Organização Industrial I– Capacidade do Processo
Process Capacity
Main concepts used in the studies of capability or capacity of machines or processes:
Definitions
• Capability or Capacity: a measure that relates the actual performance of a machine or process to its
specified yield.
• Machine: discrete operations (lathe, press, ...) with minimal outside influences.
• Process: combination of people, equipment, materials, methods and environment.
With regard to variables, the capacity of a machine is a measure of the short-term influences that the machine
alone exerts on the variability of the product. It is usual to require as a minimum requirement ± 4.s (standard
deviation) are contained within specification limits.
This means that 99,999% of cases are within tolerance. With this requirement, it is intended to ensure that the
long-term process capacity target of 99.73 % (±3.s ) is met. We will then say that a process in this situation is
in Statistical Control.
For data in the form of attributes, the capacity index is a measure of machine or process compliance.
1
Organização Industrial I– Capacidade do Processo
Specification Limits
• Specification limits measure the permitted variability tolerance of an important
product or process feature.
• These limits (tolerance) are conceptually different and in practice regardless of
control limits, are defined by the process or product designer prior to manufacture;
• The control limits, on the other hand, are calculated values of the data observed on
the factory floor and are practical values;
• Tolerance measures what it should be, while control limits measure what it really is.
2
Organização Industrial I– Capacidade do Processo
Process Capability
Capacity Indices (variables):
If the histogram of the process presents the form of a normal distribution and the process
is in Statistical Control (subject only to natural variations) it is possible to evaluate the
capability or capacity of the process through the Process Capacity index.
These indices are a means of indicating the variability of a process or machine in
relation to the tolerance of its components or attributes.
Process capacity indexes (cp and cpk)
Cp - capacity indices for processes centered in the middle of specifications
Cpk - capacity indexes for non-centered processes
3
Organização Industrial I– Capacidade do Processo
Process capacity indexes (cp and cpk)
ÍNDICE DE CAPACIDADE DO PROCESSO
Processo com Média X, Índice de Capacidade do Processo
Desvio Padrão s e Especificação
Superior (USL)
Cpk= Mínimo de
e
USL - X X - LSL
Inferior (LSL) e
3s 3s
USL - X
Superior (USL) Cp =
3s
Inferior (LSL) X - LSL
Cp =
3s
For centered processes, the capacity index (Cp) is the distance between the upper specification
limit (USL) and the lower specification limit (LSL) divided by the natural variability of the
process equal to six standard deviations
4
Organização Industrial I– Capacidade do Processo
Process capacity indexes (cp and cpk)
The following expression is the standard deviation of the estimated process
with individual values, not the values in subgroups
For values in subgroups, you can calculate a more accurate
estimate of the standard deviation s using the following
table depending on the sample size n (letters X and R):
5
Organização Industrial I– Capacidade do Processo
Process capacity indexes (cp and cpk)
ANÁLISE DA CAPABILIDADE DE PROCESSO
1.33 < Cp (ou Cpk) SATISFATORIA
1.0 < Cp (ou Cpk) < 1.33 ADEQUADA
Cp (ou Cpk) < 1.00 INADEQUADA
The Cp index indicates the dispersion inherent to the process, while the Cpk index indicates the adjustment and
dispersion of the process.
In order to ensure that the process is in Statistical Control, the histogram should be accompanied by a control
letter demonstrating that the variation in the process is due only to natural causes.
In general, the variation in the processes is due to natural and unnatural variation. The process capacity study
reveals the causes of unnatural variation present in the process and what should be done to eliminate them in
order to obtain their true process capacity.
6
Organização Industrial I– Capacidade do Processo
Example of the calculation of Process Capacity (cp and cpk)
USL = 106.09 C
LSL = 92.03 C
Chart X / AM
(since it is a singular register, the moving amplitude is calculated by the
difference between two consecutive values)
The control limits are
CL
UCL
LCL
7
Organização Industrial I– Capacidade do Processo
Example of the calculation of Process Capacity (cp and cpk)
USL = 106.09 C
LSL = 92.52 C
Cp =
Number of out-of-specification samples
p(x>106,09) 0,001006
p(x<92,028) 0,000862
p(x<92,028;x>106,09) 0,001868 -> 0,187%
Em 10000 temos = 19 peças defeituosas
8
Organização Industrial I– Capacidade do Processo
Example of calculating process capacity indices (cp and cpk)
Upper specification limit: 1050g
Lower specification limit: 950g
9
Organização Industrial I– Capacidade do Processo
Example of calculating process capacity indices (cp and cpk)
Subgrupo X R
1 1012,15 27,35 LES= 1050
2 996,76 24,37 LEI= 950
3 1016,92 33,68
4 1028,4 46,11 LCS= 1036,14
5 987,58 58,7 LCI= 981,63
6 1024,18 37,62 S= 20,31
7 985,42 55,85
8 1009,06 60,77
9 1001,15 61,06 cp= 0,82
10 1006,04 67,72
11 1009,13 51,58 cpl= 0,67
12 1014,43 40,08 cpu= 0,97
13 1012,76 13,09
14 994,51 28,47
15 1041,01 43,83 cpk= 0,67
16 1008 57,98
17 1031,26 48,26
18 1012,43 55,09
19 1019,41 59,95
20 1009,71 61,34
21 1015,74 96,69
22 999,34 50,31
23 984,3 10,31
24 1020,66 50,23
25 1013,85 51,11
média 1010,17 47,66
média 1008,88 47,24
n= 5
A2= 0,577
d2= 2,326 10
Organização Industrial I– Capacidade do Processo
Machine capacity indexes (cm e cmk)
ÍNDICE DE CAPACIDADE DA MÁQUINA
Máquina com Média X, Índice de Capacidade da Máquina
Desvio Padrão s e Especificação
$%& − &%&
!" =
Superior (USL) 6)
Cmk= Mínimo de
e
USL - X X - LSL
Inferior (LSL) e
3s 3s
USL - X
Superior (USL) Cm =
3s
Inferior (LSL) X - LSL
Cm =
3s
You can calculate an estimate of the standard deviation s by: sˆ = capacidade _ estimada
8
Machine capacity indices indicate the long-term potential of the process if it is in Statistical Control.
The minimum requirement for the Cm or Cmk index is 1.33. As the quality improvement process is
implemented, this index increases progressively.
11
Organização Industrial I– Capacidade do Processo
Capacity indices
12
Organização Industrial I– Capacidade do Processo
Capacity indices
13
Organização Industrial I– Capacidade do Processo
Interpretation of the Machine Capacity Study
The interpretation of the capacity of a machine is made taking into account the following factors:
• If extending the straight line of adjustment of the points, none of the lines of the specification limits are
intersected, it can be considered that the machine conforms (99,999 % of the product conforms to the
specifications).
• If extending the straight line of adjustment of the points intersects any of the lines of the specification
limits, the machine is not complying with the specifications, so that there will be the production of non-
compliant product. This may be due to:
§ The average is not properly centered, so the machine must be readjusted (translation of the chart line
up or down).
§ Its dispersion is large, so the machine must be improved (make the chart lines more inclined).
§ In either case it is possible to estimate which % of product is out of specification.
• The value of the mean X (adjustment) will be the intersection of the vertical 50% with the straight of the
points.
• The estimation (8s) of the capacity used to calculate the capacity indices of the machine is given by the
vertical distance between the points at which the straight line intersects the vertical lines of both ends of
the scales %.
14
Organização Industrial I– Capacidade do Processo
Interpretation of the Machine Capacity Study
Application example
Specified tolerance:240.0 - 120.0 = 120.0
Average specified : 180.0
Estimated average : 185.0
Estimated capacity (8s): 225.0 - 145.10 = 80.0
Standard Deviation : 80.0 / 8 = 10.0
Capacity index: Cm = specified total tolerance / 6s
120.0 / ( 6*10.0) = 2.0
Cmk = minimum (240.0 - 185.0 ) / 3*10.0 = 1.83 and (185.0 - 120.0 ) / 3*10.0 = 2.17
= 1.83
15
Organização Industrial I– Capacidade do Processo
Specifications and control limits
In the production processes, the characteristics of each item produced must be taken into account.
By using the Control Letters and process capability studies, samples are at stake to try to establish
the distribution of individual items.
The specifications are established at the product and process design phases with the aim of
establishing:
• restrictions on individual items (of type +/-, greater than , ... )
• in the distribution of individual items establishing a distribution characterized by a central measure and
dispersion
• in a combination of the two previous situations
When making adjustments between the limits of the Control Charts and distribution histograms, the
specification limits must be addressed to the following aspects:
• The repairs must be made on the units specified in the specification.
• In case specification limits are applied to individual items, they must be compared with each item
individually and not with the X mean or control limits.
• In case the histogram of the process is contained between the upper and lower limits of the specification,
we can say that the process is satisfactory and can be controlled by the limits of the control charts. If, on
the other hand, the histogram exceeds the specification limits, this means that the process is not
satisfactory and corrective actions will have to be taken.
16
Organização Industrial I– Capacidade do Processo
Specifications and control limits
It is essential to understand that specifications are established based on customer and user
expectations and serve to determine whether or not each item conforms.
Whereas the limits of the control charts are calculated on the basis of the natural variation of the
process and serve to establish whether or not the process is in statistical control.
A process is under statistical control when the present variation is due only to random causes (also
called common or natural) and the special causes (also called assignable or unnatural) have been
removed.
In view of these considerations, it is possible to have four situations:
• Case 1 - The process is not in statistical control and there is the production of defective items
• Case 2 - The process is in statistical control and there is the production of defective items
• Case 3 - The process is not in statistical control and there is no production of defective items
• Case 4 - The process is under statistical control and there is no production of defective items
•
17
Organização Industrial I– Capacidade do Processo
Specifications and control limits
Case 1 - The process is not in statistical control and there is the
production of defective items
Action: Improve the process, review the specifications, consider
the possibility of inspection.
Case 2 - The process is in statistical control and there is the
production of defective items
Action: improve the capability of the process (control limits and
specifications are not harmonised), review specifications
and consider the possibility of inspection until this is
achieved. It is usually necessary to decrease the
dispersion of the process (better control), but in some
cases the process may present an acceptable dispersion
and be only decentered, so that it will be adjusted.
18
Organização Industrial I– Capacidade do Processo
Especificações e limites de controlo
Case 3 - The process is not in statistical control and there is no
production of defective items
Action: in this situation the process has sufficient capacity for
specification, but its efficiency should be improved by
removing the causes of unnatural variation and
establishing statistical control.
Case 4 - The process is under statistical control and there is no
production of defective items
Action: this is the desirable situation in which the process has
adequate capability and the control limits are harmonised
with the specifications. The degree of suitability is
measured by the Capability Indexes.
19
Organização Industrial I– Capacidade do Processo
Interpretation of the ability of the case on the basis of the Control Letters
The assessment of the feasibility of the process is carried out after the
identification and correction of the special causes of variation, when the control
letters show that the process is in statistical control.
The process capacity index can then be used as the basis of the process quality
improvement measurement system and allows quantifying its effectiveness. It is
necessary to confirm the capability of the process when recalculating the control
limits.
20
Organização Industrial I– Capacidade do Processo
Exercises
Exercise 1
Check that the process under analysis is in
statistical control by determining the
capability indices
21
Organização Industrial I– Capacidade do Processo
Exercises
Exercise 2
Check that the process under analysis is in
statistical control by determining the
capability indices
22
Organização Industrial I– Capacidade do Processo
Exercises
Exercise 3
Check that the process under analysis is in statistical control by
determining the capability indices
23
Organização Industrial I– Capacidade do Processo
Exercises
Exercise 4
A process is characterized by a normal distribution with an average of 52g and a standard
deviation of 1.5g.
Knowing that customer specifications are 50 +/- 4g.
Determine:
a) The expected percentage of units out of specification.
b) Cp and Ck.
c) What recommendations would you make to improve the process?
24
Organização Industrial I– Capacidade do Processo
Capacity indices (attributes)
For control charts p (% defective) and np number (% of defective units), the capability of the process is
expressed as a percentage of compliance with the specification.
Capacity = 1 - p
For control charts c (number of defects) and u (number of defects per unit), and are usedc!! respectively
u as a
measure of process behaviour.
25
Organização Industrial I– Capacidade do Processo
Capacity indices
Limitations
The techniques of analysis of the capacitability of processes and machines are subject to some
limitations, and the respective results should be considered as approximations. This is due to:
• Existence of a certain variation present in the sampling
• No process is always in statistical control
• No production always follows an exact normal distribution
26
Organização Industrial I– Capacidade do Processo
Indications for machine capability study
Pre-study requirements.
Before starting the study, make sure that the following steps are carried outsos:
• Determine the number of parts required for the study (usually 50, or more if it is suspected that the
distribution is not normal).
• Check that all materials to be used are approved to agree to the specifications.
• Check that the measurement and test equipment is calibrated with tolerances not exceeding 1/10 of the
specifications.
• Take precautions to ensure an uninterrupted production cycle, under normal conditions and with the
machine adjusted to the nominal value.
• In multi-equipment systems, treat each station as an independent machine in assessing its capacity.
27
Organização Industrial I– Capacidade do Processo
Data collection and structuring
At this stage of the study, it is necessary to collect and structure the data in order to allow its.
For this, for example, the elaboration of a histogram and the use of henry's straight using the
properties of the "Probability Paper" which transforms the normal distribution into a bell form is
used, in a straight line, significantly simplifying the calculations.
1. Measure machine production and record data sequentially.
2. Mark the scale to cover all readings (the same accuracy is usually used as the measurements). Mark
the specification limits.
3. Group the data into a histogram and check the normality of the distribution.
Calculate frequencies, cumulative frequencies, and % cumulative frequencies, starting from
bottom to top (the upper class limit is not included in the class range itself) .
28
Organização Industrial I– Capacidade do Processo
Drawn the Henry`s Line
Mark for each top class range value, the intersection with the cumulative percentage value on
the probability paper scale.
If the marked points are arranged in a reasonably straight line, draw between them a straight
line until it joins the vertical lines of the ends of the % scales (lines +/- 4 s).
If an accurate adjustment is not possible it means that the data do not have a normal distribution,
so that the causes of what happened must be investigated.
29
