a new paradigm of manufacturing selling use instead of selling systems

8/13/2019 A New Paradigm of Manufacturing Selling Use Instead of Selling Systems

1/8

Japan-USA Symposium on Flexible AutomationJUsFa 2002

A New Paradigm of Manufacturing Selling Use Instead of Selling Systems

C. - G Seliger echnical University of Berlin German0S J. HuJ Y Koren :he University of Michigan US0

Abstract Current manufacturing system development for medium to large volume production follows acradle to grave strategy. For each new product model, a complete new manufacturing system has to bedesigned and constructed with very little of the existing system being re-used. Reconfigurablemanufacturing systems designed for product families are resource-effective systems that can adapt quicklyto product changes and be used to manufacture products of the same family for different manufacturers.Rather than selling systems to different manufacturers, the manufacturing system builder can design,construct and manage the reconfigurable system and sell parts to the manufacturers, significantly reducingthe frequent design, construction and ramp-up of the systems. On the other hand, manufacturers can leasethe reconfigurable systems from the builder to reduce cost associated with facility management and overcapacity. In this context information about the use-intensity of production equipment becomes key toenable its high availability . Usage information can be applied in the fields of preventive maintenance andthe re-use of modules form reconfigurable systems to allow cost-efficient adaptation and reconfiguration ofproduction equipment between the different usage phases. An example will be used to illustrate theeconomical benefits of such a new paradigm in manufacturing and to point out the call for action.

1 IntroductionDedicated manufacturing systems for medium tohigh volume production have traditionally beendesigned following the life cycle model ofcradle to grave . That is, given the design ofone or two products, manufacturing systems aredesigned specifically for these products. Forexample, auto manufacturers spend over $500million every three to four years to develop themanufacturing systems, mainly stamping andassembly systems, for the vehicle bodies of aspecific model or associated variants. Themanufacturing systems are usually designedfrom the ground up, with little or no componentsof the manufacturing systems being reconfiguredor reused for the next model. The developmentof such system takes time. In addition, there aremajor quality and productivity problems with thesuch designed systems since very littleknowledge learned from the operations of theprior system are transferred to the new. Thequality and reliability problems take a long timeto debug and ramp up. As a result, manufacturersmay miss the short windows of opportunity forintroduction of new products.When demand for the product s high, the cost perpart is relatively low. Dedicated manufacturingsystems are cost effective as long as demandexceeds supply and they can operate at their fullcapacity. But w th increasing pressure from globalcompetition and over-capacity built worldwide,

there may be situations in which dedicated lines donot operate at full capacity. Due to the lowadaptability, dedicated systems can not respond tomarket changes in either volume or product varietycost effectively.Flexible manufacturing systems (FMS) can producea variety of products, with changeable volume andmix, on the same system. FMSs consist ofexpensive, general-purpose computer numericallycontrolled (CNC) machines and otherprogrammable automation. Because of the singletool operation of the CNC machines, the FMSthroughput is lower than that of DML. Thecombination of high equipment cost and lowthroughput makes the cost per part relatively high.Therefore, the FMS production capacity is usuallylower than that of dedicated lines and their initialcost is higher.A new type of manufacturing system,reconfigurable manufacturing system (RMS), isnow being developed to cope with the largefluctuations in product volume and mix caused bythe changing market conditions [Koren, et al,1999]. This is achieved through: Design of a system and its machines withconfigurations which may be adjusted at thesystem level [e.g., adding machines orchanging configurations] and at the machinelevel [changing machine hardware andcontrol software; e.g., adding spindles and


2/8

axes, or changing tool magazines andintegrating advanced controllers].Design of a manufacturing system around thepart family, with the customized t1exibilityrequired for producing all parts of this partfamily. (This reduces the system cost.)

The RMS is designed to cope with situations whereboth productivity and the ability of the system toreact to change are of vital importance. Threecoordinates - capacity, functionality, and cost -define the difference between RMS and thetraditional DML and FMS approaches. While DMLand FMS are fixed in capacity-functionality, RMScapacity and functionality change over time as thesystem reacts to changing market circumstances(Figure 2).

MachineStructureSystem focusScalabilityFlexibility

Table 1: RMS combines features of dedicatedand t1exible systems (Koren eta , 1999).Independent of the types of manufacturingsystems, the current practice is the same: i.e.,machine tool builders or system integrators willdesign and construct the manufacturing systemsfor the manufacturers according to the productdesign given by the manufacturer, and then sellthe system to the manufacturer. Rather thanselling systems to different manufacturers, thispaper proposed a selling use approach: themanufacturing system builder can design,construct and manage the system and sell parts tothe manufacturers, significantly reducing thefrequent design, construction and ramp-up of thesystems. Reconfigurable manufacturing systemsis especially suited for such new paradigm.2. The Selling Use approachWhen selling product, an essential element ofprofit is the decline of marginal unit costs due tolarge lot sizes. The resulting resourceconsumption is not of too much concern. Allcosts of purchase, operation, maintenance anddisposal of the product are at the expense of theproduct buyer. f the product is not in use, theproduct buyer as owner has to bear the idlecapacity costs. The manufacturer has a reducedinterest in long products life, as replacement

products increase sales and profit. The buyer of aproduct has become its legal owner and ishimself in charge of it.In the selling use approach, the buyer paysonly for the utilization of the product and not forthe product itself. The costs of investment,operation, maintenance and disposal aremanaged by the utilization seller. The old styleproduct manufacturer and seller hence developsinto a utilization seller and service provider ofcomponents and products. The utilization selleris hence interested in a long-lasting and robustproduct that bears little costs throughout itsusage. Such a product decreases resourceconsumption compared to a product under theselling products paradigm.The disadvantage of selling products comparedto selling use is the tendency to higher costs ofunderutilization and higher resourceconsumption. Selling use gets competitive oncethese costs can be avoided .What are the criteria for a successfulimplementation of a Selling Use model forReconfigurable Manufacturing Systems? Thetwo key parameters that can be altered bychanging the systems architecture over time arecapacity and functionality. In a selling use modelthe Original Equipment Manufacturer's (OEM)interest as the proprietor and configurator of theequipment must be to realize a high utilization ofhis system as his profit is generally related to theunits being produced. Hence, he is more likely toact in highly t1uctuating market environmentsregarding the capacity and functionality requiredfor manufacturing. In this environment the RMScan best compete with conventional DML andFMS in terms of high utilization. Fig . 1 depictsthe two fundamental process chains for theSelling the Product vs. the Selling Useapproach.

Dedicated Manufacturing Lines & Flexible Manufacturing Systemsrs i iti rro2u Ci ........ ::: Operation : r a t i o n ii I ) Desi gn cons true) Sell > > c l e .

. ..: ............................................. ............. ... ~ ~ ~ : t : ~ ~ ~ ~ - ...............Reconfigurable Manulacturing Systems e f f f i e c u < i i ~ ~

OEM anufacturer

ofUsage Phase i

. . .... . . . . . . . . .. . .. .. .. Figure : Selling Use vs. Selling theProduct

2


3/8

The vital drivers for the selling use model arisefrom the option to use manufacturing equipmentmodules more than just once. The OEM as aproprietor bears the costs for the reconfigurationand is therefore tempted to re-use alteredcomponents, e.g. in other RMS.Y ct, the selling use model bears the danger ofopportunistic behavior because it separates therights to the cash flows over the Life cycle fromthe rights over the residual value at the End ofLife Cycle. Separating the rights regarding thecash flow and residual value, the Selling usemodel produces incentives for asset abuse, undermaintenance and opportunistic behavior at thepoint of the Use-Contract renewal. Bilateralagreements such as obligations and options canbe applied to overcome some of these negativeincentives, as shown in Fig . 2. Such instrumentsare not adequate to be used in highly fluctuatingmarket environments where the demand forproduction capacity and functionality changesrapidly. Mistrust comes from a lack ofknowledge about the usage of the product, i.e.the asymmetric distribution of informationbetween the manufacturer and the OEMregarding the usage time, reliability andfunctionality of the product.

Tradi t ional0 Juridical and econom ical( .) obligations and options150 Deposits and Penalty Clauses

Leads to mistrust,; . between partnerss::::cuE:::1ic:

Maintenance charge~ u s e s servicepiracy

Option to Purchaseow incentive in consequenceof rapid model change

t-t Re-use drivenUte Cycle Opportunitiesby maintenance, modernizationand re-u ti lizationLife Cycle Diagnos is

Motivation by higheravailability due to preventivemaintenance

Reuse Reconfigurati onRapid and low pricedadaptation to new Life Cycledue to modularity andreu sability of equipment

Figure 2: Methods to safeguard sellinguse contractsReconfigurable Manufacturing Systems clear theway to implement the selling use model,particularly in rapidly changing marketenvironments. KOREN defines key characteristicsfor reconfigurability which are Modularity,Integrability, Customization, Convertibility andDiagnosability. Modularity, integrability anddiagnosability reduce the reconfiguration timeand effort whereas customization andconvertibility reduce cost.Diagnosability is critical in reducing the ramp-uptime of RMS. Frequently reconfigured systemstherefore depend highly on fast and efficientdiagnosis during ramp-up time. In a selling use

model all five RMS-characteristics have a directinfluence on the cost to manipulate the capacityand functionality of the system, yetdiagnosability deserves a special focus. It is thebasis for the re-use driven selling use model forRMS.3. iagnosability1.1. Application in manufacturing systemsDiagnosis of the system-state is a keycharacteristic that deserves a specialconsideration in the selling use model, evenbeyond the ramp-up process. Whereasmodularity, integrability, customization andconvertibility are characteristics that are crucialin the reconfiguration of the RMS, diagnosabilitystays the predominant characteristic during theoperation of the system. As the proprietor of theRMS the OEM s ability to control the state andavailability of the system becomes a corecompetency. Knowledge about the usage of thesystem and each of its components allows itsreconfiguration when the needs regardingcapacity or functionality change, that is due toincreasing or intensifying workloads that couldpossibly overstrain the system. Equally thisknowledge allows the systems adaptation in caseof altering system qualities due physical changes,i.e. wear.Knowledge about the usage is therefore crucialto guarantee the required functionality and tomake the re-use of components in other systemseconomically favorable. A change in systemqualities can have its origin in the wear of theproduct, such as abrasion, corrosion, fatigue,ageing or staining. The selling use model alignsthe interests of both parties, i.e. the OEM s andthe manufacturer's. High availability is thecommon goal that can only be achieved by theappropriate reconfiguration and adaptation.The risk of production standstill due to systemfailure caused by wear of components is sharedby both parties. Postulating that risk and returnwill always find an equilibrium under freemarket conditions, the central question is whichparty can best provide the diagnosis of thesystem to guarantee the expected availability. Intraditional Dedicated Manufacturing Lines(DML) the competency lies in the hands of themanufacturer. Knowledge is acquired during therelatively long utilization of a hardly changingsystem architecture. This type of organizationseems no longer applicable in the market forRMS. Design, building and installation as well as

3


4/8

ramp-up time and each single Life Cycle of theconfigured system are comparatively short. Thesystem behavior of the RMS changes with everyreconfiguration, making it difficult for themanufacturer to acquire the professionalexperience to conduct intensive diagnosis.

DML FMS RMS' ~ ~ ~ ~ : Hardware Hardware Hardwareoftware Software Software

n - u Manufacturern c - OEM0 0 buys; : I ownsCIG>.. a.015

;

OEM;< sells Manufacturer0- pays per unitLow Medium High

Legend Reconfiguratlon,.Selling-Use Zone" j frequencyOEM Manufacturer

DML De d icated Manufacturing ineFMS - Flexible Manufacturing SystemRMS Roconfigurable Manufacturing System

Figure 3: OEM s diagnosis experienceenables the selling use model for RMSConsidering the labour intensive, i.e. costintensive diagnosis of systems one should use theOEM's experience. This will only be possible ina selling use model, since once the productbecomes the property of the manufacturer theOEM can no longer dispose components in orderto optimize their utilization.Figure 3 depicts the three common groups ofmanufacturing systems and classifies themregarding reconfiguration frequency anddiagnosis experience. Traditional DML are likelyto be serviced by the manufacturer due to thelong production experience with the hardlychanging systems architecture. In comparisonwith DML manufacturers using FlexibleManufacturing Systems (FMS) can hardly copewith the complexity of these systems when itcomes to service such as preventivemaintenance. OEM's have the design experienceregarding these universally applicable systems.Yet, the OEM knows little or nothing about theutilization of the systems as they can beprogrammed to perform an almost infinitevariety of processes. Generally the OEM has noaccess to process data, i.e. that maintenance canonly be realized according to time intervals or incase of machine failure. Thus, diagnosisexperience is only medium on both sides. DMLand FMS are consequently not recommendablefor the selling use model, simply because the

OEM can not customize the manufacturingsystem to its individual application regardingcapacity, functionality and physical condition.Reconfigurable Manufacturing Systems (RMS)leave less time for the manufacturer to becomefamiliar with the system's properties. Theindividual configuration of a RMS for veryspecific processes and for a comparatively veryshort time leaves no time for the acquisition ofprofound system knowledge by themanufacturer, that is the user of the RMS. TheOEM is predestined to become the diagnosisspecialist. As no RMS is ever likely to beidentical to any other RMS, behavior and failurepatterns are unlikely to be derived for theindividual configured system, but rather forsystem components or frequently repeatingdesign patterns. This knowledge can only beacquired by the accompaniment of the RMS overits entire Life Cycle requmng a closeinformational contact with all RMS utilized inthe market.The OEM will have to become the Life CycleManager of the RMS which comprises the

requirement specific (re-)configurationof the product,acquisition of statistical significantusage and maintenance data,interpretation and aggregation of thisdata into adaptation knowledge andconduction of preferablybased respectively timecorrective maintenance.

conditionbased orAs the Life Cycle Manager for the RMS, theOEM requires the right of disposition for theRMS components, to optimize their utilizationduring their entire Life Cycle. This demands aselling use model with the objective to reducethe costs for both parties, i .e. the OEM and themanufacturer. To understand how diagnosabilitybecomes a new core competency of the OEM wefocus on how information about the actualutilization of the RMS can be evaluated.1 2. Deriving diagnosis informationTo configure or reconfigure the RMS parametersare required. KOREN suggests possible productparameters that are wo rkpiece size, partgeometry and complexity, production volumeand production rate, required processes, accuracyrequirements and material property that must betaken into consideration. These product

4


5/8

parameters address one aspect of the utilizationof the manufacturing system. Described in aninfinite high detail they would be an adequatemean to describe the actual utilization of theRMS.Yet, they can not comprise unplannedvariations of their values nor randomenvironmental impacts. Thus, to describe theactual utilization of the manufacturing system,additional usage data is indispensable. Theprovision of this usage data is the basis for thediagnosis that itself is the prerequisite for a costefficient maintenance and profitable re-use ofproduction equipment by the OEM.KOREN proposes that diagnostics should beembedded into RMS on component level.Information from reconfigurable sensors is usedto detect faults or quality problems during rampup time. As ramp-up time has a significantlyhigher impact on RMS than on conventionalDML or FMS, new sensory measurementprinciples are more likely to be introduced inRMS. As sensors are embedded on componentlevel they can be used during the entire LifeCycle of the RMS at low cost. Thus, embeddedsensors should not only be used during ramp up

Re-)Configurationproduct parameters

Work pieceSize

RequiredprocessesAccuracyrequirementsMaterialproperty

RMSconfigurationsEmbedded modularLife Cycle Unit atcomponent level

r i v ~ s

gears

/ filters ,

daptation

time, but also be applied to meet the needs formaintenance and re-use of the system.Life Cycle accompanying diagnosis requires theidentification of use-intensity factors for theelements, i.e. modules of an RMS. Thissystematic identification of use-intensity factorsand adequate sensory measurement principles isthe first step in the Life Cycle Unit concept thatis considered to be an enabler to conductdiagnosis during the Life Cycle of a product [1].The Life Cycle Unit itself is conceived as amodular microsystem to provide usage data byits elements sensors (detection) and Life CycleBoard (storage, processing and transmission) .Sensors and Life Cycle Board (LCB) areembedded in systems on the component level ,i.e. in modules [2] . The LCU concept focuses onthe provision of usage data to be applied in thediagnosis of the system s status. The objective isto support the system s adaptation, i.e.maintenance, re-use of components and itsreconfiguration. Hereby the L U conceptcontributes to make the se lling use modelprofitable for the OEM.

RMSI \

Life Cycle Datautilization parametersProvided system andcomponent usage data

Diagnosisknowledge basedDeriving models fromexperienced malfunctionsof components andsystems

IRe-startsRMS. II i L...1'. Load, stress, torque iWeibull analysisStatistical techniquesExpert SystemsArtificial Neural Neworks

1 i to identifyi Vibration, sound Failure Patternsand to allow

~ ~ ~ ~ e r t u ~ ~ ~ ~ Feedback-learningRMS.nDiagnosisdamage symptom based

Element WearAgeingto restore properties t leari no s Abrasiona ltered by wear cjrive sFatigueReconfiguration oears Stains Model basedto install required toois informationperformance Wters DisruptionCorrosion

Figure : Diagnosis of usage data from different RMS configurations by the OEM

5


6/8

State of the Art methods concentrate highly onthe measurement of a few parameters likestructure-borne-sound (vibrations), moments ortemperature form bearings, gears and drives.Measurements are primarily realized withstandard add-on sensor systems, not withembedded modular microsystems [3] . A broaderapproach will have to identify and classify morerelevant signals and provide this sensory datawith cost-efficient systems. Furthermorestatistical techniques, e.g. Weibull analysis,graph theory or techniques based on expertknowledge or artificial neural networks have tobe applied to transform sensory data into widelyusable utilization knowledge.Figure 4 describes how the (re-)configuration ofa RMS as well as its adaptation is beingsupported by Life Cycle Data. Diagnosis can besymptom based as in the case of vibrationanalysis on a bearing that provides data for thedirect identification of wear, such as a disruptedbearing cage. Knowledge based diagnosisapplies models and several utilization parametersto derive statistical information regarding theprobable state of single components respectivelythe RMS. Symptom based and model baseddiagnosis support adaptation and reconfiguration.They require the implementation of embeddedmodular sensory systems and the Life Cycleaccompanying analysis of acquired data.4 Profitable reuseFigure 5 below summarizes the vital issuesreconfiguration, adaptation, re-use and diagnosisaddressed ab ove in a general analytic relationcomparing the selling the product vs . the selling

Se ll the Product

Profit from sold equipmentbeginning of period)/Profit from reconfiguration

end of period)

Example for a twoyear period wi threconf igurationafter two years.

use approach for a RMS. The example comparesthe net profit and cost groups over a 24 monthperiod and takes their temporal realization intoaccount. The reconfiguration of the RMS at theend of the period is included in the example.The seller of the RMS realizes his profits by thesale of equipment and service at the beginning,throughout and at the end of the period. Theseller of the use of the RMS realizes none ofthese profits from sale and service. Hisremuneration is a monthly payment related to theusage of the RMS. In the example the number ofproduced units was taken as a measure todetermine the calculation basis for the OEM sprofit.Diagnosis was identified as a crucial factorduring the entire Life Cycle of the RMS to allowcondition based maintenance and the re-use ofmodules. The extra costs for the implementationof embedded sensory systems on componentlevel and the Life Cycle accompanying diagnosisof the RMS are being assigned to the OEM .Likewise he is the beneficiary of the residualvalue of re-used components at the end of theperiod. The example limits its complexity to thefactors that are being determined by the qualityand extend of diagnosis of the system during theusage. That means that varied economicalconditions like taxation or operating expensesdue to changing economies of scale or scope arenot being considered. It is being assumed, thatthese operational effects can be measured bymeans of managerial accounting and thus willfind a bilateral equilibrium in a free marketeconomy.

Sell the Use

Profit from unit based monthly paymentsmonthly):::;.1(5'+:: Profit from re -used components~ - . . v - . . r l end of period)

U=Return for sold unitX= Units sold per month :?:-o Lt.;: Cost for diagnosis systemsv r-f beginning of period)rofit from service and sparepartsmidterm) i= Weighted average costof capital per year) i J j ~ Cost for diagnosis and maintenanceT midterm)Figure 5: Sell the RMS vs. sell the use of the RMS for a two year period

6


7/8

The focus lies on the profit from re-usedcomponents C), the initial cost for theaccompanying diagnosis system D) and the costfor diagnosis and maintenance M). Improvedcondition based maintenance during the usageperiod is monetarily accounted for by the profitfrom re-used components. Improveddetermination of the residual value of RMSmodules using Life Cycle Data allows a betterallocation of the modules.Profit from re-used components is being reducedto the residual value and the costs for adaptation.Eq u. l C=R - KC Profit from re-used componentsR Residual valueK: Cos ts for adaptationThe residual value of the components is beingdetermined by physical wear and economicaldevaluation.Equ . 2 R =I A E

nEqu.3 A T w

w= l

Equ.4

Equ.S

A- ~ ) b=e TwtE l

TEI: initial Investment

with A E 0;1)

with Aw E 0;1)withEE 0;1]

A: Total wear indicator, that is the probability ofsurvival for the componentAw Probability of survival regarding thecomponent and a single wear-classw: wear-class ageing, abrasion, fatigue, creep,stains, deformation, disruption, loss or corrosion)b Shape parameterT: Scale parametert: TimeE: Economic devaluationTE Length of economic devaluationThe Weibull distribution Equ. 4) can be used tomodel systems with decreasing failure rate,constant failure rate, or increasing failure rate.This versatility is one reason for the wide use ofthe Weibull distribution in reliability. Theevaluation of shape and scale parameters depend

on the extend and quality of the diagnosability ofthe system. The length of economic devaluationEqu. 5) is being determined by the speed oftechnical development and decreasing marketprices. The determination and utilization of theresidual value requires processes that generatecosts in inspection, disassembly, assembly,refurbishment as well as in the filed of logisticsand opportunity cost.Equ.6 K=K 1 Ko KM KA + KL KoK1: Cost for inspectionK0 : Cost for disassemblyKM: Cost for assemblyKA Cost for refurbishmentKL: Cost for logisticsK0 : Opportunity costCost for inspection, disassembly and assembly,such as labour cost, are highly dependent on thequality and accessibility of diagnosis data. Costfor refurbishment, such as machining hours,spare parts and labour cost cor relate directly withthe decisions made during inspection. Cost forlogistics and the opportunity cost of unusedequipment represent further calculative cost.These basic relations help to clarify thechallenges and define the call for action to makethe Selling use approach profitable forReconfigurable Manufacturing Systems:

Identification of a wider spectrum of wearparameters for components and selection ofappropriate sensors.Realization of cost-efficient embeddedmodular microsystems on component levelin RMS modules as presented with the L ifeCycle Unit concept. Factor D) will decreaseas micro-integration becomes state of the art.Stimulation of knowledge based diagnosisconsidering the prospect of a statisticalsignificant data basis provided bydiagnosable RMS . With such a genericdatabase on failure modes and processingparameters reliability predictions can bemade on other RMS modules and similarsystems. Thus, the factor M) will decreasewith the realization of wide knowledge fortheRMS .Life Cycle Management of used modules tomaximize the profit C) from re-usedcomponents considering the residual valueand costs for adaptation.

7


8/8

[1] Grudzien, W.; Middendorf, A*; Reichl,H.*; Seliger, G., A micro system tool forimproved maintenance, quality assuranceand recycling. In: Proceedings of the CAPE2000 7-9 August, 2000, Edinburgh,Scotland, pp. 391-399.

[2] Seliger, G. Grudzien, W., MUller, K. 1998,The acquiring and handling of devaluation.In: Proceedings of the 5th InternationalSeminar on Life Cycle Engineering,September 16-18, 1998, Stockholm, pp. 99-108.[3] Marschner,U. Elender, G., Fischer, W.-J .

Heinrich, M., Herrmann, T., Hoffmann, T.,Jossa, 1., KrompaB, A., DecentralizedMicrosystem-Based Diagnostics of Bearingsof an Electric Motor. In: European ControlConference ECC 99, Karlsruhe, August 31-September 3, 1999, 154-160.

8

a new paradigm of manufacturing selling use instead of selling systems

Documents