model validation is the existence of machine-readable data structures for experimental data, e.g. based on semantic representations. In addition, platforms to share this structured data in a searchable format in addition with reproducible scientific workflows must be established.
Based on these tools, objective methods for model parameter estimation and validation of the models are a critical step in ensuring their accuracy and usefulness in industrial applications. Bayesian inference is one approach that can be used to optimize simulation parameters, improving the accuracy of the model while being able to estimate the uncertainty. Integrating these procedures in a real-time application with a bidirectional information exchange between the physical systems and the virtual simulation model leads to the concept of digital twins.
For complex simulation models, the computational efficiency is a significant barrier both for real-time applications and in the context of model calibration with potentially millions of required forward-model evaluations. As a consequence, reduced order modeling approaches such as the proper generalized decomposition method can be used to build an efficient to evaluate model - speeding up the online computation but requiring a precomputation in an offline phase.
Finally, a perspective for integrating material and structural design into a joint design optimization is given. The integration of advanced simulation models into industrial applications is a rapidly evolving field with significant potential for innovation and progress. From additive manufacturing of concrete to digital twins and reduced order modeling, there are many tools and approaches that can be used to improve industrial processes and drive progress in a variety of industries.
We introduce a novel adaptive Gaussian Process Regression (GPR) methodology for efficient construction of surrogate models for Bayesian inverse problems with expensive forward model evaluations. An adaptive design strategy focuses on optimizing both the positioning and simulation accuracy of training data in order to reduce the computational cost of simulating training data without compromising the fidelity of the posterior distributions of parameters. The method interleaves a goal-oriented active learning algorithm selecting evaluation points and tolerances based on the expected impact on the Kullback-Leibler divergence of surrogated and true posterior with a Markov Chain Monte Carlo sampling of the posterior. The performance benefit of the adaptive approach is demonstrated for two simple test problems.
A common engineering practice is to optimize the geometry of a structure by an iterative process, in which an objective function is minimized by systematically choosing the value of design variables and computing the value of the objective function many times.
However, regarding multiscale problems, this direct numerical approach is not feasible for a number of reasons.
The model has to take into account both the structural and unit cell (UC) scale, because the objective function to be minimized is a global quantity, while the parameters are related to the shape or material parameters of the UC.
In the absence of a clear separation of scales, the direct resolution of both scales in the numerical model leads to a significant increase in computational cost, which makes it impossible to repeatedly evaluate the model during the optimization process.
The main goal of this contribution is to overcome the aforementioned limitations and develop a new efficient computational framework for the optimal design of lattice structures.
To this end, parametric MOR is combined with DD methods.
The change in geometry of each UC within the lattice is modelled by introducing a transformation that maps the reference to the physical domain.
In offline simulations, suitable localized approximation spaces, that account for the change in geometry of the UC and its neighbours, are constructed using the concept of oversampling and random sampling.
These are then integrated into a dual DD framework to derive a robust and scalable solver, i.e. a fast-to-evaluate reduced order model (ROM).
The ROM is validated on the example of optimizing a concrete arc or pre-stressed multi beam structure minimizing its mass.
