A
Classroom Exercise
To
Illustrate Lean Manufacturing Pull Concepts
Peter J. Billington
Professor of Operations
Management
Hasan School of Business
Colorado State University -
Pueblo
2200 Bonforte Blvd.
Pueblo, CO 81001
719-549-2880
peter.billington@colostate-pueblo.edu
A
Classroom Exercise
To
Illustrate Lean Manufacturing Pull Concepts
This exercise is adapted from
the Hewlett-Packard (1983) stockless production “traveling road show”
training tape, and the Cellulose Aircraft, Inc. exercise in Heineke and
Meile (1995), which provides the basis of paper airplane folding. The
HP tape is an excellent presentation of push and pull, however, the
exercise in this paper engages the students in a hands-on production
line that focuses specifically on the push vs. pull aspect of lean
manufacturing.
The goals
of this exercise are to illustrate the significant reduction in WIP and
throughput time using the pull system, to explain and illustrate the
concepts of push, pull, kanban, bottleneck, cycle time, idle time, line
balance, and to illustrate worker behavior in an operational setting.
This exercise utilizes four students in an
assembly line to build a paper airplane. The work center tasks are
assigned in such a way as to place the bottleneck at the third
workstation. The students first work in a standard push approach: each
student works at their own comfortable pace as long as inventory is
available, and they do not worry about inventory buildup. The second
run requires a pull of 3 units at a time, and the third run requires a
pull of one unit at a time. The reduction in inventory and throughput
time is readily evident to all, providing a powerful learning
experience. In addition, student discussion of a multitude of lean
concepts and behavioral issues is easily elicited.
This exercise can be completed in about 50
minutes, making it feasible to do in a standard 1 hour class time
block. The basic multi-fold paper airplane using blank, standard 8 ˝ x
11 inch paper is used. The basic folds are easily explained, as shown
in Figure 1.
Material
required:
·
100 sheets of
white paper (probably more than you will actually use)
·
3 sheets of
colored paper
·
Color marker
·
Forms for data
collection:
¨
Task time
sheets (12) – one for each work center for each of three runs – observer
indicates times for 20 units. See Form 1.
¨
Cycle time
sheet (3) – one for each run of the game – observer indicates cycle time
for 20 units that come off the end of the assembly line. See Form 2.
¨
Throughput time
sheet (1) – observer indicates the throughput time for the color sheet
to move from raw material inventory to the end of the production line;
time once each run of the game. See Form 3.
The four work centers are set in a line at
tables in the classroom, with chairs for the workers to sit, as shown in
Figure 2. About 12 linear feet of table space is required to insure
sufficient space for beginning raw material, WIP between each student,
and finished goods inventory. A stack of paper is placed before the
first work center (raw material).
Four students are asked to work on the
production line. The students are given the instruction sheet, Figure
1, and start a practice run to get any learning effects out of the way,
insure correct folding, and to push some inventory into the system.
Since the objective in this exercise is the push and pull concepts, the
quality is monitored by the instructor only to insure that they are not
rushing improperly to try to balance the line. During the practice, the
instructor can correct any quality problems. If the instructions are
“work at a comfortable pace, do not speed, this is not a contest for
maximum output, but quality will be monitored by the instructor,”
students will work accordingly.
Seven students are asked to time the
activities and collect WIP quantities. Work centers one to four will
have task times monitored using Form 1. It is important to instruct
students to time only the actual hands-on task time for the worker,
since work center 4 will have idle time and will not be working all the
time. One student will collect the cycle time at the end of the line,
the time between completion of successive units, using Form 2. The
student should be instructed to start timing when a plane is finished by
work center 4, count the production of the next 20 planes, then stop
timing when this last, 21st, plane is finished. Dividing by
20 will give the average cycle time. The sixth student will note the
throughput time of the color sheet using Form 3, discussed below. The
seventh student will count WIP on each run.
The remaining students should
gather around and be instructed to keep an eye on the workers and the
action, with knowledge that they are to provide insightful discussion at
the conclusion.
The first run is the push run: let them
make inventory as fast as they want and push to the next work center.
After a few minutes, as the inventory builds, have the observing student
count the WIP without stopping the line. We have found that writing the
results in a table on the board will provide easy comparisons, as shown
in Figure 4.
At an opportune time, insert the colored
sheet onto the top of the raw material paper. Make sure in the push run
that this sheet is not conveniently placed on the top of subsequent
piles of inventory, thus resulting in a relatively fast throughput.
Insist on a FIFO type of process between work centers; the instructor
can easily move the sheet to the bottom of a pile if the students try to
use LIFO. The student observer is instructed to track the throughput
time: the time from when the first worker takes the sheet from the raw
material pile, to the time when the last worker completes the plane and
places it in the finished goods inventory. This throughput time should
be written on the board as shown in Figure 4. When the colored sheet is
out of the line and the throughput time noted, the run can be stopped if
at least 21 units have made their way through the end of the line and
all the data collection has been completed.
At the conclusion of each run, the
observers are asked to find their average task times, cycle time, and
throughput time. These results will be added to the table on the white
board as shown in Figure 4. The tasks times should indicate that work
center 3 is the bottleneck, and since the bottleneck is really the
deciding factor in throughput, and thus the cycle time should be about
the same as the bottleneck’s task time for each run.
A
discussion at this point should result in the identification of the
bottleneck, the piles of inventory, and observation that work center 4
is idle some of the time while the other work centers worked
continuously. This is a good time to reinforce the concepts of
bottleneck, cycle time, throughput time, push, line balance and worker
behavior.
The second run is a pull 3
run. Designate a space between each work center as an imaginary kanban
box where inventory will be held as shown in Figure 3. Place 3 units of
WIP in each of the 3 kanban areas. Indicate that one worker’s in box is
the previous worker’s out box. A worker can start work at their work
center only when 1) there are 3 units in the incoming space, the kanban
box, between their work center and the previous work center (except for
work center 1, who has a stack of raw material paper), and 2) when there
is no inventory in the box out of their work center. When these
conditions are met, the empty out box is considered a kanban signal that
more production is needed. The worker can then pull the 3 units out of
the in box (only when there are 3 units in the in box) and do the work.
When they are done they place the 3 units in the out box and wait until
the box is empty again. Work center 4 can pull the 3 out of the in
kanban box only when he or she has completed the last batch of 3. This
is the time to discuss the kanban as a visual signal that is pulling the
production of planes.
Since this is very different
from the push system, students may need some observation and help to
make sure they are following the pull discipline. After a few trials to
make sure the pull is working, start the run. Repeat the insertion of
the color sheet, and the collection of the WIP quantity, throughput time
and cycle time. During the pull run, an instructor can have some fun
asking the bottleneck worker what the problem is since this is holding
up the output, while also asking the idle workers why they are sitting
around doing little work.
The pull system requires 3 of
the 4 work centers to be idle part of each cycle, a process that is
counter to our usual thinking. It is actually good, in this case, to be
idle! The alternative is a build-up of inventory. If you use
Goldratt’s The Goal, this is a perfect time to discuss similar
key concepts in that book.
After this pull 3 run and
debriefing, repeat with a kanban box holding one unit, showing even more
performance improvement. Discussion will focus on the dramatic
reduction in the performance measures of inventory and throughput time.
In addition to the discussion
of the key learning objectives noted earlier, students should be able to
discuss the following. Discussion of many other topics will follow.
·
Students are
often astounded to discover that as the inventory is reduced, the
throughput time is also reduced. In addition, three workers are idle.
How can this be?
·
Work center 3
has the longest task time and is the bottleneck. The bottleneck worker
can be asked how he or she felt as the only one working all the time
while the other three workers had idle time.
·
If this is a
work team, how will they go about resolving the obvious imbalance in
work loads?
·
In the pull
system, inventory build-up hid the fact that the first two work centers
were out of balance with the bottleneck.
·
The line can be
balanced in a variety of ways. Use the task times to determine how to
combine tasks to balance the line.
·
Work centers 1
and 2 were often idle in the pull runs, but not the push system.
·
Workers often
do not communicate with one another. Worker communication must be
enabled if this team is to improve the production process.
·
The concept of
pull is readily apparent.
Other classroom material can
be used effectively in conjunction with this exercise. Goldratt’s
The Goal is natural for bottlenecks and throughput, and the issues
of people (or machines) sitting idle because they are driven by the
bottleneck speed. Why do we keep those people (or machines) busy when
they really should be working at the pace of the bottleneck?
At the conclusion of a recent
semester, the author surveyed students on the most important learning
aspects of the course. The airplane exercise was rated one of the most
important learning activities in the class. Students noted that the
basic lessons learned were remembered through to the conclusion of the
course. The most remembered concept was that as the inventory was
reduced so was the throughput time.
References
Figure
1
Paper
Airplane Work Center Assignments
Work
Center 1
Fold up in
half long way.
  
Long fold
Work
Center 2

Fold first
corner on each side. Fold down on each side.
Work
Center
3
Long fold
 
Second
fold of wing on each side. Fold down on each side.
Draw a
star with a colored marker on one side
Work
Center 4
Third fold
of wing on each side. Place in finished goods inventory.
 
Fold down
on each side.
Figure
2
Push
Production Line Setup
Table length ~12 feet
|
|
Student |
|
Student |
|
Student |
|
Student |
|
|
Raw
Material
(Paper) |
Work Center
1 |
WIP |
Work Center
2 |
WIP |
Work Center
3 |
WIP |
Work Center
4 |
Finished goods
(Airplanes) |
Material Flow
Figure
3
Pull
Production Line Setup
Table length ~12 feet
|
|
Student |
|
Student |
|
Student |
|
Student |
|
|
Raw
Material
(Paper) |
Work Center
1 |
Kanban Area |
Work Center
2 |
Kanban Area |
Work Center
3 |
Kanban Area |
Work Center
4 |
Finished goods
(Airplanes) |
Material Flow
Figure
4
White
Board Data Collection
|
|
|
|
|
Task
|
Times
|
|
|
|
|
WIP
Units |
Throughput Time |
Work Center 1 |
Work Center 2 |
Work Center 3 |
Work Center 4 |
Cycle Time |
|
Push |
|
|
|
|
|
|
|
|
Pull 3 |
|
|
|
|
|
|
|
|
Pull 1 |
|
|
|
|
|
|
|
Form 1 (abridged)
Paper Airplane Manufacturing
Task Time
Work Center ___________
|
Unit |
Seconds |
|
1 |
|
|
2 |
|
|
.
. |
|
|
19 |
|
|
20 |
|
|
Sum |
|
|
Average |
|
Form 2
Paper Airplane Manufacturing
Cycle Time
Run _________
|
Time |
|
|
|
Time at completion of
21st unit |
|
|
Time at completion of
first unit |
|
|
Difference, in seconds |
|
|
Cycle time = difference
divided by 20 |
Form 3
Paper Airplane Manufacturing
Color Sheet Throughput Time
|
System |
Throughput time |
|
Push |
|
|
Pull 3 |
|
|
Pull 1 |
|
|