This is fundamental work with strong connections to industrial processes for materials processing and manufacturing. This work aims at applying fluid mechanics and heat transfer concepts for better understanding, analysis, and design of manufacturing processes such as spray deposition, strip casting, laser surface melting, and melt-spinning. (Photos of some of the setups can be seen on my web page.)

The earlier phases of the work focused on the melt-spinning process through simple numerical modelling and some high-speed videography experimental work which enabled us to visualize the solidification puddle and its dynamic variations in-situ during the process. Further numerical modelling quantified the important effect of heat transfer in the substrate, which had usually been overlooked, as well as the effect of various process parameters including in particular the thermal contact at the interface, again usually left out. We were also able to introduce undercooling in our numerical models, which plays an important role through recalescence but complicates greatly the boundary conditions at the interface. Improved boundary layer models were later also developed for this process, again including undercooling effects. These and other models allowed us to conduct parametric studies covering wide ranges of all the main parameters controlling the planar flow casting process, greatly improving our understanding of the thermal aspects of the problem.

Another related solidification process is splat cooling, which we modelled numerically using improved numerical techniques based on control volume integrals and taking undercooling into account to predict the solidification patterns more accurately. Some experimental work was also conducted during which we were able to use high-speed videography to quantify the impact of a molten metal drop on a substrate, and to compare the results to those of numerical models.

When developing numerical models for solidification, it is important that they be able to function for alloys as well as pure metals. The introduction of mass transfer does, of course, make the models more complicated, and that is particularly true when non-equilibrium and kinetics effects are introduced. We have nevertheless been able to develop interface-tracking models which are able to take the undercooling or superheating into account and predict the related variations in solute concentration. When applied to processes such as laser surface melting, for example, these models predicted complex concentrations profiles and interface velocity variations. The development of advanced numerical models including non-equilibrium kinetics also allowed us to predict the solute-rich cores that have been observed after rapid solidification.

One very important parameter, yet little understood, is the thermal contact coefficient at the interface between the melt and substrate, and we have undertaken an extensive study to measure it and to quantify its effect on the processes. Some experiments were conducted, which showed that this coefficient can vary greatly during the solidification process, unless the deposit becomes strongly bonded to the substrate. In the latter case, and if the substrate is brittle, there can be substrate spallation, and in an another experimental study we have quantified this spallation effect by combining acoustic measurements, high-speed videography, and numerical heat transfer modeling. Further experiments were conducted to measure the interfacial contract coefficient as a function of a number of other parameters such as melt superheat, substrate properties, substrate finish etc. In the course of these measurements we had to develop techniques to measure the emissivity of high-temperature molten metals undergoing rapid solidification, for which there was very little quantitative data available. An extensive review of values measured for this parameter also revealed that some simple correlations between the contact coefficient and the substrate speed or strip thickness hold remarkably well over surprisingly wide ranges of parameters covering a number of processes of importance to the industry, some very useful information for design or analysis of these processes. We have also developed some devices that allow us to quantify directly the solid/liquid interface thermal behavior, and we have begun to develop analytical models based on interface micro-geometry that aim at predicting the thermal contact variations at the solid / liquid interface. We believe that a better understanding of the thermal contact problem in general is essential for optimum operation of the manufacturing processes.