题 目：Some Statistical Methods in Reliability Engineering Research
主讲人：Rong Pan 博士
Informatics and Decision Systems Engineering Arizona State University
时 间：2009年12月28日上午10：00—11：00
主讲人简介：
Dr. Pan received his doctorate in Industrial Engineering from the Pennsylvania State University in 2002. Prior to coming to ASU in 2006, he was an assistant professor of Industrial Engineering at the University of Texas at El Paso. His research interests include failure time data analysis, design of experiments, multivariate statistical quality control, time series analysis and control. He is the recipient of 2008 Stan Ofsthun Award of the Society of Reliability Engineers. He is a senior member of ASQ and a member of SRE, IIE and INFORMS. He has published in Journal of Quality Technology, Journal of Applied Statistics, International Journal of Production Research, Quality and Reliability Engineering International, etc. His current research project on modeling and analysis of profiled reliability testing is funded by the National Science Foundation (NSF). His previous projects were funded by U.S. Department of Education (DoEd), Texas Department of Transportation (TxDOT) and GM.
内容简介：
Reliability engineers consider the time dimension of quality, i.e., the product or process quality over time. Like quality engineering, the research in the field of reliability engineering utilizes many statistical tools for modeling and data analysis. In this talk, I will discuss two topics from my recent research on accelerated life testing (ALT). The first topic is the data analysis of step-stress accelerated life tests. We formulate theproblem through generalized linear models and provide a Bayesian analysis solution. The second topic is the experimental design problem of accelerated life tests. I will discuss the research motivation through a practical example, which highlights the complexity and difficulty of planning industrial ALT experiments. The optimal experimental design obtained based on a D-optimal criteria is compared with a legacy design and an ordinary orthogonal design. The research to be presented in this talk is supported by two NSF grants.
