Cellular Manufacturing Systems

Cellular Manufacturing Systems
Lecture Series 8
IE 3265 POM
R. R. Lindeke, PhD
Spring 2005
CMS and its relationship to Job and
Flow Shops:
We can define the movement in a Job Shop
(mathematically) this way for any product i:
Pr(12)
i
= Pr(13)
i
= Pr(14)
i
= = Pr(1n)
I
While in a Flow Shop:
Pr(12)
i
= 1 and Pr(1n)
i
= 0 (n 2)
In developing CMS manufacturing systems
we are trying to make all part flows act like
Flow shop mathematics!
Examining a Cell in the CMS:
Notice MW or
multi-functional
workers this
team is
responsible for
all production
within their cell
CMS and Group Technology (GT)
CMS layout are based on recognizing similarities in
products similarities in geometry, size, materials
and processing requirements
This similar products are collected Grouped
instead of being treated as individuals
Leads to product families that visit similar
equipment and populate their cells production
schedule
Simpler setups like in a Job shop can follow and the
workers become multifunctional and responsible for
all aspects of a product and its quality
Cells can be scheduled to produce synchronously
bringing the various sub-assemblies in as needed at
final assembly with greater variety built in
NOTE: Step 1 is CMS a fundamental action in LEAN MFGing
Building
the
FACTORY
With A
FUTURE
Benefits of GT and CMS (Companies
Reporting):
52% Report reduction in new part design
10% Report reduction in # of new drawings thru
standardization
30% Report reduction in new shop drawings
60% Report reduction in IE time
20% Report reduction in floor space
45% Report reduced scrap
80% Report reduced production and quality costs
69% Report reduced set-up time (cost)
Note: Reported by companies in a survey of adopters of GT
Benefits of GT and CMS (Companies
Reporting):
70% Report reduced throughput time (even more
report better predictability of delivery)
82% Report reduced numbers of overdue orders
42% Report reduced raw-materials inventory
62% Report reduced WIP
60% Report reduced finished goods inventory
33% Report increased employee output/time unit
(productivity improvement)
Clustering Techniques: the
Fundamental Issue in Cell Development
We cluster parts to build part families
Part Families visit cells
Part Families share set-up ideas and equipment
(Family Fixtures)
Part Families follow the same (or similar) process
routing
These are the ideas and activities that offer
reported benefits
Clustering Techniques: the
Fundamental Issue in Cell Development
We cluster Machines to build cells:
Cells lead to Flow Mathematics
Cells contain all equipment needed to produce a part family
Cells allow development of Multi-functional workers
Cells hold work teams responsible for production and
quality They Empower the workers
Empowered to set internal schedules
Empowered to assign tasks
Empowered to train and rotate jobs
Etc, etc, etc
Building the CMS Facility
Before
Clustering
After
Clustering
Clustering Methods
Using Process Similarity methods:
Create Machine Part Matrices
Compute machine pairwise Similarity
Coefficient comparisons:

:
is # of parts (in matrix) visiting
both machines of the pair
is # of parts visiting one but not both machines
ij
jj
x
ij
ij
ij jj
here
x
x
S
x x
Example:
Part Number
M
a
c
h
i
n
e

I
D

X
1 2 3 4 5 6
A
1 1
B
1 1
C
1 1
D
1 1 1
E
1 1 1
Computing Similarity Coefficients:
Total Number is:
[(N-1)N]/2 = [(5-1)5]/2 = 10
For 25 machines (typical number in a small Job
Shop): 300 S
ij
s
Here they are:

1
.33
1 2
0
0
0 4
2
.67
2 1
AB
AC
AD
S
S
S

Continuing:
0
0
0 5
0
0
0 4
2
.67
2 1
0
0
0 5
0
0
0 5
2
.67
2 1
0
0
0 6
AE
BC
BD
BE
CD
CE
DE
S
S
S
S
S
S
S

Here, if the similarity

coefficient is .33
consider clustering
This criteria means
clustering:
A&D, A&B, B&D
C & E
Declustering:
A&C, A&E, B&C, B&E
and C&D, D&E

Continuing:
Examining our Matrix and our freshly clustered
machine cells, we develop 2 part families:
For the Cell A/D/B: Part Numbers 2, 3 & 5
For the Cell C/E: Part Numbers 1, 4 & 6
Care must be taken (in most cases) to assure that
each cell has all the machines it needs sometimes
a couple of families need a key machine
In this case, the manager must decide to either replicate the
common machine or share it between the cells creating a
bottleneck and scheduling problem for each cell
This is typically one of the cost problems in CMS systems
Summarizing:
Make Machine/Part Matrix
Compute Similarity Coefficients
Cluster Machines with positive ( .33) S
ij
s
Determine Part Families for the clusters (cells)
Decide if machine replication is cost effective
Re-layout facility and Cross Train workforce
Start counting your new found cash
Court customers to grow part families on Cell-by-Cell
basis
Other Clustering Methods:
Rank order Clustering
This method automates the cluster study by computing
Binary weights from a machine part matrix
It orders parts and machine cells automatically by
structuring and computing the matrix with binary weights
It implies a computer algorithm for solving the clustering
problem
It may not solve if machines are needed by more than one
family forces intelligence in application and hand scanning
after several ordering iterations
Rank Order Clustering Method:
1. For each row of the machine/part matrix (M/P/M) read the
pattern of cell entries as a binary word. Rank the rows by
decreasing binary value. Equal values stay in same order.
2. Ask if newly ranked rows in the matrix are the same as
previous order? Yes (STOP) No (continue)
3. Re-form the M/P/M with rows in new descending order. Now
rank the columns by decreasing binary word weight. Columns
of equal weight are left where they are
4. Are current column weights the same as current column
order? Yes (STOP), No (continue)
5. Re-form the matrix column order per rank order (highest to
left) and return to #1.
Lets try it with our earlier problem:
Part Number
M
a
c
h
i
n
e

I
D

X
1 2 3 4 5 6
A
1 1
B
1 1
C
1 1
D
1 1 1
E
1 1 1
Step 1:
Part Numbers
D. Equiv Rank
M
a
c
h
i
n
e

I
D

1 2 3 4 5 6
B. Wt: 2
5
2
4
2
3
2
2
2
1
2
0
A 1 1 2
3
+2
1
= 10

5
B 1 1 2
4
+2
3
= 24 4
C 1 1 2
5
+2
2
=36 2
D 1 1 1
2
4
+2
3
+2
1

= 26
3
E 1 1 1
2
5
+2
2
+2
0
=
37
1
Step 2: Must Reorder!
Step 3:
Part Number
B. WT. 1 2 3 4 5 6
M
a
c
h
i
n
e

I
D

E 2
4
1 1 1
C 2
3
1 1
D 2
2
1 1 1
B 2
1
1 1
A 2
0
1 1
D. Equiv
2
4
+2
3

= 24
2
2
+2
1
=
6
2
2
+2
1
+
2
0
=7
2
4
+2
3
=
24
2
2
+2
0
=
5
2
4
=16
Rank 1 5 4 2 6 3
Step 4: Must Reorder
Back at Step 1:
Part Number
D. Eqv Rank
1 4 6 3 2 5
B Wt: 2
5
24 2
3
2
2
2
1
2
0
M
a
c
h
i
n
e

I
D

E 1 1 1
2
5
+2
4
+
2
3
=56
1
C 1 1
2
5
+2
4
=
48
2
D 1 1 1
2
2
+2
1
+
2
0
= 7
3
B 1 1 2
2
+2
1
=6 4
A 1 1 2
2
+2
0
=5 5
Order stays the same: STOP!
Great
Cluster
Result!
Issues in Clustering:
R/O clustering oscillations indicating need of machine
replication (happens often!)
Presence of Outliers and/or Voids in the finished
clusters
Outliers indicate the need of machine replication
Voids indicate skipped machines in a cell
Generally speaking, these clustering algorithms are
designed to convert existing routes for facility re-
organization
They require a previous engineering study to be performed to
develop a series of routers on a core sample of parts that
represent most of the production in the shop
Alternative means to Develop
Cells/Families:
Most often companies rely on Classification and
Coding (C&C) systems for analyzing their part mix
These codes can be general purpose or company
specific
General Purpose:
Opitz is a german developed code for machined parts (see
over)
KC1, KC2 and KK1 systems Japanese government lab based
codes for machined parts
Brish a british developed code for general material use
Foundry codes have been developed by several groups (see
Lindeke & Rubinovich, 1987 in USA)
Examining Opitz Code:
This Form
code is the
Opitz Code
Solution on this
shaft- like part
Cells/Families:
They can be company specific
If so, they are typically hierarchical and list
important characteristics of the part/process
mix, physical characteristics like size,
geometric features, or material, etc.
Since they are specific they tend to be more
accurate in building part families
Cells/Families:
Using GT Classification and coding
systems, parts are coded by experts at the
company
The newly coded part is used to search
existing production databases for similarly
coded products
The new part is assigned to the family it
most closely matches
Its routing is thus set and only minor
variation needs to be considered
Using specific digits, a company can
target marketing in certain areas of their
product mix
Cells/Families:
In a greenfield shop, managers can
develop facility designs (in the form
of reasonable cells) by selecting
reasonable seed parts as
suggested by their GT C&C system
These seeds can be used to build
routers and, hence appropriate
machine clusters
Using GT C&C systems, process
clusters evolve from parts as
opposed to clustering evolving by
process
Life in the Cell Working with Family
Fixturing Ideas
Fixturing is a means to speed up part
loading and increase accuracy of
machine and mfg. processes
These are tools that:
Locate the work for geometric control of
various DOF
May also provide a means to guide the
tooling used to perform the operations
(Jigs)
Before being used these tool must be
accurately placed on the machine
often a time consuming task since their
placement tolerance must be 10x better
than part tolerance!
Fixturing Ideas
In CMS, it is often possible to build Family Fixtures
These are fixtures that can be shared among all the
parts in the family (because they are similar
geometrically and by mfg. process) thus reducing
time to set-up any part in the family
The Family Fixture is generic and may (likely)
require the addition of specific change pieces for
different members of the family but definitely not
different fixtures.
Fixturing Ideas
Example of Cost Savings:
Shop cost is $50/hour
Hand setup is 2 minutes/piece (lot is 400 parts)
Setup on Fixture is 0.03 min/part
Saving of 1.97 min = .033 hr = $1.64/part
If machine takes 5 minutes/part, Production rate increases
from 8.57 parts/hour to 11.93 parts/hr almost a 40% increase!
The company would invest in Fixturing tools if the cost of a
fixture applied to a given part over the life of the tooling and
part production is less than the $1.64 savings from reduced
setup times
Fixturing Ideas
Conventional fixturing means a separate
fixture for each part made
Family fixturing means a separate fixture for
each family made (but several adaptors for
individuals in the family)
Typically, FFixtures cost more than
conventional fixtures so lets do a cost
analysis
Conventional Fixturing
Facility Tool Costs:
1
1
leads to unit tooling cost/part:
i
i
i
tools
P
tools w
i
w d m f i h
P
w
i
u
where
C C
C C C C C C
C
C
P n
P = number of parts
needing tooling
C
d
is design cost;
C
m
is material cost;
C
f
is fabrication cost;
C
i
is inventory cost;
C
h
is handling cost
n = lifetime number of
parts to be made (est.)
Family Fixturing:
Cell Tooling Cost:
1
( )
leads to unit tool cost/part:
cost of adaptor
Q
tools FF a
i
a
FF d m f i h
tool
u
C C C i
C
C C C C C C
C
C
Q n
Q = number of parts in
Family
C
d
is design cost;
C
m
is material cost;
C
f
is fabrication cost;
C
i
is inventory cost;
C
h
is handling cost
n = lifetime number of each part
in family to be made (est.)
Lets do an Example:
Conventional GT Ideas
Main Tool $500 $1000
# F. Required 1/part 1 for family
Cost Adaptor NA $100
No. Adaptors Reqr NA .85/part
Typical Order Size 400 400
Typical Batch/lifetime 3batch/yr/3yrs = 3600 3batch/yr/3yrs = 3600
Costs of tooling a function of the
number of parts in a family!
No. Parts C. Tools Unit Cost GT Tools Unit Cost
1 $500
500/3600 =
.139
$1085*
1085/3600 =
.301 (.278)
2 $1000
1000/7200=.
139
$1170
1170/7200 =
.163 (.153)
3 $1500
1500/10800=
.139
$1255
1255/10800
= .116 (.111)
20 $10000
10000/72000
=.139
$2700
2700/72000
= .038
*Note: 1000 + .85*1*100 = 1085 (maybe should be 1000 in a family of 1!
Fixturing Ideas
Earlier we found the text author stating that the cost
of inventory in a batch is independent of schedule
here we see this may not be the case!
In a cell, setting up the family fixture is time
consuming but changing between family members
is quick and easy only the time to remove an
adaptor and addition of a new one (or not!)
This leads to the second rung of the factory with a
future SMED if scheduling is rational in the cells!
Lean Manufacturing is then
INTIMATELY tied to CMS and GT
These methods add
efficiency to the production
floor
They improve our quality
picture
They empower employees
They reduce setup and
product change time
They mean more
productivity
They JUST WORK!

Cellular Manufacturing Systems - Lecture Series 8: Ie 3265 Pom R. R. Lindeke, PHD Spring 2005

Uploaded by

Copyright:

Available Formats

Cellular Manufacturing Systems - Lecture Series 8: Ie 3265 Pom R. R. Lindeke, PHD Spring 2005

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cellular Manufacturing Systems - Lecture Series 8: Ie 3265 Pom R. R. Lindeke, PHD Spring 2005

Uploaded by

Copyright:

Available Formats

Here, if the similarity

You might also like