Title: MODELING AND ANALYSIS OF MANUFACTURING SYSTEMS Session 8 CELLULAR MANUFACTURING GROUP TECHNOLOGY
1MODELING AND ANALYSIS OFMANUFACTURING SYSTEMS
Session 8 CELLULAR MANUFACTURING GROUP
TECHNOLOGY
- E. Gutierrez-MiraveteSpring 2001
2ORIGINS
- FLANDERS PRODUCT ORIENTED DEPARTMENTS FOR
STANDARIZED PRODUCTS WITH MINIMAL TRANSPORTATION
(1925) - SOKOLOVSKI/MITROFANOV PARTS WITH SIMILAR
FEATURES MANUFACTURED TOGETHER - BURBIDGES SISTEMATIC PLANNING
3BASIC PRINCIPLE
- SIMILAR THINGS SHOULD BE DONE SIMILARLY
- THINGS
- PRODUCT DESIGN
- PROCESS PLANNING
- FABRICATION ASSEMBLY
- PRODUCTION CONTROL
- ADMINISTRATIVE FUNCTIONS
4TENETS OF GROUP TECHNOLOGY
- DIVIDE THE MANUFACTURING FACILITY INTO SMALL
GROUPS OR CELLS OF MACHINES (1-5) - THIS IS CALLED CELLULAR MANUFACTURING
5A Typical Cell
- Machining Center
- On-machine Inspection Monitoring Devices
- Tool and Part Storage
- Part Handling Robot Control Hardware
6COMMENTS
- CONFIGURING MACHINES INTO COHESIVE GROUPS IS AN
ALTERNATIVE TO PROCESS LAYOUT - GROUP CONFIGURATION IS MOST APPROPRIATE FOR
MEDIUM VARIETY, MEDIUM VOLUME ENVIRONMENTS
(Fig.1.6, p. 11)
7COMMENTS
- GROUP TECHNOLOGY AIMS TOWARDS A PRODUCT-TYPE
LAYOUT WITHIN EACH GROUP - RESULTANT GROUPS DEDICATED EACH TO A FAMILY OF
PARTS - NEW PARTS ARE DESIGNED TO BE COMPATIBLE WITH
EXISTING FAMILIES
8COMMENTS
- EXPERIENCE ACCUMULATES AND STANDARD PROCESS PLANS
AND TOOLING ARE DEVELOPED - SHORT-CYCLE, JUST-IN-TIME PRODUCTION BECOMES
POSSIBLE - SINCE NEW PARTS AND EXISTING PARTS ARE SIMILAR,
PRODUCTION IS ACCELERATED
9A GT approach to design
- COMPOSITE PART FAMILIES
- Fig. 6.1 , p. 165
10FACILITY LAYOUT
- EACH PART TYPE FLOWS ONLY THROUGH ITS SPECIFIC
GROUP AREA - WORKERS MAY BE CROSS-TRAINED ON ALL MACHINES IN
GROUP AND FOLLOW PARTS FROM START TO FINISH - MACHINE SCHEDULING IS SIMPLIFIED
- See Fig. 6.2, p. 166
11FACILITY LAYOUT TYPESFig 6.3 p. 167
- GT FLOW LINE
- ALL PARTS ASSIGNED TO A GROUP FOLLOW SAME
MACHINE SEQUENCE - GT CELL
- PARTS CAN MOVE FROM MACHINE TO MACHINE
- GT CENTER
- LOGICAL ARRANGEMENT
12BENEFITS OF GT
- EASE OF DESIGN RETRIEVAL
- DESIGN STANDARIZATION
- SETUP TIME REDUCTION
- REDUCED THROUGHPUT TIME
- INCREASING QUALITY
- REDUCED LABOR COSTS
- INCREASED JOB SATISFACTION
13Generic Benefits of GT
- SIMPLIFICATION
- STANDARIZATION
- See Table 6.1 p. 168
- See also queuing model of GT system with set-up
time reduction on p. 168
14STEPS IN GT PLANNING
- CODING
- SPECIFICATION OF KNOWLEDGE CONCERNING
SIMILARITIES BETWEEN PARTS - CLASSIFICATION
- USE OF CODES TO ASSIGN PARTS TO FAMILIES
- LAYOUT
- PHYSICAL PLACEMENT OF FACILITES
15CHARACTERISTICS OF SUCCESSFUL GROUPS
- TEAM
- PRODUCTS
- FACILITIES
- GROUP LAYOUT
- TARGET
- INDEPENDENCE
- SIZE
- See Table 6.2, p. 170
16CODING SCHEMES
- BASIS OF GT
- GOAL TO COMPACTLY DESCRIBE PART CHARACTERISTICS
AND DEFINE HOW ACTIVITIES SHOULD BE PERFORMED
17Features of Good Coding Systems
- INCLUSIVE
- FLEXIBLE
- DISCRIMINATING
18ISSUES GUIDING CODE CONSTRUCTION
- PART POPULATION
- CODE DETAIL
- CODE STRUCTURE
- REPRESENTATION
- Opitz Code (F6.5, 6.6, 6.7)
19CODE DETAIL
- EFFICIENCY
- TOO LITTLE VS TOO MUCH INFO
- SHAPE INFORMATION
- SCALE OF DIMENSIONS
- SECONDARY SHAPE INFORMATION
- STANDARD PART VS CUSTOM MADE
- PRODUCTION RATE
- LIFETIME
20CODE STRUCTURE
- CODE TYPES
- HIERARCHICAL (MONOCODE)
- CHAIN (POLYCODE)
- HYBRID
- See Fig. 6.4, p. 173
21CODE REPRESENTATION
- ALPHANUMERIC VS BINARY CODES
22THE OPTIZ CODING SYSTEM
- FIVE DIGIT GEOMETRIC FORM CODE PLUS
- FOUR DIGIT SUPPLEMENTARY CODE, PLUS
- FOUR DIGIT, COMPANY SPECIFIC SECONDARY CODE
- See Figs 6.5, 6.6, 6.7
23ASSIGNING MACHINES TO GROUPS
24GROUP ANALYSIS
- ONCE PARTS ARE CODED, GROUPS MUST BE FORMED
- GOAL
- TO ASSIGN MACHINES TO GROUPS TO MINIMIZE MATERIAL
FLOW AMONG GROUPS
25STEPS IN GROUP ANALYSIS
- 1.- DETERMINATION OF PART TYPES REQUIRED BY EACH
MACHINE TYPE - MACHINE WITH FEWEST PART TYPES IS THE KEY MACHINE
and A SUBGROUP IS FORMED OF THOSE PARTS VISITING
THE KEY MACHINE AND THOSE OTHER MACHINES NEED BY
THE PARTS - See Example 6.1, p. 178
26STEPS IN GROUP ANALYSIS
- 2.- DO THE MACHINES IN THE SUBGROUP FALL INTO TWO
OR MORE DISJOINT SETS WITH RESPECT TO THE PARTS
THEY SERVICE? - IF DISJOINT SUBSETS EXIST THE SUBGROUP IS DIVIDED
INTO SUBGROUPS - EXCEPTIONAL MACHINES ARE REMOVED
27STEPS IN GROUP ANALYSIS
- 3.- SUBGROUPS ARE COMBINED INTO GROUPS OF THE
DESIRED SIZE - SUBGROUPS WITH THE GREATEST NUMBER OF MACHINE
TYPES ARE COMBINED - EACH GROUP IS ASSIGNED SUFFICIENT MACHINES AND
STAFF TO COMPLETE ITS PARTS
28THE MACHINE-PART INDICATOR MATRIX
- A BLOCK-DIAGONAL MATRIX IN WHICH ROWS ARE PARTS
AND COLUMNS ARE MACHINES - ROWS SUMMARIZE RESULTS OF STEP 1 OF GROUP
ANALYSIS - DENSE BLOCKS OF 1S FORM NATURAL MACHINE-PART
GROUPS - See Tables 6.3a and 6.3b
29BINARY ORDERING ALGORITHM
- PROVIDES AN EFFICIENT ROUTINE FOR TAKING AN
ARBITRARY 0-1 MACHINE-PART MATRIX AND TURNING IT
INTO BLOCK DIAGONAL FORM
30BINARY ORDERING ALGORITHM
- ENVISION ROWS AS BINARY NUMBERS
- SORT ROWS BY DECREASING ORDER
- ENVISION NOW COLUMNS AS BINARY NUMBERS
- SORT COLUMNS BY DECREASING ORDER
- REPEAT UNTIL ORDERING DOES NOT CHANGE
- See Example 6.2, p. 181
31Comment on BO
- BO ignores
- Machine Utilizations
- Group Sizes
- Exceptional Elements
32SINGLE-PASS HEURISTIC
- MACHINE UTILIZATION
- COMPUTE TOTAL SETUP TIME FOR PART i , fim
- COMPUTE THE TIME AVAILABLE PER MACHINE PER PERIOD
Rm - COMPUTE VARIABLE PROCESSING TIME FOR PART i ON
MACHINE m, vim - UTILIZATION uim (fimvim)/Rm
33SINGLE-PASS HEURISTIC
- 1.- REPLACE THE 1S IN MACHINE-PART MATRIX BY
ACTUAL MACHINE UTILIZATIONS (T6.4) - 2.- USING THE PART ORDERING FROM THE BOA
ITERATIVELY ASSIGN PARTS AND MACHINES TO GROUPS
34SINGLE PASS-HEURISTIC
- 3.- ASSIGN NEXT PART TO THE FIRST GROUP THAT HAS
SUFFICIENT CAPACITY ON ALREADY ALLOCATED MACHINES - 4.- IF NO GROUP HAS CAPACITY, ADD MACHINES TO THE
MOST RECENT GROUP FORMED SO IT CAN HANDLE THE PART
35Single-Pass Heuristic Example
- See Example 6.3, p. 184
- See resulting Table 6.5, p. 185
36SIMILARITY COEFFICIENTS
- EMPHASIS ON LOCATING MACHINES WITH HIGH
INTERACTION IN THE SAME GROUP - NUMBER OF PARTS VISITING MACHINE i , ni
- NUMBER OF PARTS VISITING MACHINE i AND j ,
nij
37SIMILARITY COEFFICIENT
- sij max ( nij/ni , nij/nj)
- INDICATES THE PROPORTION OF PARTS VISITING
MACHINE i THAT ALSO VISIT MACHINE j (OR
VICEVERSA, WHICHEVER IS GREATER)
38HIERARCHICAL CLUSTERING
- 1.- EACH MACHINE IS REPRESENTED BY AN ICON (NODE)
- 2.- NODES ARE CONNECTED BY LINES (ARCS)
- 3.- ARCS ARE LABELED WITH THE VALUES OF sij
- 4.- THE FINAL GRAPH IS THE MODEL
39HIERARCHICAL CLUSTERING
- 4.- ELIMINATE ARCS WITH SMALL VALUES OF sij ( lt
T ) - 5.- ALL CONNECTED MACHINES CONSTITUTE A GROUP
- 6.- DIFFERENT VALUES OF T ARE TRIED TO GET A
RANGE OF SOLUTIONS
40Hierarchical Clustering Example
- See Example 6.4, p. 186
- See dendogram on Fig. 6.9, p. 188