Eigenverlag BAM
Filtern
Dokumenttyp
- Dissertation (2)
Referierte Publikation
- nein (2)
Schlagworte
- ultrasound (2) (entfernen)
Die Notwendigkeit und Nachfrage von zerstörungsfreien Prüfverfahren für Beton, die schnell scannend automatisiert große Messflächen abtasten können, wird zunehmend größer. Ein großes Potential bietet das Ultraschallecho-Verfahren, mit dem sich eine Vielzahl von baupraktischen Fragestellungen bei der Überprüfung des Bauwerkszustandes beantworten lassen. Durch Verwendung luftgekoppelter Ultraschallprüfköpfe (ACU-Prüfköpfen), die berührungslos über das zu untersuchende Betonbauteil geführt werden, könnte dieses Ziel erreicht werden. In der vorliegenden Arbeit wird die Machbarkeit von luftgekoppeltem Ultraschallecho (ACU- Echo) an Betonbauteilen von 20 cm Dicke gezeigt. Um den Einfluss der Parameter auf die Wellenausbreitung zu studieren, werden experimentelle Untersuchungen von ACU-Echo an speziellen Plexiglas- und Betonprobekörpern durchgeführt. Dabei kommt auf der Empfangsseite auch ein Laservibrometer zum Einsatz. Mit dessen Hilfe wird die Wellenausbreitung des eingetragenen Luftultraschalls im Beton visualisiert. Durch eine besondere digitale Signalbearbeitung können verschiedene Wellenarten voneinander getrennt und interpretiert werden. Mit den gewonnenen Ergebnissen wird ein Scanner für ACU-Echo-Messungen aufgebaut, mit dem weitere wesentliche Erkenntnisse gewonnen werden und ein erstes praxisorientiertes Anwendungsbeispiel demonstriert wird.
Ultrasonic guided waves offer a wide range of applications in fields such as non-destructive testing, structural health monitoring or material characterization. They can be excited in thin-walled structures and propagate over comparably long distances. Due to their complex and dispersive propagation behavior, numerical methods are often required in order to analyze the guided wave modes that can be excited in a given structure and to simulate their interaction with defects. In the work presented in this thesis, highly efficient numerical methods have been developed that are specifically optimized for guided wave problems. The formulation is based on the Scaled Boundary Finite Element Method (SBFEM). The SBFEM is a semi-analytical method which evolved from the concept of Finite Elements but requires the discretization of the boundary of the computational domain only. To compute dispersion curves and mode shapes of guided waves, only the cross-section of the waveguide is discretized in the Finite Element sense, while the direction of propagation is described analytically. The wavenumbers of guided wave modes and the corresponding mode shapes are obtained as the eigenvalues and eigenvectors of a frequency-dependent Hamiltonian matrix. For the discretization, higher-order spectral elements are employed, leading to very low computational costs compared to traditional Finite Elements. Particular formulations are presented for plate structures as well as axisymmetric waveguides, where only the throughthickness direction has to be discretized. For the cases where the waveguide is embedded in or coupled to a quasi-infinite medium, a dashpot boundary condition is proposed in order to account for the effect of waves being transmitted into the surrounding medium. Though this approach is not exact, it leads to sufficiently accurate results for practical applications, while the computational costs are typically reduced by several orders of magnitude compared to other Finite Element based approaches. As a particular application, an experimental set-up for material characterization is discussed, where the elastic constants of the waveguide’s material are obtained from the analysis of waves propagating through the waveguide. A novel solution procedure is proposed in this work, where each mode of interest is traced over the required frequency range. The solutions are obtained by means of inverse iteration. To demonstrate the potential of the SBFEM for non-destructive testing applications, the interaction of guided wave modes with cracks in plates is simulated in the time domain for several examples. Particularly for the modeling of cracked structures, the SBFEM is very well suited, since the side-faces of the crack do not require discretization and the stress-singularity at the crack tip does not introduce additional difficulties. Hence, the computational costs can be reduced by typically a factor 100 compared to traditional Finite Elements and the meshing is straightforward.