Experimental Design and Statistical Analysis of the Quality of Aluminum 380.1 Melt


Contents:
Executive Summary
Objectives of the Experiment
Background Information
Background Information Regarding Melt Quality
Research Design and Definitions


Executive Summary

The purpose of this experimental design was to determine which sets of conditions offered the best quality of aluminum melt used within company's foundry operations. Quality was defined in terms of specific gravity (percentage). Higher percentages yield parts of better quality. Lower percentages of specific gravity are revealed in the die cast part by bubbles (porosity) within the solidified metal. A part failure may result if excessive porosity is located where a portion of the part is subjected to stresses.

Three main areas (factors) were used in this experimental design: Temperature (High 1300F or Low 1200F), Fluxing (fluxed or unfluxed samples), and Location (Breakdown Furnace or Holding Pot). These three factors at two levels each generated eight experimental test conditions. This experimental design is referred to as a Factorial design.



Objectives of the Experiment

1. Develop an understanding of the physical sampling and testing of the metal and repeatability of the experiment.
2. Understand the effects of temperature on hydrogen content and melt cleanliness.
3. Measure the effects of the transfer and charging of the metal on the hydrogen content and melt cleanliness.
4. Quantify the effects of fluxing on metal quality.
5. Establish guidelines for the handling and treatment of metal that will maximize metal quality and minimize handling costs.

In general, the outcomes of the experiment will provide a range of standards that the company can compare to later for evaluation of metal on a consistent basis by this test procedure. In addition, the experiment will also provide guidelines for comparing metal quality from the company's existing breakdown furnace with the quality of metal melted in alternative furnaces.



Result of the Experiment

Great care was taken to insure consistency in sampling, measurement and calculations. Data reduction and analysis was performed using the MINITAB statistical software package. Based on the results of the designed experiment and accompanying statistical analysis, the following results can be offered:

1. Overall, the best quality of aluminum melt comes directly from the breakdown furnace regardless of temperature setting and the use of drossing flux.

2. The melt charging operations significantly decreases the specific gravity of the melt. The levels of specific gravity at the machine's holding pot vary widely. The turbulence created by the charging operation is probably the assignable cause.

3. The use of flux has the most impact at the machine's holding pot location. The reasons for this are not known for certain. However, speculation provides some plausible reasons. For example, the impurities within the melt are better revealed to the drossing flux after the holding pot has been charged.

4. Overall, temperature ranges have a significant impact on the levels of specific gravity.

5. This experiment and protocol can be replicated at any time.

The outcome of this experiment suggests that improvements may be needed in the area of melt charging operations. Improvement should focus around the minimization of turbulence during these operations. Additionally, efforts should be made to clear away impurities built up around the machine's holding pot, as well as the transfer ladle used to transport the melt from the breakdown furnace to the machine.



Background Information

The company is a major supplier of parts and assemblies to the pneumatic equipment industry.

Die cast parts and assemblies are made from aluminum or zinc. No alloys were used.

Sodium Fluoride Nitrate is used as a drossing flux to promote separation of the aluminum oxide (Al2O3) dross layer that forms on the surface of the melt.

Aluminum melt is obtained by melting 25 lb aluminum ingots at one end of the breakdown furnace and aluminum scrap, runners, overflow and rejected parts at the other end.

Melting temperature of the breakdown furnace is kept high to quickly form the melt. This approach is used to keep the die-casting machines at higher output levels.



Background Information Regarding Melt Quality

The high temperatures associated with nonferrous (aluminum) molten metal processes cause hydrogen gas to become soluble within the melt. During the die-casting operation the metal cools causing the hydrogen to become insoluble. The insoluble hydrogen attaches to inclusions (impurities) within the melt, creating bubbles within the cast part. Typically, higher temperatures result in higher levels of soluble hydrogen. Higher levels of hydrogen result in an increase in the porosity (bubbles), thereby resulting in a decrease in the specific gravity (density) of the cast part.

The existing quality control process involves sampling die cast parts after they have cooled, then x-raying the part for internal porosity. If the excessive porosity is located at a critical area (e.g., threads, areas under pressure, etc.) the part is rejected, documented, and returned to the breakdown furnace to reclaim the aluminum.



Research Design and Definitions

Response Variable: Specific Gracity (expressed as a precentage)

Factors: Levels:
Temperature High (1300F) (+) Low (1200F) (-)
Drossing Flux Fluxed sample (+) Unfluxed sample (-)
Location Breakdown furnace (+) Holding pot (-)

Experimental Design chosen: 23 factroial design


Result of the Experiment

1. Type of application. This design is appropriate when several factors are to be investigated at two or more levels, and interaction of the factors may be important.

2. Structure. This design has a basic structure where several factors are investigated at several levels by running all combinations of factors and levels. Blocking is not required with this design.

3. Information sought. The purpose of this experiment is to estimate and compare effects of several factors; estimate possible interaction effects; and estimate variance.

4. Economic. The time, money, and effort required to perform the sampling and data acquisition doesn't restrict the scope of the experiment. Sampling is performed off-line.


The experimental combinations below reflect three factors with two levels each.

(+ + +) High (+ - +) High (+ + -) High (+ - -) High
Fluxed Unfluxed Fluxed Unfluxed
Breakdown Breakdown Holding Pot Holding Pot


(- + +) Low (- - +) Low (- + -) Low (- - -) Low
Fluxed Unfluxed Fluxed Unfluxed
Breakdown Breakdown Holding Pot Holding Pot


Home | Approach to Quality | Corporate Services | Experience
Expectations | Executive Profile | Newsletters | Contact

P.O. Box 374 Littleton, Colorado 80160-0374
303.798.6092 sdtaylor@taylorqualitygroup.com

Web Design by dave@wildandfreed.com
©2002 Taylor Quality Group, LLC