2. 2
METHODS OF SPC
• ERROR DETECTION
(After the event)Tolerate Waste
ADJUST
• ERROR PREVENTION
(Before the event)Avoids wastage
Monitor adjust
Process Not OK
OK
InspectionMachine
Materials
Methods
Man
Process Not OK
OK
InspectionMachine
Materials
Methods
Man
Samples collected
/measured & analyzed
3. 3
What is Statistical
Process Control(SPC)?
• STATISTICAL: Taking of
measurements and arrangement
of those measurements in clear
pattern to allow prediction to be
made on performance.
» Data collection
» Data Presentation
» Data Analysis
» Data Interpretation /Drawing
Conclusions.
• PROCESS: Transformation
of input(4M & 1E) to desired
output through logical & inter-
linked Sequence of cause.
• CONTROL: Making
something to behave in way we
want it to behave.
» Define target
» Compare with target
» Measure actual performance
SPC is a technique of managing THE Quality Through error prevention.
(THE total performance of the process depends on Communication between customer & Supplier)
4. 4
Process
• About Process Performance :-
» By studying the process Output.
» By understanding the process itself
and its internal variability.
» Focus on process Characteristics.
• Action on the Process.:-
» Action on the process is frequently
Most Economical When Taken to
prevent the important
Characteristics (Process or output)
» Due to variation we
always problem, not
sleeping.
5. 5
Why use SPC?
¥ SPC is technique to control the Process .
¥ To know the special cause so as to Take corrective action.
¥ It gives Power to predict the behavior of the process .So as
to take corrective actions before actual happening (Gone out
of control).
¥ A more effective Method of working is to avoid waste by Not
Providing defective Parts in the First Place. This requires
everyone to concentrate on defect prevention Rather then
defect defection.
¥ SPC is a means by which to achieve the prevention Of
defects by highlighting situation when the output of process
is drifting.
¥ We can forecast things happening, Predict and take
corrective action I.e Preventive technique
¥ SPC is not superficial/ actual Req.. by customers.Generally
used on critical dimensions which directly may create
problem in our fitments and others.
6. 6
Process Control System
Statistical
Methods
THE way we work
/bending of resources Customers
Identifying
Changing, needs
and expectations
People
Equipment
Materials
Methods
Environment
Inputs Process/system Outputs
Voice of customers
Product/
services
Voice of
the process
1. Process control system is feed back system & SPCC is one type of that system.
2. It gives information about the performance,action on process,action on the output.
7. 7
BENEFITS OF SPC
¢ Parts can be easily assembled.
¢ Parts can be easily changed.
¢ Improves product Performance.
¢ Reduces Customer complaints
¢ Reduces rejection
¢ Increases Production
¢ Reduce Cost
¢ Reduce wastage of time
¢ Increase Confidence of
customers
• Tools of SPC
1 Data Collection.
2 Summarization of data
• Frequency
• Histogram
• Average Value
• Range
• Standard deviation
3 Control chart
• X-R chart
• Np and P chart
• C and U Chart
4 Process capability study
8. 8
BASIC CONCEPTS OF
STSTISTICS
• Variation is the LAW OF NATURE.As such No two
things are exactly alike.
• Variation in a Product or process can be measured.
• Things vary According to a definite Pattern.
• Whenever things of the same things are measured,A
large Group of the measurements Will tend to cluster
around the Middle.
• It is possible to determine the Shape Of THE
Distribution Curve for parts produced by any process.
9. 9
Nature of Variation
• There are two types of variation in a
process:
1 Due to large number of unknown causes
(Natural/Common Causes).
2 Due to special Causes(Assignable Causes).
10. 10
Assignable (Special)
Causes are Those Which:
• Are Unrelated to THE intended Design of
process.
• Do not Affect everyone.
• Are Unpredictable & Temporary.
Examples:-
Material Variation
Poor Maintenance
Electrical Power surge
New Methods/Procedures
Untrained Operator
11. 11
NATURAL (COMMON CUASES)
• Common causes of variations are inherently
part of the System/Process.
• Affect everyone associated with process.
• Predictable
Examples:-
• Procedure & Method used
• Education & Training given
• Machine Movement
12. 12
Preconditions for collecting
attribute data .
• Establish a clear reference standards (limit samples)
• Visual reference to demonstrate .”acceptable or non
acceptable”.
• Communicate the standards to appropriate personnel.
• Ensure that assessment Personal have appropriate
faculty & aptitudes .
• Train & develop Personal.
• Ensure correct environment for the task
• Verifying the gauging equipment.
• Encourage everyone to keep to the standard.
13. 13
CONTROL CHARTS
• It is a special type of trend chart
with limits specifically used to
assess and maintain stability of the
process
• This indicates whether the process
variation is natural and expected
(chance variation) or due to special
causes.
• Control limit theorem :it states
even if the outcome of the process/total
population is not normally distribution
but samples /average is always
distribution charts.
CL(Control limit)
OUT OF CONTROL PROCESS
UCL(UPPER Control lim
LCL(Lower Control limit)
1 2 3 4 5 6 7 8
ITEM NO.
5.25
5
4.75
cm.
14. 14
ADVANTAGE OF CONTROL CHART
• It gives a pictorial view of the process or
performance.
• It tells at a glance if a process is behaving
naturally or not.
• It is a tool to detect the pressure of assignable
cause in a process.
• It tells when a process should be corrected &
when it should be left alone.
15. 15
TYPES OF CONTROL CHARTS
• VARIABLE DATA: Which can
take up any value
example:Length,Weight,Temp,etc.
• Data collection is done by
measurements.
• TYPE OF CHART USED:
X-R CHART
• ATTRIBUTE DATA:
Which are classified as good
or bad,OK or Not OK
• data collection is
done by counting.
• FOR DEFFECTIVE
: P AND Np
• FOR DEFECTS : C
OR U
A DEFECT is a blemish flaw or non -conformity which causes an item to be rejected.e.g. dent, Paint
blemish, Scratch,Blow Holes. It is possible to have more than one defect on one single item.
A DEFCETIVE is a component or item which is unacceptable to use.
16. 16
Work on control charts
• Collect the data (25 Sub groups) at least.
• By interpretation of control chart we can find whether the
process is in control or out of control
» Points out side the control limit.
» RUNS: when 7 or more consecutive
readings are in one side they may be
above or bellow.
» TRENDS: A trend is described by seven or more
consecutive intervals that are continuously increasing or
decreasing.
» HUGGING: Center line,Control Limit(tracing and moving
causes.
CL
UCL
LCL
Out of control
17. 17
Benefits Of Control Charts
• Easier to understand by any person what is going on process.
• Help the process Performance Consistently,Predictability for
Quality And Cost.
• Allow the Process to Achieve.
• Higher Quality
• Lower Quality
• Lower Unit Cost
• Higher Effective Capacity
• Providing A common language for discussing the performance of
the project.
• Distinguish between Special & common Causes.
18. 18
STEPS FOR X-R CHART
• Determine the size of sample group.
• Design a suitable tally sheet.
• Collect data under standard operating
conditions(without adjustments)
• Calculate the average (x) and range(R,
the difference between the largest &
the smallest value in the group)
• Calculate the average of average-x
• Calculate the average range -R
• Calculate the control limits for for the x
chart using the formula
UCL=X + A2*R
LCL=X - A2*R
* Values of A2 for more common group size
are
• Calculate the control limits for the
R_Chart Using.
UCL = D3 * XR
UCL = D4 * XR
Values of D3,D4 for common sample group
sizes are given in the table.
• Divide the graph into two portions for 1) X
Chart & 2) R Chart
• Periodic observation are taken on X-axis and X
& R on Y-axis
• Select Appropriate scales and draw central lines
and control limits.
• Plot the observation and look out for indication
of the process going out of control or other
trends.
The steps involved in setting up control chart for variable X & R charts are:
19. 19
VALUES OF A2,D3 AND D4
NO OF
INSPECTION
TAKEN
A2 D3 D4
4 0.729 0 2.282
5 0.577 0 2.115
6 0.483 0 2.004
8 0.373 0.136 1.864
10 0.308 0.223 1.777
20. 20
PAND Np CHART
• P- Chart
C.L =P
P= Number of defectives
Total checked
UCL=P+3(P(1-P)/n)^2
LCL =P-3(P(1-P)/n)^2
Take LCL =0 when its value is negative
These charts are utilized for ‘Defective’ :
P Chart:when the sample is constant /Np Chart:When the sample size is varying
• Np- Chart
C.L =Np
Np= Number of defectives
No. of Samples Taken
UCL=Np+[3Np(1-Np)/n]^2
UCL=Np- [3Np(1-Np)/n]^2
Take LCL =0 when its value is
negative
21. 21
C And U - CHART
• C Chart:
C.L.=C
C =Total no of defects
Total Number of Samples(n)
UCL=C+(3C)^2
LCL=C-(3C)^2
Take LCL =0 when its value is
negative
• U Chart
These Charts are utilized for controlling ‘Defects’
C: Charts are used for products of constant size .
U: Charts are used for products of Varying size
C.L.=U
U =Total no of defects
Number
Checked
UCL=U+3(U/n)^2
LCL=U-3(U/n)^2
Take LCL =0 when its value is
negative
22. 22
Variation is Measured By:
Range Value
Standard Deviation(Sigma)
1. Range = Max.(value) - Min.(value)
Data: 10$$$,11,10,12,13,10.6
Range: 13 - 10 = 3
2. Standard Deviation ($) = {sum of(x-x)^2/n}^.5
23. 23
Measurement of things of
THE same kind
34.13% 34.13%
13.6%13.6%
2.14%2.14%
0.13%
0.13% X-1$ +1$ +2$ +3$-2$-3$
Area showing
total population.
24. 24
Analysis of Curve
• If a part is taken out at THE random,THE chances are that
68.16 Out of 100 Measurements (34.13%+34.13%) will fall
within two mid-section of THE graph.
• THE chances are that 27.2 out of 100 PCs. (13.6%+13.6%)
will fall in THE next sections.
• THE chances are That 4.28 out of 100 PCs. (2.14%+2.14%)
will fall into THE subsequent two sections.
• And 0.26 out of 100 PCs. (0.13%+0.13%) will fall in outside
two sections.
25. 25
WHAT IS PROCESS
CABABILITY?
It is the minimum variation the process can achieve.
It is defined by the sigma limits on either side of the mean(Target).It
is the “Inherent Capability” of the Process.
PURPOSE OF PROCESS CAPABILITY STUDY:
• Procedure for evaluating a process.
• Determine if the Process is capable Relative to its specifications.
• Determine the Centering & Stability of the Process.
METHODS OF FINDING OUT PROCESS CAPABILITY
¢ Frequency Distribution & Histogram
¢ Control Chart
26. 26
Cp & Cpk process capability
Indices
• Cp
-COMPARE PROCESS SPREAD THE
SPECIFICATION WIDTH
-IS USED WITH TWO SIDED SPECI.
-WANT TO FIT AT LEAST 8$
Cp = (Usl-Lsl)/6$
WANT Cp > 1.67
Useful measures for management to assess “health” of process.
Need to know that the process is stable (in control), the Process center(x) the process variability $(sigma)
Cp & Cpk MEASURE THE PROCESS CAPABILITY W.R.T THE CUSTOMER REQD.
Lsl Usl
6$
8$
3$
Lsl Usl
• Cpk
-ADDRESSES BOTH CENTERING AND
SPREAD V/S SPEC.
WANT CENTER AT LEAST 3-S FROM
SPEC.
Cpk = MINIMUM OF (Usl-X)/3S,(X-
Lsl)/3S
WANT Cpk > 1.33
27. 27
Methods of finding out
process capability
• FREQUENCY DISTRIBUTION &
HISTOGRAM
• CONTROL CHART
28. 28
Frequency Distribution
and Histogram
STEPS:
• Collect 50 consecutive samples From a running Process &
Record THE measured value of Quality Characteristics.
• Make frequency Distribution Chart and Histogram & check if
the Distribution is normal or Not.
• If distribution is not normal analyze the process Find out the
root cause of the problem and eliminate the root cause.
• Take fresh data of consecutive 50 Samples & repeat steps 2 & 3
stated above
• When distribution is normal calculate value of 6$,and also Cp
and Cpk.
• It can be characterized by
Location:-> Most frequently occurring value.
Spread :-> Span of values form smallest to largest.
Shape :->THE pattern of variation.
TIR:->Total indicating reading.
29. 29
• Definition:
THE voice of THE customers into
Engg.(D/G) Requirements & Measurements
Parts requirements & Measurements
Processes requirements & Measurements
Prod. Requirements & Measurements.
Quality Function
Deployment (QFD)
(Product planning)
(Per system identification)
(Process planning)
(Production planning)
Benefits:-
•Higher customer satisfaction & Excitement.
•Quick Transfer of knowledge to new Engineer's.
•Makes Good Engineer’s into excellent Engineer’s.