QTT Project Crease Stifness Problem (1)

QUALITY TOOLS AND TECHNIQUES PROJECT

TO ATTAIN CUSTOMER SATISFACTION, IMPROVE THE CREASE STIFNESS OF CIGARETTE HINGLED (HLs)

Page 1

Nazish Laraib

Page 2

TABLE OF CONTENTS

SIX SIGMA TOOLS

Define phase Project charter Deployment Map SIPOC Diagram KPIV’s KPOV’s Measure Phase Cause & Effect Diagram Sigma Level Box & Whisker Plot Process Capability Measurement System Analysis Testing Hypothesis Design Experiment Failure Mode and Effect Analysis

SIX SIGMA TOOLS FOR THE PROCESS OF FLAT CARTON CREASE STIFFNESS VALUES

Page 3

CREASE STIFFNESS: Crease stiffness is a part of Flat cartons unit’s product called Cigarette Hingled (HLs) used to store cigarettes, it is measured by sensor/transducer equipment by sensing crease bending force. There are many ups of die for paper creasing here they tested only 6 to 7 creases, Stiffness causes runnability problem on customer end. Lemanic Machine produces this problem chronically. After passing six different unit of printing, embossing, cutting and creasing units. It produce crease stiffness problem at C&C unit due to paper quality, die setting and humidity factors.

DIE SETTING: They have three plates for different HL’s product requirements counter plates, alternate counter plates and cito. Channel makes crease’s depth and grave, it also includes cutting and die ballades. A normal die setting can produce 25 million average HL’s. They manually adjust the die as per instructions. Die Crease pressure, machine speed, sheet displacement contributing elements of producing problem. According to operator, machine can produce 500 HL’s per minute.

BOARD TYPE: Usually they use three types of board but for HL’s right now they are using white coated bleech paper. Board Grammage and thickness affects paper stiffness.

DEFINE PHASE

PROJECT CHARTER DEPLOYMENT MAP SIPOC DIAGRAM KPIV’S & KPOV’S

PROJECT CHARTER

Page 4

Project Title:

To attain customer satisfaction, improve the crease stiffness of cigarette cartons, HL’s.

Business Case

Folding Carton line of Packages’ company produces Cigarette HL’s through rotogravure process by Lemanic and Riviera machines. It is being observed that Lemanic Machine produces beyond specification products that results customer objection due to runnability problem. Therefore it is required by ABC customer to improve crease stiffness specification. A single HL has seven creases, from 1 to 6 these are specified 11 Ncm/m to 18 Ncm/m and only 7 th crease has 8 Ncm/m to 14 Ncm/m.

Problem Statement:

Problem arises due to the component of paper board like grammage & thickness, and moisture of the production hall. It has to be improved that process should be lie within specification to achieve desired customer satisfaction.

Days

Cre

ase

va

lue

s

30272421181512963

20

19

18

17

16

Crease Values F1-F6April

Specification ( 11- 18) Ncm/m

Days

Cre

ase

Va

lue

s

30272421181512963

19

18

17

16

15

14

Crease Values F1-F6MARCH

Specification ( 11- 18) Ncm/m

Page 5

Days

Cre

ase

va

lue

s

30272421181512963

19

18

17

16

15

Crease Values F1-F6MAY

Specification ( 11- 18) Ncm/m

It is observed that values crossed the customer’s specification limits also indicate Process is not stable.

Object

To improve crease stiffness up to 14Ncm/m to 18 Ncm/m.

Metrics:

Primary Metric:

Crease Stiffness: Measure 6 to 7 creases of Hingled (HL’s) by using equipment of sensor or transducer unit of measure is Ncm/m.

Grammage: By using grammage balance with g/m2 unit

Thickness: Moisture %=( (A-B)/A)*100 degree centigrade.

Project Scope: Business Unit Folding Carton (BU-FC).

Page 6

DEPLOYMENT PROCESS FLOW DIAGRAM

SIPOC DIAGRAM:

Page 7

Supplier Input Process Output Customer

BSPPL Century

Wheat Straw& Waste Paper

Water Temperature

Paper board Folding Carton Unit of Packages.

Operator & Machine

Paper board Ink Solvent Varnish

Printing Sheet

Embossing Unit- 7

Operator & Machine

Pressure Embossing of requires image

Creasing& Cutting U-8

Operator & Machine

Board Creasing

Pressure Blade Cutter

HL’sXYZ Company

KPI’s and KPO’s:

Pulping Process

Ink Printing

Unit 1-6

Unit 7

Embossing

Cutting & creasing process

By setting die

Page 8

KPI’s X KPO’s YBoard Grammage X1 Crease Stiffness YBoard Thickness X2Moisture %age X3Machine Speed X4Operator X5Shifts X6

MEASURE PHASE

Cause & Effect diagram Sigma Level Box Plot Process Capability Measurement System Analysis

CAUSE & EFFECT DIAGRAM

Page 9

Page 10

MEASURE SIGMA LEVEL METRIC

2013 MARCH APRIL MAYTotal Production 95996397 100239720 40498500Defects 8877180 2721600 1209600Opportunities 23 23 23DPU 0.092474096 0.027150914 0.029867773DPO 2.1269042 0.624471018 0.686958776DPOM 2126904.2 624471.0181 686958.7763

Sigma Level 0 1 1.1

Flat carton(Lemanic) required improvement to reduce product variation and cost.

BOX WHISKER’S PLOT

Cre

ase

Ranges

ShiftsY2Y1

321321

19

18

17

16

15

14

13

12

Shifts123

15.59

16.1616.085

12.955

13.64

13.21

Boxplot of Y1, Y2 vs Shifts

Page 11

Remarks: For crease values F1 to F6 shifts 3 controlled die setting accurately as compare to others. (Considered only Y1)

Cre

ase

Values

OperatorY2Y1

ZahidJawadIsrarZahidJawadIsrar

19

18

17

16

15

14

13

12

OperatorIsrarJawadZahid

16.25

15.84516.1

13.30513.413.16

Boxplot of Y1, Y2 vs Operator

Remark: Jawad and performs consistently as compare to others operator Zahid need to improvement . (Considered only Y1)

Page 12

Remarks: Paper thickness at 283 and 288 cm affect the specification of crease values that are skewed.

Remarks: Variation is observed in values, that are less than 203 g/m squares for crease stiffness value

MACHINE SPEED (hl/ h)

45000037500030000022500015000075000

Median

Mean

340000330000320000310000300000290000

Anderson-Darling Normality Test

Variance 5635154651Skewness -1.26773Kurtosis 1.48407N 165

Minimum 53760

A-Squared

1st Quartile 258949Median 3255003rd Quartile 350350Maximum 483000

95% Confidence Interval for Mean

289330

7.40

312409

95% Confidence Interval for Median

315000 338565

95% Confidence Interval for StDev

67748 84174

P-Value < 0.005

Mean 300869StDev 75068

95% Confidence Intervals

Summary for Machine Speed

Page 13

Remarks: According to standard design speed of machine has to produce 382923 hl/hr. Average value tell that the speed is 300669 hl/hr. Even shape of the distribution is skewed.

PROCESS CAPABILITY OF CREASE STIFFNESS VALUES

F1 T0 F6

AFTER APPLYING BOX COX TRANSFORMATION:

Page 14

Page 15

Comments:

There is variation in paper crease stiffness values. Thus process is not capable according to the value of Cpk and Cp.

MEASUREMENT SYSTEM ANALYSIS

MINITAB offers several commands to help you determine how much of your process variation

arises from variation in your measurement system.May-13

Part Reference Value Crease Value

1 13 16.741 13 16.7531 13 17.6491 13 17.5251 13 19.017

Page 16

1 13 16.6471 13 18.0341 13 16.0681 13 18.2251 13 16.4372 14 16.4012 14 15.5322 14 16.8422 14 17.2162 14 16.782 14 15.6062 14 17.7522 14 15.9472 14 17.1272 14 15.1573 15 16.1953 15 15.4313 15 16.6863 15 16.263 15 18.213 15 15.6953 15 17.0363 15 16.1093 15 16.9743 15 15.2734 16 17.3814 16 16.5334 16 17.3784 16 17.9944 16 18.3444 16 16.5344 16 18.2694 16 17.064 16 18.194 16 17.2675 17 17.1165 17 15.7215 17 17.2145 17 17.475 17 17.8185 17 16.336

Page 17

5 17 18.1325 17 16.5765 17 17.8325 17 16.4616 18 18.466 18 16.0746 18 17.366 18 18.4126 18 16.9896 18 16.6856 18 19.1386 18 16.7876 18 18.3016 18 16.413

Page 18

Gage Linearity and Bias examines gage linearity and accuracy.In Gage bias section reference averages from 13 to 16 shows bias because values are less than 0.05.In Gage Linearity, slope value also less than 0.05 that shows gage is producing non linear results. (for good gage it should be linear).

GAUGE REPEATABILITY & REPRODUCIBILITY

Method for assessing repeatability and reproducibility:

One-Way ANOVA Table

Source DF SS MS F PPart 5 12.4862 2.49724 3.32420 0.011Repeatability 54 40.5664 0.75123Total 59 53.0527

Alpha to remove interaction term = 0.25

Gage R&R %Contribution Source VarComp (of VarComp)Total Gage R&R 0.751230 81.14

Repeatability 0.751230 81.14

Part-To-Part 0.174601 18.86

Total Variation 0.925832 100.00

Process tolerance = 7

Study Var %Study Var %ToleranceSource StdDev (SD) (6 * SD) (%SV) (SV/Toler)

Total Gage R&R 0.866736 5.20041 90.08 74.29

Repeatability 0.866736 5.20041 90.08 74.29

Part-To-Part 0.417853 2.50712 43.43 35.82

Total Variation 0.962201 5.77321 100.00 82.47

Page 19

%contribution Between 1% and 9% the measurement system is not acceptable and according to action group definition for gage acceptance it is considered poor.

Similarly % study variation Between 10% and 30% the measurement system is not acceptable according to action group definition for gage acceptance it is considered poor.

The percent contribution from Gage R&R ia larger than that of part to part, telling you that much of the variation is due to difference between gage R&R.

ANALYZE PHASE:

Testing Hypothesis

Page 20

Correlation & Regression Analysis

TESTING HYPOTHESIS

Crease values V’s Operator:

Comment: Less variation in Jawad performance.

Test for Equal Variances: Y versus Operator

95% Bonferroni confidence intervals for standard deviations

Page 21

Operator N Lower StDev Upper Israr 31 0.87415 1.14684 1.64223 Jawad 63 0.80205 0.97581 1.23806 Zahid 56 1.00041 1.23059 1.58710

Bartlett's Test (Normal Distribution)Test statistic = 3.18, p-value = 0.204

Levene's Test (Any Continuous Distribution)Test statistic = 0.26, p value = 0.770

One-way ANOVA: Y versus Operator

Null Hypothesis: Hο

µ(zahid)= µ(jawad) = µ(Israr)

Alternate Hypothesis: Ha

µ(zahid)≠µ(jawad)≠ µ(Israr)

One-way ANOVA: Y versus Operator

Source DF SS MS F POperator 2 1.00 0.50 0.41 0.668Error 147 181.78 1.24Total 149 182.78

S = 1.112 R-Sq = 0.55% R-Sq(adj) = 0.00%

Individual 95% CIs For Mean Based on Pooled StDevLevel N Mean StDev --+---------+---------+---------+-------Israr 31 16.701 1.147 (---------------*---------------)Jawad 63 16.505 0.976 (----------*----------)Zahid 56 16.493 1.231 (-----------*----------) --+---------+---------+---------+------- 16.25 16.50 16.75 17.00

Pooled StDev = 1.112

Inferential statistics results tell us that operators performance are almost equal because p-value is 0.668 that is greater than 0.05.

Page 22

Here histogram makes a bell shape curve pattern and residual follow a straight line thus normality assumption fulfilled.

Page 23

Test for Equal Variances: Y versus SHIFT

95% Bonferroni confidence intervals for standard deviations

SHIFT N Lower StDev UpperShift1 52 0.79187 0.98111 1.27930Shift2 55 1.00179 1.23441 1.59621Shift3 43 0.86764 1.09622 1.47390

Bartlett's Test (Normal Distribution)Test statistic = 2.73, p-value = 0.255

Levene's Test (Any Continuous Distribution)Test statistic = 0.53, p-value = 0.589

Page 24

One-way ANOVA: Y versus SHIFT

In Shift analysis, histogram makes a bell shape curve pattern and residual follow a straight line thus normality assumption fulfilled.

Page 25

Null Hypothesis: Hο

µ(shift 1)= µ(shift 2) = µ(shift 2)

Alternate Hypothesis: Ha

µ(shift 1)≠µ(shift 2)≠ µ(shift 2)

One-way ANOVA: Y versus SHIFT

Source DF SS MS F PSHIFT 2 0.94 0.47 0.38 0.685Error 147 181.85 1.24Total 149 182.78

S = 1.112 R-Sq = 0.51% R-Sq(adj) = 0.00%

Individual 95% CIs For Mean Based on Pooled StDevLevel N Mean StDev -------+---------+---------+---------+--Shift1 52 16.590 0.981 (------------*-----------)Shift2 55 16.592 1.234 (-----------*-----------)Shift3 43 16.417 1.096 (-------------*------------) -------+---------+---------+---------+-- 16.25 16.50 16.75 17.00

Page 26

Pooled StDev = 1.112

Accept Null hypothesis There isn’t any difference between shifts.

REGRESSION ANALYSIS

To find out the relationship between variables we apply multiple regression analysis technique and observe that which variable part huge effect on response value.

Y = Response variable (dependent variable) – crease value

X1 =Independent variable – Moisture

X2 = Board Grammage

X3 = Paper thickness

Regression Analysis: Y versus Moisture, Board Grammage, Paper Thickness1

The regression equation isY = 35.3 - 0.533 Moisture - 0.0564 Board Grammage - 0.0135 Paper Thickness1

Predictor Coef SE Coef T PConstant 35.324 7.073 4.99 0.000Moisture -0.5331 0.5400 -0.99 0.325Board Grammage -0.05635 0.02503 -2.25 0.026Paper Thickness1 -0.01345 0.01468 -0.92 0.361

S = 1.09139 R-Sq = 4.9% R-Sq(adj) = 2.9%

Analysis of Variance

Source DF SS MS F PRegression 3 8.881 2.960 2.49 0.063Residual Error 146 173.904 1.191Total 149 182.784

Source DF Seq SSMoisture 1 1.553Board Grammage 1 6.327Paper Thickness1 1 1.001

Unusual Observations

Page 27

Obs Moisture Y Fit SE Fit Residual St Resid 28 6.25 18.6800 16.3965 0.1094 2.2835 2.10R 55 6.10 19.5800 16.9273 0.1781 2.6527 2.46R 56 6.14 19.5800 16.9060 0.1707 2.6740 2.48R 58 6.40 19.0700 16.7674 0.1870 2.3026 2.14R 75 6.32 16.9000 16.9300 0.8389 -0.0300 -0.04 X 84 6.21 14.7200 16.9359 0.1773 -2.2159 -2.06R 95 5.70 16.2700 16.5072 0.3384 -0.2372 -0.23 X100 5.70 16.6400 16.8697 0.3089 -0.2297 -0.22 X102 6.06 18.8100 16.4094 0.2131 2.4006 2.24R103 5.68 16.6400 16.8265 0.3145 -0.1865 -0.18 X105 5.65 15.0100 16.8745 0.3445 -1.8645 -1.80 X120 6.79 15.5800 16.0145 0.3221 -0.4345 -0.42 X134 6.31 13.2400 16.0744 0.2050 -2.8344 -2.64R141 6.40 13.6100 16.2762 0.1413 -2.6662 -2.46R

R denotes an observation with a large standardized residual.X denotes an observation whose X value gives it large leverage.

Regression equation shows that there is relation between board grammage and crease value.

Thus board grammage affects crease value.

Page 28

IMPROVE PHASE:

Design of Experiment.

Failure Mode and Effect Analysis

DESIGN OF EXPERIMENT

DOE tell us that which input(variable) has effect on the crease values

Factors: Factors setting values.

Levels

1. Board Grammage: 204, 210 gram2. Paper Thickness: 285, 2913. Moisture: 6%-7%

Replicate=3

Response: Crease value

Full Factorial Design

Factors: 3 Base Design: 3, 8Runs: 24 Replicates: 3Blocks: 1 Center pts (total): 0

Generated data in minitab:

StdOrder

RunOrder

CenterPt

Blocks

Board Grammag

e

Paper Thicknes

sMoistur

e

Crease

value10 1 1 1 210 284 6 17.0018 2 1 1 210 284 6 17.4511 3 1 1 204 291 6 16.45

Page 29

4 4 1 1 210 291 6 15.5423 5 1 1 204 291 7 17.45

7 6 1 1 204 291 7 15.8721 7 1 1 204 284 7 18.0020 8 1 1 210 291 6 16.0014 9 1 1 210 284 7 15.0012 10 1 1 210 291 6 14.54

8 11 1 1 210 291 7 14.003 12 1 1 204 291 6 16.005 13 1 1 204 284 7 18.002 14 1 1 210 284 6 16.21

24 15 1 1 210 291 7 15.0019 16 1 1 204 291 6 14.2513 17 1 1 204 284 7 15.35

6 18 1 1 210 284 7 17.9517 19 1 1 204 284 6 16.00

1 20 1 1 204 284 6 15.0015 21 1 1 204 291 7 18.0022 22 1 1 210 284 7 15.0016 23 1 1 210 291 7 15.00

9 24 1 1 204 284 6 14.33

Factorial Fit: Crease value versus Board Gramma, Paper Thickness,Moisture

Estimated Effects and Coefficients for Crease value (coded units)

Term Effect Coef SE Coef T P

Constant 15.9746 0.2260 70.69 0.000

Board Grammage -0.5008 -0.2504 0.2260 -1.11 0.284

Paper Thickness -0.5992 -0.2996 0.2260 -1.33 0.204

Moisture 0.4875 0.2438 0.2260 1.08 0.297

Board Grammage*Paper Thickness -0.8225 -0.4113 0.2260 -1.82 0.088

Board Grammage*Moisture -1.2858 -0.6429 0.2260 -2.84 0.012

Paper Thickness*Moisture -0.0642 -0.0321 0.2260 -0.14 0.889

Board Grammage*Paper Thickness* 0.1692 0.0846 0.2260 0.37 0.713 Moisture

S = 1.10714 PRESS = 44.127 R-Sq = 49.55% R-Sq(pred) = 0.00% R-Sq(adj) = 27.48%

Page 30

It is being observed that paper grammage and moisture have impact on crease value of HLs

Main effect plot for crease value

Paper grammage and thickness have negative relation with crease value as both values increases crease values decreases. While moisture value increase crease value also increases.

Page 31

Interaction plot for crease value

Main effect plot for crease values shows interaction between paper grammage &paper thickness and Board grammage and moisture on the other side paper thickness and moisture are independent.

Page 32

Page 33

Pareto chart shows that the main effect on the curl values is being caused by Paper grammage and moisture combination

.

Page 34

FALIURE MODE AND EFFECT ANALYSIS

Please adjust the excel layout.

Page 35

References:

Packages Limited company (Quality control department)

Page 36

Page 37