Control Chart Presentation

7/21/2019 Control Chart Presentation

http://slidepdf.com/reader/full/control-chart-presentation 1/56

Seminar-Workshop on

ISO 8258:

SHEWHART CONTROL CHARTS

Department of Science and

Technology IXRegional Standards andTesting Laboratory

February 26,#rand $storia %otel, &amboanga

by (anice T) *ng



+resentation Outine

!#eneral+rinciple

s

2Deelopi

ng'ontrol'harts

-*ut.of.'ontrolTrends



#eneral

!rin"ipes!



Intro#u"tion

/uality 'ontrol 0 IS* !12"2"

5$% Assurin& the 'uait( o) test an# "ai*rationresuts

5$%$+ The a*orator( sha ha,e 'uait( "ontropro"e#ures )or monitorin& the ,ai#it( o) testsan# "ai*rations un#ertaken$ The resultingdata shall be recorded in such a way thattrends are detectable and, where

practicable, statistical techniques shallbe applied to the reviewing of the results.

This monitorin& sha *e panne# an# re,iee#an# ma( in"u#e. *ut not *e imite# to. the

)ooin&:



IS* 32"3 3r$ Water A$ Shehart 4+%2/

He was the frst to

propose the use o

control charts as

graphical means o

applying the

statistical principles

pertaining to thesignifcance o

control in a

production process.



IS* 32"3 Statisti"a !ro"ess Contro

A statisti"a too that in,o,es inspe"tin& aran#om sampe o) the output )rom apro"ess an# #e"i#in& hether the pro"essis pro#u"in& pro#u"ts ith "hara"teristi"sthat )a ithin a pre#etermine# ran&e$

SPC answers the question of

whether the process isfunctioning properly or not.



IS* 32"3 6ariation

The variation concept is a lawof nature in that no twonatural items in any category

are the same.

The ,ariation ma( *e 'uite ar&e an# easi(noti"ea*e

The ,ariation ma( *e ,er( sma$ It ma(appear that items are i#enti"a7 hoe,er.pre"ision instruments i sho otherise

The a*iit( to measure ,ariation is ne"essar(

*e)ore it "an *e "ontroe#



IS* 32"3 6ariation 4Cate&ories/

There are three "ate&ories o) ,ariation inpie"e part pro#u"tion:

!) 4ithin.piece ariation: Sur)a"e

2) +iece.to.piece ariation: Amon& pie"espro#u"e# at the same time

-) Time.to.time ariation: 3ieren"e inpro#u"t pro#u"e# at #ierent times o) the#a(



IS* 32"3 6ariation 4Sour"es/

Sour"es o) 6ariation in pro#u"tionpro"esses:

Input

OutputProcess

MaterialsTools

OperatorsMethods

MachinesEnvironment

MeasurementsInstruments

Human InspectionPerformance



IS* 32"3 6ariation 4T(pes/

6ariation ma( *e #ue to chancecauses 4ran#om "auses/ orassignable causes$

When only chance causes are present,t hen the process is said to be in a

state of statistical control. The process is stable and predictable.



#lossary o) Terms

$ssignable cause, n.—an identifable actor that

contributes to variation in quality and which it is easible todetect and identiy. Sometimes reerred to as a special cause.

'hance cause, n.—an identifable actor that ehibits

variation which is random and ree rom any recogni!able

pattern over time. Sometimes reerred to as a common cause.

Lot, n.—a defnite quantity o some commodity produced

under conditions that are considered uniorm or sampling purposes.

Sample, n.—a group o units, or portion o material, ta"en

rom a larger collection o units or quantity o material, whichserves to provide inormation that can be used as a basis oraction on the larger quantity or on the production process. #ay be reerred to as a subgroup in the construction

o a control chart.



#lossary o) Terms

Subgroup, n.—one o a series o groups o observations

obtained by subdividing a larger group o observations$alternatively, the data obtained rom one o a series o samplesta"en rom a series o lots or rom sublots ta"en rom a process.%ne o the essential eatures o the control chart method is tobrea" up the inspection data into rational subgroups, that is,

to classiy the observed values into subgroups, within whichvariations may, or engineering reasons, be considered to be dueto non&assignable chance causes only, but between which theremay be di'erences due to one or more assignable causes whose presence is considered possible. #ay be reerred to as a sample

rom the process in the construction o a control chart.

5nit, n.—one o a number o similar articles, parts, specimens,

lengths, areas, etc. o a material or product.

Sublot, n.—an identifable part o a lot.



#lossary o) S(m*os

nsubgroup or sample si!e, that is, the

number o units or observed values in asample or subgroup

number o subgroups or samples underconsideration

Xobserved value o a measurablecharacteristic$ specifc observed values aredesignated ( ), ( *, ( +, etc.$ also used to

designate a measurable characteristic

average arithmetic mean-$ the sum o then observed values divided by n

average value o the subgroup averages.

nsubgroup or sample si!e, that is, the

number o units or observed values in asample or subgroup

number o subgroups or samples underconsideration

Xobserved value o a measurablecharacteristic$ specifc observed values aredesignated ( ), ( *, ( +, etc.$ also used to

designate a measurable characteristic

average arithmetic mean-$ the sum o then observed values divided by n

average value o the subgroup averages.



#lossary o) S(m*os

True mean value o the process.

7e

#edian value o a subgroup . or a set on numbers ( ), ( *, ... ( n arranged in order

o ascending and descending magnitude,

the median is the middle value in the seti n is odd and the arithmetic mean o thetwo middle numbers i n is even.

/verage median value o the subgroups.

RSubgroup 0ange1 di'erence between thelargest and smallest observations o asubgroup.

True mean value o the process.

7e

#edian value o a subgroup . or a set on numbers ( ), ( *, ... ( n arranged in order

o ascending and descending magnitude,

the median is the middle value in the seti n is odd and the arithmetic mean o thetwo middle numbers i n is even.

/verage median value o the subgroups.

RSubgroup 0ange1 di'erence between thelargest and smallest observations o asubgroup.



#lossary o) S(m*os

#ean o 0 values or all subgroups

7

R

the absolute value o the di'erence o twosuccessive values plotted on a controlchart.

/verage moving range o n&) movingranges rom a series o n values

s Standard deviation o the sample1

#ean o 0 values or all subgroups

7

R

the absolute value o the di'erence o twosuccessive values plotted on a controlchart.

s



#lossary o) S(m*os

/verage value o the standard deviations

o the sub&samples.

8 True value o the standard deviation othe process within the subgroup.

2stimated value o the standard deviationo the process within the subgroup.

p3roportion or raction o nonconorming

units in a subgroup1

/verage value o the standard deviations

o the sub&samples.

8 True value o the standard deviation othe process within the subgroup.

2stimated value o the standard deviationo the process within the subgroup.

p



#lossary o) S(m*os

/verage value o the ratio or ractionnonconorming1

np

4umber o nonconorming units in asubgroup.

4umber o nonconormities in a subgroup.

/verage value o c values or all subgroups.

np4umber o nonconorming units in asubgroup.

4umber o nonconormities in a subgroup.

/verage value o c values or all subgroups.



#lossary o) S(m*os

u 4umber o nonconormities per unit in asubgroup.

/verage value o u values1

u4umber o nonconormities per unit in asubgroup.



2Deeloping

ControCharts



Nature o) a

'*9TR*L'%$RT



'*9TR*L '%$RT 3e0nitions

A control chart 4or pro"ess "hart or'uait( "ontro "hart/ is a &raph thatshos hether a sampe o) #ata )asithin the "ommon or norma ran&e

o) ,ariation$It is a &raphi"a metho# o) presentin&an# "omparin& in)ormation *ase# on

a series o) sampes representin& the"urrent state o) a pro"ess a&ainstesta*ishe# imits a)ter "onsi#erin&

the inherent ,aria*iit( o) the pro"ess$




/ control chart has upper and lower control

limits that separate common rom assignablecauses o variation.

A pro"ess is out of control hen a pot o)

#ata re,easthat one or more sampes )a outsi#e the"ontro imits$




The x a:is represents sampes taken )rom the

pro"ess o,er time$ The y a:is represents the 'uait("hara"teristi" that is *ein& monitore# 4e$&$oun"es o) i'ui#/$




The centerline 4CL/ o) the "ontro "hart is the mean. ora,era&e. o) the 'uait( "hara"teristi" that is *ein& measure#$

The upper control limit 49CL/ is the maimum a""epta*e,ariation )rom the mean )or a pro"ess that is in a state o)"ontro$ 4;< a*o,e )rom CL/

The lo;er control limit 4LCL/ is the minimum a""epta*e

,ariation )rom the mean )or a pro"ess that is in a state o)"ontro$ 4;< *eo )rom CL/




The ;arninglimits 49WL an#LWL/. hi"h are

aso "ae# 2<"ontro imits.ser,es as a arnin&o) a situation ith

imminent oss o)"ontro hene,ersampe ,aue )as*e(on# this imit$



'*9TR*L '%$RT T(pes o) Errors

Type I error o""urs

hen the "orrespon#in&pro"ess remains un#er"ontro *ut a point )asoutsi#e the "ontro imits

*( "han"e$ An incorrectconclusion that the process is out ofcontrol is made incur

to !nd non"existent problems.Type II error o""urs hen the "orrespon#in&pro"ess is out o) "ontro *ut the &enerate# point )asithin the "ontro imits *( "han"e$#ncorrect conclusion$ the process is instatistical control



Eements o) a

'*9TR*L'%$RT



Data 'ollectionSection

The "ounts ormeasurementsare re"or#e# in

the #ata"oe"tion se"tiono) the ControChart prior to*ein& &raphe#$



Title

=rie>( #es"ri*es

the in)ormationhi"h is#ispa(e#$



Legend

A ist that epainsthe si&ni0"an"e o)ea"h "ooure# ineson the "hart$ It asoin"u#e ,aues o)

important statisti"su"h as mean an#stan#ar# #e,iation.an# tra"ea*iit( o)the "ontro sampesuse#$



<ertical or =.$:is

This ais ae"ts thema&nitu#e o) the#ata "oe"te#$ The ?-ais shos the s"aeo) the measurement

)or ,aria*es #ata. orthe "ount 4)re'uen"(/or per"enta&e o)o""urren"e o) ane,ent )or attri*ute

#ata$



%ori>ontalor X.$:is

This ais #ispa(sthe "hronoo&i"a

or#er in hi"h the#ata ere"oe"te#$



'enterline

This ine is #ranat the a,era&e ormean ,aue o) a

potte# #ata$ Theupper an# oer&raphs ea"h ha,ea separate"enterine$



'ontrolLimits

These are set at a#istan"e o) ;<a*o,e an# ;< *eothe "enterine$ The(in#i"ate the

,ariation )rom the"enterine an# are"a"uate# *( usin&the a"tua ,auespotte# on theContro Chart$



4arningLimits

These are set at a#istan"e o) 2<a*o,e an# 2<*eo the

"enterine$ The(are "a"uate# *(usin& the a"tua,aues potte# on

the Contro Chart$



$cceptableLimits

These are set at a#istan"e o) +<a*o,e an# +<*eo the

"enterine$ The(are "a"uate# *(usin& the a"tua,aues potte# on

the Contro Chart$



7?$S5R$@L? Ch i i



7?$S5R$@L? Chara"teristi"s

A ariable is a pro#u"t "hara"teristi" that

"an *e measure# an# has a "ontinuum o),aues 4e$&$ hei&ht. ei&ht. or ,oume/$continuous value - there are potentia( anin0nite num*er o) ,aues *eteen a#@a"ent units$

4e$&$. *eteen + an# 2 se"on#s: +$+. +$+. +$+se"on#s. et"$/

An attribute is a pro#u"t "hara"teristi"

that has a #is"rete ,aue an# "an *e"ounte#$discrete value - the ,aue "han&es in 0e#amounts. ith no interme#iate ,aues possi*e$

4e$&$. BNum*er o) "hi#ren in a )ami(/

'*9TR*L '%$RTS ) 6 i *



'*9TR*L '%$RTS )or 6aria*es

!) 'ontrol chart for mean AB and rangeARB or standard deiation AsB

2) 'ontrol chart for indiidual

obserations AXB and moing rangeA7RB

-) 'ontrol chart median A7eB and

range ARB

These #es"ri*e pro"ess #ata in terms o)

dispersion 4,aria*iit( )rom sampe to sampe/an# location 4a,era&e o) the pro"ess/$ =e"ause o)this. "ontro "harts )or ,aria*es usua( areprepare# an# ana(se# in pairs: one "hart )oro"ation an# one )or the #ispersion$

'* * ' S



'*9TR*L '%$RTS )or ean an# Ran&e1Stan#ar# 3e,iation

!) 7ean AX barB 'hart

It is use# to monitor "han&es in the meano) a pro"ess$

and and and

where,

LetCs Reie; ?:ample 6)!

For kG 5. an# krepresents thenum*er o)repi"ates persu*&roup

'*9TR*L '%$RTS




!) 7ean AX barB 'hart

Another a( to "onstru"t the "ontroimits is to use the sampe ran&e as anestimate o) the ,aria*iit( o) the pro"ess$In this "ase "ontro imits ou# *e

"onstru"te# as )oos: and

where, /* is a actor obtained rom Table 5 o /ST#

#anual on 6uality 7ontrol o #aterials.

LetCs Reie; ?:ample 6)2

'*9TR*L '%$RTS ) # # # i i




2) Range ARB 'hart

Ran&e "harts monitor the #ispersion or,aria*iit( o) the pro"ess$ The metho# )or#e,eopin& an# usin& R-"harts is thesame as that )or -*ar "harts$

789 and

where, :; and :+ are actors obtained rom Table 5

o /ST# #anual on 6uality 7ontrol o #aterials.

LetCs Reie; ?:ample 6)-

'*9TR*L '%$RTS ) # R 1S # # 3 i i




-) Standard deiation AsB 'hart

It aso monitors the #ispersion or ,aria*iit(o) the pro"ess$

789 and

where, :; and :+ are actors obtained rom Table 5

o /ST# #anual on 6uality 7ontrol o #aterials.

LetCs Try This $ctiity

'*9TR*L '%$RTS ) Att i* t



'*9TR*L '%$RTS )or Attri*utes

!) +.'harts

are use# to measure the proportion o)items in a sampe that are defectie$4e$&$. proportion o) *roken "ookies in a *at"h an# theproportion o) "ars pro#u"e# ith a misai&ne# )en#er$/

2) '.'harts"ount the a"tua num*er o) defects$4e$&$. the num*er o) "ompaints )rom "ustomers in amonth. the num*er o) *a"teria on a petri #ish. or thenum*er o) *arna"es on the *ottom o) a *oat$/

defectie a unit that )ais to meet a""eptan"e"riteria #ue to one or more #e)e"tsdefects a )aiure to meet one part o) ana""eptan"e "riteria



-*ut.of.

'ontrol Tren#s



Rule !

An( sin&e point*e(on# ;< imits4or *e(on# 9CL an#LCL/$

ReEuires immediateaction

#easurement shouldstop until the cause othe action limit

violation is identifedand corrected.



Rule 2

To "onse"uti,epoints *e(on# 2<imits on the sameor *oth si#es o)the "enterine$7easurementshould stop until thecause o the actionlimit violations are

identifed andcorrected.



Rule -

To out o) threepoints in a ro*e(on# 2< imitson the same or *othsi#es o) the

"enterine$7easurement shouldstop until the cause othe action limitviolations are

identifed andcorrected.



Rule

Ei&ht points in aro on one si#eo) the "enterine$

There has been ashift in the mean:

wor" can continue,however the causeo the shit needs tobe identifed and thelimits re&set i <ustifed.



Rule "

Si points in a rothat are mo,in& aa(or toar# the"enterine ith no"han&e in #ire"tion4aso "ae# as Btren#rue/$

The measurements areshifting ;ith time:needs investigation, sincethe action and warninglimits will eventually beeceeded.



Rule 6

Fourteen"onse"uti,e pointsaternatin& up an##on 4satoothpattern/$Suggests a cyclicalchange inconditions: may bedue to two di'erentoperators.$



Rule 1

Four out o) 0,epoints *e(on#+< imits on the

same si#e o) the"enterine$This may indicate amean shif t$



Rule 3

Fi)teen points in aro ithin the +<

imits on either si#eo) the "enterineaso "ae#

BstratiGcationrule$

This may arise ater the procedure has beenoperated or some timeand may require the

limits to be re&set.



Rule H

Ei&ht "onse"uti,epoints outsi#e the+< imits on*oth si#es o) the

"enterine aso"ae# Bmiturerue$

This may indicate amean shif t$



Than =ou



?

@ut ;aitJ

LetJs Anser A"ti,it( 8K

Control Chart Presentation

Documents