Robust design processes for quality & reliability
Design of experiments for industrial printer design
University of Westminster, London, UK
Monday 28th - Tuesday 29th June, 2010

Course fee - GBP 740, Euro 850, US$ 1,175		 Course registration

Location & accommodation

Back to Summer School 2010
COURSE FOCUS
The Design of Experiments (DOE) course introduces the audience to robust design using the methods made popular by Dr. Genichi Taguchi. The class is unique because it is focused on the printing industry and the examples are related to printing. The robust design methodology focuses on the choice of appropriate Quality Metrics and the incorporation of noise and variability into the product design process. The design method also emphasises the efficient use of experimental and design resources. Once the important design parameters and experimental windows have been chosen, the robust design process helps us to minimise the required experimental effort required to find the optimum design and understand our sensitivity to noises present in customer environments. The course emphasises practical design examples and the audience completes multiple examples that are geared to the printing industry.

Class Objectives:
• Learn how to apply the robust design process to industrial printer design
• Learn the fundamentals of robust design process
• Learn how to efficiently plan and analyse designed experiments
• Learn how to minimise the experimental effort required to improve printer performance
• Learn how to reduce development time and increase product robustness
• Gain an understanding of the background, concepts, and terminology associated with robust design and quality engineering

Included with the Class:
• A course binder along with numerous examples related to the printing industry
• Simple Microsoft Excel templates for analysing the examples taught in class
COURSE LEADER
Dr John E LaFleche
Flexible Engineering Solutions
Tolland, Connecticut, USA

Flexible Engineering Solutions (FES) was founded in 1997 by John E LaFleche, PhD FES has a broad international client base that includes manufacturers of ink jet, thermal, laser and xerographic printing systems as well as manufacturers of paper and paper handling devices.

John earned a Ph.D. in Mechanical Engineering from the University of Rochester. His Ph.D. research concentrated on thermal and mechanical simulations of digital printing processes and was supported by Eastman Kodak Company.

John has over 15 years of experience in digital imaging technology. In addition to consulting, he designed and teaches four custom training courses: Paper handling, thermal printing, Design of Experiments for Printers and fuser technology review.

John is an expert in ink jet, thermal, and laser printing. He has trained over one thousand engineers at the best printing companies in the world, including Hewlett Packard, Xerox and Lexmark.
Course outline
Monday June 28, 2010

10:00 – 11:00 am            Registration

11:00 am            Opening session

Introduction
• Class objectives
• Basic Robust Design concepts and history
• How to apply the Robust Design Process to printer design
• Example: Improving printer output consistency

Statistics Review
• Mean, variance, standard deviation
• How to choose the appropriate metrics for evaluating printer
performance
• Example: Paper handling to minimise skew
• Example: Paper feeding to avoid jams, mis-feeds and double feeds

12:30 - 1:30 pm Lunch

Robust Design Steps
• Steps for planning, testing, and analysing printer performance
• Example: Print artifact analysis
• Planning Stage:
- Step 1: Identify the main function
- Step 2: Identify noise factors
- Step 3: Identify quality characteristic and objective functions
- Step 4: Identify control factors and levels
- Step 5: Design / Choose matrix and data analysis plan
• Experimental stage
- Step 6: Conduct the matrix experiment
• Analysis and verification stage
- Step 7: Analyse data and choose optimum levels
- Step 8: Run verification experiment
• Summary

5:30 pm Session closes

6:00 – 7:00 pm Reception

Tuesday June 29, 2010

9:00 am Session 2

Experimental Design Basics
• Approaches
- "Full factorial"
- "Fractional factorial" (orthogonal arrays)
- "Everything constant"
- Pros and cons
• Standard orthogonal arrays
• Screening experiments
– What factors have the most influence on the printer performance
• How to identify interactions between factors
• Adding noise to the experiments to evaluate robustness
• Example: How to identify interactions between factors
• Example: Experimental setup with the inclusion of noise

12:30 - 1:30 pm Lunch

Data Analysis
• Introduction
• Experimental setup
• Analysis of means (ANOM)
• Analysis of variance (ANOVA)
• Confirming experiment
• Example: Simultaneous paper de-skew and feed
• Class wrap-up

5:00 pm Adjournment