1

UNIT 5: LEAN MANUFACTURING

Manufacturing Lead Time

Push VS Pull
Kanban and ConWIP
 2

5.1. MANUFACTURING LEAD TIME

 3

NOTATION
• WLi: average workload for station i (average time required to complete the

processing of a work part at station i). There are going to be n stations, plus the
part-handling system (station n+1).

• Rp*: maximum production rate of all parts, given by the capacity of the

bottleneck station, which is the station with the highest workload per server. The
bottleneck will remain the same provided that the product mix remains constant.

• Tw: mean waiting time experienced by a part due to queues at the stations.

• MLT: Manufacturing Lead Time (time spent by a part in the system).

• N: number of work parts in the systems, i.e. work-in-process (WIP).

 4

EXTENDED BOTTLENECK MODEL

• Manufacturing Lead Time will be the sum of processing times at the stations,

transport times between stations, and any waiting time experienced by the parts in
the systems:
n
MLT  WLi  WLn 1  Tw
i 1

• Little´s formula can be expressed as:

N  R p  MLT

where Rp stands for the production rate of the system.

 5

NON-BOTTLENECK CASE
• When N is small, MLT will take on its smallest possible value, since the waiting

time of a unit will be short o even zero: n

MLT1   WLi  WLn 1
i 1

• The production rate is be less than the maximum, because some of the stations

will be idle and the bottleneck may not be fully utilized:

N
Rp 
MLT1
• The critical value of N, the dividing line between the cases:

 n

N *  R*p   WLi  WLn 1   R*p MLT1
 i 1 
 6

BOTTLENECK CASE
• If N is large, the system will be fully loaded, with parts waiting in front of stations.

Since the system is congested, the production capacity will reach the maximum
(Rp*). However, the WIP will be high and the Manufacturing Lead Time will be
long. By applying Little’s formula:
N
MLT2  *
Rp

• The mean waiting time a part spends in the system can be estimated as:

 n 
Tw  MLT2    WLi  WLn 1   MLT2  MLT1
 i 1 
 7

BEHAVIOR OF THE SYSTEM

Both Manufacturing Lead Time and Production Rate as functions of the WIP of the system:
 8

INSIGHTS

• If work-in-process (N) is below N*, MLT has a constant value (it cannot be less
than the sum of the processing and transport times), while Rp decreases
proportionally with N, because stations become starved for work.
• If work-in-process is above N*, Rp will be equal to the maximum (it cannot be
greater than the output capacity of the bottleneck station), whereas MLT increases
with N because backlogs build up at the stations.
• Although the optimum N value seems to occur at N* (minimum MLT and maximum
Rp), it is intended to be a rough-cut method to estimate performance in early
phases of the design of the system. More reliable estimates should be obtained
using computer simulations.
 9

CAPACITY MANAGEMENT 100 %

Cycle Time
• Traditional: Capacity is chosen to meet Production Production
Feasible Infeasible
demand only, at the minimum cost. Cycle constant
lead time
time is not an issue.

Utilization
• Modern: Lead times and WIP levels, i.e.
100 %
congestion, grow continuously with

Cycle Time
increasing utilization. We must consider
other measures of performance in addition
to cost and throughput.

Utilization
 10

5.2. JUST-IN-TIME
 11

INTRODUCTION

• JIT (Just-in-Time) has its roots in the Japanese industry (Toyota Motor), which
could not reduce costs by exploiting economies of scale in mass production
facilities, due to space and resource limitations. Instead, the focus was on waste
elimination.
Pillars
• Producing only what is needed, or acquiring the required materials from upstream
workstation precisely as needed.
• Autonomation: machines that are both automated, so that one worker can
operate many machines, and foolproofed, so that they automatically detect
problems.
 12

INTRODUCTION
 13

JIT GOALS: SEVEN ZEROS

Ideal goal: achieve zero inventories.

• Zero Defects: To avoid delays due to defects. (Quality at the source)

• Zero (excess) Lot Size: To avoid delays of parts waiting for the batch to be produced.

(Usually stated as a lot size of one.)

• Zero Setups: To minimize setup delay and facilitate small lot sizes.

• Zero Breakdowns: To avoid stopping lines with low WIP.

• Zero (Excess) Handling: No extra moves other than necessary.

• Zero Lead Time: Reducing queuing time to achieve perfect just-in-time flow.

• Zero Surging: Smooth production plan without sudden changes, because there are no

WIP to level changes.

 14

VALUE
STREAM
MAPPING
• This technique
allows to identify
activities that do
not add value,
and to rearrange
the processes in
order to decrease
the lead time.
 15

ENVIRONMENT AS A CONTROL

Unlike Western methodologies, which takes constraints as fixed, Just-in-Time

encourages to shape proactively the environment, so that the manufacturing
systems can be easier to manage, instead of trying to optimize isolated problems.

• Optimizing lot size VS reducing setup time or setting up long-term agreements

with vendors to make frequent deliveries feasible.

• Optimizing a production schedule VS negotiating due dates with customers and

integrating marketing and manufacturing.

• Setting up inspection procedures to find quality defects VS ensuring that both

suppliers and workers meet quality requirements.

 16

HEIJUNKA

Production smoothing: developing a repetitive manufacturing environment. It

requires relatively constant volumes and relatively constant product mix.

Example: 10,000 units per month (20 working days), 500 units per day (2 shifts)

250 units per shift (480 minutes), 1 unit every 1.92 minutes. (takt time).

Production Sequence: Mix of 50% A, 25% B, 25% C in daily production of 500 units.

• 0.5 x 500 = 250 units of A

• 0.25 x 500 = 125 units of B

• 0.25 x 500 = 125 units of C

A–B–A–C–A–B–A–C–A–B–A–C–A–B–A–C…
 17

CAPACITY BUFFERS

• Since JIT is intrinsically rigid with respect to the production volume, product mix

and sequences, there should be capacity buffers to protect against quota

shortfalls due to unanticipated disruptions.

• If production falls behind the desired rate, either because of problems in the line or

because of changes in the requirements, then the extra time is used.

• Two shifting: a popular approach is to schedule shifts “4 – 8 – 4 – 8”. The down

periods can be used for preventive maintenance or catch-up, if necessary.

 18

SETUP REDUCTION
Setup times must be reduced because small lot sequences are not feasible with large
setups.
• External setups: performed while machine is still running

• Internal setups: performed while machine is down

Approaches:
1. Separate the internal setup from the external setup.
2. Convert as much as possible of the internal setup to the external setup.
3. Eliminate the adjustment process.
4. Abolish the setup itself (e.g., uniform product design, combined
production, parallel machines).
 19

QUALITY
JIT requires high quality to work. Otherwise, there will be line stops and a
throughput decrease.

JIT promotes high quality:

• identification of problems

• facilitates rapid detection of problems

• pressure to improve quality

 20

QUALITY TECHNIQUES

• Process Control (SPC)

• Poka-yoke

• Insistence on Compliance (quality first, output second)

• Stop the line or the process (Jidoka).

• Correcting One's Own Errors (no rework loops)

• Continuous Improvement

• Small Lots

• Vendor Certification

• Total Preventive Maintenance

 21

5.3. KANBAN and ConWIP

 22

KANBAN AND PULL SYSTEMS

• Toyota Kanban system: cards were used to control the flow of materials through
the plant.
• In a push system, work releases are scheduled (and prepared in advance),
whereas in a pull system, releases are authorized (depending on the status on
the plant).
• A push system accommodates customer due dates, but has to be forced to
respond to changes in the plant (e.g. MRP must be regenerated). On the other
hand, a pull system directly responds to plan changes, but must be forced to
accommodate customer due dates (e.g. by using overtime to maintain the
production rate).
 23

MRP vs KANBAN
MRP
Lover
Level … Assem
Inven- -bly
tory

Kanban
Lover
Level … Assem
Inven- -bly
tory
…

Kanban Signals Full Containers

 24

KANBAN AND PULL SYSTEMS

• In the Kanban system, production is triggered by a demand: when a part is

removed from an inventory point (finished goods inventory or WIP), the
workstation feeding that inventory point is given authorization to replace the part.

• This workstation then sends an authorization signal to its upstream workstation to

replace the part it just used. In other words, each station replenish the
downstream void and sends authorization to the next workstation upstream.

• The Kanban system developed at Toyota made use of two types of cards:
authorizing production and movement of product.
 25

TWO-CARD KANBAN: TOYOTA

• When a workstation becomes available, the operator takes the next production

card from a box. It tells the operator that a particular part is required at a
downstream workstation. If the necessary materials are available, the operator
removes the move cards attached to them, processes the part, attaches the
production card, and places it in the outbound stock point.

• Periodically, the box containing move cards are checked: the operator will move

the materials from the outbound to the corresponding inbound stock points,
replacing their production cards with move cards. The removed production cards
will be deposited in the boxes of the workstation from which they came.
 26

TWO-CARD KANBAN: TOYOTA

Inbound Outbound Move stock to
stock point stock point inbound stock point.

Move card
authorizes
pickup of parts.

When stock is Remove

removed, place move card
production card and place in
Production in hold box. Move hold box. Production
cards cards card
authorizes
start of work.
 27

TWO-CARD KANBAN

• Two-card system is justified when workstations are spatially distributed, hence it is

not feasible to achieve instantaneous movement of parts. Therefore, there are

WIP stored in two places: inbound and outbound stock points.

• The two-card system can be modelled as an one-card system, in which the move

operations are treated as workstations. The choice of one over the other depends
on the extent to which we wish to regulate the WIP involved in move operations.
 28

ONE-CARD KANBAN
• The worker still requires a production card and the necessary materials to begin processing.

However, the worker simply removes the production card from the upstream process and
sends it back.

Outbound Completed parts with Outbound

stock point cards enter outbound stock point
stock point.

Production When stock is Production card

cards removed, place authorizes start
production card of work.
in hold box.
 29

KANBAN AND BASE STOCK MODELS

• The key controls in a Kanban system are the WIP limits at each station. The

limits can take the form of a card count, number of containers, volume limitation,
or even be supervised by electronic means.

• Kanban shares mechanics with base stock models: each time inventory in the

downstream stock point falls below m (m = number of production cards),

production cards are freed up, authorizing the station to replenish the buffer. It
corresponds with a base stock model, with the downstream station acting as the
demand an the card count m serving as the base stock level.
 30

CONWIP

• ConWIP = Constant Work in Process. It establishes a limit on the WIP in the line. It

does not allow releases into the line whenever the WIP is at or above the limit.

Finished
Parts buffer Parts buffer goods
inventory

When parts are

Remove production withdrawn, return cards to
card and attach to box at beginning of line
Production cards completed part
 31

CONWIP

• Departing jobs send production cards back to the beginning of the line to authorize

release of new jobs. ConWIP assumes that:

1. The production line consists of a single routing.

2. Jobs are identical, so WIP can be measured in units.

• If the facility contains multiple routings, a ConWIP level can be established for

each routing.

• If the jobs require different amounts of processing on the machines, the ConWIP

levels can be stated in units of “standardized jobs”.

 32

CONWIP vs KANBAN

• Kanban requires setting more parameters than does ConWIP, e.g. the user must

establish a card count for every station.

• Kanban cards are typically part number – specific, i.e. cards in a Kanban system

identify the part for which they are authorizing production. Therefore, a Kanban
system must include containers of WIP for every active part. In addition, it has to
maintain WIP on the floor for parts with low demand.

• ConWIP cards do not identify any specific part number. Instead, a release list

gives the sequence of parts to be introduced to the line. This list must by
generated by a module outside the ConWIP loop.
 33

5.4. PUSH vs PULL

 34

INTRODUCTION

Traditionally, the term pull has been misunderstood:

• Pull is not a synonym for Kanban (there are more control-based pull control
systems).
• Pull does not imply make-to-order (MRP can consist of customer orders, and
thus make-to-order, but MRP is always a push system).
The essential difference between push and pull is the mechanism that triggers the
movement of work:
• A push system (e.g. MRP) schedules the release of work based on demand.

• A pull system (e.g. Kanban) authorizes the release of work based on system
status (i.e. inventory voids signaled by cards).
 35

PUSH vs. PULL MECHANICS

• Push systems schedule work releases on the basis of information from outside the

system, whereas pull systems authorize releases based on information from inside the
system.
PUSH PULL

(Exogenous)
Schedule
(Endogenous)
Stock Void

Production Process Production Process

Job Job

Push systems do not limit Pull systems deliberately

WIP in the system. establish a limit on WIP.
 36

PUSH AND PULL EXAMPLES

• ETSI copy shop: Push

• Vending machine: Pull

• MRP system: Push

• Emergency Room at the hospital: Push into waiting room, Pull into examination

room.

• Supermarket (goods on shelves): Pull

• Runway at Seville airport during peak periods: Pull

• Order entry server at Amazon.es: Push

 37

THE MAGIC OF PULL

• By establishing a WIP cap (i.e. limit on WIP), pull systems place an emphasis on

material flows: if production stops, inputs stop.

Reduces costs:
• Prevents WIP explosions.

• Reduces average WIP, thus decreasing holding inventory costs.

• Reduces engineering changes (it is easier to customize or change an order before releasing it to

the shop floor).

Maintains flexibility:
• Avoids early release, thus facilitating engineering changes.

• Less congestion and less reliance on forecasts.

 38

THE MAGIC OF PULL

Improves Quality:
• Pressure for higher quality, because a workstation will starved if the parts in its inbound buffer do

not meet quality standards.

• Improved defect detection.

• Improved communication between workers, because downstream operators are likely to check the

quality of parts from upstream workstations.

Improves Customer Service Quality:

• Reduces cycle time explosion and variability.

• Pressure to reduce sources of process variability.

• Promotes shorter lead times and better on-time performance.

 39

INVENTORY ORDER / INTERFACE

• The inventory/order (I/O) interface describes the point in all production systems

in which the trigger for material flow shifts from make-to-stock to make-to-order.

• There is the benefit of eliminating an entire portion of cycle time seen by

customers by building up stock, at the expense of losing flexibility.

• Example: Five Guys production line:

I/O Interface
Make-to-Stock Make-to-Order

Refrigerator Cooking Assembly Packaging Sales

Customer
 40

INVENTORY ORDER / INTERFACE

• Example: McDonald’s production line:

I/O Interface
Make-to-Order Make-to-Stock

Refrigerator Cooking Assembly Packaging Sales

Warming Customer
Table

• I/O interface can differ by time of day (or season).

• I/O interface can differ by product.

 41

INVENTORY ORDER / INTERFACE

• Moving I/O interface closer to customer shortens lead time seen by customer, i.e.

increases responsiveness but (usually) also increases inventory.

• Tradeoff: responsiveness vs. inventory (time vs. money)

• The optimal position of I/O interface depends on the need for responsiveness,

along with the cost of carrying inventory and the need for customization of the
products.

• Delayed differentiation of products (for different countries) enables pooling of

safety stocks, while postponement can be used to facilitate rapid customer

response in a highly customized manufacturing environment.
 42

5.5. CONCLUSIONS
 43

CONCLUSIONS

• Production environment itself is a control. Strategies that involve reducing

setups, changing product designs, leveling production schedules, and so on, can
have a great impact on the effectiveness of the process.

• Controlling WIP is important. All benefits of JIT either are a consequence or low

WIP levels (e.g. short cycle times) or are supurred by the pressure low WIP levels
create (e.g. high quality levels).

• Flexibility is an asset, since JIT is inherently inflexible (steady rate and mix of

production). In order to respond to a volatile marketplace, flexibility must be

promoted: short setup times, capacity cushions, worker cross-training, etc.
 44

CONCLUSIONS

• Quality can come first. Systems in which quality takes precedence over

throughput and is assured at the source work and are profitable.

• Continual improvement is a condition for survival.

• Variability must be identified and reduced, by determining its root source and

eliminating the course.

• The magic of pull systems is that they establish a WIP cap, which prevents

producing unnecessary WIP that does not significantly improve throughput.

• Adjusting the location of the I/O interface alters the balance between speed and

flexibility.