The DOE training seminar begins with the fundamentals of Design of Experiments (DOE) methods and continues with advanced concepts, principles and requirements. Topics include Anova, Full Factorial Designs, Fractional Factorial Designs, robust designs, the Response Surface Methodology(RSM), reliability DOE and Taguchi Design. We will begin with screening design, process characterization and optimization.

- Participants learn to solve problems, improve yields, achieve robust processes and build models for prediction with Design of Experiments (DOE)
- Response Surface Methodology (RSM) and Multiple Regression Analysis
- The training course presents concepts and Minitab or excel based templates tools that you could use to help your organization:

- The Design experiments that are effective for studying the factors that may affect a product or process
- Analyze experimental results in order to identify the significant factors and evaluate ways to improve and optimize the design
- To determine if interactions between factors are significantly affecting the output of the process

Lecture, Discussion & Case Studies

1. Baselining Data collection

It is considered passive observation. The process is monitored and recorded without intentional changes or tweaking. In Designed Experiments, the independent variable (Response) is observed. Designed experiments are used to:

- Determine which factors (X’s) have the greatest impact on the response (Y)
- Quantify the effects of the factors (X’s) on the response (Y)
- Prove the factors (X’s) you think are important really do effect the process

2. This training includes the full understanding, applications and terms used in lectures and case studies

3. Orthogonality

Since our goal in experimentation is to determine the effect each factor has on the response independent of the effects of other factors, experiments must be designed so as to be horizontally and vertically balanced. An experimental array is vertically balanced if there are an equal number of high and low values in each column. The array is horizontally balanced if for each level within each factor we are testing an equal number of high and low values from each of the other factors. If we have a balanced design in this manner, it is Orthogonal. Standard generated designs are orthogonal. When modifying or fractionating standard designs be alert to assure maintenance of orthogonality.

4. Repetition

Completing a run more than once without resetting the independent variables is called repetition. It is commonly used to minimize the effect of measurement and to analyze factors affecting short term variation in the response

5. Replication

Duplicate experimental runs more than once after resetting the independent variables is called replication. It is commonly used to assure generalization of results over longer term conditions

6. Randomization

Running experimental trials in a random sequence is a common, recommended practice that assures that variables that change over time have an equal opportunity to affect all the runs. When possible, randomizing should be used for designed experimental plans

7. Blocking

A block is a group of “homogeneous units”. It may be a group of units made at “the same time”, such as a block by shift or lot or it may be a group of units made from “the same material” such as raw material lot or manufacturer. When blocking an experiment, you are adding a factor to the design.

8. Factorial Designs

Factorial Designs are primarily used to analyze the effects of two or more factors and their interactions. Base on the level of risk acceptable, experiments may be either full factorial, looking at each factor combination, or fractional factorial looking at a fraction of the factor combinations. Fractional Factorial experiments are an economical way to screen for vital X’s. They only look at a fraction of the factor combinations. Their results may be misleading because of Confounding, the mixing of the effect of one factor with the effect of a second factor or interactions. In planning a fractional factorial experiment, it is important to know the confounding patterns, and confirm that they will not prevent achievement of the goals of the DOE.

9. DOE Analysis:

Analysis of DOE’s includes both graphical and tabular information. It includes Pareto Analysis, Anova, Main Effects, Interactions analysis. It also include Cube Plots, Contour Plots and Optimization Plot,etc.

10. Response Surface Methodology (RSM)

Response Surface analysis is a type of Designed Experiment that allows investigations of non-linear relationships. It is a tool for fine tuning process optimization once the region of optimal process conditions is known. Using the CCD type RS Design, you will be designing an experiment that test each factor at five levels, and an experiment which can be used to augment a factorial experiment that has been completed. The CCD design will include Factorial points, STAR points, and CENTER Points.

11. Taguchi Designs

A Taguchi design, or an orthogonal array, is a method of designing experiments that usually requires only a fraction of the full factorial combinations. An orthogonal array means the design is balanced so that factor levels are weighted equally. Because of this, each factor can be evaluated independently of all the other factors, so the effect of one factor does not influence the estimation of another factor. In robust parameter design, you first choose control factors and their levels and choose an orthogonal array appropriate for these control factors. The control factors comprise the inner array (Signal). At the same time, you determine a set of noise factors, along with an experimental design for this set of factors. The noise factors comprise the outer array (NOISE). The L8 (2**7) Taguchi design (orthogonal array). L8 means 8 runs. 2**7 means 7 factors with 2 levels each. If the full factorial design were used, it would have 2**7 = 128 runs. The L8 (2**7) array requires only 8 runs − a fraction of the full factorial design. This array is orthogonal; factor levels are weighted equally across the entire design.

12. Anova

13. Essential Basic Statistics

14. Test for Normality

15. Understanding of 1-way Anova with many levels

16. Test for equal variance

17. Understanding of 2-Way Anova with one or more levels

- Engineering Manager / Executive / Supervisor/Engineers
- Process Improvement Managers / Process Engineers
- QC/QA Manager / Executive / Engineers
- Personnel involved in Quality Control & Improvement projects

## No comments:

## Post a Comment