Ingenieurwissenschaften und zugeordnete Tätigkeiten
Filtern
Dokumenttyp
- Zeitschriftenartikel (222)
- Vortrag (219)
- Posterpräsentation (81)
- Beitrag zu einem Tagungsband (42)
- Dissertation (5)
- Forschungsdatensatz (5)
- Buchkapitel (3)
- Video (2)
- Beitrag zu einem Sammelband (1)
- Tagungsband (Herausgeberschaft für den kompletten Band) (1)
Sprache
- Englisch (582) (entfernen)
Schlagworte
- Additive manufacturing (119)
- Additive Manufacturing (89)
- Residual stress (53)
- Thermography (51)
- Computed tomography (48)
- Laser powder bed fusion (37)
- Microstructure (28)
- X-ray refraction (28)
- Residual Stress (27)
- Laser Powder Bed Fusion (26)
Organisationseinheit der BAM
- 8 Zerstörungsfreie Prüfung (582) (entfernen)
Paper des Monats
- ja (4)
Since laser powder bed fusion (PBF-LB/M) is prone to the formation of defects during the building process, a fundamental requirement for widespread application is to find ways to assure safety and reliability of the additively manufactured parts. A possible solution for this problem lies in the usage of in-situ thermographic monitoring for defect detection. In this contribution we investigate possibilities and limitations of the VIS/NIR wavelength range for defect detection. A VIS/NIR camera can be based on conventional silicon-based sensors which typically have much higher spatial and temporal resolution in the same price range but are more limited in the detectable temperature range than infrared sensors designed for longer wavelengths. To investigate the influence, we compared the thermographic signatures during the creation of artificially provoked defects by local parameter variations in test specimens made of a nickel alloy (UNS N07208) for two different wavelength ranges (~980 nm and ~1600 nm).
Components produced using additive manufacturing can be marked for unique identification and secure authentication [1,2]. Serial numbers and machine-readable codes can be used to identify the component, and link digital product-related data (i.e., a digital product passport) to the actual components. The most prevailing solution consists of local process manipulation, such as printing a quick response (QR) code [3] or a set of blind holes on the surface of the internal cavity of hollow components. However, local manipulation of components may alter the properties, and external tagging features can be altered or even removed by post-processing treatments. This work therefore aims to provide a new methodology for identification, authentication, and traceability of additively manufactured (AM) components using microstructural features that are unique to each part. X-ray computed tomography (XCT) was employed to image the microstructural features of AlSi10Mg parts. Based on size and geometry, the most prominent features were selected to create a unique digital authenticator. We implemented a framework in Python using open-access modules that can successfully create a digital object authenticator using the segmented microstructure information from XCT. The authenticator is stored as a QR code, along with the 3D information of the selected features.
The integration of additive manufacturing with traditional processes, termed hybrid additive manufacturing, has expanded its application domain, particularly in the repair of gas turbine blade tips. However, process-related defects in additively manufactured materials, interface formation, and material property mismatches in dual-material structures can significantly impact the fatigue performance of components. This investigation examines the low cycle fatigue and fatigue crack growth behaviors in dual-material specimens of nickel-based alloys, specifically the additively manufactured STAL15 and the cast alloy 247DS, at elevated temperatures. Low cycle fatigue experiments were conducted at temperatures of 950 °C and 1000 °C under a range of strain levels (0.3%–0.8%) and fatigue crack growth tests were conducted at 950 °C with stress ratios of 0.1 and −1. Fractographic and microscopic analyses were performed to comprehend fatigue crack initiation and crack growth mechanisms in the dual-material structure. The results consistently indicated crack initiation and fatigue fracture in the additively manufactured STAL15 material. Notably, fatigue crack growth retardation was observed near the interface when the crack extended from the additively manufactured STAL15 material to the perpendicularly positioned interface. This study highlights the importance of considering yield strength mismatch, as well as the potential effects of residual stresses and grain structure differences, in the interpretation of fatigue crack growth behavior at the interface.
The first Non-destructive testing (NDT) method which evolved in the industrial age was radiographic testing (RT). Among all NDT methods, RT is no exception, so there are still many issues for optimizations even today. One of them is the measurement of the focal spot of X-ray tubes. The size of the focal spot is critical for imaging, because it determines the spatial resolution in the X-ray image. The classical way to evaluate focal spots of X-ray tubes is by pinhole imaging using a camera obscura. But this method has a natural lower limit, which is defined by the diameter of the pinhole used (today min. 10 µm). Therefore, focal spot sizes lower than 50 µm diameter cannot be imaged and measured correctly. An alternative approach, which permits this, was investigated here using the edge unsharpness of holes much larger than the focal spot size. The results of both methods were compared using 3 different X-ray tubes.
Additive manufacturing techniques, such as laser powder bed fusion (PBF-LB), are well known for their exceptional freedom in part design. However, these techniques are also characterized by the development of large thermal gradients during production and thus residual stress (RS) formation in produced parts. In this context, neutron diffraction enables the non-destructive characterization of the bulk RS distribution. By control of the thermal gradients in the powder-bed plane by scan strategy variation we study the impact of in-process scan strategy variations on the microstructure and the three-dimensional distribution of RS. Microstructural analysis by means of electron backscatter diffraction reveals sharp microstructure transitions at the interfaces ranging from 100-200 µm. The components of the RS tensor are determined by means of neutron diffraction and the principal stress directions and magnitudes are determined by eigenvalue decomposition. We find that the distribution of RS in the powder-bed plane corresponds to the underlying scan strategy. When the alternating scan vectors align with the x- and y sample coordinate axes, the principal stress directions co-align. In the present geometry, nearly transverse isotropic stress states develop when the scan vectors are either aligned 45° between x and y or continuously rotated by 67° between each layer.
Many modern ultrasonic methods in the fields of Non-Destructive Testing (NDT) and Structural Health Monitoring (SHM) require simulations in research. Researchers either use simulation data initially during development to investigate certain aspects, or the simulation process is directly part of the research task. Examples of the second case are inverse methods for parameter estimation, model-assisted probability of detection analysis or the generation of training data for AI algorithms. All these applications require algorithms that are as efficient as possible. For methods based on explicit time-step methods, a significant increase in efficiency can be achieved if a so-called lumped mass matrix can be used, which approximates the consistent mass matrix but is easier to invert.
The finite element method has been the subject of many studies on approximations of the mass matrix. In contrast, the lumped mass matrix in the context of the Scaled Boundary Finite Element Method (SBFEM) is a current field of research [1,2]. In the time domain, the semi-analytical SBFEM is notable for its flexibility to be applied to polygonal meshes. In particular, image-based mesh generation using a quadtree algorithm is possible. In general, polygonal meshes have the same flexibility as triangular meshes, but polygonal meshes can have additional advantages such as greater tolerance to distortion.
In this contribution, the SBFEM formulation based on bubble functions [3] for the time domain is presented for two-dimensional elastic waves. The adjustments necessary for a good approximating lumped mass matrix are emphasized. Several grid generation methods for polygonal elements are shown. Figure 1 depicts the difference between the consistent mass matrix and the lumped mass matrix for a normal polygonal mesh. Finally, the accuracy of mass lumping for linear, quadratic and cubic shape functions is presented and the computational efficiency is demonstrated using exemplary waveguide geometries.
Additive manufacturing (AM, also known as 3D printing) of metals is becoming increasingly important in industrial applications. Reasons for this include the ability to realize complex component designs and the use of novel materials. This distinguishes AM from conventional manufacturing methods such as subtractive manufacturing (turning, milling, etc.). The most widely used AM process for metals is laser powder bed fusion (PBF-LB/M, also known as selective laser melting SLM). Currently, it has the highest degree of industrialization and the largest number of machines in use. In PBF-LB/M, the feedstock is present as metal powder in an inert gas atmosphere inside a process chamber where a laser melts it locally. By repeatedly lowering the build platform, applying a new layer of powder, and then selectively melting it with the laser, a component is built up layer by layer. The local temperature distributions that occur during this process determine not only the properties of the finished component, but also the possible formation of defects such as pores and cracks. Due to the high relevance of the thermal history for precise geometries and defect formation, a temporally and spatially resolved measurement of quantitative (or real/actual) temperatures would be optimal. Quantitative values would ensure comparability and repeatability of the AM process which would also positively affect the quality and safety of the manufactured component. Furthermore, it would also contribute to the validation of simulations and to a deeper understanding of the manufacturing process itself.
At present, however, only qualitative monitoring of the thermal radiation is performed (e.g., by monitoring the melt pool using a photodiode), and safety-relevant components must be inspected ex situ afterwards which is time-consuming and costly. A reason for the lack of quantitative temperature data from the process are the challenging conditions of the PBF-LB/M process with high scanning speeds and a small laser spot diameter. Furthermore, the emissivity of the surface changes at high dynamics (temporally/spatially) as well as with temperature and wavelength. This specifically makes contactless temperature determination based on emitted infrared radiation challenging for PBF-LB/M. Although classical thermography offers very good qualitative insights, it is not sufficient for a reliable quantitative temperature determination without a complex temperature calibration including image segmentation and assignment of previously determined emissivities.
For this reason, this publication presents the hyperspectral thermography approach for the PBF-LB/M process: The emitted infrared radiation is measured simultaneously at many adjacent wavelengths. In this study, this is realized via a fast hyperspectral line camera that operates in the short-wave infrared range. The thermal radiation of a line on the target is spectrally dispersed and detected to measure the radiant exitance along that line. If the melt pool of the PBF-LB/M process moves through this line at a sufficient frame rate, a spatial reconstruction of an effective melt pool is possible.
One approach to determine the desired emissivities and the quantitative temperature from this hyperspectral data are temperature-emissivity separation (TES) methods. A major problem is that n spectral measurements are available, but n+1 parameters are required for each image pixel (n emissivity values + one temperature value). TES methods offer the possibility to approximate this mathematically underconstrained problem in a reliable and traceable way by analytically parameterizing the spectral emissivity with a few degrees of freedom. Using this approach, setup and method are applied to a research machine for PBF-LB/M, called SAMMIE (Sensor-based Additive Manufacturing Machine). First results under AM process conditions are shown which form the basis for the determination of quantitative temperatures in the PBFLB/M process. This marks an important contribution to improving the comparability and repeatability of production, validating simulations, and understanding the process itself. When fully developed and validated, the presented method can also provide reference measurements to evaluate and optimize other, more practical monitoring methods, such as melt pool monitoring or optical tomography. In the long run, this will help to increase confidence in the safety of AM products.
This work provides an investigation of the influence of low temperature heat treatments on the fatigue behavior of a PBF-LB AlSi10Mg alloy. Fatigue specimens are produced in form of round bars on a build platform preheated at 200 ◦C. The specimens have been tested in three different conditions: as-built, and after heat treatments at 265 ◦C for 1 h and 300 ◦C for 2 h. Prior to the fatigue testing, the defect distribution is analyzed by means of micro computed tomography. Subsequently, the peak over threshold method is successfully applied to provide a prediction of the size of killer defect. The defect population was of gas porosity type. No clear improvement of the fatigue performance is observed after the heat treatments. The fatigue strength predicted using fracture mechanics-based approaches is in good agreement with the experimental data. Among the studied approaches, short crack models provided the most conservative predictions.
This article presents a method to use the dispersive behavior of ultrasonic guided waves and neural networks to determine the isotropic elastic constants of plate-like structures through dispersion images. Therefore, two different architectures are compared: one using convolutions and transfer learning based on the EfficientNetB7 and a Vision Transformer-like approach. To accomplish this, simulated and measured dispersion images are generated, where the first is applied to design, train, and validate and the second to test the neural networks. During the training of the neural networks, distinct data augmentation layers are employed to introduce artifacts appearing in measurement data into the simulated data. The neural networks can extrapolate from simulated to measured data using these layers. The trained neural networks are assessed using dispersion images from seven known material samples. Multiple variations of the measured dispersion images are tested to guarantee the prediction stability. The study demonstrates that neural networks can learn to predict the isotropic elastic constants from measured dispersion images using only simulated dispersion images for training and validation without needing an initial guess or manual feature extraction, independent of the measurement setup. Furthermore, the suitability of the different architectures for generating information from dispersion images in general is discussed.