A model of direct-drive friction welding has been developed, which can be used to predict the time-temperature histories, the resultant microstructure, and the microhardness distribution across the weld interface of direct-drive friction-welded AISI/SAE 1045 steel bars. Experimentally measured power and axial displacement data were used in conjunction with a finite-element transient thermal model to predict the time-temperature history within the heat-affected zone (HAZ) of the weld. This was then used with a microstructure evolution model to predict the volume fraction of the subsequent microconstituents and the microhardness distribution across the weld interface of welds produced using three significantly different welding conditions: one with optimal conditions, one with a long burn-off time, and one with high axial pressure and rotational speed but short burn-off time. There was generally good agreement between the predicted and the measured time-temperature histories, volume fraction of the resultant microstructures, and microhardness distribution in the HAZ of AISI/SAE 1045 steel friction welds produced using these three significantly different welding conditions.

There is an industry-wide desire to produce aluminum alloy tailor welded blanks for automotive applications to facilitate light-weighting of automotive body structures; however, the ductility and formability of blanks made using conventional fusion welding processes have not yet been acceptable due to the effects of internal and surface weld defects. The objective of this study was to identify techniques that might be used to reduce weld defects and improve ductility and formability in autogenous fusion welded AA5000 series aluminum tailor welded blanks. The double-sided arc welding (DSAW) process was used to produce autogenous welds between 1.0- and 1.5-mm-thick AA5182-O aluminum sheets. While visually acceptable conduction-mode DSAW welds could be made using a range of powers and welding speeds, weld metal. ductility and formability were not acceptable due to the presence of internal weld defects including hydrogen porosity, solidification microporosity, and solidification porosity and cracking at the weld centerline. A comprehensive set of techniques was developed that minimized the occurrence of these weld defects. The resulting defect-free fusion welds were found to have significantly improved ductility and formability with final fracture occurring in the thin sheet of the tailor welded blank and not in the fusion weld.

A diode laser welding/brazing technique was used for lap joining of 5754 aluminum alloy to DP 980 steel with Al-Si filler metal. The correlation between joint interfacial microstructure, wettability of filler metal, and mechanical properties was systematically investigated. At low laser power (1.4 kW), a layer of intermetallic compounds, composed of θ-Fe(Al,Si)3 and τ 5 -Al7.2Fe1.8Si, was observed at the interface between fusion zone and steel. Because of the poor wettability of filler metal on the steel substrate, the joint strength was very low and the joint failed at the FZ/steel interface. When medium laser power (2.0 kW) was applied, the wettability of filler metal was enhanced, which improved the joint strength and led to FZ failure. With further increase of laser power to 2.6 kW, apart from θ and τ 5, a new hard and brittle η-Fe2(Al,Si)5 IMC with microcracks was generated at the FZ/steel interface. The formation of η significantly degraded the joint strength. The failure mode changed back to interfacial failure.

To improve casting and product yields of Fusion™ cast AA3003-core/AA4045-clad laminate ingots, a fundamental understanding of the wetting and interface formation process between the oxidized AA3003 cast surface and the AA4045 melt is required. In this study, a laboratory scale analog/immersion test was developed which mimics the wetting and interface formation process that occurs during Fusion™ casting. The effects of reheating and remelting of the AA3003 cast surface, the degree of surface oxidation present during initial contact of the two alloys, and the atmospheric conditions (air or argon) on interface formation when dipped into an AA4045 melt were examined. Results indicated that in an argon atmosphere, wetting and dissolution of the solid, oxidized AA3003 samples by liquid AA4045 occurred at temperatures both above and below the measured onset of AA3003 remelting. AA3003/AA4045 interfaces were similar to those produced during Fusion™ casting, with a thin layer of AA4045 forming an oxide-free, metallurgical interface to the AA3003 after immersion. The AA3003 surface oxides were not an impediment to wetting and bond formation. Mg surface segregation was observed on the oxidized AA3003 surfaces and may play a role to help penetrate the oxide layer. For tests conducted in air, wetting of the sample by AA4045 liquid was generally poor regardless of temperature.

Fusion™ casting is a unique Direct Chill continuous casting process whereby two different alloys can be cast simultaneously, producing a laminated ingot for rolling into clad sheet metal such as AA3003/AA4045 brazing sheet. Better understanding of the wetting and interface formation process during Fusion™ casting is required to further improve process yields and also explore use of other alloy systems for new applications. In this research, AA3003-core/AA4045-clad ingots were cast using a well-instrumented lab-scale Fusion™ casting system. As-cast Fusion™ interfaces were examined metallurgically and by mechanical testing. Computational fluid dynamic analyses of the FusionTM casts were also performed. It was shown that the liquid AA4045-clad alloy was able to successfully wet and create an oxide-free, metallurgical, and mechanically sound interface with the lightly oxidized AA3003-core shell material. Based on the results of this study, it is proposed that the bond formation process at the alloys interface during casting is a result of discrete penetration of AA4045 liquid at defects in the preexisting AA3003 oxide, dissolution of underlying AA3003 by liquid AA4045, and subsequent bridging between penetration sites. Spot exudation on the AA3003 chill cast surface due to remelting and inverse segregation may also improve the wetting and bonding process.

The brazeability of AZ31B-H24 magnesium alloy sheet to Sn-coated plain carbon steel sheet using a Mg-Al-Zn alloy filler metal and a diode laser heat source has been investigated. While the Sn coating promoted good wetting between the molten filler metal and the steel sheet, it did not play a role in forming the final metallic bond. Its primary function appeared to be in maintaining an oxide-free steel surface until the molten Mg filler metal could come in direct contact with the steel surface. In all cases, failure of transverse tensile shear test specimens of the joint occurred in the steel base metal. Metallic bonding between the magnesium alloy and the steel was facilitated by the formation of two transition layers, including a Fe(AI) solid solution formed on the surface of the steel and a nano-scale layer of Al-8(Mn,Fe)(5) phase on the Fe(AI) surface layer. Examination of the Fe(Al)-Al-8(Mn,Fe)(5) and Al-8(Mn,Fe)(5)-Mg interfaces using HR-TEM showed that orientation relationships (OR) with a low angle of rotation of the matching planes and low interplanar mismatch and therefore low interfacial energy density existed at the Fe(Al) Al-8(Mn,Fe)(5) interface (i.e., when [10 (1) over bar1](Al8Mn5)//[(1) over bar 11]Fe(Al), {110}(Fe(Al)) was 4.2 deg from {30 (3) over bar3}(Al8Mn5) with 5.2% interplanar mismatch) and a high angle of rotation of the matching planes and large interplanar mismatch, and therefore, high interfacial energy density existed at the Al8Mn5-Mg interface (i.e., when [10 (1) over bar1](Al8Mn5)//[10 (1) over bar0]Mg, {30 (3) over bar3}(Al8Mn5) was within 47.4 deg of the {0002}(Mg) with 16.8% interplanar mismatch). These results were further validated using an edge-to-edge crystallographic matching model of the Fe(Al)-Al-8(Mn,Fe)(5) and Al-8(Mn,Fe)(5)-Mg interfaces.

Laboratory-scale experiments were conducted to cast AA3003/AA4045 clad ingots via Fusion™ Technology, a novel process developed by Novelis Inc. for the production of aluminum clad materials such as brazing sheet. Experimental results were used to validate a steady-state thermofluids model of the Fusion™ Technology co-casting process. The numerical model was able to accurately predict the temperature field within the AA3003/AA4045 clad ingot as well as the shape of the AA3003 liquid sump. The model was also used to quantify the temperature, fraction solid, and velocity fields in a clad ingot cast with an asymmetrical molten metal-feeding system. Feeding of core and clad molten metals at opposite corners of the mold was found to reduce the risks of hot spots and liquid metal breakthrough from the core sump to the clad side of the Fusion™ Technology mold. The use of a diffuser for the AA3003 core molten metal and of a vertical feeding tube for the AA4045 clad produced different flow patterns and liquid sump shapes on either side of the mold. The quality of the metallurgical bond at the core/clad interface appeared good near the clad inlet and at the ingot centerline, but poor near the edges of the ingot. SEM–EDS analysis of the chemical composition across the interface showed that a 1 to 20-μm-deep penetration of silicon from the AA4045 clad into the AA3003 core had occurred at visually acceptable interfaces, whereas silicon diffusion across poor interfaces was very limited. A study of the model-predicted fraction solid history at different points along the interface indicated that reheating of the AA3003 core is not required to form a visually acceptable metallurgical bond. However, a sufficient amount of interaction time between the solid AA3003 core shell and the silicon-rich AA4045 clad liquid is required to chemically dissolve the surface of the core and form a good metallurgical bond. An approximate dissolution depth of 750 to 1000 μm was observed along the visually good interface. Partial dissolution of the Mn-rich AA3003 core led to the formation of Al(Mn,Fe)Si intermetallic particles in the AA4045 clad and an increased manganese concentration near the core/clad interface.

In this study, wetting has been characterized by measuring the contact angles of AZ92 Mg alloy on Ni-electroplated steel as a function of temperature. Reactions between molten Mg and Ni led to a contact angle of about 86 deg in the temperature range of 891 K to 1023 K (618 °C to 750 °C) (denoted as Mode I) and a dramatic decrease to about 46 deg in the temperature range of 1097 K to 1293 K (824 °C to 1020 °C) (denoted as Mode II). Scanning and transmission electron microscopy (SEM and TEM) indicated that AlNi + Mg2Ni reaction products were produced between Mg and steel (Mg-AlNi-Mg2Ni-Ni-Fe) in Mode I, and just AlNi between Mg and steel (Mg-AlNi-Fe) in Mode II. From high resolution TEM analysis, the measured interplanar mismatches for different formed interfaces in Modes I and II were \( 17{\kern 1pt} \;{\text{pct}}_{{\{ 10\overline 11\}_{\text{Mg}} //\{ 110\}_{\text{AlNi}} }} \)-\( 104.3\;{\text{pct}}_{{\{ 110\}_{\text{AlNi}} //\left\{ {10\overline{1}0} \right\}_{{{\text{Mg}}_{ 2} {\text{Ni}}}} }} \)-\( 114\,{\text{pct}}_{{\left\{ {0003} \right\}_{{{\text{Mg}}_{ 2} {\text{Ni}}}} //\{ 111\}_{\text{Ni}} }} \) and \( 18\,{\text{pct}}_{{\{ 10\overline 11\}_{\text{Mg}} //\{ 110\}_{\text{AlNi}} }} \)-\( 5\,{\text{pct}}_{{\left\{ {110} \right\}_{\text{AlNi}} //\{ 110\}_{\text{Fe}} }} \), respectively. An edge-to-edge crystallographic model analysis confirmed that Mg2Ni produced larger lattice mismatching between interfaces with calculated minimum interplanar mismatches of \( 16.4\,{\text{pct}}_{{{\text{\{ 10}}\overline 1 1 {\text{\} }}_{\text{Mg}} / / {\text{\{ 110\} }}_{\text{AlNi}} }} \)-\( 108.3\,{\text{pct}}_{{{\text{\{ 110\} }}_{\text{AlNi}} / / {\text{\{ 10}}\overline 1 1 {\text{\} }}_{{{\text{Mg}}_{ 2} {\text{Ni}}}} }} \)-\( 17.2\,{\text{pct}}_{{{\text{\{ 10}}\overline 1 1 {\text{\} }}_{{{\text{Mg}}_{ 2} {\text{Ni}}}} / / {\text{\{ 100\} }}_{\text{Ni}} }} \) for Mode I and \( 16.4\,{\text{pct}}_{{{\text{\{ 10}}\overline1 1 {\text{\} }}_{\text{Mg}} / / {\text{\{ 110\} }}_{\text{AlNi}} }} \)-\( 0.6\,{\text{pct}}_{{{\text{\{ 111\} }}_{\text{AlNi}} / / {\text{\{ 111\} }}_{\text{Fe}} }} \) for Mode II. Therefore, it is suggested that the poor wettability in Mode I was caused by the existence of Mg2Ni since AlNi was the immediate layer contacting molten Mg in both Modes I and II, and the presence of Mg2Ni increases the interfacial strain energy of the system. This study has clearly demonstrated that the lattice mismatching at the interfaces between reaction product(s) and substrate, which are not in direct contact with the liquid, can greatly influence the wetting of the liquid.

The brazeability of AZ31B-H24 magnesium alloy and steel sheet with a microlayer of electro-deposited Ni in a single flare bevel lap joint configuration has been investigated. The macro- and microstructure, element distribution, and interfacial phases of the joints were studied by optical microscopy (OM), scanning electron microscopy (SEM), transmission electron microscopy (TEM), and X-ray diffraction (XRD). Wetting of the steel by the Mg-Al brazing alloy was improved significantly through the addition of a Ni electroplated interlayer. Bonding between the magnesium brazing alloy and the steel was facilitated by the formation of a transition layer composed of a solid solution of Ni in Fe on the steel followed by a layer of alpha-Mg + Mg2Ni eutectic. A band of AlNi intermetallic compound with different morphologies also formed along the steel-fusion zone interface, but was not directly responsible for bonding. Ni electroplating was found to significantly improve the brazeability and mechanical performance of the joint. The average fracture shear strength of the bond reached 96.8 MPa and the joint efficiency was 60% with respect to the AZ31B-H24 Mg alloy base metal. In all cases, failure occurred in the fusion zone very close to the steel-fusion zone interface.

For the last 70 years, direct-chill (DC) casting has been the mainstay of the aluminum industry for the production of monolithic sheet and extruded products. Traditionally, clad aluminum sheet products have been made from separate core and clad DC cast ingots by an expensive roll-bonding process; however, in 2005, Novelis unveiled an innovative variant of the DC casting process called the Fusion™ Technology process that allows the production of multialloy ingots that can be rolled directly into laminated or clad sheet products. Of paramount importance for the successful commercialization of this new technology is a scientific and quantitative understanding of the Fusion™ casting process that will facilitate process optimization and aid in the future development of casting methodology for different alloy combinations and ingot and clad dimensions. In the current study, a numerical steady-state thermofluids model of the Fusion™ Technology casting process was developed and used to simulate the casting of rectangular bimetallic ingots made from the typical brazing sheet combination of AA3003 core clad with an AA4045 aluminum alloy. The analysis is followed by a parametric study of the process. The influence of casting speed and chill-bar height on the steady-state thermal field within the ingot is investigated. According to the criteria developed with the thermofluids model, the AA3003/AA4045 combination of aluminum alloys can be cast successfully with casting speeds up to 2.4 mm s−1. The quality of the metallurgical bond between the core and the clad is decreased for low casting speeds and chill-bar heights >35 mm. These results can be used as a guideline for improving the productivity of the Fusion™ Technology process.

The thermodynamic stability of precipitated phases at the steel-Ni-Mg alloy interface during laser brazing of Ni-plated steel to AZ31B magnesium sheet using AZ92 magnesium alloy filler wire has been evaluated using FactSage thermochemical software. Assuming local chemical equilibrium at the interface, the chemical activity–temperature–composition relationships of intermetallic compounds that might form in the steel-Ni interlayer-AZ92 magnesium alloy system in the temperature range of 873 K to 1373 K (600 °C to 1100 °C) were estimated using the Equilib module of FactSage. The results provided better understanding of the phases that might form at the interface of the dissimilar metal joints during the laser brazing process. The addition of a Ni interlayer between the steel and the Mg brazing alloy was predicted to result in the formation of the AlNi, Mg2Ni, and Al3Ni2 intermetallic compounds at the interface, depending on the local maximum temperature. This was confirmed experimentally by laser brazing of Ni electro-plated steel to AZ31B-H24 magnesium alloy using AZ92 magnesium alloy filler wire. As predicted, the formation of just AlNi and Mg2Ni from a monotectic and eutectic reaction, respectively, was observed near the interface.

The weldability of coated sheet steels by Nd:YAG lasers has been examined using a 250-W pulsed laser, a 1-kW pulsed laser and a 2-kW continuous wave (CW) laser. Seam welds were produced in 0.75-mm-thick (23gauge) galvanized and galvannealed sheet steels using a modified lap-joint configuration consisting of a groove-shaped projection in the top sheet of the joint. Experiments were performed to assess the effects on weld quality of coating type, groove-projection dimensions and laser process parameters. Good quality welds, which failed at levels comparable to the base metal in shear tensile tests, were made over a wide range of conditions using the CW Nd:YAG laser, but could only be produced using a limited range of process conditions with the two pulsed lasers. Using the CW laser at a mean power of 1600 W, good quality welds could be produced at speeds up to 60 mm/s (144 in./min). In comparison, the maximum acceptable welding speed when using the 250-W pulsed laser at 220 W laser mean power was just 2.4 mm/s (5.4 in./min). The melting ratios of the welds produced with the CW laser were found to be about 0.25, while they were approximately 0.10 for the welds produced using the pulsed lasers. Finally, the quality of welds produced was not affected by the dimensions of the groove projection or by the coating type. M. R GRAHAM, D. C. WECKMAN, and H. W. KERR, are with the Department of Mechanical Engineering, University of Waterloo, Waterloo, Ontario, Canada. D. M. HIRAK is with Imperial Oi l Resources. Ltd, Calgary, Alberta, Canada. Introduction The automotive industry uses zinc-rich coated sheet steels extensively in auto body components for enhanced corrosion resistance. Laser beam welding is being evaluated as an alternative joining technique for these sheet steels in the lap-joint configuration, because it offers a number of advantages over traditional resistance spot welding practices. For example, much less flange material is required for laser welding, resulting in a potentially significant weight reduction. As well, no direct contact is required between the welding machine and the workpiece and access is only required from one side of the weld joint. In addition, lasers, particularly Nd:YAG lasers whose output can be transmitted through fiber-optic cables, can be easily integrated into automotive robotic welding cells. Laser beam welding of zinc-coated sheet steel in the lap-joint configuration is illustrated schematically in Fig. 1A. The well-known problem with welding these materials in this configuration is related to the low boiling point of zinc

Novelis Inc. recently developed and patented a unique Direct Chill (DC) casting process whereby two different alloys can be cast simultaneously, producing a laminated aluminum ingot, known as Fusion™1 casting. In this process, a high quality metallurgical bond is formed between the two simultaneously cast alloys; the mechanism of how this bond formation occurs between the two alloys is not clearly understood. Since direct observation of wetting and interface formation in the Fusion casting process is difficult to obtain in practice, due to temperature and proximity issues, an analog test was designed to mimic the processes on the laboratory scale. The apparatus provides a means of controlling the basic variables thought to influence bond formation in Fusion casting, namely: i) the relative temperatures of the two alloys in contact, ii) the sample contact time, and iii) surface oxidation. The results of two test series will be presented, highlighting the relative importance these variables have on as-cast interface quality.

The brazeability of AZ31B-H24 magnesium alloy and steel sheet using two different types of micro-interlayers (Al-12Si and Ni) in a single flare bevel lap joint configuration has been investigated. The macro- and microstructure, element distribution, and interfacial phases of the joints were studied by optical microscopy (OM), scanning electron microscopy (SEM), transmission electron microscopy (TEM), and X-ray diffraction (XRD). The results of this study suggest that the feasibility of this process depends strongly on the pre-existing Al-12Si coating or Ni electro-plating layer on the steel sheet. Both interlayers promoted wetting of the Mg-Al filler alloy. Use of the Al-12Si interlayer resulted in the formation of a brittle layer of theta-Fe(Al,Mg)(3) intermetallic compound along the braze/steel interface, which was responsible for bonding. In comparison, using a Ni interlayer led to formation of a band of AlNi intermetallic compound with different morphologies along the steel-fusion zone interface. However, this phase was not directly responsible for bonding. Bonding between the magnesium brazing alloy and the steel was facilitated by the formation of a transition layer composed of a solid solution of Ni in Fe at the steel interface. The average fracture shear strength of the bond using the Al-12Si and Ni interlayers reached 49.3 MPa and 96.8 MPa, respectively, which gave joint efficiencies of 31% and 60% with respect to the AZ31B-H24 Mg alloy base metal. Thus, the Ni interlayer resulted in joints with significantly improved joint strength relative to those made using the Al-12Si interlayer.

Prized for its excellent strength-toweight ratio, magnesium and its alloys are currently under intense investigation for use in many applications in the automotive and aerospace industries (Refs. 1–3). However, steel sheet is still the most commonly used material in the automotive industry for fabrication of autobody structures. The ability to make hybrid structures of magnesium alloy and steel sheet would be desirable for many applications in the automotive industry, because the overall weight of the autobody could be reduced resulting in better fuel efficiencies and lower environmental impact. Therefore, there is increasing interest in identifying and developing new techniques and processes that can be used to make dissimilar joints between magnesium alloys and steel sheet (Refs. 3–9). Joining magnesium alloys to steel by conventional fusion welding technologies is difficult due to the large difference in the melting points between Mg (649°C) and Fe (1538°C). In addition, the boiling point of magnesium is only 1091°C, so direct contact with molten steel causes catastrophic vaporization of the magnesium (Refs. 3–8). Moreover, the maximum solid solubility of Fe in Mg is estimated to be only 0.00041 at.-% Fe (Ref. 6) and wetting of steel by molten magnesium is very poor (Ref. 8). The weldability of magnesium to steel using the hybrid laser-arc welding (Refs. 5, 6, 8, 9) and resistance spot welding (Refs. 4, 7) processes has been examined. Zhao et al. (Ref. 5) used a hybrid laser-gas tungsten arc welding (GTAW) process to join AZ31B magnesium alloy and 304 stainless steel. However, oxides that formed at the interface were found to cause joints with poor tensile strength. Using the same welding technique, Liu et al. (Refs. 8, 9) studied lap joining of AZ31B Mg alloy to Q235 steel with Sn and Cu interlayers. Mg2Sn and Mg2Cu intermetallic compounds were found to form along the grain boundaries of the Mg alloy when using the Sn and Cu interlayers, respectively. The use of Sn and Cu interlayers was reported as the main reason for the elimination of gaps along the steel-fusion zone interface and the improvement of wetting properties of the steel by molten magnesium alloy (Refs. 8, 9). Finally, in a recent study, Liu et al. (Refs. 4, 7) used resistance spot welding to join AZ31B magnesium alloy to DP600 Zn-coated steel. They found that a preexisting transition layer of Fe2Al5 between the Zn coating and the steel improved wetting and bonding between the steel and the magnesium alloy. Review of the literature suggests that joining Mg alloys to steel will be possible provided the temperatures required for joining are kept below the boiling point of the magnesium alloy (1091°C) and provided another interlayer element is used that can interact and promote wetting and bonding between both immiscible alloys. For this reason, brazing can be a superior choice in joining dissimilar metals such as magnesium and steel because brazing SUPPLEMENT TO THE WELDING JOURNAL, JANUARY 2013 Sponsored by the American Welding Society and the Welding Research Council

The heat transfer coefficient (HTC) associated with primary cooling during direct-chill casting of AA6111 aluminum alloy was investigated by conducting casting experiments with an instrumented mould. Mould temperature measurements obtained at various positions were used as input in an inverse heat conduction analysis in order to calculate the heat flux between the ingot and the mould. Heat transfer coefficient profiles versus vertical position within the mould were obtained for casting speeds of 1.39 and 1.83 mm/s. Relatively low HTC values of 93 to 576 W/m(2).K were attributed to the formation of an air gap between the ingot surface and the mould. The experimental heat transfer coefficients were compared to theoretical values predicted using a 3-dimensional CFD model of direct-chill casting with a simple one-dimensional density-based model of shell deformation, which calculates the air gap thickness and the HTC within the mould.

An accurate thermofluids model of aluminum direct-chill (DC) casting must solve the heat-transfer equations in the ingot with realistic external boundary conditions. These boundary conditions are typically separated into two zones: primary cooling, which occurs inside the water-cooled mold, and secondary cooling, where a film of water contacts the ingot surface directly. Here, a simple model for the primary cooling boundary condition of the steady-state DC casting process was developed. First, the water-cooled mold was modeled using a commercial computational fluid dynamics (CFD) package, and its effective heat-transfer coefficient was determined. To predict the air-gap formation between the ingot and mold and to predict its effect on the primary cooling, a simple density-based shrinkage model of the solidifying shell was developed and compared with a more complex three-dimensional (3-D) thermoelastic model. DC casting simulations using these two models were performed for AA3003 and AA4045 aluminum alloys at two different casting speeds. A series of experiments was also performed using a laboratory-scale rectangular DC caster to measure the thermal history and sump shape of the DC cast ingots. Comparisons between the simulations and experimental results suggested that both models provide good agreement for the liquid sump profiles and the temperature distributions within the ingot. The density-based shrinkage model, however, is significantly easier to implement in a CFD code and is more computationally efficient.

Thermal modeling of the direct-chill casting process requires accurate knowledge of (1) the different boundary conditions in the primary mold and secondary direct water-spray cooling regimes and (2) their variability with respect to process parameters. In this study, heat transfer in the primary cooling zone was investigated by using temperature measurements made with subsurface thermocouples in the mold as input to an inverse heat conduction algorithm. Laboratory-scale experiments were performed to investigate the primary cooling of AA3003 and AA4045 aluminum alloy ingots cast at speeds ranging between 1.58 and 2.10 mm/s. The average heat flux values were calculated for the steady-state phase of the casting process, and an effective heat-transfer coefficient for the global primary cooling process was derived that included convection at the mold surfaces and conduction through the mold wall. Effective heat-transfer coefficients were evaluated at different points along the mold height and compared with values from a previously derived computational fluid dynamics model of the direct-chill casting process that were based on predictions of the air gap thickness between the mold and ingot. The current experimental results closely matched the values previously predicted by the air gap models. The effective heat-transfer coefficient for primary cooling was also found to increase slightly with the casting speed and was higher near the mold top (up to 824 W/m2·K) where the molten aluminum first comes in contact with the mold than near the bottom (as low as 242 W/m2·K) where an air gap forms between the ingot and mold because of thermal contraction of the ingot. These results are consistent with previous studies.

A diode laser brazing procedure has been developed for joining AZ31B-H24 Mg alloy sheet to aluminum-coated, cold-rolled carbon steel sheet in the single flare bevel lap joint configuration using a Mg-Al based welding wire. In this process, the Mg-Al based filler metal and a shallow surface layer of the Mg alloy sheet were melted simultaneously by a diode laser beam, while no melting of the steel sheet occurred. The results of this study suggest that feasibility of this process depends strongly on the pre-existing Al-12Si coating layer on the steel sheet that promotes wetting of the Mg-Al filler alloy as well as formation of a layer of theta-Fe(Al,Si)(3) intermetallic compound along the braze/steel interface. From the middle part of the braze/steel interface to the root of the joint, the Al-Si layer melted and mixed into the braze alloy and the intermetallic layer grew up to 8 mu m thick. From the middle part of the braze/steel interface to the top of the joint, both the Al-Si and the intermetallic layer were dissolved. These two simultaneous phenomena led to an intermetallic layer with nonuniform thickness ranging from 0 to 8 mu m along the braze/steel interface. The average fracture load of the joint was 767 N, representing a 72% joint efficiency relative to the steel sheet. Failure occurred when cracks propagated along the intermetallic layer starting at the root of the bevel joint and moved into the braze metal at the upper part of the joint.