Reflecation paper

- Module 9 - Using Prototypes for

Product Assessment  Overview

The purpose of prototype assessment is to conduct an evaluation to

(1) provide answers to design questions,

(2) improve the prototype or design,

(3) review the usefulness of evaluation provided by the prototype,

(4) transform evaluation results into design or revision

recommendations.

1Dr.Munther Hermez

- Module 9 – Continue

Prototype structure set-up to include:

1. Review prototype objective:

2. Identify the key issues:

3. Review the level of approximation of the prototype:

4. Review the experimental plan:

2Dr.Munther Hermez

- Module 9 – Continue

5. Review the schedule for procurement, construction,

and testing:

6. Complete test and measurement preparation

7. Analyze the measurements:

8. Perform the assessment and report including proposal

for improvements. Reflect on the design process:

9. Reflect on the design process:

(3Dr.Munther Hermez

- Module 9 – Continue

4Dr.Munther Hermez

Prototyping outcome could be:

(1) design validated as is.

(2) minor adjustments needed, overall approach validated.

(3) concept still worth investigating but serious problems

identified.

(4) design approach not validated.

(5) decide whether the prototyping result has changed in the

design perspective.

- Module 9 – Continue

 DESIGN OF EXPERIMENTS

• Quality Engineering by Design (QED)

• Taguchi Technique

can change many variables at the same time and still retain

control of the experiment.

5Dr.Munther Hermez

- Module 9 – Continue

 Standard Deviation

> is a statistic that tells how tightly all the various examples are

clustered around the mean in a set of data.

> can define MSD (mean squared deviation) = total deviation

squared / number of samples

> can also define V, or Variance,= total deviation squared =

(number of samples 1)

6Dr.Munther Hermez

- Module 9 – Continue

• Define:

σ : standard deviation,

7Dr.Munther Hermez

- Module 9 – Continue

 LOSS FUNCTION

Taguchi loss function is a way to show how each imperfect part

produced, results in a loss for the company.

Defined as:

8Dr.Munther Hermez

- Module 9 – Continue

9Dr.Munther Hermez

- Module 9 – Continue

10Dr.Munther Hermez

Loss function

- Module 9 – Continue

 ORTHOGONAL ARRAYS

• OA that is a set of well balanced (minimum) experiments in which

all parameters of interest are varied over a specified range

• Levels are values that a factor assumes when used in the

experiment.

• The number of factors and their levels determine the choice of an

OA.

11Dr.Munther Hermez

- Module 9 – Continue

12Dr.Munther Hermez

- Module 9 – Continue

 TAGUCHI’S DOE PROCEDURE

Taguchi method procedures sequence are:

1. Identifying the control factors and the noise factors and

evaluating the interactions.

2. Choosing the levels for the factors.

3. Selecting an appropriate ORTHOGONAL ARRAYS (OA).

4. Assigning the factors and interactions to columns of the OA.

13Dr.Munther Hermez

- Module 9 – Continue

5. Conducting the experiments.

6. Analyzing the data and determining the optimal levels.

7. Predicting the responses for the optimal levels.

8. Running the F-tests and estimating the confidence.

9. Conducting the confirmation experiment.

14Dr.Munther Hermez

- Module 9 – Continue

 ANALYSIS OF VARIANCE (ANOVA)

ANOVA is a technique that can be used to test the hypothesis

that the means among two or more groups are equal

• ONE-WAY ANOVA

• TWO-WAY ANOVA

• THREE-WAY ANOVA

15Dr.Munther Hermez

- Module 9 – Continue

• INTERACTION EFFECTS

• TWO-WAY ANOVA AND ORTHOGONAL ARRAYS

• SIGNAL-TO-NOISE RATIOS

16Dr.Munther Hermez

- Module 9 – Continue

 ANOVA USING EXCEL

 SINGLE-FACTOR (ONE-WAY) ANOVA

Step 1: Enter data in a spreadsheet

Step 2: Select Data Analysis from the Tools Menu

Step 3: Fill in the required information in the dialog boxes

Step 4: ANOVA output summary

17Dr.Munther Hermez

- Module 9 – Continue

 TWO-FACTOR (TWO-WAY) ANOVA WITHOUT REPLICATION

The procedure to conduct a two-way ANOVA is similar.

Step 1: Enter data in a spreadsheet.

Step 2: Select Data Analysis from the Tools Menu. Choose ANOVA:

Two-Factor Without Replication.

Step 3: Fill in the required information in the dialog boxes.

Step 4: ANOVA output summary.

18Dr.Munther Hermez

- Module 9 – Continue

 TWO-FACTOR (TWO-WAY) ANOVA WITH REPLICATION

The procedure to conduct two-way ANOVA with replication is similar.

• Step 1: Enter data in a spreadsheet.

• Step 2: Select Data Analysis from the Tools Menu.

• Step 3: Fill in the required information in the dialog boxes.

• Step 4: ANOVA output summary.

19Dr.Munther Hermez

- Module 9 – Continue

 QUALITY CHARACTERISTIC (QC)

• Is any dimensional, mechanical, and physical property.

• predefine the prototype objectives:

-Why is the prototype made ?

-What are the objectives of the prototype ?

-What are the main features to be measured ?

-How are they going to be measured ?

20Dr.Munther Hermez

- Module 9 – Continue

 OVERALL EVALUATION CRITERION (OEC)

.OEC comes into the picture when there is more than one

performance criterion or response in the experiment.

. The QCs must be normalized and weighted in order to give a single

OEC

21Dr.Munther Hermez

- Module 9 – Continue

Wx is the weight percentage allotted to the QCx.

OEC will be evaluated based on the magnitude of deviation from

the nominal value.

(22Dr.Munther Hermez

- Module 9 – Continue

 PREDICTIVE MODEL

• predictive model is constructed from the results of the OA analysis

where, Y is a QC, A, B, C, D,

YA is the best response value of A, and so on.

23Dr.Munther Hermez