Continuous Improvement: Six Sigma and the DMAIC Process (Part 2)

Written by: Jim Gilbert, Senior Consultant

Our last blog post ended with the promise to look at some 6 Ϭ projects and determine if the reader has an interest in learning more about 6 Ϭ. 

I will walk you through two of my 6 Ϭ projects.   The first case is a manufacturing project and the second case is an administrative project.  Both required the application of some 6 Ϭ tools to understand the excess variation and ultimately eliminate the problems that were causing the poor performance.

Process

We used the DMAIC process where we engaged in each of the following project phases:

• Define – Describe the problem to better understand it
• Measure – Measure appropriate features to better understand the excess variation
• Analyze – Determine what can be causing the excess variation
• Improve – Permanently eliminate the root causes of the excess variation
• Control – Develop a Control Plan to provide counter-measures if the process begins to experience excess variation

Projects

Case #1: A production shop was experiencing excessive cycle times for straightening some of the shafts that come out of heat treat.  This caused takt time problems downstream and required both a safety buffer to maintain production and some overtime to catch up.  It was originally suspected that the more run out in the center of the shaft the longer the straightening time.  However, the data did not support that.  The table displays the Coefficient of Determinations (R²) for the average run outs at the three locations on the shaft and reveals that the run out at the flange is much more influencing of the total straightening time.  This led us to understand that the shaft adjacent to the flange necks up to a larger diameter and is therefore more difficult to straighten.  Upon going to the work station it was observed that the shafts were coming out of an in-line induction heat treat coil warped.  Further investigation showed that the shaft turned inside the coil and that some of the shafts turned as an eccentric.  This meant that one side of the shaft was always closest to the coil surface and the opposite side of the shaft was always farthest away from the coil surface.  The surface of the shaft closest to the coil heated at a rate faster than the surface farthest from the coil and a bow was therefore being created in the heat treat process.  This was then traced to how the center holes were drilled.  Once center holes were properly drilled the problem disappeared. The problem was corrected, safety buffers were reduced and most overtime was eliminated.

Case #2: A Defense Contractor required to submit an Earned Value Report each week to a Prime Contractor was consistently unable to do so during the beginning of the probationary period of the contract.  Analysis of the process revealed 26 consecutive steps with no concurrent work being performed.  Thus, if a mistake was made the process would have to start over.  It was also found that 69% of data entry errors were due to three errors that could be poka yoked (mistake proofed) by re-programming the entry screens.  The process was redesigned to produce results in table. Project saved direct costs over the first year and protected a $4,000,000 award bonus for perfect performance. Contractor controller calculated overall first-year savings at $695,200.

Conclusion

These two processes, while very different, were dramatically improved using the 6 Ϭ methodology and some of its tools.  It is highly unlikely that these improvements could have been made without this approach.

In our next blog, we will elaborate on the DMAIC process and discuss what Capable and Control means.  If a process is Capable and in Control, you will, by definition, get the outcome that the process was designed to produce.  Once we understand this we can discuss some of the tools that enable us to achieve this condition.