AchievingSupplyAvailability intheFaceofHighlyUncertainDemand
ACaseStudyofSparePartsforMilitaryAviation
WilliamKillingsworth ExecutiveDirector,MITForumforSupplyChainInnovation 6November2008
1
Itstoughtomakepredictions, especiallyaboutthefuture.
YogiBerra
2
ThereAreSeriousChallenges inDemandForecasting forMilitaryAviationSpareParts
VeryLongProductionLeadTimesand AssociatedForecastHorizons UncertainOperatingEnvironments UncertainOperatingLevels UncertainImpactsofPoliticalProcessand PublicOpinion AdequacyofFundingforPurchases
ACaseStudy
ATypicalChallenge inProvidingSpareParts
AHelicopterBladeCostsRoughly$175,000; RecentDemandHasBeenRunningatRoughly22/Month; ProductionLeadTimeisRoughlyTwoYears; HowManyNeedtobeOrderedNowtoMeetLikelyDemand inMayof2010?(DontForgettoFactorintheForecastthe ElectionResultsofNovember2008!) IsThereAnythingThatCanBeDonetoDevelopAdaptability andProvideRiskMitigation?Whatmighthavebeendone in2000? 5
StepsintheAnalysis
IdentifyCurrentRoleofDemandForecastin SupplyPlanning;AreThereProblemsinthe Process? UnderstandStructureandDetailsofthe SupplyChain; DevelopStrategyforRiskMitigationand MinimizingImpactsofForecastError; AssessImpacts,CostsandBenefits
6
TheRoleofForecastDemandintheSupplyProcess forAviationSpares
Assets Applicable to Repair Review
Repair Action Point
Repair Actions
Forecast Demands Historical Demands
+
Demands
Total Available Assets
Procurement Reorder Point
Procurement Buys
Total Net Assets
DataElementsUsedintheSupplyProcess
Assets Applicable to Repair Review Repair Action Point
Below Depot Requirements War Reserves Repair Lead Time Quantity Repair Safety Level Quantity One Month of Gross Demands Serviceable Stock on Hand Dues In from Repair Action Dues In from Procurement (RLT+1 Months) - Dues Out
Repair Actions
Forecast Demands
Historical Demands
Demands
Total Available Assets
On Hand Dues in Dues out - Serviceable Inv - Recoverable Unserv Inventory
Procurement Reorder Point
Below Depot Requirements War Reserves Safety Level Requirements ALT Requirements PLT Requirements Repair Cycle Requirement Procurement Cycle Requirements
Procurement Buys
Total Net Assets
Serviceable Stock on Hand Dues In From Repair Dues In From Procurement Recoverable Unserviceable Stock on Hand - Dues Out
Observations HowDoYouSpellTrouble?
Theorderingprocessisdrivenbyforecastdemand, typicallycalculatedasarollingtwentyfourmonth average; Recommendedbuysandoverhauleffortsareeach calculatedasthedifferenceoftwolargenumbers; Theorderingprocessisextremelysensitiveto forecasterrorandcommondataerrorssuchasthe productionleadtime; Shortagesandbackorders,andinsomecases, excessinventories,arenotuncommon.
9
StructureofSupplyChain forAviationSpareParts
PROGRAM DEMANDS Component One Supply Chain Component 1 Tier 3 Tier 2 Tier 1 Orders Component Two Supply Chain Component 2 Tier 3 Tier 2 Tier 1 Orders Component Three Supply Chain Component 3 Tier 3 Tier 2 Tier 1 Orders Component Four Supply Chain Component 4 Tier 3 Tier 2 Tier 1 Orders Component Five Supply Chain Component 5 Tier 3 Tier 2 Tier 1 Orders Component Six Supply Chain Component 6 Tier 3 Tier 2 Tier 1 Orders Component Seven Supply Chain Component 7 Tier 3 Tier 2 Tier 1 Orders Component Eight Supply Chain Component 8 Tier 3 Tier 2 Tier 1 Orders Wholesale Maintenance Supply Programs Orders for Components C O M P O N E N T P A R T S Orders for Components Orders O F Components New Spares A L L O C A T I O N DUE IN SUPPLY CHAIN CONTROL CENTER ON HAND RECURRING DEMANDS CONUS Aircraft A L L O C A T I O N O F F I N A L P A R T S Inventory Inventory
Orders
SWA Inventory Aircraft
Europe Aircraft
Components Commercial Overhaul Orders for Components Orders
K, J, & H Inventory Aircraft
Components Depot Overhaul
Return to Overhaul
10
Supply Chain Map for Blade Assembly (1 of 2)
Anodize
7(M) 490 (M) 100(M) 165(M) LowerTip 93(M) 7(A)
NoseCap
56(PA) Prime 180(M) Final Assembly 56(PA) TubeAssembly
56(PA) 56(PA)
Weights 87(M) 100(M)
HeatTreat& Mag Part
14(M) 200(M)
Notes: (A) Administrative Lead Times provided by supplier (calendar days) (M) Manufacturing Lead Times provided by supplier (calendar days) (PA) Prime Administrative Lead Times including dock to stock time (calendar days)
130(M) 10(A)
11
Supply Chain Map for Blade Assembly (2 of 2)
14 (M) 14(M) 98(M) 7(A) 107(M) 5(A) Bracket 56(PA) 42(M) TieDown
Processing
81(M) 7(A)
Endcap 56(PA)
56(PA) 28(M) 56(PA) Strip
Prime 180(M) Final Assembly
28(M)
112(M) 7(A)
Notes: (A) Administrative Lead Times provided by supplier (calendar days) (M) Manufacturing Lead Times provided by supplier (calendar days) (PA) Prime Administrative Lead Times including dock to stock time (calendar days)
12
LogicTools InventoryAnalyst
CreatingAPushPullBoundary
13
KeyAssumptions
Thereareninecriticalitemsinthebladesupplychain(NoseCap, TwoTubeAssemblies,TipLower,EndCap,Strip,TieDown, Weight,Bracket) AllotheritemsarecategorizedasRemainingItems Thecostofthebladeis$175,790
TheOEMpays50%ofthatcostforthepartsneededtoassemblea blade,brokendownasfollows:
TheNoseCapandEndCapeachcomprise15%oftheOEMss costforthe entireblade ThetwoTubeAssembliesandtheTipLowereachcomprise10%oftheOEMs costfortheentireblade TheStrip,TieDown,Weight,andBracketeachcomprise5%oftheOEMs costfortheentireblade Theremaining20%isthecostoftheremainingitems
Firsttiersupplierspay50%oftheOEMscostforthepartsneededto assembletheirproducts
14
StructureofIAModel
CommittedServiceTime
OEM New Spare
Govt
OEM Overhaul
CommittedServiceTime:750days
NoseCap: 0 16.2 0.127 0.127
Weights:
1.27
TubeAssy:
0.254
0.127
0.127
CommittedServiceTime:240days
NoseCap: 0 53.7 0.127 0.127
Weights:
197
142
TubeAssy:
28.2
20.7
0.127
CommittedServiceTime:30days
NoseCap: 0 53.7 0.127 23.1
Weights:
197
214
TubeAssy:
28.2
0.127
25.6
MovetoNewPerformanceCurve ThroughOptimization
Readiness or Supply Availability
OptimizationApproach: MovetoaNewCurve
CurrentApproach: Increase$toIncreaseReadiness
$Investment
19
[Link]
Demand Forecast Error = 6
Working Capital (Million) $50 $40 $30 $20 $10 $0 0 200 400 600 800 1000 1200 Maximum CST (Days)
Nostocking Stocking Stocking,OEMPLT=120days
[Link] [Link]
Demand Forecast Error = 6
$10.00 Change (Million) $8.00 $6.00 $4.00 $2.00 $0.00 ($2.00) 0 200 400 600 800 1000 1200
Maximum CST (Days)
NoStocking Stocking
Demand Forecast Error = 6
$20.00 Change (Million) $15.00 $10.00 $5.00 $0.00 ($5.00) 0 200 400 600 800 1000 1200
Maximum CST (Days)
NoChangeinPLT(180days)
PLTReducedto120days
InventoryAnalystConclusions
Increasingsafetystocklevelsinthe manufacturingsupplychaincanbothreduce leadtimesandreducetheamountofworking capitalinvestedtoachievedesiredservicetimes; Increasingsafetystocklevelsevenfortheone componentwiththegreatestleadtime producesnoteworthyresults; Increasingsafetystocklevelsfurthermore reducestheriskofshortagesandlongerlead timesintheeventofanunexpectedincreasein demand,aproblemthathasexistedforaviation spares.
23
StrategyEvaluation UsingDynamicModeling
24
StructureofSupplyChain forAviationSpareParts
PROGRAM DEMANDS Component One Supply Chain Component 1 Tier 3 Tier 2 Tier 1 Orders Component Two Supply Chain Component 2 Tier 3 Tier 2 Tier 1 Orders Component Three Supply Chain Component 3 Tier 3 Tier 2 Tier 1 Orders Component Four Supply Chain Component 4 Tier 3 Tier 2 Tier 1 Orders Component Five Supply Chain Component 5 Tier 3 Tier 2 Tier 1 Orders Component Six Supply Chain Component 6 Tier 3 Tier 2 Tier 1 Orders Component Seven Supply Chain Component 7 Tier 3 Tier 2 Tier 1 Orders Component Eight Supply Chain Component 8 Tier 3 Tier 2 Tier 1 Orders Wholesale Maintenance Supply Programs Orders for Components C O M P O N E N T P A R T S Orders for Components Orders O F Components New Spares A L L O C A T I O N DUE IN SUPPLY CHAIN CONTROL CENTER ON HAND RECURRING DEMANDS CONUS Aircraft A L L O C A T I O N O F F I N A L P A R T S Inventory Inventory
Orders
SWA Inventory Aircraft
Europe Aircraft
Components Commercial Overhaul Orders for Components Orders
K, J, & H Inventory Aircraft
Components Depot Overhaul
Return to Overhaul
25
SharpIncreaseinDemand2003
26
BaseCase:BladeInventories
27
AlternativeCaseA: InventoriesWithStockingPolicy
28
AlternativeCaseB:InventoriesWith StockingPolicy&OEMPLTReduction
29
Summary&Conclusions
ForecastsAreAlwaysWrong; TheLongertheForecastHorizon,theWorse theForecast; HoldingInventoryofFinalGoodsisaVery ExpensiveWayofDealingwithUncertainty; PushPullBoundariesEnhanceAbilitiesto beAdaptiveandResponsiveandEfficiently MitigateRisksofForecastErrors
30
QuestionsorComments?
BillKillingsworth billk@[Link]
31
