Printer Friendly

6 sigma eases strain of converting tooling to nobake molding.

Inside This Story

* The Harrison Steel Castings Co. decided to convert its existing green sand molding line to nobake molding to produce a more consistent casting and eliminate the overhead costs of the green sand process.

* Detailed within is the Six Sigma program that Harrison used to complete the conversion of tooling from green sand to nobake in 18 months and increase total line savings by 14%.

In the late 1990s, The Harrison Steel Castings Co., Attica, Ind., came to the realization that in order to stay competitive, it needed to find ways to produce a more consistent casting while eliminating some of the overhead costs associated with the green sand molding process it was running.

In 2001, the firm concluded that replacing its existing green sand molding system with an automated molding line (AML) was the only way to drive down the costs of the steel castings it produces primarily for the construction/mining industry. The problem was that converting to nobake molding required all of the existing green sand patterns to be reengineered. A total of 196 patterns required conversion to nobake molding over an estimated three-year period in order to make the switch.

With the huge capital investment involved, Harrison needed to speed up the tooling conversion process in order to start producing castings with the new line. The goal was to cut the projected conversion time in half. At the same time, Harrison was beginning the process of implementing a Six Sigma program, which is a combination of several quality initiatives blended together with a clear problem-solving process that anchors all quality activities to dollars.

This article examines the Six Sigma steps of define, measure, analyze, implement and control that a team at Harrison used to convert its patterns and tooling production from green sand to nobake molding in 18 months.

Getting Started

Harrison's first step was to define what needed to be accomplished. With the goal of "reducing the tooling conversion time and/or double the rate of conversion in order to shut down the green sand molding system 18 months ahead of schedule," a map of the entire process for converting the patterns needed to be developed to establish the baseline data.

The define stage involved determining the Voice of the Customer (VOC), which asked who the customer is and who the beneficiary of the final product (a converted pattern) will be? The team determined Harrison had several internal customers and began talking to them to convert their "pains" into critical customer requirements (CCR) that included:

* improve the casting "exit rate" to more than 90%;

* increase pattern changeovers from 1/wk. to more than 2/wk.;

* decrease the time it takes to move the casting from shakeout through final inspection;

* more than double the rate at which the pattern shop was changing patterns over;

* improve the accuracy.

The voice of the business (VOB) also needed to be determined. This involved identifying what concerns the business has in the conversion process. These "pains" were converted to critical business requirements that included increasing the pattern conversions to move the target completion date forward by 18 months and maintaining an overall average of sand-to-net-metal ratio of 4.5:1.

The next stage was to establish the tasks involved to convert a pattern. A top-down charting method enabled the firm to list the major steps involved in the process and break them down into smaller steps. A schedule was then set up starting with the date in which the vice president of operations scheduled the conversion and ending when the vice president approved the pattern for production. The schedule was considered a "live" document as it was updated weekly to reflect possible customer order changes and to allow for more flexibility.

Functional deployment maps were then created to examine pattern shop processes and the preliminary design process, marking the completion of the define stage.

Moving Ahead

During the measurement stage, the VOCs were used to determine what needed to be measured because there were no previous metrics in place for many of the items required. From there, each "performance measure" was examined to form an "operational definition." Using the functional deployment maps and the CCRs, a measurement system was shaped and distributed among the team.

Each team member was assigned the task of gathering data in his/her area. Data was collected on all subsequent changeovers starting March 7, 2002 and was examined on April 19, 2002. The time frame allowed for nine patterns to make it completely through the process and be analyzed.

To start the analysis, all of the data was combined to produce a Pareto Chart (Fig. 1) to see what was contributing the most time to the overall conversion process. The chart showed that the cleaning room and pattern shop used up more than 80% of the time to convert a pattern. Following the Six Sigma rules, the focus quickly narrowed to the cleaning room process.

[FIGURE 1 OMITTED]

Historical data used by Harrison showed that it would take an average of 42.5 days to convert each pattern from sample to date of approval. It would then take an average of 14.5 days to work the sample casting through the cleaning room and quality control department (QC). By looking deeper at the cleaning room process, it was concluded that it took a maximum of 4 hrs. to clean the casting (not including heat treat) and prepare it for QC. Once the casting was ready for QC, it took an average of 9.2 hrs. to perform each process (gamma-ray and layout). Forty-three percent of the time during this 18-hr. span was spent waiting or performing other tasks not related to the sample casting.

By simply putting in the metrics without making any changes to the process itself, the cleaning room and QC department process time went from an average of 14.5 days down to 6.6 days (Fig. 2). There was enough data to support that it should take no more than five days to complete this part of the process.

[FIGURE 2 OMITTED]

While looking at the data for the cleaning room, the team noticed the process could be broken down into smaller increments because the measurement system looked at the days to do the process, not the actual hours spent working on the component.

All "non-value-added time" was removed from the data in the cleaning room, and the measures were converted into hours. The Pareto Chart was reproduced following those steps. The pattern shop process ascended to the top on this chart (Fig. 3). It was No. 2 on the original Pareto, but now it stood out as the major contributor to process time.

[FIGURE 3 OMITTED]

The process of building the boxes and boards took 40% of the pattern shop's time. From there, nothing was clear-cut as to why it took so long to convert a pattern. The team decided to speak with the employees doing the actual conversions and questioned the length of time required to convert a pattern. From the answers received on a survey, a ranking sheet was developed, and the team asked all the employees who performed the hands-on work to fill out and return.

The results from this survey made the team focus on the pattern boxes. Four out of the top five rankings were processes involved in building the boxes and boards. The corner blocks, quality of plywood and the rails were then analyzed. All of these were involved in building the boxes and boards and took 100 hrs. of work.

The patterns were not brought into the pattern shop to convert until Harrison had enough inventory to satisfy orders. Once that was done, the vice president would schedule the pattern into the pattern shop for conversion. Rehashing the original pattern shop process functional deployment map, the team noticed that work could be done prior to the pattern being ready to convert. The preliminary design drawings could be completed months in advance, and the pattern shop could build the boxes and boards without having the pattern available. At this time, the pattern shop riggers were slightly ahead of the game. A measurement system determined that over any given span, 17% of their time was spent on the boxes and boards. Thus, the team asked if they could outsource the boxes and boards, which would increase the cost by an average of $600.

Before moving any further, the team made the decision to propose a new process in the pattern shop. Hoping to run a small pilot, the team wanted to simulate having the boxes and boards completed prior to the pattern being available for conversion. They hoped that the savings in time was worth the additional $600.

During a review, the team was permitted to run its pilot. A component was chosen as the pilot pattern, and the decision was made to incorporate the entire process with proposed changes made by the team.

The cost to convert the pilot pattern decreased by 1% compared to the overall average, but increased 9% compared to patterns of similar size. The cost increase was attributed to the use of plank board, which is a more expensive material, on the sample. If the boxes and boards had been outsourced, it would have cost an additional 14%.

Two options then were proposed for decreasing the amount of time it takes to rebuild a pattern. In option 1, one group of three pattern shop employees would be assigned to work on the boxes and boards. The group's responsibility would be to stay at least four patterns ahead of the conversion schedule. In option 2, the boxes and boards would be outsourced. Harrison would need to provide the drawings, plank board and plywood in a timely manner in order for the supplier to keep the in-house inventory at a minimum of four patterns ahead of the conversion schedule.

Implementing the Process

Using the data gathered during the analysis, Harrison allocated money to purchase 10 boxes and boards from an outside source to begin the implementation. Then, the control part of the process was set up to monitor and report feedback on pattern shop hours, dollars associated with pattern conversion, sample time and feedback time.

Through a total-system overview driven by Six Sigma philosophies, Harrison Steel achieved its goal. The last green sand mold was produced on June 27, 2003--18 months ahead of schedule. As a result of meeting the deadline, the conversion project was credited with increasing the total line savings by 14%.

Troy Sheridan, The Harrison Steel Castings Co., Attica, Ind., is a Six Sigma Black Belt and led Harrison's conversion project.

For More Information

Visit www.moderncasting.com to view, "Six Sigma--Converting Patterns from Green Sand to AML Molding," T. Sheridan, Proceedings from the 2003 Steel Founders' Society of America T&O.

"Harrison's Super Size Nobake Conversion," K. Bauer, MODERN CASTING, Oct. 2004, p. 18-22.
COPYRIGHT 2005 American Foundry Society, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2005, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Author:Sheridan, Troy
Publication:Modern Casting
Geographic Code:1USA
Date:Jan 1, 2005
Words:1818
Previous Article:Getting more from your cleaning room: cleaning and finishing operations in low- and high-volume facilities have their problems, from automation to...
Next Article:Marrying Almag 535 to the permanent mold process.
Topics:


Related Articles
Flaskless Molding Machine Passes Test at Gray Iron Foundry.
Nobake Molding and Pouring System Eases Changes at Ferrous Foundry.
Rowe's Nobake Expansion Lifts Productivity, Lowers Costs.
Willman: Making Big Green Sand Castings for a Lower Overall Cost.
Troubleshooting Casting Defects in Nobake Molding.
Harrison Steel to Switch to Nobake in Three Years.
A tale of two nobake start-ups.
Core Blower saves small foundry $30,000/Yr, eliminates coreroom personnel. (Case History).
Harrison's super size nobake conversion: instead of gradually adding nobake molding lines and phasing out its existing green sand lines, Harrison...
Inoclulating iron through filters.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters