Experimental Melting Research Articles

Glass-ceramic binder of cordierite composition for low-temperature sintering of high-strength ceramic materials based on oxygen-free silicon compounds

The mechanical properties of composite ceramic materials obtained based on oxygen-free silicon compounds are largely determined by the properties of the glass binder. This paper presents the results of studies aimed at determining the most fusible glass in the pseudo-binary system 2MgO2Al2O35SiO2–2MnO2Al2O35SiO2 with a high tendency to crystallize as a glass-ceramic binder for low-temperature sintering of high-strength ceramic materials based on oxygen-free silicon compounds. The crystallization tendency of the experimental melts decreases with an increase in in the content of manganese cordierite, as confirmed by X-ray and infrared spectroscopic studies. Based on experimental studies, a melting diagram was constructed, which was used to determine the ratio between magnesium cordierite and manganese cordierite (50:50 wt.%), ensuring a minimum melt temperature of 12750C. The melting point of the glass of the specified composition is 14500C. The synthesized glass is characterized by a softening point of 8000C and crystallizes intensively at 10300C. The The thermal coefficient of linear expansion of the crystallized glass samples is 20.810–7 0C–1. X-ray diffraction and electron microscopic studies have shown that the developed glass is almost completely crystallized during heat treatment for 2 hours, forming a cordierite solid solution 2(Mg,Mn)O2Al2O35SiO2. The size of the cordierite phase crystals ranges from 0.5 to 3.0 m. Due to its fusibility and high crystallization tendency, the developed glass, can be proposed as a promising glass-ceramic binder for the production of high-strength ceramic materials (wear and impact resistant) based on SiC and Si3N4 with reduced sintering temperatures.

Improving melt pool depth estimation in laser powder bed fusion with metallic alloys using the thermal dose concept

AbstractLaser-matter interactions in laser powder bed fusion for metals (LPBF-Ms) significantly impact the final properties of the fabricated components. Critical process parameters, such as the linear energy density (LED), the ratio of laser power to scan speed, modify the energy input and consequently modify the melt pool geometry. LED strongly influences the melt pool cross-sectional profile, which dictates the thermal effects, microstructure, and mechanical properties of the finished part. Recognizing the crucial role of the melt pool in additive manufacturing, researchers have developed predictive models to estimate its dimensions and morphology. These models aid in tailoring part properties, optimizing process parameters, and reducing the number of experimental trials. However, existing models are either computationally expensive or analytically overly simplified for general LPBF-M applications. This study proposes an improved model that incorporates the Rosenthal equation as described by Tang to increase the accuracy of melt pool depth prediction. By using the thermal gradient per unit time, termed the “thermal dose” in this paper, corresponding to the LED value that produces experimental near-semicircular melt pool shapes for each studied material, we can improve the melt pool depth estimation. The trend revealed a good fit across the LED range compared with experimental measurements, suggesting the model’s effectiveness.

Deep learning for melt pool depth contour prediction from surface thermal images via vision transformers

Anomalous melt pools during metal additive manufacturing (AM) can lead to deteriorated mechanical and fatigue performance. In-situ monitoring of the melt pool subsurface morphology requires specialized equipment that may not be readily accessible or scalable. Therefore, we introduce a machine learning framework to correlate in-situ two-color thermal images observed via high-speed color imaging to the two-dimensional profile of the melt pool cross-section. We employ a hybrid CNN-Transformer architecture to establish a correlation between single bead off-axis thermal image sequences and melt pool cross-section contours measured via optical microscopy. Specifically, a ResNet model embeds the spatial information contained within the thermal images to a latent vector, while a Transformer model correlates the sequence of embedded vectors to extract temporal information. The performance of this model is evaluated through dimensional and geometric comparisons to the corresponding experimental no-powder melt pool observations. Our framework is able to model the curvature of the subsurface melt pool structure, with improved performance in high energy density regimes compared to analytical models. Additionally, the use of ratiometric temperature estimates improves the accuracy of the model predictions compared to monochromatic imaging. This work establishes a framework extensible towards powder-based AM builds.

Central Tibetan adakitic rocks archive the critical impact of water on the Nb/Ta variation in deep crustal melts

Nb and Ta are geochemical twins, yet their ratio varies considerably in igneous rocks. It is generally supposed that felsic melts coexisting with residual rutile will have elevated Nb/Ta ratios relative to their basalt precursors, which is pivotal to our understanding of deep crustal melting and continent formation throughout Earth’s history. This study reports petrological, bulk-rock and mineral geochemical, and Sr-Nd-Mo-Hf isotopic data on Eocene adakitic rocks in central Tibet. These rocks can be petrologically and geochemically divided into an amphibole-bearing high-Nb/Ta and an orthopyroxene-bearing low-Nb/Ta group, which were formed by partially molten eclogitized lower crust based on the geochemistry of their early-crystallized minerals. The two groups have similar Sr–Nd–Hf isotope compositions and abundances of strong incompatible elements, suggesting that their source protoliths share identical compositions. Besides, the values of TiO2, Ba, Zr/Sm, and La/Yb are comparable between the two groups of rocks, eliminating the possibility of variations in Nb/Ta-sensitive minerals (e.g., rutile, amphibole, biotite, and titanite) in the source as the cause of their Nb/Ta variation. The contrasting mineral assemblages of the two group rocks, along with the higher Ce/Mo ratios and lower δ98/95Mo values in the low-Nb/Ta group, indicate the source of the low-Nb/Ta contains less water than that of the high-Nb/Ta group. These findings, combined with experimental results and trace element partial melting modeling, lead us to propose that the Drutile Nb/Ta responsible formation of the low-Nb/Ta group is higher than that of the high-Nb/Ta group, due to the decreased water content in the source. Thus, the central Tibetan adakitic rocks provide the first natural example demonstrating that the Nb/Ta ratios in felsic rocks derived from rutile-bearing eclogite depend on the physical conditions of source melting, which are not always higher than the protolith. Our study also highlights the potential of employing Mo isotopes in adakitic rocks to test this hypothesis for those that are not petrologically distinguishable.

Experimental constraints on iron and sulfur redox equilibria and kinetics in basaltic melt inclusions

The oxidation states of dissolved iron (Fe) and sulfur (S) in silicate melts were studied experimentally by heating Hawaiian olivines containing S-bearing melt inclusions (MIs) in a 1 atm gas-mixing furnace. The MIs are closed with respect to S and C, but equilibrate with the oxygen fugacity (ƒO2) and water fugacity (ƒH2O) of the furnace. We also conducted experiments using a S-free synthetic Hawaiian MI composition held on Pt-loops. Experiments were run at 1225 °C for 8–96 hr in H2-CO2 gas mixtures corresponding to ƒO2 values that varied over seven orders of magnitude and then drop-quenched into water. Time-series experiments show that MI Fe3+/Fe2+ and S6+/S2− ratios reached constant values within ∼24 hr. The H2O contents of the Pt-loop glasses and the dehydrated MIs overlap and are proportional to the fH2O in the furnace gas. The slope of log10(Fe3+/Fe2+) vs. log10ƒO2 is 0.25±0.02(2σ) for the Pt-loop glasses and 0.23±0.01(2σ) for the quenched MIs; the slopes of the two sets of experiments are consistent with equilibrium having been reached via the reaction FeO + ¼ O2 = FeO1.5. S-bearing quenched MI glasses have Fe3+/Fe2+ ratios that are systematically low compared to ratios measured in the S-free Pt-loop glasses run at the same T and ƒO2; this may reflect an effect of S on Fe3+/Fe2+, although further experiments are needed to explore this possibility. S6+/S2− ratios in the experimental melt inclusions vary systematically as a function of ƒO2, and the measured ratios agree with independent calculations of S oxidation state in basaltic melts.After being held at 1225 °C for 24 or 48 hr, a subset of olivine-hosted MIs were cooled at rates of 1700–40,000 °C/hr and quenched at temperatures between 900 °C and 1150 °C to explore whether Fe and S speciation are modified during cooling. Fe and S speciation in isothermal and cooled MIs held initially at the same ƒO2 overlap within uncertainty, suggesting that the T dependence of the homogeneous reaction 8Fe3+ + S2− = 8Fe2+ + S6+ is small and/or its kinetics are sluggish relative to cooling rates required to quench basaltic liquids to glasses. The centers of unheated, naturally quenched MIs from Papakōlea, Hawaii with syneruptive cooling rates ranging from ∼50 to 12,000 °C/hr were also measured for Fe and S speciation. MIs that experienced the lowest cooling rates have higher Fe3+/Fe2+ and S6+/S2− ratios relative to rapidly cooled inclusions of comparable size. The increase in Fe3+/Fe2+ is due to olivine crystallization on the inclusion walls during cooling and diffusion of the resulting oxidized boundary layer into MI interiors. Diffusive modification is minimized in MIs that have been rapidly cooled and/or are sufficiently large such that the oxidized boundary layer is restricted to a narrow region adjacent to the inclusion walls. Our experiments and their comparison to naturally quenched MIs show that measurements of Fe3+/Fe2+ and S6+/S2− ratios in quenched basaltic glasses and in the centers of rapidly cooled and/or large MIs are likely to be representative of the redox state in the melt at magmatic temperatures prior to quenching to glass.

Intraplate alkali basalts related to end-Cretaceous Deccan magmatism: implications for tectonomagmatic processes

This study aims to investigate the earliest imprint of Deccan rift magmatism as preserved in alkali basalts from the northwestern Indian shield. The alkali basalts are petrographically classified as nephelinites and basanites. They are silica undersaturated and their high Mg numbers, CaO/Al 2 O 3 , Cr and Ni indicate their primitive character. Geochemically, they are similar to global ocean island basalts; their bulk rock trace elements and Sr–Nd–Hf isotopic signatures suggest their derivation from a garnet-bearing peridotite field. However, their elevated values of Sr/Sm, Sm/Hf, Zr/Hf, Nb/Ta; positive Ba, Sr and negative Zr, Hf spikes suggest that the magma source represents a mixture of garnet peridotites and carbonated melts. Estimated primary melt compositions closely follow the trajectory defined by the high-pressure experimental partial melting trend of a low-carbonated peridotite source. The melting environment approximates to a high mantle potential temperature. A low 87 Sr/ 86 Sr ratio and a negative correlation between 176 Hf/ 177 Hf and 143 Nd/ 144 Nd of the alkali basalts suggest that the mantle source is a mixture of a depleted Indian MORB-type mantle and an enriched mantle type 2. We correlate this event with the melting of the leading edge of the Réunion plume head during Gondwana break-up in a relatively short span of the Cretaceous/Paleogene boundary.

Ti-6Al-4V alloy printing — correlations between experimental and numerical modelling melt pool data

Ti-6Al-4V alloy printing — correlations between experimental and numerical modelling melt pool data

A theoretical equation of state to formulate the melting curve of metals with varying pressure

The objective of this study has been formulate mathematical equations for the melting temperature of metallic solids using various thermodynamics equation of state, specifically those proposed by Modified generalized Lennard – Jones(mGL-J) equation of state (EOS) and Usual-Tait EOS. The derived equations of state establish the relation between pressure, volume, melting temperature, bulk modulus and first-order pressure derivative at zero pressure. These equations of state are used to calculate melting curves for the metallic elements Pd, Cd, Pt, Mn, Au, Al, Zn and Ag up to 30 GPa. The obtained results from Modified generalized Lennard – Jones(mGL-J) equation of state display a significant rise in the melting curve because of a large increase in bulk modulus. The results of mGL-J EOS is the most consistent with the experimental data. The outcomes obtained from Usual-Tait equation of state diverges to a great extent with the experimental melting temperature. The study contributes to a better understanding of the quantitative effect of pressure on the melting temperature of various solids. The derived equations of state are useful to calculate melting temperature at larger pressures.

Geometrically-informed predictive modeling of melt pool depth in laser powder bed fusion using deep MLP-CNN and metadata integration

The laser powder bed fusion (LPBF) process has the ability to manufacture intricate geometries. However, the fabrication of complex geometries using the LPBF process for industrial applications remains challenging in terms of achieving consistent microstructure and mechanical properties. Experimental investigations show that in addition to process parameters, geometrical factors such as shape and size of parts impact the thermal history, microstructure, defect structure, and thus mechanical and fatigue properties. As a result, the standardization, qualification, and certification of additively manufactured (AM-ed) parts encounters substantial obstacles posed by the obscure impact of design parameters on component quality. In this paper, an artificial intelligence framework is proposed to predict the melt pool depth of the additively manufactured parts considering the geometrical information of the parts. More specifically, a deep Multilayer Perceptron and Convolutional Neural Network (MLP-CNN) framework is designed to predict melt pool depth for any part geometries when printing metallic parts via LPBF using metadata (i.e., numerical and image data). Quite often the melt pool depth generated during the AM processes is employed as a surrogate for thermal, defect- and micro-structural signatures. In melt pool depth prediction of complex geometries in LPBF, a digital twin environment using a finite element model (FEM) is developed and validated using experimental melt pool measurements for different geometries within multi-tracks and layers. The FEM was in agreement with experimentation with an error of less than 15%. Through the process simulation, a large dataset of melt pool depths was obtained for various part geometries and different process parameters. Next, the MLP-CNN framework was established to identify the parallel impact of part design and process parameters on melt pool depths. To enhance the performance and robustness of this model, data augmentation has been implemented by rotating and transferring geometries to artificially expand the dataset. Data augmentation helps to mitigate overfitting and promote better generalization, especially in the context of limited training data. After training, the proposed model is found to give accurate melt pool prediction even for new geometries not considered during training. The fusion of CNN and MLP has led the melt pool predictions for unseen geometries with an accuracy of 95%.

CALPHAD-guided design and experimental study on a novel Al-Ti-Nb alloy

We have used a CALPHAD-guided design methodology to develop a novel lightweight Al-Ti-Nb alloy with nearly 50 vol% of in-situ formed trialuminide reinforcement. After compositional optimization, we selected Al87.5Ti6.25Nb6.25 (at%) as the experimental alloy composition and subjected it to a 24-hour homogenization heat-treatment at 475°C for attainment of equilibrium, guided by DSC and CALPHAD-derived melting point data. Phase and microstructure analysis revealed a microstructure with approximately 50 vol% of a FCC Al-rich matrix and 50 vol% of trialuminide phases for strengthening. The experimental onset melting point of the Al-rich matrix was determined to be 647°C, which significantly surpasses Al-Si based high-temperature alloys (577 ℃), showcasing the potential of this alloy as a high-temperature structural material. Nanoindentation tests demonstrated exceptional mechanical properties. The alloy was found to have a microhardness of 3350 MPa, which not only outperformed 7075 Al alloy, A390 Al-Si alloy, and commercially pure Ti but also exceeds the levels of Ti-6Al-4V alloy at nearly 28% lower density. This study's significance lies in its integration of CALPHAD predictions and meticulous microstructural investigation, offering an effective approach for the design and characterization of novel Al-based alloys.

Experimental melting of phlogopite websterite in the upper mantle between 1.5 and 4.5 GPa

Abstract Reaction experiments have confirmed that phlogopite websterite can be formed by interaction of peridotite with hydrous alkaline or silica-rich melts. Phlogopite websterites commonly occur as xenoliths in orogenic and intraplate volcanism but do not receive much attention. We have experimentally investigated the melting behaviour of a phlogopite websterite at 1.5 GPa (1050–1300 °C), 3.0 GPa (1100–1500 °C), and 4.5 GPa (1200–1500 °C) to contribute to understanding the sources of ultrapotassic rocks that occur in different settings. The solidus temperature rises with increasing pressure, bracketed between 1050 and 1100 °C at 1.5 GPa, 1100 and 1150 °C at 3.0 GPa and between 1200 and 1250 °C at 4.5 GPa. At 1.5 GPa, phlogopite websterite melts incongruently to form olivine and melt, whereas orthopyroxene, garnet and melt are formed at 3.0 and 4.5 GPa. The transition of orthopyroxene from reactant to product with increasing pressure results in changes in the SiO2-content of melts. The experimental melts reach a maximum K2O content when phlogopite is consumed completely at temperatures ~150 °C above the solidus. The melting reactions are similar to those of phlogopite lherzolite, but the low-Al2O3 starting materials result in lower Al2O3 in the melt than in melts of phlogopite lherzolite. Comparison with natural ultrapotassic rock compositions reveals that the sources of ultrapotassic rocks in convergent settings may be dominated by phlogopite websterite, phlogopite lherzolite and phlogopite harzburgite. Sources of ultrapotassic rocks in intraplate settings are more likely to include phlogopite clinopyroxenite ± CO2 and K-richterite. In all melting experiments on phlogopite-bearing rocks, K2O from phlogopite passes into the melt, and hence the highest K2O contents in ultrapotassic rocks must be an indication of the minimum stoichiometric coefficient of phlogopite in the reaction. In cases where phlogopite websterite or phlogopite lherzolite is identified as the source, the minimum modal percentage of phlogopite in the source can be inferred from the highest K2O content. When applied to the Milk River minettes and New South Wales leucitites, the estimated modal proportion of phlogopite in the sources is greater than 20 wt.%. Phlogopite can survive the subduction process and melt later in the post-collisional environment, whereas thermal perturbations are necessary to trigger melting of phlogopite-bearing assemblages at the base of the lithosphere in intraplate settings.

Boron removal in seawater desalination by progressive freezing-melting.

Desalination plays a crucial role in addressing water scarcity and promoting sustainable development. However, the presence of high boron content in seawater poses a significant challenge. This study introduces a progressive freezing-melting method that effectively removes boron while desalinating seawater. The experimental results indicated that salinity and boron rate of removal increased with freezing temperature and decreased with freezing duration. Among the experimental melting methods, ultrasonic melting (UM) and oscillatory melting (OM) were superior to natural melting (NM) for boron removal and desalination, with oscillatory melting proving to be the most effective. Specifically, when seawater was frozen at - 20 °C for 44h followed by OM of 55% of the ice, salinity and boron removal rates reached 96.79% and 97.60%, respectively. The concentrations of boron and salinity in the treated seawater were only 0.777‰ and 0.149mg/L. Moreover, the estimated theoretical energy consumption for treating 1 m3 of seawater was calculated to be 5.95 kWh. This study not only contributes to environmental sustainability but also holds significant potential due to its high efficiency in desalination and boron removal.

Numerical simulation approach for Laser Melting of lunar regolith using an enthalpy-based model

The Moon is considered a future destination for international space exploration. Transport costs to the Moon are enormous, requiring the use of available resources to build infrastructure. Lunar regolith, predominantly found in the form of dust, covers the entire Moon’s terrain. A promising approach to process regolith is laser-based additive manufacturing, which aims to melt it directly on the surface. This paper presents a finite element simulation model to investigate the effects of laser power and feed rate variation on the melting process, with the objective to predict experimentally unexplored parameter ranges. The focus of this study is the phase transition from solid to liquid, implemented by an enthalpy model. Validation is carried out using experimental melting tests with regolith simulants. Despite limited experimental data, the observed effects are consistent with the simulation results, confirming that sample thickness increases with laser power and decreases with feed rate.

Petrogenesis of the Morobe Granodiorite and their shoshonitic mafic microgranular enclaves in Maramuni arc, Papua New Guinea

Abstract The Miocene tectonics of Papua New Guinea, where subduction, arc-continent collision, and changes in subduction direction are considered to have occurred, is very complex and various tectonic models have been proposed. The Maramuni arc, active in the Miocene, is composed of a chain of granitoid bodies. As the chain-like distribution indicates the generation of igneous activities in a wide range of the same tectonic settings, the study of the Maramuni arc magmatism is important for elucidating the geologic events of the time. We provide data on the petrological and geochemical characteristics of the Morobe Granodiorite that form part of the Maramuni arc. The Morobe Granodiorite consists of metaluminous I-type granitoids, belonging to the medium-K to high-K series. The whole-rock major element variations in the granitoids can be explained by the fractionation of hornblende and plagioclase. They are generally within the composition range of experimental partial melts of amphibolites, and the whole-rock trace element compositions have characteristics of slab failure magma rather than arc. This suggests that the granitoids were generated by partial melting of the torn slab after slab failure. The mafic microgranular enclaves (MMEs) in the granitoids are classified as shoshonite, and their trace element compositions suggest that they were formed by partial melting of phlogopite-bearing mantle. The occurrences of native gold and barite within the MME show that MMEs transport Au from the mantle metasomatized by slab-derived sediment melt and/or fluid to the crustal magma chamber.

Effect of Alloying Additives and Moulding Technology on Microstructure, Tightness, and Mechanical Properties of CuSn10 Bronze.

Thise research was conducted to determine the impact of the applied casting technology, mould and alloying additives on the tightness of the CuSn10 cast alloy. Under industrial conditions, a series of experimental melts were made that were characterised by varying the concentrations of the main alloying element (Sn) and the introduced alloying additives (Si, Zn, Zr). The mould was made from green sand and used the CO2 moulding process. To assess the influence of the alloying additives, a metallographic analysis of the studied alloy was carried out, and the alloy's microstructure was examined using optical and scanning electron microscopy. The introduced alloying additives affected the properties and microstructure of the studied alloy. As alloying additives, zirconium resulted in a visible refinement of the microstructure, while silicon improved the fluidity and quality of the casting's external surface. The use of alloying additives and moulds made using different technologies is intended to improve the structure of the tin bronze castings produced and to find the best solution to significantly eliminate the lack of leakage of the castings. The castings were subjected to mechanical processing, and a leak test was performed using the pressure drop method. The conducted research allowed us to determine which technology, applied to production, will bring about a reduction in the problem and will inform further investigations.

Petrological and Geochemical Constraints on the Origin of Strongly Peraluminous Granitoids from the Triassic Guangtoushan Pluton in South Qinling

AbstractAs an important part of the early Mesozoic granites in the South Qinling tectonic belt (SQTB), the Guangtoushan pluton provides a material basis for research on the composition of magma sources and the effects of peritectic assemblage entrainment (PAE) on the changes in the granite composition. As shown by the results of LA‐ICP‐MS zircon U‐Pb dating, the Guangtoushan pluton was emplaced during the Late Triassic (214–212 Ma) and was formed in the post‐collision stage between the SQTB and the Yangtze plate. The collected samples had high SiO2 content and low Cr and Ni contents, indicating that the magmas did not undergo significant crust‐mantle mixing during their evolution. The Guangtoushan granitoids were distributed along the trend line of magmatic fractional crystallization in the F–An–Or diagram. This result, combined with the relatively homogeneous Sr‐Nd isotopic composition, implies that the Guangtoushan pluton underwent slight assimilation and contamination. As can be inferred from the comparison between the compositions of the Guangtoushan granitoids and various fluid‐absent experimental melts, the magma sources of the Guangtoushan granitoids contain a variety of materials, such as graywackes, pyroclastic graywackes, and pelites and are not derived from lower crustal mafic rocks. The correlation between the maficity and the major and trace elements further indicates that the strongly peraluminous granitoids from the Guangtoushan pluton was formed by the partial melting of biotite‐bearing crustal rocks and its magmatic evolution was accompanied by the entrainment of clinopyroxenes and accessory minerals.

Modeling and measurements of creep deformation in laser-melted Al-Ti-Zr alloys with bimodal grain size

Elemental powders of Al, Ti, Sc, and Zr are blended to create binary Al-0.5Ti and Al-0.25Ti-0.25Zr (at%) alloys via laser powder-bed fusion (L-PBF), with Al-1Ti and Al-0.25Sc-0.25Zr alloys produced for comparison. The Al-Ti alloys formed primary D023-Al3Ti precipitates upon solidification, whereas the Al-(Sc/Ti)-Zr alloys formed primary L12-Al3(Sc,Ti,Zr) precipitates which nucleated bands of fine grains. Upon aging, the Al-Ti alloys do not harden (consistent with most Ti remaining in solid solution), unlike the Al-(Sc,Ti)-Zr alloys, as expected from precipitation of secondary L12-Al3(Sc,Ti,Zr) nanoprecipitates. Upon compressive creep testing at 300–400 °C, Al-0.25Ti-0.25Zr exhibits steady-state strain rates comparable to those of Al-0.5Zr at low stresses (up to ∼14 MPa), but about an order of magnitude faster at higher stresses. The effects of bands of fine grains (formed by L12-Al3(Sc,Ti,Zr) primary precipitates at the bottom of the melt pools) on hardness and creep resistance is investigated via finite element modeling of simple slab models, with varying geometries approximating experimental melt pool geometries. The degree of connectivity of coarse- and fine-grain regions is a greater factor in the resulting creep strain rates than the volume fraction or spatial orientations of these two regions. This suggests that reducing the concentration of Zr (in Al-0.5Zr), or partially substituting with a less expensive element such as Ti (as for Al-0.25Ti-0.25Zr), may maintain the added l-PBF processability benefits of Zr and reduce the volume fraction of fine grains without sacrificing creep resistance, allowing for further tailoring of custom microstructures.

Prediction of melt pool geometry by fusing experimental and simulation data

Accurate prediction of the melt pool geometry in laser-directed energy deposition (L-DED) is crucial for ensuring product quality. Although machine learning methods have offered aid in addressing this challenge, existing surrogate models are typically based on either simulation or experimental data. Surrogates trained using numerical simulation data reflect modeling assumptions in their accuracy. Conversely, surrogates developed using experimental data would demand a substantial investment of effort and resources to generate adequate data for training. As a possible solution, in this paper, a multi-fidelity Gaussian process (MFGP) framework is developed to fuse experimental and simulation data. Experimental melt pool dimensions are obtained from single-layer, single-track deposits fabricated via powder-fed L-DED. These constitute the high-fidelity/ground truth data for the MFGP surrogate. An analytical Eagar-Tsai model is calibrated and queried at sampled input points within a predefined process parameter window to yield near-accurate estimates of low-fidelity melt pool data at a significantly lower effort. Thereafter, the MFGP surrogate is designed on the combined sources of data using an appropriate kernel and adequate calibration. An elaborate assessment of the trained surrogate on unseen experimental data is shown to yield results with high accuracy and low uncertainty. The proposed method of blending experimental data with simulation data has the potential to reduce the volume of experimental data required for creating surrogates without compromising accuracy. While the current work focuses on the L-DED of a popular alloy, SS316L, it can be easily extended to other advanced manufacturing processes to design reliable surrogates.

Development of a comprehensive model for predicting melt pool characteristics with dissimilar materials in selective laser melting processes

Selective laser melting (SLM) is a popular additive manufacturing (AM) technique for the fabrication of precise components. The success of SLM relies on accurate control of the melt pool, where the metal powder melts and solidifies. Fluid flow inside the melt pool is driven by forces like Marangoni force and recoil pressure. SLM process begins with the fabrication of support, which involves the process of joining dissimilar metals. A strong support-base substrate connection ensures the success of subsequent additive manufacturing. In this study a comprehensive model was developed to predict the dimensions and characteristics of the melt pool with a focus on the first layer of the SLM process. An initial numerical simulation was done to obtain transient melt pool temperature and velocity distribution during the SLM process. The results were then coupled with a theoretical analysis development of the proposed model. The effectiveness of the model was validated by experimental results and melt pool characteristics from previous publications. The model shows Root Mean Square Error (RMSE) of 0.53 and 0.49 for height-to-width ratio and normalized area, respectively. The Mean Absolute Percentage Error (MAPE) of normalized area is 15.7% for the majority of the data, which demonstrates a good accuracy of the model. In addition, the transition from conduction mode to keyhole mode can be identified with the model. This model can be used to rapidly and effectively predict the melt pool characteristics for SLM and other similar AM processes with a wide range of processing parameters.

The origin of Na-alkaline lavas revisited: new constraints from experimental melting of amphibole-rich metasomes+lherzolite at uppermost mantle pressure

We present a new experimental dataset for reaction experiments between natural amphibole-clinopyroxene metasomes (hornblendite) and synthetic lherzolite that produced Na-rich alkaline melts. Experiments were conducted at 1, 3 and 4 GPa and 1000–1300 °C. The generated melts range from foidite over basanite to phonotephrite. At 1 GPa between 1000 and 1100 °C amphibole decompression-breakdown products generate a phonotephritic melt. Among the breakdown components rhönite was found to be stable up to 1100 °C and 1 GPa. At 3 and 4 GPa the melt compositions are affected by phlogopite melting and shift to more foiditic compositions. We find that the melting of hornblendites and the reaction of the melt with the lherzolite produce wehrlitic residues with different olivine/clinopyroxene ratios. Wehrlite formation does not always require separate metasomatic processes but can be a direct by-product of alkaline volcanism. We applied a metasome melting model to the magmas of the Kula volcanic province, Turkey, and show that at 1 GPa basanite melts and phonotephrite melts cover the whole range of known Kula lava compositions. The Kula lava compositional trend can be therefore generated by basanite-phonotephrite melt mixing. A comparison of high-pressure (3–4 GPa) melts with natural nephelinite data shows overlap with many major, minor, and trace elements but differences in SiO2, FeO, and TiO2 argue that the natural nephelinite data do not represent primary metasome melts.