Page 1

TECHNICAL UNIVERSITY BRNO

BRNO UNIVERSITY OF TECHNOLOGY

FACULTY OF MECHANICAL ENGINEERING

DEPARTMENT OF METROLOGY AND TESTING

FACULTY OF MECHANICAL ENGINEERING

INSTITUTE OF METROLOGY AND QUALITY ASSURANCE TESTING

APPLICATION OF SPC METHODS TO THE PRODUCTION PROCESS

APPLICATION OF SPC METHOD TO OPERATION CONDITION

THESIS
DIPLOMA THESIS

AUTHOR OF THE WORK Ing. PETR KOPL

AUTHOR

HEAD OF WORK doc. Ing. ALOIS FIALA, CSc.

SUPERVISOR

BRNO 2008

Page 2

Brno University of Technology, Faculty of Mechanical Engineering

 Department of Metrology and Testing
Academic year: 2007/08

ASSIGNMENT OF DIPLOMA THESIS

student: Kopl Petr, Ing.

who is studying in a master's degree program

field: Quality Management (3911T023)

The director of the institute, in accordance with Act No. 111/1998 on Higher Education Institutions and the Study and
The examination regulations of BUT determine the following topic of the diploma thesis:

APPLICATION OF SPC METHODS TO THE PRODUCTION PROCESS

in English language:

APPLICATION SPC METHOD TO OPERATION CONDITIONS

Brief characteristics of the problem:

1. Defining work goals.

2. Problem analysis and literature search.
3. Proposed solution.
4. Design verification.
5. Discussion of results.
6. Conclusions and recommendations.

Objectives of the diploma thesis:

Process analysis and determination of suitable performance indicators.

Design of usable tools.
Efficiency verification.

Page 3

List of professional literature:

ČSN ISO 8258: 1994 Shewhart control diagrams

Statistical Process Regulation (SPC), 2nd edition 2005 Czech Society for Quality, Prague, 2006
(Czech edition)
 Thesis supervisor: doc. Ing. Alois Fiala, CSc.

The deadline for submitting the diploma thesis is set by the time schedule of the academic year 2007/08.

Brno, 30.11.2007

LS

___________________________ _______________________________
Doc. Ing. Leoš Bumbálek, Ph.D. Doc. RNDr. Miroslav Doupovec, CSc.
Director of the Institute Dean of the Faculty

Page 4
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

ANNOTATION

Ing. Petr Kopl

Application of the SPC method to the production process
Diploma thesis, Department of Metrology and Testing, BUT FME

Principles of the SPC system. Dispersion influences affecting the production process. SPC application
into the production process. Design of SPC methodology for spindle production line. Verification of the SPC method
in operating conditions. Processing and interpretation of control cards. Summary of results
and proposed measures.

Keywords: Statistical process control, control diagram, mean value of range,

mean value of the process, limits of intervention.

ANNOTATION

Ing. Petr Kopl

Application SPC method to operation conditions
Diploma dissertation, Institute of metrology and quality assurance testing, Brno University
of technology Faculty of mechanical engineering

Principles of SPC system. Spreading effects influencing the production process. Application
of SPC to operation conditions. Proposal of SPC methodology for a spindle production line.
Testing SPC methodics in operative conditions. Processing and interpretation of regulation
cards. Total of results and suggestioning measure.

Key words: Statistic regulation of the process, regulation diagram, average value of the range,
average value of the process, limit sof interference.
 Brno, 2008 -3- Ing. Petr Kopl

Page 5
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

BIBLIOGRAPHIC CITATIONS ACCORDING TO ISO 690

KOPL, P. APPLICATION OF SPC METHODS INTO PRODUCTION PROCESS. Brno: University
Technical University in Brno, Faculty of Mechanical Engineering, 2008. 74 pp. Thesis supervisor doc. Ing.
Alois Fiala, CSc.
 Brno, 2008 -4- Ing. Petr Kopl

Page 6
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

DECLARATION

I declare that I have a diploma thesis on the application of the SPC method to production
process developed independently using the literature and sources listed
on the list that accompanies this work.

Date ………………………………….
Signature of the graduate

Brno, 2008 -5- Ing. Petr Kopl

Page 7
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management
 UNDER THE HARDWARE

Thank you Doc. Ing. Alois Fiala, CSc. for his valuable comments and advice. Further
I would like to thank the employees of Jihomoravská armaturka spol. s ro for providing
information for the processing of this diploma thesis.

Brno, 2008 -6- Ing. Petr Kopl

Page 8
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

CONTENT
Title page
Diploma thesis assignment
Annotation ………………………………………………………… .. …………………………… 3
Bibliographic citation according to ISO 690 …………. …………………. …………………………… .4
Declaration … .. …………………………………………………………………………… ...… 5
Pod of forging ......... ... .................................................................................... .6
Contents ………………………………………………………………………………………… .7
Introduction ………………………………………………………………………………………… ... 9
1. Statistical process control …………………………………………. …………. ………………………………………… 11
1.1 Prevention versus detection …………………………………………………… ... 11
1.2 Process control system …………………………………………………… ..11
1.3. Fluctuations: accidental and special causes ………………………………………… 12
1.4 Local and system measures …………………………………… ..… .14
1.5 Process control and process capability .15 ..15
2. Regula No the diagrams ........................ .. ............................... .................. 21
2.1 Control diagrams: tools for process control 21 .21
 2.2 Effective use and benefits of control charts .2. ………… ..23
2.3 Control diagrams during control by measuring ………………………………… ... 24
2.4 Elements of control diagrams 24 .24
2.5 Processes controlled by control diagrams ……………………………………… ..28
2.5.1 regulatory mechanism No. them diagram of ................................. .. ...... 29
2.5.2 Determination of regulatory No them between ................................................... .31
2.5.3 Interpretation of statistical regulation …………… ... ……………………… ..31
2.6 Determination of signals of statistically uncontrollable process …………………………… 33
2.6.1 Point outside regula No limits ................................................... ... 33
2.6.2 Non-random clusters or trends inside of regulatory No. them between ............ ... 33
2.6.3 Special criteria for any special of í No. ins ... ... ................................. ..34
2.7 Formulas for control diagrams ……………………………………………… 35
2.8 Testing of process stability, assumption of application of control diagrams 37 .37
3. Sou No Current status checks of ethene in JMA ................................................... ..38
3.1 Spindle production line …………………………………… ... ……………… ..38
3.2 Inspection of spindles at the workplace SP 12 CNC …………………………………… 40
4. Draft SPC methodology …………………………………………………. ………………………………………………… 41
4.1 Choice of representative of manufactured spindles …………………………………… .... 41
4.1.1 Extract from the technological process for the row spindle ............... ... ............ ..43
4.1.2 Existing m measured ICI as of devices ............................................. ... ...... ... 44
4.1.3 Draft m measured -enhancing and monitoring for of equipment ........................... ... ... ..44
4.2 Selection of control level ……………………………………………………… ..44
4.2.1 Selection of R-type regulatory No. whom you diagram ..................................... ...... ..45
4.2.2 Management regulatory No. whom you diagram ( X , R) .................................... .. ... 45
5. Implementation of the SPC method into the production process ............................................ ..... 46
5.1 Specification of process conditions ……………… ... …………………………… ..46
5.2 Calculation of values for the first control diagram - 1. Implementation ……… ...… 46
5.3 Calculation of values for the second control diagram - 2. Implementation… ....… ...… 53
5.4 Evaluation of control diagrams ………………………………………… ... 57
6. Economic evaluation ................................................ ........................................... 58
7. Conclusion of r ..................................................................... ... ..................... .59
List of used sources ………………………………………………………………… ..60

Brno, 2008 -7- Ing. Petr Kopl

Page 9
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

List of abbreviations and symbols used ………………………………………………… ..… 61

List p of attachments ....................................................................................... ..... 62
 Brno, 2008 -8- Ing. Petr Kopl

Page 10
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

INTRODUCTION
To succeed in today's economic climate requires all manufacturers, suppliers
and sales organizations are dedicated to continuous improvement. More efficient ways still need to be sought
for the production of products and services. Products and services must continue to increase their
values. Producers must constantly focus on customers, both domestic and foreign,
and customer satisfaction must be considered a primary business goal. To meet this goal
everyone in organizations must focus on improving and using more effective methods.
If the view is taken that a company that wants to compete must constantly strive
about increasing quality while reducing costs, then naturally the question arises
how these objectives are to be achieved. Specifically, at what stage of the production process and at what stage
resources. Are there any universal methods to achieve these goals?

The core of the traditional method of quality assurance is the quality control of process outputs
and sorting units that do not meet specifications. This quality assurance strategy
it is uneconomical as it is a form of "ex post" control, where the resources needed to create it
of the required product have already been spent and, in the case of a non-compliant product
were spent unnecessarily. Modern approaches to quality assurance are built
to avoid wasting resources. This prevention can be provided along the way
constantly obtaining information about the behavior of the process and their analysis in order to influence the process
so that the outputs have the required properties and functions.
Statistical Process Control (SPC) represents
preventive tool of quality management, because on the basis of early detection of significant
deviations in the process from a predetermined level allows to implement interventions in the process
in order to keep it at an acceptable and stable level in the long term, or to enable the process
to improve.

This work deals with the application of the SPC method to the production process. It is processed
for Jihomoravská armaturka spol. s ro The study is specifically focused on the production process
spindles of water fittings. The work is focused on one typical representative of spindles
- spindle of the closing flange slide valve EKO - Plus PN 16; DN 40. Theoretical part
the study deals with the area of statistical process control. It sets out the basic approaches
and procedures for applying the SPC methodology. The practical section includes a description of the current state
performing spindle quality control and designing the SPC method. For SPC verification they are
Shewhart control diagrams implemented in operating conditions. Based
conclusions and recommendations are made of the evaluated diagrams.

Basic information about Jihomoravská armaturce spol. s ro

The company Jihomoravská armaturka is a manufacturer of important elements in infrastructure
water supply. It belongs to the top in its field.
JMA has more than one hundred and twenty years of production tradition. The company was founded in 1881 by Mr.
Karel Kudrna under the name "Karel Kudrna Foundry and Engineering". The current name was created
in 1950. After 1990, a partner for privatization was sought. The result was a framework
privatization of the company into several companies and their subsequent entry in 1996 into a significant one
German joint-stock companies Bopp and Reuther AG Mannheim. In 2002, Czech co-owners
they leave. The German company VAG-Armaturen GmbH merges with Bopp and Reuther AG
to VAG Armaturen Gmbh. It is also a subsidiary of VAG Armaturen GmbH
JMA also included.
Fittings for water mains have become the main thing for JMA in recent years
 subject of business. The result is products that rely on many years of experience

Brno, 2008 -9- Ing. Petr Kopl

Page 11
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

high durability and reliability. The JMA offer includes more than one hundred product types. After 125
The largest Czech manufacturer of water fittings is based on the foundations of a once small company.
JMA has the most modern foundry of gray and ductile iron in the Czech Republic, advanced
technologies for machining parts. It also handles heavy surface treatment technology
anti-corrosion treatments according to GSK.
For the last fifteen years, this company has been making leap miles among renowned European companies
valve manufacturer.

Brno, 2008 - 10 - Ing. Petr Kopl

Page 12
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

1. STATISTICAL REGULATION OF THE PROCESS (1,2,3,4)

The essence of statistical quality control lies in the fact that the parameters of the examined file
 products are assessed according to the selection characteristics determined from a particular selection of this
file. Statistical regulation is the application of statistical quality control in operational control,
its findings serve as information for the management of the production process.

1.1 Prevention versus detection

In the past, the creation of a marketable product depended on the production of such a product
effective quality management to have this product inspected and discarded
those products that did not meet the specifications. The case includes a detection strategy that
it is uneconomical because it allows time and material to be spent on producing such
products that are not always usable. It is far more effective to prevent scrap by
that in the first place no unusable output will be produced - a prevention strategy.

1.2 Process control system

The process control system can be described as a feedback system. The SPC is one such
type of feedback system.

Process
The process is a complete combination of suppliers, manufacturers, people, equipment, input
material, methods and environment, ie all the elements that together contribute to the output
from the process, as well as customers who use the output (see Figure 1.1). Overall process performance
it depends on the communication between the supplier and the customer, on the way the process is designed
and implemented, as well as how it works and how it is managed. Process control system
it is only useful if it contributes to maintaining the perfect level of the process
or to improve the overall performance of the process.

Performance information
Much information about the actual performance of the process can be obtained by studying the outputs
process. However, the most useful information about process performance comes from perfect knowledge
own process and its internal variability. Process characteristics (such as temperature, time
cycle, number of interruptions,…) should be the end point of the effort. The target must be determined
values for those features that contribute to the most productive running of the process,
and then monitor how close or far it is from these targets. If he gets
with such information and interpreted correctly, it may be shown whether the process is working normally
or in an unusual way. If necessary, then your own activities can be focused on correction
process or to the output just produced. If action is needed, it must be timely
and reasonable, otherwise the effect of the information obtained is wasted.

The abbot of learning process

Measures in the process are often the most cost-effective if they are oriented to work
preventively and prevent important features (process or output) from being too large
fluctuations and deviations from the required target values. This guarantees the stability of the process
and process output fluctuations are kept within acceptable limits. Such interventions include:
• Changes in operations - eg operator training, changes in input material.
• Changes in the basic elements of the process itself - eg equipment, people communication
and their relationships, design of the process as a whole.

Brno, 2008 - 11 - Ing. Petr Kopl

Page 13
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

If necessary, the impacts of the measures should be monitored together with further analysis
and the action taken.

The abbot of on output

Measures at the output of a process are usually the least economical if they are limited only
to detect and repair a product that does not meet the specifications without resolving the issue
in process. Unfortunately, if the output of the process does not meet the customer's requirements, it is necessary to sort
all products and any nonconforming either "sweep" or repair. In this activity must
continue until corrective action has been taken and verified in the process
its effectiveness. Measures only at the exit should be used strictly only as temporary
measures in the event of an unstable or ineligible process.

FEEDBACK MODEL OF THE PROCESS CONTROL SYSTEM

VOICE OF THE PROCESS STATISTIC METHODS

 PEOPLE
DEVICE THE WAY OUR PRODUCTS OR SERVICES
MATERIAL WORK /
METHODS MIXING CUSTOMERS
MEASUREMENT
RESOURCES
ENVIRONMENT

IDENTIFICATION OF CHANGING
INPUTS PROCESS / SYSTEM OUTPUTS NEEDS AND
EXPECTATION

CUSTOMER VOICE

Figure 1.1 Process control system

1.3 Fluctuation: random and odd p of s No. ins

No two products or two process values are exactly the same, because for each
The process causes many sources of fluctuation. The differences between the products can be large or they can
be immeasurably small, but they always exist. For example, the diameter of the manufactured shaft should be sensitive
respond to any fluctuations caused by the machine (play, bearing wear), tool
(strength, wear rate), material (diameter, hardness), maintenance (lubrication) and environment
(temperature, mains voltage stability) and the measurement system. Some sources cause fluctuations in the process
short-term differences from one piece to another, eg machine idling, machine storage,
component clamping. Other sources of fluctuation cause a change in the output only in the longer
time interval. Thus, the time period and conditions under which the measurements are performed are
critical because they indirectly affect the magnitude of the overall fluctuation that is observed.

Brno, 2008 - 12 - Ing. Petr Kopl

Page 14
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

While the individual measured values may all differ from each other, as a whole
they try to create a certain shape, which can be described as a division (see Figure 2.1). This
the division can be characterized by:
• position parameter (usually the mean value),
• the scattering parameter (usually the range, ie the difference between the largest and smallest
value),
• shape parameters (symmetry or skewness).
In terms of minimum requirements, the problem of fluctuations is often simplified: components
inside the tolerance field are acceptable, components outside this field are unacceptable. However,
should aim of lo be to maintain the position of the process target value of p of the minimum variability of.
To handle any process and reduce fluctuations, this fluctuation should be traced
to its sources. The first step that should be taken is to distinguish between random and special
causes of fluctuations.
Random of í No. ins fluctuations related to multiple sources of variation that are always active
on the process. The result of their action is a stable and recurring distribution over time
the probability of the monitored quality characteristic. This state of the process is called “statistically
managed state of the process ", less aptly" process under statistical control ", or briefly
"Process under control". Random causes act as a stable system of random causes.
If only random causes are present and do not change, the output of the process is predictable.
Special p of s No. ins (also definable p of s No. ins ) are related to each of the factors
which causes fluctuations affecting only some products from the process outputs. These causes
they usually act intermittently and are unpredictable. Presence of special causes
is signaled by one or more points outside the control limits or non-random
by grouping points within regulatory limits. Until all the special causes of fluctuations
identified, may continue to influence the output of the process unpredictably
ways. If these causes are present in the process, the process is unstable over time.
Changes in the distribution of the characteristic characterizing the process caused by special causes,
they can be either harmful or beneficial. The harmful ones must be identified and removed.
The beneficial ones should also be identified and should become a permanent part of the process.
 Brno, 2008 - 13 - Ing. Petr Kopl

Page 15
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Figure 2.1 Fluctuations: random and special causes

1.4 Local measures of learning and measures of learning in the system

There is an important link between the types of sources of fluctuation (Chapter 1.3) and the types of measures
needed to reduce their fluctuations.
Simple methods of statistical control can detect special causes of fluctuations.
It is usually the responsibility to identify the specific cause of fluctuations and take appropriate action
 Brno, 2008 - 14 - Ing. Petr Kopl

Page 16
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

who is in direct contact with the operation. Although management may be involved from time to time
until the conditions are remedied, resolving such a cause of fluctuation usually requires local action,
ie the participation of people who are directly linked to the production operation.
The same simple statistical methods can also show the range of random causes
fluctuations, but a more detailed analysis is needed to isolate these causes. Correction of these non-random ones
The causes of fluctuations are usually the responsibility of management. In general, the solution
problems associated with accidental causes of fluctuations usually require action in the system.
Only a relatively small proportion of excessive process fluctuations - industry experience
states 15% - it can be solved by local measures by workers directly connected with the production operation.
The majority - the remaining 85% - can be managed only by management measures in the system.

5.1 Regulation of the process and method of Eligibility Process

A process is in a "statistically controlled state" if the only sources are process fluctuations
only random causes. One of the functions of the process control system is then to issue a statistical
signal if special causes of fluctuations are present and to avoid false signals,
when special causes are not present. This allows appropriate action (s) to be taken
against these special causes (either to eliminate them or, if they are useful, to make them permanent).
The process control system can be used as a one-time device
for process evaluation. However, its real benefits only become tangible when
will become a permanent identification tool instead of a conformity assessment tool (distinction
products to good and bad, characterizing the process to stable and unstable
or eligible and ineligible, etc.).

Process control versus way of Eligibility

When it comes to the suitability of a process, two somewhat contradictory ones must be considered
terms:
• process capability and
• process performance.
Zp of Eligibility process is determined by the fluctuation, which is caused only by random
causes. This generally represents the best performance of the process itself. This condition manifests itself,
when the process is operating in a statistically controlled state regardless of the requirements of the specification.
However, customers, both external and internal, focus more on process performance , which is
the overall output of the process and how this process corresponds to their defined requirements
technical specification regardless of process fluctuations.
In general, a process in a statistically mastered state can be described as predictable
by dividing the monitored quality characteristic and the proportion of components meeting the technical requirements
The specification can be estimated from this division. If the process remains statistically
state and there is no change in position, dispersion or shape, it continues
the process of creating the same division of components that meet the requirements of the technical specification.
Once the process is in a statistically mastered state, the first action to be performed
is the setting of the process position to the target value. If the distraction of the process is unacceptable,
this strategy allows to minimize the number of manufactured components that do not meet the specifications.
Measures in the system (to reduce fluctuations caused only by accidental causes)
are required to improve the ability of the process (and its output) to comply with the specification.
First of all, the process must be brought to a statistically mastered state by being
detected and subsequently eliminated by measures all special causes of fluctuations.
Then the performance of the process is predictable and its ability to perform can be estimated
customer requirements given by the specification. This is the basis for continuous process improvement.

Brno, 2008 - 15 - Ing. Petr Kopl

Page 17
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Each process can be classified in terms of its capability and state of mastery. Any
the process can be classified into one of the four classes shown in the following table:

Table 1.1 Eligibility and status of process management

 Statistical regulation
Status statistically Status statistically
mastered unmanaged
P of case 1
P R acceptable salt thereof P of case 3
Zp of Eligibility
P of case 2
Nep of acceptable salt thereof P of case 4

For a process to be acceptable, it must be in a statistically controlled state and its fluctuations induced
random causes must occur in a narrower field than the tolerance field given by the specification.
The ideal situation occurs in Case 1, when the process is statistically mastered and its ability to perform
requirements specification is acceptable. In Case 2, although the process is statistically mastered,
but shows too much fluctuation due to random causes that must be reduced.
The process in Case 3 meets the requirements of the technical specification, but the process is not
statistically mastered; special causes need to be identified and eliminated. In Case 4, the process
it is neither statistically mastered nor acceptable in terms of eligibility. Both random and strange
the causes of fluctuations must be reduced.

Brno, 2008 - 16 - Ing. Petr Kopl

Page 18
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management
 Figure 3.1 Process control and process capability

Process indicators
The accepted practice is to calculate the eligibility of a process only when the process has been demonstrated
is in a statistically controlled state. These results are then used as a basis for prediction,
what will be the performance of the process. It is almost worthless to make predictions based on the data that was
obtained from a process that is not stably repeatable over time. Special causes are responsible
for changes in the shape, scattering, or position of the distribution of the character characterizing this process
and thus can quickly invalidate the process capability prediction. So to use different

Brno, 2008 - 17 - Ing. Petr Kopl

Page 19
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Process pointers and indexes as tools to predict its behavior are required to get the data used
to calculate these characteristics was obtained from processes that are statistically mastered
condition.
The output of a stable process can be described by a probability distribution distribution function
the monitored quality characteristic (in a statistically controlled state). To make this division
probabilities can be used to predict the future state of the process, the process must be stable.
The probability distribution is described by parameters whose magnitude can be estimated using
statistics calculated from data obtained from the process.
The most common parameters are estimates of the position (center of gravity) and dispersion parameters in accordance
with customer requirements. Normally, the position is estimated using a sample average
or sample median. Dispersion is most often estimated using a sampling range
or sampling standard deviations.
The centering of the process and its dispersion mutually influence the production of an acceptable product.
If the division of the monitored character shifts from the center, the space for the possible decreases
fluctuations in variability. Shift in the position of the process, increase in the dispersion of the process or a combination of these
two factors can lead to the production of products outside the technical limits
specifications. A process with such a division could not be classified as satisfying
customer needs.
The most frequently used indicators and indices include C p and C pk , which are indicators
respecting both the fluctuations and the position of the center of gravity of the process and technical specifications. Then there are
process performance indicators P p and P pk .

Process rates for predictable processes - indicators

Cp Cp is an eligibility indicator. Compares process capability with maximum
possible process fluctuations due to the given tolerance field. This indicator expresses peace
how well the process meets the requirement for variability. The Cp pointer is determined by
formulas:
USL - LSL USL - LSL
C = =
, (1.1)
6p
p
C 6 R
d2
where USL - upper limit dimension, LSL - lower limit dimension,
^
p - estimate of standard deviation for X , d 2 - coefficient for calculation,
C

R- average range.
The indicator is not affected by the position of the process. It can only be calculated in the case of a prescription
 both limit values.

Brno, 2008 - 18 - Ing. Petr Kopl

Page 20
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

LSL USL Cp = 1
USL - LSL = ⇒
xi 1 USL - LSL = 6 p
6p C
C

σ 6C

Cp> 1
xi USL - LSL > ⇒
1 USL - LSL > 6 p
6p C
C
σ6
C

Cp <1
xi USL - LSL < ⇒
1 USL - LSL < 6p
6p C
C
σ6
C

Figure 4.1 Three values of C p

A good idea of the significance of the various C p shown in Figure 4.1

it also gives the inverse of C p , expressed as a percentage. E.g. for C p = 0.8 is 100 (1 / 0.8) =
125 which means that the tolerance interval is "used" to 125%, ie significantly exceeded,
while for C p = 1.2 it is 100 (1 / 1.2) = 83%.

C pk C pk is also an indicator of process capability. It also takes into account the position of the process. When
of the regulation of bilateral limit values, C pk is always at most equal to C p , ie it applies,
that C p ≥ C pk . Equality occurs only in a centralized process. The indicator C pk is obtained
as the smaller of the CPU and CPL values , where:
USL - X = USL - X
CPU = and (1.2)
3p C
3 R
d2

X - LSL = X - LSL
CPL = , (1.3)
3p C
3 R
d2
^
where USL , LSL , σ ,C d 2 and R - viz. above, X - overall average.

The indicators C pk and C p should always be evaluated and analyzed simultaneously. Value
A C p indicator greater than the C pk value indicates an opportunity to improve process centering.

Pp P p is an indicator of process performance. Compares process performance with maximum

permissible fluctuations with a given tolerance field. This indicator expresses the degree to which
how well the process meets the requirement of process fluctuations. The indicator P p is determined by
formulas:
USL - LSL = USL - LSL
Pp = , (1.4)
6p p 6 with
where USL, LSL - see above, s - sampling standard deviation,
σ - total fluctuation,
p

n
()
x -x
2

p p
= ≈with ∑ and
. (1.5)
1
n -1
 Brno, 2008 - 19 - Ing. Petr Kopl

Page 21
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

The value of P p is not affected by the process position.

P pk P pk is also an indicator of process performance. It also takes into account the position of the process. When
of the regulation of two-sided limit values, P pk is at most equal to P p , so it holds that P p ≥
P pk . P pk is obtained as the smaller of the PPU and PPL values , where:
USL - X = USL - X
PPU = and (1.6)
3p p 3 with

X - LSL X - LSL
PPL = = (1.7)
3p p 3 with

where USL , LSL , p σ , s , X - see above.

The indicators P p and P pk are to be evaluated and analyzed simultaneously. Greater value
indicator P p than the value of the indicator P pk shows an opportunity to improve the centering of the process.
For a statistically mastered process, the value of the process capability is very close to the value
its performance. The large difference between indicators C and P shows the presence of special (ch)
causes (causes).

Brno, 2008 - 20 - Ing. Petr Kopl

Page 22
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Second REGULATORY NO THEREOF DIAGRAMS (1.2)

Control diagrams can be used to monitor or evaluate a process.
There are basically two types of control charts. One for quantitative data, the other
for qualitative data. The process usually determines which type of control diagram is appropriate
 use. If the process data are discrete, then control charts should be used
for qualitative characteristics (comparison check). If the data are continuous,
then a control chart for quantitative data (measurement control) must be used.
One of the most commonly used control charts is the average charts ) ( X

and range (R) , for individual values (I) and sliding range (MR) , etc., which belong to the family
control diagrams during control measurements. Diagrams based on counting mismatched pieces
or discrepancies on the pieces ( p , np , c , u -diagrams) belong to the family of diagrams during inspection
by comparison.

Regula 2.1 No the diagrams: tools P R and process control

There must be effective means to effectively manage fluctuations during production
to detect special causes.
They provide statistical and probabilistic methods for time-dependent data
necessary and sufficient methods to determine whether special causes are present. Although for a solution
various methods are useful for this task, a class of control diagrams is universal and robust,
which was developed and introduced for the first time in the 1920s by Walter
Shewhart of Bell Laboratories studying the data. He was the first to distinguish between regulated fluctuations
from unregulated, ie fluctuations caused only by random causes from fluctuations
caused by special causes. He also developed a simple but effective detection tool
of these two types of causes, namely the control diagram. The control diagrams are oriented directly
effectively to the specific causes of fluctuations, if they occur, and to reflect the extent of the fluctuations induced
only by random causes that must be reduced by the system or process of improvement.
The errors described above cannot be reduced to zero. Shewhart was aware of that
and proposed a graphical approach to minimize the economic losses caused by
both types of errors mentioned above.
If process control activities ensure that no resources are present
For special reasons, the process we say is under statistical control or in statistically
managed state. Such processes behave as stable, predictable and consistent,
since process performance can be predicted.
The real presence of any special cause will push the process out of the state
statistically mastered or directly into a statistically unmanageable state. The performance of such
the state will always be unstable and cannot be predicted.

Regula No. her limits

When Shewhart designed control charts, he also dealt with the economics
process control, ie a measure to be taken in a process only when
there is a special cause. This is achieved by comparing appropriate sampling statistics
with regulatory limits. But how to set these limits?
The normal probability distribution of the observed quality characteristic by which
the behavior of the process can be described. Then the goal is to define when special causes affect the process. Other
the way the same problem is expressed - the process has changed since it was last researched
or during the data collection period?

Brno, 2008 - 21 - Ing. Petr Kopl

Page 23
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Because the normal distribution is described by the process position parameter (mean value)
and the variability parameter (range or standard deviation) raises the question: Has changed
is the position of the process or its variability?
Only a change in position is considered. What procedure can be used to determine
has the position changed or not? One option would be to look at each piece made
process, but it is usually uneconomical. An alternative is to remove the selection from the process
and calculate the sample average.
If the process has not changed, will the sample average be the same as the mean of the process?
The answer is that this happens absolutely exceptionally. But how is that possible? The process does at all
did not change. Does this not mean that the mean value of the process remained the same? The reason for this statement
is hidden in that the sample average is only an estimate of the mean value of the process.
Shewhart used the distribution of sample averages to determine the operational definition
process in a statistically controlled state. It begins with the assumption of the process
is in a managed state. Then the value of the sample average is compared with its distribution
using limits determined as three times its standard deviation. (Shewhart chose limits
in the form of three times the standard deviation as a limit economically advantageous for regulation
process). These limits are called regulatory limits. If the sample average falls outside of these
limits, there is reason to believe that a special cause has arisen. It can also be expected
that all selection averages will show a random arrangement within these limits. When
the group will form some non-random grouping, again this is a reason to believe that it is present
special cause.
 The data is compared to control limits to determine if process fluctuations
it is stable and corresponds to the fact that it was caused only by accidental causes. If it is obvious
The presence of a particular cause of fluctuations, the process is analyzed to determine what causes it.
After the implementation of measures into the process, additional data from the process are obtained, the control limits are recalculated
and, if necessary, respond to the occurrence of other possible special causes.
If all special causes have been captured and eliminated and the process is running
in a statistically mastered state, the application of control diagrams as a tool continues
for process monitoring and its capability can also be calculated. If there is fluctuation
caused only by random causes too large, the process can not produce output,
which permanently meets the customer's requirements. Then the process itself must be analyzed
and management measures usually need to be taken to improve the system.

Errors in regulatory No. them diagrams

Two types of errors are possible when applying control diagrams. The first type occurs,
if the relevant production process remains in a statistically controlled state, but the point accidentally falls
outside the regulatory limits. As a result, it is incorrectly inferred that the production process is no longer present
in a statistically managed state and there are costs for trying to find the causes of the non-existent
problem.
An error of the second kind occurs when the production process is in a statistically unmanageable state,
but the clogged point happens to be within the control limits. In this case, it is incorrect
concludes that the production process is in a statistically controlled state, and then costs are incurred
associated with the failure of the ability to detect an increase in nonconforming units at the output of production
process. However, the risk of error of the second kind is a function of three factors: the width of the regulatory limits,
the extent of selection and the degree to which the production process enters as statistically unmanageable.
The nature of these factors is such that the magnitude of the risk of error of the second kind can only be deduced
general conclusions.
Shewhart's system takes into account only the error of the first kind and within limits
3x standard deviation is the magnitude of this error 0.3%. Because in a given situation it is difficult to get
meaningful estimate of the cost of error of the second kind and because it is convenient to choose arbitrarily

Brno, 2008 - 22 - Ing. Petr Kopl

Page 24
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

subgroup of small scale 4 or 5, it proves appropriate and feasible to use limits

3x standard deviation and focus on managing and improving one's own behavior
production process.

2.2 Effective use etc. of iNOS regulatory No. them diagram of

Important benefits can be gained from the effective application of control charts.
The benefits and benefits of control charts include:
Management philosophy: The way a company is managed can directly influence
effectiveness of the SPC. The following are examples of what to respect:
• Focus on the organization to reduce volatility.
• Creating an open environment that supports the work of interdisciplinary teams.
• Applying the SPC in order to stimulate interest in deepening knowledge about fluctuations
in production processes.
• Application of SPC in the field of data management and use of conclusions of their processing
in everyday decision making.
Philosophy technician s : Engineers use data in developing proposals that can and will be
influence on the level and type of fluctuations on the final product. Here are some ways to
which technicians can demonstrate the effective use of the SPC:
• Focus on process fluctuations during its design, ie. for example, on the number
design changes, design for production and assembly, movement of workers, etc.
• Support and financially provide training for management and staff for proper use
and the SPC application.
• Require an understanding of the behavior of the process in terms of its fluctuations and stability in the relationship
for measurement and data that are used for design development.
• Support technical changes proposed on the basis of information obtained from the analysis
using SPC aimed at reducing fluctuations.
Production philosophy: As production develops and works, machines and transportation systems affect the level
and type of variation of the final product and therefore it is necessary:
• Focus the interest of the production organization on the reduction of fluctuations, eg on the number of different ones
processes, the impact of multi-purpose equipment, tools and machine maintenance.
• Do not allow work with control diagrams to end until the process is stable.
The transfer of responsibility for the production process should always take place after the start of the process
to a stable state.
• Ensure the correct localization of data outputs from the SPC for optimal use by employees.
Ø Quality Management: The quality is a critical component of the introduction of SPC.
 • Support training in SPC methods for management, engineers and workers
organization.
• Advise key people in organizing the correct implementation of SPC methods.
• Assist in identifying and reducing sources of fluctuations.
• Ensure optimal use of data and information from the SPC.
Production: Production workers are directly connected to the process and can influence
process fluctuations. Therefore, they have:
• Be properly trained in SPC and troubleshooting.
• Understand fluctuations and stability in relation to the measurements and data that are used
for regulating and improving the process.
• Be vigilant and communicative when conditions change in the process; update,
maintain and draw control diagrams within the scope of their competencies.
• Get to know the process and internal connections based on the collected information.

Brno, 2008 - 23 - Ing. Petr Kopl

Page 25
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

• Use information obtained from the SPC in real time while the process is running.

Using the approaches outlined above can create the right environment for understanding
and reducing process fluctuations. Minimal use of SPC for process monitoring will result in maintenance
process at the current level of performance. However, real improvement can be achieved when SPC
is used for a direct path to process analysis. Proper use of SPC methods can
direct the organization to improve product and process quality.

2.3 Regula No. her diagrams p of and control m measured tions

1
Control diagrams during measurement control represent a typical statistical application
process control, where processes and their outputs can be characterized by measurement values
quantitative sign. These control charts are useful for several reasons:
• The quantitative value (eg diameter is 16.45 mm) contains more information
than a simple yes-no statement (eg “the diameter dimension is within tolerance
field").
• Although collecting quantitative data is usually more expensive than collecting qualitative,
the decision to control the measurement can be achieved faster with a smaller range
subgroups (selection). This then leads to lower total measurement costs due to
higher yield of information in quantitative data.
• Because fewer pieces need to be inspected before reliable decisions can be made,
there is usually a shorter delay between the process change signal and the corrective action.
• Quantitative data can be used to evaluate process performance and improvement
can be quantified even if all individual values are within the limit values
given specifications. This is very important for the continuous improvement of the process.

The control diagram during measurement control can better explain the data from the process
in connection with process fluctuations, fluctuations from one piece to another and by means of diameter
process. For this reason, control diagrams are usually prepared during measurement control
and analyze in pairs, one diagram for process average and the other for process fluctuations.
The most commonly used pairs of diagrams are the X and R diagrams, with X
is the arithmetic mean of the values obtained from the small subgroups - a measure of the mean of the process, and R
is the sampling range of values within each subgroup (largest value found minus
smallest) - the rate of process fluctuations. However, there are other types of control charts,
which may be much more useful in certain circumstances.

1 Note: Due to the application of control diagrams in the control of measurements in the practical part
only the above-mentioned control diagrams are described in more detail in the theoretical part.

2.4 regulatory elements No them diagram of

There is no single approved way to display control charts. But still
it is necessary to be aware of the reasons for the use of control diagrams. Any format is acceptable,
if it contains the following:
• (A) Corresponding scale
The scale on the control chart should be such that it is possible to show the natural ones
process fluctuations. A scale with too narrow a scale on the control diagram does not allow
necessary analysis and regulation of the process.
• (B) UCL, LCL

Brno, 2008 - 24 - Ing. Petr Kopl

Page 26
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

The ability to determine values outside the regulatory limits that signal the occurrence of special causes,
is given to the control diagrams by plotting the control limits calculated by
distribution of selection characteristics. In no case may it be for analysis and regulation
process, the values given in the specification are used instead of the valid regulatory limits.
• (C) Central line CL
It requires to identify non-random groupings that signal the presence of a special cause
control diagram plotting the central line based on the sampling division
characteristics.
• (D) A sequence of subgroups / No Timeline
Maintaining the sequence in which the data was collected allows identification,
when the special cause occurred and whether this special cause is time-oriented.
• (e) identification of the point of where the process is not handled statistically
On the control diagram, the drawn points should be identified when the process was not
in a statistically mastered state. Once such a point is drawn on the diagram, it should be done
analysis of special causes and their identification.

Figure 1.2 Elements of the control diagram

Regulatory diagrams that are incorporated into the report as part of it, and those that are
kept manually, should contain the following information:
• What : part / product name and identification number.

Brno, 2008 - 25 - Ing. Petr Kopl

Page 27
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

• Where : process operation / step information and identification name.

• Who : name of the operator and evaluator.
• How : measurement system used, name, number, units (scale).
• How many : subgroup range, same or different.
• When : collection scheme (frequency and time of collection)
 Brno, 2008 - 26 - Ing. Petr Kopl

Page 28
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management
 Figure 2.2 Front of the manual control chart for diameter and span

Brno, 2008 - 27 - Ing. Petr Kopl

Page 29
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

2.5 The processes of controlled regula No. them diagrams

P of preparatory steps
Some preparatory steps should be taken before applying the control charts:
✓ Create an environment suitable for the implementation of SPC.
✓ Define the process.
✓ Identify the features or quality traits to be observed in the diagrams and are
based on:
∎ customer needs,
∎ existing and possible problem areas,
∎ link (correlation) between quality features.
✓ Defining a quality mark.
The quality mark must be defined by an operational definition so that the results can be obtained
communicable to all concerned, in a way that makes the same sense today as well
like yesterday. It should contain a specification of what information is to be collected, where, how
in what way and under what conditions.
The operational definition describes the quality feature to be evaluated and whether it is
a qualitative or quantitative sign.
✓ Defining the measurement system
The total (total) variability of a process consists of the variability it expresses
variability from one piece to another and from the variability of the measurement system. It's very
important to estimate the share of variability caused only by measurement on total variability
process and determine whether it is acceptable. The performance of the measurement system must be predictable
in terms of accuracy, consistency and stability. To validate the capability of the measurement system
periodic calibration alone is not sufficient for the intended use. The measurement system must be
in addition to calibration, it is also assessed for suitability for the intended use of the MSA.
MSA
One of the most common reasons for low quality data is that it is too large
variability. Interactions between can be a large part of this variability in the measurement set
measurement system and its environment.
If the interaction causes too much variability, then there may be quality
data so low that they are not useful. For example, a measurement system with great variability
may not be suitable for analyzing the production process because of system variability
the measurement may mask the variability of the manufacturing process.
When evaluating the measurement system, attention must be paid to three
basic questions:
1) The measurement system must demonstrate adequate sensitivity.
✓ Does the device have an appropriate sensitivity threshold? The sensitivity threshold (or class) is specified
 design and serves as a basic starting point for choosing a measurement system. Normally
the "rule of ten" or "10 to 1" rule is used, according to which the sensitivity threshold is to be divided
tolerance (or process variability) to ten or more parts.
✓ Does the measuring system demonstrate effective resolution? In relation to the sensitivity threshold
it shall be determined whether the measurement system has a detectable sensitivity in case of application and conditions
product changes or process variability effective resolution.
2) The measurement system must be stable.
✓ In repeatability conditions, the variability of the measurement system is caused only by
accidental causes and not special causes.
3) Statistical properties (errors) coincide in the expected range and are
adequate for the purpose of the measurement. In the case of product management, there must be system variability
measurement small compared to the limit values given in the specification. When

Brno, 2008 - 28 - Ing. Petr Kopl

Page 30
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

process control, the variability of the measurement system should prove effective
distinguishability and be small compared to the variability of the production process.

✓ Minimize unnecessary fluctuations

Unnecessary external causes should be reduced before monitoring the process
fluctuations as little as possible. This simply means paying attention to the process
worked the way he should. The aim is to avoid obvious problems,
which could and should be corrected without the use of control charts. This applies
setting up the process or interfering too often with its operation.

2.5.1 regulatory mechanism No. them diagram of

The steps in using control charts are:
1. Data collection
2. Determination of regulatory limits
3. Interpretation of statistical regulation
4. Extension of regulatory limits for existing regulations

Collection of r data
Control diagrams are based on the measurement of a certain quality feature. These measurements are
used to calculate regulatory statistics (selection characteristics such as average, range,
standard deviation), which describes a certain element characterizing the distribution
the probability of a regulated trait. Measurement data are collected from samples taken
from the process flow. Samples are taken in the form of subgroups, which can be formed once or
more units. In general, the larger the subgroup, the easier it is to detect
small process shifts.

Making ODB plan of ru samples s

To ensure the operation of the control diagrams, a logic sampling plan should be determined
subgroups. A logical subgroup is a group of samples taken to minimize
the chances of fluctuations caused by special causes manifesting within the subgroup
and at the same time maximized the chances for fluctuations between subgroups caused by special ones
causes. A guide to designing a sampling plan is to keep existing fluctuations between
subgroups was comparable to fluctuations within subgroups. Including consecutive ones
samples into a logical subgroup, the opportunity for the process to change is minimized
and to minimize fluctuations within subgroups. The frequency of subgroup sampling is usually
provides for the possibility of capturing the presence of specific causes
between subgroups.
Fluctuations within a subgroup represent fluctuations from one piece to another
in a relatively short period of time. Any significant variation between subgroups should be reflected
changes in the process to be investigated in terms of appropriate action.

Subgroup Range - How large a subgroup range should be is determined by the type itself
the process under investigation. The appropriate scope of the subgroup is determined by the responsible team. If
a small shift in the process is expected, then a larger range of subgroups will be needed
when the expected process shift will be large. The range of subgroups should be constant. Calculation
regulatory limits depend on the scope of the subgroup.

HF frequencies of interference subgroups - subgroups are taken one after the other successively
at a time, eg every 15 minutes or twice per shift. The goal is to detect changes in the process
in time. Subgroups should be removed as often and at appropriate times as possible

Brno, 2008 - 29 - Ing. Petr Kopl

Page 31
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

reflect possible opportunities for change. Possible causes of a change in the process may depend
from the differences characterizing work shifts, from the change of operators, the start-up phase of the process,
batches of material, etc.

After No et subgroups - the number of groups necessary to determine the control limits, should fulfill
the following criterion: remove such a large number of subgroups from the process that
that most of the possible sources of specific causes of fluctuations that could have affected were manifested
process. Generally 25 or more subgroups containing about 100 or more individual readings
it is sufficient for a good process stability test and if this process is stable, it also gives good estimates
location and dispersion of the process. This number of subgroups ensures that the impact is minimized
any extreme values in the range or standard deviation.

ODB scheme of ru samples s - In an unpredictable occurrence of specific causes in the process

the corresponding sampling scheme is a random sample. Random selection is like that
a selection in which each selection element (here a subgroup element) has the same chance of being selected.
Random sampling is systematic and planned, meaning that all samples are determined
before the actual data collection.

Regulatory structure No. whom you chart

The control diagram consists of the following parts:
✓ Header describing the sampling process and plan.
✓ Part for recording / displaying acquired numerical data.
This section should contain the date, time and other identifying information for each subgroup.
✓ Part for calculations.
This section should also contain space for calculations using observed readings
and calculations of statistics for process control.
✓ Part for graphical display of statistics (selection characteristics) for control
process.
The value of the selection characteristic intended for process control is usually shown
time is recorded on the vertical axis and on the horizontal axis. Measured values
and the displayed values of the characteristics in the diagram are to be drawn vertically. Scale
it should be wide enough to cover the variations of the plotted sampling characteristic.

Recording of raw data

✓ All individual measured values from each subgroup are recorded
and are identified for each subgroup.
✓ Next, each event in the process is recorded.

Calculation No et selection of rových characteristics (statistics) for each group

The statistics used for control, which are to be displayed in the diagram, are calculated from the data
in the subgroup. These statistics can be sampling mean, sampling range, etc.
The calculation is performed using formulas depending on the type of control diagram used
(Chapter 2.7).

Drawing selection of Rove characteristics (statistics) to regulatory NO whom you diagram

The calculated values of the statistics used for process control are plotted.
The values of the statistics calculated from one subgroup must be arranged vertically below each other.
To better highlight possible trends or non-random groupings, the drawn points are combined
solid line. To identify possible problems, the data should be checked immediately after plotting.

Brno, 2008 - 30 - Ing. Petr Kopl

Page 32
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

2.5.2 Determination of regulatory no them between

Regulatory limits are determined based on their natural fluctuations in the statistics used
for process control. They define the space in which the calculated values can move
statistics if the process is affected only by random causes of fluctuations. If it is counted
arithmetic mean of two subgroups from the same process, it is reasonable to expect both sample
the averages will be approximately the same. But because these statistics were calculated from different values
samples, both averages cannot be expected to be identical. Although both averages are different,
 yet there is a limit to what difference - caused by the presence of chance - can be made between them
expect. This fact determines the position of the control limits.
All types of control diagrams are based on this principle. When there is a process
stable (ie it works only under the influence of random causes), then there is a large one
the probability that the calculated value of the selected statistic for any subgroup will fall
between regulatory limits. If the value statistically exceeds the control limits, this indicates possible
the presence of a special cause of fluctuations in the process.

There are two phases in the analysis of statistical process control.

1. The first phase concerns the identification and elimination of special causes of process fluctuations.
The aim is to stabilize the process. Stabilized predictable processes are called
statistically mastered processes.
2. The second phase concerns the prediction of future measurements, ie the verification of existing ones
process stability. During this phase, the data is analyzed and responded to specific causes
in real time. The stabilized process can be analyzed and it is determined
whether it is capable of producing in accordance with the customer's wishes.

Ur No. learning center P of ímky and regulatory No. them limits the diagram
To support graphical analysis of plotted selection characteristics (statistics)
for process control, it is necessary to determine an estimate of the position of the process center of gravity on the control diagram
(central line) and control limits. To construct a control diagram, it is necessary to calculate
(formulas see chapter 2.7):
✓ central line CL,
✓ upper control limit UCL,
✓ lower control limit LCL.

2.5.3 Interpretation of statistical regulation

If there are no special causes in the process affecting its variability,
then the values of the selection characteristics (statistics) used for its regulation will be
randomly distributed between regulatory limits.
Special causes can affect either the process position (average) or the fluctuations (range).
The purpose of analyzing a control chart is to determine any evidence that process variability or
The position of the process does not work at a constant level - that is, one or the other of these
characteristics of the process or both at the same time are in a statistically uncontrollable state - and perform
appropriate measures.
Because the control limits for position statistics depend on the fluctuation statistics, they must
first analyze the stability of the process variability statistics. Statistics for variability
and for the position of the process are analyzed separately, but comparisons of non-random groupings between the two
Diagrams can sometimes provide new insights into the specific causes affecting the process.
The process cannot be described as stable until any of the control diagrams
signals features of a statistically uncontrollable process.

Brno, 2008 - 31 - Ing. Petr Kopl

Page 33
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Analysis of values plotted in the R -diagram

Because the ability to interpret the behavior of both the range and the average of the subgroups depends
to estimate the variability from one piece to another, it is necessary to first analyze R -
diagram. The drawn points are compared with the control limits in terms of whether the points are not lying
outside regulatory limits or do not show trends and non-random groupings (Chapter 2.6).

Identification and sown of measurement of specific P of s No in ( R Diagram)

A run analysis must be performed each time a special cause is signaled on the R- diagram
process and identify sources of causes and better understand the process. The conditions need to be corrected
leading to a special cause and, if possible, to prevent it from recurring.
The control chart is intended to analyze the problem in the process and to determine when the change occurred
and how long it lasted. However, not all special causes have a negative effect. Some strange
the causes can have a positive effect on improving the operation of the process by causing a reduction in fluctuations
span. These special causes should also be assessed for their possible incorporation
into the process when it proves desirable.

P of EPO No et regulatory No. them limit ( R Diagram)

When performing an input process analysis or when reassessing process capability
the control limits should be recalculated in order to exclude the influence of all time periods,
when the process was not in a statistically controlled state and when there were all special causes
identified, removed or incorporated into the process documentation. Only after exclusion
those subgroups - which were influenced by specific causes that were identified
and remove or incorporate into the documentation of the regulation as beneficial - it is possible to recalculate
 and draw a new average range and new control limits for the range. It is necessary to make sure
whereas all points of the margin lie within these new limits; otherwise it must be repeated
the whole sequence of activities: identification of special causes, correction and recalculation of regulatory ones
between.
If a subgroup was excluded from the R- diagram because it was found in it
influence of a special cause, the same subgroup should be excluded from the sample average diagram.
The revised values of the average range R and the overall diameter X are to be used
to recalculate the experimental control limits for arithmetic means.

Identification and sown of measurement of specific p R s no IN ( X Diagram)

Once a special cause is identified on the R -diagram and its effect has been removed,
the diagram for the averages can be evaluated in terms of the presence of special causes.
At each indication of the condition when the control diagram for the averages is the process
out of control, it is necessary to perform an analysis of the process and determine the reason for its presence
special causes, correct this condition and prevent it from recurring. Data on the diagram
shall be used as a tool to determine when such conditions have taken effect and for how long
the action took place. The timeliness of the analysis is important, both for diagnosis and for minimization
inconsistent outputs.

P of EPO No et regulatory No. them limit ( X Diagram)

When performing an initial analysis of the process and reassessing its competence,
all points on the diagram for which they have been identified and removed must be excluded
special causes. At the same time, these points must also be excluded from the R -diagram, because it does not exist
certainty that these discovered causes on the X - diagram did not affect the level of variability here as well.
Then it is necessary to recalculate and draw the overall diameter and the new control limits for the averages.
It is necessary to make sure that all averages are within new limits. If this is not the case,

Brno, 2008 - 32 - Ing. Petr Kopl

Page 34
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

it is necessary to repeat the whole sequence of activities: identification of possible special causes, their
correction and recalculation of regulatory limits.

2.6 Ur No measurement signal of statistical process Uncontrolled

2.6.1 Point outside regula No. her limits
The presence of one or more points outside some of the regulatory limits is the main evidence
the presence of a special cause of fluctuations. A special cause could have occurred before this point.
Because the occurrence of points outside the control limits is relatively rare if it is present only
fluctuations caused only by accidental causes, it is assumed that a special cause falls
at the expense of the extreme value. Thus, any point outside the control limit is a signal for analysis
process due to the presence of a special cause. When such a special cause occurs
all points outside the regulatory limits and in the process shall be marked for examination
appropriate measures shall be taken.
A point outside the control limit is generally a sign for the occurrence of one or more
from the following phenomena:
• The control limit or point was incorrectly calculated or plotted.
• Fluctuations from one piece to another or scattering of distributions increased (ie worsened)
either at a single point in time or as part of a trend.
• The measurement system has changed (eg due to another evaluator or meter).
• The measurement system lacks sufficient resolution.
A point outside the control limit is generally a signal that the process has shifted either
point or as part of a trend.

2.6.2 Non-random clusters or trends inside of regulatory No. them limits

Presence of unusual groupings or trends, even if all points are inside
regulatory limits, may be evidence of the effect of any particular cause during a non-random
grouping or trend. This may be the first warning of the presence of adverse conditions
in the process to be corrected. Conversely, certain groupings or trends may have
beneficial effect on the process and should be analyzed in order to make them possibly permanent
by improving the process.

Iteration
Iteration - Each grouping of the following is a sign of a process shift or beginning
the effect of some trends:
• seven points in a row on one side of the X and R .
• 7 points in a row form an increasing or analogously decreasing sequence.
The point that initiates the decision must be marked. It can be useful for marking back
 to the beginning of the iteration. The analysis should assess the approximate time at which he could start first
show a shift or trend.

An iteration above the span average (CL) or an iteration upwards indicates the presence
one or both of the following phenomena:
✓ Greater scatter in output values, which can be caused irregularly
manifesting cause (such as equipment failure or loosening of the clamping device)
preparation) or by shifting one of the process elements (eg new, less homogeneous
batch of bulk material).
✓ Change in the measurement system (eg new controller or meter).

Brno, 2008 - 33 - Ing. Petr Kopl

Page 35
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

An iteration below the span average (CL) or a downward iteration indicates that it may have occurred
one or both of the following phenomena:
✓ Smaller scatter in output values, which is usually a good condition to have
investigate for wider application and process improvement.
✓ A change in the measurement system that could mask the actual performance of the process.

Iteration relative to the process average is generally an indication that one or both may have occurred
the following phenomena:
✓ The process diameter has changed - and may change.
✓ The measuring system has changed.

Out of evidencing the non-random clusters

In addition to the occurrence of points outside the control limits or long iterations, they may occur
other different groupings of data that provide guidance for identifying specific causes.
Particular attention should be paid to incorrect interpretation of data, because it is random data
they can sometimes give rise to a presumption of non-randomness. Examples of non-random groupings are
obvious trends, cycles, overall scatter points within regulatory limits or even relationships
between values within the subgroup.

2.6.3 Special criteria for any special of í No. ins

There are several criteria for identifying specific causes. They are the most commonly used
listed in Table 1.2. The decision on which criteria to use depends on the process to be taken
analyzed.

Table 1.2 Overview of typical criteria for detecting special causes

P of an overview of typical criteria for detecting specific P of s No in

1 1 point is more than 3 standard deviations from the central line

2 7 points in a row lie on the same side of the central line

3 6 points in a row form a descending or ascending sequence

4 14 points in a row fluctuate alternately up and down

2 of the 3 points are greater than 2 standard deviations from the central line
5
(on the same page)
4 out of 5 points lie at a distance greater than 1 standard deviation from the central line
6
(on the same page)
7 15 points in a row lie within the middle third of the band between the regulatory limits

8 8 points in a row lie in the two outer thirds in the zone between the control limits

Brno, 2008 - 34 - Ing. Petr Kopl

Page 36
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

UCL UCL

CL CL

LCL LCL

Criterion 1 Criterion 2
UCL UCL

CL CL

LCL LCL

Criterion 6 Criterion 8
Figure 3.2 Example of some criteria for detecting special causes on the control card

In the examination described above, it is worth recalling that not all considerations are focused
to interpret the regulation can be applied in operation. It's easier to use them all
benefits of the computer than requiring the evaluator to remember too many criteria, and therefore
not everything is feasible directly in the workshop or in operation. So often a thorough analysis of the process
it can be done offline rather than in real time. This supports the need and importance
workshop logs and records in them of all phenomena occurring in the process and implementation
subsequent thorough analyzes.
Another consideration concerns the training of operators. Use of other criteria for process control
should be carried out directly in the workshop, if possible, but not before the operator is able
handle it. Both are combined with appropriate training and equipment. During time
and by gaining experience, the operator will be able to recognize non-random groupings in real life
time.

2.7 Formulas regulatory No the diagrams

Due to the application of control diagrams for diameter and span ( X , R ) in practical
only the formulas for the above diagrams are given in this chapter.

xx1 + + ... x n
X =
2
Pr of m of r subgroups: (2.1)
n
n = number of samples in the subgroup

Margin of three subgroups: xR= max

-x
min (2.2)
x max , x min - max., min. value from subgroup

Brno, 2008 - 35 - Ing. Petr Kopl

Page 37
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

X 1 + X 2 + ... X
X =
to
The (total) diameter of m of R (2.3)
to
k = number of subgroups used for the calculation
overall diameter and average range

R 1 + R 2 + ... R to
Pr of m of confidential margin of three: R = (2.4)
to
 ^
Sm estimate of standard deviation for X : p R
(2.5)
C= d2

^
^ p
Sm estimate of standard deviation for X : σ= C
(2.6)
X
n
Diagram parameters:
Central line : Regula No. her limits:

CL X = X (2.7) UCL X
+ =RAX
2 (2.8) LCL X
- =RAX 2 (2.9)

CL R = R (2.10) UCL R = RD4 (2.11) LCL R = RD

3 (2.12)

Table 2.2 Constants for calculation of control diagrams

ST Ø mean value - the margin THE VOLTAGE
n A2 d2 D3 D4
2 1,880 th most1,128
common
th most common
- 3,267 th most common
3 1,023 th most1,693
common
th most common
- 2,574 th most common
4 0.729 2,059 th most common
- 2,282 th most common
5 0.577 2,326 th most common
- 2,114 th most common
6 0.483 2,534 th most common
- 2,004 th most common
7 0.419 2,704 th most0,common
076 1,924 th most common
8 0.373 2,847 th most0.136
common 1,864 th most common
9 0.373 2,847 th most0.136
common 1,864 th most common
10 0.308 3,078 th most0.223
common 1,777 th most common
11 0.258 3,173 th most0.256
common 1,744 th most common
12 0.266 3,258 th most0.283
common 1,717 th most common
13 0.249 3,336 th most0.307
common 1,693 th most common
14 0.235 3,407 th most0.328
common 1,672 th most common
15 0.223 3,472 th most0.347
common 1,653 th most common
16 0.212 3,532 th most0.363
common 1,637 th most common
17 0.203 3,588 th most0.378
common 1,622 th most common
18 0.194 3,640 th most0.391
common 1,608 th most common
19 0.187 3,689 th most0.403
common 1,597 th most common
20 0.180 3,735 th most0.415
common 1,585 th most common
21 0.173 3,778 th most0.425
common 1,575 th most common
22 0.167 3,819 th most0.435
common 1,566 th most common
23 0.162 3,858 th most0.443
common 1,557 th most common
24 0.157 3,895 th most0.451
common 1,548 th most common
25 0.153 3,931 th most0.459
common 1,541 th most common

Brno, 2008 - 36 - Ing. Petr Kopl

Page 38
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

2.8 Stability Testing of process P of assumption by regulatory No. them diagram of

A prerequisite for the application of control diagrams is a statistically controlled process
and a normal probability distribution of the monitored quality trait. That is, studied
the random variable (quality mark) does not change over time. The shape of its division does not change either
parameters - mean value (process center of gravity) and variance (process variability). If so
no, the risk of a false alarm, ie the risk that the selection characteristic used for monitoring
process, accidentally occurs outside one of the regulatory limits, can significantly exceed the value
0.00135. Thus, if the assumption of a statistically controlled process is observed, this can happen
on average, one of 740 control subgroups.
Verification that the process is statistically mastered and therefore stable presupposes verification of the hypothesis,
that all selected subgroups come from basic normally distributed files
with the same mean and the same variance. If this hypothesis does not apply, it cannot be applied
control diagrams and modified, extended control limits must be used.
In the practical part, process stability testing will be performed using software
MINITAB Release 14, which is statistical software for Windows.
 Brno, 2008 - 37 - Ing. Petr Kopl

Page 39
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Third SOU No Current status CHECKS Rom ETEN The JMA

The third chapter of the study deals with the line for the production of spindles for water fittings.
It describes the individual workplaces on the line. The next section is devoted to the inspection of spindles
at the SP 12 CNC workplace, where the SPC method will be further applied.

3.1 Production line for the row eten

Several dozen types of spindles for water fittings are manufactured on-line. In the year 2004
the line has undergone a thorough modernization, thanks to which it can respond to the fulfillment of specific orders,
both in terms of quantity and quality.
The beginning of the line is a horizontal machining center, where the bar material is cut
first machined from the foreheads and provided with centering dimples. This is followed by induction heating of the semi-finished product
to forming temperature. The hydraulic press creates a shape on the rod for the future collar. As
due to the buckling forces during pressing, the axis of the bar bends, a lathe is located next to the press,
where the axis of the blank is compared. Next, the bars are placed in baskets and moved to the furnace
to normalization annealing. Another link in the line consists of a pair of semi-automatic CNCs
lathes that give the spindle a characteristic shape. The part of the spindle under the seal is here
completed on done. Subsequently, the spindle is moved by a roller conveyor
to a single-purpose machining center, where the pyramid is machined, the production of internal
thread and drilling a hole for a cross pin. They are moved from the machining center by robots
spindle to the last stage of production, which is the rolling of an isosceles trapezoidal thread.
The manufactured spindles are equipped with protective sleeves made of plastic mesh for protection. Thus
the packed spindles are stored in europallets and transported to the assembly workplace. Line diagram
is shown in Figure 1.3.
 Brno, 2008 - 38 - Ing. Petr Kopl

Page 40
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Legend:
9
→ Material flow direction

8
1. Horizontal machining center

(HELLER BEA 05)

2. Induction heating of semi-finished product

7 (ADBOTERM R 50,000)

3. Hydraulic press

(PYE 63 S1)

6
4. Lathe with straightening pulley

5. Annealing furnaces

(EL 5 KLO R2V)

6. Semi-automatic lathes

(SP12 CNC)
5

7. Single-purpose machining center

(TOS JÚS)

8. Industrial robots

4 (RP 16)

9. Thread rolling machines

1
3 (UPW 25)
2

P
AT
T
WITH
IN
Figure 1.3 Diagram of the spindle production line

Brno, 2008 - 39 - Ing. Petr Kopl

Page 41
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

3.2 Checking in of ethene workplace SP 12 CNC

Machining operations at this workplace are crucial for the geometric shape
spindle before thread production, on the one hand for the functional completion of the sealing part of the spindle to done
(dimension for SPC application). The principle of self-control of production is introduced in the company directly
production workers at individual workplaces. Because the company has been certified for many years
according to the EN ISO 9001 system, marking and sorting of products by color is performed at the workplace
identification cards with the signatures of the production worker concerned. Identical products
with the customer's requirements are marked with white identification cards, products do not match
but repairable are marked with green cards and products are inconsistent and unrepairable
marked with red cards. Repairable products and non-repairable products each have their own
color-coded palette and are isolated from other productions. At the end of the shift they will
production operators record the performance and status of work in progress in the CAPP system through
terminals at their workplaces.

Check the reference dimension of RU - pr of m of r in row eten under seal

The technological manual (see chapter 4.1.1) prescribes the inspection of every third piece
and prescribes a digital caliper micrometer of the appropriate range as a meter (see chapter 4.1.2).
The current incidence of non-conforming and unrepairable work-in-progress spindles is approximately one
non-conforming product for 200 pcs of production.

Brno, 2008 - 40 - Ing. Petr Kopl

Page 42
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

4. DRAFT SPC METHODOLOGY

The fourth chapter deals with the design of the application of the SPC method to the production process.
A typical representative of the spindles is selected and a specific dimension on it, for which the methodology will be
SPC designed. Further in this section are other requirements for the application of SPC.

4.1 Selection P of REPRESENTATIVE manufactured of out in the row eten

Based on the findings of the total production of all manufactured pieces of spindles for the previous year
 (2007)
flangedspindle no. 4HA
gate valve EKO 3503-038
- Plus PN (Figure
16; DN 1.4), which
40. This is part
slide of the shut-off
has happened in thevalve
last ten
years the best-selling product from the JMA offer. For its quality and favorable price
has gained its customers in more than fifty countries around the world.
In 2007, 46,000 pieces of this type of spindle were produced. They are so most common
manufactured type on the line. In one shift, 200 pieces are produced on one SP 12 CNC machine
spindles (another type is usually produced on the second machine).

Brno, 2008 - 41 - Ing. Petr Kopl

Page 43
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management
 Figure 1.4 Spindle no. 4HA 3503-038 with marked regulated sign

Brno, 2008 - 42 - Ing. Petr Kopl

Page 44
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

4.1.1 Extract from the technological process for the row spindle
Table 1.4 contains an extract from the technological process for spindle no. 4HA 3503-
038. The highlighted section in the table describes the SP 12 CNC workplace.

Table 1.4 Extract from the technological process with a description of control operations and means at the spindle
Work operations and NO Frequency
Control Used free from the control of Formulation
job description controls
Alignment and drilling: lengths, 251231 caliper 250, each
- clamp to average 251238 caliper 150 tenth piece
preparation
- to balance
- align,
drill both drilling visually after a dose
cell
- clean up
diameter
Pressing: lengths,
251238 caliper 150,
- warm up to dimension
251124 steel gauge 300
forging collar each
temperature tenth piece
251811 dial indicator 60,
- press throwing
251858 magnetic stand
collar
Annealing: three pieces of
hardness 000020 hardness tester RB1
- anneal benefits
Turning: 47-252-02 turning gauge ∅ 20
- load d11,
pr of m of interference,
spindles to digital micrometer 0-25,
length each T of third
tray 251238 caliper 150,
piece
- turn 251280 sliding depth gauge 160
according to NC roughness
250163 Rugotest 2632 roughness standard
program surface
251811 dial indicator 60,
throwing after a dose
251858 magnetic stand
Turning: 47-283-03 pyramid gauge 14.3,
pyramid
- load 251238 caliper 150,
spindles to 48-012-02 thread depth gauge M8,
thread every fifth
tray 254020 Mzt M8-6H,
piece
- work roughness
250163 Rugotest 2632 roughness standard
according to NC surface
program precipitation visually
Rolling length 251420 micrometer 0-25
trapezoidal thread:
- insertion
 spindles to each
tray 254220 Mzk Tr 18x4L-Se12,
thread tenth piece
- roll 254225 Mzk Tr 18x4L-Se12 scrap
thread
grooving
way
Note: Batch = factory design or one part of a design that is in production

Brno, 2008 - 43 - Ing. Petr Kopl

Page 45
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

4.1.2 Existing m measured ICI as of devices

For the purpose of testing the SPC methodology in this particular case, the use is appropriate
digital caliper micrometer. For the spindle according to drawing 4HA 3503-038 is for measurement
- 03.0
dimension ø 20 - 06.0 mm caliper prescribed a digital caliper micrometer
with a range of 0 ÷ 25mm (caliper micrometer DIGIMATIC series 293-240; manufacturer MITUTOYO).
The value of one scale division is 0.001 mm. The manufacturer states the maximum permissible error
gauges for this type and range value ± 2µm (accuracy exceeding the requirements of DIN 863).
The use of this meter therefore ensures compliance with the tolerances prescribed in the drawing.

Evaluation of the measuring system (micrometer) according to chap. 2.5 paragraph MSA :
1) Rule ten
T - tolerance given by the drawing [mm] - 0.03 mm,
c - sensitivity [mm] - 0.001 mm.
T 03.0
= = 30 > 10
C 001.0
the measuring device has a corresponding sensitivity threshold. The tolerance is divided into more than ten
parts. The meter is an order of magnitude more accurate than the manufacturing tolerances.
2) Stability of m measured measurement
3) Statistical properties
For Parts 2 and 3, it is assumed that all requirements for measuring equipment are met.
These points are not specifically addressed in the study, because the work deals with the application of the SPC method
into the production process and not a detailed analysis of the measurement system - MSA.

4.1.3 Draft m measured -enhancing and monitoring for of devices

In the case of subsequent integration of the SPC methodology into the production process, it is appropriate
to suggest the use of measuring devices with data output - digital caliper
micrometer with data output, and monitoring equipment - evaluated microprocessor
unit with printer and connecting cable. The set can be further supplemented with an interface
for possible data transfer to PC. All parts of the measuring chain come from the Internet
Mitutoyo meter catalog. Detailed technical data of the mentioned components
are listed in Annex 1 at the end of this study.
These measuring and monitoring devices make the work of the production operator very easy
with control card management. He deals only with production and measurement. The system prevents errors,
which may arise from operator intervention. After the system is installed, only the measurement changes
means, the method of measurement remains the same.
If a signal cable is added to the measuring device together with the interface
for data transfer to a PC, then the data could be further processed using a software system
MCOSMOS, which contains MeasurLink. This program is specially designed for statistical
data evaluation.

4.2 Selection of r degree of control

The SPC method has not yet been introduced at the SP 12 CNC workplace or elsewhere in the company,
therefore, statistical control on control diagrams is proposed for initial verification in practice
guided by hand.
After manual verification of the use of the SPC method on the spindle line, processing is offered
control diagrams using computer technology see. chapter 4.1.3.

Brno, 2008 - 44 - Ing. Petr Kopl

Page 46
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

4.2.1 Selection of R-type regulatory NO whom you diagram

Due to the characteristics of the production process, it seems to be the optimal design of the application
control diagram for mean values and range - ( X , R ) according to figure 2.2. To the necessary
a pocket calculator will be used in the calculations on the diagram.
After possible introduction of a digital meter with data output into the control
process, the calculations will be performed automatically in the microprocessor unit or
on the computer see. chapter 4.1.3.

4.2.2 Management regulatory No. whom you diagram ( X , R )

Ur no measurement of the regulated character
- 03.0
The regulated feature is the diameter of the spindle under the seal ø 20 - 06.0 mm. To meet
the sealing function must adhere to the prescribed dimensions.

Selects the No 's ODB and size of r s

Since it is necessary to capture all possible influences entering the process, it is proposed
make twenty-five possible samples that are recommended for positioning and variance
process. The recommended order size is five consecutive products. Proposed frequency
sampling is every thirty minutes. Thus, a sample of 125 pcs is obtained. So we can talk
about a good test of the stability of the monitored process.
The next procedure for keeping the control diagram is already described in chapters 2.4 and 2.5.

Brno, 2008 - 45 - Ing. Petr Kopl

Page 47
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

5. IMPLEMENTATION OF THE SPC METHOD IN THE PRODUCTION PROCESS

The fifth chapter verifies the SPC methodology in operating conditions on the spindle production line.
It lists all the calculations needed to keep control diagrams. The diagrams are
subsequently processed and evaluated.

5.1 Specification of process conditions

During the manufacturing process, the following cutting edges were used to machine the spindle
plates:
 Roughing: MITSUBISCHI
one cuttingDNMX 150608 -approx.
edge machines MW US607020, four-edged,
spindles.
Clear: ISCAR DNMG 110404 - NF IC 907, four-edged, one-edged
machines about 40 spindles.
BLASTER BC-25 emulsion was used for cooling.

The measurement of the controlled character was performed with a digital micrometer of range
0 ÷ 25mm, e. No. 3/54, calibrated (ČSN 01 0115) according to the internal calibration procedure JMA
Hodonín.
Samples were taken at half-hourly intervals, one sample counting five samples.
Data collection was performed twice - for results after the 1st implementation, see. further - in two
consecutive shifts. Based on the obtained data, three regulatory ones were processed
diagrams.

5.2 Calculation No et values for the first regulatory No the diagram - 1. Implemented
The formulas and coefficients given in section 2.7 (all
the measured values are in the control diagram given in Annex 2.). In front of themselves
calculations, a test was performed to see if the measured values had a normal distribution. Testing was
performed using the MINITAB Release 14 program, using the Kolmogrov-Smirnov test.
Since the P- value is greater than 0.150, a normal distribution can be assumed
the probability of the monitored quality characteristic. Figure 1.5 shows the test result. Giant. 2.5 displays
histogram of measured values. Based on the measured and calculated values, it was processed
control diagram - complete is given in appendix no. 2, simplified is shown in fig. 3.5.

Brno, 2008 - 46 - Ing. Petr Kopl

Page 48
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Figure 1.5 Test of normal distribution according to Kolmogrov-Smirnov - 1. Implementation

 Figure 2.5 Histogram for spindle diameter - 1. Implementation

Brno, 2008 - 47 - Ing. Petr Kopl

Page 49
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Calculation No. You:

Pr of m of r subgroups (first):
x 1 + x 2 + ... x5 + 954,19 + 962,19 th+most + 948,19
X1 = = 963.19 958.19
common = 957.19 mm
5 5
Margin of subgroups (first):
R1 = x 1 max - xmin1 = 963.19 - 948,19 = 015.0 mm
Calculations of the values of other samples are no longer performed here, the values of the following
consumption are shown directly in the control diagram.
The (total) diameter of m of R
X1 + X + ... X + + 9612.19 ...
= 957.19 9592.19
25
X = = , 1995962 mm
2

25 25
Pr of m of confidential margin of three:
R1 + R 2 + ... R25 + +
= 015.0 006.0 012.0 ... = , 000856 mm
R =
25 25
Sm estimate of standard deviation for X :
^
pC = R = , 000856 = , 000368 mm
d2 326.2

Diagram parameters:
Central line :

CL = X = 9596.19 mm CL R ==R , 000856 mm

X
Regula No. her limits:

UCL = RAX
+ = 9596.19 + , 0.577.0 00856 = 9646.19 mm
X 2

LCL = RAX
- = 9596.19 - , 0.577.0 00856 = 9547.19 mm
X 2

UCL R = RD
4
= , 0.114.2 00856 = 0181.0 mm

LCL R = RD3 = , 0.0 00856 = 0 mm

 Brno, 2008 - 48 - Ing. Petr Kopl

Page 50
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Figure 3.5 Simplified control diagram - 1. Implementation

Interpretation regula No. whom you chart - 1. Implementation of

Looking at the range diagram, it can be seen that the individual values are even
distributed around a line representing the mean. No special causes are apparent.
The mean value tab shows a sharp deviation in the time interval between 13:30
to 14:00. The process is statistically unstable because the upper limit of the intervention has been exceeded
(production dimensional tolerances were not exceeded). The cause was excessive wear of the blade
tool cutting inserts. After replacing the insert, a manual correction was made in the controller
machine system. The parts remained fine. It is not assumed in other control diagrams
repeated exceeding of the intervention limits due to excessive wear of the tool edge,
because the excessive wear of the blade was probably due to a material defect
inserts.
Due to the occurrence of points outside the intervention limits, the values for the control diagram must be
recalculated and the control diagram must be re-constructed (1. Implementation -
repair). Measured values in the time intervals 13:30 and 14:00 are not taken into account during the conversion.
The calculation procedure is similar to chapter 5.2. Complete control diagram 1.
Implementation - the correction is given in Annex No. 3.

Brno, 2008 - 49 - Ing. Petr Kopl

Page 51
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Figure 4.5 Test of normal distribution according to Kolmogrov-Smirnov - 1. Implementation - correction

Figure 5.5 Histogram for spindle diameter - 1. Implementation - correction

Brno, 2008 - 50 - Ing. Petr Kopl

Page 52
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

P of EPO No. those regulatory No. whom diagram:

Pr of m of r subgroups (first):
xx1 + + ... x5 + + + +
= 963.0 958.0 962.0 958.0 958.0 = 9598.0 mm
X1 =
2

5 5
Margin of subgroups (first):
R1 = xmax1 - xmin1 = 963.0 - 958.0 = 005.0 mm
Pr of m of r subgroups (first):
xx1 + + ... x5 + + + 958.19
= 963.19 954,19 962,19 th most
+ 948,19
common
X1 = = 957.19 mm
2

5 5
Margin of subgroups (first):
 R1 = xmax1 - xmin1 = 963.19 - 948,19 = 015.0 mm
The (total) diameter of m of R
X 1 + X 2 + ... X + +
= 957.19 9592.19 9612.19 ... = , 1995906 mm
23
X =
23 23
Pr of m of confidential margin of three:
R1 + R 2 + ... R23 + 006.0 + 012.0 ...
R= = 015.0 = , 000857 mm
23 23
Sm estimate of standard deviation for X :
^
pC = R = , 000857 = 0037.0 mm
d2 326.2

Diagram parameters:
Central line :

CL X ==X 9591.19 mm CL R ==R , 000857 mm

Regula No. her limits:

UCL + =RAX 2
= 9591.19 + , 0.577.0 00857 = 964,19 mm
X

LCL - =RAX 2
= 9591.19 - , 0.577.0 00857 = 9542.19 mm
X

UCL R = RD
4
= , 0.114.2 00857 = , 001812 mm

LCL R = RD
3
= , 0.000857 = 0 mm

Brno, 2008 - 51 - Ing. Petr Kopl

Page 53
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Figure 6.5 Control diagram - 1. Implementation - repair

Interpretation regula No. whom you chart - 1st Implementation - Correction

The control diagrams provide evidence that the process is in a statistically controlled state
(after excluding subgroups 16 and 17). The process probably has a normal distribution (Fig. 4.5 P -
 value> 0.150),
respectively P ptherefore
and P pk the
. values of the process capability indicators C p and C pk can be calculated

Calculation No et values of C p and C pk :

USL - LSL = USL - LSL = 97.19 - 94.19 03.0
Cp = = = 36.1
6p C , 06 00857 , 002211
6 R 326.2
d2
C pk = min CPU , CPL

USL - X = USL - X = 97.19 - 9591.19 = 0109.0 = 99.0

CPU =
3p C , 03 00857 , 001105
3 R 326.2
d2

X - LSL = X - LSL = 9591.19 - 94.19 = 0192.0 = 74.1

CPL =
3p C 3 R
, 03 00857 , 001105
d2 326.2
C p = 1.36 C pk = 0.99
Calculation No et values P P and P pk :
USL - LSL = USL - LSL = 97.19 - 94.19 = 03.0 =
Pp = 38.1
6p p 6 with , 0.6003616 0217.0

Brno, 2008 - 52 - Ing. Petr Kopl

Page 54
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

n ()
xx -
2 23 (x1 - 9591.19 )2
p p
= ≈with ∑ and

n -1
= ∑ 123-
= , 0003616
1 1
P pk = min PPU , PPL

USL - X = USL - X = 97.19 - 9591.19 = 0109.0 =

PPU = 00.1
3p p 3 with , 0.3003616 , 001085

X - LSL = X - LSL = 9591.19 - 94.19 = 0191.0 =

PPL = 76.1
3p p 3 with , 0.3003616 , 001085
P p = 1.38 P pk = 1.00

Interpretation of index s way of eligibility

C p is not approximately equal to C pk , P p is not approximately equal to P pk . These two facts show
that the process is not well centered. Because C p and P p are almost the same, there are only very small ones
fluctuations between subgroups. A large difference between C pk and P pk could signal the presence
excessive fluctuations between subgroups. The values of C p and P pk are relatively high, i.e.
that if the process remains in a statistically controlled state, it is able to produce practically
zero proportion of non-conforming products.
From the above, the possibility of improving the centering of the process results. This will be achieved
by resetting the machine. The machine is set to clear diameter when machining the spindle diameter
19,955, this should improve the centering of the process, and thus increase the indicators C pk
and P pk . After this intervention in the process, another control diagram was prepared - 2. Implementation.

5.3 Calculation No et values for the second regulatory No the diagram - 2. Implementation
To verify the improvement of the process centering after resetting the machine tool
another regulatory diagram was developed, together with the calculations of the competency indices
(2. Implementation).
The sampling procedure is based on Chapter 5.1. Formulas were again used for the calculations
and the coefficients given in section 2.7 (all measured values are in the control diagram
listed in Annex 4). Before the calculations themselves, the test was performed again to see if they were measured
values have a normal distribution. Testing was performed using the MINITAB program
Release 14, using the Kolmogrov-Smirnov test. Since the P -value is greater than 0.150,
a normal probability distribution of the observed quality characteristic can be assumed.
Figure 7.5 shows the test result. Giant. 8.5 shows a histogram of the measured values. Based
measured and calculated values, a control diagram was processed - the complete one is given
in Annex No. 4, a simplified one is shown in Fig. 9.5.
 Brno, 2008 - 53 - Ing. Petr Kopl

Page 55
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Figure 7.5 Test of normal distribution according to Kolmogrov-Smirnov - 2. Implementation

Figure 8.5 Histogram for spindle diameter - 2. Implementation

Brno, 2008 - 54 - Ing. Petr Kopl

Page 56
 VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

Calculation No. You:

Pr of m of r subgroups (first):
xx1 + + ... x5 + + + +
= 957.0 953.0 954.0 954.0 964.0 = 9564.19 mm
X1 =
2

5 5
Margin of subgroups (first):
R1 = xmax1 - xmin1 = 964,19 - 953,19 = 011.0 mm
The (total) diameter of m of R
X + X 2 + ... 25
X = 9564.19
+ 955.19 + 9568.19 ...
X = 1 = , 1995519 mm
25 25
Pr of m of confidential margin of three:
R1 + R 2 + ... R25 + +
R= = 011.0 007.0 01.0 ... = , 000808 mm
25 25
Sm estimate of standard deviation for X :
^
pC = R = , 000808 = , 000347 mm
d2 326.2

Diagram parameters:
Central line :

CL X ==X 9552.19 mm CL R ==R , 000808 mm

Regula No. her limits:

UCL + =RAX 2
= 9552.19 + , 0.577.0 00808 = 9599.19 mm
X

LCL - =RAX 2
= 9552.19 - , 0.577.0 00808 = 9505.19 mm
X

UCL R = RD
4
= , 0.114.200808 = , 001708 mm

LCL R = RD
3
= , 0.000808 = 0 mm

Brno, 2008 - 55 - Ing. Petr Kopl

Page 57
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management
 Figure 9.5 Control diagram - 2. Implementation

Interpretation regulatory No. whom you diagram - 2. Implementation

Looking at the range diagram, it can be seen that the individual values are even
distributed around a line representing the mean. No special causes are apparent.
The same is true for the mean value diagram, again no special effect on the process
causes. The process is therefore able to meet customer requirements.
The control diagrams provide evidence that the process is in a statistically controlled state.
For these data, it is therefore possible to calculate the process capability indicators C p and C pk and P p and P pk, respectively .

Calculation No et values of C p and C pk :

USL - LSL = USL - LSL = 97.19 - 94.19 03.0
Cp = = = 47.1
6p C , 0.6 00808 , 002084
6 R
d2 326.2
Cpk = min CPU , CPL

USL - X = USL - X = 97.19 - 9552.19 = 0148.0 =

CPU = 44.1
3p C , 0.300347 , 001041 , 001041

X - LSL X - LSL 9552.19 - 94.19 0152.0

CPL = = = = = 46.1
3p C , 0.300347 , 001041 , 001041
C p = 1.47 C pk = 1.44

Calculation No et values P P and P pk

USL - LSL = USL - LSL = 97.19 - 94.19 = 03.0
Pp = = 56.1
6p p 6 with , 0.6003174 , 0019044

Brno, 2008 - 56 - Ing. Petr Kopl

Page 58
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

n
()xx - 2 25
(x1 - 9545.19 )2
p p
= ≈with ∑ and

n -1
= ∑ 125-
= , 0003174
1 1

Ppk = min PPU , PPL

USL - X = USL - X = 97.19 - 9552.19 = 0148.0 =

PPU = 53.1
3p p 3 with , 0.3003174 , 0009522

X - LSL = X - LSL = 9552.19 - 94.19 = 0152.0 =

PPL = 19.2
3p p 3 with , 0.3002312 , 0006936
P p = 1.56 P pk = 1.53

Interpretation of index s way of eligibility

C p is approximately equal to C pk , P p is also approximately equal to P pk . These two facts
show that the process is well centered. Because C p and P p are approximately the same, there is only a lot
small fluctuations between subgroups. The values of all indicators are relatively high,
that is, if the process remains in a statistically controlled state, it is able to produce
virtually zero proportion of non-conforming products. A big difference between C pk and P pk could
signal the presence of excessive fluctuations between subgroups. The big difference between C p and C pk
(similarly between P p and P pk ) could signal a process centering problem.

5.4 Evaluation regulatory No. them diagram of

When implementing control diagrams into the production process, it was necessary to process
three control diagrams. The first control diagram (given in Annex 2 - 1. Implementation)
tells of the process without prior statistical regulation. There are two distinct points outside
limits of impact, which occurred due to excessive wear of the tool edge (the cause was
 probably a defectmastered
in a statistically in the insert
state.material).
Because The watched
the limits process waswere
of intervention not found
exceeded, they had to be
recalculated UCL , LCL and CL . Thanks to this, the second control diagram was prepared (shown in
in Annex No.3 - 1. Implementation - correction). The second control diagram is based on the first.
The monitored process is already in a statistically controlled state and could therefore be calculated
values of eligibility indicators. After their analysis, the machine was reset to
insufficient process centering. After the intervention in the process, the last regulatory one was processed
diagram (given in appendix no. 4 - 2. Implementation). The third control diagram with indicators
eligibility is fine in all respects. The process is in a statistically controlled state
and is able to meet customer requirements.
Based on practical verification in operating conditions, it can be said that the methodology
designed for the production of spindles is valid and has sufficient for the regulation of the given process
telling ability. Monitoring of the process allows in addition to preventive measures
also mediated information about tool life and the actual wear of the machine. When
limitation of systematic influences on the given workplace, sampling will be sufficient for control
at hourly intervals of 5 consecutive samples.
All the above reasons suggest for further mechanization of the SPC methodology using
monitoring and measuring equipment proposed in Chapter 4.1.3. Regulation of course
does not imply complete elimination of nonconforming products in this workplace. It remains a possibility
the occurrence of a non-conforming product due to defects in previous sections of the production line. Significant
inaccurately drilled centering dimples are the causal factor affecting the production of the non-conforming spindle.
The consequent loss of the geometric shape of the spindle is no longer repairable. This fault occurs
for one piece out of about 400 pieces produced.

Brno, 2008 - 57 - Ing. Petr Kopl

Page 59
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

6. ECONOMIC EVALUATION
In one shift of the machine workplace, 200 pieces of spindles type 4HA 3503 - 038 are produced.
According to long-term statistical monitoring of the occurrence of scrap, 1 piece per 200 is non-conforming
pieces produced. One SP 12 CNC machine workplace has 230 shifts per year (number
pieces produced in 2007 divided by the number of pieces produced per shift).
Following the introduction of the SPC methodology, a reduction in the number of non-conforming products is expected,
that the number of nonconforming products drops to 1 piece per 400 pieces produced. Price of one spindle
in progress (after turning) is CZK 247. Investment in measuring and monitoring
equipment for one workplace is according to the current price list of the Mitutoyo company CZK 15,640 with VAT
(micrometer 5229 CZK, miniprocessor 8633 CZK and connecting cable 1779 CZK).
A comparison of the current state with the proposed method is in Table 1.6.

Table 1.6 Economic evaluation - current status / SPC method

Pointer Current status Pointer SPC method
Annual production Annual production
46 000 46 000
Q T [pcs / year] Q T [pcs / year]
Inconsistent Inconsistent
products Q n1 230 products Q n2 115
[pcs / year] [pcs / year]
Q n1 / Q T 100 [%] 0.50 Q n2 / Q T 100 [%] 0.25
Costs Costs
mismatched 56 810 mismatched 28 405
products N n1 [CZK] products N n1 [CZK]
Investments in
new equipment 15 640
N inv [CZK]

Calculation of the SPC method

Savings
AT = N
sp n1
- N
n2
= 56 810 - 28 405 = 28 405 To no

Profitability
ATsp 28 405
RE = 100. = 100. = % 62.181
N inv 15 640

Return
N inv 15 640
N and= = = 55.0 year
ATsp 28 405

The above economic calculations confirm that the investment in new equipment
 will be paid and the costs incurred will be reimbursed in about six months. To these facts
it is still necessary to add the labor savings of the production operator, which lies not only
in saving time spent on control measurements, but it is also given by saving work consumed
for possible repair of non-conforming but repairable pieces.
These savings in the overlapping time of the production operator can be used e.g.
for the handling of material between the workplace and the intermediate warehouse, which it currently provides
handling worker operating a machine shop.

Brno, 2008 - 58 - Ing. Petr Kopl

Page 60
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

7th CONCLUSION OF R
The aim of this work is to design the application of SPC method in the production process for the company
Jihomoravska armaturka spol. s ro Specifically, it is the use of SPC method for production
line that produces spindles for water fittings. Because they are made different on the line
types of spindles, so (thanks to the largest annual production) one specific type of spindle was chosen -
spindle for flanged shut-off valve EKO - Plus PN 16 DN 40. It was selected on it
one dimension, which is the most important in terms of the function of the whole valve (spindle diameter
under seal) for the application of the SPC method. It has been designed and verified for the above-mentioned dimension
SPC methodology. The study also included the design of measuring and monitoring equipment
in the case of permanent application of SPC to the production process.
During the implementation, three control diagrams were processed. In the first regulatory
there were two points outside the limits of the diagram. Exceeding the limits was caused
excessive wear on the tool edge. Following the first control diagram
the second was processed (exclusion of subgroups causing exceedance limits).
After the elimination of subgroups, the process was already in a statistically controlled state. Subsequently, they were
calculated and evaluated indicators C p , C pk , P p and P pk . After their analysis, an adjustment was proposed
machine tool to improve process centering. The last control diagram that was
processed after interventions in the process, testifies to a statistically managed and well-centered
process. The practical implementation proved that the monitored workplace is eligible for statistical
regulations. The proposed methodology has sufficient sensitivity and explanatory power to the factors
affecting the quality of production.
In case of inclusion of the proposed method and measuring equipment in the quality control system
production process is supported by a technical and economic evaluation in favor of the SPC method
with the following results:
- annual savings on non-conforming products is CZK 28,405,
- the investment in the new facility amounts to CZK 15,640,
- the economic return on investment is six months,
- the time required for control measurements is saved.

In the future, it would be possible to apply the SPC method in various modifications not only
for other types of spindles but also in other parts of JMA production. However, this extension is already necessary
count on the connection of individual measuring and monitoring devices to the company's CAQ
system containing a suitable SPC module.

Brno, 2008 - 59 - Ing. Petr Kopl

Page 61
VUT in Brno Department of Metrology and Testing
 Faculty of Mechanical Engineering THESIS Metrology and quality management

LIST OF SOURCE OF
[1] MICHÁLEK, Jiří. Statistical process control s (SPC) p R handbook No. ka . 2nd edition Prague:
Czech Society for Quality, 2006. 216 pp. Terminological Dictionary. ISBN 80-02-
01810-9.

[2] ČSN ISO 8258: 1994. Regulatory Shewharovy No the diagrams . 1994.

[3] TOŠENOVSKÝ, Josef, NOSKIEVIČOVÁ, Darja. Statistical methods for improvement

quality . Ostrava: Montanex, 2000. 362 pp. ISBN 80-7225-040-X.

[4] MLČOCH, Lubomír, SLIMÁK, Ivan. Ø Quality management and engineering metrology:
national university with no textbooks for engineering faculties of technical universities .
1st edition Prague: Státní nakladatelství technické literatury, 1987. 330 p.

[5] The analysis system of m measured measurement (MSA) p R handbook No. ka . 3rd edition Prague: Czech Society for Quality,
2003. xvi, 233 pp. Terminological Dictionary. ISBN 80-02-01562-2.

[6] KOPL, P. Evaluation of the control system of the production of fittings . Brno: University
Technical in Brno, Faculty of Mechanical Engineering, 2006. 83 pp. Thesis supervisor
doc. Ing. Jiří Pernikář, CSc.

[7] Materials of the company Jihomoravská armaturka spol. s ro

[8] Catalog of Mitutoyo Česko spol. sro

Brno, 2008 - 60 - Ing. Petr Kopl

Page 62
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

LIST OF ABBREVIATIONS A SYMBOL OF

A2 [-] coefficient for calculation
CL [mm] Central line
CL R [mm] Central line for the control chart for the span
CL X
[mm] Central line for control diagram for diameter
Cp [-] process capability indicator
C pk [-] process capability indicator
DN [mm] nominal diameter
JMA Jihomoravska armaturka spol. s ro
LCL [mm] upper control limit
LCL [mm]
R upper control limit for control diagram for span
 LCL X
[mm] upper control limit for the control diagram for the diameter
LSL [mm] lower limit dimension
MR [mm] sliding range
N inv [CZK] costs for investments in new equipment
N n1 [CZK] costs for non-conforming products
PN [mm] nominal pressure
Pp [-] process performance indicator
P pk . [-] process performance indicator
Qn [pcs / year] quantity of non-conforming products
QT [pcs / year] quantity of annual production
R [mm] subgroup span
R [mm] average span
Re [%] profitability
SPC Statistical control and regulation of the production process
T [mm] tolerance
X [mm] diameter of subgroup
X [mm] overall diameter
UCL [mm] lower control limit
UCL R [mm] lower control limit for the control diagram for the range
UCL X
[mm] lower control limit for the control diagram for the diameter
USL [mm] upper limit dimension
C [mm] sensitivity
d2 [-] coefficient for calculation
to [-] the number of subgroups used to calculate the overall average and the average
span
AND [mm] individual value
n [-] number of samples in the subgroup
with [mm] selection standard deviation
x [mm] arithmetic mean
^
p [mm] standard deviation estimate for X
C

p p [mm] total variation

^
p X [mm] standard deviation estimate for X

Brno, 2008 - 61 - Ing. Petr Kopl

Page 63
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

LIST P Ø ADDENDUM
P of Appendix No .1: The proposed measuring and monitoring equipment
P of Appendix No. .2: Control Chart - 1. Implemented
P of Appendix No. .3: Control Chart - 1st Implementation - Correction
P of Appendix No. .4: Control Chart - 2. Implementation
 Brno, 2008 - 62 - Ing. Petr Kopl

Page 64
VUT in Brno Department of Metrology and Testing
THESIS
Faculty of Mechanical Engineering Metrology and quality management

P Ø ANNEXES
 Brno, 2008 -4- Ing. Petr Kopl

Page 65

P Ø ANNEX NO .1 Suggested m measured ICI and for monitoring of devices

Page 66
Page 67

P
Ř
ÍL
O
H
AN
C
.2
R
eg
la
C
n
in
ia

-1
 p
he
en
ta

Page 68

P
Ř
ÍL
O
H
AND
C
.3
R
egu
la
C
n
in d
iagr

- 1. I

p
hem
en
tace

rava

Page 69

P
Ř
ÍL
O
H
AN
.C4
R
eg
la
C
n
in
ia

-2

p
he
en
ta