Quality Management – Prof. Schmitt
Lecture 06
Lecture Quality Management
06 Quality Management in Manufacturing
Prof. Dr.-Ing. Robert Schmitt
© WZL/Fraunhofer IPT
© WZL/IPT
Quality Management in manufacturing
L 06 Page 0
Quality Management – Prof. Schmitt
Lecture 06
Contents
„ Detection and Prevention of Waste
„ Development of Rapid Saving Potentials with the 5S
„ Efficient Modelling of the Value Stream with Value Stream
Design
„ Securing of the Process Capability with Statistical Process
Control (SPC)
© WZL/Fraunhofer IPT
Page 1
Literature:
Kleppmann, W.: Taschenbuch Versuchsplanung: Produkte und Prozesse optimieren. Brunner, F.J. (Hrsg.); Carl
Hanser Verlag; München, Wien, 2003
Pfeifer, T.: Qualitätsmanagement Strategien, Methoden, Techniken; Carl Hanser Verlag; München, 2001
Pfeifer, T.: Quality Management, Strategies, Methods, Techniques; Carl Hanser Verlag; München, 2002
Pfeifer, T.: Praxisbuch Qualitätsmanagement Aufgaben, Lösungswege, Ergebnisse; Carl Hanser Verlag; München,
2001
Quentin, H.: Versuchsmethoden in Qualitäts-Engineering; Friedr. Vieweg & Sohn; Braunschweig, Wiesbaden, 1994
© WZL/IPT
Quality Management in manufacturing
L 06 Page 1
Quality Management – Prof. Schmitt
Lecture 06
Waste and Added Value
Added value is increased by elimination of waste
ed v
Add
alue
V
W
Initial situation
Was
te
Wrong!
Increase in output
by compression!
Added value
Waste
Added value
W
W
Added value
=
W
Correct!
Replacement of
waste by added
value!
V
© WZL/Fraunhofer IPT
Page 2
Waste and Added Value
- The added value of the manufacturing process is increased by the elimination of waste.
- Increase in output by compression not only increases the added value, but also the waste.
- With the replacement of waste by added value the output can be increased effectively and
sustainably.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 2
Quality Management – Prof. Schmitt
Lecture 06
Areas and Starting Points of Waste
There are at least 7 starting points to detect waste ...
Overproduction
Failures/
Repair
Transports
Waitingperiods
Depots
Stocks
Distances
...which can be found in all areas:
Process
© WZL/Fraunhofer IPT
Product
Machine
Page 3
Areas and Starting Points of Waste
Waste can appear everywhere, in processes as well as in the product itself of in production machines.
The following starting points help to detect waste:
- Overproduction
- Failures / Repairs
- Transports
- Waiting periods
- Depots
- Stocks
- Distances
In the following, waste caused by unnecessary stocks and transports will be discussed more precisely.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 3
Quality Management – Prof. Schmitt
Lecture 06
Stocks Cover Problems
Production
Stocks
Problems
Problems are „covered“
„ Missing material
„ Bottlenecks in manufacturing
„ Machine breakdown
„ Quality lacks
© WZL/Fraunhofer IPT
Produc
tion
Problems
Problems are uncovered
„ In time detection
„ Easy identification
„ An obligation for fast
problem solving is the result
Production
Problems
Problems
Problems are solved
„ The solution of outstanding
problems allows a production
with minimal stocks
Page 4
Stocks Cover Problems
- Oversized stocks cover problems. Missing material, bottlenecks in manufacturing, machine
breakdown, quality defects, etc. remain undetected.
- With small stocks, problems during the manufacturing process are uncovered immediately and need fast
solutions.
- These solutions enable an optimised production, which works without big stocks.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 4
Quality Management – Prof. Schmitt
Lecture 06
Stocks Prevent the Possibility for Improvement
Stocks
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
Cover problems
Cause problems
Require depots
Hamper work flow
Prevent real control of the processes
Mislead to routine and comfort
Prevent the possibility for improvement
Only produce what is needed – in the correct amount - to the correct time!
© WZL/Fraunhofer IPT
Page 5
Stocks Prevent the Possibility for Improvement
- Apart from the problem covering, discussed earlier, stocks cause additional costs, need depots, hamper
the work flow, mislead to routine and comfort and prevent real control of the processes and thus the
possibility for improvements.
- A production on demand with the KANBAN system (pull principle) only produces on demand and does
not need stocks.
- Clean and efficient processes are possible.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 5
Quality Management – Prof. Schmitt
Lecture 06
Transports lead to Costs, Damages, Search Procedures
ACTUAL
IST
TARGET
Assembly
Material station
Truck 1
Handoff 20
Handoff 24
18 20 22
24
Handoff 30
26
30
32
Truck n
Open space for further equipment
variant
Subassembly
During the new installation of an
assembly department the available
areas were completely used. The
results are long ways, temporary
buffers...
© WZL/Fraunhofer IPT
The installation of a flow production for
the subassembly leads to small area
requirements, with short ways and
small stocks...
Page 6
Transports Lead to Costs, Damages, Search Procedures
- Lean production does not imply using up the available space completely, but using is wisely and
effectively.
- Furthermore the arrangement of the stations is important. Machines that are arranged in the U-layout
e.g., allow the operator to reach every point in the process chain quickly.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 6
Quality Management – Prof. Schmitt
Lecture 06
Contents
Detection and Prevention of Waste
„ Development of Rapid Saving Potentials with the 5S
„ Efficient Modelling of the Value Stream with Value Stream
Design
„ Securing of the Process Capability with Statistical Process
Control (SPC)
© WZL/Fraunhofer IPT
Page 7
Quality Management in Manufacturing
A further starting point for increasing the added value are the 5S. Their application is described in the
following.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 7
Quality Management – Prof. Schmitt
Lecture 06
"5S" Increase Safety, Quality and Productivity
Safety
Factory
Rejections
5S as basis of production
Direct
connectivity
Productivity
Safety, Quality, Productivity
© WZL/Fraunhofer IPT
Page 8
"5S" Increase Safety, Quality and Productivity
- The 5S must form a basis for the whole production. They directly connect place safety, quality and
productivity.
- By applying the 5S consequently, the safety will be improved, rejections will be minimized and the
production will be increased.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 8
Quality Management – Prof. Schmitt
Lecture 06
Definition of 5S
Definition
5S are based on the five Japanese words:
SEIRI, SEITON, SEISO, SEIKETSU, SHITSUKE
ƒ Sorting out means taking out unnecessary items and
disposing them
ƒ Systematizing means arranging necessary items in good
order
ƒ Sweeping means cleaning the workplace
ƒ Sanitising means maintaining high standard of housekeeping
ƒ Self-discipline means doing things spontaneously without
being told
© WZL/Fraunhofer IPT
Page 9
Definition of 5S
5S are based on the five Japanese words:
SEIRI, SEITON, SEISO, SEIKETSU, SHITSUKE
They can be translated like:
- Sorting means taking out of unnecessary items and disposing them.
- Systematizing means arranging necessary items in good order.
- Sweeping means cleaning the workplace.
- Sanitising means maintaining high standard of housekeeping.
- Self-discipline means doing things spontaneously without being told.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 9
Quality Management – Prof. Schmitt
Lecture 06
What Do the 5S Do?
Aim of the method
ƒ The 5S are the starting point for every successful company. In the
Japanese understanding lack of 5S means inefficiency.
ƒ The 5S method originates from production companies, but can also
be applied to the service industry.
The aim is to eliminate everything from a working process that
ƒ complicates
ƒ delays or even
ƒ prevents
it.
© WZL/Fraunhofer IPT
Page 10
What Do the 5S Do?
- The 5S remove inefficiency from the working process, which means everything that complicates, delays
or prevents the process.
- Although this method was developed by (Japanese) production companies, it can also be applied to
service industries.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 10
Quality Management – Prof. Schmitt
Lecture 06
SEIRI – Sorting Out Unnecessary Things
SEIRI - Sorting out unnecessary things
ƒ Everything, that is not going to be used within the next days, is
removed from the workspace.
ƒ The workspace only contains things, which are needed directly for
value adding activities within the next thirty days.
Before SEIRI Many unnecessary things within the
workspace
After SEIRI Only necessary things within
the workspace
© WZL/Fraunhofer IPT
Page 11
SEIRI – Sorting Out Unnecessary Things
- All tools, documents, racks etc., which are not used within the next days, are removed from the
workspace, or even disposed if necessary.
- Thus, there are only things remaining within the workspace, which will be needed for value adding
activities during the next 30 days.
- The work space becomes leaner and cleaner.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 11
Quality Management – Prof. Schmitt
Lecture 06
SEITON - Cleaning Up the Remaining Things
SEITON - Cleaning up the remaining things
ƒ Everything that remains within the workspace is stored that way, that it is
easily accessible at any time.
ƒ Further, ways and depots have to be labelled to avoid blockings
ƒ A certain maximum capacity will be assigned to each depot.
Before
© WZL/Fraunhofer IPT
After
Page 12
SEITON - Cleaning Up the Remaining Things
- All tools, documents, racks etc., which remain within the workspace, are stored that way, that it is easily
accessible at any time.
- Further, ways and depots have to be labelled to avoid blockings.
- It is important that everything has its own spot. Each individual tool will be assigned a fixed address and
a maximum capacity will be assigned to each depot.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 12
Quality Management – Prof. Schmitt
Lecture 06
SEISO - Keeping the Workspace Clean
SEISO - Keeping the workspace clean
ƒ Everything that remains within the workspace is cleaned and
maintained constantly.
ƒ Thereby, flaws which would lead to errors during the process, are
uncovered and eliminated in advance.
ƒ The aim is to create a tidy, functional and safe work environment
© WZL/Fraunhofer IPT
Page 13
SEISO - Keeping the Workspace Clean
- After all useless things are removed and the remaining things are properly stowed away, they are
cleaned and maintained in constant intervals.
- Error potentials are supposed to be uncovered by regular maintenance to avoid defects, down-times or
damages.
- Besides reducing losses of production, the work safety will be increased.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 13
Quality Management – Prof. Schmitt
Lecture 06
SEIKETSU - Making Arrangements to Rules
SEIKETSU - Making arrangements to rules
ƒ Seiketsu means creating a clean and healthy work environment by
wearing working clothes and safety equipment, as well as daily applying
and improving Seiri, Seiton und Seiso.
ƒ With Seiketsu a once achieved level is kept and does not drop by “resting
on the laurels”
ƒ The management must determine, how often Seiri and Seiso are applied
and who participates in it.
5S
© WZL/Fraunhofer IPT
Page 14
SEIKETSU - Making Arrangements to Rules
- Seiketsu means two different things:
- To create a clean and healthy work environment by wearing working clothes and safety equipment, as
well as daily applying and improving Seiri, Seiton und Seiso.
- Only with Seiketsu a once achieved level is kept and does not drop by “resting on the laurels”
- The management must determine, how often Seiri and Seiso are applied and who participates in it.
It is important that the 5S become a part of the corporate culture and not a once used experiment.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 14
Quality Management – Prof. Schmitt
Lecture 06
SHITSUKE - Keeping and Constantly Improving All Points
SHITSUKE - Keeping and constantly improving all points
ƒ Shitsuke means self-discipline.
ƒ After the introduction and regular application of Seiri, Seiton and
Seiso, Seiketsu has to be kept.
ƒ Therefore the application of the individual components must be
measured and evaluated. The evaluation is done by employees,
superiors or external advisors.
ƒ Another possibility is, to turn the evaluation into a competition.
Seiso
Seiton
Seiketsu
Seiri
© WZL/Fraunhofer IPT
Page 15
SHITSUKE - Keeping and Constantly Improving All Points
- Shitsuke means self-discipline.
- After the introduction and regular application of Seiri, Seiton and Seiso, Seiketsu has to be kept.
Therefore the application of the individual components must be measured and evaluated.
- There are different possibilities, for example an evaluation by employees, superiors or external advisors.
Another possibility is, to turn the evaluation into a competition.
- The management has to pay attention, that the acquired self-discipline keeps its level
A regular evaluation helps that the acquired self-discipline keeps its level and the company does not return
to the conditions before the application of the 5S.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 15
Quality Management – Prof. Schmitt
Lecture 06
Applying Phases of the 5S – Expanded by a Sixth S
Seiri
(Sorting out)
Seiton
(Systematizing)
Seiketsu
(Sanitising)
Shitsuke
(Self-discipline)
© WZL/Fraunhofer IPT
Seiso
(Sweeping)
Phase 1
Phase 2
Shukan
(Familiarisation)
Phase 3
Page 16
Applying Phases of the 5S – Expanded by a Sixth S
- The 5S are extended by Shukan (familiarisation) as the sixth S.
- In phase 1 the steps Seiri (sortingout), Seiton (systematizing) and Seiso (sweeping) are implemented.
- The 2. phase is Seiketsu (sanitising).
- Phase 3 consists of Shitsuke (self-discipline) und Shukan (familiarisation). The rules and standards of
the 6S have to be assimilated.
- Shukan takes care, that the rules and standards of the 6S are assimilated and become an integral part
of the company culture.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 16
Quality Management – Prof. Schmitt
Lecture 06
Contents
Detection and Prevention of Waste
Development of Rapid Saving Potentials with the 5S
„ Efficient Modelling of the Value Stream with Value Stream
Design
„ Securing of the Process Capability with Statistical Process
Control (SPC)
© WZL/Fraunhofer IPT
Page 17
Quality Management in Manufacturing
Elementary control levers within the production are the value streams. They can be analyzed and modeled
with the Value Stream Design. This method is described in the following.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 17
Quality Management – Prof. Schmitt
Lecture 06
Value Stream Design
Definition
ƒ With „Value Stream Design“ the value stream is illustrated fast and
easily.
ƒ It enables an intervention which is more rational than it is with single
process improvements.
ƒ Designing a plan for the whole value stream is missing in many
improvement efforts.
Aim of the method
ƒ The Value Stream Design allows the illustration of value streams.
Thereby the flow of material and information is illustrated.
ƒ With a defined symbolism, an easy understandable illustration with
focus on the weaknesses of the value stream is achieved.
ƒ Theses bottlenecks can be eliminated specifically and the whole
process can be redesigned.
© WZL/Fraunhofer IPT
Page 18
Value Stream Design
A value stream describes all activities, that are necessary to bring a product through the main streams:
1. The manufacturing stream from the raw material into the hands of the customer
2. The development stream from the concept of a product through to the start of its sale.
- To take a value stream perspective means to work on the overall picture, not only on single processes.
Simplified Value Stream Design means the following:
- Go along the whole production way of a product and create a rough image of every process in the
material and information flow.
- Afterwards ask some key questions and draw a target state, which is an image of the value stream,
how it should look like in future.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 18
Quality Management – Prof. Schmitt
Lecture 06
Example of Use: Transparency in the Illustration
Corporation:
15t Stahlcoil
Termin:_____________
_
90D/60D/30D
-preview
Production planning
6 –weeks
preview
Daily
fax
Weekly
fax
Weekly
planning
Assembly plant
750p./day
Container: 20p.
Daily
planning
Tue. and Thu.
Termin:_____________
_
1x daily
B
B
5 days
5D
2600 p.
Blanking
CT = 1 Sek.
PT = 1 Std.
U = 85%
3,5D
1 sec.
© WZL/Fraunhofer IPT
B
B
Welding 1
CT = 40 Sek.
PT = 10 min.
U = 95%
Cap. = 55200s
40 sec.
1000 p.
1,2D
Assembly 1
CT = 62 Sek.
PT = 0 min.
U = 100%
62 sec.
Shipment
1000 p.
1,2D
LT = 10,9 T
PT =103sec
Page 19
Example of Use: Transparency in the Illustration
- Value Stream Design allows drawing an image of the actual state of a value stream.
- In Value Stream Design single processes like “blanking” or “welding” are wrapped up into process
categories.
- Processes and flows are illustrated with a group of symbols.
- As soon as the whole stream of the manufacturing facility becomes visible, you can increase or
decrease the zoom factor, so you can either examine one single process or take a look at the supply
hain beyond your facility.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 19
Quality Management – Prof. Schmitt
Lecture 06
Example of Use from a Workshop
„ Value stream illustration
– Visual illustration of the
material and information
streams
© WZL/Fraunhofer IPT
Page 20
Example of Use from a Workshop
This slide shows an example of Value Stream Design in a workshop.
Material and information streams are clearly illustrated and allow the development of improvement
potentials.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 20
Quality Management – Prof. Schmitt
Lecture 06
Result from Regular Workshops
before
after
„ Specific application: 2 container system, smaller containers
© WZL/Fraunhofer IPT
Page 21
Result from Regular Workshops
The results of regular workshops are clearly visible:
- tidy,
- clearly defined stock piles
This way, a clear more effective working environment has been created.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 21
Quality Management – Prof. Schmitt
Lecture 06
Contents
Detection and Prevention of Waste
Development of Rapid Saving Potentials with the 5S
Efficient Modelling of the Value Stream with Value Stream
Design
„ Securing of the Process Capability with Statistical Process
Control (SPC)
© WZL/Fraunhofer IPT
Page 22
Quality Management in Manufacturing
Finally in this lecture unit the Statistical Process Control is regarded. The procedure will be discussed in a
practical example.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 22
Quality Management – Prof. Schmitt
Lecture 06
Process Capability
cp = 1
cpk = 1
cp = 1,33
cpk = 1,33
cp = 1,33
cpk < 1
© WZL/Fraunhofer IPT
Page 23
Process Capability
- The meaning of the cp- und cpk-values can be demonstrated by a truck driving through a gate. In this
example the walls of the gate indicate the range of tolerance and the middle of the bottom of the truck
represents the mean value of the measured values. The width of the truck represents the variation of
the measured values.
- If the cp- und cpk -values are equal the truck is driving in the middle of the gate.
- If cp und cpk =1, the truck is able to drive through the gate theoretically; Practically, it has to be assumed
that the truck is going to hit the wall and gets damaged in case of a small impact (like a road hole in this
example).
- The bigger the values get, the smaller is the variation. Thus, the truck is smaller and, therefore, is able
to react on outer impacts without hitting the walls immediately.
- In case, at least one of the values is smaller than 1, the truck would bump into the wall of the gate. For
the process, however, this means, that several values lie outside the range of tolerance.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 23
Quality Management – Prof. Schmitt
Lecture 06
Capability Indices
22 -21 --
3
>
UL 20 --
19
18
17
16
NV 15
14
13
12
11
LL 10
9
8
-------------
s
X
X
X
X
cp = 0.71
cpk = 0.71
X
cp = 2.50
cpk = 1.00
-
LL
cp = 1.67
cpk = 1.33
cp = 5.00
cpk = 5.00
Capability indices
cp
cpk
=
=
Tolerance range
Process variation
=
Minimum distance to tolerance limit
Half process variation
>
legend:
s : Mean standard deviation of the samples
σ : Estimation of variation of the population
x : Mean of means of the samples
>
© WZL/Fraunhofer IPT
>
=
UL - LL
2• (3 σ )
| x - LV| min
3σ
UL
LL
LV
NV
:
:
:
:
Upper limit
Lower limit
Limiting value
Nominal value
Page 24
Capability Indices
- The capability of a process can be described using the so-called capability indices cp and cpk.
- The cp index is a measure of the extent of the process variation.
- The cpk index additionally takes account of the position of the variation.
- In general, the process is regarded as capable when both indices are greater than 1. If cp is greater than
cpk, the process can be improved by being centred.
- The minimum demand for the process capability indices are: cp and cpk > 1. Today the demand cp and
cpk > 1,33 33 is more common.
- In short-term-examinations, the short-term process capability indices pp and ppk are applied, which are
calculated just like cp and cpk. (In this case the demand pp and ppk > 1,67 67 is commo).
- The correlation between the characteristic of the variation and the capability indices is illustrated with
some couple of examples.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 24
Quality Management – Prof. Schmitt
Lecture 06
Procedure for Determining Process Capability
Elimination of systematic
influences
Correction
Basics for calculation
(normal distributed)
Preparatory run
UL - LL
6σ
zkrit.
cpk=
3
cp=
Test for normal
distribution
Construction of
probability net
x and s
x and s
Calculation
Graphic
Calculation of capability
indices cp and cpk
cp 1
cpk 1
no
cp 1
cpk<1
yes
yes
Centering of process
Potential for
improvement?
Process
capability
Arbeitsplan
no
Process is
not capable
Process is capable
© WZL/Fraunhofer IPT
σ
yes
no
Analysis
zkrit.= Min
=
UL - x
σ
=
x - LL
Drehen
σ= s
c4
σ= R
d2
n
d2
c4
2
3
4
5
1,128
1,693
2,050
2,326
0,7979
0,8862
0,9213
0,9400
.
.
.
.
.
.
.
.
.
Page 25
Procedure for Determining Process Capability
- The general way, how to run capability analysis is illustrated above.
- After analyzing the process and eliminating systematic and special disturbing influences a preparatory
run is accomplished in order to acquire the character of the distribution and to calculate the mean value
and the variation. On the basis of these parameters the capability indices cp and cpk are calculated.
- The capability of the process is a requirement for using control chart techniques.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 25
Quality Management – Prof. Schmitt
Lecture 06
Statistical Process Control and Statistical Quality Surveillance
Statistical Process Control
Statistical Quality Surveillance
Process
Process
Batch
n.o.k.
Inspection
Inspection
o.k.
Inspector
decision
Operator
Key
Material flow
Information flow
Control chart
Histogram
UCL
x
LCL
08:00 09:00 10:00
11:00
08:00 09:00 10:00
11:00
UCL
R
LCL
© WZL/Fraunhofer IPT
LL
UL
Page 26
Statistical Process Control and Statistical Quality Surveillance
-
-
The aim is, to minimise inspection efforts during and after the manufacturing process. When tests are
unavoidable, it makes sense to ensure that deviations are detected as early as possible, so that the
manufacturing process can be adjusted before defective parts are produced.
Therefore, statistical methods such as Statistical Process Control (SPC) or Statistical Quality
Surveillance are increasingly used in order to assure product and process quality.
Due to SPC, a machine-oriented quality control loop is closed, which is supposed to detect and correct
failures.
The control chart, which serves to detect systematic and special process influences at an early stage,
is a useful tool. Using control charts requires the knowledge of the natural process variation.
A further application of statistics is the determination of the test decision for a production lot by
evaluating the characteristic values x and s or on the basis of histograms. This kind of inspection is
called Statistical Quality Surveillance (SQS).
© WZL/IPT
Quality Management in manufacturing
L 06 Page 26
Quality Management – Prof. Schmitt
Lecture 06
Stable and Unstable Process
Unstable process
Stable process
8:00
10:00
9:00
11:00
•••
• •
••
••
••
•
• •
8:00
9:00
••
•
•
10:00
11:00
••
••
••
•
© WZL/Fraunhofer IPT
•
••
• •
•
• •
•
• •
••
• •
Page 27
Stable and Unstable Process
- Each of the control charts, shown on the slide, reflects the values from samples containing four parts,
taken from the process at hourly intervals. The boundary lines shown, represent the random variation
range of the process that is determined in a preliminary investigation. The centre line is the mean value.
- A stable (undisturbed) process is only subject to natural variation.
- The appearance of special influences (non-natural alteration of the process-state) leads to an unstable
process.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 27
Quality Management – Prof. Schmitt
Lecture 06
Implementation of Statistical Process Control (SPC)
Preparatory run
Analyse
process and
eliminate
failures
Determination of control limits
Calculate and plot
control limits
Keep control charts
Measure
samples
Measure
sample
yes
Further
samples
no
Determination
of capability
Mark values
with systematic
influences and
calculate the control
limits again,excluding these factors
Calculate
and plot
characteristics
Determine
failure cause
and eliminate
failure
Systematical failure
no
yes
no
Capable
yes
© WZL/Fraunhofer IPT
Page 28
Implementation of Statistical Process Control (SPC)
Process control consists of the phases:
- Data collection
- Data analysis
- Control
Requirements for using control charts are:
- Determination of the process capability by a preparatory run (usually 10 samples each containing 5
parts).
- Definition of the control limits, which describe the area of the natural process variation.
- Samples are removed from the ongoing manufacturing operation at specified intervals.
- The characteristic statistical values are identified and entered in the charts for location and variation.
- Then the graphs are analysed in order to detect special (disturbing) influences.
- When special disturbing influences appear in the control chart, the process must be stopped and
investigated. Once the disturbing influence has been eliminated, the control limits are calculated again
and the process will be restarted (control loop).
© WZL/IPT
Quality Management in manufacturing
L 06 Page 28
Quality Management – Prof. Schmitt
Lecture 06
Preparatory Run
Process analysis
material
person
• Determination of disturbance variables
method
feature
• Determination of setting parameters
environment
process
Conduct preparatory run
• Sampling (e.g. 50 parts)
• Analysis of variation and mean
_ 1n
x = ∑ xi
n i=1
s=
1
_
n
∑ ⎛⎜ xi − x ⎞⎟
⎠
2
n − 1 i=1 ⎝
LL
UL
Capability analysis
Determination of capability indices pp and ppk
(short-time examination)
respectively. cp and cpk
Key :
s : Standard deviation
σ : Estimation of variation of the population
x : Mean of means of the samples
pp, ppk, cp, cpk : Capability indicators
© WZL/Fraunhofer IPT
UL - LL
2•(3 σ )
pp, cp =
ppk, cpk =
| x - LL/UL|min
3σ
UL
LL
LV
x
:
:
:
:
Upper limit
Lower limit
Limiting value
Mean
Page 29
Preparatory Run
The preparatory phase consists of 3 steps:
- Process analysis
- Conduction of the preparatory run
- Capability analysis
Objective of the preparatory run is to check the suitability of the process for SPC and to acquire the process
capability.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 29
Quality Management – Prof. Schmitt
Lecture 06
Determination of Control Limits for Mean/Range Charts
x
UCL X
Range
R = xmax- xmin
x
LCL
X
R
Time
x1 + x2 + ... + xm
m
R=
R1 + R2 + ... + Rm
m
Process mean
UCL R
R
LCL
x=
R
Time
Control limit
mean
UCL x = x + A2 * R
Control limit
range
UCLR = D4 * R
LCL x = x - A2 * R
LCLR = D3 * R
© WZL/Fraunhofer IPT
Page 30
Determination of Control Limits for Mean/Range Charts
- The individual types of control charts are subject to different calculation formulas.
- In the figure these calculation formulas are exemplarily shown for the mean value/range chart.
© WZL/IPT
Quality Management in manufacturing
L 06 Page 30
Quality Management – Prof. Schmitt
Lecture 06
Test Procedure for Non-Random Progressions in the Control Chart
Standard tests
More than 7 values
on one side of
mean (run)
Exceeding control
limit
More than 7 values
with the same direction
of slope (trend)
Additional tests
3σ
2σ
1σ
2 of 3 values on
the same side
more than 2σ
away from mean
© WZL/Fraunhofer IPT
3σ
2σ
1σ
4 of 5 values on
the same side and
more than 1σ
away from mean
Page 31
Test Procedure for Non-Random Progressions in the Control Chart
- There are a number of test criteria that may be applied to detect the occurrence of non-random events in
the control charts; some of them are exemplarily shown in the figure.
- Over 30 other test techniques are known in addition to those described. The situation is similar to
“Russian Roulette”: the more bullets loaded in the revolver, the greater the chance that a shot will be
fired. Therefore, it is advisable to limit the number of techniques used, to those which are suitable
for the task and to investigate very thoroughly the events which are revealed!
© WZL/IPT
Quality Management in manufacturing
L 06 Page 31
Quality Management – Prof. Schmitt
Lecture 06
Examples of Different Types of Control Charts
Mean
Mean
Original value
X
X
Standard deviation
X
Range
Moving range
R
S
Median
R1
Median / original value
Cumulated sums
X
III
II
Range
X, X 1
II
II
III
R
© WZL/Fraunhofer IPT
I
II
III
Cusum
Page 32
Examples of Different Types of Control Charts
- Besides the mean value/range chart there are a number of different types of control charts.
- Beyond control charts for variable test characteristics (measurable) there are also various types of
control charts for attributive test characteristics (countable or good/poor).
© WZL/IPT
Quality Management in manufacturing
L 06 Page 32
Quality Management – Prof. Schmitt
Lecture 06
Process Control and Monitoring Methods
SPC CPC
Statistical Process Control
Regulative
time discrete sampling
Continous Process Control
Regulative
continuous sampling
Process
Process
Inspection
Inspection
Controlhardware
Operator
Process
o.k.
Process
Key:
o.k.
= Flow of material
= Flow of
information
Inspection
robot
n.o.k.
n.o.k.
Inspection
Supervisory
time discrete sampling
Statistical Quality Surveillance
© WZL/Fraunhofer IPT
Decision
SQS CQS
Cupervisory
continuous sampling
Continuous Quality Surveillance
Page 33
Process Control and Monitoring Methods
Besides the Statistical Process Control (SPC) the following further methods exist for process control and
surveillance:
- Continuous Process Control (automated 100%-inspection with regulating intervention via control
electronics).
- Statistical Quality Surveillance (statistical quality inspection without regulating process intervention).
- Continuous Quality Surveillance (automated 100%-inspection without regulating process intervention).
© WZL/IPT
Quality Management in manufacturing
L 06 Page 33