Filtern
Dokumenttyp
- Beitrag zu einem Tagungsband (21)
- Zeitschriftenartikel (12)
- Vortrag (12)
- Forschungsdatensatz (1)
Schlagworte
- SBFEM (12)
- Scaled Boundary Finite Element Method (8)
- Lamb waves (5)
- Scaled boundary finite element method (5)
- Ultrasonic guided waves (5)
- Ultrasound (5)
- NDT (4)
- Composites (3)
- Enriched finite element method (3)
- Geführte Ultraschallwellen (3)
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (12)
An essential task in many industries, e.g. food, petrol or chemical industry, is the precise and accurate characterization of liquids. Therefore, the development of innovative in-line sensors is of great interest. New concepts based on periodic structures, so-called phononic crystals (PnCs), are an interesting idea for the design of innovative sensors.
A PnC-based sensor can be designed by introducing a resonance inside a bandgap, a frequency region where no wave propagation is allowed. High-Q measurement systems using PnCs are already reported in the literature. However, existing designs cannot be implemented into a piping system directly, but need special fittings, openings or by-passes to be in contact with the liquid.
To circumvent this issue, we develop a new sensor based on PnCs, which can be directly implemented as part of the piping system. For this purpose, we use a PnC consisting of hollow cylinders with a periodic change of the outer diameter.
A bandgap could be found for the described geometry without fluid in simulation and measurement. However, simulations show, that a bandgap for fluid-filled cylinders can only be obtained for quasi-longitudinal modes. Hence, we propose a mode selective excitation for the sensor.
Data-driven analysis for damage assessment has a large potential in structural health monitoring (SHM) systems, where sensors are permanently attached to the structure, enabling continuous and frequent measurements. In this contribution, we propose a machine learning (ML) approach for automated damage detection, based on an ML toolbox for industrial condition monitoring. The toolbox combines multiple complementary algorithms for feature extraction and selection and automatically chooses the best combination of methods for the dataset at hand. Here, this toolbox is applied to a guided wave-based SHM dataset for varying temperatures and damage locations, which is freely available on the Open Guided Waves platform. A classification rate of 96.2% is achieved, demonstrating reliable and automated damage detection. Moreover, the ability of the ML model to identify a damaged structure at untrained damage locations and temperatures is demonstrated.
In Non-Destructive Testing, ultrasonic waves are commonly used to identify flaws and cracks.
In plates, shells, pipes and other geometries guided waves can be used to test the whole structure at once. In these tests, the input and response signal can have a nonlinear relationship due to cracks. At least for higher deflections, the propagating wave excites each side of the crack in such a way that it hits the other side. This clapping generally leads to a generation of higher harmonic waves and is referred to Contact Acoustic Nonlinearity (CAN). To get a better insight into the salient physics of the effect numerical simulations are necessary.
In the recent years, the Scaled Boundary Finite Element Method (SBFEM) was introduced to efficiently simulate wave propagation. The main advantage of the method is an easy grid generation process because the domain is discretized with arbitrary polygons instead of the triangles and rectangles. Another advantage is the possibility to model crack tips elegantly without additional workload. The SBFEM approach is still related to the Finite Element Method and uses similar techniques. The method is very efficient using high-order-spectral elements.
In this contribution, we present a short introduction into the basics of SBFEM formulation of the dynamic elastic wave equation. The SBFEM is then extended for modeling the non-linear behavior of crack clapping. Different approaches with increasing complexity are presented and assessed with respect to numerical stability.
In the field of guided waves for non-destructive testing, the interaction of these waves with damages or other discontinuities in a structure is critical. When a guided wave mode travels and hits a defect, it scatters in all directions, converting to other modes and reflecting the existing one. These interactions are captured in scattered far field complex amplitudes. The amplitudes are stored in scattering matrices, which characterise the elastodynamic behaviour of a defect completely. Scattering matrices are also useful to simulate backpropagation from a defect using ray-tracing methods.
Simulating these interactions is challenging, and analytical solutions only exist for simple geometries. Still, using general tools like the finite element method results in large, usually costly models. Recently, researchers proposed a method based on a numerical implementation of the Kirchhoff–Helmholtz integral that allows the computation of the scattering matrices using a model containing only the damaged region. However, classical methods to resolve the far field and low-order elements were used, leading to large models yet more efficient than using other techniques.
We propose using the SBFEM as an alternative to enhance the computation of the far field scattering. The damaged region is discretised using high-order polyhedral elements, while the far field is constructed using a modified version of the SBFEM. Examples compared to the literature demonstrate the validity of the approach.
In der Zerstörungsfreien Prüfung und der Zustandsüberwachung sind geführte Wellen von großem Interesse, um Fehlstellen zu finden und zu charakterisieren. Die Interaktion der Wellen mit den Fehlstellen kann dabei aufgrund ihrer Komplexität häufig nicht analytisch beschrieben werden. Dies macht numerische Programme neben Experimenten unabdingbar. Aufgrund der kleinen Wellenlänge im Verhältnis zur Bauteilgröße ist jedoch eine effiziente Simulation noch immer Teil der aktiven Forschung.
Um die Effizienz der Simulationsalgorithmen zu steigern, ist es möglich, analytische Annahmen in die Finite Elemente Methode (FEM) zu integrieren. Beispiele für solche Methoden sind die Scaled Boundary Finite Element Method als ein semi-analytisches Verfahren und eXtended Finite Element Method. Diese beiden Möglichkeiten erlauben es, die Wechselwirkungen effizient zu simulieren.
In diesem Beitrag werden verschiedene Wechselwirkungen und Auswertungsmöglichkeiten von geführten Wellen mit Fehlstellen vorgestellt. Hierbei liegt der Fokus auf linearen und nicht-linearen Effekten. Zunächst wird auf Modenkonversion eingegangen und der Frage nachgegangen, ob diese ausreicht, um einen Riss zu charakterisieren. Diese Untersuchungen motivieren dann ein inverses Verfahren, mit dem einige Parameter eines simulierten Risses in einer Folge von Simulationen wieder rekonstruiert werden. Ein zweiter Teil beschäftigt sich mit nicht-linearen Risseffekten. Diese Risseffekte erzeugen im allgemeinen höhere harmonische Wellen. Hier werden Ergebnisse und Filtermethoden zur Auswertung vorgestellt.
The dispersive properties of Lamb waves can be utilised for material characterisation because the frequency-wavenumber-relationship, as well as the group velocity, depend on material parameters. These dependencies make a non-destructive estimation of an elastic constant possible. This preliminary study investigates the sensitivity of dispersion curves caused by a change in elastic constants. The Scaled Boundary Finite Element Method is used to compute special dispersion curves, which show the sensitivity value of the frequency and group velocity as a colour value. This representation allows for easy identification of patterns and local effects. Two sets of dispersion curves are presented, one set for a steel plate and the other set for a plate made of a carbon fibre reinforced polymer. In general, we notice that the sensitivity often increases with the frequency and that higher-order modes seem to be more suitable for material characterisation. Moreover, specific modes respond to material changes while others are relatively unaffected, which must be taken into consideration for material characterisation.
Ein Hauptziel der zerstörungsfreien Prüfung und der Strukturüberwachung (engl. Structural Health Monitoring - SHM) mit Ultraschallwellen ist die Charakterisierung von Schäden in Bauteilen. In vielen schalenförmigen Bauteilen, wie zum Beispiel Rohrleitungen, Laminaten und Platten, breitet sich der Ultraschall in Form geführter Wellen aus. Zwar erlauben geführte Wellen eine großflächige Prüfung durch das langsame Abklingen der Wellen. Jedoch breiten sich die Wellen in verschiedenen dispersiven Moden aus, was die Analyse der vom Schaden erzeugten Reflexionen erschwert. Eine Möglichkeit, die Messsignale zu interpretieren, um Schäden zu charakterisieren, ist der direkte Vergleich mit einem Simulationsmodell. Die Rekonstruktion des Schadens stellt ein inverses Problem dar. Das inverse Problem kann als Optimierungsproblem formuliert werden. Für die Optimierung werden mehrere Vorwärtsrechnungen gebraucht, um das Schadensmodell an die Messdaten anzupassen.
Aufgrund der kurzen Wellenlängen von Ultraschallwellen sind klassische Methoden für die Vorwärtsrechnung, wie z.B. die Finite Elemente Methode (FEM), rechenintensiv. Eine Möglichkeit den Rechenaufwand zu reduzieren, bietet die Approximation der Wellenausbreitung mittels der semi-analytischen Scaled Boundary Finite Element Method (SBFEM). Frühere Untersuchungen haben gezeigt, dass die benötigten Freiheitsgrade im Vergleich zur FEM wesentlich geringer sind [1].
Im Beitrag wird eine Optimierung basierend auf einem Gradientenverfahren in Kombination mit der SBFEM vorgestellt und an verschiedenen Schadenstypen in 2D-Querschnittsmodellen von Stahlplatten getestet. Der Gradient des Vorwärtsmodells wird durch Algorithmisches Differenzieren berechnet, wodurch eine genaue und schnelle Optimierung ermöglicht wird. Es werden Untersuchungen zum inversen Problem und das Finden einer geeigneten Zielfunktion präsentiert. Es wird verdeutlicht, dass der entwickelte Algorithmus robust gegenüber von Rauscheinflüssen ist. In diesen Untersuchungen werden zunächst „Messdaten“ aus unabhängigen Simulationen verwendet [2]. Erste Schritte für die experimentelle Validierung und Erweiterung auf 3D Modelle werden anschließend vorgestellt.
Structural health monitoring techniques associate strongly with damage detection and characterization. Ultrasonic guided waves (UGW), for such scope, arise as one of the most promising methods for many reasons i.e. UGW are able to travel long distances and they have high sensitivity to damage. In this context, the necessity to model realistic wave-defect interaction occurs to be critical.
Realistic damage scenarios can be modeled through the usage of image-based quadtree meshes. Images, such as the outcome from X-ray scans, C-scans, etc., can be converted into meshes for further integration in a computational domain. Quadtree meshes are created by converting the intensity of the pixels to quadrilateral cells. Homogeneous regions inside one image result in one quad, whereas fine features such as discontinuities can be described with smaller quads.
This contribution proposes an efficient methodology to model wave defect interaction, using as a framework the scaled boundary finite element method (SBFEM) and quadtree meshes. Problems as non-conforming regions in the mesh due to the space tree decomposition can be easily avoided using SBFEM’s polygonal elements. Moreover, the semi-analytical nature of the SBFEM allows the modeling of arbitrarily long prismatic/undamaged regions of the waveguides without an increase in the computational burden.
Wavefield measurements by a scanning laser Doppler vibrometer are generally carried out in a cartesian coordinate. As a piezoelectric transducer generates Lamb waves following radial paths, the use of a polar coordinate can be a suitable alternative to the use of a cartesian coordinate. Therefore, in the proposed method, using a single transducer placed on the center of the specimen, the measured wavefields are transformed into polar coordinates, making several identical radial line inspections from the center in a direction of incident waves. Taking advantage of the properties of the polar coordinates, a signal processing technique is proposed through a frequency-wavenumber filtering process in these coordinates. In this technique, by using proper filters, unwanted wave modes of the incident wave along with all reflected waves are filtered out. In addition, the conventional features of RMS and Euclidean distance are adapted for the polar coordinate system to image the bonded plate. The proposed signal processing and damage imaging are first introduced through a numerical simulation. Then, the performance of the proposed technique is presented by experimental measurements of two specimens including adhesively bonded carbon fiber-reinforced plastic composite plates and bonded aluminum plates.