Six Sigma Green Belt Certification ProjectUniversity of Michigan

Finding the right temperature[1]

Rubberland, Inc. is a producer of liquid rubber used in the production of molded parts. This company has been receiving negative feedback from their customers due to excessive variation in the curing initiation temperature of their liquid rubber. The manufacturing process is illustrated in the following figure.


Currently, their customers need to adjust the settings for each new batch of material. These continuous changes cost the customers an average of one hour per day per machine, which is equivalent to a process downtime of 6.25%. The management at Rubberland has organized a multidisciplinary team consisting of Process Engineers, Chemists, Operators and Supervisors to identify a solution to this problem. This team has identified several major sources of variation during various brainstorming sessions. The results were compiled and are attached in a data file. In addition, the team has conducted an observational study of the process. They collected data on several process input and response (output) variables, including the cure initiation temperature (CIR), which has a specification of 140 oC +/- 5 oC. These data have been obtained for 118 samples taken over four months.

Assignment Deliverables

Using your six-sigma green belt skills, prepare a report that follows the DMAIC process and identifies possible solutions to reduce the amount of variation in the process output variables. Your report should include the following minimum requirements:

  • Definition of the problem (based on the written case description above),
  • Identification of a performance metric(s) KPOVs.
  • Assessment of the Current State of your metric.
  • Analysis which identifies those input variables which are contributing to lack of performance of your metric as well as those variables which are not affecting performance (robust input variables). For such input variables, identify robust operating windows. For the output variables, use an appropriate analysis technique to determine whether the number of response variables that are currently monitored could be reduced. (Hint: correlation.)
  • List Potential Improvements and Recommendations to improve performance. Develop a control plan to insure they are followed. (Note: the assignment involves some creativity here as you do not have the ability to actually verify your recommendations or control plan.)
  • Prediction of future state if you were to remove the sources of variation. Feel free to sort the accompany data sets and only examine data inside the region (operating windows) you plan to operate. For example, if you were to eliminate one of your defect categories through some recommendation, re-compute performance without this defect category. Again, this assignment will involve some creativity, as you do not have the ability to actually verify future performance under your new set of recommendations. Ultimately, try to make an assessment of the potential improvement from the current to future state using the available data.

Green Belt Certification Projects should follow the template provided on the course web site. This template is intended to minimize formatting questions. Your report should have a mixture of effective tables and/or graphs in addition to written text. Detailed calculations should be included in the appendix. The length of your report should be approximately 3-6 pages, single-spaced (includes an allowance for 2 or 3 graphs and/or tables). This length does not consider the appendix. Please do not include overly lengthy appendices. These guidelines are not absolute requirements; they are intended to simplify the report writing needs.

Note: If your wish to follow another report template, please obtain approval from course instructors. In addition to help you organize your thoughts, the report template also simplifies the grading process for us.

Data Analysis Strategies

Clearly, multiple approaches exist to analyzing these data sets. Thus, one of your principal tasks in this assignment is to determine an efficient analysis strategy in addition to choosing the appropriate analysis tool.

Accompany Data Files

  • process-data.xls – Excel Data File with process variables and cure initiation temperature for 118 different lots. Process Variables include: Duro Shore, Tensile Strength, 100% modulus, Elongation, Tear Strength, Catalyst weight, etc. Process outputs include various cure rates and initial curing temperatures for the catalyst and the liquid rubber.
  • brainstorm.xls - Excel Data File with the list of most plausible causes of the problem based on the opinion of an interdisciplinary team.

Project Due Dates

Our intention was to provide an assignment whose data analysis requirements could be completed in approximately 8 hours. If you are taking significantly longer, feel free to contact your instructors for direction.

1

[1] This case study has been prepared as a teaching tool to explore various data analysis skills used in Six Sigma problem-solving. The data used in this case study are based on real-world examples; however, the data has been manipulated to provide a teaching case study for a manufacturing Six Sigma project.