Presentation on theme: "Optimizing High-Mix Low-Volume Operations"— Presentation transcript:
1Optimizing High-Mix Low-Volume Operations Shahrukh A. IraniDepartment of Integrated Systems Engineering The Ohio State University Columbus, OH 43210PHONE: (614)WEBSITE:Lean is widely popular only because the TPS (Toyota Production System) is a revolutionary, successful and globally adopted production system. But, TPS is best suited for assembly line manufacturing because it was developed by an automobile manufacturer. Toyota is a LOW-mix (they produce a limited range of highly-similar products in any one of their facilities) HIGH-volume (they produce thousands of the same car model in any year at any one of their facilities) manufacturer that uses assembly lines to produce its products.In contrast, there are thousands of SME’s like Hoerbiger Corporation of America that are HIGH-mix LOW-volume manufacturers of a wide variety of products that are ordered in LOW/MEDIUM/HIGH volumes.I started studying Lean for HIGH-mix LOW-volume manufacturers around In 2001, I obtained funding from the Department Of Defense to adopt/adapt/enhance TPS (aka Lean) for custom forge shops that are mostly, if not all, HMLV defense suppliers. The result of my research is JobshopLean, a methodology that integrates appropriate Lean tools with Industrial Engineering methods to implement Lean in HMLV facilities.One of the characteristic features of JSL is the reliance on computer-aided analytics because the popular Lean tools:are pencil-and-paper (manual) methods that cannot solve large versions of the problems that they solve --- They can easily be replaced by optimization methods that can be implemented using available software tools or computer programs written by IE/OR professionalscannot solve complex problemsmake us ignore important problems that jobshops struggle with
2Agenda Types of Manufacturing Systems What type of manufacturing system do you manage?What makes jobshops (high-mix, low-volume, MTO) different from any “Toyota-type” plant?Limitations of the (Toyota)Lean toolsExamples of high-mix low-volume (HMLV) manufacturing systemsCASE STUDY: Optimization of the layout of an entire plantCASE STUDY: Optimization of the setup on a pressA Toolbox for High-Mix Low-Volume Manufacturers
3Spectrum of Manufacturing Systems QUANTITY(Volume)Assembly Lines or Transfer LinesFlexible FlowshopsManual CellsHighFlexible Mfg.CellsJobshopsThe two criteria that are typically used to classify manufacturing systems are Volume and Variety (Mix)Notice that the Assembly Line and the Jobshop are at extreme ends of the spectrum?Then why do we persist with implementing Lean in jobshops using many Lean tools that are primarily suited for LOW-mix HIGH-volume assembly lines?LowMIX(Variety)LowHigh
4What is Your Manufacturing System? MIX (R)QUANTITY (Q,T)OutliersPRUNNERSREPEATERSIn a Part FamilySTRANGERSExcept for the OEM’s who have repetitive assembly lines, many manufacturers are actually operating two or more manufacturing systems in the same factory and trying to manage all of them using a single combination of manufacturing strategy, workforce skills, ERP system, facility layout, etc.!How does one identify these different manufacturing systems? I use a method called PQR$T Analysis to segment the product mix of the factory. A typical PQR$T Analysis locates every product in the entire product mix produced by the factory on three axes:$ALES ($): This is how you classify your products as being Cash Cows versus DogsVOLUME & TIME (Q,T): You study both the quantity ordered for each product and the pattern of ordering i.e. Q(t) to determine if it is a Runner, Repeater or Stranger --- You DO NOT manage all three segments the same way!MIX (R): You analyze the routings of all the products to identify how many clear-cut part families exist – One? Many? Outliers? It is important to know the outliers because they do NOT belong in even one of the existing part familiesWhere would products produced in an assembly facility lie and where would products produced in a jobshop lie on this 3-dimensional grid?LOWHIGHPROTOTYPES$ALES ($)
5Lean for Toyota ≠ Lean for Jobshops Jobshops ≠ Assembly FacilitiesHigh mix of products i.e. many different routingsPart families may not be knownProduct mix segmentation must be doneSetup times, cycle times, lot sizes, etc. vary significantlyWide variety of product designs and equipment typesTypical facility has a Process Layout (= batch-and-queue)Different (and Difficult) Business EnvironmentDemand is unstableLot sizes changeCustomer loyalty and sanity are non-existentProduction schedules are driven by due dates (not Takt Time)Shifting capacity constraintsDue dates are different and subject to frequent changesAn assembly facility and a jobshop are radically different manufacturing systems that SHOULD NOT be designed and operated using a generic set of Lean tools just because they have been passed down to us for free from Toyota
6Many Lean Tools are NOT Universal Strategic PlanningTop-Down LeadershipMotivated Workforce5STotal Productive MaintenanceSetup ReductionError-ProofingQuality at SourceVisual WorkplaceRight-sized EquipmentStandardization of WorkX Right-sized (= Inflexible) MachinesX Kaizen Events (Mainly by Operators)X 20th (not 21st) Century ManagersX Pencil-and-Paper Problem SolvingX Value Stream MappingX One-Piece Flow CellsX Product-specific KanbansX FIFO Sequencing of OrdersX Pacemaker SchedulingX Inventory SupermarketsX Scheduling using Takt TimeX Heijunka/Load LevelingX Assembly Line BalancingUSE the tools in the left hand column but DO NOT USE the tools in the right hand column. There are enough existing tools that were NOT developed at/by Toyota that could easily replace, if not are superior to, the standard Lean tools. So why don’t we learn them and use them to improve jobshops?
7How to Recognize a High-Mix Low-Volume Manufacturing System? A Spaghetti Diagram maps the routings of all the products that you make on the existing facility layout. It will tell you whether or not you are a HIGH-mix LOW-volume manufacturer. But it is not a trivial task to produce a Spaghetti Diagram for the product mix of any jobshop that could consist of hundreds of unique manufacturing routings!Let us now look at some examples of this simple but effective method from past projects that we did.
8Forge Shop (≈500 Routings) What does this tell you about the product mix and the facility layout?
9MTO Industrial Scale Fabrication Facility Here we mapped the complete Bill Of Routings for a complex fabricated product. What does this map tell you about the existing facility layout?
10Flexible Machining Cell This is a multi-product machining cell. Does it have the traditional U-shaped layout? Notice that there is no single linear material flow pathway?
11Finish Grinding Department HMLV manufacturers have to deal with high flow complexity instead of ignoring it. Using optimization, they can quit relying on the manual Lean tools that are incapable of visualizing this complexity and fail to suggest strategies to reduce or eliminate this complexity.
12Factory Layout Optimization at Ulven Forging This case study utilized optimization methods to study and improve an entire factory!
13Current StateThis custom forge shop produced 450+ different forgings for commercial and defense customers.
14Future StateWe could not completely convert the facility into a set of independent manufacturing cells because there were many monuments i.e. equipment that could not be relocated; so we did the next best thing and implemented partial cells around the monuments.
15Actions TakenAn additional processing area was created in the Drop Hammer building where cleaning,finishing,packaging and shipping were consolidated.The 158 ton Trim Press was replaced by a 440 ton press that was positioned next to the 5000# Hammer. This eliminated the transportation of large forgings to a distant 350 ton Trim press. Also, a 350 kW induction heater and conveyor were purchased and co-located with this press to form an Upset Forging cell.A new 2.5” Upsetter was purchased and positioned next to the 3000# Hammer to form an Upset Forging cell.By implementing CELLS based on the concept of part families, Ulven Forging agreed that only a major layout change would allow them to reduce travel distances and promote teamwork.
16Actions Taken (contd.)The 1.5” Upsetter was replaced with a faster machine and positioned next to the 700 ton Press to form an Upset Forging cell.A crane was installed over the 5000# Hammer to reduce piston change-out time, reduce die key tightening time and to facilitate product movement in the area.A portable Marvel Hacksaw and 1.5” Bar Shear were acquired.A CNC Mill was acquired and positioned next to the EDM machine to reduce vendor costs and lead times for die sinking.(CONTINUED FROM PREVIOUS SLIDE) By implementing CELLS based on the concept of part families, Ulven Forging agreed that only a major layout change would allow them to reduce travel distances and promote teamwork.
17Benefits Approximately 5% cost savings on annual sales of $ 6 million WIP reductions were significantLead times quoted to customers were reducedThroughput ($ales) increasedSince this company is a defense supplier, they were reluctant to release financial and delivery performance data that could be seen by their customers (Defense Logistics Agency and Department Of Defense)
18Press Setup Optimization at Hirschvogel In the previous case study, we looked to optimize the entire factory; in this case study, we utilized optimization models to study and improve workflows at a single forging press in an effort to reduce its setup time
19Observe and Document the Press Setup SMG 14 is one of the “workhorses” at Hirschvogel’s facility in Columbus, OH
20Work Locations around the Press This is a list of all the locations around the press that the video showed the operator visiting and working at throughout the setup process.
21Sequence of Tasks in Press Setup This spreadsheet captures the entire sequence of 225 steps that constitute the setup of the press. Each step is performed at one of the locations around the press that are listed in the previous slide.
22Operator Motion Traffic around the Press PFAST, one of the optimization packages that we use at OSU, converted the 225-step setup process into this aggregated traffic matrix between all pairs of locations around the press.
23New Workstation Layout for the Press STORM, another one of the optimization packages that we use at OSU, took the From-To Chart produced by the other package and generated this conceptual solution for how the work locations around the press ought to be re-arranged.
24Operator Motion in Previous Layout This is how the operator moved during the setup process BEFORE we did our optimization study.
25Operator Motion in New Layout This is how the operator moved during the setup process AFTER we did our optimization study.
26Press Setup Time: Before vs. After ActivitiesCurrent ProcessRedesigned ProcessDifferenceNo.Time%Operations2471:32:0771%2351:28:5379%-12-0:03:14Inspections320:11:289%310:11:2310%-1-0:00:05Transport.1350:17:5414%920:12:4511%-43-0:05:09Storage0:00:000%Delays120:09:017%-0:09:01Total4262:10:30100%3581:53:01-68-0:17:29IN ADDITION, we did a classical Time and Motion Analysis of the entire setup process which helped us to reduce the time per setup done on this forging press by 17 minutes.So how did this time saved per setup benefit Hirschvogel? Simple! Use the capacity that was freed up to produce more forgings and sell them to make more money!How did we do this? Capacity that was already available on the press had now become available. So Hirschvogel could afford to do more setups even if they had to lose some of that freed up capacity to do the extra setups. But there was sufficient leftover capacity that could be utilized to make more forgings!
27Estimation of Increased Production Potential Increase in Production = [EPT – (# of Setups * Time/Setup)] / Cycle Time per PartRelationship between Number of Setups and Number of Parts Produced (of same type) that could be produced using the Extra Production Time (EPT)
28Estimation of Increased $ales Potential Increase in Revenue = [EPT – (# of Setups * Time/Setup) / Cycle Time per Part] * $/PartRelationship between Number of Setups and Revenue from Extra Production that could be earned (based on various part prices) using the Extra Production Time (EPT)
29Benefits Savings in Setup Time Potential Increase in Revenue Category InitialProposed% DecreaseSetup Time (min)130.560.8653.3%Number of Activity Steps42635816.0%Distance travelled by Operator (steps)116779731.70%Potential Increase in RevenuePart #DemandPrice/ PartRevenueA12346,000$$ ,300.00B34563,300$$ ,100.00D67557,500$$ ,750.00Extra Parts16800Extra Revenue$ 210,150.00The setup reduction project freed up enough available capacity whereby more setups could be done to process additional/extra orders because, despite losing some of the freed up capacity in setups, there was still time left to make forgings to complete the additional orders
30Parallel Task Scheduling for Press Setup 1 Operator min (Current State)2 Operators min (Parallel Execution of Activities)3 Operators min (Unnecessary, 2 operators are OK)Gantt Chart for Optimized 2-Person Setup ProcessOp #1Op #2We even utilized a project scheduling algorithm that told us that we could convert a sequential one-person setup process into a parallel two-person process which could be completed in HALF the time compared to the original setup
31Optimization could Enhance Every Lean Tool Why do I use use optimization to enhance and extend the Lean tools? Because the manual pencil-and-paper Lean tools have handicapped and limited HMLV manufacturers from achieving significant gains that could be earned by solving complex Continuous Improvement problems using optimization-aided methods!
32Multi-Period Slotting of Orders Value Stream MappingDemand ForecastingMulti-Period Slotting of OrdersWork Order ReleaseCurrent Project at Pompano BeachWarehouse DesignIn this slide, using a Value Stream Map obtained from the Learning To See book, I have shown all the areas of any manufacturing system for which optimization models have been successfully developed to design, analyze and improve each of those areas.Scheduling Supplier DeliveriesCell DesignWork Center SchedulingWIPInventory ControlScheduling Material Handlers
33A Toolbox for HMLV Manufacturers “4H” LEADERSHIPWalks the GembaKnowledgeableBeyond TLSSCompetitiveRespects EmployeesEfficient FireFighters (RiskManagers)Change ManagersInvested in Talentand TechnologyENGAGED WORKFORCELean Thinkers“Factory ofOne” ProblemSolversEmpoweredMulti-skilledTeam-orientedIT SavvyCollaborativeFUTURE STATECURRENT STATEMORE….COMPETING THROUGH INNOVATIONValue NetworkMappingM3 FacilitiesFlexible Focus(GT)Product MixSegmentationERP+FCS+MESIT-enabled JITCommunicationsVirtual Cells andDistributed TeamsWater StridersReal-time OrderTrackingPLM+CAPPAgile SuppliersUNIVERSITY PARTNERSHIPSCo-CurricularProjectsCo-opsInternshipsFaculty R&DExecutives onLoanFactories as“Test Beds”INDUSTRIAL ENGINEERINGWork MeasurementsFacility LayoutErgonomicsSchedulingVariety ControlValue Analysis/ValueEngineeringDFMAFlexible Mfg. CellsProduct/ProcessStandardizationProduct MixRationalizationetc.LEARNING ENTERPRISECafeteria ChatsE-newsletterWikisKaizensOnline ChatGroupsIdea BoardsAnnualConferenceResource Center- E-books- Case Studies- Videos- Online Apps- etc.Why the conveyor representation for a manufacturing enterprise? Because every wheel needs to be working in order for the conveyor to work and carry items forward from one end to the other. Any HMLV manufacturer like Hoerbiger Corporation of America needs to be doing many things correctly and well in order to go from good to great!
34AcknowledgementsThe PRO-FAST Program is enabled by the dedicated team of professionals representing the Defense Logistics Agency, Department of Defense and industry. These team mates are determined to ensure the Nation’s forging industry is positioned for the challenges of the 21st Century. Key team members include: R&D Enterprise Team (DLA J339), Logistics Research and Development Branch (DLA – DSCP), and the Forging Industry Association (FIA).
35Acknowledgements Project Champion: Craig Kaminski Project Engineer: Haydn GarrettProject Champion: Kevin ShawProject Engineer: Greg MuniakProject Champion: John WilburProject Engineers: Thomas SlautaProject Champion: Dick JohnstonProject Engineer: Todd SheppardJon TirpakRussell BeardVicky McKenzieProject Champion: Joe KracheckProject Engineer: John LucasProject Champion: Andrew UlvenProject Engineer: Jim HuirasDan GearingProject Champion: Thomas StysProject Engineer: Jorge AlvarezPFAST Development TeamDr. Rajiv RamnathDr. Rajiv Shivpuri