Manufacturing Systems Research Chris Hicks © C.Hicks, University of Newcastle Computer Aided Production Management Systems in Engineer to Order Companies • ACME grant in collaboration with NEI Objectives • Identify the characteristics of companies in ETO/MTO sector • Evaluate the status of CAPM • Identify common CAPM problems • Develop methods for modelling CAPM systems in ETO/MTO environments • Apply modelling methods to identify solutions to CAPM problems © C.Hicks, University of Newcastle Identification of Company Characteristics and CAPM problems • Nine one day visits • Three long visits (3 days - 1 month) • Developed semi-structured “audit” methodology • Developed methods for company classification © C.Hicks, University of Newcastle Audit • • • • • Markets Products Processes Manufacturing Systems CAPM Systems © C.Hicks, University of Newcastle Company Classification Jobbing Company Type “A” Main product Batch Spares Subcontract Flow Deep Shallow Product Structure © C.Hicks, University of Newcastle Company Type “B” Jobbing Batch Valves & Pumps Motors Cabs Flow Deep Shallow Product Structure © C.Hicks, University of Newcastle Control Approaches Project Management Jobbing MRP MRP + JIT Batch JIT Flow Deep Shallow Product Structure © C.Hicks, University of Newcastle Audit Conclusions • Markets - demand highly variable and lumpy • Products - complex, highly customised, mix of products • Processes - wide range yet all areas tend to be controlled in same way • Manufacturing systems - mainly functional layouts, high capital employed • CAPM systems - poor integration, wide variety of subsystems, incorrect data structures, poor operational procedures, generally unsuccessful © C.Hicks, University of Newcastle Systems Modelling • Functional models decompose complex systems using a hierarchical top-down approach. They provide a means of understanding activities and interrelationships • Information models enable structure of information to be described • Dynamic models show changing behaviour over time. © C.Hicks, University of Newcastle Company X - Context Diagram a Customer a Customer ITT Contract Awarded Progress Report Tender Company X ITT Quote b Supplier © C.Hicks, University of Newcastle Order b Supplier Data Flow Diagram - High Level 4 Gen. Manager D 1 Approval Of Tender Supplier Tender Quote Quote ITT (copy) 1 Designs, TPS, PP Recommend Suppliers Design for Tender Tendering Designs M 5 Pref. Suppliers M 6 Historic Designs ITT (copy) CQAR 5 Projects Progress Report Project Plan Update Plan & Coordinate Project Quality Project Report 9 Approve P.O. Contract Progress File (copy) Report Drawings, Manuals 7 Engineering 6 Conceptual & Detailed Design New Designs M 5 M 6 Supplier Approval Supplier Pref. Suppliers Historic Designs Purchasing Supplier Selection, Ordering & Expediting M 3 Contract File D/M 4 Client Corresp BRR Action Drawings Progress Report Customer G.M Prepare CQAR Contract File (copy) Contract Awarded Tender ITT Contract File Supplier Details Suppliers & Costing ITT (copy) Supplier 3 D 2 Prepare Tender Engineering 2 ITT & Tender Progress Report Drawings ITP Contract File (copy) Purchase Order (copy) Supplier Approval 8 Quality ITP & Supplier Approval Supplier M/D Suppliers 7 Supplier Purchase Expedite Quote Order D Previous Suppliers Designs 8 Supplier © C.Hicks, University of Newcastle Inspection report Data Flow Diagram - Low Level : Supplier Selection, Ordering & Expediting 7 . Supplier Selection, Ordering & Expediting Projects Drawings 7.1 Purchasing Generate Bill of Materials Components (grouped) 7.2 Contract File (Copy) Supplier Details Possible Supplier Previous Suppliers M/D 7 Supplier Literature New Supplier Purchasing Select Supplier(s) D 8 Approve Supplier Engineering Quote Supplier Selected 7.3 Purchasing Order Component Agreed Delivery Date, Location 7.4 Purchasing Purchase Order Project Plan (Delivery Date) Purchase Order (Copy) Approve Order Supplier Projects Project Progress Check Expedite Progress © C.Hicks, University of Newcastle Supplier Method Limitations • Audit, systems analysis and data modelling provide static “snapshot” views. Longitudinal studies are a series of snapshots. No model of system dynamics. • At best enable “best practice” or potential solutions to be described and documented. • Not possible to perform experiments to examine alternative configurations and evaluate them in terms of specified performance criteria © C.Hicks, University of Newcastle Simulation • Allows modelling of system dynamics • Very expensive in terms of model building and computational resources • Validation often a problem • Predominantly used for either small scale models or rough-cut high level models © C.Hicks, University of Newcastle Simulation Model CAPM Modules PP MPS FIN MRP CP BOMP P/A P INV POC SFC CAPM Modules Existing Developing SIMULATION © C.Hicks, University of Newcastle STATIC DATA OPERATIONAL DATA RESOURCE DATA RESOURCE DATA Resource code Name Optional resource Location (x,y) Rectangular size (w,h) Coordinates Company reference Shift pattern Audit period Data update period Availability Dispatching rule Optional resource rule Batch splitting rule Batch sizes Transfer device Minimum set up time Minimum processing time Minimum transfer time Stochastic distribution Efficiency Overhead cost per hour Cost per hour PRODUCT DATA PART CODE Part code Name Component codes Component quantities PRODUCT STRUCTURE Product structure codes Product instance codes MINOR RESOURCE DATA Minor resource code Location Quantity Cost per hour Minimum use time SHIFT DATA CONDITIONAL DATA RESOURCE DATA Resource status Work in progress (non added value) Work in progress (added value) Queue length Maximum queue length Operations per family Sum set up time Sum processing time Last part Batch counter PRODUCT DATA PRODUCT STRUCTURE Current operation Current operation type Current value Actual operation start times Queue times before each operation Actual set up times Actual processing times Queue times after each operation Actual transfer time Shift number Daily work pattern PRODUCT DATA PART CODE Routing Operation types Minor resources used Minor resource quantities When minor resources used Planned set up times Planned operation times Planned transfer times Lead time Lot size Lot sizing rule Safety stock Safety lead time Planned product structure PRODUCT STRUCTURE Product family Original due time Planned due time Planned operation start times MASTER PRODUCTION SCHEDULE MPS Item Number Part code Quantity Static data - time invarient Operational data - specified configuration Conditional data - dynamic stochastic data DEFAULT DATA Refers to Minimum set up time Minimum processing time Minimum transfer time Minimum queue before operation Minimum queue after operation © C.Hicks, University of Newcastle Key Features • Large scale model allows whole manufacturing facilities to be represented • Models facilities, products, processes and planning and control systems • Many product families can be represented with shallow, medium or deep product structure • Data structures match ETO/MTO requirements • Allows variety of planning and control methods to meet local requirements • May be used as a research tool or for planning and simulation. © C.Hicks, University of Newcastle Simulation Case Study • Heavy Machine Shop used for case study • Static configuration used layout and other resource constraint information • BOM information obtained for products • Process data and planning data obtained for 18 month period. © C.Hicks, University of Newcastle Experimental Design • Series of full factorial experiments Factors • Process - assembly lead time, minimum set-up, machining and transfer • CAPM, scheduling methods, BOMs, dispatching rules, capacity planning • Product mix / load • Operational - data update period © C.Hicks, University of Newcastle Conclusions • Assembly planning and capacity planning important CAPM subsystems • “Dispatching” rules not very significant • Manufacturing performance sensitive to transfer times. • Significant advantages gained through having close control of key resources • Real time data recording led to improved manufacturing performance © C.Hicks, University of Newcastle Consequent research topics • Manufacturing Layout (Hinrichs, Wall) • Capacity planning (Tay, Holmes, Hines, Pongcharoen) • Assembly planning (Sullivan ..) • Planning of product development activities • Planning under uncertainty (Wall, Brand, Song) • Integration of project planning methods with MRP type approaches (EPSRC proposal) • Plan Optimisation through Genetic Algorithms © C.Hicks, University of Newcastle Manufacturing Layout • All data required for layout analysis, clustering and generation in simulation data structures • Work started with Chris Lee who was interested in improving the layout of Vickers’ Scotswood Road factory • Much work in layout has focused on moving from functional layout to cellular layouts © C.Hicks, University of Newcastle Methods • Clustering – Matrix based methods – Similarity coefficient methods • Layout generation – Starting with some candidate solution generate new layout that minimises (maximises) some objective function – Simulated annealing – Genetic algorithms © C.Hicks, University of Newcastle Capacity Planning • • • • • Finite / infinite loading Re-planning rules (Tay) Finite loading rules (Holmes) Interactive tools (Hines, Poncharoen) Schedule “optimisation” (Poncharoen) © C.Hicks, University of Newcastle Supply Chain Management • ETO companies moving towards buy rather than make • Business process analysis approach (McGovern, Earl, Harrison, Hamilton) • Agent based modelling of supply chains (Harvey, McLeay, Hines) © C.Hicks, University of Newcastle IT Implementation • Embodies audit, business process analysis and requirements definition • Transfer of computing expertise • 3 Teaching Company Schemes © C.Hicks, University of Newcastle Layout design & effect on benchmarks • Tony Wells Siemens Semiconductors • Data from North Tyneside, US, Germany, Taiwan. • Pareto analysis of costs • Identification of cost drivers • Relating cost drivers to plant design configurations • Results so far: potential cost reduction of 50% on £30m/annum -pity the plant has closed! © C.Hicks, University of Newcastle Management of Knowledge • Data modelling / systems analysis based upon thematic knowledge that is formal, explicit and easily shared • Knowledge management requires knowledge of product and process structure • “Tacit” or embedded knowledge that is disorganised, informal, context dependent and relatively inaccessible often important. • Interested in developing methodologies for systems integration that include both thematic and tacit knowledge • LABS Proposal with Paul Braiden © C.Hicks, University of Newcastle Other Activities • £180k EDF grant with ISRU to develop distance learning material (Dave Stewardson & Mark Gary) • TCS with House of Hardy aimed at improving manufacturing efficiency (Rob Davidson & Paul Braiden) © C.Hicks, University of Newcastle Summary • Wide portfolio of manufacturing systems research • Various types of research undertaken from the theoretical through to applied. • “Market led” rather than “product led” © C.Hicks, University of Newcastle