Eaton Cuts Costs with Minitab

Achieving Operational Excellence Through Statistical Analysis:

How three Eaton facilities used Six Sigma and Minitab Statistical Software to cut costs

Eaton's Galesburg, MI, Truck Components Operations Headquarters

The "Power of One" is a philosophy that is embraced by Eaton. They believe that one project, one division, or one person can make a difference, not only in their local work environment, but also across the entire company. According to Eaton, the way to achieve operational excellence is to build upon success by sharing "lessons learned," effective management tools, and best practices. They value and encourage the transfer of information and ideas throughout the company's many divisions and operating units.

s article illustrates how a project that started at one Eaton facility turned into three projects that ultimately impacted several divisions and the bottom line.

Kings Mountain, North Carolina

"It started with a challenge: find ways to increase the output at our truck components plant in Kings Mountain, NC, without any capital investment," said Patrick Watts, senior manufacturing engineer and Certified Six Sigma Black Belt at Eaton. "The plant was producing parts at maximum capacity and our outsourcing options were expensive."

While looking for ways to improve processes and efficiency, Watts saw an opportunity to reduce the number of changeovers and decrease time spent setting up new machines between production runs. The plant used two similar components; both parts were steel gears produced in Eaton's forging operation in South Bend, IN. The principle difference between the two parts was the presence of holes pierced into the web of one of the gears during the forging process. Research within Eaton determined that piercing the holes had been added to the forging operation years before to facilitate the flow of oil during the quenching process during heat treatment. The oil quench process controls the cooling rate and hardens the metal.

Watts explained the project idea saying, "Each time we switched between these two similar parts during production, we'd have to reconfigure the setup of the machines. If we could consolidate two different forgings into one common forging, then we'd reduce the number of setups and only have one common part number to keep in inventory."

Team members from the forging operation determined that they would also benefit from the elimination of the gear with pierced holes. The piercing/ punching operation was a frequent source of breakdown and the cost of production was increased due to tooling and maintenance costs. Removing the piercing step from the production process would allow them to ship parts more quickly, and tool and die life would be extended.

Watts and a project team set out to compare the two gears to decide if they could stop producing the gear without holes. They focused on a particular gear geometry feature known as "lead range" as a measure of the gear's distortion during quenching. Minitab Statistical Software was used to statistically compare the two gears and determine that there was no difference in the average lead range error or any difference in the variation of the lead range data between the two forgings.

Watts explained the role of Minitab statistical software in the project saying, "Minitab has been used on virtually all Six Sigma projects and any projects that require either data analysis or data comparison. For this project, we used it to analyze data taken directly from our heat-treatment operation. The information we obtained provided us with a definitive direction in which to move. Our decisions were not based on anecdotal information or opinions."

After analyzing the data, the team concluded that the holes could be eliminated. The project prompted Eaton to make obsolete one of the forgings and utilize a common part number, which reduced the gear forging costs by approximately 10 percent.

John O'Neill, plant manager and Six Sigma Project Champion from the South Bend forging operation said, "Anytime we can eliminate waste, we save money. When we stopped producing the gears with holes, we decreased waste in two ways: we removed the need for making and replacing the pierce tooling used to make the holes and we also eliminated the down-time of the press cell caused by changing our tooling or adjusting tooling. Minitab was very useful to us during the project. The software saved us time and created telling graphics that helped us to make decisions."

The South Bend team suggested to managers and design engineers at the Truck Components Operations Headquarters in Galesburg, MI, that there were other parts produced at the forging operation they should consider for consolidation into one part number. Management decided to have Shelbyville's Six Sigma Black Belt conduct a follow-up to the team's suggestion.

Shelbyville, Tennessee

Kevin Colby, Certified Six Sigma Black Belt, was charged with identifying opportunities to apply the knowledge gained from the project in Kings Mountain to Eaton's Heavy Duty Truck Transmission plant in Shelbyville, TN. Colby decided to focus on gears that were components used in manual seven-speed truck transmissions, a high volume part at Shelbyville. As with the project done in Kings Mountain, there were two similar steel gears in inventory, one with holes and one without holes.

Colby had to do some preliminary work to verify that the project met Eaton's Six Sigma project requirements. At Eaton, all Six Sigma projects are subjected to a rating system to quantify projects before they are approved.

Several questions are considered, including: Is the proposed project important to customers?
Will there be a cost savings?
Is the time until completion six months or less?
Does the project use Six Sigma tools?
Is there a project team?
Does the project have a high probability of success?

Once Colby proved that the project met the criteria for Six Sigma projects, a project team was selected. The team included members from both the Eaton Forge in South Bend and the plant in Shelbyville, and included plant operators, supervisors, a materials manager, engineers, quality inspectors in the Eaton Gear Lab, and a Six Sigma Champion.

The Analysis At Shelbyville

The team had to prove that there was no difference in part performance between gears with and without holes. Twenty-five samples of each of the gears were machined for analysis. The samples were measured in the Eaton Gear Lab prior to heat treatment (green) and after the heat treatment operation. The difference from the green values and the post heat treatment values were used for the final analysis.

As in the previous studies at Kings Mountain, the lead range (the variation between gear teeth caused by warping created during the heat treatment) was selected as the key variable to analyze. This variable was considered to be an indirect measurement of the possible distortion caused during the heat treatment process. Lead range analysis was performed on the left (drive) and the right (coast) sides of the gear teeth.

Summary of Minitab Tools Used:

• Basic Statistics of Variables Data
• Test for Stability, Individuals and Moving Range Charts
• Test for Normality
• Boxplots
• Comparison of Means, Two Sample t-Test
• Comparison of Variation, Test for Equal Variances

The statistical analysis assumed no difference in the gears unless proven otherwise. An alpha level of 0.01 was used for all significance tests. Four basic analyses were conducted: Stability, Normality, Comparison of Variation, and Comparison of Means (averages).

Stability: Minitab's Individuals (I) and Moving Range (MR) charts were used to illustrate the stability of the Right and Left Lead Ranges on both gears. Individuals and Moving Range charts allow you to track the process level and variation at the same time. The values for the Right Lead Range fell within the control chart limits indicating stability (Figure 1). One value fell outside the control limits of the Left Lead Range indicating fair stability (Figure 2). The team investigated the value and did not find a definitive reason why the value was out of control, except that the operator may have insufficiently cleaned the machine and part prior to machining the gear. They thought this explanation made sense and decided to continue with the analysis.

Normality: The normality of the Left and Right Lead Range sample data for both gears was tested. Results showed there was no evidence of a systematic deviation from normality for either gear, indicating that the data roughly followed a normal distribution.

Comparison of Variation: A test for equal variances was conducted to determine if the lead range variation was equal for gears with and without holes. No evidence of significant difference in variation was found.

Comparison of Means: A two-sample t-test was performed to determine if the lead range means were equal for gears with and without holes. Results showed no evidence that the Lead Range means were statistically different.

Based on the statistical analyses, the team recommended that the gears should be produced without holes and that forging for the gears with holes be eliminated. This change provided approximately a 10 percent reduction in the gear forging costs.

Figure 2: One value fell outside the control limits of the Left Lead Range, indicating fair stability.

Automated Products

While Kevin Colby was working on the project in Shelbyville, the Truck Components Automated Products Division was examining cost savings opportunities. The division produced transmissions that included speed sensors, which measure shaft speeds and work in conjunction with the gears produced in Shelbyville. The gears with holes caused signal fluctuations that impacted the sensors. Two electronic control units (ECUs) with different circuit speeds were manufactured to allow the sensor to work with both types of gears.

Engineers within the division's design group who were involved with the Shelbyville gear project realized they could simultaneously impact two divisions (Shelbyville and the Automated Products Division). Jerry Ganski, principal engineer, who led the effort to eliminate the second ECU said, "We realized that removal of the holes in the gears would allow the Automated Products Division to eliminate the special ECU we had to manufacture to deal with the holes. We now use a common ECU for all our platforms and thus save the money it took to build, stock, and handle two ECU styles where the only difference between them was the speed sensor circuit. The savings is estimated at approximately 12 percent." Based on the improvements realized from these three projects, Eaton is investigating other gear-related projects for potential improvement opportunities.

Approximate Percent Cost Reduction Summary:

• 10 percent cost reduction for a high volume gear manufactured in Kings Mountain, NC, and forged in South Bend, IN.
• 10 percent cost reduction for a family of gear forgings manufactured in Shelbyville, TN, and forged in South Bend, IN.
• 12 percent cost reduction for electrical control units (ECUs) used in certain seven-speed transmissions by the Automated Products Division, Galesburg, MI.

This article was originally published in the July 2002 issue of Scientific Computing and Instrumentation).