Discipline: Mathematics and Statistics
Subcategory: Civil/Mechanical/Manufacturing Engineering
James D. Finnie II - Virginia State University
Co-Author(s): Britton Bean, Virginia State University, Petersburg , VA
EPT is a German manufacturing company in Virginia, which produces injection molded electronic components for automotive systems to ensure the safety of motorists as well as to prevent the loss of life in dangerous situations. Due to the important task these systems perform, it is critical to ensure that they are consistently produced accurately and reliably. As part of SPC Statistical Process Control, the MSA Measurement System Analysis is an experimental method used to determine how fluctuation within a measurement system contributes to the overall variability of a production process.
In our NSF project we focused on Type I and Type II Gage Repeatability and Reproducibility, also known as Gage R&R. The Type I study focuses on repeatability which is a measure of precision with one operator measuring the same part using the same tool. Type II Reproducibility study focuses on different operators measuring with the same tool. Repeatability is a study of standard deviation, while reproducibility is a study of mean. In our study we primarily used a Wenzel LH65 CMM coordinate measurement machine. By testing the measurement process of the Wenzel CMM using two different methods, we could examine the potential causes for a rise in bias. Using several statistical analysis software packages such as Minitab, Q-Das and ProLink SPC, we discovered bias introduced due to probe rotation and gravitation; although overall, the error due to measurement itself was low. The Reproducibility Gage Type II studies with three different operators on the Wenzel CMM demonstrated that the composite error due to tool and evaluator was also acceptably low. Further Type II studies using a manual digital height indicator also verified the measurement process to contribute only a small fraction of part tolerance. Thus the MSA Measurement System Analysis procedures of EPT’s Statistical Process Control methodology were satisfactory.
Funder Acknowledgement(s): NSF
Faculty Advisor: Shahzad Akbar,