Productivity Improvement of a Special Purpose Machine

Hindawi Publishing CorporationJournal of Quality and Reliability EngineeringVolume 2013 Article ID 752164 13 pageshttpdxdoiorg1011552013752164

Research ArticleProductivity Improvement of a Special Purpose MachineUsing DMAIC Principles A Case Study

Sunil Dambhare1 Siddhant Aphale2 Kiran Kakade2 Tejas Thote2 and Atul Borade3

1 Department of Mechanical Engineering PVPIT Bavdhan Pune Maharashtra 411021 India2Mechanical Engineering PVPIT Bavdhan Pune Maharashtra 411021 India3 Department of Mechanical Engineering JDIET Yavatmal Maharashtra 445001 India

Correspondence should be addressed to Atul Borade atulboraderediffmailcom

Received 5 April 2013 Revised 16 July 2013 Accepted 22 July 2013

Academic Editor Shey-Huei Sheu

Copyright copy 2013 Sunil Dambhare et alThis is an open access article distributed under theCreativeCommonsAttribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Six Sigma is one of the popular methodologies used by the companies to improve the quality and productivity It uses a detailedanalysis of the process to determine the causes of the problem and proposes a successful improvement Various approaches areadopted while following Six Sigmamethodologies and one of them is DMAICThe successful implementation of DMAIC and FTAis discussed in this paper In this study the major problem was of continuous rework up to 16 which was leading to wastage ofman hours and labor cost Initially fault tree analysis (FTA) was used to detect the key process input variables (KPIVs) affecting theoutput Multivariable regression analysis was performed to know the possible relationship between the KPIVs and the output TheDMAIC methodology was successfully implemented to reduce the rework from 16 bores per month to 220 bores per monthThe other problem of nonuniform step bores was also reduced significantly

1 Introduction

Diesel engines have very wide applications in this moderntechnological era They are designed to cater for the needof construction mining power generation locomotivesmarine transport compressors and so forth market seg-ments These are heavy power requirement applications Forsuch applications the output requirement may vary from 10horsepower to 3500 horsepower The engine with this hugepower output is also huge The firm in which this studywas conducted was involved in manufacturing 1 V- and 16V-cylinder diesel engines The sand casted engine block wentthrough various operations and then was assembled Oper-ations like undercutting water clearance chamfer boringstep-boring and so forth were performed on each cylinderThe step-boring operation was the most difficult operation Itneeds special attention as the liner is resting in it A cylinderliner is pressed into an engine block and houses the pistonThe cylinder liner is much harder than the engine block andprevents the piston from wearing out through the cylinderbore Typically used in aluminium engine blocks and diesel

engines the cylinder liner is either pressed into position orheld in place by the cylinder head In large engines such asthe engines found in diesel locomotives the cylinder liner ispart of an assembly containing a new piston piston rings anda connecting rod During scheduled maintenance or repairthe liner is changed as a complete unit In aluminium engineblocks the block material is too soft to house a piston Thefriction of a piston moving up and down inside the alloyblock would make the piston wear out resulting in a lossof compression and severe oil consumption [1 2] In suchcases the steel cylinder liner is pressed into the engine blockand then the engine block is machined to assure that thecylinder head mating surface is smooth and flat With thismodification the engine is able to operate for many yearswithout failureThe flat surface resulting from the machiningof the engine block assures a proper seal of the head betweenthe cylinder head and the engine block An improperly sealedhead gasket will result in overheating of the engine lossof power and the potential to ruin the block and cylinderhead It is clear that the depth of a V-cylinder engine isone of the most crucial parameters for the efficient working

2 Journal of Quality and Reliability Engineering

of an engine A slight variation in depth of the step candamage the cylinder and piston as well as hamper the engineoperation A cylinderrsquos head and liner rest on the bored depthThus maintaining the depth is the necessity for any enginemanufacturing firm

Six Sigma is a business management strategy first devel-oped byMotorola in 1986 which seeks to improve the qualityof process outputs [3] It identifies and removes the cause ofdefects andminimizes variability in manufacturing and busi-ness processes [4 5] A set of quality management methodsand statisticalmethods is used Each Six Sigmaproject carriedoutwithin an organization follows a defined sequence of stepsand quantified financial targets According to the definition aSix Sigma process is one in which 9999966 of the productsmanufactured are statistically expected to be free of defects(34 defects per million) [6ndash8] Six Sigma uses a group ofimprovement specialists for problem solving and improv-ing the process continuously [9 10] Six Sigma techniqueshave two main methodologies DMAIC and DMADV [11]Define Measure Analyse Improve and Control (DMAIC)methodology was followed for reducing the rework [12 13]The reasons for the main problem can be detected usingagain two methodologies FTA and failure mode and effectsanalysis FTA deals with identifying all possible causes relatedto a particular problem It is a stepwise approach to identifycauses and all the parameters related to every cause It can beused in almost every application which involves cause-effectanalysis Once the problem is defined it is to be measuredso as to collect statistical data about the problem Once thestatistical data is collected it is analysed using various analysistechniques like chi-square test regression analysis ANOVAand so forth Minitab statistical software was used for theanalysis of various phases of the project

2 Review of the Literature

Six Sigma methodologies have become a top agenda formany companies which are continuously trying to improveproductivity at lesser costs It was shown previously that SixSigma continued to be a predominate target to try and obtaina competitive advantage [11] Many Fortune 500 companieshave adopted Six Sigma to improve the productivity andreduce cost Six Sigma has been described as a data drivenapproach for problem solving business process disciplinedstatistical approach and a management strategy [14ndash18] SixSigma methods can prove to be beneficial when applied tolabour-intensive repeatable processes according to Swinkand Jacobs [14] Six Sigma benefits are significantly cor-related with intensity in manufacturing and with financialperformance before adoption in services [14] The Six Sigmaimprovementmethod is problem-focused and itsmain objec-tives are decreasing scrap earning income and creating valuePrevious researches have shown the effective implementationof Six Sigma techniques in problem solving and processimprovements FTA was successfully utilised to analyse thebridge erecting tripping problem [19] FTA can be used toanalyse hazards and calculate system reliability for simple aswell as complex systems [20] Six Sigma methodologies havebeen previously used for reducing the rejection level [21]

FTA can be effectively used for finding the effective causesfrom accident cases to scrap reduction [22] According toWang et al [23] FTA is a simple effective reliable methodwhich is recognised internationally and is used on guidingsystem optimization and analysing and repairing the systemof weak links FTA helps deepen the research to find out thepossible causes for the fault [24] FTA has previously beenused successfully to establish priorities for themanufacturingplant for future projects aimed at improving the manufactur-ing plant [25]

According to Buyukozkan and Ozturkcan [4] and deKoning and de Mast [26] Six Sigma program offers a widerange of tools and techniques which might be statisticaland nonstatistical that are intended to assist the projectleader Swink and Jacobs [14] showed solid support for thehypothesis that Six Sigma adoption tends to produce signif-icant benefits for firmrsquos profitability Positive return on asset(ROA) changes were frequently observed in latter periods(years +3 and +4) and moreover these benefits also appearto be persistent These findings hint at potential differencesin how Six Sigma programs are possibly being applied infront-office versus back-office contexts Findings suggest thatSix Sigma methods may be most beneficial when applied tolabour-intensive repeatable processes However less labour-intensive quality experienced manufacturing firms will notexperience the profit impact from Six Sigma adoption thesame way that others will Their outcomes also revealedmarginally significant positive effects on sales growth SixSigma projects can accomplish successfully using FTA FTAhas been used extensively by the military the space programand nuclear industry FTA is a very structured systematicand rigorous that lends itself well to quantification [25] FTAcan be used to express the logical relationship between thepossibility of certain accident and causes of undesired eventsor accidents in fault tree diagram [19] It could be successfullyimplemented to find out all the possible causes for the prob-lem or fault that has occurred According to Shalev and Tiran[20] FTA method analysts apply top-down logic in buildingtheir models The problem of reducing process variation andthe associated defect rate can be solved using the DMAICmethodology of Six Sigma Six Sigma is a useful problem-solving methodology and provides a valuable measurementapproach Six Sigma focuses on some vital dimension of busi-ness processes reducing the variation around the mean valueof the process [20] The original task of Six Sigmarsquos DMAICMethodology is variation reduction As stated by deMast andLokkerbol [27] Six Sigma and itsDMAICare built on insightsfrom the quality engineering field incorporating ideas fromstatistical quality control total quality management andTaguchirsquos offline quality control It has also been used for gen-eral tasks like quality improvement efficiency improvementcost reduction and other pursuits in operations Thus SixSigma is a generic method and its original task domain wasvariation reduction typically in manufacturing processesLi et al [28] successfully implemented DMAIC approach toimprove the capability of the solder paste printing processby reducing thickness variations from a nominal valueThe DMAIC approach has shown a wider application andhow the engineering organisation can achieve competitive

Journal of Quality and Reliability Engineering 3

Figure 1 Ingersoll special purpose machine and engine block

advantages efficient decision making and problem-solvingcapabilities within a business context According to Li andAl-Refaie [29] adopting theDMAICprocedure includingGRampRstudy turns out to be an effective method in improving thequality system including measurements

3 Background for the Study

The case study was conducted at a leading manufacturerof 12 V- and 16V-cylinder diesel engines The sand castedengine block is processed with operations like rough boringwater clearance chamfer surface milling finish boring andso forth before accomplishing the engine assembly Criticaloperations which demand precise dimension control areperformed on special purposemachines Engine block boringis one of the critical operations performed on special purposemachines under study

The finished boring operation is completed in 3 stagesInitially the V-surface of the engine is milled using millinghead Considering the milled surface as datum step-boringoperation and water clearance chamfering operation areperformed As the engine is a V-cylinder engine performingmachining operations on the block is a tedious task If theseoperations are performed on a CNC machine it requiresnearly 5 hours and tiresome programming Thus there wasa need of special purpose machine

Ingersoll machine performs the mentioned operationsin 17 minutes approximately The machine uses hydrauliccircuits for performing operations which were designedaround 40 years back The machine performs three opera-tions on each bore namely undercutting chamfering andstep-boring after performingmilling on theV-surface Oper-ations are carried out in a sequence as undercutchamferingstep-boringMachine starts its forward stroke at 800 psi pres-sure acting in forward direction In this stroke it performs twooperations undercutting and chamfering Both operationsare carried out at 800 psi pressure and flow control valves areused to control speed of sliding tool during operations Aftercompletion of chamfering tool slide completes the forwardstoke and at the same time a lever attached to the tool posttouches the inclined milling surface actuating the boringtool slide mechanism This provides forward motion to theboring tool Simultaneously the lever actuates return pressurevalve and 1000 psi pressure acts to carry out backward strokeof the tool post slide Step-boring finishing operation is

completed in return stroke of tool postOnce tool post returnsto its original position another engine bore gets lined for thesame set of operations Figure 1 shows the Ingersoll specialpurpose machine and the engine block machined using thismachine

4 Case Study

The machine under consideration is a special purposemachine which was specially developed for performing thespecified operations on the V-engines manufactured by thefirm The operations performed are milling of the V-surfacewhere the material on the surface is removed After millingcrevice chamfering is completed Step-boring is performedat last in which the existing hole is enlarged to the designeddepth The step bore is a critical dimension as the liner of thecylinder rests on it The boring tool performs the operationand after it reaches the required depth a lever senses itsposition As soon as the lever touches the surface the boringtool retracts and the operation is completed

During the study readings for the step bore depth werenoted for every bore of each engine block The allowabletolerance for the step bore was 0719010158401015840 plusmn 0001310158401015840 From thedata it was observed that there was a variation in the readingsand many of the readings were outside the tolerance limitof 0719010158401015840 plusmn 0001310158401015840 This was a serious problem Bringingthe dimension within the tolerance limits means reworkon the bores was necessary The rework data was gatheredand it was found that average rework per month was 16of the bores machined For performing rework operations369 man-hours per month and approximately INR 60400were spent This hampered the productivity of the firm Thestudy showed that rework was needed mostly on the leftbank of the engine block indicating need of the improvementactions on this section of engine boring operation Thereforethe objective of the study was set to minimize the reworkpercentage per month close to zero without affecting theoperations and cycle time To improve the productivity andreduce the rework expenses the Six Sigma technique wasselected

5 Methodology

As the study aimed at improving the existing businessprocess DMAICmethodology was considered [11 12 21 30]


Figure 2 Step bore and depth variation

It consists of phases namely Define Measure AnalyseImprove and Control [31 32] The whole Six Sigma projectstarts withDefine phase and is defined based on the customerrequirement and company strategy and mission [33 34]Measure phase helps the project team to refine the problemand begin the search for various causes of the failure InAnalyse phase the causes found are analysed using variousdata analysis tools and the data is validated for Improvementphase Improvement phase helps in finding solutions andimplementing them so that the problems can be elimi-nated In Control phase the performance of the processafter Improvement is measured routinely and accordinglyadjustments are made in operations If the Control phase isnot implemented it may revert the project to its previousstate

In the case study presented the DMAIC methodologywas applied to identify the probable sources of deviation inmachined surface and successfully reduced the rework to220 from an initial 16 per month The following sectionsexplain the methodology applied for the purpose

51 Step ImdashDefine

Problem Statement Reduce engine block liner bore counterdepth rework close to zero from 16bores permonthwithoutadversely affecting the cycle time

The special purpose machine was in regular use withheavy production for a long time Due to the continuouscourse of action and heavy load parts of the machine areworn out Thus a variation in the depth of step bore wasobserved as shown in Figure 1 This variation occurred ona number of blocks leading to increased rework The majorconcern was the unpredictable behaviour of the machineEach V-block has two sides left bank and right bank whenlooked to from the rear end of the engine The data showedthat the majority of the rework was required on the left bankof the block If there is a variation in the depth of the boreeither if it can be reworked or if the depth is out of reworkrange then the entire block is scrapped Reworking of thecylinder bore is possible at the expense of 369 man-hours permonth and approximately INR 60400

The rework demands skilled manpower due to precisetolerances man-machine-hours and other considerableresources Thus to increase productivity and reduce therework cost there was a need to reduce this rework Itis expected that the depth of each bore must lie within

0717710158401015840 to 0720310158401015840 Based on the lower limit and upper limit

of the bore they are categorised as undersize in size oroversize bores Each bore failing to achieve these toleranceswas subjected to reworkThe depth of the bore was measuredusing a depth gauge which was calibrated before data wascollected The depth gauge indicated the measured depthabout the mean depth that is 0719010158401015840 The depth wasmeasured at two points of the bore upside and downside ofthe bore on both banks as shown in Figure 2

The machine behaviour was unpredictable because of thefollowing reasons

(1) Dimensional variation was observed mostly on theleft bank of the block even under the samemachiningconditions on both sides

(2) A fix pattern in dimensional variation was notobserved in the finished bores Some blocks wereoversize while some were undersize Some cases werereported where all the bore categories were involved

(3) There was no pattern repetition in dimensional varia-tion of the bores If on a particular block all the boreswent out of tolerance then for the immediate nextblock it could be a block without any fault

The following objectives were set to achieve the target

(1) to reduce rework of bores from 16 bores per monthclose to zero without adversely affecting the cycletime

(2) to improve the overall quality of the process(3) to reduce the energy consumption involved in the

process by reducing the rework(4) to reduce the cost of rework

52 Step IImdashMeasure In the proposed study a variation inthe depth of step bore was observed on the blocks used forV 12- and V 16-engines These variations were not uniformand of same pattern

The detailed data for total number of bores producedfrom themonth ofApril 2012 toDecember 2012was collectedThe up and downmeasurements of both banks were recordedfor six months The measurements for the engine bore whichwere not in the specified tolerances were also counted in allThe total number of bores produced per month was countedand accordingly the rework percentage was found out byplotting the I-MR chart as shown in Figure 3Theboreswhichrequired rework were classified into two major categoriesoversize bores and undersize bores These two categorieswere again divided into two subcategories depending on thebanks of the block where variation was recorded that is leftbank and right bank For the collected data individual valueand moving range chart (I-MR) was plotted using Minitab16 software I-MR chart plots individual observations onone chart accompanied with another chart of the range ofthe individual observations normally from each consecutivedata point Figure 3 is the I-MR chart of the rework datacollected during AprilndashDecember 2012


DecNovOctSepAugJulJunMayApr

302010

0

Observation


201510

50

Observation

I-MR chart of no of bores reworked

Indi

vidu

al v

alue

Mov

ing

rang

e

UCL = 3234

LCL = minus014

UCL = 1996

LCL = 0

X = 1610

MR = 611

Figure 3 I-MR chart of rework data

It can be inferred from Figure 3 that the average monthlyblock liner bores reworked for DC are 16 of total produc-tion Each undersize bore required 10 minutes of reworktime and each oversize bore required 60 minutes of reworktime The manual rework cost incurred per bore whetheroversize or undersize was INR 233 Accordingly eliminatingrework would save monthly 369 man-hours and INR 60400It also saved average sleeve rework cost of INR 30000 permonth Hence the total average monthly cost saving could beINR 90400 The projected annual cost saving could be INR1084800 or USD 193700 approximately

53 Step IIImdashAnalyse The Analyse phase is the third andusually the longest phase in the Six SigmamethodologyMostof the crucial data analysis is performed in this phase Thiseventually leads you to isolate the root causes of the problemand provides insight into how to eliminate them

The operational working of the machine was consideredfor the FTA FTA is not a cause and effect diagram FTAcan be used when the problem has already occurred in thecurrent business process As the case of the project was ofcurrent business process FTA was used instead of FMEAFMEA or failure mode and effect analysis is used for ldquowhatcan happenrdquo whereas FTA is used for ldquowhat has happenedrdquoFTA is amethod to analyse a failuremode in order to identifypossible assignable causes and find the failure mechanism[25] FTA connects failure mode to assignable causes

In this case study the fault tree was started from thedefinition of problem and then it was directed to primarycauses and secondary causesThis procedure was followed tillall possible causes were listed FTA provided all areas to beimproved in single view and helped in stepwise analysis Thecritical parameters were segregated from experience of thepersons using the machine and further analysis was carriedout on these key input parameters Figure 4 represents thefault tree drawn for the case study

Factors that were considered the most influential keyinputs are shown in Figure 5

Once the key inputs were obtained from the FTA therewas a need to check the reliability of all the readings taken

by the operators This was done by performing measurementsystem analysis Three inspectors measured two blocks sepa-rately once in a serial order and then in a randomorderThesereadings were analysed using Minitab software to check thegage reproducibility and repeatability [6] Figure 6 was theoutcome of the measurement system analysis

From the above results around 90 confidence level wasobtained Thus there was no error in measurement systemand now all the readings can be called Data

The third objective of this phase was to find out how theyare relatedThe continuous key inputs namely slide pressurelever pressure ambient temperature and oil temperaturewere analysed using one of the multivariable regression anal-yses Tests that can be used in this phase are regressioncorrelation analysis of variance hypothesis testing 119905-testschi-squared tests graphical analyses GLM logistic regres-sion and so forthThese tests come underMulti-Vari StudiesBefore proceeding to select the test type of data wasanalysed

Multi-Vari analysis is a graphical tool which throughlogical subgrouping analyzes the effects of categorical Xrsquos oncontinuous Y rsquos The graphical results of Multi-Vari analysiscan be quantified using nested analysis of variance

Multi-Vari was chosen because of the following reasons

(1) to determine with high statistical confidence thecapability of the KPOVs of a process

(2) to identify assignable causes of variability(3) to obtain initial components of variability (shift-to-

shift run-to-run and operator-to-operator)(4) to get a first look at process stability over time(5) to provide direction and input for design of experi-

ments (DOE) activities

Selection of Test to Be PerformedThe selection of test dependson the type of data whether it is continuous or discrete singleinputs or multiple inputs single outputs or multiple outputsand so forth In this study depth variation that is Y and allthe Xrsquos were continuous so multiple regression analysis wasperformed as seen from Figure 7

The general equation of approach was

119910 = 119891 (1199091 1199092 1199093 119909

119896) (1)

Depending on the key inputs obtained from the FTA thedata was sorted into continuous and discrete data Multi-Vari regression analysis was performed for the continuousdataThe key inputs varying continuously with time includedslide pressure lever pressure ambient temperature and theoil temperature As the data was categorized as continuousthe data collection was done depending on time Data wascollected during all three shiftsThe slide pressure lever pres-sure oil temperature and ambient temperature were notedfor every bore Per shift 2 engine blocks were considered forthis data collectionThe data was collected at the start of eachshift and at the end of each shift

Before performing the regression analysis null hypothe-sis (H

0)was set Null hypothesis (H

0) is equal to the specified


Variation in DC depth

Error in level sensor

Vibrations Wear out and tear

Variation in time lag

Variable acting pressure

Leakages in lever line

Damage to cylinder

Piston and cylinder wall wear out

Piston ring and seal wear out

Frequent use

Friction

Insufficient pressure built up

Decrease in pump efficiency

Less pressure developed

Leakages

Faulty lines and valves

Loose connections

Filter malfunctioning

Environmental factors

Temperature variation

Dust

Lack of cleanliness

Improper filing disposal

Tools

Defects in tool holding devices

Misfit in tool slide mechanism

Unwanted stresses due to delay in insert

change

Misalignment in tool post

Angular misalignment in milling and boring

Wear out and tear in

pump parts

Figure 4 FTA


DC depth variation

Ambient

temperature

Leakages in circuit lines

Oil temperatureSlide pressure

Error in milling surface

Lever pressure

(ie 800psi)

(ie 100psi)

Figure 5 FTA key inputs

value or parameter from another population Alternativehypothesis (Ha) is not equal to the specified value or parame-ter fromanother population119875 value is the value used to rejector fail to reject the null hypothesis Α is the probability thattrue null hypothesis is rejected

If 119875 le 120572mdashReject H0

If 119875 gt 120572mdashFail to reject H0

(2)

The statistical analysis is done with the development of atheory null hypothesis The analysis will ldquofail to rejectrdquo orldquorejectrdquo the theory

Null Hypothesis (H0) data are independent (not

related)Alternative Hypothesis (Ha) data are dependent(related)If the 119875 value is ge005 then accept the H

0(no

statistical relationship)If the 119875 value is lt005 then reject H

0(a statistical

relationship exists)

According to this theory it was assumed that the inputsambient temperature oil temperature slide pressure andlever pressure were not affecting the process that is the Xand Y are not relatedThus it was called null hypothesis (H

0)

Once the null hypothesis was set the very first step was tofind the correlations between each of the four inputs thatis ambient temperature oil temperature slide pressure andlever pressure Figure 7 shows the correlation results providedby Minitab software From Table 1 it can be easily seen thatcorrelation exists only between ambient temperature and oiltemperature Thus for regression slide pressure and leverpressure can be neglected

Once the correlation test was done the next step wasto perform multiple regression using the terms obtainedfrom correlation test that is oil temperature and ambienttemperature

For the left bank themultiple regression failedThe resultsand residual plots obtained are shown in Figure 8 andTable 2

Table 1 Correlation results fromMinitab

Correlations amb temp oil temp slide pr lever prAmb temp Oil temp Slide pr

Correlation for left bankOil temp 0790Slide pr lowast lowast

Lever pr lowast lowast lowast

Correlation for right bankOil temp 0091Slide pr lowast lowast


Cell Contents Pearson correlationlowastAll values in column are identical

Table 2 Minitab results for regression LB

Regression analysis down versus amb temp oil tempThe regression equation is

down = 0712 + 0000077 amb temp + 0000126 oil temp21 cases used 1 case contain missing values

Predictor Coef SE Coef 119879 119875

Constant 0712296 0002422 29407 0000Amb temp 000007696 000003924 196 0066Oil temp 000012577 000009384 134 0197119878 = 0000293334 119877-Sq = 600 119877-Sq (adj) = 555

It can be observed from Table 2 that the 119875 value is zerothat is le005 Thus null hypothesis (H

0) is accepted But the

variance that is R-Sq value is just 60The R-Sq value mustbe at least 80 for multiple regressions to be successful

Figure 9 and Table 3 show the residual plots and resultsfor Right Bank of the engine block It can be seen that theR-Sqvalue is very low for right bank which is just 39Thus theseresults are strictly rejected The multiple regression analysisfor left as well as right banks is not successful

The null hypothesis (H0) set that there is no relation

between the Xrsquos and the Y was true Thus we fail to reject


Figure 6 MSA plots

Multiple XrsquosX data

X data X dataSingle X

DiscreteDiscrete ContinuousContinuous

Multiplelogistic logisticregression regression

2 3 4 waymiddot middot middotANOVA

Medians tests

Multipleregression

Multivariate analysis(Note this is not the same as Multi-Vari charts)

MultipleLogistic

Regression

regressionDisc

rete

Con

tinuo

us

Disc

rete

Con

tinuo

us

Y d

ata

Sing

le Y

Mul

tiple

Yrsquos

Y d

ata

Y d

ata

Chi-square

meansmedians tests

ANOVA

Figure 7 Selection of test for analysis [35]


000100000500000

99

90

50

10

1

Residual071900071875071850071825071800

Fitted value

8

6

4

2

0

Residual

222018161412108642

00005

00000

Observation order

Normal probability plot Versus fits

Histogram Versus order

Residual plots for LB down

minus00004

minus00002

00000

00002

00004

00006

00008

minus00005

Resid

ual

Resid

ual

Freq

uenc

y(

)

minus00005

00005

00000

minus00005

Figure 8 Residual plots of LB

000080000400000

99

90

50

10

1

Residual

071880071865071850071835071820

00005

00000

Fitted value

000040000200000

48

36

24

12

00

Residual222018161412108642

00005

00000

Observation order



Residual plots for RB down

minus00004minus00008

()

Resid

ual

Freq

uenc

y

minus00002minus00004minus00006

minus00005

minus00005

Resid

ual

Figure 9 Residual plots for right bank


Table 3 Minitab results for multiple regression of RB


down = 0718 + 0000097 amb temp minus 0000060 oil tempPredictor Coef SE Coef 119879 119875

Constant 0718100 0003310 21697 0000Amb temp 000009655 000002780 347 0003Oil temp minus000005955 000009589 minus062 0542119878 = 0000320196 119877-Sq = 390 119877-Sq (adj) = 326

the null hypothesis (H0) It was concluded that the four key

process input variables do not really dominate in contributionto key process output variable Hence some other parameterswere dominating the output This could be easily seen fromthe regression equation as well as the residual plot Two rifleshot inputs were obtained from the FTA namely leakages inthe hydraulic circuit and errors in milling surface

54 Step IVmdashImprovement As the hydraulic circuit workedwith high pressure of 800ndash1000 psi leakage was the factorresponsible for pressure drop in the circuit Hydraulic circuitwas of very complex nature involving many directionalcontrol valves and many pressure switches There were manypressure switching actions causing stress on various jointsThese leakages were creating problem to maintain pressureLeakages were observed on both banks of the machine Aspressure maintaining was critical all leakages were removedThese leakages were removed by cleaning all the pipes andvalves in the circuit and changing pipes which were cut Afterremoving the leakages the rework percentage dropped but thechange was not significant Thus it was decided to check formilling surface error and remove if any

Boring operation needs to be performed precisely as itcan go wrong very easily It requires precise alignment withdrilled hole as well as surface on which hole is drilled Align-ment with drilled hole never creates a problem Alignmentwith surface was another critical issue involved especiallywhen the surface was inclined Surface alignment with boringtool was perfectly perpendicular when the machine wasmanufactured But in course of time due to vibrations andother undesirable actions misalignment was produced in themilling head that is milling surface and boring tool whichcreated undesirable difference between up and down depthsof step bore On performing the analysis it was observed thatthe difference between up and down readings was 0004010158401015840which was almost 40 of total allowable tolerance Thus asleeve of 0004010158401015840 was manufactured and inserted behind themilling head The front view of sleeve is shown in Figure 10where ldquo119887rdquo = 0004010158401015840

This solved the problem significantly and the results wereproven by plotting I-MR charts for rework

After making the suggested improvements the reworkdata was collected similarly as collected in Define phase TheI-MR chart was plotted and both charts were compared tostudy the results obtained before and after making improve-ments

120579

(a)

(b)

Figure 10 Milling head slide sleeve front view

The I-MR chart of the revised rework for 3 monthsis shown below It can be seen that the rework has beenreduced to approximately 0 in March 2013 The value ofmean rework for the months of April 2012ndashDecember 2012was 16 After doing the improvements to the machine therework reduced continuously from January 2013 In Januarythe rework was 422 For the month of February 2013 therework was further reduced to 233 The main objectivewas achieved in March 2013The percentage rework droppeddrastically to 033 Thus the target to make rework close tozero was successfully achieved

At the beginning there was a lot of variation in the depthsof bores of a single blockWith reduction in rework the otheraim was to reduce this variation in the depth of the boresThis variation in depthswas nonuniformAs discussed earlierthe variation for a bore may go oversize and the very nextbore would be undersize On completion of the ImprovementPhase the box plot was plotted to compare the moving ragesof the depths of the blocks which is shown in Figure 12It could be observed that the moving range in March 2012varied from 0717010158401015840 to 0721010158401015840 for a block In November2012 thismoving rangewas decreased It varied from 0718310158401015840to 0720510158401015840 Thus the improvement could be seen In March2013 this range was drastically decreased and the new rangevariation was between 0718710158401015840 and 0719510158401015840 thus making thevariation uniform within less moving range

55 Step VmdashControl After completing the Improve phasefactors affecting the depth variation of the step bore wereproposed The actions proposed were implemented in themanufacturing process The results of these improvementswere monitored in Control phase A control plan was pre-pared which is the major action of this phase This controlplan consisted of all the actions that were proposed fordecreasing the rework of the blocks It included trainingand certifying the operators employees maintenance planpreparation regular inspection and preparation of controlcharts And thus from Figure 11 it can be observed that thegoal set of reducing the rework to zero percent was achieved

6 Results

The case study was carried out on a special purpose machinedeveloped by Ingersoll The machine was in continuous


MarFebJanDecNovOctSepAugJulJunMayApr

302010

0

Observation

Define Improvement

201510

50

Define Improvement

I-MR chart of no of bores reworked by project phase

Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210

Figure 11 I-MR chart after improvement

production for last 40 years The hydraulic circuit compo-nents seals hoses were worn out resulting in the inefficientworking of the machine This was leading to high percentagerejection of the engine blocks The study was carried outin phases and the principles of DMAIC were proved to beuseful for reducing the rework rate and hence improving theproductivity of themachine As themachinewas in operationand there were many factors contributing to deformation ofthe surface it was difficult to carry out the experimentationfor finding the reasons for the rework Hence the FTA wasselected for the purpose

At the first stage the goals were set to reduce the reworkfrom 16 to theminimumpossible value as the cost of reworkwas very high and small deviation in work could reject theentire engine block Later at Measurement phase the actualmeasurement of deviation was carried out It was found thatthe left bank of the enginewas prone to deviation as comparedto the right bank Hence concentration was focused onthis part of the engine block The analysis of the deviationwas an important issue Probable reasons for the deviationwere listed and categorized and the FTA was performedAfter discussion with the experienced staff actually workingon the machine the principle factors contributing more tothe deviation were identified for the study The key inputsthus obtained from the FTA were needed to be checkedfor the reliability of all the readings taken by the operatorsThis was done by performing measurement system analysisThe results are shown in Figure 6 The multivariable regres-sion analysis was performed to understand the relationshipbetween the parameters Figure 8 shows the residual plotsfor the left bank from Table 2 it was observed that the R-Sqvalue is just 60 for the left bank Figure 9 shows residualplots for right bank and fromTable 3 it was observed that theR-Sq value is just 39 For the multivariable regression testto be successful the R-Sq value must be at least 80 Thus itwas found that the four key process input variables were not

0720

0719

0721

0718

0717

Depth variation inMarch 2013

Depth variation inNov 2012


Dat

a

Box plot of depth variation

Figure 12 Box plot for depth variation

dominating the key process output variable on both banksThus some other parameters were dominating the outputFTA provided the inputs namely leakages in the hydrauliccircuit and errors in locating milling surface which were thenstudied for the performance

The entiremachine was operating on a complex hydrauliccircuit with an oil pressure in the range of 800ndash1000 psi Theleakages in the hydraulic circuit were traced and removedThe results of this step showed improvement in the processbut were not significant Another parameter was the millinghead location error It was removed by inserting a sleeveas shown in Figure 11 This time the efforts worked and therework was reduced drastically close to 220 on averagefor 3 months A control plan for all the contributing factorswas prepared for reducing the rework It included trainingand certifying the operators employees maintenance planpreparation regular inspection and preparation of controlcharts for further reduction in the rework Figure 12 shows


the improvement in the process after the DMAIC wassuccessfully implemented for the machine

7 Conclusion

Industries have to deal with a host of problems related toquality control Substandard quality hampers the productiv-ity of the plant which directly affects the company targetsOrganizations have to suffer huge losses which are not easy tocope upwithThus there is a need to improve the productivitysimultaneously keeping in mind the quality of the productSix Sigma can be effectively applied and the existing businessprocesses can be improved and made error free Six Sigmaprovides statistical proof to each and every action thushelping making decisions more efficient It can work evenwith less number of readings in the database Thus Six Sigmais completely an industry oriented methodology of qualityand productivity improvement

In the presented case study the rework percentage wasmuch higher that is 16The firm had to sustain the reworkcost and the man-hours required in the reworking decreasedthe productivity Establishing the relationship between theinput parameters and the output parameter is a challenge in acomplex system like the one discussed above The decisionof using Six Sigma methodology proved to be facile FTAwas implemented to find all the key inputs that are affectingthe output The KPIVs were categorised as continuous anddiscrete depending on their property whether they vary withtime or not and were analysed using multivariable regressionanalysis The Multi-vari regression analysis proved that theselected continuous parameters were not dominating theoutput failing to reject the null hypothesis (H

0) that is these

input variables did not affect the output Hence the rifle shotinput parameters leakages in hydraulic circuit and millingsurface errors were checked Leakages were observed in thecircuit for the left bank These leakages were removed Themilling surface error was removed by inserting a sleeve asshown in Figure 10 After these errors were removed therework reduced to 220 per month thus achieving the setgoalThe rework time of 369 hours permonth was reduced to42 man-hours per month The cost of rework was reduced toUSD 3500 per year Thus there was significant improvementin the productivity and losses the firm incurred

It is thus concluded that Six Sigma methodologies couldbe applied successfully in small firms Practitioners couldrefer this case study and implement it in a similar kind ofstudy With the help of case study we try to prove that SixSigma tools help to reduce the wastage and help improvequality of product In this study we had considered limitedparameters that affected the output Depending upon theexpertrsquos experience the study and improvements were doneon milling surface errors and leakages By performing exper-imentation on other parameters from the KPIVs furtherimprovements could be done This would reduce the reworkcloser to zero in the future DOE can be planned for theremaining KPIVs which can provide detailed effects onthe output

Abbreviations

DMAIC Define Measure Analyse Improve and ControlDMADV Define Measure Analyse Design and VerifyROA Return on assetFTA Fault tree analysisFMEA Failure mode and effect analysisMSA Measurement system analysisDOE Design of experimentsH0 Null hypothesis

Ha Alternative hypothesisR-Sq R-squared valueKPIV Key process input variablesKPOV Key process output variable

References

[1] Andrew Spencer Optimising Surface Textures for CombustionEngine Cylinder Liners Lulea University of Technology 2010

[2] ldquoWhat is Cylinder Linerrdquo httpwwwwisegeekcomwhat-is-a-cylinder-linerhtm

[3] J Antony and R Banuelas ldquoA strategy for survivalrdquo Manufac-turing Engineer vol 80 no 3 pp 119ndash121 2001

[4] G Buyukozkan and D Ozturkcan ldquoAn integrated analyticapproach for six sigma project selectionrdquo Expert Systems withApplications vol 37 no 8 pp 5835ndash5847 2010

[5] A Y T Szeto and A H C Tsang ldquoAntecedent to successfulimplementation of six sigmardquo Journal of Six Sigma and Com-petitive Advantage vol 1 no 3 pp 307ndash322 2005

[6] H C Hung and M H Sung ldquoApplying six sigma to manufac-turing processes in the food industry to reduce quality costrdquoScientific Research and Essays vol 6 no 3 pp 580ndash591 2011

[7] A Saghaei H Najafi and R Noorossana ldquoEnhanced rolledthroughput yield a new six sigma-based performancemeasurerdquoInternational Journal of Production Economics vol 140 no 1 pp368ndash373 2012

[8] F W Breyfogle Implementing Six Sigma Smarter SolutionsUsing Statistical Methods John Wiley amp Sons New York NYUSA 1999

[9] X Zu L D Fredendall and T J Douglas ldquoThe evolvingtheory of quality management the role of six sigmardquo Journalof Operations Management vol 26 no 5 pp 630ndash650 2008

[10] C T Su and C J Chou ldquoA systematic methodology for thecreation of six sigma projects a case study of semiconductorfoundryrdquo Expert Systems with Applications vol 34 no 4 pp2693ndash2703 2008

[11] U D Kumar D Nowicki J E Ramırez-Marquez and DVerma ldquoOn the optimal selection of process alternatives in asix sigma implementationrdquo International Journal of ProductionEconomics vol 111 no 2 pp 456ndash467 2008

[12] A A Junankar and P N Shende ldquoMinimization of reworkin belt industry using dmaicrdquo International Journal of AppliedResearch in Mechanical Engineering vol 1 no 1 2011

[13] Y H Kwak and F T Anbari ldquoBenefits obstacles and future ofsix sigma approachrdquo Technovation vol 26 no 5-6 pp 708ndash7152006

[14] M Swink and B W Jacobs ldquoSix sigma adoption operatingperformance Impacts and contextual drivers of successrdquo Journalof Operations Management vol 30 no 6 pp 437ndash453 2012

[15] J A Blakeslee Jr ldquoImplementing the six sigma solutionrdquoQuality Progress vol 32 no 7 pp 77ndash85 1999


[16] G J Hahn W J Hill R W Hoerl and S A Zinkgraf ldquoTheimpact of six sigma improvementmdasha glimpse into the future ofstatisticsrdquoThe American Statistician vol 53 no 3 pp 208ndash2151999

[17] M Harry and R Schroeder Six Sigma The Breakthrough Man-agement Strategy Revolutionising the Worldrsquos Top CorporationsDoubleday New York NY USA 2000

[18] M J Braunscheidel J W Hamister N C Suresh and H StarldquoAn institutional theory perspective on six sigma adoptionrdquoInternational Journal of Operations and Production Manage-ment vol 31 no 4 pp 423ndash451 2011

[19] X Li J Zhan F Jiang and S Wang ldquoCause analysis of bridgeerecting machine tipping accident based on fault tree andcorresponding countermeasuresrdquo Procedia Engineering vol 45pp 43ndash46 2012

[20] D M Shalev and J Tiran ldquoCondition-based fault tree analysis(CBFTA) a newmethod for improved fault tree analysis (FTA)reliability and safety calculationsrdquo Reliability Engineering andSystem Safety vol 92 no 9 pp 1231ndash1241 2007

[21] E V Gijo J Scaria and J Antony ldquoApplication of six sigmamethodology to reduce defects of a grinding processrdquo Qualityand Reliability Engineering International vol 27 no 8 pp 1221ndash1234 2011

[22] M Xia X Li F Jiang and S Wang ldquoCause analysis andcountermeasures of locomotive runway accident based on faulttree nalysis methodrdquo Procedia Engineering vol 45 pp 38ndash422012

[23] Y Wang Q Li M Chang H Chen and G Zang ldquoResearchon fault diagnosis expert system based on the neural networkand the fault tree technologyrdquo Procedia Engineering vol 31 pp1206ndash1210 2012

[24] R McClusky ldquoThe rise fall and revival of six sigmardquoMeasuringBusiness Excellence vol 4 no 2 pp 6ndash17 2000

[25] A E Summers ldquoAchieving six sigma through FTArdquo in Proceed-ings of the Process Plant Reliability Symposium Houston TexUSA October 1997

[26] H de Koning and J de Mast ldquoA rational reconstruction ofsix-sigmarsquos breakthrough cookbookrdquo International Journal ofQuality and Reliability Management vol 23 no 7 pp 766ndash7872006

[27] J de Mast and J Lokkerbol ldquoAn analysis of the six sigmaDMAIC method from the perspective of problem solvingrdquoInternational Journal of Production Economics vol 139 no 2pp 604ndash614 2012

[28] M H C Li A A Al-Refaie and C Y Yang ldquoDMAIC approachto improve the capability of SMT solder printing processrdquo IEEETransactions on Electronics PackagingManufacturing vol 31 no2 pp 126ndash133 2008

[29] M H C Li and A Al-Refaie ldquoImproving wooden partsrsquoquality by adopting DMAIC procedurerdquoQuality and ReliabilityEngineering International vol 24 no 3 pp 351ndash360 2008

[30] D Starbird ldquoBusiness excellence Six Sigma as a managementsystemrdquo in Proceedings of the Annual Quality Congress pp 47ndash55 Milwaukee Wis USA May 2002

[31] httpenwikipediaorgwikiSix Sigma[32] ldquoWhat is Six Sigmardquo httpwwwisixsigmacomnew-to-six-

sigmagetting-startedwhat-six-sigma[33] G W Frings and L Grant ldquoWho moved my sigma effective

implementation of the six sigma methodology to hospitalsrdquoQuality and Reliability Engineering International vol 21 no 3pp 311ndash328 2005

[34] R McAdam and A Evans ldquoChallenges to six sigma in a hightechnology mass-manufacturing environmentsrdquo Total QualityManagement and Business Excellence vol 15 no 5-6 pp 699ndash706 2004

[35] httpwwwsix-sigma-materialcomHypothesis-Testinghtml

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery


Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014


Shock and Vibration


Mechanical Engineering

Advances in


Civil EngineeringAdvances in

Advances inAcoustics ampVibration



Electrical and Computer Engineering

Journal of


DistributedSensor Networks


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


thinspJournalthinspofthinsp

Sensors

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Active and Passive Electronic Components

Advances inOptoElectronics


Volume 2014

RoboticsJournal of


Chemical EngineeringInternational Journal of


Control Scienceand Engineering

Journal of



Antennas andPropagation

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014


Navigation and Observation



of an engine A slight variation in depth of the step candamage the cylinder and piston as well as hamper the engineoperation A cylinderrsquos head and liner rest on the bored depthThus maintaining the depth is the necessity for any enginemanufacturing firm

Six Sigma is a business management strategy first devel-oped byMotorola in 1986 which seeks to improve the qualityof process outputs [3] It identifies and removes the cause ofdefects andminimizes variability in manufacturing and busi-ness processes [4 5] A set of quality management methodsand statisticalmethods is used Each Six Sigmaproject carriedoutwithin an organization follows a defined sequence of stepsand quantified financial targets According to the definition aSix Sigma process is one in which 9999966 of the productsmanufactured are statistically expected to be free of defects(34 defects per million) [6ndash8] Six Sigma uses a group ofimprovement specialists for problem solving and improv-ing the process continuously [9 10] Six Sigma techniqueshave two main methodologies DMAIC and DMADV [11]Define Measure Analyse Improve and Control (DMAIC)methodology was followed for reducing the rework [12 13]The reasons for the main problem can be detected usingagain two methodologies FTA and failure mode and effectsanalysis FTA deals with identifying all possible causes relatedto a particular problem It is a stepwise approach to identifycauses and all the parameters related to every cause It can beused in almost every application which involves cause-effectanalysis Once the problem is defined it is to be measuredso as to collect statistical data about the problem Once thestatistical data is collected it is analysed using various analysistechniques like chi-square test regression analysis ANOVAand so forth Minitab statistical software was used for theanalysis of various phases of the project

2 Review of the Literature

Six Sigma methodologies have become a top agenda formany companies which are continuously trying to improveproductivity at lesser costs It was shown previously that SixSigma continued to be a predominate target to try and obtaina competitive advantage [11] Many Fortune 500 companieshave adopted Six Sigma to improve the productivity andreduce cost Six Sigma has been described as a data drivenapproach for problem solving business process disciplinedstatistical approach and a management strategy [14ndash18] SixSigma methods can prove to be beneficial when applied tolabour-intensive repeatable processes according to Swinkand Jacobs [14] Six Sigma benefits are significantly cor-related with intensity in manufacturing and with financialperformance before adoption in services [14] The Six Sigmaimprovementmethod is problem-focused and itsmain objec-tives are decreasing scrap earning income and creating valuePrevious researches have shown the effective implementationof Six Sigma techniques in problem solving and processimprovements FTA was successfully utilised to analyse thebridge erecting tripping problem [19] FTA can be used toanalyse hazards and calculate system reliability for simple aswell as complex systems [20] Six Sigma methodologies havebeen previously used for reducing the rejection level [21]

FTA can be effectively used for finding the effective causesfrom accident cases to scrap reduction [22] According toWang et al [23] FTA is a simple effective reliable methodwhich is recognised internationally and is used on guidingsystem optimization and analysing and repairing the systemof weak links FTA helps deepen the research to find out thepossible causes for the fault [24] FTA has previously beenused successfully to establish priorities for themanufacturingplant for future projects aimed at improving the manufactur-ing plant [25]

According to Buyukozkan and Ozturkcan [4] and deKoning and de Mast [26] Six Sigma program offers a widerange of tools and techniques which might be statisticaland nonstatistical that are intended to assist the projectleader Swink and Jacobs [14] showed solid support for thehypothesis that Six Sigma adoption tends to produce signif-icant benefits for firmrsquos profitability Positive return on asset(ROA) changes were frequently observed in latter periods(years +3 and +4) and moreover these benefits also appearto be persistent These findings hint at potential differencesin how Six Sigma programs are possibly being applied infront-office versus back-office contexts Findings suggest thatSix Sigma methods may be most beneficial when applied tolabour-intensive repeatable processes However less labour-intensive quality experienced manufacturing firms will notexperience the profit impact from Six Sigma adoption thesame way that others will Their outcomes also revealedmarginally significant positive effects on sales growth SixSigma projects can accomplish successfully using FTA FTAhas been used extensively by the military the space programand nuclear industry FTA is a very structured systematicand rigorous that lends itself well to quantification [25] FTAcan be used to express the logical relationship between thepossibility of certain accident and causes of undesired eventsor accidents in fault tree diagram [19] It could be successfullyimplemented to find out all the possible causes for the prob-lem or fault that has occurred According to Shalev and Tiran[20] FTA method analysts apply top-down logic in buildingtheir models The problem of reducing process variation andthe associated defect rate can be solved using the DMAICmethodology of Six Sigma Six Sigma is a useful problem-solving methodology and provides a valuable measurementapproach Six Sigma focuses on some vital dimension of busi-ness processes reducing the variation around the mean valueof the process [20] The original task of Six Sigmarsquos DMAICMethodology is variation reduction As stated by deMast andLokkerbol [27] Six Sigma and itsDMAICare built on insightsfrom the quality engineering field incorporating ideas fromstatistical quality control total quality management andTaguchirsquos offline quality control It has also been used for gen-eral tasks like quality improvement efficiency improvementcost reduction and other pursuits in operations Thus SixSigma is a generic method and its original task domain wasvariation reduction typically in manufacturing processesLi et al [28] successfully implemented DMAIC approach toimprove the capability of the solder paste printing processby reducing thickness variations from a nominal valueThe DMAIC approach has shown a wider application andhow the engineering organisation can achieve competitive









4 Case Study



5 Methodology
























302010

0

Observation


201510

50

Observation


Indi

vidu

al v

alue

Mov

ing

rang

e

UCL = 3234

LCL = minus014

UCL = 1996

LCL = 0

X = 1610

MR = 611



















119910 = 119891 (1199091 1199092 1199093 119909

119896) (1)












Damage to cylinder



Frequent use

Friction




Leakages


Loose connections




Dust

Lack of cleanliness


Tools




change




pump parts

Figure 4 FTA


DC depth variation

Ambient

temperature




Lever pressure

(ie 800psi)

(ie 100psi)





(2)




0(no


0(a statistical



0)























Figure 6 MSA plots






Medians tests

Multipleregression


MultipleLogistic

Regression

regressionDisc

rete

Con

tinuo

us

Disc

rete

Con

tinuo

us

Y d

ata

Sing

le Y

Mul

tiple

Yrsquos

Y d

ata

Y d

ata

Chi-square

meansmedians tests

ANOVA



000100000500000

99

90

50

10

1

Residual071900071875071850071825071800

Fitted value

8

6

4

2

0

Residual

222018161412108642

00005

00000

Observation order




minus00004

minus00002

00000

00002

00004

00006

00008

minus00005

Resid

ual

Resid

ual

Freq

uenc

y(

)

minus00005

00005

00000

minus00005


000080000400000

99

90

50

10

1

Residual

071880071865071850071835071820

00005

00000

Fitted value

000040000200000

48

36

24

12

00

Residual222018161412108642

00005

00000

Observation order





()

Resid

ual

Freq

uenc

y


minus00005

minus00005

Resid

ual













120579

(a)

(b)





6 Results




302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of
















4 Case Study



5 Methodology
























302010

0

Observation


201510

50

Observation


Indi

vidu

al v

alue

Mov

ing

rang

e

UCL = 3234

LCL = minus014

UCL = 1996

LCL = 0

X = 1610

MR = 611



















119910 = 119891 (1199091 1199092 1199093 119909

119896) (1)












Damage to cylinder



Frequent use

Friction




Leakages


Loose connections




Dust

Lack of cleanliness


Tools




change




pump parts

Figure 4 FTA


DC depth variation

Ambient

temperature




Lever pressure

(ie 800psi)

(ie 100psi)





(2)




0(no


0(a statistical



0)























Figure 6 MSA plots






Medians tests

Multipleregression


MultipleLogistic

Regression

regressionDisc

rete

Con

tinuo

us

Disc

rete

Con

tinuo

us

Y d

ata

Sing

le Y

Mul

tiple

Yrsquos

Y d

ata

Y d

ata

Chi-square

meansmedians tests

ANOVA



000100000500000

99

90

50

10

1

Residual071900071875071850071825071800

Fitted value

8

6

4

2

0

Residual

222018161412108642

00005

00000

Observation order




minus00004

minus00002

00000

00002

00004

00006

00008

minus00005

Resid

ual

Resid

ual

Freq

uenc

y(

)

minus00005

00005

00000

minus00005


000080000400000

99

90

50

10

1

Residual

071880071865071850071835071820

00005

00000

Fitted value

000040000200000

48

36

24

12

00

Residual222018161412108642

00005

00000

Observation order





()

Resid

ual

Freq

uenc

y


minus00005

minus00005

Resid

ual













120579

(a)

(b)





6 Results




302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of






























302010

0

Observation


201510

50

Observation


Indi

vidu

al v

alue

Mov

ing

rang

e

UCL = 3234

LCL = minus014

UCL = 1996

LCL = 0

X = 1610

MR = 611



















119910 = 119891 (1199091 1199092 1199093 119909

119896) (1)












Damage to cylinder



Frequent use

Friction




Leakages


Loose connections




Dust

Lack of cleanliness


Tools




change




pump parts

Figure 4 FTA


DC depth variation

Ambient

temperature




Lever pressure

(ie 800psi)

(ie 100psi)





(2)




0(no


0(a statistical



0)























Figure 6 MSA plots






Medians tests

Multipleregression


MultipleLogistic

Regression

regressionDisc

rete

Con

tinuo

us

Disc

rete

Con

tinuo

us

Y d

ata

Sing

le Y

Mul

tiple

Yrsquos

Y d

ata

Y d

ata

Chi-square

meansmedians tests

ANOVA



000100000500000

99

90

50

10

1

Residual071900071875071850071825071800

Fitted value

8

6

4

2

0

Residual

222018161412108642

00005

00000

Observation order




minus00004

minus00002

00000

00002

00004

00006

00008

minus00005

Resid

ual

Resid

ual

Freq

uenc

y(

)

minus00005

00005

00000

minus00005


000080000400000

99

90

50

10

1

Residual

071880071865071850071835071820

00005

00000

Fitted value

000040000200000

48

36

24

12

00

Residual222018161412108642

00005

00000

Observation order





()

Resid

ual

Freq

uenc

y


minus00005

minus00005

Resid

ual













120579

(a)

(b)





6 Results




302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of















Damage to cylinder



Frequent use

Friction




Leakages


Loose connections




Dust

Lack of cleanliness


Tools




change




pump parts

Figure 4 FTA


DC depth variation

Ambient

temperature




Lever pressure

(ie 800psi)

(ie 100psi)





(2)




0(no


0(a statistical



0)























Figure 6 MSA plots






Medians tests

Multipleregression


MultipleLogistic

Regression

regressionDisc

rete

Con

tinuo

us

Disc

rete

Con

tinuo

us

Y d

ata

Sing

le Y

Mul

tiple

Yrsquos

Y d

ata

Y d

ata

Chi-square

meansmedians tests

ANOVA



000100000500000

99

90

50

10

1

Residual071900071875071850071825071800

Fitted value

8

6

4

2

0

Residual

222018161412108642

00005

00000

Observation order




minus00004

minus00002

00000

00002

00004

00006

00008

minus00005

Resid

ual

Resid

ual

Freq

uenc

y(

)

minus00005

00005

00000

minus00005


000080000400000

99

90

50

10

1

Residual

071880071865071850071835071820

00005

00000

Fitted value

000040000200000

48

36

24

12

00

Residual222018161412108642

00005

00000

Observation order





()

Resid

ual

Freq

uenc

y


minus00005

minus00005

Resid

ual













120579

(a)

(b)





6 Results




302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of









DC depth variation

Ambient

temperature




Lever pressure

(ie 800psi)

(ie 100psi)





(2)




0(no


0(a statistical



0)























Figure 6 MSA plots






Medians tests

Multipleregression


MultipleLogistic

Regression

regressionDisc

rete

Con

tinuo

us

Disc

rete

Con

tinuo

us

Y d

ata

Sing

le Y

Mul

tiple

Yrsquos

Y d

ata

Y d

ata

Chi-square

meansmedians tests

ANOVA



000100000500000

99

90

50

10

1

Residual071900071875071850071825071800

Fitted value

8

6

4

2

0

Residual

222018161412108642

00005

00000

Observation order




minus00004

minus00002

00000

00002

00004

00006

00008

minus00005

Resid

ual

Resid

ual

Freq

uenc

y(

)

minus00005

00005

00000

minus00005


000080000400000

99

90

50

10

1

Residual

071880071865071850071835071820

00005

00000

Fitted value

000040000200000

48

36

24

12

00

Residual222018161412108642

00005

00000

Observation order





()

Resid

ual

Freq

uenc

y


minus00005

minus00005

Resid

ual













120579

(a)

(b)





6 Results




302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of









Figure 6 MSA plots






Medians tests

Multipleregression


MultipleLogistic

Regression

regressionDisc

rete

Con

tinuo

us

Disc

rete

Con

tinuo

us

Y d

ata

Sing

le Y

Mul

tiple

Yrsquos

Y d

ata

Y d

ata

Chi-square

meansmedians tests

ANOVA



000100000500000

99

90

50

10

1

Residual071900071875071850071825071800

Fitted value

8

6

4

2

0

Residual

222018161412108642

00005

00000

Observation order




minus00004

minus00002

00000

00002

00004

00006

00008

minus00005

Resid

ual

Resid

ual

Freq

uenc

y(

)

minus00005

00005

00000

minus00005


000080000400000

99

90

50

10

1

Residual

071880071865071850071835071820

00005

00000

Fitted value

000040000200000

48

36

24

12

00

Residual222018161412108642

00005

00000

Observation order





()

Resid

ual

Freq

uenc

y


minus00005

minus00005

Resid

ual













120579

(a)

(b)





6 Results




302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of









000100000500000

99

90

50

10

1

Residual071900071875071850071825071800

Fitted value

8

6

4

2

0

Residual

222018161412108642

00005

00000

Observation order




minus00004

minus00002

00000

00002

00004

00006

00008

minus00005

Resid

ual

Resid

ual

Freq

uenc

y(

)

minus00005

00005

00000

minus00005


000080000400000

99

90

50

10

1

Residual

071880071865071850071835071820

00005

00000

Fitted value

000040000200000

48

36

24

12

00

Residual222018161412108642

00005

00000

Observation order





()

Resid

ual

Freq

uenc

y


minus00005

minus00005

Resid

ual













120579

(a)

(b)





6 Results




302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of



















120579

(a)

(b)





6 Results




302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of










302010

0

Observation

Define Improvement

201510

50

Define Improvement


Indi

vidu

al v

alue

LCL = minus337


Observation

Mov

ing

rang

e

UCL = 685

LCL = 0

X = 220UCL = 777

MR = 210




0720

0719

0721

0718

0717




Dat

a







7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of










7 Conclusion



0) that is these



Abbreviations



References






































VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of






























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of







Sensors






Volume 2014

RoboticsJournal of





Journal of








Productivity Improvement of a Special Purpose Machine

Documents

Transcript of Productivity Improvement of a Special Purpose Machine