Lean Production Systems QEM-1103-2

Continuous Improvement
Kaizen
Chapter 2
Leila Emamitaba

1
 Continuous Improvement
• Improvement: Implies that something about a process has
changed for the better while trying to minimize cost and adverse
consequences.

• Continuous improvement is measured in terms of producing

things better, faster, and cheaper, and being more agile.

• Paralleling the idea of continuous improvement is the concept of

elimination of waste .

• A main reason for Toyota’s success: Continuous Improvement is

embedded in Toyota’s vision, goals, expectation of employees,
suppliers, daily operations, and management practices and culture

2
 PDCA Process Improvement Cycle
(aka Deming Cycle or Shewart Cycle)

P
Plan

A Act Do
D

Check

3
 PDCA
PLAN

1. Collect data. Systematic observation and

documentation to understand and analyze situation
2. Define problem. Clear statement of the problem and
the root cause. e.g., use the “Five-why”.
3. State the goal. Define desired situation after the
problem has been solved.
4. Analyze and solve the problem. Prepare plan to
implement the solution. Include schedule, people
responsible, etc.

4
 The Five-Why Analysis

Taiichi Ohno
“You can eliminate the source of a problem, but unless
eliminate the root cause, the problem will just keep coming
back.”

Five-Why Analysis: a method for separating root causes from

symptoms:

Ask “why?” at least five times.

5
 The Five-Why Analysis
Example
Problem: Stone at the Jefferson Memorial was crumbling

1.Why was it crumbling? Frequent washings of the stone was the primary cause.

2.Why was it washed so often? To remove bird droppings.

3.Why were there so many birds? There is an abundant food supply, hundreds of
spiders.

4.Why were there so many spiders? The spiders were attracted to the many midges.

5.Why were there so many midges? Every evening at dusk they emerge in a mating
frenzy. At the same time the Park Service turns on its powerful spotlights. The
midges are then attracted to the lights.

Solution: Delay lighting until one hour after sunset.

Result: Midge population dropped by 90%, breaking the food chain!
6
 The Five-Why Analysis
Example
Problem: Defective parts

1. Why are the parts defective? The machines on which they are produced do not
maintain the proper tolerance.
2. Why do the machines not maintain the proper tolerance? The operators of the
machines are not properly trained.
3. 3. Why are the operators not properly trained? The operators keep quitting and
have to be replaced with new ones, so the operators are always new.
4. 4. Why do operators keep quitting? Working at the machines is repetitive,
uncomfortable, and boring.
5. Why is the work at the machines repetitive, uncomfortable, and boring? The tasks
in the operator’s job were designed without considering their effect on human
beings.

Suggested Solution: Redesign the job so that it is less uncomfortable and more
fulfilling for the operator
7
 PDCA
Do
Implement the plan.

Check
Monitor and measure results. Compare to goals.

Act
Take appropriate action
keep plan and expand it
revise plan
dump plan; try something else
Go See: Return to Plan for further improvement.
8
PDCA-Example

Plan Do
Plan for the new Create a prototype, test it
product development and collect data from the
& production process customer

ACT Check
Fully Implement the Analyze the collected
new design/act on data to measure
collected feedback customer satisfaction

9
 6-Sigma Quality-Improvement Process:
DMAIC
Method to enact the GE’s six-sigma philosophy.
Used by employee teams in projects to tackle specific problems.
Goal: to minimize process variation.
Define (D)
Define problem, the customer of problem, and critical-to-quality attributes (CTQ’s) that
customer considers most important.
Measure (M)
Identify processes that influence CTQ’s. Measure their performance.
Analyze (A)
Determine causes of problems and poor performance in those processes; determine
key factors causing large or erratic process variation.
Improve (I)
Confirm impact of key factors on CTQ’s. Determine methods for measuring variation,
maximum acceptable range of variation, and methods to make processes acceptable.
Control (C)
Employ methods to ensure that processes stays within acceptable range
10
 PDCA and DMAIC

Both rely heavily on data analysis

and use the “Magnificent Seven”
data collection and analysis tools

11
 Magnificent Seven data
collection and analysis tools

 Check sheet
 Histogram
 Pareto analysis
 Scatter diagram
 Process flowchart
 Cause-and-effect (fishbone) diagram
 Control charts and run diagrams

12
 Check Sheet Inspector/Operator
Department FF #6
Geeves, R. S Date 5 March 07
Part/Product OS 94/95

Type of Defect Tally Total

1
• Customized for each situation Cracked switch
4
• Clear terms Cracked cap

• Special training for observer Missing cap pin 1

Paint defects on housing 11

Dent on housing 4

Loose housing 1

Loose handle 5

Other, please specify

2
Table shows a check sheet for Missing cap. Chipped handle

tracking several possible types of

defects observed during the final Total units inspected 132 total defects 29

inspection of a product Special cases (please specify)

Observations

13
 Histogram

Histogram graphically shows the frequency distribution of a variable.

Histogram do not suggest the cause of variation or problems.

This histogram shows the distribution of total daily defects recorded for 31
days of observations.

14
 Histogram
This histogram shows the distribution of types of defects observed over 20 days.

15
 Histogram
This histogram shows the distribution of paint defects for each of the last six months.

16
Histogram

17
 Pareto Analysis
Pinpoint the “vital few” out of the “trivial many”

It seeks to identify are those relatively few

problems that occur with:

• The greatest frequency

• The biggest dollar loss,
• The greatest number of defects

18
 Scatter diagram

Scatter diagram shows relationship between two variables

This scatter diagram shows plant temperature
and number of blurry-edge defects occurring in a 40-day observation period

19
 Process flowchart

• A process flowchart shows the steps in a process and is useful for

analyzing a process to pinpoint sources of problems.

• The chart should show all relevant activities, value added as well as
nonvalue added.

20
 Process Flowchart- paint process

This figure shows the flowchart for the

portion of the paint process that is relevant
to the blurry-edge problem.

21
Process Flowchart
“Supply Flow, Operating Room”

22
Flowchart (Value stream map)

23
 Cause-and-effect Analysis
(Ishikawa diagram or fishbone diagram)

• Cause-and-effect analysis is used to identify all of the possible

contributors (causes) to a given outcome (effect).

• As is typical, causes are divided into the categories of manpower,

materials, methods, equipment, and environment.

• The team brainstorms to generate as many ideas as possible about

causes for the problem.

24
 Cause-and-effect diagram-Example
This diagram showing possible causes for blurred edges around the
product logo

25
Cause-and-effect diagram-Example
Detailed analysis

26
Cause-effect diagram: Midway Airlines-
Causes of Flight Departure Delays

27
Cause-effect (fishbone) diagram:
“Over shipments”

28
Cause-effect: “Late for Work”

29
 Cause-effect diagram and brainstorming

Benefits

• Educational and stimulating

• Helps keep focus on the issue at hand
• Stimulates active search for causes
• Motivates data collection
• Demonstrates high level of understanding: the more complex the
diagram, the better the understanding

30
 Control charts

• Involves sample inspection

• Shows control limits as determined by statistical formulae
• Shows process variation
• Assess whether variation is from random or nonrandom causes

31
 Run diagrams

• A run diagram shows the results of observations taken at

prescribed intervals (e.g., every 10th unit or one unit every 10
minutes)
• Involves 100% inspection
• Shows process variation
• Assess patterns in variation

32
 Run diagrams-Example

• This diagram shows the results of 100% inspection for two

shifts.
• The diagram indicates no clear pattern, which is often the case
when the period of observation is short.

33
 Run diagrams-Example

• This diagram shows the results of 100% inspection for 10 days. It

suggests a general increase in severity of blurred edges during the
day shift and a general decrease during the night shift.

34
 Example: “Causes of missed schedules”

1. Brainstorm possible causes, place on diagram and identify most

feasible causes.
 Example: “Causes of missed schedules”

2. Collect data. Use Pareto to identify major causes.

Example: “Causes of missed schedules”
 Example: “Causes of missed schedules”
4. Brainstorm reasons for the major cause.
 Example: “Causes of missed schedules”
5. For each reason, do detailed analysis: collect data, identify root
causes (5-why, fishbone, Pareto, etc.)

• Rely on facts not perceptions, or preconceptions

• Before doing anything, collect the data, get the facts
• Use the data as a baseline to drive the initiative and assess progress.
• In practice, the basic problem-solving tools are used in combination, as
needed

The Data Don’t Lie!