Seminar on Industrial Mathematics and Statistics
Friday, March 31, 2000
8:00 A.M. - 12:00 Noon
Department of Mathematics and Statistics
College of Arts and Sciences
Oakland Room at Oakland Center
  • 8:15 A.M. Refreshment Welcome by the Chair, Marc Lipman


  • 8:45 A.M. Presentation I: Mathematical and Computational Questions in Body Manufacturing


  • 9:30 A.M. Presentation II: Assembly Modeling for Quality and Productivity


  • 10:30 A.M. Refreshment Break


  • 11:00 A.M. Presentation III: Variation Drivers Analysis -- A Variance Decomposition Method for Multi-Stage Data


  • 12:00 Noon Seminar Adjourns


  • Each presentation will consist of a 30 to 45 minute talk of the presenter followed by a 15 minute discussion.
  • Please bring your colleagues who may be interested.
  • Please offer suggestions for potential topics or appropriate presenters for future seminars. Your volunteering is appreciated!
  • Please give all of your suggestions to: James Pan: (248) 370-3449, pan@oakland.edu, or Devadatta Kulkarni: (248) 370-4032, kulkarni@oakland.edu.
  • Presentation I: Mathematical and Computational Questions in Body Manufacturing

    Abstract: The goals of reducing vehicle mass and shortening vehicle development time are forcing significant changes in materials and process for body manufacturing. The ability of manufacturing designers and engineers to efficiently implement these technology changes is enhanced by the availability of mathematical and computational tools to synthesize, model and analyze manufacturing operations prior to hardware installation. In this presentation we review, for the area of body manufacturing, mathematical and computational issues which emerge as virtual body manufacturing capability is sought. Several specific examples will illustrate progress as well as gaps. Our efforts to develop solutions to some of these problems through a recent academic partnership with University of Michigan will also be briefly reviewed.

    Speaker: Dr. Samuel P. Marin, General Motors Research and Development Center.

    Sam received his Ph.D. in Mathematics from Carnegie Mellon University in 1978. He is the Laboratory Group Manager of the Body Assembly Group in the GM R&D Center's Enterprise Systems Lab. In this role, he is responsible for research focused on the development of mathematical tools to improve joining and assembly operations in body manufacturing. His research interests are in geometric design and approximation, and in the numerical solutions of PDEs

    Presentation II: Assembly Modeling for Quality and Productivity

    Abstract: Assembly is critical to successful product realization. In this talk, I will review various researches that are currently available in the general area of mechanical assembly. Research at the University of Michigan in modeling assembly quality and productivity will be presented with emphasis on automotive body assembly. New techniques have been developed for predicting the variation of compliant, non-rigid part assembly by combining engineering structure analysis with statistical techniques. In addition, analytical models have also been developed for evaluating the productivity of assembly systems with various configurations.

    Speaker: Dr. S. Jack Hu, The University of Michigan.

    Dr. Hu received his Ph.D. from the University of Michigan in 1990. Currently he is an associate professor in the Department of Mechanical Engineering and Applied Mechanics. His research interest is in assembly and joining, and engineering statistics. He is currently the Director of the National Science Foundation Industry/University Cooperative Research Center at UM, and is also co-directing the Advanced Body Design and Manufacturing Division of the General Motors Satellite Research Lab at UM.

    Presentation III: Variation Drivers Analysis - A Variance Decomposition Method for Multi-Stage Data

    Abstract: When an item is assembled in a series of distinct steps, each step can contribute variation to the final product. If the final variation bute variation to the final product. If the final variation is unacceptable or if process improvement is a goal, how does one determine the stage that make the biggest contribution to the final variation? The variation drivers methodology provides an empirical model-based approach to answer that question by partitioning the variance at each stage into added variation and transmitted variation. This talk will discuss the model and the computations and give simple graphical techniques for analyzing the data and presenting results. Complications, such as missing data and measurement error, that frequently arise in practice will also be discussed.

    Speaker: Dr. Michael Wincek, General Motors, Research and Development Center.

    Mike is a staff research scientist at the GM R&D center. He received his PhD in statistics from the University of Wisconsin, Madison. His recent work involved visualization techniques for and analysis of dimensional data. His interests include linear models, time series analysis, and wavelets.