**ABSTRACT**

Powder metallurgy is a widely used process specializing in the creation of unique alloys and metal matrix composites (MMCs) that cast and wrought alloy processing techniques cannot achieve. Production of metal powders, however, can be cost prohibitive when considering small production runs.

Some techniques, such as gas atomization, are necessary for producing clean, spherical powders, but these pristine powders are not always required, depending on the post-processing technique that will be employed. For example, during cryomilling the pristine nature of the metal feedstock powders is destroyed from ball milling in a slurry of liquid nitrogen or argon.

Building on the work of previous research, we have designed a machine capable of generating inexpensive, irregularly shaped powder metals suitable for further processing. Design of the centrifuge was based off the findings from investigating two different analytical models with the results verified by computational fluid dynamics (CFD)for powder formation in the centrifuge chamber.

The analytical models agree within 3% of one another and within 11% of the CFD model. The machine has been designed and developed to be built for under $15,000 to meet the criteria given to be considered “inexpensive.”

**BACKGROUND**

**Powder Metallurgy Process:**

Powder metallurgy is a culmination of several production technologies that work together to create parts from powder metals. In the last few decades, the atomization of metal powders has taken a great interest due to the scalability of the process to create cost-effective parts from metal powders of variable size and properties. The range of metals and their alloys are not limiting by the methods of PM, and different materials like ceramics, other alloys, and organic fibers can be added in post-processing techniques to create metal matrix composites.

**Applications of Powder Metallurgy:**

After covering the basic principles of PM, there should be a discussion on the applications of PM in the industry and the economics behind decisions to use such. Around 80% of all known applications of powder metallurgy are in the automotive sector. That said, most of the automotive sector examples are related to transmission components like gears.

**TYPES OF ATOMIZATION**

Types of atomization processes that exist to create metal powders are discussed. There are four main types of atomization that occur. A brief overview of each process will be given along with the materials that best utilize each method. The four types of atomization that commonly occur today are water atomization, inert gas atomization, plasma atomization, and centrifugal atomization.

**Water Atomization:**

This process involves a molten melt dropping by its own accord through a jet of water that is directed downwards at an angle surrounding the entire melt area. This process forms small particles often around 50-2000 microns. Metal powders manufactured via this process , e.g. Fe, Cr, Mn, and Si, experience oxidation because the energy of formation of the metal oxide is higher than that of the water.

**Gas Atomization:**

Gas atomization is one of the more favorable methods for atomization of a wider variety of materials than water atomization. It also helps to reduce the number of oxides that would have formed in a wateratomization process. The main effort for gas atomization is to break up and freeze the metal stream with an inert gas. Two types of gas atomization can occur; those being closed-couple gas atomization or free-fall gas atomization.

**MODELING OF CENTRIFUGAL ATOMIZATION**

Arguably the most exciting part of this research is the modeling of the centrifugal atomization process. The complexity that is involved with the fluid dynamics and the different dynamic forces enacted upon the liquid metal that sheer it off into atomized particles make this of an interesting research topic. Beyond the modeling of the atomization process is creating the boundary conditions that must surround the rotational disk and melt. With these boundary conditions, the effort made to build a machine can be carried out.

**MODELING OF CENTRIFUGAL ATOMIZATION**

Arguably the most exciting part of this research is the modeling of the centrifugal atomization process. The complexity that is involved with the fluid dynamics and the different dynamic forces enacted upon the liquid metal that sheer it off into atomized particles make this of an interesting research topic. Beyond the modeling of the atomization process is creating the boundary conditions that must surround the rotational disk and melt. With these boundary conditions, the effort made to build a machine can be carried out.

**Alternate Method of Atomization Behavior by Zhao:**

To broaden the discussion on the atomizer disk, this section will focus on ways of eliminating or justifying some of the assumptions made in the previous section while introducing others. This alternate method is to offer evidence to suggest improvements the estimation of the melt stream and the trajectory.

**Methodology Behind Computational Fluid Dynamics:**

To further the study of melt flow from the nozzle to the atomization disk, there are programs like ANSYS and SOLIDWORKS Flow Simulation that offer advanced computing capacities to determine the characteristics of fluid flow in the atomization process. Many of the techniques that are at the disposal to researchers have been done for processes like gas atomization, but few for methods like centrifugal atomization. K. H. Ho and Y. Y. Zhao are some of the only researchers to have findings using simulation software for modeling the melt as it hits the atomization disk.

**EXPERIMENTAL CALCULATIONS**

**Setting Up Boundary Assumptions:**

This section will use all previously discussed equations in a logical order to describe the atomization process. There will be fully calculated solutions shown when needed along will all constants and assumptions were given.

However, drawn out or repetitive calculations will be provided in graph form using plots done in MATLAB to aid in a better discussion of results. The order in which the results will be presented are: (i) given assumptions of system, (ii) constants used for all calculations, (iii) simplified atomization calculations, (iv) refined atomization calculations, and (v) trends between both method.

**First Method of Modeling Atomization:**

The constants used for all future calculations. Most variables are variable with future estimates. Global values, where applicable, have been assigned on commercially pure aluminum, with the as sumption that these can be defined as close approximations for the application of this research to a physical machine.

**Second Method for Modeling Atomization:**

Now having to see the results from the first method the following will be a discussion on the results from the second method for calculating the particle formation properties. The other method first focused on defining the criteria for which powder can be formed. Using Equation 14 and Equation 15 the results show that for the desired range of particle size from 100-1000 microns that the minimum atomizer disk radius must be at least 6 inches and no greater than 2.4 inches.

**DESIGN OF THE CENTRIFUGAL ATOMIZER**

**Machine Design Principles:**

To fully develop the centrifuge as a system-a basic understanding of power, heat transfer, and design constraints will be discussed here for a reference to nomenclature and fundamental equations. The result will be a brief overview of the equations used to formulate the design limitations and the requirements to complete the build. The following section will be constructed as follows with a discussion on (i) basic principles of heat transfer, and (ii) designing around power constraints.

**Discussion on the Design:**

Now that all the initial conditions have been defined, here will be the overall design and the components that will go into the build of the atomizer. The initial project began with the atomizing chamber since this component of the project was a donated item, and everything else had to be sized to fit inside of the chamber.

**CONCLUSION**

The objective of this thesis was to analyze and develop a centrifugal atomizer for small runs creating powder metals. The results of this thesis found the following:

- A comprehensive look at two different analytical models that took different approaches to analyze the atomization technique, yielding results within 3% or fewer of each other.
- Validation using CFD to examine the cooling time from the velocity outputs of the two analytical approaches, yielding 11% differences or fewer.
- Using the results from the three methods, a model of the dontated centrifuge with constraints was created.
- The resulting MATLAB code can be used to construct the necessary conditions for other materials to be atomized beyond aluminum.
- The costs of the machine components make it a feasible exploratory option to manufacture powder metals on campus.

Future work will be to complete the machine and to conduct trial experiments to test the validity of the findings within this thesis.

Source: California State University

Author: Christopher Wayne Potts

>> Automobile based Matlab Project Topics with Free Base Papers Downloads