Course Number 032 - Design of Experiments I

This 10-week course uses a simple and straightforward approach, Taguchi and traditional Design of Experiments will be presented.  Hand's-on workshops will illustrate how Design of Experiments are used for solving problems and optimizing processes.

TOPICS include:

• 2 & 3 Factor Experiments

• Screening Designs

• Taguchi L4 L8 orthogonal arrays

• Excel templates will be used for data analysis

Syllabus

 Week 1

            Introduction to DOE

            Lecture:            Barrentine Chapter 1

            Workshop:         2 Factors at 2 levels

            Homework:        Read Chapter 1; Read Chapter 2, pages 5-21;

                                    Develop a straight line model, (y = mx + b) for the workshop data

                                    Read George Box paper: ‘Do interactions really matter?’

Week 2

            Review:             Week 1 Homework

            Review:             George Box paper: ‘Do interactions really matter?’

            Lecture:            Barrentine Chapter 2, pages 5 – 21

            Workshop:         The Eight Steps for the Analysis of Effects using 2 Factors at 2 Levels

            Homework:        Read Chapter 3, pages 27-33;

Read ‘Communicating DOE to Non-Statisticians’, Steve Schmidt, author, pages 1-5; input the FORCE experiment into the Excel DOE templates. Using the Excel templates compare and contrast analysis methods between Barrentine and Schmidt.

Week 3

            Review:             Week 2 Homework

            Review:             Steve Schmidt paper, ‘Communicating DOE to Non-statisticians’, Force experiment;

            Lecture:            Barrentine Chapter 3, pages 27-33

            Workshop:         Experiments with three factors

            Homework:        Barrentine Exercise 3, Ink Transfer, using 3 factor Excel template;

                                    conduct an analysis for both effects and variation (variance).

Read ‘Communicating DOE to Non-statisticians’ pages 5 – 9; input the Ohm experiment into the 3 factor DOE template.  Compare and contrast analysis methods between Barrentine and Schmidt.  Is there a significant factor, or factors with respect to variance?

 Week 4

            Review:             Week 3 Homework

            Review              Steve Schmidt paper, ‘Communicating DOE to Non-statisticians’, Ohm Experiment;

            Lecture             Barrentine Chapter 3, pages 34 to 36, ‘Analysis of Unreplicated Experiments.

                                    DOE analysis using the difference between two replicates.

            Workshop          Experiments with three factors, one and two replicates.

            Homework:        Analyze workshop DOE and determine significant factors for both effects and variability.

                                    Read ‘Report No 35, The necessity of Modern Quality Improvement and Some Experience With its Implementation in the Manufacture of Rolling Bearings’, C. Hellstand, author.   Treating the data as a single replicate, perform an analysis as per Barrentine, pages 34 – 36.

                                    Read Barrentine Chapter 4, Screening Designs, pages 37 – 40 up to Example 4

 Week 5

            Review              Week 4 Homework

            Review              C. Hellstand paper ‘The necessity of Modern Quality Improvement and Some Experience With its Implementation in the Manufacture of Rolling Bearings’, and analysis of unreplicated experiment.

            Lecture             Screening Designs, aka Fractional Factorials

            Workshop          Experiments with 4 Factors

            Homework         Analyze workshop DOE and determine significant factors for both effects and variability.

                                    Analyze 4 factor Popcorn DOE treating data as a single replicate, include analysis of residuals.

                                    Read Barrentine pages 45 – 47 up to table 4.8                                 

Week 6

            Review:             Week 5 Homework

            Lecture:            Unconfounding Interactions using Reflection, aka Fold-over designs

            Workshop          Fold-over Experiments using 4 factors.

Homework         Analyze workshop DOE and determine significant factors for both effects and variability using Reflection method, ref Barrentine pages 46 – 49. Include analysis of residuals.                      

Week 7

            Review:             Week 6 Homework

            Lecture:            Blocking 8 run Designs

            Workshop:         Experiments with 4 factors and 1 Block

Homework         Analyze workshop DOE and determine significant factors for effects, blocks and variability.

                        Read ‘What We Can Learn from 8 Run Designs’, George Box author.

Week 8            

Review              Week 7 Homework

            Lecture:            ‘What We Can Learn from 8 Run Designs’

            Workshop:         Experiments with 7 factors, standard order matrix

Homework         Analyze workshop DOE and determine significant factors for both effects and variability.  Propose what the next logical 8 run design would be based upon the data analysis

Read Barrentine Example 4, pages 40 – 46.

Read ‘Pick a Design, Any Design’, C.J. Keller and Richard Scranton, authors

Week 9

            Review              Week 8 Homework

            Lecture             ‘Pick a Design, Any Design’

            Workshop          Experiments with 7 factors Plackett Burnham matrix

Homework         Analyze workshop DOE and determine significant factors for both effects and variability.  Propose what the next logical 8 run design would be based upon the data analysis

                        Prepare for final project. 

Week 10                       Final Project

      ==============================================================     

Please email about classes, registration, or any other questions or concerns you may have.