Statistical Design of Experiments

 

January 21, 22, 23, 2009
University of Delaware (Newark Campus)

Register Online!

Course Description

We are extremely pleased to have Dr. Babatunde Ogunnaike once again offer this very practical and much-needed course.

Are you an engineer or scientist who is well-trained in the “art” of problem formulation and problem-solving - but only when all entities involved are deterministic in nature?

Is your background rusty (at best) when it comes to applying the fundamentals of probability and statistics - especially when it comes to handling randomly varying phenomena?

Are you concerned that dealing with randomly varying phenomena must involve calculus? (Guess what - algebra is sufficient!)

Are you able to generate lots of data - but then you're uncertain how to interpret that data and design the next experiment?

This intensive and extremely practical course - now expanded based on feedback from previous participants - provides fundamental knowledge for using the tools of probability and statistics to design experiments, efficiently interpret large amounts of experimental data, cope with random variability and uncertainty, and use the results to improve quality and reliability.

Benefits

This course enables you to:

• learn from an instructor whose unparalleled background includes industrial applications of probability and statistics to industrial processes and experimental design;
• apply statistical inference and understand the role of carefully designed experiments;
• analyze large quantities of data and then design new experiments using that new knowledge;
• bring theory to practical application;
• use computer software to solve numerous real-life examples;
• interact and share experiences with peers from other companies;
• understand methods that can be put to use immediately.

There will be ample time for hands-on computer exercises in the UD computer lab!

The Presenter

Babatunde A. Ogunnaike joined the University of Delaware 's full-time faculty in 2002 and recently was named the William L. Friend Professor of Chemical Engineering. Since 1989, Tunde had served as part-time faculty, teaching Process Control & Dynamics, as well as Random Phenomena-Applied Probability and Statistics for Engineering Problem-solving, and participating in various research programs. In 1989, he had joined the DuPont Company's Advanced Control and Modeling group, where he was a research fellow. His work at DuPont involved on-line dynamic modeling for various processes, identification and control of nonlinear systems, applied statistics, and reverse engineering biological control systems for process applications. Tunde holds an MS in statistics and PhD in chemical engineering, specializing in process control, both completed in 1981 at the University of Wisconsin-Madison . From 1981-82, he was a research engineer in systems development at the Shell Development Corporation, where he designed and implemented advanced control schemes at two refineries and developed statistics-based modeling techniques for “dynamic matrix control”. He then spent six years as a professor of chemical engineering and statistics at his undergraduate alma mater, the University of Lagos, Nigeria. He is author or co-author for numerous texts and book chapters, has published extensively in technical journals, and is a frequent seminar speaker at universities around the world.

Agenda

Wed January 21 , 8:00 a.m. - Registration;
Wed January 21 and Thurs January 22, Class times: 8:30 a.m. - 4:30 p.m.
Fri January 23, Class time 8:30 a.m. - 1:00 p.m.

PART I: Why?

A. Motivation and Introduction

• Objectives and Nature of Experimental Research
• Random Variability and the Role of Probability and Statistics
• Statistical Inference and the Role of Carefully-designed experiments
• Phases of Efficient Experimental Studies
• The Role of Computer Software in DoE

B. Background

• Basic Concepts of Data Analysis
• Review of Statistical Inference

PART II: What & How?

A. Fundamental Concepts and Applications

• Terminology
• Single-factor Experiments:
• One-way Classification
  • Completely Randomized Designs
  • Fixed Effects and Random Effects
  • Latin Square and Related Designs
• Single-factor Experiments:
  Two-way Classification
  • Randomized Complete Block Designs
  • Latin Square and Related Designs
• Multi-factor Experiments
  • Two-factor Experiments
  • General Multi-factor Experiments
• 2k Factorial Designs

B. More Advanced Concepts

• Introduction (Going on from 2k Factorials)
• Screening Designs
  • Fractional Factorial Designs
  • Plackett-Burman Designs
• Response Surface Methodology
• Intro. to Taguchi Robust Parameter Designs

C. Summary, Conclusions and Going on from Here

Registration and General Information

Program Fee:  Groups of 3 or more from same company: $695/person.  Individual registrations: $745/person.  Includes continental breakfast on DAY ONE, all breaks, UD parking, course notes.  Lunches are not included, but Newark offers many convenient options.

REGISTRATION DEADLINE: January 6, 2009

REGISTRATION CONFIRMATION:  A confirmation letter including directions to campus and parking will be mailed to your home on or before January 6, 2009.

CANCELLATIONS and SUBSTITUTIONS: Refunds granted if the request is received in writing by the course registration deadline. No refunds after that date, but substitutions are permitted up to the first day of class.
ACCOMMODATIONS and TRANSPORTATION: Participants are responsible for making their own housing and transportation arrangements.  Air transportation should be arranged to either Philadelphia or Baltimore airport.  Call early and ask if the University of Delaware affiliated rates are available for the dates you are requesting from the following hotels:
Courtyard by Marriott - UD Campus Hotel - 400 Pencader Way, Newark, DE, 1-302-737-0900 ($144+tax/night)
The following are each about one mile from campus:
Sleep Inn - 630 S. College Ave.(Rt.896N), Newark, DE, 1-302-453-1700 ($79+tax/night)
Howard Johnson Inn - 1119 S. College Ave., Newark, DE, 1-302-368-8521 ($62+tax/night)