FG Hybride Fertigung
Refine
Document Type
- Doctoral thesis (14)
- Scientific article (4)
- Book (1)
- Habilitation thesis (1)
Has Fulltext
- yes (20)
Is part of the Bibliography
- no (20)
Year of publication
Keywords
- Additive manufacturing (4)
- Selektives Laserschmelzen (4)
- Additive Fertigung (3)
- Blechumformen (3)
- L-PBF (3)
- Creep (2)
- Deformation (2)
- Hybrid manufacturing (2)
- Kriechen (2)
- Leichtbau (2)
Institute
Das Laser Powder Bed Fusion-Verfahren (L-PBF) ist ein etabliertes additives Fertigungsverfahren zur Herstellung geometrisch komplexer Strukturbauteile mit hohem Leichtbaupotenzial. Einer breiteren industriellen Anwendung stehen jedoch die vergleichsweise geringe Produktivität des Prozesses und die damit verbundenen hohen Bauteilkosten entgegen.
Vor diesem Hintergrund untersucht die vorliegende Arbeit Beschleunigungsansätze entlang der Fertigungskette von L-PBF-Bauteilen, von der Fertigung über die thermische Nachbehandlung bis hin zur Entfernung der Stützstrukturen. Zunächst wurde die Anpassung der L-PBF-Prozessparameter mithilfe einer an den Anwendungsfall angepassten Optimierungsmethode untersucht. Ziel war es, eine definierte Zielporosität von 0,1 % zu erreichen bzw. zu unterschreiten. Auf Basis der experimentell ermittelten Ergebnisse konnten für Schichtstärken von 30, 60 und 120 µm energiebezogene Parameterräume identifiziert werden, in denen Bauteile mit geringer Porosität mit erhöhter Wahrscheinlichkeit gefertigt werden können. Gegenüber dem Referenzparametersatz ließ sich die Aufbaurate dabei um den Faktor 3,2 steigern.
Zur weiteren Beschleunigung wurde eine L-PBF/HIP-Kombinationsstrategie untersucht, bei der Bauteile zunächst mit nochmals erhöhter Aufbaurate und damit erhöhter Porosität gefertigt und anschließend durch heißisostatisches Pressen (HIP) in den Zielbereich verdichtet werden. Die Bewertung der Produktivität in Abhängigkeit vom Bauteilvolumen zeigte, dass diese kombinierte Prozessroute ab einem Volumen von etwa 1421 cm³ der konventionellen Prozessführung überlegen ist.
Ergänzend wurde das elektrochemische Verfahren des Hirtisierens® zur Entfernung von Stützstrukturen sowie hinsichtlich seines Einflusses auf die Oberflächenqualität bewertet. Zudem wurde eine hierfür geeignete Stützstrukturgeometrie entwickelt. Die Qualität der beschleunigten Prozessrouten wurde durch quasistatische und zyklische Prüfungen nachgewiesen, wobei die Anforderungen gemäß der ASTM F2924 durchgängig erfüllt wurden.
Zur weiteren Reduktion des Stützstrukturbedarfs wurde eine numerische Methode zur Bestimmung des kritischen Überhangwinkels entwickelt. Dabei wurde ein Wert von 25° ermittelt und experimentell bestätigt. Die Validierung der entwickelten Ansätze erfolgte abschließend an einem Cross-Over Duct, an dem sowohl die Übertragbarkeit der beschleunigten Prozessparameter auf eine komplexe Geometrie als auch das numerische Modell nachgewiesen und das Beschleunigungspotenzial in unterschiedlichen Aufbauorientierungen gegenüber dem Referenzparametersatz aufgezeigt wurden.
With a combination of desirable properties such as low density, high specific yield strength, low material cost,
and excellent oxidation and corrosion resistance, iron aluminide (Fe-Al) has shown considerable potential to be
an alternative to high-alloy chromium steels, and in some cases even nickel-based superalloys, in hightemperature
applications. Due to these features, it is especially suitable for the aerospace and automotive industries.
Recent advancements indicate an increasing interest in Fe-Al within the additive manufacturing industry,
particularly in directed energy deposition (DED) processes. Despite this progress, processing of Fe-Al
materials using the laser directed energy deposition (L-DED) has not been sufficiently investigated. In this study,
Fe-Al powder material was produced from a commercial Al rod encased in a commercial low alloy-steel tube by a
plasma-based ultrasonic atomization eliminating the need to cast an alloy ingot in advance. Subsequently, the
produced powder was used in a L-DED process to fabricate an additively manufactured sample. The sample was
investigated in terms of mechanical property, microstructure, chemical composition, and phase structure by
scanning electron microscope (SEM) / energy dispersive X-ray spectroscopy (EDX), X-ray diffraction (XRD),
electron backscatter diffraction (EBSD) and microhardness analyses.
The progression of the automotive industry has introduced a slew of complex designs to meet the personalized style expressions of consumers. However, these intricate designs in premium automobiles present significant manufacturing challenges, especially in detecting and prioritizing the varying severities of cosmetic surface defects. Traditionally, defect identification is performed at the end of the development phase, which is a costly and time-consuming process. This thesis seeks to enhance early defect detection efficiency in the concept phase using artificial intelligence (AI). Through examination of unpainted parts, auditors’ perceptual practices, and deep draw simulations, the research aims to establish a virtual prediction method to accurately classify surface defects. The findings of this thesis prove that the geometry of unpainted parts sufficiently predicts potential defects post-painting, refuting the need for physical painted part analysis. The identification process adopted by auditors, grounded in analyzing light distortions on surfaces, is replicated within a neural network. Furthermore, by incorporating physiological optical aspects, the thesis improves the virtual representation of defects, creating a visible and analyzable dataset for AI algorithms. The accuracy of the neural networks trained on such datasets is substantiated by the successful translation of auditors’ classification methods into a machine learning environment. In particular, a machine learning approach using Random Forest (RF) algorithms excelled in prefiltering areas of interest based on curvature and strain values. A convolutional recurrent neural network (CRNN) is developed to classify the severity of defects, with the introduction of an annotation application to label data by experts. The CRNN demonstrated a remarkable accuracy of 86 % in classifying defect severity based on simulation data. Moreover, the study assessed the frame-by-frame localization of defects, where the PatchCore anomaly detection algorithm proves optimal, achieving an F1 Score of 0.86. Overall, the research successfully shows the transferability of expert auditors’ perception into a neural network architecture, highlighting the feasibility of implementing AI in early phases of automotive design to predict and classify defects, which has the potential to significantly reduce development costs and improve manufacturing efficiency.
The demands of high-performance industries such as aerospace, automotive, tool manufacturing, oil, and gas industries are driving the innovation in high-performance materials and their production methods. This study explores the impact of hybrid manufacturing, specifically the effect of the addition of tungsten carbide (WC/W2C) via Laser-Directed Energy Deposition (L-DED), on the hot workability, hardness, and microstructure of nickel-based superalloy Inconel 625 (IN625). IN625 is known for its high temperature and high corrosion resistance, and tungsten carbide for its high wear resistance and grain refinement effect. The integration of WC/W2C particles into the IN625 matrix, in addition to the use of the hybrid approach of additive manufacturing followed by a hot–forming process, significantly influences the microstructure and mechanical behavior of the material. Thus, while incorporation of the WC/W2C can strengthen the material and extend the mechanical limitations, its full impact, including any potential usages, should be thoroughly evaluated for the intended application of the materials. To understand the effect of WC/W2C, additive manufacturing of IN625 both with and without WC/W2C and isothermal hot compression was carried out. The objective is to analyze the differences in microstructure and properties between L-DED manufactured IN625, and WC-reinforced IN625, and their hot-forming behavior, focusing on the effects of WC addition and post-deformation on microstructure and mechanical properties. This work represents the first investigation into the effect of WC/W2C hard particles on the hot-forming process of additively manufactured Ni-based metal matrix composites.
Within the scope of these investigations, the feasibility of a material bond between Ti-6Al-4V and the magnesium alloy AZ91 is analyzed. Ti-6Al-4V is frequently used for implants due to its biocompatibility, corrosion resistance, and specific strength. However, depending on the surface quality, the attachment behavior of the bone to the implant varies. Magnesium implants promote the regeneration of bone tissue and biodegrade as the bone tissue heals. Combining the properties of both materials in one implant enables a reduced implant volume and increased stability. For this reason, this study aims to demonstrate the feasibility of creating a material bond between the materials Ti-6Al-4V and AZ91. For this purpose, Ti-6Al-4V truncated cones and AZ91 sleeves were produced using the additive manufacturing process of laser powder bed fusion (L-PBF). The as-built sleeves were then pressed onto machined truncated cones. Since zinc serves as a lubricant and has good diffusion properties with the materials used as a result of heat treatment, a comparison was made between zinc-coated and the as-built Ti-6Al-4V samples. This showed that a bond was created after hot isostatic pressing and that the push-out force could be increased by more than 4.5 times. Consequently, a proof of feasibility was demonstrated, and a high potential for applications in medical technology was shown.
Flanges are essential elements in sheet metal parts, where they perform important functions such as increasing structural stiffness, serve as bearing seats and positioning aids. Currently, flanges are mainly formed by the conventional hole-flanging process, which uses dedicated dies and punches. The limitations of conventional hole-flanging include the high cost involved in die design and manufacture, poor accessibility to form flanges in some complicated dies, and low formability. High-speed tool rotation is introduced in hole-flanging processes to increase the process temperature and forming limit of hard-to-form materials. This is done using paddle-shaped tools and spherical punches in existing production lines. In addition, strategies are explored to enhance the geometrical accuracy in robot-based hole-flanging by single-point incremental forming (SPIF).
This thesis investigates new hole-flanging process variants by using experiments and finite element (FE) analysis. Empirical material models are used to study the effects of the process parameters on flange shape, forming limits, process temperature, and forming mechanics. The process formability is highest in hole-flanging with punch rotation and reduces for paddle forming and hole-flanging by SPIF. The least formability is observed in conventional hole-flanging. High-speed tool rotation and tool feeds form flanges with high formability. Surface strain measurements of the flanges show that low-speed tool rotation combined with high tool feeds favor crack formation. Tensile tests revealed that the material's formability increases with temperature and strain rate. From FE analyses of the process variants, the deformation in hole-flanging was determined to be mainly because of membrane stretching and bending. Some shear was present in paddle forming and SPIF. Based on FE analyses and experiments, the temperature and strain rate are determined to be the main parameters that account for the difference in process formability.
Hole-flanging by SPIF is conducted on a robot manipulator to enhance geometrical accuracy. New tool paths are explored to prevent flange conicity. A parametric study is conducted to determine the process's feasibility. The stiffness of a robot is compensated to achieve the target flange geometry.
Additive technologies are now widely used for the production of complex precise parts and have high potential for the production of forming dies. In this work, hot-forming dies optimized for additive manufacturing were developed and produced with wire arc direct energy deposition (WA-DED) and laser powder bed fusion (L-PBF) technologies. The concept of lightweight hot-forming dies with a 2D-lattice structure was developed, which reduced the weight of each die by 56%, from 14.2 kg to 6.1 kg, in production using L-PBF. Maraging/precipitation-hardened steel 17-4PH was used as an alternative to traditional hot-working steels with slightly lower mechanical properties and a much higher processability in the additive manufacturing process. The workability of the manufactured dies was confirmed by forging tests on an industrial screw press.
Additive Manufacturing of metals has become relevant for industrial applications. The near net-shape production of components produced by Laser Powder Bed Fusion (PBF-LB/M) enables new possibilities in component design combined with a reduction of the amount of needed material. Omitting the extra material, that was part of conventionally produced components due to machining constraints, results in components which in consequence lack the inherent additional safety margins provided by the higher material consumption of conventionally produced components. Therefore, to use PBF-LB/M metals in safety critical applications an in-depth understanding of porosity and internal stresses in parts made by PBF-LB/M is needed. Only non-destructive testing methods—such as computed tomography and residual stress analysis using neutrons—enable the assessment of porosity and stresses in the whole part. In this thesis I tackled creep and tensile static deformed specimens to fill research gaps in this field in terms of analysing PBF-LB/M stainless steel AISI 316L: from structural properties to in-situ behaviour. The initial void population of AISI 316L is studied using X-ray and synchrotron micro computed tomography. Specimens produced with different process parameters were analysed to quantify the influence of process parameters on the initial void population. The possibility to close voids using the laser illumination of subsequent layers is discussed by a quantitative study of the ability of the laser to melt different multiples of the applied layer thickness. The formation of internal stress is inherent to components produced by the PBFLB/ M process. These stresses remain in the components after production as residual stresses. In this thesis a study is presented which aims to unravel the mechanisms that define the spatial distribution of the residual stresses, and their magnitude. In the end, the population of internal voids during mechanical testing is studied by X-ray micro computed tomography. The evolution of damage accumulation in creep specimens is studied at different stages of the creep test. Results are compared to a creep tested conventionally made specimen and to a PBF-LB/M specimen from a tensile test. An interconnection between the PBF-LB/M microstructure and the pattern of damage is revealed.
Effiziente Fertigungssysteme sind wichtiger denn je, um die steigenden Klimaauflagen und die damit verbundenen Anforderungen an die Ressourceneffizienz erfüllen zu können. Klassische Fertigungsverfahren haben dabei den Nachteil, dass diese entweder mit einem hohen Ressourcenverbrauch verbunden sind oder Produktvarianten nicht wirtschaftlich herstellen können. Zwar können additive Fertigungsverfahren wie Selective Laser Melting (SLM), Wire Arc Additive Manufacturing (WAAM) oder Laserpulverauftragschweißen (LPA) durch ihre hohe Flexibilität dieses Problem teilweise beheben, aber bei großvolumigen Bauteilen oder größeren Stückzahlen sind diese Verfahren mit langen Prozesszeiten verbunden und weisen damit eine geringe Produktivität auf.
Dahingegen können hybride Fertigungsverfahren bestehend aus einer Kombination aus einem werkzeuggebundenen Prozess (z. B. Schmieden) und einem werkzeuglosen Verfahren (z. B. Auftragschweißen) eine effiziente Lösung darstellen. Die Herausforderungen sind dabei, die Wechselwirkungen der Fertigungsverfahren zu untersuchen und die Teilprozesse aufeinander abzustimmen. Ziel war es, die negativen Auswirkungen einer Wärmeeinbringung und deren Effekte auf die Mikrostruktur sowie die daraus resultierende Härte und Festigkeit aufzuzeigen. Des Weiteren wurde die Prozessreihenfolge untersucht, mit der eine optimale Festigkeit erreicht werden kann.
Anhand von Beispielen wurden anschließend Richtlinien und Strategien für das Hybridverfahren abgeleitet. So konnte gezeigt werden, dass durch das Hybridverfahren der Werkstoffeinsatz mit bis zu 53,2 % gegenüber der reinen Zerspanung gesenkt werden kann. Die Prozesszeit liegt mit 4,4 bzw. 2,4 h deutlich unter der eines rein additiven Prozesses, was auf eine konkurrenzfähige Wirtschaftlichkeit hindeutet.
Among variants of AM technology, wire-arc additive manufacturing (WAAM) process suitably produces bulky metal parts with a medium complexity. Although WAAM shows a great potential, this process has not been fully explored, and it is therefore a worthwhile subject to further investigate. This monograph engages with tool path planning for the WAAM process and existing problems in WAAM are subsequently addressed, as following:
The existing overlapping models yield an uneven surface due to the inner beads overlapping with two neighbor beads while the outers have only a one-sided overlap. New mathematical models are established to obtain optimal distances between adjacent weld beads and to make surfaces more even.
Lightweight structures have high strength as compared to their weight. Producing these structures by WAAM results in imperfections because of uneven weld beads - an inevitable phenomenon when starting a new track. A new tool path strategy to produce lightweight structures is developed using contour patterns, which transforms any arbitrary geometry into that of a continuous path. Unfortunately, voids are formed at junctions when using contour patterns. An adaptive correction using machine learning is then developed to overcome this defect.
Large overhangs and inclined features are out of reach of WAAM when operating in the x-y plane in Cartesian system. This limits the ability of the WAAM process. To overcome that, each overhang feature should be welded with a distinct direction. This approach is realized within the robot based WAAM process. Establishing the kinematics of the FANUC robot arm and the positioner helps compute the new coordinates, the angular displacements of the positioner, and the orientation of the torch. The proposed strategy is examined through the construction of an overhang part.
The capabilities of the multi-axis deposition are extended by constructing curved thin-walled structures. Irregular thickness layers are deposited by controlling the travel speed, resulting in a different deposition rate. The curved thin-walled structures can be properly fabricated by combining a multi-axis deposition with non-uniform thickness layers. Bonding quality as well as geometric accuracy are also to be investigated. To ensure geometric accuracy, a compensation strategy is developed. The performance of the proposed algorithm is validated on different geometries.
This monograph will conclude with a summary of this work’s main achievements and contributions as well as outlooks on future research.
