Department of Product and Production Development CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2016

Decision Support Model for Production Layout Improvement A Case Study at Brose Sweden AB Master’s thesis in Production Engineering

KIM CHENG CARL HÖJER

3

4

DecisionSupportModelforProductionLayoutImprovement

ACaseStudyatBroseSwedenAB

KIMCHENGCARLHÖJER

DepartmentofProductandProductionDevelopmentCHALMERSUNIVERSITYOFTECHNOLOGY

Göteborg,Sweden2016

5

Decision Support Model for Production Layout Improvement A case study at Brose Sweden AB Kim Cheng Carl Höjer © Kim Cheng & Carl Höjer, 2016. Department of Product and Production Development Chalmers University of Technology SE-412 96 Göteborg Sweden Telephone + 46 (0)31-772 1000 Chalmers Reproservice Göteborg, Sweden 2016

6

ABSTRACTThis report is a master thesis from Chalmers University of Technology’s ProductionEngineeringmasterprogram.ItwaswrittenforBroseSwedenABandsupportedbyÅFAB.Thepurposewastoinvestigatetheprospectsofdevelopingasupportivemodelthatfacilitatestheprocessofdevelopinglayoutsinproductionsystems.Providingacommonstructureforthelayoutdevelopmentprocessgivesclaritytoaprocesscontainingmanysources of information and a large amount of parameters. Streamlining the processcouldalso lead topotential timesavings.The researchwasconductedasa case studyconsisting of a literature review, familiarization of Brose’s production system, datagathering, layout generation, different methods of evaluating layouts and lastly thegeneration of a decision supportmodel. Themodel was developed based on existingacademic researches of production systems and production layout development incombinationwithanactuallayoutimprovementcaseatBroseSwedenAB.Theresultingmodelis intheformaflowchartdescribingwhichstepstotakeinordertodevelopanexisting layout. Themodel emphasizes evaluation of solutions and iterative stages ofevaluation and correction.While themodel yielded positive results for Brose furthertesting at other facilities is required in order to verify the model’s effectiveness.Productionfacilitiesinotherindustriesthanautomotiveisofparticularinterest.Keywords:Casestudy,layoutproblem,layoutevaluation,layoutgeneration,productionsimulation

7

8

ACKNOWLEDGEMENTSThisthesiswasmadepossiblewiththehelpofafewpeople.Wewouldliketoexpressourgratitude toHansFrohlund,whoseexpertise fromsimilarprojects in the industryprovedtobeinvaluableforus.WewouldalsoliketoextendourthankstoÅFABwhichsupplied us with software licenses. We would also like to thank our supervisor JonBokrantz for supportingus inwriting this report.His support gaveus structure in anunstructuredworld. Last but not leastwewould like to thank Bengt Fransson, PeterSeuring,andtherestofBroseinTorslandaforthecollaboration.Theirsupportandtrustfromdayonemeantalotforthispaper.

9

TerminologyThefollowinglistconsistsoftermsandabbreviationsthatareusedinthethesisreport.BMT=BasicMotionTimeStudyCAD=ComputerAidedSoftwareDSM=DecisionSupportModelDSS=DecisionSupportSystemERP=EnterpriseResourcePlanningKPI=KeyPerformanceindicatorMODAPTS=ModularArrangementofPredeterminedTimeStandardsMTM=MethodsTimeMeasurementSLP=SystematicLayoutPlanningSWAG=ScientificWild-AssGuessWF=WorkFactor

10

TableofContents1 Introduction...........................................................................................................................................121.1 Background..............................................................................................................................................................121.2 Purposestatement................................................................................................................................................121.3 Casedescription-BroseSwedenAB..............................................................................................................131.3.1 Productdescriptions.............................................................................................................................131.3.2 Brose’sProductionSystem.................................................................................................................15

1.4 ObjectiveandResearchQuestions..................................................................................................................161.5 Limitations...............................................................................................................................................................16

2 Theoreticalframework......................................................................................................................172.1 ProductionLayout.................................................................................................................................................172.1.1 SystematicLayoutPlanning(SLP)..................................................................................................182.1.2 Studiesofimpactoflayoutdesignonwork................................................................................21

2.2 ProductionEvaluation........................................................................................................................................212.2.1 Measuringwork......................................................................................................................................212.2.2 Productionflowevaluation................................................................................................................24

2.3 DigitalTools.............................................................................................................................................................242.3.1 ComputerAidedDesign-AutoCAD2016....................................................................................242.3.2 Materialflowsimulation-ProPlanner:FlowPlanner.............................................................252.3.3 Workbalancing-AviXresourcebalance......................................................................................25

2.4 DecisionSupportModel......................................................................................................................................252.4.1 Thedefinitionofadecisionproblem.............................................................................................262.4.2 Differentapproachesofdecisionproblems................................................................................262.4.3 Flowchart...................................................................................................................................................272.4.4 QuantitativedataandQualitativedata.........................................................................................29

3 Methodology..........................................................................................................................................303.1 Researchdesign......................................................................................................................................................303.2 Preparatorywork..................................................................................................................................................323.2.1 Literaturereview....................................................................................................................................323.2.2 IntroductiontoBrose’sproductionsystem................................................................................32

3.3 DataCollection.......................................................................................................................................................333.3.1 Interviews..................................................................................................................................................333.3.2 Timestudies..............................................................................................................................................343.3.3 Distancemeasurements......................................................................................................................353.3.4 Machinedata.............................................................................................................................................353.3.5 DatacollectionfromERPsystem.....................................................................................................35

3.4 Bufferreplenishmentrate..................................................................................................................................353.5 WorkbalanceandWorkdesign......................................................................................................................353.5.1 MethodTimeManagement(MTM).................................................................................................363.5.2 Recordingactivities...............................................................................................................................363.5.3 Calculationofworkforce.....................................................................................................................36

3.6 Generatingproposalsforfuturelayout.......................................................................................................373.7 MaterialFlowSimulation..................................................................................................................................393.8 DevelopmentoftheDecisionSupportModel.............................................................................................393.8.1 DefiningthecharacteristicsoftheDSM........................................................................................393.8.2 FindingasuitableDSMstructure....................................................................................................41

4 Results......................................................................................................................................................434.1 Datacollection........................................................................................................................................................434.1.1 Interviews..................................................................................................................................................434.1.2 ERP-system...............................................................................................................................................43

4.2 Brose’sproductionsystem.................................................................................................................................44

11

4.2.1 Theproductionsystem........................................................................................................................444.2.2 RailRiveting..............................................................................................................................................454.2.3 PlateSealing..............................................................................................................................................474.2.4 CoolingFanAssembly...........................................................................................................................49

4.3 SLP................................................................................................................................................................................544.4 Logisticssimulation..............................................................................................................................................544.5 Workdesign.............................................................................................................................................................544.5.1 Operatorresponsibleforbothrailandsealingmachine.......................................................54

4.6 Thelayoutdevelopment.....................................................................................................................................594.6.1 Creativegeneration...............................................................................................................................594.6.2 Firstgeneration.......................................................................................................................................594.6.3 Secondgeneration..................................................................................................................................604.6.4 Thirdgeneration.....................................................................................................................................604.6.5 Fourthandfinalgeneration...............................................................................................................60

4.7 TheDSMforproductionlayoutimprovements........................................................................................614.7.1 ComponentsoftheDSM......................................................................................................................634.7.2 Stage1:Pre-study...................................................................................................................................634.7.3 Stage2:Datacollection........................................................................................................................654.7.4 Stage3:CreativeProcess....................................................................................................................664.7.5 Stage4:RefinementandScreening................................................................................................66

5 Discussion...............................................................................................................................................685.1 Resultsdiscussion..................................................................................................................................................685.1.1 Workdesign..............................................................................................................................................685.1.2 Thequalityoftheworksampling....................................................................................................685.1.3 TheDSMapplicability...........................................................................................................................695.1.4 Layoutdesignimpactonwork.........................................................................................................69

5.2 Methodologydiscussion......................................................................................................................................695.2.1 Theeffectivenessoftheresearchdesign.....................................................................................695.2.2 Interviewformat.....................................................................................................................................705.2.3 Bufferprioritization..............................................................................................................................705.2.4 SystematicLayoutPlanning(SLP)..................................................................................................705.2.5 Materialflowsimulation.....................................................................................................................715.2.6 Choiceofmethodformeasuringwork..........................................................................................71

5.3 Contributiontothefield.....................................................................................................................................726 Conclusion..............................................................................................................................................73APPENDIX.......................................................................................................................................................78AppendixA............................................................................................................................................................................78AppendixB............................................................................................................................................................................80AppendixC............................................................................................................................................................................82AppendixD...........................................................................................................................................................................85AppendixG............................................................................................................................................................................90AppendixH...........................................................................................................................................................................94AppendixI.............................................................................................................................................................................96

12

1 IntroductionThepurposeofthisinitialchapteristointroducethethesis.ItincludesabackgroundoftheresearchareafollowedbyadescriptionofthecompanyBroseSwedenABuponwhichacasestudyisperformedon.Inoneofthefinalpartstheobjectiveofthethesisisstatedalongwiththeresearchquestions.Afinalsectionofthechapterliststhelimitationsofthethesis.

1.1 BackgroundThefrequentoccurrenceoftechnological innovationshasledtocontinuouslychangingmarketsandcompanieshavetoquicklyadaptinordertostaycompetitive(Wecketal.,1991). The oneswho are able to change alongwith themarket are consequently theonesthatareabletoacquireacompetitiveedge.Forthemanufacturingindustrythishasresultedinchangedrequirementsoftheproductionprocesses.Studiesshowthatthereis aneed for flexibility in theproduction,which isdue to the frequent changesof themarket (Benjaafar et al., 2002). Facilitating the process of developing factory layoutscould support this notion. A streamlined development process that incorporatesimportant considerations could potentially shorten the time that it takes formanufacturing industries to adapt to the prevailing circumstances of their markets.Furthermore, studies suggest that production layout development for manufacturingcompanies isanareawheregreat reductions inoperationalexpensescanbeachievedwhen done correctly. According to Drira et al. (2007) the cost reductions can be assignificant as 50%. Finding the optimal layout is however a complexmatter and as aresult a vast amount of studies has been carried out on the subject during the lastdecades. The issue has come to be known as the “facility layout problem”, where“facilities” refer to the production units such as machines, workstations, and buffers(Driraetal.,2007).Thepotentialcostreductionscombinedwithastreamlinedlayoutdevelopmentprocesssuggest that a decision supportmodelwith the application area of production layoutdevelopment,couldbeapotentialtoolforthemanufacturingindustry.Themodelwouldfacilitate theprocessofdevelopinga layoutwhilesimultaneouslymakingsure thatnoimportantconsiderationsaremistakenlyexcludedinthedevelopmentprocess.BroseSwedenABisafirst-tierautomotivesupplierthatmanufacturesdoorsystemsandwindowregulators.Thecompanyisundergoingchangesintheirproductionsystemanddemandsanefficientproductionlayoutthatperformsonalevelthatmeetstheneedsoftheir customer. The improvement process at Brose is therefore a suitable foundationuponwhichadecisionsupportmodelcanbebasedon.

1.2 PurposestatementThe purpose of the thesis is to investigate the prospects of developing a supportivemodel that facilitates theprocessofdeveloping layouts inproductionsystems.Layoutdevelopment can be a complicated and time-consuming endeavor. Streamlining theprocesswithastructuredmodelcouldsavetimeandsupporteffectivedecision-making.

13

1.3 Casedescription-BroseSwedenABThe subject of the thesis is the German automotive supplier Brose Sweden AB inTorslanda,whichwillbereferredtoasjustBroseduringthecontinuationofthisthesisreport.Thecompanyislocatedinmultiplecountries,butthefacilityinTorslandaworksexclusively as a supplier for the automotive manufacturer Volvo Car Corporation.Working close to the automotive industrymeans thatBrose is active in a competitiveindustrywhere it is crucial that theperformanceof a supplier isup topar.Notbeingabletodelivertheexpectedperformancecouldpotentiallymakethecustomerschooserivalingsuppliers.AsapartoftheimprovementprocessBrosehasinvestedinamachinewith shorter cycle time and that takesup a smaller area than theprecedingmachine.Theextra space that it clearsupshouldbeutilized inanoptimalwaywith regards toproductionperformance.Thetransportdistancebetweenstorageandproductionislongandsomeconsecutivestationsareplacedfarapart.Brosebelievesthataredesignofthelayout is in order. They havemade similar changes in the past with positive results.Howevertheirchangeswereoftenmadebasedon“gutfeeling”,oftenwithoutanydatatosupporttheimprovementwork.Asaresultthefullcapacityoftheproductionmightnotbecaptured.Brosedesiresthenewlayouttobebasedonfactsandinordertodosoadeeperunderstandingoftheproduceddataisrequired.

1.3.1 ProductdescriptionsBrose’s production facility in Torslanda supplies two types of products to Volvo CarCorporation. The first product is a door module that handles a number of door andwindow mechanisms such as the window regulators, locks, and loud speakers(Brose.com,2016).This thesis isconcernedwithtwocomponentsof thedoormodule,namelytherailsandtheplates.SeeFigure1andFigure2.

14

Figure1.Anassembledrail.Thetwopulleysoneachendoftherailguidethewirethatenablethewindowregulationsystem.

Figure2.Asealedplate.Thethinlinealongtheedgeoftheplateistheappliedsealantthatprotectsthecomponentsofthedoormodulefromgettingdamagedbyexternalmoistordirtparticles.

15

ThesecondproductthatBroseproducesisacoolingfan(CF)thatcoolsdowncarengines,seeFigure3.Processesinvolvingbothoftheproductsareincludedwithinthescopeofthethesis.

Figure3.Acompletecoolingfan.Anelectricalmotorinthecenterofthemoduledrivestherotationofthefan.

1.3.2 Brose’sProductionSystemThethesiscoversthreeproductionprocesses.Twoofthemmanufacturecomponentsforthe car doormodule. They are called the sealingprocess and the rail-rivetingprocess.Both of them are semi-automatic. The required manual work consists of feedingmaterial into themachinesandhandling thematerialafter ithasbeenprocessed.Thethirdproductionprocessisthecoolingfan(CF)process,whichinvolvesassemblingthecoolingfanandisfollowedbyaqualitycontrol.

16

1.4 ObjectiveandResearchQuestionsTheobjectiveofthisthesisistocreateaDecisionSupportModel(DSM)thatisgenerallyapplicabletoanyproductionsystem.Inordertoachievethepurposeandobjectiveoftheresearchthetwofollowingresearchquestionsandtheirrespectivesub-questionsaretobeanswered.

RQ1:Howisaproductionlayoutoptimized?• HowisBrose’sproductionsystemoptimized?• Howtomaximizetheutilizationofprocesseswithregardstoproduction

layout?

RQ2:Whatshouldbeincludedinadecision-makingmodelforlayoutimprovements?

• Howaretheessentialparameterswithregardstoproductionlayoutidentified?

• WhichmethodscanbeincludedinaDecisionSupportModelforlayoutimprovements?

• Howisalayoutdevelopedinastructuredmanner?

ThequestionswillbeansweredthroughacasestudyattheautomotivesupplierBrose.ThethesisgroupwillproposeanewproductionlayoutofBrose’sproductionsystembyconsideringanumberoffactorssuchastheplacementofmachines,levelofworkforce,materialstorage,etc.Basedontheimprovementprocess,amodelwillbedevelopedtohelpsupportfuturedecision-makingrelatedtolayoutimprovements.

1.5 LimitationsThefollowinglistincludesthelimitationsofthethesis:

• The thesis does not cover Brose’s complete production system. Only a limitedsectionofthesystemisthesubjectofstudy.

• Relocation of material storage locations outside of the production is notconsidered.

• Thetimescopeofthethesisislimitedtofivemonth.• The thesis group has not been provided with any monetary resources. Used

equipment and tools are borrowed from the faculty office and/or the externalsupervisor.

• The development of the Decision Support Model is only based on a singledevelopment process, which is the one done at Brose. Therefore the resultingmodelisonlyafirstdraft.

• The amount of previous studies on the area of Decision Support Models incombinationwithproductionlayoutdevelopmentislimited.

17

2 TheoreticalframeworkThispartof the report covers theories that together create the theoretical frameworkfor the thesis. The theoretical framework includes theory about production layout,which is themain topicof the thesis.The theoryalso includesmeansof assessing theperformanceofaproductionsystem followedbydescriptionsof software that isusedforevaluationandanalysispurposes.ThefinalchapterisabouttheoryrelatedtoDSMs.

2.1 ProductionLayoutWu(1997)observedthatmostproductionsystemdevelopmentprocesseswerecarriedoutthroughtwomainapproaches.Thefirstapproachistodesignaproductionsystembasedonpredeterminedobjectivesforthedesign.Thisapproachdoesnotconsideranypreviousdesignsystems.Asa result thegeneratedsystemmight requireanextensivesystemredesign,whichisnotalwaysfinanciallyviable.Thesecondapproachconsidersthe current system. The positive effect of this approach is that it takes into accounthistoricalsystemdevelopmentprocesses,thuspreviouslygainedexperienceisnotlost.Thedownsideisthatitimpedescreativitysincethedevelopmentprocessissolelybasedon previous work. In order to capture the benefits of both approaches Wu (1997)suggestsageneralframework,seeFigure4.

Figure4.Wu's(1997)generaldesignframework.

Theframeworkismadeupofanumberofphaseswheretheworkingprocessgoesbackandforthbetweenthephasesiterativelyinordertogenerateanoptimalsolution.

18

Themethod thatwasused in the research is called Systematic LayoutPlanning (SLP)(Muther&Wheeler,1977)andcomprisesfeaturesthatmatchthefirstapproachwherethecurrentproductionsystemisnotconsideredinthedevelopmentprocess.Followingsectionexplainsthemethodandbreaksdowneachofthesteps.

2.1.1 SystematicLayoutPlanning(SLP)Systematic layout planning, or SLP, is a structured approach to layout design. It wasdevelopedbyRichardMutherinthe1960’s,butisstillwidelyusedasatoolandisstillcited by others even though it was developed decades ago (Kusiak & Heragu,1987)(Bellgran&Säfsten,2010)(Meller&Gau,1996).SLPisasystemofsixstepsthat,when followed, will result in a systematic and objective layout (Muther & Wheeler,1977).Thesixstepsarethefollowing:(Bellgran&Säfsten,2010)

1. ClarifytheconnectionInthefirststepallaffectedfunctions,areasandactivitiesareidentifiedandtheirconnectionsnotedinaso-calledtriangularconnectionchart.SeeFigure5

Figure5.Atriangularconnectionchart.Thefunctions,areas,andactivitiesarelistedinthecolumntotheleft.Thedesireddistancesbetweenthemaregradedinthesquares.

19

Thefunctionsare listed insequentialorder.Thenthedesireddistancebetweensaidfunctionsisdecided.ThedesireddistanceisgradedinthescaleofA,E, I,OandU,whichdescribesascalefrom“absolutelynecessary”to“notnecessary”.Adistance can also be graded X, which implies that the two functions must beplacedfarapartfromeachother.Inaddition,thereasonforthegradingshallbeexplainedforeachfunction,enablingfollow-uplater.

2. EstablishfunctionaldemandsStep twonotes thedemandsof the functions fromstep1.Considereddemandscanberelatedtotheworkarea,constructionalfactors,serviceneeds,sensitivitytovibrationsetc.

3. LinefunctionalconnectionInthisstep,avisualdrawingofthefunctionslinkageismadewhileconsideringtheconnectionsandrequirementsfromsteponeandtwo.Thedrawingismadeby linking the functions with different number of lines, corresponding thedemandsof closeness,wheremoreparallel lines representsa linkageofhigherimportance, i.e. fourlinesrepresentsanA-grading(absolutelynecessary),whilethreelinesrepresentsanE-grading(highlyinfluential),etc.Theonlygradingthatdoes not follow this pattern are the X-connections, which are represented bywavy lines. The process starts with the A-connections and proceeds in adescending order until all connections have been drawn. The chart is iterateduntil all functions are placed in a structured manner. Finally, the arearequirementsofeachfunctionarewrittenoutnexttoeveryfunction.SeeFigure6.

20

Figure6.Exampleofasimplifiedlinefunctionalconnectiondiagram.Thecircleswithnumbersarethevariousfunctions.Thelinesaretheirconnectionsandthenumberoflinescorrelatestohowclosetheyshouldbeplacedtoeachother;morelinesrepresentsabiggerneedforcloseplacement.

4. Drawalternativemasterplans

Stepfourgroupsthefunctionsgeographicallyonadrawingwithregardstospacerequirements.Constructionalandtechnicaldemandsarealsoconsidered.Severalplansaremade,desirablywithdifferentsolutionsofmaterialandproductflows.

5. AssessthedifferentalternativesStepfiveistheevaluationstage.Here,thedifferentlayoutproposalsareassessedaccordingtoseveralcriteriaidentifiedasimportantforthefinalsolution.Typicalcriteriacouldbe:

• Investmentcost• Effectivenessofflows• Flexibility• Utilizationoffloorarea

21

The criteria are givena certainweightdependingon their relative importance.The layoutproposalsareassessedaccording the identifiedcriteria toa scaleofabsolutely perfect (A=4), efficient solution (E=3), interesting solution (I=2),ordinary solution (O=1), unimportant (U=0), unwanted (X=-). Theappropriatenessofthesolutionisgainedbymultiplyingthecriteriaweightwiththe layout proposals solution for said criteria. For example, if the criteria lowinvestmentcost isgiven theweight3andsolutionA is regardedasanefficientsolution(E=3)forthataspect, thentheresultingvaluewillbe3x3=9.Thesameprocedure ismade for all thedifferent criteria and solutions.At the endof theassessment the score for the different solution is summed, with the highestscoring solution as themost appropriate. For the purpose of assessing a chartsuchastheonebelowistypicallyused.

6. DesignthechosenplansolutionThesixthandfinalstageiswherethechosenlayoutbecomesthefinaldrawingofthe solution. Machines, equipment and buffers are all clearly marked.Construction details are checked, truck paths, door clearance, etc. Details areinvestigatedandthencorrected,resultinginthecompletelayoutdrawing.

2.1.2 StudiesofimpactoflayoutdesignonworkThere are several studies of the effect of layout design and its influence on theworkefficiency (Kamarudin et al., 2011)(Cohen et al., 2008). Kamarudin et al. (2011)investigatetheeffectofheadcountandlayouttypeontheperformanceofalayout.Thestudy found that for small batch and high variety a higher headcount is needed tocounteracttheincreasednumberofsetupsandotherrequiredtasks.Cohenetal.(2008)investigate the effect of layout and work allocation on the “makespan” of batches ofdifferentproductscharacterizedbya learningcurve.“Makespan” isacommonterminmanufacturingmeaningthetimedifferencebetweenstartandfinishofatask(Cohenetal.,2008).Thestudyresearchespotentialguidelinesforreducingthemakespanforasetnumberofworkers.The results showed that small batchproductionof ahighvariantnumbercouldbeeffectiveifrepetitionishighthusshorteningthelearningcurve.Bestresultiswhentherearestationsinalineproducingtheproduct,furtherincreasingtherepetitionfortheworkers.

2.2 ProductionEvaluationEvaluatingtheperformanceofaproductionisimportantwhendeterminingifconductedchanges on a production system yield positive or negative results. Appropriatemeasurescanthenbetakeninorderto improvetheperformance.Thischaptercoverspreviouslydevelopedtheorythatisusedforthesetypesofperformanceevaluationsandmeasuresforimprovingit.

2.2.1 MeasuringworkMeasuring and quantifying work is a way of determining if a process is balanced. Aprocess is balanced with the purpose of creating work standards that enhance theworkflow. Standards are in general developed through one out of three ways:estimation,directobservation,andtheuseofstandarddata(Lawrence,2004).Furtherexplanationofthethreewayscanbefoundinthesectionbelow.

22

EstimationEstimation can typically be done in two ways. The first way is when an individual,believedtobeknowledgeableabout the taskorworktobeevaluated,giveshisorherestimationofhowlongitwilltake.Theflawsaremany,asitisnotanaccuratemethodandproblemssuchasbottlenecksareoftenformedduetotheinaccuracy.Thistypeofestimation is known as the acronym SWAG (Scientific Wild-Ass Guess) (Lawrence,2004).Thesecondwayofestimationistolookathistoricaldataandmakeacomparison.Historical data is however subject to Parkinson’s Law (Lawrence, 2004). Parkinson’sLaw applied to industrial engineering states that the amount of time required tocomplete a task is directly proportional to the time available.Which implies that thestandardtimeincreasesordecreasesbasedontheamountofavailabletime.DirectobservationAnotherwayofmeasuringworkisthroughdirectobservation.Directobservationsaredifferentiatedbetweentimestudies,worksampling,andphysiologicalworkmeasurement(Lawrence,2004).In time studies an operation is analyzed and necessary steps of the operation areidentified. The chronological order of the steps and the time that it takes to performeach of the steps effectively are also observed (Lawrence, 2004). Time studies areusuallyperformedusingatimemeasuringdeviceandismostsuitablewhenappliedonhighly repetitive tasks where the cycle time is relatively short. The method can becombinedwithcomputationalsimulations,whichenablemoreadvancedanalysisoftheworktasks.Section2.3.3goesintofurtherdetailoncomputationalsimulationsofworktasks.Beforethetimestudyisperformeditisimportanttoknowthatthequalityoftheresultsisdependentontheoperatorthatisselectedforthestudy(Lawrence,2004).Theoperator ispreferably skilledandworksaccording to theestablishedworkmethod inordertoreducethevarianceoftheworkperformance.Whenintroducingthetimestudytotheoperatoritisimportanttoconsiderthewaythatitisconducted.Lawrence(2004)emphasizesbeingtransparentwiththestudy.Conveyinghonestyandclarityaboutthepurposeofthestudymakestheoperatormorereceptivetoit.Work sampling is the second standard that is based on direct observation. It is astatistical method that is carried out by making a large number of observations atrandom intervals (Lawrence, 2004). Based on the observations a probabilisticconclusionisdrawnconcerningtheworktasks.Themethodissuitableforworkthathasa relatively long cycle time and is normally more time consuming compared to timestudiesduetothelargenumberofsamplesthatisrequired(Lawrence,2004).Thethirdand finalwayofmeasuringwork throughdirectobservation isbydoingphysiologicalworkmeasurements.Themethod isbasedon the fact that thehumanbodyconsumesdifferent amountof energydependingon thework task that is performed (Lawrence,2004).Theenergyconsumptionisregardedasan“energycost”whichmeansthat it isdesired to keep the energy levels on a low level. According to themethod the energyconsumptioncanbederivedfromaperson’sheartrateandoxygenconsumption.Thusbymeasuringthesetwoparameterswhiletheworkerisperformingvaryingworktaskstheenergycostscanbedetermined.

23

StandarddatasystemsThethirdwayofmeasuringworkistousestandarddatasystems(Lawrence,2004).Astandarddata system isessentiallyadatabasewitha listof standardmovements thatoftenoccurinmanualworkandthetimethatittakestoperformeachmovement.Thesystem is then used for creating a time standard for a specific work. If the personmeasuring the work knows which movements are needed the time consumption foreachmovementcanbecollectedfromthestandarddatasystemandatimestandardcanbedeveloped.Thedatainthedatasystemhasbeencompiledthroughpreviousstudiesofmanualwork.Thefundamentalideaofstandardtimesystemsisthatitisarelativelyeasy and quick way to set standard times since it does not require any timemeasurementsoftheworktasks.(Lawrence,2004)There are two types of standard data: macroscopic standard data and microscopicstandarddata(Lawrence,2004).Macroscopicstandarddataconsistsofworkelementsthatoccur indifferentworks.Lawrence(2004)uses theexampleofwalkingassuchaworkelement.Walkingisanelementthatisperformedinthesamewaynomatterwhichtypeofworkitisrequiredin.Theonlydifferencewillperhapsbethedistancethattheworkerwalks.Insteadofmeasuringevery“walkingelement”thetimedataforwalkingcanbeeasilyacquiredfromthedatasystem.Thealternative,microscopicstandarddata,emphasizes motions. The standard data type is also known as predetermined timesystems.Byanalyzingaworktaskanditsmovements,entireprocessescanbeevaluatedusing predetermined time systems. The most widely recognized predetermined timesystem was first developed in USA during the 1940s and is called Methods TimeMeasurement,commonlyknownastheabbreviationMTM-1(Lawrence,2004).MTM-1ismadeupofalistof“basicmotions”togetherwithadescriptionofeachmotionandthetimethatittakestodothesemotions.WhenanoperationisanalyzedusingMTM-1itisbroken down into smaller basic motions that fit with the basic motions that aredescribedinMTM-1.ThetimedatainMTM-1isthenassignedtothemovementsintheoperationandatimestandardcanbedeveloped.Examplesofdifferentbasicmovementsin MTM-1 are reach, move, turn, position, and grasp. The effectiveness of thispredeterminedtimesystemissupportedbythefactthatsystemsthatweredevelopedafter its introduction were heavily based on it. Other predetermined time systemsaccordingtoLawrence(2004)areWorkFactor(WF)system,BasicMotionTimeStudy(BMT),andModularArrangementofPredeterminedTimeStandards(MODAPTS).Thesetime systems are generally accepted as alternatives for MTM-1, but are not asinternationallyrecognized.TheWFsystemwasthefirstpredeterminedtimesystemthatwaseverdeveloped.Itincludesa“workfactor”thatisaddedtothebasicmotiontimeinorder to factor in the additionalwork time causedby anyweight or resistance in theworktask.BMTalsoincludesafactorthatitcallsthe“forcefactor”,makingitsimilartotheWFsystem.Thedifference ishowtheydefinewhatabasicmovement isandwhatthe force factor is supplemented for. MODAPTS is a system that is oriented towardshuman work rather than mechanical work. The work times that MODAPTS arecomprisedofarebasedonthefactthatalargerpartofahumanbodytakesalongertimetomovethanasmallerbodypart.

24

2.2.2 ProductionflowevaluationInalayouttheflowefficiencyisacrucialcriterion.Whatvaluesflowefficiencyactuallyconstitutesof,variesdependingonliterature.Lin&Sharp(1997)haveidentifiedthreecriteria:

• Materialflow• Spacerelationship• Robustness&flexibility

For thematerialflow subgroupdigitalsimulationtoolscanwithadvantagebeapplied,seesection2.3.2.Spacerelationshipandrobustness&flexibilityontheotherhandhaveamorequalitativenatureasopposedto thematerial flow’squantitativenature(Lin&Sharp,1997).Spacerelationship isconcernedwiththevisibilityandaccessibilityoftheproduction, which is an effect of the machine and material placement. In practice itmeans that paths should not be obstructed or too narrow for neither trucks noroperators (Lin & Sharp, 1997). Robustness & flexibility describes the level of futureproofingandreadinessforchangesintheflowlayoutintermsofcapacityforchangesinbothspacecapacityforanincreasednumberofvariantsaswellasincreasedproductioncapacity(Lin&Sharp,1997).

2.3 DigitalToolsForthepurposeofevaluationsandanalysisdigitaltoolscanbeused.Presentedinthissectionarethedigitaltoolsutilizedinthethesis.

2.3.1 ComputerAidedDesign-AutoCAD2016A tool that isused for technicaldrawings is the software toolComputerAidedDesign(CAD).Itisadesigntoolthatisusedforcreating,modifying,analyzing,andoptimizingengineering designs (Lalit Narayan et al., 2008). CAD-softwares facilitate the productdevelopment process and reduce the time that it takes to synthesize, analyze, anddocumentdesigns.ThereareavarietyofdifferentCAD-softwaredevelopersinthemarket.CAD-softwaresincludeAutoCAD,CATIA,andSolidworks just tomentiona few(AutoCAD.com,2016a)(3DS.com,2016)(Solidworks.com,2016).Therelevantsoftwareinthisstudyis2016’sversionofAutoCAD. It is adesignanddrafting software that canbeused for creatingvisual and virtual representations of products in either 2D or 3D depending on thepurposeofthedrawing(Autodesk.com,2016a).ThesoftwareiscreatedbythecompanyAutodesk that develops software for 3D designs, engineering, and entertainment(Autodesk.com, 2016b). AutoCAD can be used for facility drawings, which makes itsuitableforthestudy.

25

2.3.2 Materialflowsimulation-ProPlanner:FlowPlannerInordertoevaluatetheeligibilityofthematerialflowinafactorylayout,simulationcanbeapplied.FlowPlannerbyProPlanner isoneof such tools (ProPlanner,2016a).Thesoftwareisanadd-ontoAutoCADandcanbeusedforanalyzingtaskssuchas:

• Aisledesign(locationandsizing)• Dockquantityandplacement• Off-linestorageplacement• Workstationplacementandreplenishmentdesign• Indirectlaborrequirementsestimation• Planttrafficforintersection/aislesafetyevaluation• Efficienthandlingmethodsselection(Fork,Conveyor,Crane,etc.)

(ProPlanner.com,2016b)FlowPlannercanhandleandsimulateasetofinputdataforamaterialflowsimulation.Thecomplexityofthesimulationdependsontheamountofinputparametersandwhichparameters that are selected.Examplesof input informationare transportationpaths,quantityofproducts/materialper flow, transportationtype,andnumberofcontainersper delivery, just to name a few. By applying this information to a technical drawingwherethelocationsandpathsaredefinedflowdiagramsandnumericalresultscanbegenerated.

2.3.3 Workbalancing-AviXresourcebalanceThe software AviX resource balance uses MTM-based standard times for quantifyingworktaskscreatingdatathatcanbeusedincalculationsandanalyseswhenbalancingproductionsystems.With theaidofAviX theproductionprocesscanbeplanned foraconceptual system evaluating the possibilities and optimizing the effectiveness.(AviX.com,2016)

2.4 DecisionSupportModelInthisthesisreporttheDecisionSupportModelwillbereferredtobyitsabbreviationDSM.Mostcommonly,aDSMisusedintheformofdigitaltoolsknownasDSSs(DecisionSupportSystems).TwoofthemostcommonandclassicaldefinitionsforDSSsare:“Interactivecomputer-basedsystems,whichhelpdecisionmakersutilizedataandmodelstosolveunstructuredproblems.”(ScottMorton&Gorry,1971)It was later described by Scott Morton & Keen (1978) as a system that couplesindividual’s intelligence with that of a computer to further improve the quality of adecision.Acomputer-basedsupportfordecisionmakerswhodealwithsemistructuredproblems. However the term DSS and DSM are content-free expressions. There is noofficialdefinition,meaningthatitcanbedefineddifferentlybydifferentpeople(Turbanetal.,2014).

26

Thesedefinitionscanbemisleading.ADSMisnotstrictlyadigitaltool,itcanalsobeanon-digitaltool.Forexample:

• Evaluationmatrices• Qualityfunctiondeployment(QFD)(Bergman&Klefsjö,2010)• Kanomodel(Bergman&Klefsjö,2010)• Muther’sSLPmethod(Muther&Wheeler,1977)

2.4.1 ThedefinitionofadecisionproblemAdecisionproblemariseswhenapersonorgrouphasanideaofapossiblefuturestate.Theimaginedfuturestateisdifferentfromthecurrentorwillbedifferentinthefuture.Inordertoachievethefuturestatethedifferencesofthefutureandcurrentstateneedto be eliminated or minimized. This alone will not make a decision problem. To beconsideredadecisionproblemtheremustbeseveralwaystoeliminate/minimizethesedifferences (Grünig & Kühn, 2013). In short, a decision problem is when there areseveralwaysofachievingatargetstate.Decisionproblemscanbeclassifiedintotwogroupsofdifficulty.Itcanbeeitherasimplechoicedecisionofwhatcolortochooseforacaroritcanbeacomplexdesignquestionsuch as “what features should be included in our product” (Grünig & Kühn, 2013).AccordingtoGrünigandKühn(2013)aproblemcanbeconsideredtobecomplex if itfulfillstwoormoreofthefollowingcriteria:

• Severaltargetsaretryingtobereachedsimultaneously• Thereisalargenumberofdecisionvariables.Resultinginmanypossible

problemsolvingoptionsandthussolutions.• The future development of several uncontrollable situation variables is

unknown.• Thedecisionmakerisinexperiencedand/orlackstheappropriatetoolsto

determinetheconsequencesofthedecision.Simpleproblemsareoftenwell-structured.Allthebackgroundinformationrequiredformaking a sound decision is given. Further the options are clearly defined and theirinfluenceisknown.

2.4.2 DifferentapproachesofdecisionproblemsTherearemanywaysofsolvingadecisionproblem.Choosingatrandom,relyingonapastsolution,unquestioninglychoosinganexpert'ssolution,choosingbyintuition,orbyrational and systematic research. These approaches can be gathered into two groups,“heuristic” and “analytical” approach. These can be further subdivided into “general”and“specific”approach.(Grünig&Kühn,2013)

27

HeuristicapproachTheword“heuristic”originatesfromancientGreekandroughlytranslatesto“toseek”or“tofind”.Themaindifferenceandadvantageofworkingheuristicallyistheminimalamount of formal application restrictions (Grünig & Kühn, 2013). A heuristic way ofworkingistosimplify,useruleofthumbandusetrickstoreachagoodenoughsolution.Which means that since it is not an exact method there is no need to follow stricttemplatestoconfirmaccuracy.Approachingaproblemheuristicallyisawaytoreduceeffortorcomplexity.Asaresulttheoptimalsolutioncannotbefound;onlysatisfactorysolutionsarepossible(Grünig&Kühn,2013).AnalyticalapproachTheoppositeoftheheuristicapproachistheanalytical“rational”approach.Forapplyingan analytical solution the problem needs to bewell-structured according to Simon&Newell (1958). A problem can be considered well-structured if it fulfills threeconditions:

1. The problem contains only quantitative aspects or that the qualitative aspectscanbereducedtoquantitative.

2. What an acceptable solution constitutes is clearly defined in a quantitativemanner.

3. Itispossibletodevelopaprocedurewhichletstheactorfindtheoptimalsolutionwithinacertaintime-scopeandexpenditure.

Ananalyticalapproachfindstheoptimalsolutionthroughmetrics.Thisallowssolutionstobecomparedobjectivelyandimprovementstobequantified.GeneralandspecificapproachesA general decision making procedure can be adapted for many problems and aparticularsituation.Aspecificprocedureontheotherhandisnarrowlydefinedand isspecificallydesignedforaprobleminaparticularsituation.

2.4.3 FlowchartAs mentioned, a decision support tool can take many forms. One of the forms is aflowchart.Aflowchartgraphicallydisplaysasequenceofactivitiesoractions,creatingachain. The purpose of the chain is to create, provide or produce a specific output.Howeveritcanhaveseveralinputsandoutputsthatarerequirementsfor,orasaresult,of the actions (Damelio, 2011). Figure 7 shows an example of a flowchart. Thequestions-boxinthechartiscalleda“decisionpoint”.

28

Figure 7. A schematic example of a flowchart.The questions-box is a decision point thatdeterminesthefollowingstepoftheprocess.

Whyuseflowcharts?Thephrase“Apicturetellsmorethanathousandwords”isfamiliarto most people. This is also the strength of flowcharts, an accessible visualrepresentationofaprocessdescription.Insteadofreadingpagesofwordsaquickglanceof a well-designed flowchart is enough to decide the next step in the process.Furthermore a flowchart is an excellent method of presenting a new process to anaudience.(Andersen,2007)FlowchartdesignThere aremanyways of drawing a flowchart, there is no unified practice. There arehowevercommonlyusedsymbols.AccordingtoAndersen(2007)atypicalgenerationofaflowchartgoesbythefollowingsteps:

1. Definetheboundaries,start&endpoints2. Mapactivities3. Constructtheflowchartgraphically4. Resolveanyissues5. Redraw

Thestartandendpointof the flowchart togetherdefinetheboundariesandshouldbewell-defined outputs or events. The activities are identified by finding out all of thenecessarystepsthatneedtobeincludedintheflowchart.Oncetheactivitieshavebeenidentifiedthegraphicalconstructionoftheflowchartiscreatedbyplacingtheactivitiesinthesequenceinwhichtheyareperformed.Eventual“decisionpoints”areidentifiedatthis stageaswell.Once the firstdrafthasbeen created thedevelopment team resolveanyissues that thechartmighthave.The finalstep is toredrawtheflowchartwith thepurposeofincreasingthereadability.

29

2.4.4 QuantitativedataandQualitativedataThe data that is used when creating the DSM is categorized using the commoncategorizationofdatawherethedataisdeemedtobeeitherqualitativeorquantitative.Qualitativedataisthedatatypethatiscollectedinaqualitativeresearch(Bryman&Bell,2003). The following features of qualitative data are defined as described by EblingLibrary(2016):Incontrasttoaquantitativeresearch,aqualitativeresearchisbasedonsomething that is impossible to accurately and exactlymeasure. The collecteddata ofthistypeisnormallyusedtoformatheory,comparedtoquantitativedata,whichisusedfortestingexistingtheory.Quantitativedataisthedatatypethatiscollectedinaquantitativeresearch(Bryman&Bell,2003).Following featuresofquantitativedataaredefinedasdescribedbyEblingLibrary(2016):Asimplicatedbyitsname,quantitativedataisquantifiableandexactlymeasurable.Afurtherfeatureofquantitativedataisthatitisusedfortestinganexistingtheoryasopposedtodevelopingatheory,whichisthecaseforqualitativedata.

30

3 MethodologyThischapterdescribestheworkprocessofthestudyandincludesthevariousmethodsthatwereusedthroughouttheproject.Incaseswheretherearealternativemethods,amotivationofwhycertainmethodswerechosenisalsostated.Thechapterconsistsofeightpartsthatdescribethemethodologyinchronologicalorderexceptforchapter3.7which was given its own chapter due to its extensiveness. Themethodology chapterincludes detailed explanations of the research design, the preparatorywork, the datacollectionprocess,thebufferreplenishmentrate,theworktaskanalysis,thegenerationoffuturelayouts,thematerialflowsimulations,andtheprocessofdevelopingtheDSM.

3.1 ResearchdesignTheresearchdesignthatwasusedthroughouttheresearchwasthecasestudymethod,as explained by Yin (2003). The method is commonly used for creating anunderstandingofcomplexphenomenonaboutindividuals,groups,organizations,socialsituations, political situations, economics, and related phenomena (Yin, 2003). Itconsistsof six sequential stages:plan,design,prepare, collect, analyze,andshare.Themethodwasnotfollowedstrictlyintheresearch,butservedmerelyasaguidelinethatstructured the work. The reason for choosing case study as the method was itsapplicability to research projects concerning social, organizational, and managerialsystems.Further,acasestudyisappropriateincaseswheretheresearchgrouphasnoability tocontrol thesystemthat theresearch isstudiedupon(Yin,2003).This fit theresearch since the production system that was studied at Brose was an establishedsystemthatthethesisgrouphadnoabilitytocontrolormanipulate.Anotherfactorthatisrequiredforcasestudiesisthattheresearchiscarriedoutonthecurrentstateofthestudy subject, as opposed to its historical states (Yin, 2003). The thesis groupwill bestudyingtheexistingproductionsystemsolelyinitscurrentstateandbasethefindingson it,which is in linewith the requirementwhenselectinga case studyasa researchmethod.Thelistbelowfollowstheactivitiesthattookplaceineachofthestagesofthecasestudy,includingshortdescriptionsofeachstagebasedonYin’s(2003)definitionsofthem. 1.PlanThe first stage of a case study is to plan the research. This is a crucial stage forstructuringtheconsecutiveworkthatfollowsintheresearch.Thestepsthatweretakenintheplanningphaseofthethesisarelistedbelow:

• Problemdefinition• Researchquestions• Timeplan• Deliverables

31

2.DesignThe design is the link between the initial question and the collected data, whichultimately results in the conclusion of the study. Itworks as a guide for the researchgroupandfacilitatestheprocessofachievingtheresearchgoal.Stepsofthethesisarelistedbelow:

• Literaturestudy• FamiliarizationwithBrose’sproductionsystem• Identifygapsandrequireddata• Identifyinformationsources

3.PrepareThepreparationstageregardsthepreparationofthedatagatheringprocess.Thedegreeofwhichthedatamatchestherealityisheavilydependingonhowwellthepreparationsaremadepriortothedatacollection.Measuresthatweretakeninordertoassureafullyprepareddatacollectioncanbefoundinthelistbelow:

• Designinterviewguides• Constructtimestudy

4.CollectThefourthstageofacasestudyistocollectthedatathatwillbeusedasafoundationandevidenceforthefutureresults.Therearesixmajordatatypes:documents,archivalrecords,interviews,directobservations,participant-observation,andphysicalartifacts.Themethodologythatisusedwhencollectingthedatavariesdependingonthetypeofdatathatismeanttobecollected.Thedatacollectingstepsthatwerecarriedoutintheprojectarelistedbelow:

• Timestudies• Interviews• Collectionofdatafrominternaldocuments

5.AnalyzeFollowingthedatacollectionstageisananalysiswherethecollecteddataisanalyzedinordertoformaresultoftheresearch.Therearemanydifferentapproachesfordoingitandtheproblemisdecidingwhichstrategyisthebestsuitedfortheresearch.Iflackingastrategythereistheoptionof“playingaround”withthedataasanintroductiontoamoresystematicapproach.Thestepsthatweretakeninordertoanalyzethedatacanbefoundbelow:

• Combinedataandgeneratelayoutproposals• Evaluateeligibilityofproposed layoutsbysimulationandrequestsofBrose

representatives.• Analysis of previous work where the results are used as a base for

constructingdecisionsupportmodel.

32

6.ShareThesixthandfinalstageofacasestudyistoreportandsharetheresultsoftheresearch.This is the closing stage of the research and concerns the way that the results arepresented. This stage involves a number of considerations such as determining thetarget audience, structuring the report, and following certainprocedures. Toworkonthereportastheresearchprogressesisgenerallyrecommended.Themeansofsharingthefindingsofthethesisarelisteddownbelow:

• Report• OralpresentationforBroseandtheinstitute

3.2 PreparatoryworkBeforeanyactualdesignofanewproductionsystemstartedabackgroundstudyofthecurrentandfuturerequirementshadtobeestablished(Bellgran&Säfsten,2010).Theimportanceofthepreparatoryworkwastoavoidtoorapidconclusionsandmakingthewrong decisions. If the current layout were not considered there was a risk of goodsolutionsbeinglost.Abackgroundstudytypicallyentailsevaluatingthecurrentsystem(if any), identifying the needs of the future system, and studying other productionssystems.(Bellgran&Säfsten,2010)

3.2.1 LiteraturereviewA literature reviewwas conducted in order to create a theoretical framework for theresearch. The main scope of the review was to find methods of structured layoutplanning. General knowledge of production systems was already established. Theliteratureexaminedwasmainlypracticalguidesandindustrialengineeringhandbookswith topics such as evaluation techniques, interview guides, and software tools.Moretheoretical topicswere also examined, such as theory of data collection and researchmethods.Thefollowingonlinedatabasesweresearchedforliterature:

• Books24/7• AccessEngineering• Emerald

Keywords that were used when searching on the online databases were: Functional,layout, structured, planning, problem, method. Besides online catalogues Chalmers’(physical) library was frequently visited for access to reference literature and othercopiesthatwerenotavailableonline.

3.2.2 IntroductiontoBrose’sproductionsystemThe first analysis of the production systemwas conducted in the early stages of thestudy.AfewdayswerespentatBrose’sproductionfacilitywiththepurposeofgettingfamiliarizedwiththestudiedproductionsystem.Anoverallpictureoftheprocesseswasestablishedand the general characteristics of thevariousprocesseswere capturedbyworkingside-by-side theoperators.Furthermore, the thesisgroupperformedsomeofthework tasks. The flowofmaterial in between the processeswas also examined. Inaddition, a blueprint of the facility was provided by Brose, which helped to furtherimprovetheunderstandingofthefactory.

33

3.3 DataCollectionDataworksasthefoundationforanyresearch,thustheoutcomeofaresearchisheavilydependentonthequalityofthecollecteddata(Yin,2003).Followingsubchapterscoverthedatacollectionactivitiesofthethesis.

3.3.1 InterviewsOnepartofthedatacollectionprocesswastointerviewemployeesofthecompanythatwereaffectedbythelayoutofthefactory.Thepurposeoftheinterviewwastoacquireinformation for three areas: (a) the SLP, (b) themanufacturing processes, and (c) thematerial flow simulation of the layouts. The following sections include more detailedinformationabouttheperformedinterviews.IntervieweesThechosenintervieweeswereemployeesthatwereaffectedbythelayoutofthesectionof the production that was within the scope of the research. These included (a) theproductionmanager,(b)thesupplychainmanager,(c)machineoperators,and(d)truckoperators. Machine operators from all three concerned production processes (i.e.sealing,railingriveting,andCFassembly)werealsoselected.InterviewtypeBefore an interview guide could be developed, the interview type needed to bedetermined.Intheorytherearethreetypesofinterviews(Quetal.,2011).Theseare:● Structured

Thisinterviewtypeischaracterizedbyusingasetofpreparedquestionswithalimitedamountofanswers.Open-endedquestionsareavoided.

● Semi-structuredAsemi-structuredinterviewisformalandusesaguideoftopicstobecovered.Usuallythesetopicsfollowaspecificorderandwhenappropriateallowsformorein-depthdiscussion.Inthistypeofinterviewtheintervieweeisnormallyonlyinterviewedonce.

● UnstructuredThisapproach,whilestillformal,usesstrictlyopen-endedquestions,tryingtogettherespondentstoopenup.

Thetypethatwaschoseninthisresearchwasthesemi-structuredapproach.Thiswasappropriate for two reasons. The first reason was due to (a) the time-limit of theresearch.Sincetheinterviewwasonlyonepartofthedatacollectionprocess,thethesisgroup sought to use a time-efficient interview form in order to fit in the other datagatheringprocessesintothetimeplan.Semi-structuredinterviewsareinlinewiththisreasoningsincetheyaredesignedforcaseswhereeachintervieweeisonlyinterviewedonce. The secondmotivation for selecting semi-structured interviewswas (b) itsmixbetween structure and flexibility. Semi-structured interviews are normally structuredwith a prepared interview guide, but the included questions are often open-ended inorder to enable the interviewer explore areas that were previously unknown. Thestructure was necessary since one of the goals of the interview was to acquireinformation that could be used in the SLP-method and the simulation, thus specificquestions needed answer. The open-endedquestionswere necessarywhen forming adeeperunderstandingoftheproductionsystem,whichwasoneofthethreepurposesof

34

the interview.As a result, some received answerswere not relevant for the research.Theyweredocumented,butnotfurtherconsideredduringtheprojectwork.InterviewguidedevelopmentThedevelopmentprocessoftheinterviewguidecanbebrokendownintothreemajorsteps.Inthefirststepadraftoftheguidewasformed.Thesecondstepwastohaveanexternal person go through the interview guide and give his or her opinion on itscontent and structure. It was preferred that this person was experienced withinterviews. The academic supervisor of this thesiswas therefore chosen for the task.Oncethesupervisor’sopinionhadbeentakenintoconsiderationitwastimetomoveonto the third and last step of the interview guide development process, which was toadjusttheinitialguideinawaythatincorporatedthesupervisor’sfeedback.InterviewguidestructureThe interviewswere carried outwith the purpose of collecting information for threeareas: (a) the SLP, (b) the three manufacturing processes, and (c) the simulation ofmaterial flows related to the three manufacturing processes. Consequently, theinterviewguidewasstructuredbasedonthesethreepurposes.Thepurposeofthefirstsection was to acquire information that was necessary for the SLP-method. Thequestions were formed after studying the method thoroughly and needed data wasidentified. The purpose of the second sectionwas to get both a general and a deeperunderstandingofthethreeproductionprocesses.Thequestionswereformulatedbasedon the preparatory work that took place prior to the data gathering stage of theresearch.Thethirdand finalsectionof the interviewguidewasmadewiththegoalofobtaining information that could be used for the simulation of the production flow.These questions were formulated based on the type of input data required for FlowPlanner.

3.3.2 TimestudiesAnother chosenmethod for collecting datawas a time study. The alternativemethod,work sampling, requiresmanyobservationsbefore an estimation canbemadeon thesystem.Thismakes themethod excessively time consuming for the time scope of thethesis.Itwasthereforenotchosenasthemeansofcollectingtimedataoftheproductionsystem. Inorder toattain timedata thataccurately reflect the realproduction systemthe involved work tasks were video recorded and timed using the recorded footage.Thus, the time studies were performed on the footage rather than on the physicalproductionsystem. Inaddition,a limitedamountof timedatawasprovidedbyBrose.Thisdatawasusedasacomplementtothemeasuredtimesfromtherecordings.

35

3.3.3 DistancemeasurementsIn order to place objects at the right distances from each other accurate distancemeasurements were needed. This includes width and length of machines, buffers,equipment etc., but also themetrics of the factory itself. Exactmeasures of pillar andpathplacementswereneededforpreventinganymisplacement.Thesedistancescanbemeasured by a variety of methods, measuring tape, optics, and scaled drawings tomention a few. In this project the measures were gathered from the Brose plantblueprint,measured in AutoCAD 2016, and later controlled bymeasuring tape at theactualfacility.

3.3.4 MachinedataMachinedatasuchasdowntimes[%]andfaultyproducts[%]werepartlyacquiredfromthe automatic logging programs of the machines. Further data were gathered fromproductionrecords.Cycletimesandsetuptimeswereacquiredfrommachineandtaskspecifications.

3.3.5 DatacollectionfromERPsystemData for the three production processes were collected from Brose’s EnterpriseResource Planning (ERP) system. The data that were collected were the number ofvariants that were in production for the years 2018. The total amounts that wereproducedforeachvariantwerecollectedaswell.

3.4 BufferreplenishmentrateThereplenishmentratesofthebuffersaffecttheirplacementinthelayout.Buffersthatneed to be replenished more frequently should be placed in a way that is easilyaccessible for the logistics. The replenishment rate could be derived from the buffercapacities, the receivedorderquantity, and theproduct specifications; abufferwithalowcapacityandahighconsumptionrate(basedonorders)needstobetendedtomorefrequentlythanabufferwithahighcapacityrelativetoits lowconsumptionrate.Thismeans that it isnot thereplenishment rateof thebuffer themselves thatare relevant,butrathertheirmagnitudeinrelationtoeachother. The replenishment rate required both data collection and some calculations. Buffercapacitiesweremeasured by simply observing the various buffers in the production.Dataregarding theorderquantitywerecollected fromtheERPsystem.Data forsomecomponentswerenotavailableinthesystem,butwerederivedfromtheorderquantityand the product specification. The product specifications show the number ofcomponents thatgo intoeachproductand theorderquantity shows theconsumptionrate of the products, enabling the consumption rate of the various components to becalculated.Once the replenishment rateshadbeencalculated theycouldbe comparedandthebufferprioritizationwasconducted.

3.5 WorkbalanceandWorkdesignInordertomakesurethattheworkloadoftheworkersissettoanappropriatelevelandto analyze the possibility of aworker to service severalmachines on his or her own,workbalancingwasused.Thisstepwascarriedoutinparallelwiththeflowsimulation.Partof theanalysiswasbasedonworksampling in the formofvideorecordings.TherecordedactivitieswerebrokendownintotasksinthesoftwareAviX.Thisallowedfor

36

exact estimations ofwork operations. Tasks that do not yet exist and are going to beimplementedinfuturesolutionswereestimatedusingMTM-codes.MovementssuchaswalkingandturningwerealsoestimatedusingMTM-codes.Inordertomeasuretheoperator’performancesinthelayouts,therequiredworktasksand movements were calculated with mentioned methods and later also activityclassified. As the future demands of Brose’s products were known it was possible tocalculate how many operators that were needed for running the sealing and railingprocesses. Granted that the placement of the machines enable the operator to moveefficiently between them. Different solutions were tried and evaluated, thus coveringseveralpossiblelayoutvariants.

3.5.1 MethodTimeManagement(MTM)In this project the MTM-1 system was employed as it allows for the most exactestimations, although at the highest time cost of the MTM-systems. All operatormovementbetweenprocessesandworkstationswereestimatedwiththismethod.AlsosomeminortasksthatwerenotcapturedbytherecordingswereestimatedusingMTM-codes.Inaddition,somenewsolution-specifictaskswereconstructedinthesameway.

3.5.2 RecordingactivitiesFor achievingbestpossible resultwithAviX, the analyzedoperations shouldbe videorecorded.Videos canbedirectly imported intoAviX.They can thenbe edited and cutintopartsforbetteranalysisdirectlyintheprogram.Whenrecordingtheactivitiestherearecertainthingstokeepinmindforachievingthebestresult:(SolmeAB,2015)

• Theoperatorsshouldbeinformedbeforehandsothattherecordingdoesnotcomeasasurprise.

• Thevideoshouldbefocusedontheobjectbeingmanipulated.Agoodruleofthumbfornotaccidentallymissinganyhandlingistofocusontheareabetweentheoperator’schinandwaist.

• Iftheoperatorismovinglongdistancesremembertorecordhisfeet.• Donotjustrecordoneproductioncycle,doseveralinordertocapturethe

naturalvariation.• Ifpossiblelettheoperatortalkaboutwhatproblemsthatmightoccur.

Furthermore, the used method should be recorded and not the one defined ininstructions.However,itisimportanttomakesurethatthecorrectproductionaidsarebeingutilized.Theoperatormighthaveabettermethodofdoingthingsthandescribedin the instructions. The method should only be considered if it is ergonomic and ofsufficientquality.

3.5.3 CalculationofworkforceFor calculating theworkforce thework balancing suite of AviXwas used. In AviX theworkstations are first defined and then operators and machine resources can beassigned to it. Theseoperators andmachine resources can thenbe assigned tasks.Bysetting rules for the tasks such as “start after” and “end before” interrelationshipsbetween theresourcescanbeset.The frequencyof the taskscanbederived fromthedemand.

37

Some support tasks such as quality checks and container changes can however occurirregularlyasinthecaseofthisprojectwheretherearetasksthatarenotasfrequentasthemain value-adding tasks that seemingly occur at random. The possibility of tasksoccurring beneficially as well as colliding needs to be investigated, a “best case” and“worst case” scenario. Unfortunately, unplanned production stops such as materialshortage, lackingqualityandmachinefailuresarehardtoincludeduetotheirrandomnature.Thus,appropriatetimemarginsshouldbeset.Thismarginshouldbebasedonhistoricdataaswellasadesiredsafetyfactor.Forthepurposeofanalyzingwheretimecouldbebetterutilizedfortheworkersofthecurrentstate,allworkprocessesweresetupinAviX.Duetothetasksrepetitivenaturepatternsquicklyemerged.Inthesepatternsreoccurringtimegapscouldbefound.Theyaretheresultofaninternalprocessworkingfasterthanothersorduetowaitingforamachine resource to finish processing. By moving the processes closer workers cansupportotherprocesseswhiletheirtimegapoccurs.Anotherscenarioanalyzedthiswaywasthepossibilityoflettinganoperatorbeinchargeoftwoprocesses.

3.6 GeneratingproposalsforfuturelayoutSLPwasusedasamethod forgeneratingprovisional layoutproposals. Itworkedasafoundationforthe layoutdevelopmentprocess.Material flowsimulationwasaddedtothemethodandisexplainedindetailinthenextchapter.SLPwaschosenforitsgeneralacceptancewithin the fieldof layoutdevelopment; the literature studies in this thesisshowed that Muther’s (1977) SLP method was frequently cited. Furthermore, itsstructured “step-by-step” approach facilitated the advancement of themethodwhereadditionalstepscouldbeaddedrelativelyeasy.Thischaptercoversthesixstepsofthemethodthatwerefollowedinordertogeneratetheprovisionallayout.Step1:ClarifyingtheconnectionThepurposeof the first stepof theSLP-method is to identify theconcerned functionsand assessing the connections between them. Information needed for this step wascollectedthroughinterviewsinthedatagatheringprocessandthroughcommunicationwith Brose via e-mail. The analysis of the data was visualized using a triangularconnection chart. The chart lists the involved functions and rates the connectionsbetween them. The ratings show which functions that needed to be placed in closeproximitytoeachother.Step2:EstablishfunctionaldemandsThe second step of generating a layout proposal was to identify the demands of thedifferent functions. This informationwas collected during the interviews just like thepreviousstep.

38

Step3:LinefunctionalconnectionThe third step was conducted to create a visual representation of the functionconnectionsthatwereestablishedinthefirststep.Theestablisheddemandsofsteptwoneedtobeconsideredaswellwhendoingthis.Thefunctionsareplacedonebyoneandrearrangeduntilaclearlayouthasbeengenerated.Step4:DrawalternativemasterplansThe fourth step is a creative process where the actual layouts were generated anddifferent solutionsand ideascouldbe tested.Thecreatingprocesswassteeredby thedemands and function connections found in previous steps. By the end of this step anumberofdifferentlayoutsolutionswithvaryingapproacheshadbeengenerated.Thisstepislargelycreative.Tosupporttheprocess,theplantfloorwasdrawninscaleonawhite board.All objects of the facility,machines, buffers, and support functions,weredrawninthesamescaleasthefloor,cutoutwithscissor,andfastenedtothewhiteboardwithadhesiveputty.Truckpathsweredrawnwithawhiteboardpen.Step5:AssessthedifferentalternativesAlargenumberoflayoutsweregeneratedinstepfour.Asthefocusofthisstagewastocreateasmanyandasdifferentsolutionsaspossiblemanywereofinsufficientquality.Thesesolutionswerescreenedduringthefifthstepinordertonarrowthemdown.Thescreeningprocesswasmadebasedonthewishesanddemandsthatwereinitiallystatedby the company during the interviews. Themost suitable layouts were then decideduponandbroughttothenextstepofthelayoutassessmentprocess.In the next step iterative screening meetings were held with a focus group ofstakeholders. The group consisted of the CEO of Brose, the production manager, thelogistics manager and the chief industrial engineer. At these meetings the generatedlayout proposals were discussed, those of insufficient quality were discarded. Othersolutionsthatshowedpotentialwerediscussedandissueswerebroughtup.Beforethenext meeting layouts were updated according to the given feedback. This screeningprocesswasdoneiterativelyuntilonlyoneapprovedsolutionremained.Step6:DesignthechosenlayoutplantsolutionTheselected layoutswere furtherdevelopedbasedon the feedback thatwasreceivedfrom Brose in the previous step of the layout generation process. Complete CAD-drawingsofthefactoryfloorwiththeproposedlayoutsweredrawninordertogetmoreaccuratemeasurements of the layouts. The CAD-software thatwas used in this studywasAutoCAD2016. Itwas chosen for two reasons. The first reason is that (a) itwasBrose’schoiceofCAD-software.ThedigitaldrawingofthefactorythatwasprovidedbyBrosewasinthefileformatthatisusedinAutoCAD.UsingthesameCAD-softwarewasthustimeefficientsincethefilesdidnotneedtobeconvertedbeforetheycouldbeused.Furthermore,once theproposed layoutshadbeen finalized theycouldbedelivered toBrose ina format that theyarealready familiarwith.The second reason for choosingAutoCADinsteadofotherCAD-softwaresisthat(b)thereisalargeamountofexternaladd-ontoolsforitsuchasFlowPlanner,whichwasusedinthenextstageofdesigningthelayout.

39

3.7 MaterialFlowSimulationAsapartofassessingthedifferentlayoutproposals’eligibilitymaterialflowsimulationswere carriedout on the layouts. This is an addedpart of the development that is notconsidered in SLP. However, it is relevant to consider since the layout does have aneffect on thematerial flow. The software that was used for the simulationwas FlowPlanner.The softwarewas chosen for its compatibilitywithAutoCAD.Thepurposeofthesimulationwastoseehowthematerialflowwasaffectedbythelayoutsbyusingthecurrent state of the factory layout as a reference, i.e. the simulation was a way ofassessingifthelayoutsresultedinimprovedorworsenedmaterialflow.TheextentofwhichthesimulationsweredonewaskeptrathersimplesinceBrosewasonly interested in howactivities of themanufacturing processeswere affected by thevariouslayouts.Theanalysiswasthereforelimitedtoonesignificantparameter,whichwasthetransportationtimebetweendifferentpickupanddeliverypoints.Furthermore,the generated results of the simulationswere analyzedby comparing themwith eachother, thus the individual results were of lesser importance. Consequently thesimulationscouldbekeptsimpleandtheamountofinputdatathatwasneededcouldbelimited to a select few. Detail-level data that did not need to be included were forexample the acceleration of the trucks, the dimensions of the containers, and theefficiencyof the transportations.Theused input information for the simulationswerelimited to pick up/delivery points, truck speed, and operator walking speed (fortransportationofmaterialbyfoot).

3.8 DevelopmentoftheDecisionSupportModelThelayoutproposalsofthecasestudylaidthefoundationfortheDSM.Thedevelopmentincluded defining the characteristics of the DSM and finding an appropriate form forpresentingthemodel.AtheoreticaldevelopmentinstructionforthechosenpresentationformatwasusedforcreatingtheDSM.Followingtwosectionsexplainthedevelopmentprocessindetail.

3.8.1 DefiningthecharacteristicsoftheDSMBeforetheDSMcouldbedevelopeditscharacteristicsweredefinedbasedonthetheorythatwascollectedduringtheliteraturestudies.Thecharacteristicsdimensionsincludeaclassificationofthedecisionproblemandchoosinganappropriateapproach forsolvingthedecisionproblem.ClassificationoftheproblemAs explained in the theory chapter of this report, decision problems are classified aseithersimpleorcomplex.Aproblemisconsideredtobecomplexifitfulfillstwooutofthefourcriterialistedinthetheory.Thefacilitylayoutproblemwasthereforeanalyzedforthecriteriainordertoclassifyit.The first criterion is if several targets are attempted to be reached simultaneously. Aproductionsystemofahighperformancethatiscausedbyitslayoutisnotcharacterizedbyonevalue,butratherbyapaletteofvariables.Assuchwhendevelopingalayout,thegoal is to reach multiple targets simultaneously. This results in a complex problemwhereseveralgoals,someofthemevencontradictory,aretargeted.

40

The facility layout problem further fulfills the criteria of having a large number ofdecision variables. There are many decision variables in the layout problem. Somevariables are shared among all production layouts such as distance and cycle time.Therearealsovariableswhoseimportancedependsonthesituation.Healthandsafetyregulationscandifferdependingontheindustry.Medicalandfoodindustryforexamplehavestrictrestrictionsoncleanliness.Inalayoutthiscouldinfluencetheplacementsofprocesses andmachines (Motarjemi& Lelieveld, 2014).With two criteria fulfilled thefacility layout problem can already be classified as a complex problem. However, ananalysis was done on the remaining two criteria as well in order to cement theclassification.Thethirdcriterionofthelayoutproblemwasanalyzedwithregardstoifthe future development of several environmental variables is unknown. Environmentalvariablesforamanufacturingcompanycouldconstitutevariablesrelatedtothemarket.Thebehaviorofamarketmightbepossibletopredictwithinthenearfuture,howeverthe market gets harder to predict further into the future, thus the third criterion isfulfilledaswell.Thefourthandlastcriterionisnotassimpleasthepreviousones.Itisaboutthedecisionmaker’sabilitytodeterminetheoutcomeofamadedecisionwithhisorherpossessedexperienceandtools.Theanswerishighlydependingontheindividualthatismaking thedecision and the company that thedecisionmakerworks for. Since theDSMthatwasdevelopedinthisstudyismeanttobegenerallyapplicablethereisnotaspecific decision maker or company that it is made for. The decision maker that isresponsible for a facility layout is most likely a production manager of some sort.Althoughthemanagerpossessesthegeneralexperienceonproductionmanagement,theexperience of the specific production system at hand could be limited, thus it is notpossible to give a definite answeron the criteria. Furthermore, the available tools foracquiringknowledgeabouttheoutcomeofmadedecisionsdependsonwhichtoolsthecompanyhasmadeavailabletothedecisionmaker.Theanalysisshowsthatthefacilitylayoutproblemfulfillsthreeofthefourcriteria,whichclassifiesitasacomplexproblem.ChosenapproachtotheproblemAsexplainedinsection2.4.2,therearetwoapproachesforsolvingadecisionproblem:analytically or heuristically. The suitable approach was determined by evaluating thethree conditions that characterize awell-structured problem,which indicates that ananalytical approach is most suitable. Although these goals may not traditionally bemeasured in a quantitative manner their qualitative aspects can be reduced intoquantitative ones allowing for a structured approach. Reducing qualitative measuresintoquantitativeonesalsohelpssettingtherequirementsforanacceptablesolution.Ascan be concluded by the theory chapter of this report, a layout problem is a complexproblemthereforeageneralanalyticalproblemapproachisbestsuited.Certain parameters identified by Brose stakeholders and the project groups researchserveasmeasuresofeligibility.

• Amountofworkforcerequired• Efficientlogisticstransport• Productionworksindependentlyfromlogistics• Minimizenon-valueaddingtasks(walkingetc.)• Accessibility• Visibility

41

Sinceananalyticalapproachisappliedthesequalitativelyexpressedparametersneedtobe converted into a quantitative value which allow solutions to be compared in astructuredmanner.By conducting interviews on the affected personnel qualitative information of theproduction systemwas gained. Opinions on current situation and of proposed futurestates. In the interviews mainly the opinion based parameters, such as visibility andaccessibility, are touched upon. The qualitative information was then converted intoquantitative through ranking. In practice this is done by presenting alternatives thenaskingtheintervieweetorankthemfrombesttoworst.Theamountofworkforcerequiredisalreadyaquantitativemeasure,althoughonethatiscomplicatedtomeasureorcalculate.For thispurposeamixof timestudies,MTM-1andworkbalancingwasused.Thisprojectused theworkevaluationsoftwareAviX toorganize evaluate the processes of the involved areas. Activity classification, i.e. valueadding/non-valueadding,isincludedintheAviXevaluation.With theAutoCADadd-onFlowPlanner all transportdistances in the layout couldbemeasuredandanalyzed.Highlightingwherethelargesttransportlossesarelocatedandgivingindicationsofwherepossiblecongestionsmightbeaswellassupportingtheaisledesignandmachineplacements.

3.8.2 FindingasuitableDSMstructureFinding an appropriate form for presenting the DSM in a way that fits the model’spurpose was crucial for its usability and effectiveness. Themodel structure that waschosen for the DSMwas a flowchart due to its clarity and structure that it is able toprovidewith its visual representationofworkprocesses. Furthermore, the sequentialprocess representation of flowcharts suits the inherently sequential work process oflayoutplanning.The case studyon layout creationatBrose showed that it essentiallyconsistedofasequenceofactivitiesthatweredonestepbystep,whichisthesamewaythatflowchartsarestructured.DevelopingtheflowchartThe flowchartdevelopmentprocesscombined thegatheredexperiences fromthecasestudy atBrose andAndersen’s (2007) flowchartdevelopmentprocess as explained inthetheory.Figure8illustratesthefivestepsoftheflowchartdevelopment.Agroupingoftheactivitieswasaddedtotheflowchart.Thegroupsofactivitiesarecalledstagesandtheirpurposewastoclarifytheprogressofthelayoutdevelopmentprocess.Further,anadditional element was added to the flowchart that was given the name “secondaryobjective”. These are objectives that affect the layout development and should beconsideredduringthedevelopmentprocess.

42

Figure8.TheincludedstepsofdevelopingaflowchartasdescribedbyAndersen(2007).

43

4 ResultsThischapterpresentstheresultsofthethesis.Theinformationthatispresentedherearethegeneratedresultsoftheeventsthatareexplainedinthemethodchapter.Thisincludesresultsfromthedatacollection,Brose’sproductionsystem,theSLP,thelogistics,theworkdesign,thelayoutgeneration,andtheresultingDSM.

4.1 DatacollectionTheresultingdatathatwascollectedduringthedatacollectingactivitiesarepresentedinthissection.

4.1.1 InterviewsTheresultsoftheinterviewscanbefoundinAppendixA.Belowisalistthatroughlysummarizestheresults:

• Alloftheintervieweesconsidereditmostimportanttokeepmaterialandmachineclosetoeachother.

• Therearefewrestrictions.• Expectationsoftheprojectisimprovedmaterialflowandhigheroperator

utilization.

4.1.2 ERP-systemVariant data for the three production processes collected from the ERP-system ispresentedinthispartofthereport.VariantdatarailsThe produced rails are for four different car models where three of the car modelsrequiresixdifferentrailsandonecarmodelrequiresfourdifferentrails,resultingin22railvariantsthatareinproductionyear2018.AppendixBshowsthedemandforeachoftherailvariant.VariantdatasealingTheplatesaresealedforfourcarmodelswhereeachcarmodelrequiresfourdifferentplates,resultinginatotalof16plate-variants.AppendixBshowsthedemandforeachoftheplatevariantVariantdataCFIn year 2018 there will be four variants of CFs in production. All four variants havedifferentmotors,twoofthemsharethesametypeofshroudandallofthemsharethesame type of fan. Appendix B shows the demand of the different CF-variants and theresultingdemandsoftheircomponents.

44

4.2 Brose’sproductionsystemThis chapter describes Brose’s production system based on the information thatwascollected during the preparatorywork and the data collection stage. The first sectionexplains how and if the three processes are connected. The remaining sections covereachoftheprocessesindetail.

4.2.1 TheproductionsystemThree production processes are included in the study: the rail riveting process, thesealingprocess,andtheCFprocess.Figure9illustratestheflowofthethreeprocesses.Therailrivetingstationandthesealingstationprocesscomponentsthatarepartsofthedoormodules.AsaresulttheirflowsconvergeintothelineassemblyasseeninFigure9.TheCF stationprocessesadifferentproductand is thereforedecoupled from the twootherflows.

Figure9.Themappedflowsofthethreeproductionprocesses.Thecirclesinthefigurerepresenttheprocessesandthetrianglesrepresentbuffers.

45

4.2.2 RailRivetingThissectionpresentstheresultsregardingtherailrivetingprocess.Thisincludesdescriptionsoftheprocessedpart,theprocess,theprocessedpartvariants,thelogisticalactivitiesrelatedtotheprocess,andthebufferprioritization.RailsThe part that is processed in the rail riveting process is a component of the doormodulesthatenablesthewindowregulationfunction.Itisametalrailwithoneplasticpulley on each end of it. The pulleys can rotate and lead the wire that regulates thewindows,seeFigure1.TheprocessAmachineintheprocessrivetsthepulleysontotherailingsinordertokeepthepulleysinplace.Thisisthemainpartoftheprocessanditisalmostcompletelyautomatic.Themanual work that the workstation requires consists of fetching unriveted railings(storedinlargeboxescalled“gitterboxes”),feedingthemintothemachine,packingtherivetedrailingsintobins,andmovingthebinstoabufferwherethelogisticsdepartmenttransportsthemtotheassemblyline.Oneoperatorperformsthewholeprocess.Figure10visualizestheflowthroughtherailrivetingprocess.

Figure10.Therailrivetingprocess.

Themanualworkfurtherincludesmanualqualitycontrolsoftherivetedrailsisdoneintwo different intervals. The first quality control is done at the beginning of a newvariant.Atthatpoint,thecontrolisperformedonfiverivetedrails.Thecontrolintervalisdoneonceevery80thrivetedrailandisonlyperformedonasinglerail.Thecountto80 rails starts at thebeginningof theproductionof anewrailingvariant.Thequalitycontrolisdonebyrotatingthepulleysmanuallytomakesurethattheyrotatesmoothly.Further,theheightandthediameteroftherivetsaremeasuredwithacaliper.Allofthemeasuredvaluesaredocumented.

46

RailvariantsTheprocesshandlesanumberofdifferent railvariants.Thegoalof theoperator is tobuildupastockforallofthevariants.Itishandledthroughasystemthatusesthebinsthattherailsaretransportedin.Eachvarianthasitsowndesignatedbinandbylookingat the number of empty bins for a variant the operator can determinewhich type toprocessnext;alargestackofemptybinsismoreurgenttofill.Figure11showswhattherailbinslooklike.Theunrivetedrailingsarestoredneartheworkstation.Theoperatorusestheamountofunrivetedrailingsthatisneededtofillupthestockandreturnstheremainingrailingstothestorage.

Figure11.Thebinsthatareusedfortransportingrivetedrails.

LogisticsBelowisalistofresponsibilitiesofthelogisticsworkersthatarerelatedtotherailrivetingprocess:

• Transportbinsfilledwithrivetedrailstotheassemblyline• Deliverpulleysandrivets• (Emptyrecyclingbins)• (Delivergitterboxes)

According to the supply chain manager the delivery of gitter boxes and emptyingrecyclingbinscouldbeexcludedfromthestudysincetheyareseldomhandledandthelocationofthegitterboxesarefixed.

47

BufferprioritizationThe required times for the rail riveting process to empty the buffers for incomingmaterial and to fill thebuffers foroutgoingmaterial canbe found inAppendixC.Theusedraildemandandworkhoursareshownbelow.Raildemand2018:33000pcs/weekWorkhours:108,75hrs/weekAppendix B shows that the buffer for outgoing rails is filled significantly faster thanconsumingtheincomingrivetsandpulleys.Thustheaccessibilitytotherailsshouldbeprioritized.

4.2.3 PlateSealingThissectionpresentstheresultsregardingtheplatesealingprocess.Thisincludesdescriptionsoftheprocessedpart,theprocess,theprocessedpartvariants,thelogisticalactivitiesrelatedtotheprocess,andthebufferprioritization.PlatesTheplateisthepartofthedoormodulethatthecomponentsaremountedontointheassemblyline.Itismadeofplasticandhasarubbersealingthatgoesalongtheedgesoftheplates inorder toprevent theelectronicson themodule fromgettingdamagedbyexternalmoist.Figure12showsaclose-upofaplate’ssealing.

Figure12.Aclose-upoftheplateandthesealant.Thesealantisthesubstancethatcanbeseenontheedgeoftheplate.

48

TheprocessMuchliketherivetingstationthemainprocessisautomated.Oneoperatorperformsallof the manual work. The manual work consists of surrounding supporting activities.This includes fetching theunsealedplates fromabuffer, feeding them into thesealingmachineoneatatime,unloadthesealedplatesfromthemachine,andmovethemtoabufferzonewherethelogisticsdepartmenttransportsthemtotheassemblyline.Figure13illustratestheflowofthesealingprocess.Thetransportationofplatesisenabledbyracksthatcanhold132plateseach.Figure14showsarackfilledwithplates.

Figure13.Theflowofthesealingprocess.

Figure14.Arackfullofplates.

49

Themanualworktasksalsoincludeaqualitycontrolofthefirsttwoplatesineachrack.The quality control consists of taking variousmeasurements of the plates in order todetect any irregularity. The results of the controls are analogically documented inbinders.The company ishoweverplanningondoing thedocumentationdigitallyonacomputerinthefuture.PlatevariantsThe system used for decidingwhich plate variant to process next works in a similarfashion as in the rivetingprocesswhere the goal is to keep the stock level full.Rackswith sealed plates are placed in buffer queues where each queue is reserved for aspecificvariantanddependingontheamountofemptyslotsinthequeuetheoperatorknowswhichtypetoprocessnext.LogisticsThelogisticsdepartmentisresponsibleforthreeareasinthesealingprocess.Thesearelistedbelow:

• Transportunsealedplatesfromthestorageandplacingtheminthebufferforthesealingstation

• (Replenishthebarrelsthatcontainthesealant)According to theproductionmanager thesealantbarrelscanbeexcluded in thestudyduetotheinfrequencythattheyarereplenished.Notethatthequeueswithsealedplatesare transported to the assembly lineby theworkers in the assembly line andnot thetruckdrivers.BufferprioritizationThe required times for the rail riveting process to empty the buffers for incomingmaterial and to fill thebuffers foroutgoingmaterial canbe found inAppendixC.Theusedraildemandandworkhoursareshownbelow.Platedemand2018:24000pcs/weekWorkhours:108,75hrs/weekAppendixCshowsthattheincomingplatesshouldbeprioritizedintermsofaccessibilitysinceitsbufferisemptiedsignificantlyfasterthantheoutgoingplates.

4.2.4 CoolingFanAssemblyThissectionpresentstheresultsregardingtheCFassemblyprocess.ThisincludesdescriptionsoftheCF,theassemblyprocess,theCFvariants,thelogisticalactivitiesrelatedtotheprocess,andthebufferprioritization.ThecoolingfanTheCFisasecondproducttypethatBroseproduces.Itisanelectricfanwhosepurposeistocooldowncarengines.ACFconsistsofashroud,anelectricmotor,andafan.

50

TheprocessThe cooling fan process consists of two assembly processes, followed by a qualitycontrolstation.Theprocessisrelativelymanualcomparedtotherailingrivetingandtheplatesealing.Thefirststepofthecoolingfanprocessismountingtheshroudontothemotor.Theprocessismanual,butisfacilitatedbyafixturethatholdsthemotorandtheshroud.Ascrewdriversuspendedabovetheworkstationisusedforfasteningthemotor,whichfurtherfacilitatestheassemblyprocess.Figure15showsthescrewstationwherethemotorsarescrewedontotheshrouds.

Figure15.Thecoolingfanscrewstation.

AfterthefirstassemblyprocessthepartiallyassembledCFsareplacedinabufferwhereasecondoperatortakesandplacestheminapressmachinethatpressesthefansontothem.Itisasemi-automaticprocess;theassemblerplacesafanontothemachine,placestheshroudandthemotoronafixture,andexecutesthepressingfunctionthatpressesthefanontotheshroud.Figure16showsthepressmachine.

51

Figure16.Thefanpressmachine.

Thelastandfinalstageofthecoolingfanprocessisaqualitycontrol.Itisperformedbyoneoutoftwoidenticalmachinesthatworkinparallel.Theassembledcoolingfansareplaced in one of the machines where it is accelerated in order to detect rotationalimbalances that are caused by production defects.Whenever themachine detects animbalance, the operatormanually attaches clip-onweights in away that counters theimbalance. The balance control is then executed once again and eventual additionalbalancingadjustmentsareperformeduntilthefanpassesthebalancecontrol.SeeFigure17forthetwobalancecontrolmachines.

52

Figure17.Thetwobalancecontrolmachines.

Oncethecoolingfansarebalanced,theyareloadedontoacontainerinwhichtheyarestored while being delivered to customers. The operator places cardboard boards inbetween the cooling fans to prevent them for getting damaged during thetransportation.OncethecontainershavebeenloadedtheyaretransportedtostoragebythelogisticsbeforetheyaresenttoBrose’scustomers.Figure18showstheprocessflowoftheCFprocess.

Figure18.Thecoolingfanprocess.

53

CoolingfanvariantsOnly one type of cooling fan is processed in the current process. However additionaltypes are to be introduced in the future. A digital systemwill inform the worker onwhich variant to produce, as opposed to the stock-filling system that is used in therivetingprocessandthesealingprocess.Furtherinformationaboutthevariantscanbefoundinsection4.1.2wherevariantdatafromtheERPsystemarepresented.LogisticsCompared to the riveting and sealing processes the logistics department has moreresponsibilitiesregardingtheCFprocess.Followingisalistoftheirresponsibilities:

• Delivermotors• Delivershrouds• Deliverfans• Deliver small and large cardboard boards for the packaging of the finished

goods• DeliveremptycontainersforfinishedCFs• TransportfinishedCFstoastorageforoutgoinggoods

BufferprioritizationTherequiredtimesfortheCFprocesstoemptythebuffersforincomingmaterialandtofill thebuffersforoutgoingmaterialcanbefoundinAppendixC.TheusedCFdemandandworkhourscanbeseenbelow.Basedonthereplenishment/fillrateofthebuffers,theprioritizationofthebuffersispresentedinTable1.CFdemand2018:4040pcs/weekWorkhours:108,75hrs/weekTable1.PrioritizationofthebuffersintheCFprocess.Thehighestrankedbuffersneedtobetendedtomoreoftenbythelogistics.

Buffer

1 FinishedCFs

2 Shrouds

3 EmptyCFcontainers

4 Fans

5 Motors

6 Smallcardboardboards

7 Largecardboardboards

54

4.3 SLPThis sectioncovers the resulting triangular connectionchartandvisualized functionalconnections that are created during the process of generating layout proposals usingSLP.ThediagramscanbefoundinAppendixDandAppendixE.

4.4 LogisticssimulationThe logistics simulations show that the three layouts in the final stages of the layoutdevelopmenthadapproximatelythesametransportationdistancescomparedtoBrose’scurrentlayout.AtableofthevarioustransportationdistancescanbefoundinAppendixF.

4.5 WorkdesignOne of Brose’s wishes for the layout proposals is to investigate the possibility ofrationalizingthepartsoftheworkforce.Althoughtheinvestigationrequiresanumberofassumptions,hourlydemandisderivedfromyearlydemanddividedbythetotalyearlyeffectiveworkinghours.Thebalancingtimespanissetto30minutesinordertomakeitmanageable.Fortasksthatdonotyetexist,aswellaswalking,MTM-1isused.Inordertoincreasethecertaintyofthecalculationsasmalltimemarginisappliedtomosttasks.Furthermore all acceptablebalance set-ups shouldhave a timemarginof at least oneminute per half an hour (the investigated time span). This helps as a precautionarymeasureforunplannedeventssuchasmachinefailureandmaterialshortage.

4.5.1 OperatorresponsibleforbothrailandsealingmachineOne of Brose’s identified areas for improvement is the rail process. Since the railmachine only requires periodical handling, unloading/loading and occasional qualitycontrols,thereisapotentialofbetterutilizingtheoperator,seeFigure19.

55

Figure19.Thecurrentstateworkbalancefortherailrivetingprocess.

Theideaistousethesameoperatorforboththesealingandrailprocess.Theplanistomovetherailingprocessesclosetothesealing,minimizingthedistance,andintroducinga loading buffer for the sealing in order to also make the process need periodicalhandling. This enables the operator to tend both processes, while one machine isprocessingtheoperatorcanload/unloadtheother.

56

Severalscenarioswithvaryingbuffersizesanddistancesaretestedinordertocounterthis. They all however share the same result. Due to lengthy and frequent qualitycontrolsthemachinescannotbeproperlyutilizedandthedemandisnotreached.Ifthetwo most time consuming quality control processes of the sealing and rails wouldcoincidetherewouldbealmosttenminutesofdowntime,seeFigure20.

Figure20.Theworkbalanceofoneoperatorforboththerailandthesealingprocess.

Even if thequalityprocesses couldbe rationalized suchas allocating them to anotherworker or removing them all together, the scenariowould still only barely reach thedemandlevel.Inordertorealizethisideamoreextensiveautomationimprovementsareneeded.

57

4.5.2RailoperatorsupportingcoolingfanprocessAnotherattemptatbetterutilizingtherailoperatoristoinsteadsupportthecoolingfanprocess. SeeFigure21 for the current state. Two solutions are investigated. The firstsolutionentailsusingtherailoperatorassupportforoperator2ofthecoolingfaninanattempt to even out the workload. The second solution is that the rail operatorcompletelytakesoveroperator1’stasksatthescrewstation.

Figure21.Theworkbalanceofthecurrentstatecoolingfan.

Thefirstsolutionaimstoevenouttheworkloadforcoolingoperator1and2bylettingthe rail operator support operator 2. The calculations shows that letting the railoperatorsupportCFoperator2couldpotentiallyincreasetheoutputofthecoolingfanwithout affecting the rail process output significantly, see Figure 22. The disparitybetweenoperator1and2isstilltoohighhowever.

58

Figure22.Solutionone.

The second solution removes CF operator 1 altogether. Instead the rail operatorwillmanagethescrewstation.SeeFigure23,theresultsshowthatthisisarealisticsolution.Demandismet,theoperatorsfewerandtheutilizationoftheoperatorsissignificantlyhigher.

59

Figure23.Solutiontwo.

4.6 ThelayoutdevelopmentThe layout generation was done roughly through five steps. Presented here are thedifferentgenerationsoflayoutsandtheirdevelopmentprocesses.

4.6.1 CreativegenerationThe first step in generating layout proposals was the creative process described insection3.6.Theseproposalsarecreatedwithafocusonquantity.Thusmanyfalloutsideof specified requirements and can be discarded immediately in an internal screeningprocess.

4.6.2 FirstgenerationThe firstgenerationofsolutionspresentedatBrose.Thesesolutionsare theones thatpasstheinternalscreening.Thesesixsolutionsarepresentedintheformofdrawingsona whiteboard and are still in a concept stage. See Appendix G for complete list anddescription.

60

4.6.3 SecondgenerationOutofthe8solutionsofthefirstgeneration5aredeemedinsufficient.The3solutionsthatpassarefullydrawninCADandperformanceevaluated.The3solutionsthatpassare also updated according to feedback given during the assessmentmeeting. At thisstagecloserinvestigationalsoshowsthattheideaofusingoneoperatorforthesealingandrailingprocessesisnotpossible.Onefurtherrequestbythefocusgroupismade,touse the production data for 2018 instead of 2016, since the number of variants willincreasebythen.SeeAppendixHforcompletelistanddescriptionofthisgeneration.

4.6.4 ThirdgenerationInthethirdgenerationonly2solutionsremain. Oneofthesolutionsinvestigateshowthe rail process operator can support the CF process with great success. The othersolution, which includes no added truck paths, shows flaws. With 2018’s productiondataallvariantscannotfitinthedesignatedarea.Thereforeitisswiftlydiscardedandawinneremerges.Somefeedbackisgivenforafinalupdate.SeeAppendixIforcompletelistanddescriptionofthisgeneration.

4.6.5 FourthandfinalgenerationThefinallayoutistheresultoffouriterationsofassessmentandfeedback.Itfulfillstheneedsandrequirementsofnotonlytoday,butalso2018,seeFigure24.

Figure24.Thefinallayout.

61

Thewinningsolutioncapturesmanyoftheneedsandwishesofthestakeholders.Duetothe added truckpath the accessibilitywith truck is significantly improved.The addedpath allows racks being delivered close to the sealing process, reducing the walkingdistance for the operator. Even the doubled number of rack variants for 2018 can bedeliveredclosetotheprocess.Furthermorethecloseplacementoftherailandcoolingfanprocessesallowtherailoperatortosupportthecoolingfanprocess,increasingtheutilizationoftheworkforce.

4.7 TheDSMforproductionlayoutimprovementsTheDSM for layout improvements ismadeupof four stages. Each stage consists of anumberofactivities.Theimprovementprocesscanonlyproceedtothenextstageoncethepreviousonehasbeenproperlycarriedout.Thegeneralideaistofulfillthepurposeof the layoutdevelopmentwithaconcretesolutionthat isdevelopedthroughthe fourstages and the activities that they include. Following sub-chapters give detailedexplanationsofthestagesandtheactivitiesthattheycomprise.ThefinishedDSMcanbeseeninFigure25.

62

Figure25.TheresultingDSM,whichconsistsof four stages.Eachof the stage includesanumber of activities. Once all of the stages and their activities have been performedcorrectly,afinallayoutisgenerated.

63

4.7.1 ComponentsoftheDSMTheDSMisamodelthatisbasedonanumberofpreviouslydevelopedtheoriesinareasthatarerelatedtoproductionlayoutdevelopment.Table2liststheactivitiesoftheDSMandtheirorigins.Table2.TheactivitiesoftheDSMandtheirorigins.Activity OriginPurposeDefinition CasestudyFamiliarization CasestudyIdentificationofFunctionsandRelevantParameters

SLP/Casestudy

Interview SLP/CasestudyHistoricaldata CasestudyDatacompletion CasestudyDataprocessing SLPLayoutgeneration SLP/Casestudy1stand2ndstakeholderevaluation CasestudyPerformanceevaluation CasestudyAdjustment Casestudy

4.7.2 Stage1:Pre-studyThepre-studyconsistsofactivitiesthatsetupthegroundworkandgeneralstructureofthe improvement process. The expected outcome of the stage is to understand thepremiseoftheprocessandtoattainageneralunderstandingoftheproductionsystem.PurposeDefinitionThe initial activity is todefine thepurposeof the layout improvement. It will provideguidancethroughouttheimprovementprocessandfacilitateanydecisionsthatneedtobemadealongtheway.Thedecisionsthataremadecorrectlyaretheonesthattaketheprocessonestepclosertofulfillingthepurposeoftheprocess.Thisactivityalsoincludesdefining any secondary objectives that have an effect on the layout improvementprocess.

64

Thepurposeisdefinedbyansweringwhatthegoalis,whytheimprovementisneeded,andwhenthe improvementprocess isdue tobeperformed.Thepurposeof the threequestions can be found below. The formulated questionswere created based on thepurposedefiningprocessofthecasestudy.

1. Whatisthegoaloftheimprovement?Explicitly states the goal of the improvement. Could be to improve theperformanceofacertainKPIforexample.

2. Whyisthelayoutimprovementneeded?Understandingtherootcauseof the layout improvement,whichwillhelpguidetheprocess.Canbeusedassupportformakingeffectivedecisions.

3. Whendoestheimprovementprocesstakeplace?Sets thedeadline for the improvementprocessanddetermineshowmuchtimethat is allocated for it. Connected to the “what”-question; a goal that requiresextensiveworkwillmostlikelyalsorequiremoretime.

FamiliarizationThecasestudyshowsthatithelpstobefamiliarwiththeproductionsystembeforeanyimprovementworkisperformed.Itdoesnotinvolveanydeeperanalysis,butthegoalisrathertoattainageneralunderstandingofthesystem.Therefore,nospecificdataneedsto be collected at this point. The improvement team can get familiarized with theproduction system simplyby observing theproduction systemand itsworkers,muchlike in the case study, see section 3.2.2. The need for the familiarization activity isdeterminedthroughadecisionpointintheDSM.IdentificationofFunctionsandRelevantParametersOncethegeneralelementsoftheproductionsystemareunderstooditistimetomapit.Thisisdonethroughthreesteps.

1. Identifythevariousfunctionsthatareaffectedbytheimprovementprocess.Thisincludesobjectssuchasmachines,buffers,andrecyclingbins.

2. Findtheconnectionsbetweenthefunctionsinordertoseehowtheyarerelatedtoeachotherbyobservingthematerialflowsbetweenthevariousfunctions.

3. Identify the parameters that are related to the functions such as cycle times,buffer replenishment frequency, demand level, various variant data, etc. Notethattheparametersareonlyidentifiedandnotmeasuredatthispoint.

Itispossiblethatallthenecessaryinformationcannotbeidentifiedduringthisactivity.Themissing information can be found later in the data collection stage. The first twosteps come from the initial stage of SLP, where functions and their connections areidentified.Thethirdstepisbasedontheexperiencesfromthecasestudy.

65

4.7.3 Stage2:DatacollectionThedatacollectioniswheretheinformationanddatathattheimprovementworkwillbebasedonaregathered.Itincludestwotothreedatacollectionactivitiesfollowedbyadataprocessingactivity.Italsoincludesdatathatisneededforfulfillinganysecondaryobjectives. The general idea is to collect information that determines the variousrelationship strengths of the functions since this is the type of information that thelayoutgenerationprocessprimarilywilluse.InterviewThefirstactivityofthedatacollectionstageinvolvesconductinginterviewsonvariousstakeholders about the production system. The goal is primarily to collect qualitativedata such as descriptions of various work tasks or product design information thataffectstheproduction.Inordertoprovideguidancefortheinterviews,threeinterviewtopicsshouldbecovered.SteponeandthreewereacquiredfromSLP.Thesecondstepcomesfromthecasestudy.

1. StakeholderdesiresImportant to consider, as theywill be used for selecting themost appropriatelayoutlaterinthelayoutgenerationstage.

2. ProductinformationthatconcernstheproductionHighly correlated to thematerial flow of the components that are included inproducts.Forexample, it is important toknowthenumberofcomponents thatgoesintomakingtheproductsinceallofthematerialflowswillbeaffectedbytheimprovementprocess.

3. Facilityrequirements.Requirementsof the facility thatmustbe taken intoconsideration,but that theprocessisnotabletochange.Itcanincludetheplacementoffireexits,widthoftruckpaths,andplacementsofroof-supportingpillars.

The information that is collected through the interviews will indicate what type ofquantitativedata that isneeded for the improvementprocess.The interviewershouldbeopen-mindedwhenconductingtheinterviewandgatheranydatathatisdeemedtobeuseful for the layout improvement.The interview is thereforepreferablydone inasemi-structuredfashion,asitopensupforconversationsthatenabletheinterviewertodiscusssometopicsindetailifitisneeded.HistoricaldataThe collection of historical data primarily concerns quantitative data that can begatheredfromvariousdatasystemssuchasanERP-system.Basedontheinterviewandthe identification of important parameters the type of data that is needed should beknown. The source for the data can be found through the interview in the previousactivity.Thisactivityoriginatesfromthecasestudy.

66

DatacompletionTheinterviewandtheuseofhistoricaldatamightnotprovideenoughinformationanddata. If this is the case, the data needs to be completed through other means. Thisactivityisnotstrictlydefinedasitvariesdependingonthetypeofdatathatismissing.Forexample,missinginformationofamachine’sprocessingtimecouldbecompletedbytimingitmanuallywiththeuseofastopwatch.Thedatacompletionactivityisbasedontheexperienceofthecasestudy.DataprocessingThe final activityof thedata collection stage is toprocess the collecteddata,which isdonethroughtheuseoftwoofthetoolsthatareusedintheSLP-method.Thetoolsthatareusedarethetriangularconnectionchartandthelinefunctionalconnectiondiagram.Instructionsforthetwotoolscanbefoundinsection2.1.1ofthisthesisreport.

4.7.4 Stage3:CreativeProcessThecreativeprocessiswheretheproductionlayoutsarecreated.Thefocusisontestingallpossibleideas.Thelayoutsarelaternarroweddownintherefinementandscreeningstage.LayoutgenerationThe generation is done similarly to the layout generation process of the SLP-method.The idea is to test various ideas and generate multiple. CAD-softwares arerecommended as tools for testing the layouts visually, but simpler ways such as thewhiteboardapproachofthecasestudycanalsobeused,seesection3.6.Onceallideasareexhausted,thelayoutsarebroughttothenextstage.

4.7.5 Stage4:RefinementandScreeningThelayoutsthataregeneratedinthecreativeprocessarescreenedinordertofindthemost optimal solution. This is done through an iterative refinement process thatcombines feedback from stakeholders with computational evaluations of the layouts’performances.Multipleiterationsaremadeuntilasatisfactorylayoutisgenerated.FirststakeholderevaluationThefirstscreeningactivityinvolvesanopendiscussionaboutthelayoutstogetherwithstakeholders.Thelayoutsarediscussedoneatatimeandeachlayoutiseitherdiscardedor kept for further evaluation. Evaluations are done until the layouts have beennarroweddowntoa fewthatneed furtherevaluationbeforeacertaindecisioncanbemade.Theevaluationinvolvingstakeholdersisderivedfromthecasestudy.PerformanceevaluationInorder tobeable to createa strongerdecisionbasis the remaining layouts from thestakeholderevaluationareevaluatedbasedontheirperformance.Themeansofdoingitcanbethroughcomputationalflowsimulationiftheflowofmaterialisinfocus.Thecasestudyistheoriginoftheperformanceevaluationactivity.

67

SecondstakeholderevaluationThe consecutive step after the performance evaluation is a second stakeholderevaluation.Thisisanotherstepofthescreeningprocesswherethelayoutsarediscussedonce again, but by considering the generated performance results of the previousactivity. The stakeholders should have more concrete information at this point formakingamoreeducateddecisioncomparedtothefirststakeholderevaluationactivity.Foreachlayoutthefocusgroupneedstodooneofthreethings:excludethelayoutfromfurther development, choose it as the final layout, or keep it for further developmentthroughaniterativeprocess.TheDSMinFigure25showstheiterativeprocessduringtherefinementandscreeningstage.AdjustmentWhen a winning layout cannot be decided upon the stakeholders identify necessaryadjustments. The layouts are then adjusted and aligned with the feedback of thestakeholdersbeforetheyareonceagainevaluated,creatingaiterationloop.Thelayoutsareiterateduntilthestakeholdersagreeonalayoutthatbestfulfillstheirdesires,whichis the one that will be implemented into the production. The adjustment activity isderivedfromtheexperiencesfromthecasestudy.

68

5 DiscussionThis chapterof the report covers thediscussionsof the thesis. It isdivided into threesections that discuss the produced results, the used methods, and the thesis’contributiontothefieldoflayoutdevelopment.

5.1 ResultsdiscussionThissection includesdiscussionsof theproducedresults.Thediscussedtopicsaretheworkdesign, thequalityof thework sampling, theDSM’sapplicability, and the layoutdesign’simpactonwork.

5.1.1 WorkdesignIn the current state layout the rail process operator is under-utilized. Throughworkbalancing potential ways of better utilizing the operator was investigated. Thediscussionofthoseresultscanbefoundhere.OneoperatorfortherailandsealingprocessAsevidentbythebalanceresultin4.5Brose’swishofusingthesameoperatorforboththe sealingand railprocess iswithBrose’s current technology isnot achievable.Bothprocessesrequirefrequentattentionwhichlongqualitycontrolsmakeimpossible.Iftheneedforqualitycontrolwouldcoincideforthetwoprocessestheproductioncouldstopforuptosixminutes. Investing intechnology isneededinordertorealizethis idea. Itcouldlettheoperatorfocusonthesupportingtaskssuchasqualitycontrol,whichtodayis very time consuming. One alternative solutionwould be to let only one shiftworkwith one operator for the two processes and let the remaining two shifts work at ahigher rate,making up for the losses. This scenario needs some further investigationhowever. As Brose needs to work on a just in time basis with Volvo the demandvariationmightmakethisideanull.Railoperatorsupportingcoolingfanprocess.The solution, which gives the rail operator responsibility over the cooling fan screwstationshowsgreatpromise.Beingabletoreachtheaveragehourlydemandwithonelessoperatorper shift isnodoubtanexpenditure saver.This solution isnothoweverwelladjustedtoincreasesindemandastherailmachinealreadyisrunningclosetoitscapacity.

5.1.2 ThequalityoftheworksamplingThequalityofaworksampling isdependentonthenumberofsamples fora task, theworkers experience with the task and on which worker who is performing the task.(Lawrence,2004) In thisprojectswork sample study thenumberof iterative samplesaverage at around four for each task. It is hard to say whether this is an adequatesamplesizebut thevariationbetween thesampleswasminimal.Apotentiallygreatersource of error is the fact that for most tasks only one operator was sampled. It isimportanttorememberthatdifferentlevelsofexperienceandphysicalattributesexist.Thevariation in task timebetweenworkerswas foundtobegreater thanthe internalvariation.

69

5.1.3 TheDSMapplicabilityAs Wu (1997) stated that most production development processes are eitherapproached through a predetermined set of objectives that do not consider previousdesigns, or approached by improving upon a system already in place. The modeldevelopedinthisprojectimprovesuponanalreadyexistingsystem.Apositiveeffectofthis is that it takes into account historical system development processes, thuspreviouslygainedexperienceisnotlost.Thedownsideisthatitimpedescreativitysincethe development process is solely based on previous work (Wu, 1997). In order tocounteract the impeded creativity and to give the layout generation structure SLP asdescribedbyMutter&Wheeler(1977)wasapplied,whichwhenfollowedshouldgiveamoreobjectiveresult.It is hard to comment the resulting DSM’s applicability for layouts with no previousdesign,asnosuchscenarioshavebeentried.Howeveraccordingtotheorythestructureit provides should still make it relevant as a tool. There are of course otherconsiderations that needs to be captured In order to consider the DSM as a tool forgeneral layout improvement.Furtherdevelopmentof themodel shouldbebasedonalarger number of production systems. Basing the decision supportmodel solely on asingle systemas in this studymeans that its general applicabilityought tobe limited.However, the created model in this study lays the groundwork for future studies indecisionsupportmodelsforproductionlayoutplanning.

5.1.4 LayoutdesignimpactonworkTheproposedlayoutforBroserequireslessoperatorsthanthecurrent.Kamarudinet.al. (2011) found that an increased headcount might be needed to counteract anincreasednumberofvariants.Due tomore frequentset-upsandother required tasks.The proposed layout for Brose have an increase of variants across all the threeinvestigated processes, as it includes the variant number of 2018, yet it requires asmallerworkforce. This can be explained by the characteristics of Brose’s productionsystem.While it is true that thenumberofvariants increase thesetupprocess forallthreeprocessesisshort.Forthesealingprocessthetimeconsumingfirstqualitycheckmightbetroublesome,thisprocessishowevernotaffectedbythenewworkbalance.AsCohenetal.(2008)broughtupthelearningcurvemightbeaproblemwithanincreasednumberofvariants.InthecaseoftheproposedlayoutforBrosetherailoperatortendstotwoprocessesfurthercomplicating.Anincreaseofvariantsdoesnothavetoleadtoan increased learning curve. The produced parts require the same type of work nomatterwhich variant. Thedifferences are rather in shape and size. It is therefore theprojectgroup’sopinionthatthisisanon-issue.

5.2 MethodologydiscussionThissectiondiscussestheusedmethodologieswhendevelopingtheDSM.Thediscussedtopics regarding the methodology are the effectiveness of the research design, theinterview format, the buffer prioritization, SLP, thematerial flow simulation, and thechoiceofmethodformeasuringwork.

5.2.1 TheeffectivenessoftheresearchdesignThecasestudymethodologywasusedasawaytostructuretheapproachtothelayoutproblem,butwasnotfollowedstrictly.Forthatreasonitonlyneededtoprovidesomeguidance for it to be satisfactory for its intended purpose. Its ability to provide the

70

research with guidance was helpful as it helped setting the course of the projectthroughoutitsprogression.

5.2.2 InterviewformatAsemi-structuredformat forconducting interviewswasproventobeappropriatedueto its time efficiency. Planning the interview and creating the interview guidemerelyrequiredafewdaysofwork.Additionally,theapproach’scombinationofstructureandflexibilitywassuitableasithelpedacquiringspecificinformationabouttheproductionprocessesatthesametimeasdeeperdiscussionscouldbedoneinanyoftheinterviewtopics.Normallyitissufficienttoconductastructuredinterviewonlyonceifitissemi-structured according to Qu et al. (2011). In the research, the interview enabled theacquisitionofmostdata,butmissingdatahadtobe informallycollected later throughforexamplee-mailexchanges inorderto fillout themissinggaps in thedataset.Thiscould potentially have been caused due to the research group’s inexperience ofdevelopmentwork of this kind. It should be added that an open communication flowwiththestakeholdersisnotnegative,butshouldbedonewiththepurposeofupdatingthemaboutthedevelopmentprogressandreceivingfeedback.Regarding the topics of the interview guide, collecting simulation data throughinterviews was proven to be an inefficient way since concerned stakeholders do notpossessthequantitativedatathatissoughtafter.Gettingaccesstorelevantdatathroughan ERP system as done in the researchwas a significantly efficient way of collectingsimulationinputdata.

5.2.3 BufferprioritizationThebufferprioritizationyieldedarankingofthevariousbuffersregardinghowoftentheyneededtobetendedtobythelogistics.Buffersthatneedtobereplenishedmoreoftenwerethenplacedmoreeasilyaccessible.ThemethodwassupportedbyLin&Sharp’s(1997)theoryaboutefficientproductionflow,where“spacerelationship”(includesaccessibilityandvisibilityofmaterialandmachines)isoneoutofthethreeimportantcriteriaforefficientflow.

5.2.4 SystematicLayoutPlanning(SLP)ThisprojectisinspiredbyMuther’ssystematiclayoutplanning(1977).Whileoverhalfa century old approach it is still relevant today. The strength is in its simplicity. Themethodisaccessibleandstructured.Steponetothree,Clarifytheconnection,Establishfunctional demand and Line functional connection offer a structured start of a layoutproject,acategorizationofthefunctionsandtheirconnections.Stepfourtosix,Drawalternativemasterplans,Assessthedifferentalternatives,Designthechosenplantsolution,isconcernedwiththegeneration,evaluationandselectionoflayout. (Muther&Wheeler, 1977) These steps iswhere themajor development havetaken place. Today digital tools help speed up the process. CAD programs allowredrawingof a layout inminutes.Amodelof amachine can simplybedragged in thelayout rather than redrawn by hand as was needed before the digitalization. Themethods of evaluation have also greatly developed. Today complex models can besimulated relativelyquickly.Material flowandworkdesignwhichwas investigated inthisprojectisonlytwoofmanypossiblesimulations.

71

The accuracy of the simulations is dependent on the quality of the data input.Furthermorethedataneedstobequantitative,puttingemphasisonthedatagatheringandconversionofqualitativedataintoquantitative.SLP describes an evaluation of applicability through subjective measures. Selectedlayout weighted criteria are valued according to the scale absolutely perfect, (A=4),efficientsolution(E=3),interestingsolution(I=2),ordinarysolution(O=1),unimportant(U=0),unwanted(X=-).Theappropriatenessofthesolutionisgainedbymultiplyingthecriteriaweightwiththe layoutproposalssolutionforsaidcriteria(Muther&Wheeler,1977). Evaluating by estimation brings certain problems. As Lawrence (2004) states,estimations risk bringing problems such as the creation of bottlenecks. Furthermoremaking an estimation based on historical datamakes the result prone to Parkinson’sLaw, which states that the amount of time required to complete a task is directlyproportional to the time available.Which implies that the standard time increases ordecreasesbasedontheamountofavailabletime.(Lawrence,2004)Intheprojecttherewasalsoanemphasisoniterativeevaluationandcorrection,tryingtobalancethewishesanddemandsofthestakeholders.SLPisagreattoolforgeneratingideasbutlacksinrefiningideas.SteponetofourinSLPisfordevelopingalargenumberofideas.Stepfiveandsixisselectingwhichideashouldberealized.(Muther&Wheeler,1977)Asmentionedinthemethodchapterwehadfouriteratingevaluationstepswithimprovementsofthelayoutsateachstep.

5.2.5 MaterialflowsimulationThe material flow was evaluated for the logistics flow and the production flow. ThematerialflowwasevaluatedthroughflowsimulationsusingthesoftwareFlowPlanner.Its effectiveness as anevaluation toolwas satisfactory for its intendedpurpose in theresearch. Only simple simulations with a low level of detail were needed since thegenerated data was only used for comparing the material flow performance of thecurrentlayoutwiththematerialflowofthevariousgeneratedlayouts.Thesimulation’sability to mirror the real material flow’s performance was therefore of lesserimportance. The results that it was able to produce gave an indication of the flowperformanceof thedifferent layouts, proving thatmaterial flow is indeedan effectiveparameterforassessingtheproductionflowasstatedbyLin&Sharp’s(1997).

5.2.6 ChoiceofmethodformeasuringworkAs stated in theory according to Lawrence (2004) there are threeways ofmeasuringwork. Direct observation, estimation and standard data. The approach in this projectwas a hybrid of direct observation and standard data, namely a video recordedtime/workstudyandMTM-1.Usingestimationasawaytomeasuretheworkwasquicklyruledout.Theflawsofusingestimation are many. Problems like bottlenecks are often formed as a result ofestimation.(Lawrence,2004)Iftheestimationismadewithhistoricaldataasbasistheresult is also prone to Parkinson’s Law (Lawrence, 2004). The ideawas to useworksamplingformappingtheoperator’staskswhenpossible.Taskswhichdidnotyetexistwere constructed by MTM-1. The operator's movement in the layout was alsoapproximatedwithMTM-1.Fewprocesseswerealsotimedbystopwatch.

72

Theworksamplingwasdonemainly throughvideorecording,eventhoughstopwatchtechniques are faster and require less analysis, thus allowing for more iterations.(Lawrence, 2004) The choice of using video recordings despite this is because itcaptures much more information, such as ergonomics. It is also possible to post-recording analyze the video for something that was missed during preparation. Afurther argument for video recordings is that AviX support frame by frame analysis,perfectly timing tasks (Solme AB, 2015). MTM-1 was selected due to previousexperienceofusingthemethod.AtChalmerstherearealsocertifiedMTMpractitioners,whichcouldsupporttheprojectifanyproblemswouldarise.MTM-2and-3wasruledoutbecauseoftheirmorebasicnature.(Lawrence,2004)

5.3 ContributiontothefieldPrevious research indicates that production layout development can yield potentialsavingsof greatmagnitude (Drira et al., 2007).Themostwidely acceptedmethod forsupporting a development process isMuther andWheeler’s (1997) SLP-method. It isreferencedoftenintheliterature,whichindicatesthatthemethodiswidelyaccepted.Itwashowevercreateddecadesagoandhasnotbeenupdatedeversince.TheDSMusesthecorepartsoftheSLPandfurther incorporatescontemporarytoolssuchasvariouscomputational simulations. Furthermore, theDSMdeveloped in this research seeks toprovide a more complete model that supports the whole development process. Forexample,SLPassumesthattheuseralreadypossessesinformationthatisnecessaryforassessing therelationshipstrengthsof thevarious functions.TheDSMguides theuserthroughthedatacollectingactivitiesthattakeplacepriortoassessment.

73

6 ConclusionThis isthefinalchapterofthereportanditsummarizesthefindingsoftheperformedstudy. The research sought to answer two research questions thatwere stated at thebeginning of the research. These questions are answered in this chapter. The twoquestions are followed by a final section that includes suggestions for future workwithinthefieldoflayoutdevelopment.RQ1:Howisaproductionlayoutoptimized?Itwas found that a production layout could be optimized through using a structuredapproach,asprovedbytheprojectsresultingDSM.Anoptimizationgenerallystartswitha thoroughpreparation.Theproductionsystem,which is tobedevelopedneeds tobeunderstood. Why does a new layout need to be developed? What potential areas ofimprovementsexist?Statingthepurposeservesasaninitialactivitycanhelpmaketherightdecisions later.Following thepurposeabenchmarkingof thecurrent layoutwillhelpdefine the current state. For this thedataneeds tobe gathered.Thedata shouldreflecttheproductionsystem’sperformanceaswellasthedesiresofthestakeholders.TheBenchmarkingshouldwithbenefitbedonewiththesameevaluationtoolsintendedfor the new layout proposals.When the current state is defined a creative process ofgenerating layoutproposals can start. It canbe a good idea to focusonquantity overqualityatfirst.Throughaniterativeprocessofevaluation,filteringandrefinementtheselayoutscanbedevelopeduntila(orseveral)winneremerges.HowwasBrose’sproductionlayoutoptimized?Brose’s layoutwasoptimized inawaymuch like theDSMdescribes. Inaddition therewasafocusonworkdesignandhowtobetterutilizetheoperators,thecase’s“specialobjective”.Howcantheutilizationofprocessesbemaximizedwithregardstoproductionlayout?Awayofmaximizingtheutilizationofprocesseswithregardstoproductionlayoutscanbetominimizedistances.Notonlycanitdecreasethetimespentonnon-valueaddingtaskssuchaswalkingitalsoallowscross-processsupport.AsinthecaseofBrosewherethe rail operator could support the cooling fan process so that one less operatorwasneededfortheprocess.RQ2:WhatshouldbeincludedinaDSMforlayoutdevelopment?ThegoaloftheDSMwastosupportthelayoutdevelopmentprocessfromstarttofinish.Thatis,fromaninitialpurposestatementoftheimprovementtotheimplementationofafinallayout.ThecontentoftheDSMarethestepsthatturnthepurposestatementintoatangibleproductionlayout.Thedevelopedmodelinvolvesfourconsecutivestages.

1. Pre-study2. Datacollection3. Creativeprocess4. Refinement

74

Eachstageismadeupofanumberofactivities.Explanationsaregivenforeachactivityin order to create a tangible output that is neededwhen performing the subsequentactivity.Inaddition,thereareoptionalactivitiesthatareperformedifthecollecteddatais not sufficient. These can be found in the pre-study stage and in the data collectionstage. A so called decision point is used for checking if the collected information issufficient or not. If not, the user must perform the optional activities to fill out theinformation gaps. Once the user has performed all of the stages and their attachedactivitiesproperlyasatisfactorylayoutisgenerated.Howaretheessentialparameterswithregardstoproductionlayoutidentified?Thewaythatimportantparameterswereidentifiedwasbyobservingthepurposeofthelayout development. A simplified approach is to consider all parameters that areaffectedoraffectingthepurposeasrelevantparameters.Forexample,ifthepurposeistominimize the lead time then transportation distance is a parameter that should beconsideredessential.WhichmethodscanbeincludedinaDSMforlayoutimprovements?ThemethodthattheDSMisbasedonisMuther&Wheeler’s(1977)SLP-method.Thisismainly since its intended purpose is the same as for the DSM, making it easilycompatible. The various steps of the SLP were analyzed and the methods that couldeither enhance the existing steps or expand the method were included. Examples ofenhancingmethodsissimulationsandamethodthatexpandsthemodelareinterviews.Howisalayoutdevelopedinastructuredmanner?A layout can be developed in a structured manner by basing it on a pre-developedmethod.Themethodensuresthatthenecessarystepsaretakenandthatnoimportantconsiderationsareleftout.Anotherimportantaspectofstructuredworkistoestablishthe core of thedevelopmentproperly during its initial stages. The “what”, “why”, and“when” questions in the purpose definition stage of the DSM were created for thispurpose.All of thedevelopment activities throughout theproject shouldbe answeredusing the purpose statement. This will help guide the work and provide structurethroughoutthedevelopmentwork.

75

FurtherworksuggestionThissectionincludessuggestionsoffuturestudieswithinthefieldofproductionlayoutdevelopment.I.ForthecaseatBrose OneinitialwishofBrosewastoinvestigatethepossibilityofusingthesameoperatorforboth the rail and sealing process. Due to machines needing regular attendance andlengthy quality checks this was not possible. What can be done to possibly make itreality? Lengthyqualitycontrols Dotheyneedtobesolengthy?Coulditbepossibletoshortentheprocesses?Mostofthecontrolsinvolveseveralsamples,woulditbeacceptabletodecreasethesamplesizeandinstead make more frequent controls? From a work balance perspective these areinterestingquestions.Whetheritispossibleornotunfortunatelyfallsoutsidethescopeofthisproject.Butthetopiccouldbeinterestingforfurtherstudies. Possibilitiesforincreasedautomation Anotherwayof solving theworkbalancing issuescouldbe toautomate theprocessesfurther. A pick and place robot for the railing processes could decrease the need foroperatorattendancedrastically. Findingaway toautomate the loadingprocessof thesealingprocess is alsoan interesting topic.A solutionwhere theplate rack isdirectlyunloadedbyarobotcouldbenefittheworkbalancingoftheoperatorgreatly.Increasingtheautomationlevelcouldmakeitpossibletokeepthecurrentlevelofqualitycontrolanduseoneoperatorforbothprocesses. II.RegardingfurtherdevelopmentoftheDSM. AsofnowtheDSMhasonlybeentriedforoneproductionsystem.Inordertoconfirmitsusefulnessthepriorityshouldbeapplyingittootherlayoutdevelopmentcases.Allcasesareofinterestbutthosewhichlackanypreviousdesignareofspecialinterest,theDSMismeantasageneraltoolforlayoutdevelopmentinfactories.Facilitiesofvaryingsizesare also of interest. Brose is amedium sized production facility, if the DSM is fit forlargerfacilitiesremainstobetested. Another interesting topic to investigate is the socalled “specialobjective”of theDSM.Depending on the situation the stakeholders might have special requirements whichrequire an investigation that is outside of conventional layout work. The Brose caseinvolvedworkdesignandhowoperatorscansupporteachotheracrossprocesses.Howwill other special objectives interact with the DSM? Some objective might even bebeneficialtomakepermanent.

76

References3DS.com,(2016).CATIA.[online]Availableat:

http://www.3ds.com/products-services/catia/[Accessed17May2016] Andersen,B.(2007).Businessprocessimprovementtoolbox.ASQQualityPress,2nd

edition. Autodesk.com,(2016a).AutoCADvideo.[Online]Availableat:

http://www.autodesk.com/products/autocad/autocad-video[Accessed16May2016]

Autodesk.com.(2016b).AboutAutodesk.[Online]Availableat:http://www.autodesk.com/company[Accessed16May2016]

Avix.com.(2016).AviX-ResourceBalance.[online]Availableat: http://www.avix.eu/produkter/avix-resource-balance/[Accessed7June2016]

Bellgran,M.&Säfsten,K.(2010).ProductionDevelopment:DesignandOperationofProduction Systems.London:Springer

Brose.com.(2016).Products-BroseFahrzeugteile.[online]Availableat: http://www.brose.com/en/Products/[Accessed16February2016].

BrymanA.&BellE.,(2003).Businessresearchmethods.Oxford,NewYork:OxfordUniversityPress

Cohen,Y.,Dar-El,E.M.,Vitner,G.,Sarin,S.(2008).Optimallayoutandworkallocationinbatchassemblyunderlearningeffect.IndersciencePublishersLtd.

Damelio,R.(2011).Thebasicsofprocessmapping.CRC/ProductivityPress2ndedition. Drira,A.,Pierreval,H.,Hajri-Gabouj,S.(2007).Facilitylayoutproblem:Asurvey.Annual

ReviewsinControl. EblingLibrary,(2016).NursingResources:QualitativevsQuantitative.[online]Available

at:http://researchguides.ebling.library.wisc.edu/c.php?g=293229&p=1953453[Accessed18February2016]

Grünig,R.&Kühn,R.(2013).SuccessfulDecision-making,Asystematicapproachtocomplex problems.Thirdedition.Springer

Kusiak,A.&Heragu,S.S.(1987).Thefacilitylayoutproblem.Winnipeg,Manitoba:UniversityofManitoba.

LalitNarayan,K.,MallikarjunaRao,K.,Sarcar,M.M.M.(2008).ComputerAidedDesignandManufacturing.NewDelhi:Prentice-Hallof IndiaPrivateLimited.

Lawrence,S.(2004).Maynardsindustrialengineeringhandbookchapter5.1–MeasurementofWork.Fifthedition.McGraw-Hill.

Lin,L.&Sharp,G.(1997).Quantitativeandqualitativeindicesfortheplantevaluationproblem. EuropeanJournalofOperationalResearch116.

Meller,R.D.&Gau,K.Y.(1999).Thefacilitylayoutproblem:RecentandemergingtrendsandPerspectives.Auburn,Alabama:AuburnUniversity.

Motarjemi,Y.&LelieveldH.(2014).FoodSafetyManagment-Apracticalguideforthefoodindustry. Academicpress

Muther,R.&Wheeler,J.D.(1977).Förenkladsystematisklokalplanläggning. Stockholm:SverigesRationaliseringsförbund

Benjaafar,S.,Heragu,S.S.,Irani,S.A.(2002).Nextgenerationfactorylayouts:Researchchallengesandrecentprogress.Linthicum:InstituteforOperationsResearchandtheManagementSciences.

Solidworks.com,(2016).SolidworkProducts.[online]Availableat: http://www.solidworks.com/sw/3d-cad-design-software.htm[Accessed17May2016]

SolmeAB.(2012).AviXMethodTheoryManual.SolmeAB,Göteborg.

77

SolmeAB.(2015).AviX4UserManual.SolmeAB,Göteborg. Simon,H.A.&Newell,A.(1958).Heuristicproblemsolving;thenextadvancein

operationsresearch. OperRes1–10. Turbanetal.(2014).DecisionSupportandBusinessIntelligenceSystems-Ninthedition.

Pearson ProPlanner.com,(2016a).FlowPlanner.[online]Availableat:

http://www.proplanner.com/en/products/flow_planner/[Accessed4April2016]

ProPlanner.com,(2016b).TechniquesfortheEvaluation,Reduction,andEliminationofExcess MaterialFlowwithinIndustrialFacilities.[online]Availableat:http://www.proplanner.com/documents/filelibrary/documents/papers_case_studies/SLY_2012_Flow_Techniques_BF8BD45519999.pdf[Accessed17May2016]

Qu,S.Q.&Dumay,J.(2011).Thequalityresearchinterview.QualitativeResearchinAccounting&Management.

ScottMorton,M.S.&Gorry,G.A.(1971).AFrameworkForManagementInformationSystems. SloanManagementreview,Vol13,No.1,pp.55-70.

ScottMorton,M.S.&Keen,P.G.W.(1978).Decisionsupportsystems:AnOrganizational Perspective.Reading,MA:Addison-Wesley.

Weck,M.,Eversheim,W.,König,W.,Pfeifer,T.(1991).Chapter1.1–Strategiesforthefactoryofthefuture.Aachen:ElsevierLtd.

Wu,B.(1997).ManufacturingSystemsDesignandAnalysis:ContextandTechniques.London:Chapman&Hall.

Yin,R.(2003).Casestudyresearch:DesignandMethods.London:SagePublications.

78

APPENDIXAppendixAInterviewresultsUnansweredquestionsduetointervieweesinabilitytoansweriseithermarkedwithadash-symbol(“-”)orcompletelyexcludedifnoneoftheintervieweescouldgiveananswer. SLPanswersInterviewresultsoftheproductionmanagerandthesupplychainmanagerregardingSLP.Answersfromeitherofthemanagersthatarenotconsideredinthelayoutdevelopmentarelistedinthecolumntothefarright.

Productionmanager Supplychainmanager Notconsideredinthelayoutdevelopment(excluded)

Functionsneededtobekeptneareachother

material,machine material,machine Truckchargingstationneedsaccesstoelectricoutputandwater

Structuralrequirementsoffactory

Fewrestrictions.emergencyexitseasilyaccessible.pillars,meetingcornerandsealingmachinearefixed.Norestrictionsforelectricsupply,ventilation,andopenspace.

two-waytruckpaths:3,6mwide,one-waytruckpaths:2,7mwide,emergencyexitseasilyaccessible,0,8mopenspaceinfrontofelectricalcabinets.Pillarsarefixed.meetingcornerpreferablycentrallylocated.Norestrictionsforelectricsupplyandventilation.

-

Expectationsofthelayoutdevelopment

higheroperatorutilization,optimizedmaterialflow

optimizedmaterialflowforanincreasedamountofproductvariants,optimizedassemblytimes,decouplingofproductionandlogistics

-

Miscellaneousanswers

- productionareaspreferablyaccessiblefrommultiplesidesforlogistics

-

InterviewresultsoftheoperatorsregardingSLP.

Operator

Functionsneededtobekeptneareachother material,machine,motorbufferinCF

79

ManufacturingprocessanswersInterviewresultsoftheproductionmanagerandthesupplychainmanagerregardingthemanufacturingprocesses.Answersfromeitherofthemanagersthatarenotconsideredinthelayoutdevelopmentarelistedinthecolumntothefarright.

Productionmanager

Supplychainmanager

Notconsideredinthelayoutdevelopment(excluded)

Bufferplacement Canbeplacedfreely,butwithinthenewproductionarea.

Canbeplacedfreely

Morevisualaids

Alternativemeansfortransportation

Ifitresultsinsavings

Doesnotresultinanysignificantsavings

-

Set-uptimes - Insignificant -

Miscellaneousanswers

Sealingplatesmightbefedintenpiecesatatimeinfuture

- Anadditionalbalancingmachinewillbeinstalledinthefuture.Notconsideredduetouncertaintyoftheinvestment

80

Interviewresultsoftheoperatorsandthelogisticsdeoartmentregardingthemanufacturingprocesses.Answersfromeitheroftheworkersthatarenotconsideredinthelayoutdevelopmentarelistedinthecolumntothefarright.

Operators Logistics Notconsideredinthelayoutdevelopment(excluded)

Eventualimprovementareas

workareaaroundsealingiscluttered,clearnessofworkspace,placemotorsnearassemblystationofCF

biggermarginsfortruckstoturn

unergonomicloading/unloadingofracksinsealing,hardtotransportracks,fewerunbalancedfans,shortercycletimesofbalancingmachinesinCF,unergonomictopickcardboardboards,moreandclearerfloormarkings

Workpace goodpacewhentherearenoproblemsinsealing,stressfulwhenworkingaloneinCF,goodpaceinrailingduetothelongcycletimeofthemachine

requiresexperienceoftheproductionsystembeforeareasonableworkpaceisachieve

-

Miscellaneousanswers

pathwidthsofsealingaregood,

- -

AppendixBVariantdatafromERPsystemRailrivetingDataonthedemandoftherailvariants.Railvariant Demand2018(qty/variant/week) Relativedemand

1-22 1500 5%pervariant

Sealing

Dataonthedemandoftheplatevariants.

Platevariant Demand2018(qty/variant/week) Relativedemand

1-16 1500 6,25%pervariant

81

CoolingfansDataonthedemandofthefourCFvariants.

CFvariant Demand2018(qty/week) Relativedemand

1 200 5%

2 2100 52%

3 1200 30%

4 540 13%

82

Dataonthedemandofthefourmotorvariants.

CFvariant

Motorvariant

Motordemand2018(qty/week)

Relativedemandofmotors

1 1 200 5%

2 2 2100 52%

3 3 1200 30%

4 4 540 13%

Dataonthedemandofthetwoshroudvariants.

CFvariant Shroudvariant Shrouddemand2018(qty/week) Relativedemand

1 1 2300 57%

2

3 2

1740 43%

4

Dataonthedemandoffans.

CFvariant Fanvariant Fandemand2018(qty/week) Relativedemand

1 1 4040 100%

2

3

4

AppendixCBufferprioritizationRailriveting

Demand(pcs/hr)

Containercapacity(pcs/container)

Requiredtimetoempty/fillacontainer(h)

No.ofcontainersinbuffer(pcs)

Requiredtimeforemptying/replenishingthebuffer(h)

83

Rails 303,45 40 0,13 5 0,66

Rivets 606,90 40000 65,91 4 263,64

Pulleys 606,90 40000 65,91 4 263,64Platesealing

Demand(pcs/hr)

Containercapacity(pcs/container)

Requiredtimetoempty/fillacontainer(h)

No.ofcontainersinbuffer(pcs)

Requiredtimeforemptying/replenishingthebuffer(h)

Incomingplates

220,69 132 0,60 1 0,60

Outgoingplates

220,69 132 0,60 5 2,99

84

Coolingfan

Demand(pcs/hr)

Containercapacity(pcs/container)

Requiredtimetoempty/fillacontainer(h)

No.ofcontainersinbuffer(pcs)

Requiredtimeforemptying/replenishingbuffer(h)

Motors 37,15 180 4,85 1 4,85

Shrouds 37,15 22 0,59 3 1,78

Fans 37,15 60 1,62 2 3,23

Largecardboardboards

2,06 100 48,45 1 48,45

Smallcardboardboards

19,26 560 29,07 1 29,07

EmptyCFcontainers

0,69 1 1,45 2 2,91

FinishedCFs

37,15 54 1,45 1 1,45

85

AppendixDMuthertrianglechart

RailingandSealingprocess

86

CoolingFan

87

AppendixELinefunctionalconnectionRivetingandSealingprocess

88

Coolingfan

89

AppendixFTransportationtimes

AvgTripTime(Mins)

Paths Current Layout1 Layout2 Layout3STORAGE_PLATES+BUFFER_SEALING 0.77 0,78 0,74 0,66SEALING_OUT+ASSYLINE 0.46 0,47 0,47 0,47STORAGE_RAILING+MACHINE_RAILING_IN 0.47 0,6 0,53 0,47RAILING_OUT+ASSYLINE 0.53 0,64 0,53 0,6STORAGE_MOTOR+BUFFER_MOTOR 0.39 0,43 0,4 0,52STORAGE_FAN+BUFFER_FAN 0.44 0,4 0,43 0,61STORAGE_SHROUD+BUFFER_SHROUD 0.53 0,51 0,46 0,43CF_OUT+STORAGE_CF 0.37 0,57 0,48 0,52

90

AppendixGFirstgenerationandscreeningThefirstgenerationoflayouts.Herethefocuswasonquantity.Thelayoutsweredrawnin scale on a whiteboard. These layouts passed the requirements and were thuspresented.SeebelowforthefirstgenerationlayoutswhichwerepresentedatBrose.

Layout1–Discardedduetounoptimizeduseageoffloorarea

91

Layout3–Passedtothenextstage

Layout5–Discardedduetoangledtruckpath

92

Layout6-Discardedduetounoptimizeduseageoffloorarea

Layout7-Discardedduetoangledtruckpath

93

Layout8-Passedtothenextstage

94

AppendixHSecondgenerationandscreeningThesecondgenerationwascompletelydrawninAutoCadandincludedallcurrentobjects.InordertogivemorebasisforevaluationthelayoutswerealsoperformancemeasuredthroughmaterialflowsimulationinFlowPlannerandtheworkbalancewascheckedinAviX.

Layout3–Discardedduetobadworkbalance

95

Reinventedlayout3–Passedtothenextstage

Layout8–Revisedandpassedtothenextstage

96

AppendixIThirdgenerationandscreeningThisisthelastofthescreeningprocesses.Herethevariantdataof2018wasadded.Thereforethereisasignificantincreaseinobjectsinthelayout.Thesesolutionswerejustasgenerationtwoperformancemeasured.

Reinventedlayout3-Passedtothenextstage

97

Revisedlayout8–Discardedduetolackofspacefor2018variantlevels