Search

Metal additive manufacturing is a recent breakthrough technology that promises automated production of complex geometric shapes at low operating costs. However, its potential is not yet fully exploited due to the low reproducibility of quality in mass production. The monitoring of parts quality directly during manufacturing promises to solve this problem, while machine learning showed efficient performance correlating versatile manufacturing measurements with different quality grades. Today, most monitoring algorithms are based on semi- or supervised learning, thus, requiring a collection and ground-truth validation of training sets. This is costly and time consuming in real-life conditions. Our work is a feasibility study of the application of unsupervised machine learning to monitor different manufacturing regimes and quality in metal additive manufacturing. The algorithm combines the kernel Bayes rule for inference and Bayesian adaptive resonance for structuring the incoming data. Airborne acoustic emission from laser powder bed fusion is used as an algorithm input. The recognition of the main manufacturing regimes (conduction mode, stable, and unstable keyholes) are shown on real-life data, while the self-learning accuracy of developed algorithm exceeds 88 %. Our approach promises future development of plug-and-play quality monitoring systems for laser processing technology, requiring minimum modifications of the existing machines, reducing time/cost for algorithm preparation and providing continuous data driven adaptation of the algorithm to changes in manufacturing conditions.

Laser powder bed fusion (L-PBF) has emerged as an additive manufacturing technique that offers unprecedented design freedom. Besides being capable of producing complex and near net shape objects, L-PBF can impact tremendously the engineering materials community due to the possibility of locally manipulating metallic microstructures. Here we exploit the latter potentiality of L-PBF, to produce site-specifically tailored stainless steel components, in terms of their crystallographic texture. The tailored materials are tested and exhibit superior energy dissipation capabilities under bending deformation compared to uniformly textured materials. This is enabled by the strong dependence of the secondary hardening mechanisms, namely the deformation twinning and/or martensite formation, of these materials on the locally tuned microstructures. With the aid of finite element simulations, it is possible to identify the stress state and hence, the crystallographic orientations that facilitate twinning or martensite formation. Then, by engineering favorable crystallographic textures, matched to the complex stress state during bending, enhanced work hardening behavior is obtained. This site-specific microstructure design enabled by L-PBF provides a new pathway for the design of “smart” components that exhibit superior mechanical response under complex stress states.

Tensile and creep rupture properties of crack-free CM247LC alloy, processed via laser powder bed fusion, have been characterised in this work at temperatures up to 1000°C. The subject alloy matches or even exceeds the tensile performance of its directionally solidified counterpart up to 700°C, beyond which both the strength and ductility drop off relatively rapidly with increasing temperature. It was found that the agglomeration of discrete carbides – that essentially pin the grain boundaries – facilitate the nucleation, linking and propagation of microcracks along the adjacent grain boundaries at elevated temperatures. Relatively short-term creep-rupture tests conducted at 800°C show the need for greatly improving the microstructure of this alloy to compete with its directionally solidified or conventionally cast counterparts, especially considering that this material finds extensive application up to ∼950°C in turbomachinery components. Increasing the grain size, incorporating grain boundary engineering concepts and a better understanding of carbide/γ' precipitate evolutionary characteristics can greatly help to improve the material's suboptimal creep response.

Fe-Mn-Si shape memory alloys are materials, whose functional properties strongly depend on microstructural factors such as grain size and orientation, phase fraction and chemical composition. The present study demonstrates the possibility of microstructure modification via process parameter variation on a Fe-Mn-Si based shape memory alloy fabricated by laser powder bed fusion. By varying the scan speed, samples characterized by coarse elongated grains with strong orientation along the build direction or by finer equiaxed grains without preferential orientation can be fabricated. Changes in the volume phase fraction of bcc-δ ferrite and fcc-γ austenite are also introduced by selective Mn evaporation. A direct correlation between the generated microstructure and achieved mechanical and shape memory properties is found.

Non-destructive characterization methods to observe phase transformations and thus, gain insights to transformation mechanisms in representative volumes are key for the development of advanced materials and manufacturing. Conventional methods are constrained to the surface and small sizes, thus, access to the bulk often implies tedious destructive approaches and hinders in-situ observations of phase evolution. In this work, we introduce a non-destructive technique that overcomes key limitations prevailing today for mapping the spatial distribution of magnetic phases in bulk materials. The use of polarized neutrons, being sensitive to sub-percent fractions of ferromagnetic phases and able to penetrate centimeter sized samples, enables micrometer-scale spatial and second-scale time resolutions. We demonstrate ex-situ and in-situ quantitative mapping of magnetic phases, in particular the evolution of martensite induced by uniaxial- and biaxial deformation in metastable 304 steel. The quantitative results obtained during in-situ deformation testing prove polarization contrast neutron imaging to be particularly effective in detecting small fractions of martensite. Especially during early stages of deformation where neutron diffraction, which in contrast does not provide full field spatial resolution, fails. The short exposure times and high sensitivity renders the method well suited for rapid 3D tomographic mapping and/or operando investigations of phase distributions.

Laser-powder bed fusion (L-PBF) has gained significant research interest, not only for its profound advantage of producing near-net shape complex geometries of metallic parts, but also for the possibility of producing tailored microstructures. Here we exploit the capability of manipulating the crystallographic texture by L-PBF to tailor the deformation behavior of austenitic stainless steels. In specific, by adjusting the laser power and the laser scanning speed, tailored crystallographic textures can be obtained, along the uniaxial loading direction in 304L stainless steel samples produced by L-PBF. In situ neutron diffraction and uniaxial tension and compression tests are undertaken to investigate the extent of the transformation induced plasticity effect and to correlate it with the tailored macrostructures. The influence of the initial and the evolving crystallographic texture on the deformation behavior is demonstrated and elaborated accordingly. The observed asymmetry in the deformation behavior between tension and compression is also discussed in detail.

Neutron dark-field imaging (DFI) was used to investigate the microstructure of additive manufactured steels. Several DFI methods were combined to assess the microstructure over more than two orders of magnitude in size. Different degrees of porosity and other building features were found depending on the parameters of the selective laser melting additive manufacturing process. A sample built with processing parameters yielding the lowest porosity was deformed which induced a phase transformation of the austenitic phase (fcc) into the martensitic phase (bcc). In the deformed sample an increased dark-field contrast was observed which can only be explained by accounting for the fcc-bcc phase distribution and the magnetic properties of the martensitic phase. We demonstrate that neutron dark-field imaging is well suited to not only detect build flaws like cracks but quantitatively characterize the microstructure in additive manufactured steels.