TITLE: Statistical Shape Analysis of Manufacturing Data

SPEAKER: Professor Enrique del Castillo

ABSTRACT:

We show how Statistical Shape
Analysis, a set of techniques used to model the shapes of biological and other
kind of objects in the natural sciences, can be used also to model the
geometric shape of a manufactured part. We first review Procrustes-based
methods, and emphasize possible solutions to the basic problem of having
corresponding, or matching, labels in the measured ``landmarks", the
locations of the measured points on each part acquired with a CMM or similar
instrument. The analysis of experiments with shape responses is discussed
next.  The usual approach in practice is
to estimate the form error of the part and conduct an ANOVA on the form errors.
Instead, an F ANOVA test due to Goodall and a new permutation ANOVA test for
shapes are presented. Real data sets as well as simulated shape data of
interest in manufacturing were used to perform power comparisons for 2 and 3
dimensional shapes. The ANOVA on the form errors was found to have poor
performance in detecting mean shape differences in circular and cylindrical
parts. The ANOVA F test and the Permutation ANOVA  test provide highest power to detect
differences in the mean shape. It is shown how these tests can also be applied
to general "free form" shapes of parts where no standard definition
of form error exists in manufacturing practice. New visualization tools,
including main effect and interaction plots for shapes and deviation from
nominal plots are presented to help interpreting the results of experiments
where the response is a shape.

Bio: 

Dr. Castillo is
a Distinguished Professor of Industrial Engineering.  Dr. Castillo also holds a joint appointment
with the Department of Statistics.  He is
an author of over 85 journal papers and 2 textbooks and a former NSF CAREER awardee,
Fulbright Scholar, and the Editor in Chief of the Journal of Quality Technology
(2006-2009).