Computational fluid dynamic investigation of the primary and secondary atomization of the free-fall atomizer in electrode induction melting gas atomization process

Abstract

Nickel-based superalloy is mainly used for fabricating the important high temperature parts including the turbine disk, turbine baffle, compressor disk, and other critical components. Ceramic inclusions in powder metallurgy (PM) superalloy could promote fatigue crack initiation, and thus accelerating the crack propagation under certain conditions. In this case, the ultra-clean nickel-based superalloy powder is critical for PM superalloy components. Generally, there are two well-known methods of fabricating superalloy powders, i.e., argon gas atomization (AA) and plasma rotating electrode process (PREP). Electrode induction melting gas atomization (EIGA) process is a newly developed method of preparing ultra-clean metal powders. The EIGA process is a completely crucible-free melting and atomization process developed by ALD vacuum technologies. In this process, a slowly rotating prealloyed bar is fed into a conical induction coil. The end of the bar is inductively heated and molten alloys falls into an atomizer where the liquid alloy is atomized with a high-pressure inert gas. The EIGA prepared powders possess the advantages of AA (more fine powders) and PREP (ultra-clean powders) processes. Generally, there are two key issues in EIGA process, and the free-fall gas atomizer design is one of the critical issues for the powder yield and quality. Free-fall gas atomizers are some of the first two fluid atomizer designs to be used for molten metal atomization. In a simple open (unconfined stream) design a melt stream falls from a tundish exit via gravity into the convergence of focused atomization gas jets where it is disintegrated. The gas-melt interaction is complex, and it is difficult to characterize the interaction process directly. To have a good understanding of the atomisation technology, the physical break-up process instead of correlating the gas dynamics with droplet fragmentation indirectly must be able to be examined. And it will be desirable, if we input the atomization parameters, we can obtain the particles' distributions directly. In this work, a computational fluid dynamic approach to simulating the primary and secondary atomization processes is developed by using the volume of fluid method and discrete phase model. By integrating the metal stream break-up (in primary atomization) with the flow field and particles distribution simulation (in second atomization), this numerical simulation method is able to provide the direct assessment for the atomisation process. To verify the method performance, the melt stream is initialized into a 4 mm-diameter stream, which is then injected into the gas flow field for further fragmentation. The experimental results show that the simulated particles' diameter distribution is consistent with the experimental results in the same conditions.