Filtern
Dokumenttyp
Volltext vorhanden
- nein (4)
Gehört zur Bibliographie
- ja (4)
Schlagworte
- Baugrund (1)
- Building Information Modeling (1)
- Deep Learning (1)
- Diskrete-Elemente-Methode (1)
- Gauß-Zufallsfeld (1)
- Geostatistik (1)
- Geotechnik (1)
- Lagrange-Methode (1)
- Open Source (1)
- Python <Programmiersprache> (1)
Institut
In infrastructure planning and construction, modeling the subsoil and its associated uncertainty is a fundamental task of geotechnical engineers. However, probabilistic methods and tools for quantifying and displaying the uncertainty of the subsoil models are rarely used in practice where deterministic interpolation dominates. In digital planning using Building Information Modeling (BIM), the probabilistic approach supports creating a discipline model in which the uncertainties of the spatial layer structure are statistically quantified to evaluate the georisks in the design and execution of civil constructions. This article presents a case study using a combination of Sequential Gaussian Simulation (SGSIM) and Sequential Indicator Simulation (SISIM) to account for uncertainties in soil layer geometry. In a case study at the Munich Town Hall, a geostatistical approach is applied and validated based on 70 bore logs, whereby the probabilities for the occurrence of a particular layer are spatially quantified. The case study illustrates the methodology‘s great potential and benefits compared to the conventional deterministic approach based on interpolation procedures.
Die Prognose der Unsicherheiten in 3D-Baugrundmodellen für BIM verbessert die Risikobewertung und Entscheidungsfindung und ermöglicht eine wirtschaftlichere und nachhaltigere Planung und Ausführung von Baumaßnahmen. Basierend auf Open-Source-Software wird ein Ansatz zur Implementierung probabilistischer Baugrundmodelle im Industry-Foundation-Classes- (IFC-) Datenschema vorgeschlagen. Die Grundlagen von IFC, die für die Erstellung von Fachobjekt-Geometrien sowie deren Georeferenzierung und Attribuierung erforderlich sind, werden erläutert. Zwei probabilistische 3D-Modelle, die mit Bohrprofil- bzw. Cone-Penetration-Test- (CPT-) Daten erstellt wurden, dienen als Anwendungsbeispiele; diese zeigen die prognostizierte Geometrie der Bodenschichten mit den zugehörigen Unsicherheiten basierend auf Volume Pixel (Voxel). Solange Voxel-Modelle nicht in IFC integriert sind, bieten Isoflächen eine praktikable Zwischenlösung zur Darstellung von Bodenschichten und Unsicherheiten. Eine konsistente Georeferenzierung gewährleistet eine korrekte Positionierung im Koordinationsmodell. Die Modellierung von Baugrund- und geotechnischen Daten wird für verschiedene IFC-Versionen gezeigt. Die aktuelle IFC-Version 4 verbessert die Oberflächenmodellierung und reduziert die Datenredundanz. Der vorgestellte Ansatz ermöglicht eine maßgeschneiderte Implementierung probabilistischer Baugrundmodelle in IFC und eine effizientere Zusammenarbeit der an der Erstellung des BIM-Modells beteiligten Experten.
In conventional constitutive models for granular materials, calibration involves estimating a few parameters within known mathematical expressions. In contrast, data-driven constitutive models couple the model structure and parameters. Addressing this fundamental difference, the development of constitutive models based on Physics-encoded Neural Networks (PeNN) is guided from the perspective of conventional model development, highlighting similarities and differences. The crucial physical information that influences PeNN is explained, and the incorporation of three key state boundary surfaces in pressure-porosity space - critical state, loosest state, and densest state - via physics-informed deep learning is detailed. Physics-informed calibration is performed using the augmented Lagrangian method; then, the calibrated models undergo extensive drained and undrained simulations. Results indicate that using only physical information from state boundary surfaces, without data within these boundaries, fails to calibrate data-driven models; thus, boundary surface information represents partial physical information. While combining partial physical information with reasonably distributed data can improve model development under limited experimental data, adding more partial physical information and data does not necessarily enhance the results. The finding aims to bridge the gap between conventional and data-driven constitutive models, hopefully increasing the reliability and interpretability of data-driven models.
In this study, we investigate with the Discrete Element Method (DEM) the mechanical behavior of a cohesionless granular material under undrained true triaxial conditions, considering both monotonic and cyclic loading. We link the microstructure evolution within the granular assembly to its macroscopic cyclic response. To capture the mechanical response of our reference material (Karlsruhe fine sand), a rolling resistance linear contact model along with spherical particles is calibrated through a trial-and-error process, adjusting the model parameters to capture the experimentally observed behavior as close as possible. A series of cyclic undrained triaxial tests were simulated to investigate the micromechanical processes underlying liquefaction of sand under cyclic shearing. We analyzed the evolution of various fabric indices, including the redundancy index, contact normal orientations, and fabric anisotropy in relation to the pre- and post-liquefaction responses. The results reveal that a redundancy index below unity provides a unified criterion for the loss of the isostatic condition within the granular assembly, which triggers the onset of liquefaction. Throughout the cyclic loading process, sliding-dominant contact-yielding mechanisms remain prevalent. Additionally, significant changes in contact normal orientation and increasing fabric anisotropy dependent on the induced axial strain occur as the sample undergoes post-liquefaction deformation.