Table of contents

Preface

The 16^th International Conference on the Modelling of Casting, Welding, and Advanced Solidification Processes (MCWASP XVI) was held from June 18 to 23, 2023, in Banff, Canada, at the Banff Centre for Arts and Creativity. Founded in 1933, the Centre in Treaty 7 Territory within Banff National Park—Canada's first National Park—is a learning organization built upon an extraordinary legacy of excellence in artistic and creative development. The "all-inclusive" nature of the conference and the remote setting meant that participants dined, attended oral and poster presentations, and participated in social activities as a group, fostering outstanding opportunities for networking.

Given that the MCWASP community had not met in person since 2015 in Japan (the 2020 edition of MCWASP was virtual owing to COVID-19), the 2023 conference provided the opportunity to renew old friendships and make new ones as well as discuss the science of solidification and related processes—all within the backdrop of the beautiful Canadian Rocky Mountains.

The technical program comprised more than 70 oral and poster presentations. In addition to content related to modelling of casting, welding, and advanced solidification processes, keynotes were invited to talk about related subjects (artificial intelligence/machine learning, and permeability modelling in shale rock) as well as the rich diversity of fossils, especially dinosaurs, found in Alberta.

The oral technical program was organized with as a single session (i.e., no concurrent presentations). It featured all aspects of solidification modelling, including solidification process technologies (continuous and semi-continuous casting, shape casting, additive manufacturing, and welding), coupled multi-physics simulations, defect formation, fluid flow, micro- and macro-structure formation, numerical methods, and related experimentation, especially in-situ observation of solidification.

The four-day technical program was spread over five days to give participants the opportunity to explore the stunning Canadian Rocky Mountains.

In these proceedings, the papers are organized by major theme. The dominant topics are Additive Manufacturing and Welding and Microstructure Formation, followed by Continuous Casting – Shape Casting, Heat Transfer and Fluid Flow, Alloy Segregation, Defects, Imaging of Solidification, Thermomechanics, and Materials Properties. In these themes, the authors report advances in numerical modelling techniques, new scientific and process developments in solidification, and related in-situ experimentation.

Although significant progress has been made over these past 16 MCWASP conferences covering 43 years, it is clear that the complexity of advanced solidification phenomena as related to conventional and emerging manufacturing technologies still attracts a great deal of scientific and industrial interest to support technological innovation.

André Phillion

Banff, Canada, June 2023

MCWASP XVI 2023

List of Peer Reviewers, Sponsors, MCWASP XVI Organizers, International Scientific Committee are available in this Pdf.

All papers published in this volume have been reviewed through processes administered by the Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.

• Type of peer review: Single Anonymous

• Conference submission management system: Morressier

• Number of submissions received: 79

• Number of submissions sent for review: 78

• Number of submissions accepted: 71

• Acceptance Rate (Submissions Accepted / Submissions Received × 100): 89.9

• Average number of reviews per paper: 1

• Total number of reviewers involved: 31

• Contact person for queries:

Name: Janice Burke

Email: jburke@cim.org

Affiliation: Metallurgy & Materials Society of the Canadian Institute of Mining, Metallurgy and Petroleum

To reduce computation time, the inherent strain (IS) method is popular. It consists in adding layers, at room temperature, with an "inherent strain" representing the plastic deformation undergone during deposition. An IS method is developed, including a direct determination of the IS tensor based on the transient thermo-elastic-viscoplastic (TEVP) simulation. Validation is achieved: taking full-field inherent strains, the IS method allows retrieving TEVP predictions. However, application to a full part, for which a set of inherent strains must be conserved all along the construction, leads to poor results, far from TEVP. As an alternative, a new "inherent strain rate" (ISR) method is proposed, consisting in linearizing TEVP resolutions. Combined with an on-line learning technique, this strategy leads to results identical to TEVP reference, with still a significant speed-up. This makes the proposed ISR method very promising.

A transient 3D multiscale model has been created to simulate the evolution of solidification microstructures rapidly and effectively in Ti6Al4V parts produced through Laser Powder Bed Fusion (LPBF). The microstructure simulation tool has been enhanced to account for rapid solidification conditions in Ti6Al4V alloys during processing of multi-layer multi-track LPBF parts. The simulation tool can evaluate the impact of part geometry, lase power input, laser speed, and laser beam shape on the formation of the microstructure in LPBF-processed Ti6Al4V alloy. The multiscale model considers several factors, including preferential crystallographic growth direction, isomorphism, epitaxy, melt pool motion, and temperature gradients to generate the observed texture and morphology of the microstructure in Ti6Al4V components. The model can evaluate various microstructure characteristics, such as grain size and texture. Consequently, it could aid in controlling the formation of solidification microstructures in Additive Manufacturing (AM) processed parts. The simulation tool has been previously validated by using IN625 laser remelting experiments made by National Institute of Standards and Technology. The 3D simulation tool can also be utilized to predict the microstructure formation in Ti6Al4V components produced by the Electron Beam AM processes.

Selective laser melting is of great expectation to be used in additive manufacturing of aerospace components with complex geometry. However, there are still defects in the built parts, such as solutal segregation and unexpected microstructure, which contribute to cracks and lead to failure. At present, most of the simulations focus on the macroscopic grain structure, and the solute transport process has not been well demonstrated yet. In the present work, we develop a two-way fully coupled model based on cellular automaton and finite volume method to simulate the solute transport and dendritic structure evolution during the melting and solidification of the SLM process. The results reveal the microstructural evolution and solute transport during the melting, spreading, and smearing of the powder. The proposed model framework shows good potential to be applied to further numerical investigation on the solidification behaviours of the SLM process.

Although the observation of wetting phenomena has a long history both in science and in everyday life, nanoscale wetting properties leave open a multitude of questions. Improving the properties of Direct Energy Deposition (DED) products through reinforcement with nanoparticles (NPs) is attractive. Due to the involved NPs, the wetting behaviour of the deposition differs from that in macroscales. Modelling and validation of a laser-based DED process reinforced with NPs in view of the effect of wetting is demonstrated in this study. Metal alloys reinforced with NPs (Ti6Al4V- and Al7075-TiC) are melted by laser light and deposit on a substrate. Heating of the alloys by laser light is tackled using a volumetric heat source model. A three-phase (solid, liquid and gas) melting and solidification simulation methodology is established in a continuum approach. Nanoparticles are treated in the Lagrangian framework. Effect of nanoparticle reinforcement on the wetting behaviour of melt bead during the printing process is investigated numerically, taking into account the influence of nanoparticles on contact angle and surface tension. Simulation results are compared with the experimental deposition of Ti6Al4V and Al7075 alloys through dimensional analysis of the tracks by macrography.

A thermo-mechanical finite element model is developed for additive manufacturing by directed energy deposition (DED). The simulation is conducted at the part scale by modelling the progressive deposition in the way of the fraction of the layer. To incrementally resolve the displacement, strain, and stress fields, a theoretical formulation for the kinematic positioning is proposed to minimize the distortion of the non-constructed fraction by considering current displacement and strain in the constructed part, which can resolve the discontinuous problem at the interface for a patch of material depositions. Moreover, it is found that the kinematic hardening couldn't be ignored for the back-and-forth deposition mode of the DED process. The application for a turbine blade with a strong curvature is adopted, and the distributions of distortion and stress during mid-construction and final construction are studied.

Laser clad overlays provide enhanced corrosion resistance and wear performance to components across a wide range of industries. Powder fed laser processes provide flexibility in material selection, however have traditionally yielded lower material usage efficiency (60–80%) compared to wire fed processes. A general model was developed to predict powder catchment efficiency based on the overlap between the molten pool and powder cloud distribution. Mathematical techniques of asymptotic analysis and blending were used to obtain closed-form expressions relating tabulated material properties and the key process inputs with the critical melt pool dimensions of leading length (x_f) and maximum width (y_m). An improved experimental technique was implemented to measure the powder cloud distribution, a key process input which previously was not well understood. For baseline cladding conditions, melt pool dimensions were smaller than the powder cloud leading to catchment inefficiency. A custom laser optics module was designed to enable adjustment of the relative position of the melt pool and powder cloud independently from other process parameter changes. Combining optimized parameters with a shift of 1mm, catchment efficiency >90% has been achieved and sustained in production cladding operations. This work is part of a broader program at Apollo-Clad, using simple, accurate, and fast analytical modelling techniques to generate engineering design rules for laser-based processes.

In welding and wire-arc additive manufacturing (WAAM), a mobile arc is the heat source that enables the deposition of metals and the resulting properties of the final product. Because the arc involves temperatures of 20 000 K, and gas velocities of the order of 300 m/s, there are only a few experiments and models available to determine optimal, or at least acceptable, parameters for the operation such as current, voltage, and arc length. On the other hand, there is a lack of engineering guidance to optimize the processes resulting in costly and time-consuming trial-and-error optimization methods, which also involve wasteful use of energy and scrap parts. In this work, a numerical model of the gas-tungsten arc welding (GTAW) arc was created and validated against experiments. The model considers the arc interactions between a non-consumable electrode and the weld pool and accounts for multiple coupled heat transfer mechanisms: Joule heating, conduction, advection, radiation, and Thomson effect. The conditions considered cover the vast majority of GTAW welding operations. The results are generalized in the form of engineering expressions suitable to be embedded in metamodels, in which the heat source is just a part. Applications include penetration and width of welds and deposition rate in external-wire WAAM.

Laser powder bed fusion (L-PBF) is an additive manufacturing method which involves local laser melting of powder particles, a partial remelting of previously deposited layers, and subsequent re-solidification under high thermal gradients and cooling rates. The transition between melting and re-solidification becomes visible as melt pool boundaries in optical micrographs and plays a crucial role: Apart from creating a strong segregation zone, the transition determines whether the microstructure is inherited and carried over to the next layer, or whether new grains with new orientations are formed. While heterogeneous nucleation is suppressed due to the lack of seeding particles at the small length scales inherent to L-PBF, alternatively, new grains can form via dendrite fragmentation, as demonstrated in this paper by phase-field simulations using the software MICRESS^®. By strong coupling between the phase-field equation and a thermal 1D-cylinder approach for the long-range temperature field, consistency between latent heat and microstructure is ensured. To allow for a systematic variation of the orientation relationship between the dendrite growth direction and the respective temperature gradient, a two-step simulation procedure for two overlapping tracks with variable gradient directions is developed. Growth conditions which promote fragmentation and formation of new grains are analyzed and discussed.

Laser powder bed fusion (LPBF) is an extensively used additive manufacturing process that can build metal parts with complicated geometric designs. However, because of the rapid solidification conditions and the layer-by-layer building, its application is challenged by products having poor surface quality and reduced mechanical properties. The laser rescanning process is often used as a refinement method during LPBF to improve the quality of products. In this study, grain structure formation during the LPBF laser rescanning process is modelled by a 3D cellular automaton based microstructure model coupled with finite element analysis. The coupled model considers different nucleation mechanisms, including epitaxial growth, which are applicable to rapid solidification in melt pool. The model is adapted to reproduce the grain structure and evolution of an Al-Si alloy manufactured by LPBF utilizing a laser rescanning strategy. The effects of laser remelting on the characteristics of the grain structure (i.e., the grain size, aspect ratio and orientation) are evaluated. The mechanisms that enable unique grain structure (i.e., grain refinement) are discussed.

Heat transfer and stress distribution are studied using computational fluid dynamic and structural simulation for a conformal cooling die insert made of Maraging M300 steel by selective laser melting. The phase transformation of cooling water is studied to understand the cooling effects and stress concentration on die life. The results of simulation are validated in production. The failure of the die insert has been revealed a result of enhanced stress concentration and corrosion directly related to the formation of vapor in the conformal cooling channel.

Grain growth competition during solidification determines microstructural features, such as dendritic arm spacings, segregation pattern, and grain texture, which have a key impact on the final mechanical properties. During metal additive manufacturing (AM), these features are highly sensitive to manufacturing conditions, such as laser power and scanning speed. The melt pool (MP) geometry is also expected to have a strong influence on microstructure selection. Here, taking advantage of a computationally efficient multi-GPU implementation of a quantitative phase-field model, we use two-dimensional cross-section simulations of a shrinking MP during metal AM, at the scale of the full MP, in order to explore the resulting mechanisms of grain growth competition and texture selection. We explore MPs of different aspect ratios, different initial (substrate) grain densities, and repeat each simulation several times with different random grain distributions and orientations along the fusion line in order to obtain a statistically relevant picture of grain texture selection mechanisms. Our results show a transition from a weak to a strong ⟨10⟩ texture when the aspect ratio of the melt pool deviates from unity. This is attributed to the shape and directions of thermal gradients during solidification, and seems more pronounced in the case of wide melt pools than in the case of a deeper one. The texture transition was not found to notably depend upon the initial grain density along the fusion line from which the melt pool solidifies epitaxially.

High maintenance costs due to significant abrasive wear of components is experienced in the energy and mining sectors despite the current use of tough and hard coatings. During the coating process significant tensile residual stresses may build up and result in premature failure of a component. These tensile stresses can be reduced by adopting functionally graded structures of the composite. The main goal of the present study is to design an ideal material gradient in the additively manufactured composite using the plasma transferred arc (PTA) with a WC-Ni alloy graded in WC. To develop a comprehensive analysis of the functionally graded deposit, the thermal history of the WC and Ni alloy powders are first simulated as they travel through the plasma and deposit on the substrate. The initial temperature of each deposited bead depending on the WC content is deduced. These results are used as an input to predict the temperature and stress history of the solidifying deposit. Thermal stresses are computed and trends of residual stresses are obtained as a function of the WC gradient selected. The trends obtained are compared quantitatively before concluding on the most favorable gradient for this wear resistance application.

Numerous studies have used FEM simulations to assess the effects of the heat source parameters on the melt pool volume during metal powder bed additive manufacturing. However, considerable debate still exists on how to incorporate the evolution of the thermophysical properties used to describe the powder as it undergoes heating, melting, consolidation and finally solidification. For single layer studies, since powder volume is much smaller compared to the substrate volume, highly detailed, computationally expensive powder property descriptions may not provide a commensurate increase in accuracy of simulation results. This study aims to quantify the effect of powder properties on the melt pool volume created during electron beam melting of Ti6Al4V powder using predictions from a FEM-based heat conduction model. The dependence of thermal conductivity, specific heat, and density of the powder on temperature and beam power density will be studied. Additionally, the relevance of the powder properties with changing layer height and beam power and speed will also be quantified.

Advanced modelling of additive manufacturing often requires the combination of models at multiple scales and multi-physics. Therefore, building the modelling workflow describing the process is complicated. The modelling is also only a part of the innovation process and must be connected to practices, experimental work, and characterisation. Efficient communication and data exchange between the different actors could quickly become a challenge. Recent developments in the frame of the EMMC (European Material Modelling Council) and in the EU project OntoTrans points toward the early integration of semantic description and the creation of dedicated domain ontologies. This require an unambiguous and consistent use of terms and definitions for various concepts within each field of technology, and international standards is an available source for structured technical terms and definitions. For additive manufacturing (AM) the international standard ISO/ASTM 52900 "Additive manufacturing - General principles - Fundamentals and vocabulary" is the internationally recognised source for terms and definitions. Basing the ontology on the AM terminology standard will greatly facilitate integration of AM processes as a part of an industrial manufacturing system. Therefore, the present work attempts to harmonise the standard terminology and the ontology concepts. Then, to improve the impact and connection to material science, the concepts will be connected to a microstructure domain ontology and to the top- and middle-level ontology EMMO. The conceptualisation and application of the ontologies will be illustrated through simple examples of process and material modelling.

Inoculation is an important method to generate fine equiaxed grain structures in alloys during solidification, which is widely used in shape casting and direct chill casting processes of aluminium alloys. Different from the normal casting processes, welding and additive manufacturing can have ultrahigh cooling rate and temperature gradient during solidification. This makes the grain refinement by inoculation method more difficult. Therefore, a deeper understanding on the influences of cooling rate and temperature gradient on the nucleation and growth of grains is necessary. In this work, a numerical model is developed to quantitatively address the above influences. The model has been successfully applied to predict the grain size in the welding of aluminium alloys as a function of locations.

Understanding the development of a number of defects found in components fabricated by the metal Additive Manufacturing (AM) processes requires an understanding of the evolution in the thermal field within the component at both the macro- and meso-scales. As a first step, in this work, the agglomeration method was used in combination with a time-averaged input of energy to simulate the macro-scale evolution in temperature. Two example processes: 1) laser-based powder-fed directed energy deposition; and 2) electron beam powder bed fusion, are used to demonstrate the modelling methodology. The approach employed focuses on ensuring the conservation of heat and is applied using ABAQUS. The two applications have been validated by comparing the predicted thermal behaviour with process-derived data. The results indicate that this method is an efficient strategy to predict the thermal field at the scale of the component being fabricated.

A recently-developed [1] multi-component phase field model has been utilized to investigate microstructure evolution during in-situ alloying of a blended elemental Ti-1Al-8V-5Fe alloy powder via the Laser Powder Bed Fusion process. The process of in-situ alloying, where elemental powder is used instead of pre-alloyed powder, was studied by performing two simulations having: (1) a uniform initial composition, and (2) a spatially varying initial composition to represent different powder particles. Specifically, the grain morphology, solute distribution, competitive growth and nucleation under the two different scenarios were simulated and compared. To assist the microstructure simulations, a macro-scale finite element model was developed to simulate the heat transfer during LPBF process. The thermal history data calculated by the finite element model was provided to the phase field model in order to simulate transient dendritic growth behaviour. The results show that a set of evenly-spaced columnar dendrites form in the uniform initial composition case, whereas when the initial composition is spatially varying, non-uniform dendrites having elongated shape can develop. It is also shown that competitive growth between dendrites is influenced by nucleation. For the spatially varying initial composition case, the results indicate that full alloying is difficult to achieve during the LPBF printing process; this incomplete alloying greatly influences the dendrite morphology and solute distribution.

Laser welding can be used to join dissimilar materials to produce lightweight structures, and electric vehicle battery systems, which are important means of limiting the carbon emissions in the transport industry. Due to the differences in melting temperatures, thermal conductivities, and mutual solubility of dissimilar materials, it is still challenging to create defect-free joints with high mechanical strength, or low contact electrical resistance. In this work, we present a state-of-the-art numerical model of laser welding, developed within the Computational Fluids Dynamics (CFD) paradigm. The multi-physics model simulates melting, flow, and solidification of the alloys and accounts for the laser-material interactions, phase change, temperature and alloy-dependent thermophysical properties, recoil pressure, buoyancy force, and Marangoni effect. The simulation predicts weld penetration depth and width, alloy mixing, as well as the temperature gradient and cooling rate during the solidification that can be further fed into a micro-structure prediction model. The model is coupled with an optimization tool, which iterates over different process parameters to optimize the joint. The methodology is presented for steel to aluminium welding in lap configuration, but it can be used for other materials such as steel-copper, aluminium-copper, or steel-nickel, and in arbitrary geometrical configuration.

Numerical modelling methods (e.g. finite element) can provide accurate descriptions of long-range temperature fields in laser or electron-beam melting processes, however the high computational costs at part-scale make them unsuitable for process modelling in additive manufacturing (AM). Alternative methods such as semi-analytical solutions based on a moving heat source reduce the computational expense but at the cost of unrealistic assumptions. Radiation, temperature-dependent physical properties and latent heat are not considered in the semi-analytical approach but can have a significant effect on the thermal history. In this study, the error associated with each of these contributions are assessed against the conduction-only semi-analytical solution for a range of processing parameters for surface melting on solid Ti-6Al-4V. The semi-analytical model is then "enhanced" using results from finite element simulations to better account for the heat transfer in the AM process.

This article explores the reason behind the higher deposition rate in Mg-containing aluminum alloys in gas metal arc welding (GMAW). Experiments performed on ER1100 (Mg-free alloy) and ER5183 (high Mg alloys) measured current, wire feed speed, and droplet temperature. A non-linear analysis of Joule heating at the electrode extension and an analysis of the power needed to heat the consumable to the droplet temperature were performed. Analysis revealed that Joule heating is two orders of magnitude smaller than what was necessary for the higher wire feed speed. The droplet temperature measurements were done for the first time for both consumables under identical conditions, which revealed that ER5183 had a lower droplet temperature than ER1100 by almost 350 K.

This paper summarizes multiple engineering expressions that enable the prediction of thermal magnitudes of interest associated with moving heat sources. The expressions use only fundamental parameters available before performing any experiments, and their calculation is algebraic, without the need for numerical methods. All expressions are based on the fundamental governing equations of heat transfer in the solid. The magnitudes predicted include maximum width and its location, maximum penetration, thickness of HAZ, maximum temperature and its location, leading and trailing edge of an isotherm, heating and cooling rate, aspect ratio of an isotherm, melting efficiency, cooling time from 800°C to 500°C, solidification time, and maximum distribution of a heat source to reach a target temperature. Parameters involved include heat source power and speed, thermal conductivity and diffusivity of the substrate material, temperature of interest and preheat or interpass temperature. Temperature-dependent properties are accounted for by the use of effective properties. The expressions proposed can be extended into sophisticated geometries for welding and specific additive manufacturing cases.

Meltpool modulation in Selective Laser Remelting Additive Manufacturing via an oscillating magnetic field generates Thermoelectric Magnetohydrodynamics (TEMHD) flow. Numerical predictions show that the resulting microstructure can be significantly altered. A multi-scale numerical model captures the meso-scale melt pool dynamics coupled to microscale solidification showing the microstructure evolution and solute redistribution. The results highlight the complex interaction of the various physical phenomena and also show the method's potential to disrupt the epitaxial growth defect. The model predictions are supported by preliminary experimental results that demonstrate the dependency of the melt pool depth on magnetic field orientation. The results highlight how a time-dependent field has the potential to provide an independent control mechanism to tailor microstructures.

Heat input is a key component of a welding procedure, which is dependent on the several fall voltages composing a total voltage loss. As current practice is primarily reliant on trial and error to determine voltage settings for a desired heat input, a means to understand and predict the voltage loss is of interest to welding engineers. Voltage, amperage, and arc length measurements using ER4043 1.2 mm at varying voltages were used to break down fall voltage constituents for a given weld with GMAW-Sp. Arc length was defined as the distance from the weld pool to the point where the metal vapour and ionized gas boundary attach to the consumable, and measured over 5 droplet cycles to obtain a time average. Aluminum procedures in the Lincoln Procedure Hand Book (LPHB) with 1.2, 1.6, and 2.4 mm consumables were analyzed to predict the individual fall voltage constituents using experimental results. Expected arc lengths ranged from 13-26 mm depending on voltage prescribed in the procedure, and agreed with comparative experiments. Combined anode/cathode, and the arc column were major contributors to overall fall voltage with 64±3%, and 34±3%, respectively. Contact tip, electrode extension, and lead cables were minor contributors, each contributing less than 1.2% to the overall voltage loss.

Additive manufacturing (AM) may have many advantages over traditional casting and wrought methods, but our understanding of the various processes is still limited. Computational models are useful to study and isolate underlying physics and improve our understanding of the AM process-microstructure-property relations. However, these models necessarily rely on simplifications and parameters of uncertain value. These assumptions reduce the overall reliability of the predictive capabilities of these models, so it is important to estimate the uncertainty in model output. In doing so, we quantify the effect of model limitations and identify potential areas of improvement, a procedure made possible by uncertainty quantification (UQ). Here we highlight recent work which coupled and propagated statistical and systematic uncertainties from a melt pool transport model based in OpenFOAM, through a grain scale cellular automaton code. We demonstrate how a UQ framework can identify model parameters which most significantly impact the reliability of model predictions through both models and thus provide insight for future improvements in the models and suggest measurements to reduce output uncertainty.

Clogging of submerged entry nozzle (SEN) during continuous casting of steel is an undesirable phenomenon leading to different problems like flow blockage, slag entrainment, nonuniform solidification, etc. A transient numerical model for nozzle clogging based on an Eulerian-Lagrangian approach was developed and it covers the main steps of clogging: (a) formation of the first oxide layer by chemical reactions on the steel-refractory interface; (b) motion of non-metallic inclusions (NMIs) due to the turbulent melt flow towards the SEN wall; (c) interactions between the melt, the NMI, and the wall; (d) formation and growth of the clog by the deposition of NMIs on the clog front and the flow-clog interactions; and (e) detachment/fragmentation of a part of clog due to the flow drag force. Clogging in an industrial scale SEN was simulated. The simulated clog front was compared with real as-clogged SENs. The modeling results have successfully explained the SEN clogging induced transient flow phenomenon in the mold region, i.e. the transition from the stable to an unstable and non-symmetrical flow.

The thin slab casting (TSC) of steel is a type of the continuous casting (CC) with a narrow funnel-shaped mold, characterized by the rapid solidification and fast production rates. A highly turbulent flow impacts on a growing solid shell due to the constant feeding of the fresh hot melt. That strongly affects the solidification profiles and final quality of the TSC slabs. The presented work numerically investigates the solidification inside the TSC mold with the asymmetric flow pattern caused by the misalignment (tilting) of the submerged entry nozzle (SEN). These effects were considered with and without the applied electromagnetic brake (EMBr). The influence of the adjustable EMBr on the asymmetric flow and solidification profiles including turbulent and magnetohydrodynamic (MHD) effects were studied. During consistent series of simulations, the EMBr was varied between the magnetic poles and the time-averaged velocity and temperature fields were collected. The results showed that symmetric EMBr of a local type could partially improve the asymmetry. An optimal braking scenario was found for the casing speed of 5.5 m/min and maximum EMBr value of 180 mT. The solidification and MHD models including turbulence were developed using OpenFOAM®.

In the automotive industry there is a trend towards large components that are manufactured using high pressure die casting process (HPDC), in particular when many previously welded individual parts are to be replaced with one cast part to save costs and energy. In the case of large components, incipient solidification during mold filling cannot be ruled out. Casting defects due to cold running, air pockets and porosity cannot be separated spatially and temporally and influence each other. This stronger linking of the defects requires a realistic depiction of the casting process in the simulation, since the interactions between effects are more difficult to depict through approximations. The multi-phase approach presented here offers various options for this: The air is calculated as a compressible gas that is separated from the melt by a sharp boundary surface. Reduced melt flow due to solidification is represented by a porous media approach, followed by a complete flow stop at high solids. The formation of porosity due to volume shrinkage is coupled with a gas evaporation model. The multi-phase approach was validated by casting trials using a specially designed test geometry for thin-walled aluminum HPDC applications. First results of an industrial application are shown.

A model of turbulent multiphase flow, heat transfer, and particle entrapment during continuous casting of steel is presented. The model includes the top 7m of the vertical and curved strand, and considers the effects of argon gas injection and thermal buoyancy. RANS model flow results are compared with LES. Lagrangian particle transport through the Eulerian-Eulerian multiphase flow field from the k-ɛ RANS model is based on random walk and features advanced particle capture criteria, improved from previous work by including anisotropic turbulent velocity fluctuations near the walls. Particle capture via 3 mechanisms is included: capture by solidified hooks at the meniscus, entrapment between dendrites, and engulfment by surrounding large particles. The fluid flow and bubble/inclusion capture results are validated with plant measurements, including nail board dipping tests and ultrasonic tests of particle locations, and good agreement is seen. The superheat has negligible effect on flow in the mold region but causes complex flow in the lower strand by creating multiple recirculation zones due to the thermal buoyancy. With high (30 K) superheat, this leads to less penetration, and slightly fewer and shallower capture of particles. Lower (10 K) superheat may enable significant top surface freezing, leading to very large internal defect clusters. Lower superheat also leads to deeper meniscus hooks, and more surface capture. Capture bands occur near the transition from vertical to curved, where downward liquid flow velocity balances the particle terminal velocity.

Macrosegregation presents a considerable defect in the continuous casting of billets and can critically affect the final properties of the product. The numerical modelling can help to predict and better understand the segregation and flow patterns inside the mould. The process is modelled with a physical model described by a set of conservation equations describing the t heat transfer, turbulence, fluid flow, solidification and segregation. A two-equation low-Re k-epsilon model and Abe-Kondoh-Nagano closures are used to close governing equations in this incompressible fluid flow example. The Boussinesq approximation is applied to account for the thermo-solutal buoyancy effects, and the Darcy approximation is applied for the description of the flow through the porous mushy zone. On a microscale, a lever rule solidification model is used to couple liquid fraction, temperature and concentration. The three-dimensional model is solved with the method based on local collocation with multiquadric radial basis functions on seven-nodded subdomains. The aim of this contribution is to explore the three-dimensional macrosegregation patterns of 0.51 wt% carbon steel in the solidified shell of the steel in the mould.

State-of-art simulation models provide quantitative insights into flow, solidification and stress formation for continuous casting processes. This includes the entire process, from the tundish and the flow into the mold to the solidifying strand, which is withdrawn through various cooling zones. Process simulation and optimization provides important information about quality and productivity to evaluate process alternatives. This is only be possible if all relevant process parameters can be taken into account. The use of electromagnetic stirring is a technology which plays a significant role in the majority of continuous casting processes worldwide and its effects cannot be neglected in simulation models. This paper will discuss the modeling of electromagnetic stirring (EMS) and its impact on steel slab continuous casting.

Two cases with and without EMS are presented. The theoretical background to calculate the Lorentz force are described. The EMS calculation described here works with traveling (linear) magnetic fields. The effect of the EMS on the flow behavior, solidification and macrosegregation is shown on an industrial-scale slab casting.

This information leads to a better understanding of the EMS process in industrial applications to avoid casting defects, improve the quality of the final product, and increase the efficiency of the casting process.

Clogging of the submerged entry nozzle (SEN) is a common and persistent issue in the continuous casting process for steel alloys. The dominant clogging mechanism has been attributed to the deposition and accumulation of non-metallic inclusions (NMI) in the steel melt. Prior studies of NMI deposition assume that every collision between an inclusion and the nozzle wall results in adhesion, which is unrealistic. In this study, a macroscopic transport model for fluid and NMI motion is combined with a microscale model for NMI adhesion and applied to a slide-gate nozzle. Eight NMI sticking probabilities (S) and three slide-gate linear opening positions are explored. Simulation results indicated that the more closed the slide-gate, the greater the total deposition of NMI. When the slide-gate was partially open, the particle area number density was highest above, within and just below the slide-gate. But when the slide-gate was fully opened the deposition was concentrated in the upper tundish nozzle. Deposition behaviour fell into two regimes based on S. When S ≤ 0.05, the particle deposition was low and increased rapidly with sticking probability. When S ≥ 0.05, the particle deposition was high but changed very little with sticking probability. Changes to sticking probability did not significantly affect the deposition locations or relative distribution of particles within the nozzle.

In the continuous casting of peritectic steels, unevenness and cracks of the solidifying shell are typical casting defects. This study focuses on the ferrite–austenite transformation and the influence of the transformation on the solidifying shell deformation. In situ observations revealed that the austenite hardly nucleated and, consequently, the undercooled ferrite massively transformed to the austenite in the solid state. Simultaneously, the austenite rapidly solidified in the remaining liquid. The volume changes due to the ferrite–austenite transformation and the solidification shrinkage were measured using time-resolved tomography. On the basis of the observations and measurements, a model that includes the ferrite–austenite transformation and the austenite solidification was proposed to evaluate the volumetric strains in the solidifying shell. There was a sharp peak in the volumetric strains derived from the model at hypo-peritectic compositions. This work demonstrates the need to include the ferrite–austenite transformation in the solid state and the rapid austenite solidification for understanding the solidifying shell deformation in peritectic steels.

In this work, the Phillips PW 3020 X'Pert Diffractometer and PANalytical X'Pert Pro MPD X-ray Diffractometer were used to perform in-situ monitoring of quenched microstructures to define the defect evolution of Inconel 718 additive manufacturing powder. Dislocation density analysis based on X-ray Diffraction (XRD) measurement is sensitive to XRD peak broadening. The X-ray line profile is affected by the instrumental effects of the diffractometer, so quantifying and analyzing XRD instrument performance is important for accurate dislocation density analysis. The average measurement bias of the Phillips diffractometer was found to be -1.5490·10⁻³ Å while the precision was found to be 6.1000·10⁻⁵ Å. The average measurement bias of the PANalytical diffractometer was found to be -1.7633·10⁻⁴ Å while the precision was found to be 7.6917·10⁻⁵ Å. The average dislocation density calculated from the data was 3.93·10¹⁴ m⁻² for the smaller particle size range and 2.58·10¹⁴ m⁻² for the larger particle size range. The number of diffraction peaks utilized in dislocation density analysis was found to be a more significant factor than the instrumental differences. This work confirmed that differences in defect structure density can be observed across differently sized particles sourced from additive manufacturing powder.

The ESR and VAR processes are very often carried out under less-than-ideal conditions where the procedure involves inadvertent instabilities such as interruptions in metal flow or abrupt changes in the process heat balance. In this work we demonstrate the effects of typical examples of such instability as related to the remelting processes. It is suggested that a major direction in modelling development should be towards an understanding of the extent to which instability can be tolerated relative to the product requirements. Such an understanding would assist in the establishment of suitable quality standards governing the production of high-quality alloys.

A model is presented that predicts the amount and location of oxide inclusions in steel castings. A number and size distribution of inclusions, defined about a mean diameter, enters the casting system at its inlet during the filling process and are transported to their final locations in the casting. Model parameters for inclusion density, drag and wall friction are used to calculate the motion and locations of the oxide particles. Model results are presented to study the effects of casting shape and surface orientation on the final inclusion locations and distributions within castings. These results are compared with inclusion tracking experiments where the geometry of the gating system and orientation of casting cope surfaces affect the final distribution of inclusions in the castings. Measured and simulated inclusion area percent coverage, inclusion count and mean diameter are compared for a range of modelling parameters and inclusion size distributions. The size and number distribution at the casting system inlet, and other model parameters, are determined which provide the best agreement between measured and simulated inclusion area, count, and size.

Previously, the authors have used a mixed columnar-equiaxed solidification model, successfully 'reproduced' the solidification benchmark experiments on the Sn-10wt.%Pb alloy under natural/forced convections (travelling magnetic stirring) as performed at SIMAP laboratory [Int. J. Heat Mass Transf. 85 (2015) 438-54]. The current contribution is to address the flow-effect on the remelting of settling/floating crystals during the mixed columnar-equiaxed solidification. The re-melting or growth is controlled by diffusion of solute in the liquid boundary layer. The diffusion length due to the flow-effect is modelled as a function of Schmidt and Reynolds numbers. The modelling results show that remelting rate of the floating/settling crystals, which originate from fragmentation and then brought to the superheated region by the forced flow, can be enhanced by the flow. In turn the released latent heat can reduce the temperature locally (even globally), hence to speed up the solidification of the columnar structure. Additionally, the solidification-migration-remelting of equiaxed grains present an important macrosegregation mechanism. By solidification of a crystal in the cold region it rejects solute, while by remelting of the crystal it dilutes the surrounding melt. These phenomena are found critical in many engineering castings with mixed columnar-equiaxed solidification.

Castings are predestined for the application of structural optimization, but to date, the integration of process simulation into structural optimization is limited due to high computational cost and is therefore often neglected at the beginning of the design process. This leads to the need for surrogate models, which allow a fast and simplified evaluation of design proposals during the optimization in order to improve the integration. This article introduces a novel approach that estimates the solidification time of randomly created geometries solely based on the casting geometry. The approach uses ray-tracing methods to calculate the distance function along preset directions. The estimated solidification time is calculated using a Spherical Convolutional Neural Network (CNN). The training data is obtained by several thousand solidification simulations using the optimization toolkit of a commercial casting simulation software combined with further data augmentation. The model is experimentally validated for five different geometries in the sand casting process.

Dendritic deformation induced by convection of thermal fluid is one of the factors leading to dendrite fragmentation and plays a crucial role in grain structure of alloy, which lacks in-depth understanding. In this paper, we simulate the flow-induced mechanical deformation of dendrites during the solidification of Al-4.5wt.%Cu alloy by combining the cellular automaton-finite volume method (CA-FVM) for the dendrite growth and the finite element method (FEM) for handling the dendritic deformation with the complex boundary conditions given by CA-FVM results. It shows that the dendritic deformation strongly depends on the flow velocities of melt and dendritic morphology. The dendrites can undergo visible bending above the critical flow velocities for dendrite yield (ranging from 0.023 m/s to 0.126 m/s as the inlet velocity increases) and the von Mises stress increases as the flow is enhanced during the growth process.

We have extended the existing two-dimensional rigid solid phase benchmark for binary substance with the solid phase motion in the present paper. Incompressible laminar Newtonian flow is assumed, and a standard mixture formulation is used for the mass, momentum, energy, and solute transport. A coherency solid motion model accounts for the free-floating grains, assuming that the solid velocity is proportional to the mixture velocity and the liquid fraction. The lever rule is used to describe the mass fractions of the phases. A two-dimensional benchmark is solved using the semi-implicit meshless diffuse approximate method with an adaptive subdomain upwinding strategy. The results of the meshless method are compared to the finite volume method results with a reasonable agreement. The new benchmark results show that the solid motion has an essential effect on the macrosegregation pattern.

Investment casting is a highly dynamic process during which multiple competing physical phenomena are at work. Those seeking to understand and simulate such processes computationally are confronted with a considerable task, balancing accuracy with efficiency. Approximations and models based on well-understood and documented fundamental physics are powerful tools in a modeller's arsenal. Driven by observed discrepancies between experimental thermocouple measurements and simulation predictions of casting temperatures, this work explores the additional alloy cooling mechanism of mould transparency to infrared radiation, targeting a new mathematical approximation applicable in such situations. Direct attenuation, scattering from coarse sand, sand distribution in the mould and material temperatures play a role in the extent of radiation transparency that must be considered. From this model, estimation of the additional cooling rate resulting from expected mould transparency can be determined and applied as a corrective measure to computation fluid dynamics (CFD) simulation results that do not capture this phenomenon.

This work presents measurements of the droplet temperature for the Gas Metal Arc Welding process. Four different aluminium electrodes were used for experiments performed under the same conditions. The approach used yielded higher droplet temperatures than were obtained in previous experiments with the same raw data because the energy balance approach differed. However, trends were similar, with a minimum value close to the transition current and a tendency to increase with the current. Anode fall voltages behaved similarly, increasing with the current. Differences in temperature were found between the four electrodes. This can be explained by the different alloying elements present on the electrodes, which evaporate and take energy out of the droplet.

In this work, a volume averaging technique based Eulerian multiphase flow CFD (Computational Fluid Dynamics) model is developed to simulate the semi-solid solidification of the cooling slope generated Al-15Mg₂Si-4.5Si composite slurry. The four phases considered here are liquid (l), primary solid (Mg₂Si), secondary solid (α-Al) and air (a). Whereas latent heat release during solidification is tackled via a temperature correction scheme. The effect of recalescence is also observed in the present solidification model. The simulation findings include shape and size of the α-Al and primary Mg₂Si grains, solid fraction, temperature, melt viscosity etc. The developed multiphase flow model of cooling slope casting (semi solid solidification) is applied to a test case having melt pouring temperature (into the slope) of 650°C (923K). The simulation findings are then validated with the help of measurements of microstructural features observed in the experimental (optical) micrographs of the solidified samples.

The production and manufacture of aluminium castings typically involves multiple heat treatment process steps. Simulation tools are very useful to help predict the evolution of mechanical properties, to improve process parameters and as a guideline for casting design. In this study, a kinetic model, based on Nucleation and Growth Theory (CNGT), was developed and implemented for modelling heat treatment in a casting process simulation tool. Starting from information generated from the simulation of solidification and cooling of the casting, simulation was carried out to predict further precipitation behaviour of multiple Cu or Mg-Si -containing phases during solution treatment, quenching, and artificial ageing processes in heat treatment of Al alloys. Model validation was carried out within the research cluster "Advance Metals and Processes" in Aachen, Germany. Heat treatments were performed on two different Al-Si-Cu-Mg alloys with different quench methods and at different ageing temperatures. Experimental investigations including atom probe, wavelength dispersive X-ray spectroscopy, and tensile tests were conducted to calibrate and validate the model. The simulations showed good agreement for the studied conditions and a promising modelling approach for integrated simulation of casting and heat treatment for industrial applications.

Ti-Al alloys have replaced Ni-based superalloys in the last stages of some aircraft engines to improve fuel efficiency. In order to improve their properties, grain refinement has been investigated via isomorphic inoculation with Ti-Al-Nb particles. This inoculation method is orders of magnitude more efficient on a particle-by-particle basis than traditional inoculation, rather than multiple inoculant particles added to form a solidified bulk phase grain, in isomorphic inoculation each particle added results in the formation of multiple grains. As the particles are indistinguishable from the matrix after solidification, a model was used to elucidate this mechanism. Two phenomena were considered to calculate the number of particles acting during solidification: particle breakup along grain boundaries and complete particle dissolution. The grain size of the particles was calculated with an empirical model from initial TKD analysis of the particles and high temperature molten salt heat treatments. Particle dissolution was estimated via mass transport of the slowest diffusing Nb species. This showed the population of isomorphic inoculant particles which can act during solidification is near a 1:1 ratio with the number of grains formed, confirming the mechanism of grain refinement by direct epitaxial growth from the particles.

The impact of structural mechanics is often overlooked when modelling the solidification of dendritic microstructures, despite experimental observations that the interaction between these processes can be a factor leading to the development of crystal mosaicity throughout the microstructure which can itself lead to more serious defects. When considered at all, the structural mechanical behaviour of columnar dendrites is often considered as being analogous to a cantilever beam both in interpretations of experimental results and in existing numerical modelling. While this is not an unreasonable assumption when considering a dendrite in isolation, this is a scenario that infrequently occurs. In this paper a parametric study is presented using a Cellular Automata solidification solver coupled to a Finite Volume Structural Mechanics solver. These results highlight the complex non-linear behaviour that arises when considering dendrite interaction, demonstrating the significantly different microstructures that can be obtained by varying only the force experienced by the system.

A cellular automata-lattice Boltzmann coupled model was established to simulate the grain formation process of aluminum alloy under convection conditions. The influence of flow on crystal grain growth was simulated. Results show that the existence of convection prevents the crystal from growing symmetrically compared with the pure diffusion condition. Moreover, the growth of dendrite at the upstream side is promoted, whereas that at the downstream side is inhibited because of solute segregation. The greater the flow velocity is, the stronger the asymmetry of the grain is. The greater the undercooling is, the greater the dendritic growth driving force is, the faster the dendritic growth is, the dendrite arms coarsen, and the secondary dendrite arms become increasingly developed, thereby increasing the solute concentration at the solid-liquid interface at the tip of the dendrite and aggravating the solute segregation.

The performance in hydro-electric turbine casting and repair requires understanding of how process parameters and chemistry selection affect solidification microstructures. The aim of this study is to provide a quantitative phase-field formulation for process-microstructure relationships that seeks to model stainless steels. We have developed a phase-field model to simulate austenitic stainless steel solidification under experimental thermal histories. To this end we look at a pseudo-binary approximations for numerical efficiency. The pseudo-binary formulation is underpinned by the alloying element equivalent value, a metallurgical tool used to analyze the microstructural impact of "minor" alloying elements in stainless steels. For model validation we develop thin wall casting experiments to measure the thermal history and chemistry controlled microstructure. The models incorporate a thermodynamic parameterization and are linked to a thermal-phase transformation model which represents the experimentally measured thermal history. The results display a good agreement with the primary branch spacing and cellular to dendritic transition of the casting experiments. These models and software provide the basis for future expansion to include more complex microstructures.

To provide quantitative predictions, multiscale models of dendritic solidification (e.g., GEM, DNN, CAFE) need to be validated and require model parameters, which can be calculated by phase-field simulations. We report on a multiscale modeling of dendritic solidification in samples that are cooled homogeneously at a constant rate. We consider three Al-Cu alloys and samples from thin to bulk thickness. We investigate how the alloy composition, the distance between the equiaxed dendrites and the sample thickness influence the transient growth velocity of the primary tips. Using 3D phase-field simulations, we calculate the tip selection parameter based on the microsolvability theory. We show that the selection parameter depends principally on the ratio between the sample thickness and the smallest tip diffusion length during the transient growth (D/v_m, where v_m is the maximum tip velocity). The extracted tip selection parameters are then used as inputs for three-dimensional grain envelope model (GEM) simulations. The comparison between TIPF and GEM shows that the GEM can reproduce transient growth of interacting equiaxed dendrites during cooling and can account for sample confinement effects.

The dendrite solidification process has been observed and simulated using state-of-the-art techniques, such as time-resolved X-ray tomography (4D-CT) and high-performance phase-field (PF) simulations. 4D-CT has enabled the direct observation of the 3D dendrite growth in opaque alloys. However, the spatiotemporal resolution is not sufficient for investigating fast phenomena because a 3D solidification structure is obtained using hundreds of transmission images during the 180° rotation of a sample. High-performance PF simulations have enabled the simulation of multiple 3D dendrite growth phenomena. However, the material properties required in PF solutions of alloys are often unavailable. Therefore, integrating in situ X-ray observations with PF simulations using data assimilation is a promising approach for simultaneously solving these issues. In this study, we developed a data assimilation system with an ensemble Kalman filter, in which the solid fraction along the thickness of a sample was used as observation data to enable data assimilation using X-ray transmission images. The performance of the developed data assimilation system was evaluated via twin experiments for columnar dendrite growth during the directional solidification of a binary alloy in a thin film. The results showed that data assimilation using the solid fraction as observation data estimated the material properties and solidification morphologies with reasonable accuracy.

Recent solidification experiments identified an oscillatory growth instability during directional solidification of Ni-based superalloy CMSX4 under a given range of cooling rates. From a modeling perspective, the quantitative simulation of dendritic growth under convective conditions remains challenging, due to the multiple length scales involved. Using the dendritic needle network (DNN) model, coupled with an efficient Navier-Stokes solver, we reproduced the buoyancy-induced growth oscillations observed in CMSX4 directional solidification. These previous results have shown that, for a given alloy and temperature gradient, oscillations occur in a narrow range of cooling rates (or pulling velocity, V_p) and that the selected primary dendrite arm spacing (Λ) plays a crucial role in the activation of the flow leading to oscillations. Here, we show that the oscillatory behavior may be generalized to other binary alloys within an appropriate range of (V_p,Λ) by reproducing it for an Al-4at.%Cu alloy. We perform a mapping of oscillatory states as a function of V_p and Λ, and identify the regions of occurrence of different behaviors (e.g., sustained or damped oscillations) and their effect on the oscillation characteristics. Our results suggest a minimum of V_p for the occurrence of oscillations and confirm the correlation between the oscillation type (namely: damped, sustained, or noisy) with the ratio of average fluid velocity $\bar{V}$ over V_p. We describe the different observed growth regimes and highlight similarities and contrasts with our previous results for a CMSX4 alloy.

During cooling of steels and cast irons, austenite can decompose by a eutectoid transformation into pearlite, a two-phased mixture of ferrite and cementite. Since the internal lamellar structure is commonly too fine to be distinguished on the scale of the austenite grain structure, pearlite is often modelled as an effective, pseudo-single phase. Such a pragmatic treatment would also be desirable to reduce the computational effort of large-scale multi-phase-field simulations, but a fundamental hindrance is that no consistent thermodynamic description exists for effective pearlite in multicomponent databases. Alternatively, we here propose to model pearlite as diffuse mixture of two phases with individual local fractions and concentrations, such that solute partitioning and thermodynamic driving forces can be consistently derived from standard Calphad databases. The essential computational advantage is that only the outer interfaces of the pearlite nodules have to be numerically resolved, which allows for increased grid spacing and time-steps. The impact of the unresolved lamellar structure on the curvature undercooling is modelled analytically based on a characteristic spacing, which may be calibrated either experimentally or by small-scale simulations. The potential and the limitations of the new approach, implemented in the frame of the Micress^® software, shall be discussed.

Ductile cast iron, also known as nodular cast iron, is a graphite-rich cast iron with high impact and fatigue resistance, due to its nodular graphite inclusions. Ductile cast iron is produced by incorporating additives (often FeSi alloys) to the iron base metal at different production steps to obtain the desired graphite shape. A crucial step is the addition of Magnesium to promote the spheroidization of the graphite. The most common method is by adding crushed and sized Ferro-Silicon-Magnesium (FSM). The alloy composition, microstructure, and sizing are assumed to affect the key parameters of this reaction, namely, reactivity, recovery, and slag formation. Therefore, the study of the solidification of FSM is important to understand and predict its performance at the foundries. The present work aims at understanding and predicting numerically the formation of the major phases during the solidification process. Two approaches have been used: thermodynamic calculations through Thermo-Calc solver and phase field modelling using MICRESS. The models have been calibrated by comparison with advanced statistical characterization of the microstructure. The results indicate a competitive growth of the major phases and transformation of phases in solid state that can be emulated by the model.

Understanding the motion and growth behaviors of equiaxed dendrites during solidification is important for predicting macrosegregation. In this study, we develop a phase-field lattice Boltzmann (PF-LB) simulation method for the settling and growth of an equiaxed dendrite during the nonisothermal solidification of a binary alloy. The PF-LB computations are accelerated by employing parallel computation using multiple graphic processing units (GPUs) and the octree block-structured adaptive mesh refinement method, which incorporates multiple mesh and time increment methods. By using the developed method, we can simulate the three-dimensional long-distance settling dendrite while considering the effects of latent heat release and natural convection. From the simulation results, we confirm that the natural convection due to the high solute concentration around a dendrite reduces the settling velocity. In addition, we observe that the temperature increase owing to latent heat release slows dendrite growth, which in turn slightly slows the settling velocity. From these results, we confirm that the effects of latent heat release and natural convection are not negligible in the quantitative evaluation of settling dendrites.

Grain growth during the solidification of an alloy is accompanied by the enrichment of the surrounding liquid phase in chemical species. The resulting diffusion fields induce interactions between the grains that are strongly dependent on their spatial arrangement. The influence of these diffusive interactions is not limited to the shapes and sizes of grains but also controls the overall kinetics of the transformation. In this study we investigate the role of the spatial arrangement of equiaxed grains in the average growth kinetics of an ensemble of grains. We perform full-field simulations of different grain ensembles using the mesoscopic grain envelope model (GEM). We show that the growth kinetics of randomly arranged ensembles is fundamentally different from that of periodic arrangements. We provide an explanation of this difference through analyses of the behaviour of individual grains in their local neighbourhood, defined by a Voronoi tessellation. Finally, we compare the full-field GEM simulations to state-of-the-art mean-field models and we point out the limitations of the latter in the prediction of dendritic growth.

Aluminum alloys commonly contain Si as an impurity or alloying element. The energetic behavior of Si within multiple compounds and solutions is incorporated inside thermochemical packages, such as FactSage. This tool allows determining the Si partitioning within complex multiphasic systems. Recent experimental research suggests that Si can be found within Al₃Zr-based intermetallics. Nevertheless, current FactSage databases do not consider the potential substitution of Si within the Al₃Zr-D0₂₃ solid solution. In this work, Si substitution within the (Al,Si)₃Zr-D0₂₃ phase was investigated by means of first-principles calculations. Replacement of Al atoms by Si resulted in a negative enthalpy of mixing, indicating that Si substitution is energetically enabled. The density of states (DOS) for both a Si-diluted (Al,Si)₃Zr and a non-Si-doped (Al₃Zr) simulation cells were analyzed. It is shown that (even in dilution), Si significantly impacts the electronic structure of the Al₃Zr-D0₂₃ structure. Specifically, the presence of Si localizes electrons in the p orbital of Al, and increases the DOS of the d_xy, d_xz, and d_yz sub-orbitals of Zr at low energies. Thus, yielding a coupled effect that stabilizes the D0₂₃ intermetallic. These findings are a benchmark for the future integration of a Si-based end-member within the Al₃Zr-D0₂₃ solid solution of FactSage databases.

The effect of natural convection on dendrite morphology is investigated through three-dimensional large-scale phase-field lattice Boltzmann simulations using a block-structured adaptive mesh refinement scheme with the mother-leaf method in a parallel-GPU environment. The simulations confirmed that downward buoyancy enhances the growth of the primary and secondary arms, and upward buoyancy delays the growth of those arms. In addition, the effect of natural convection on the solidification morphologies gradually decreased as the primary arm tip reached the top of the computational domain and finally stopped. Furthermore, in the longer simulation under purely isothermal diffusive conditions, detachment of the secondary arms owing to curvature-driven fragmentation was observed. A large-scale non-isothermal dendrite growth simulation was also conducted, wherein it was observed that the tip growth rate of the primary arm was delayed, and the secondary arm spacing was larger than that in the isothermal condition.

We have developed a 2-D numerical meshless adaptive approach for phase-field modelling of dendritic solidification. The quadtree-based approach decomposes the computational domain into quadtree sub-domains of different sizes. The algorithm generates uniformly-distributed computational nodes in each quadtree sub-domain. We apply the meshless radial basis function generated finite difference method and the forward Euler scheme to discretise governing equations in each computational node. The fixed ratio between the characteristic size and the node spacing of a quadtree sub-domain ensures space adaptivity. The adaptive time-stepping accelerates the calculations further. In the framework of previous research studies, we used the approach to solve quantitative phase-field models for single dendrite growth in pure melts and dilute binary alloys. In the present study, we upgrade the solution procedure for the modelling growth of multiple differently oriented dendrites. Along with the space-time adaptive approach, we apply non-linear preconditioning of the phase-field equation to increase computational efficiency. We investigate a novel numerical approach's accuracy and computational efficiency by simulating the equiaxed dendrite growth from a dilute binary alloy.

The present study discusses the development of a two-dimensional phase field model of semi solid microstructure evolution of the novel Al-15Mg₂Si-4.5Si composite during cooling slope rheoprocessing. The superheated melt of the said composite undergoes partial solidification during its shear flow over the slope free surface, prior to get collected within the isothermal slurry holding furnace. Seed undercooling based nucleation model is adopted in the present PF model to simulate slurry microstructure of the composite. Experimentally determined cooling rate values are considered to simulate slurry microstructure for different process conditions. Micrographs of melt samples collected during experimentation establishes the developed PF model as an accurate microstructure prediction tool for the composite slurry. Apart from morphological features such as size and degree of sphericity values of the evolving primary Mg₂Si and α-Al grains, the model predictions includes melt temperature as well as solid content at any given instant of slurry processing.

Freckles, a significant issue encountered during the directional solidification of superalloys, are recognised by a trail of equiaxed grains parallel to the direction of gravity accompanied by local eutectic enrichment. In the current study, a mixed-columnar-equiaxed multiphase volume-average solidification model was employed to study the formation of freckles in superalloy casting. Fragments produced via flow-driven and capillary-driven fragmentation mechanisms are considered as the source of spurious grains. The transport and the growth/remelting of the fragments are considered. According to the simulation results, some segregation channels develop at the corners of the casting. Flow-driven fragments are produced in/around the segregation channels, whereas capillary-driven fragments are produced at a certain depth of the mushy zone across the entire section of the casting. The fragmentation rate caused by the flow-driven mechanism is several orders of magnitude larger than that caused by the capillary-driven mechanism, i.e. the flow-driven fragmentation mechanism is dominant for the currently investigated sample. After the solidification process, four freckles formed at the casting corners on the shadowed side, whereas it was freckle-free on the bright side.

Multi-phase field modelling using MICRESS® has been used to perform simulations of the nucleation and evolution microstructure in a 2D domain as function of fundamental thermo-physical parameters, time, cooling rate and chemical composition during solidification. In order to have a better understanding of primary precipitates evolution during solidification in a case hardening steel and the precipites chosen were niobium-carbides (NbC) and manganese sulphides (MnS), simulations were done including the nucleation and growth of precipitates and coupled through thermodynamic databases. The results show a good comparison with the experimental microstructure measured with SEM-EDX measurements. The simulated time dependent segregation profiles of Nb and Mn in the interdendritic regions, and the precipitates composition, have a good similarity with the measurements. Even when the size and morphology of the NbC and MnS particles has a difference in the both simulated and experimental situations, this can be explained by the low resolution used in the simulation. This publication shows the modelling results and the comparison with the experimental measurements.

Mesosegregation appears during the solidification of low-alloyed steels at the scale of a few grains. It causes chemically segregated bands in forgings manufactured from steel ingots. As this may affect the mechanical properties of the produced parts, it is crucial to understand the formation of such segregation patterns. While both microsegregation and macrosegregation are well understood and can be readily characterized, little is known about the formation of mesosegregation. In this paper we present a data analysis method to identify the characteristic scales and patterns of mesosegregation. Segregation was mapped on centimetric samples, using micro X-Ray fluorescence (µXRF). The fine sampling grid used reveals segregation patterns at different scales. To identify and distinguish the mesoscale segregation patterns from smaller and larger present patterns, a numerical data analysis technique based on spatial filtering was used. It can identify characteristic scales of mesosegregation patterns on 2D serial cut samples, leading to 3D reconstruction. This approach combined with simulation studies, will ultimately pave the way to a comprehensive understanding of the formation of mesosegregation.

Laser Powder Bed Fusion (L-PBF) is seen as a process of interest by aeronautical industry to develop new engine components. Nevertheless, the reliability and durability of parts produced by L-PBF depend on the possibility to suppress the occurrence of defects. Among them, hot cracking represents a key issue. These cracks are due to the liquid film remaining between grains at the end of the solidification stage combined with stresses and strains endured by the mushy domain. A microsegregation model providing relevant prediction of the solidification path during L-PBF is thus required for coupling with a thermomechanical analysis. As an answer to the industrial need, a new model is proposed and applied in cooling conditions encountered in L-PBF. It includes the initial solidification conditions and follows the phases, and their composition in the interdendritic liquid region to predict the brittle temperature range. Both dendrite tip growth model and kinetic phase diagram due to non-equilibrium interface phenomena are considered. Cross-diffusion of solute species in the liquid phase is accounted for, as well as thermodynamic coupling with CALPHAD. The model will be applied to IN718, a nickel-based superalloy widely used in the aeronautic industry.

In the search of an adequate real time strain measurement method in aluminum casting, the use of Fiber-Bragg-Grating (FBG) is being investigated with great interest. In order to do so, the behaviour of glass fiber sensors in a liquid aluminium alloy at temperatures up to 750°C is experimentally analysed in a laboratory environment. For better process understanding a simulation of the fiber alloy composite is conducted. FBG is an optical measurement method, which uses engraved Bragg reflectors in a 125 µm in diameter thick glass fiber. This reflector transmits most of the wavelengths but only reflects one specific wavelength. This specific wavelength can be measured and changes due to the axial strain on the grating by the fluid alloy reaction and by the changes in temperature. Using a so-called mirror furnace, several experiments with the fiber alloy composite are evaluated. These measurements are also the basis for the further understanding of hot tearing. The data gathered during the measurement campaign - both numerical and experimental - is used to parameterize a simulation. As a result, the understanding of the fiber alloy composite behaviour is expanded and a digital twin is modeled with MATLAB's partial differential equation toolbox.

Three dimensional models of dendritic structures during solidification are valuable for building physical models, validating simulated results, estimating some properties such as permeability in the mushy, simulating semisolid deformation and so on. Thus, it is of interest to observe microstructure evolution in situ. Time-resolved tomography combined with X-ray diffraction has allowed us to observe the evolution of dendritic structures and to measure crystallographic orientation in situ. Reconstruction still proves to be difficult for some alloy systems because of the tradeoff between time and spatial resolution. This paper demonstrates the reconstruction of dendritic structures for three different alloy systems (Al-10mass%Cu alloy with a diameter of 4 mm, CrMnFeCoNi alloy with 1 mm, and Zn-4mass%Al alloy with 0.7 mm). The observations were performed in a synchrotron radiation facility SPring-8. A filter using a phase field model was introduced to reconstruct the three-dimensional images. Parameters used in the filtering were consistently determined based on the raw reconstruction images. Evaluation of solid-liquid interface area and curvature was significantly improved by the filter. For the Al-Cu alloy, a three-dimensional model containing approximately 300 million voxels was obtained. For the CrMnFeCoNi alloys, the preferred growth direction <100> was confirmed by tomography and X-ray diffraction. For the Zn-Al alloy, the observed 14 growth directions were not simply defined by the crystallographic orientations, although the directions were consistent with the hexagonal symmetry. This study verifies that time resolved tomography, X-ray diffraction and the filter using a phase field model provide three dimensional models for light metal alloys with rather large diameters and 3d transition-metal alloys with rather large X-ray absorption coefficients. The models are expected to be used for further studies.

Hot tears are detrimental defects forming during the final stage of solidification when the remaining liquid loses the capacity to compensate for liquid to solid volume shrinkage. Although a mature semi-quantitative description of hot tearing has been developed, little is known about the dynamic evolution of hot tears as experimental studies have been conducted mostly post-solidification or in semi-static in-situ conditions. Here, we present a methodology to investigate the evolution of hot tears with high spatial and temporal resolution using synchrotron-based X-ray radiography. We develop a novel hot tear detection and tracking algorithm for quantification of hot tear density, area fraction and merging from the analysis of radiographic sequences of the solidification of thin metal samples. The methodology is demonstrated for an Al-5wt%Cu alloy and examples of the results and new insights that can be achieved are described.

Pores have been the main focus of quality assurance in castings. Latest research has shown that in aluminium castings pores can form only if there is existing entrainment damage, i.e., pores are merely the visible parts of the entrainment damage, and usually invisible damage is much more extensive. However, its effect on deformation behaviour has not been previously established or observed in-situ. This work applies 2D Digital Image Correlation (DIC) to an in-situ full-field stress-strain analysis of tensile samples with a non-conventional heterogeneous stress distribution. The observations reveal that the effect of hidden damage extends far beyond its impact on fracture behaviour and is responsible for initiating local strain concentrations during deformation. By extracting local stress-strain data, FE simulations have been performed to mimic the effect of local hidden damage on the heterogeneous stress-strain field. SEM and FIB-SEM analysis has been applied to investigate the cause for the strain concentrations. The combined results show that hidden damage in the form of oxide films is not only responsible for premature fracture, but also affects the deformation behaviour of tensile samples by introducing dispersed strain concentrations.

In this work, an integrated simulation approach previously developed for static FE analyses is extended to microstructure- and defect-based fatigue life assessments of castings. The approach, the closed chain of simulations for cast components, combines casting process simulation with microstructure modelling and local material characterisation to generate heterogeneous material data for FE analysis and fatigue life assessment. The method is demonstrated on a High-Pressure Die Cast aluminium component. Areas with a high risk of defects are identified based on the simulated solidification conditions, and heterogeneous material data for the fatigue life analysis is generated. Fatigue testing has been performed with different levels of porosities to quantify the effect of defects on the element-specific Wöhler curves. Pore characteristics are assessed using 2D X-ray, fracture surface analysis and Kitagawa diagram. The results highlight the importance of taking the risk of defect formation into consideration when designing industrial aluminium castings subjected to fatigue loads.

The main defects in heavy steel castings are related to hot tear formation during solidification. Depending on the steel grade, design, and local solidification conditions, it is possible to predict regions with higher risk of hot tear formation during the casting process. However, steels containing Boron show more complex crack and defect patterns compared to common steel casting alloys. The mechanisms behind the Boron induced hot tearing is investigated in this work to understand the influence of Boron enrichment during solidification and the influence on hot tearing. The experimental work includes the investigation of phase diagrams and the corresponding fractions of the solid and liquid phases depending on temperature using thermal analysis e.g. DSC and HT-LSCM. The hot tearing sensitivity and mechanical properties during solidification are obtained in the Submerged Split Chill Test, SSCT. In addition IMC-B 3-point bending tests are performed to determine high-temperature material properties in the solid state. The work is part of a research project where the final goal is to improve the hot tear predictions based on experimental work and carry out a benchmark simulation of a real sized casting and use it to show the agreement between the numerical results and extensive non-destructive testing from industrial observations.

Inorganic bonded sand cores are getting increasing attention in industry due to their environmental advantages, and they are now widely used for series production in automotive applications. The advantages of these binder systems are nevertheless associated with a greater sensitivity of the core quality in relation to the manufacturing and storage process before casting, which is also having a major impact on the core strength. This paper presents the current work in integrating the modelling of the binder decomposition and evaporation of binder water with the mechanical performance of partially hardened and dried sand cores. This allows a local description of mechanical properties, considering different behaviour in compression and tension which depends on temperature and different levels of moisture content through the core. Recent work has shown the importance of including creep effects in the numerical modelling of 3D Printed (3DP) inorganic sand cores. Measurements of the mechanical behaviour of 3DP cores compared to shot cores have shown some fundamental differences in strength due to different density and binder content. This has been investigated in detail in collaboration with a major automotive manufacturer and the application of the model extension was done on a complex water jacket core.

This paper presents a coupled model of heat transfer and plastic deformation in friction stir welding (FSW), accounting for the temperature profile in the substrate near the pin. This approach is analogous to the boundary layer analysis in fluid mechanics and is based on the methodology of scaling and calibration based on published data. A model focusing on common conditions in FSW, such as relatively slow translation and high rotation velocities, a thin shear layer and the influence of the shoulder on the maximum temperature was reformulated. This paper extends previous work by considering the heat flow into the pin and an improved criterion for determining the temperature at the edge of the shear layer. The results are a set of updated closed-form expressions for the maximum temperature, the thickness of the shear layer, the shear stress around the pin, torque and thermal effect of the shoulder, applicable to all metals. The predictions from this model are verified against a comprehensive database of published experiments. Applications of this model also include the accelerated determination of procedure variables and the generalization of maps of process limits.