18
© 2007 The MathWorks, Inc. ® ® Design for Six Sigma with MATLAB ® Kevin Cohan The MathWorks, Inc.

Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

  • Upload
    others

  • View
    4

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

©20

07 T

he M

athW

orks

, Inc

.

® ®

Design for Six Sigma with MATLAB®

Kevin Cohan

The MathWorks, Inc.

Page 2: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

2

® ®

Design for Six Sigma

http://www.dtic.mil/ndia/2003test/kiemele.pdf

1

10

100

1000

Rel

ativ

e C

ost t

o M

ake

aD

esig

n C

hang

e

Product StageResearch Design Development Production

“Classic” Six Sigmafocuses here

DFSS focuses here

Page 3: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

3

® ®

DMAIC Methodology

Define the problem / defects

Measure the current performance level

Analyze to determine the root causes of the problem / defects

Improve by identifying and implementing solutionsthat eliminate the root causes

Control by monitoring the performance of the improved process

M

D

A

I

C

http://www.dtic.mil/ndia/2003test/kiemele.pdf

D M A I C

Page 4: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

©20

07 T

he M

athW

orks

, Inc

.

® ®

Example

Optimization of anEngine Cooling Fan Design

Image courtesy of Novak Conversions

Page 5: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

5

® ®

Define:Define the Problem

Existing engine isn’t cooled sufficientlyunder difficult driving conditions

New design requirementAirflow > 875 cubic feet per minute

ApproachOptimize design factorsfor maximum airflowUse the MATLAB product family toimplement the DMAIC methodology

D M A I C

Image courtesy of Novak Conversions

Page 6: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

6

® ®

Measure:Measure Cooling Fan Performance

Import historical datafrom Excel

Establish baselinefor comparison

Mean = 842 ft3 / minStd Dev = + 2 ft3 / min

D M A I C

Page 7: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

7

® ®

Analyze:Analyze Factors that Affect Performance

Three design factorswe can modify

Clearance Distance

Pitch

Min MaxDistance from radiator (d) 1.0 1.5 inchesBlade pitch angle (p) 15 35 degreesBlade tip clearance (c) 1.0 2.0 inches

FactorsRange

Units

D M A I C

Page 8: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

8

® ®

Analyze:Analyze Factors that Affect Performance

Gather additional data from existing systemModify our design factors and measure performanceFit a model to the performance dataEvaluate the model to understand theeffect of each design factor

What tests do we run?… and how many?

D M A I C

(d)

(p)

(c)

Test Run Dis

tan

ce f

rom

rad

iato

r

Bla

de p

itch

an

gle

Bla

de t

ipcl

eara

nce

Perf

orm

an

ce

1 ? ? ?23…n

Page 9: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

9

® ®

Analyze:Analyze Factors that Affect Performance

Use a designed experiment to gather performance dataChoose the Box-Behnken function to determinepoints in the design space

D M A I C

Page 10: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

10

® ®

Analyze:Analyze Factors that Affect Performance

Fit a quadratic model to the data (regstats function)

D M A I C

AF = B0 + B1X1 + B2X2 + B3X3 +B4X1X2 + B5X1X3 + B6X2X3 +B7X12 + B8X22 + B9X32

AF = Airflow (ft3/min)X1 = Distance from radiator (inches)X2 = Fan pitch angle (degrees)X3 = Tip clearance between fan blades

and shroud (inches)

Page 11: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

11

® ®

Analyze:Analyze Factors that Affect Performance

Use rstool to visually inspect interaction between all three design factors (distance, pitch, clearance) simultaneously

D M A I C

Page 12: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

12

® ®

Improve:Improve the Cooling Fan Performance

Use optimization to automate the task offinding the maximum airflow

fmincon function

D M A I C

Page 13: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

13

® ®

Improve:Improve the Cooling Fan Performance

Use uncertainty analysis to ensure a robust designTwo main contributors

Model uncertaintyManufacturing variability

D M A I C

Model noise 0.00 +/- 0.96 ft3/minDistance from radiator (d) 1.00 +/- 0.005 inchBlade pitch angle (p) 27.3 +/- 0.5 degreesBlade tip clearance (c) 1.00 +/- 0.005 inch

Factors Nominal Value and Tolerance

Page 14: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

14

® ®

Improve:Improve the Cooling Fan Performance

Use Monte Carlo simulation to determineimpact of these variations

Mean = 882 ft3 / minStd Dev = + 2 ft3 / min

D M A I C

Page 15: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

15

® ®

Control:Control the Manufacturing and Installation

Use Statistical Process Control (SPC)techniques to monitor andcontrol manufacturing andinstallation of the fan

D M A I C

Page 16: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

16

® ®

Summary

Continued emphasis on quality initiatives

Data analysis increasingly becoming anintegral part of the engineering design process

MATLAB product family cansupport these initiatives

Page 17: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

©20

07 T

he M

athW

orks

, Inc

.

® ®

Questions?

Page 18: Kevin Cohan The MathWorks, Inc.€¦ · 8 ® ® Analyze: Analyze Factors that Affect Performance Gather additional data from existing system Modify our design factors and measure

©20

07 T

he M

athW

orks

, Inc

.

® ®