TABLE OF CONTENT CONTENT
PAGE
Abstract
2
Introduction
3
Theory
4
Resut
!
"iscussion
#2
Concusion
#4
Re$erences
#%
A&&endices
#'
ABSTRACT
1
The e(&eri)ent is about o&en oo& test and cosed oo& test *ith oad disturbance and set &oint test *as conducted+ There are three ty&es anaysis *hich is Contro Contro Process ,LIC ##-. Fo* Contro ,FIC 2#- and Li/uid Fo* Process ,FIC 3#-+ The resut o$ cur0e obtain $ro) usin1 three oo& anaysis+ A$ter that. the cacuated 0aues 0a ues o$ Res&onse Rate ,RR-. "ead Ti)e ,Td- and Ti)e constant ,Tc- by usin1 Tan1ent ethod+ There$ore. the o&ti)u) 0aues *hich is P and I is then cacuated usin1 ie1erNichos+ A$ter that. the 0aues o$ P and I obtain by usin1 ie1erNichos *as used $or cosed oo& test *ith oad disturbance d isturbance and set &oint test+ The 0aues o$ 5c. I and PB )ay be di0ided or ti)es by 4 to stabii6e the &rocess+ Finay. the resut o$ cosed oo& test cur0es obtain $or each anaysis as sho*n in resut and a&&endices+
INTRODUCTION
2
Contro is used in )any as&ect o$ o $ en1ineerin1 such as )echanica. eectrica and che)ica+ Process contro is one o$ the )ost used che)ica en1ineers that a&&ied to che )ica reactors. heat e(chan1er and )ass trans$er e/ui&)ent+ The e(a)&e *as sho*s *hen the o&eration o$ a e/ui&)ent )ust be consider *hen desi1nin1 because &ant ne0er o&erate in steady state+ The consideration )ust incude the disturbance to )a7e sa$e and &ro$itabe che)ica &ant desi1n+ In &rocess contro is basicay )eans re1uate the set &oint 0aue to )aintain the /uantity o$ &rocess substance+ There are sin1e 0ariabe &rocesses in *hich ony one 0ariabe is to be controed as *e as )uti0ariabe &rocesses )ay re/uire re1uation+ The e(a)&e o$ the &rocess 0ariabe are te)&erature. &ressure and $uid $o* rate and i/uid e0e and other &ro&erties o$ )aterias such as )ateria strea) co)&osition. density. &8 0iscosity. and s&eed+ The &rocess contro de0ices used in this e(&eri)ent is "istributed Contro 9yste) ,"C9o$$ "eta : E))erson and Fo(boro+ There are t*o oo&s *hich is o&en oo& and cosed oo& test+ The o&en oo& test &rocess can be used as sin1e &use and doube &use test+ For o&en oo&. the out&ut is )o0ed directy )ean*hie the cosed oo&. out&ut is )o0ed indirecty because o$ chan chan1in 1in1 1 in set set &oin &oint+ t+ The The &roce &rocess ss )o0e) )o0e)ent ent ar1 ar1er er i$ *e abe abe )eas )easur uree and and $act $actor or the the disturbance and set &oint+ That cause by disturbance+ The controer is set to auto)atic i$ the &rocess chan1ed into the cosed oo& syste)+ The controer set to )anua i$ the &rocess chan1e into o&en oo& syste)+ I$ the &rocess in auto)atic )ood. the &rocess *i stabii6ed to desired 9P+ The 0aue o$ P and I are inserted by usin1 cacuation 0aues o$ ie1erNichos+ 8o*e0er. the &oor co)bination o$ 0aues P and I *i resuts unstabe osciation+ 8ence. chan1e the 0aues o$ I. PB and 5c by di0ide or ti)es o$ 4 to contro the unstabe osciation to beco)e )ore stabe cur0es+
THEORY
3
;hen the controer in )anua )ode the &rocess *i chan1e into o&en oo& &rocess+ The ani&uated :ariabe :ariabe ,:- is increase to %<+ The res&onse cur0es obtain to $ind the 0aues o$ RR. Td and Tc+ Tc+ A the 0aues cacuated usin1 the tan1ent )ethod b y dra*in1 the tan1ent ine to stee&est &oint o$ the cur0e+ Fi1ure beo* sho*s the ste& in&ut ,=:- and the associated ste& res&onse cur0e *ith the tan1ent ine+
Fi1ure # RR is the res&onse rate *ith the unit o$ #>s+ # >s+ The 0aue o$ =t is the 0aue o$ Ti)e Constant+ The dead ti)e &eriod is $ro) the initia>start o$ o&en oo& test to the cross section bet*een tan1ent ine and initia steady state baseine+ The ti)e constant is )easured $ro) the intersection o$ initia steady state baseine and the tan1ent ine to the intersection o$ $ina steady state baseine and tan1ent ine as sho*n in Fi1ure # Then. the 0aues o$ Tc and Td cacuated as $or)ua beo*?
Td ,ti)e- @ Td ,en1th- ( b ,ti)e>en1th-
4
Tc ,ti)e- @ Tc ,en1th- ( b ,ti)e>en1th-
The units o$ Td and Tc are in second+ Then. the 0aue o$ Td and RR obtain cacuated by usin1 ie1erNichos tunin1 rue to 1et the o&ti)u) controer settin1s+
For nu)erica techni/ue. the dia1ra) sho*s the tan1ent anaysis o$ &otted ste& res&onse data+
The data o$ nu)erica )ethod *as coected by usin1 distributed contro syste) ,"C9-. su&er0isory contro and data ac/uisition ,9CA"A- and &a&eress recorder+ By usin1 the 3&oint centra di$$erence $or)ua. the so&e at to is as $or)ua beo*?
5
;here the P: is Process P rocess :aue :aue and to as Td+ Then. the RR. Td and Tc are de$ined as?
RESULT
CALCLATION OF OPTI P AN" I OF 9ELF REGLATING PROCE99 9ING TANGENT ET8O" , FIC2#6
P:i @ 2+2<
Θ = 78o
P:$ @ 2'+#<
a = 10%/17mm
=: @ ,43%-< @ %<
b = 15s/12mm
tan θ a
RR @
∆ MV b
/ 15 s / 12 mm
tan tan 78 10 17 mm
@
5
Td @ #)) ( ,#%s>#2))
@ +442>s
@ #+2%s
Tc @ 2)) ( ,#%s>#2))- @ 2+%s
I @ 3+33,Td@ 3+33,#+2%- @ 4+#'3s
PB @ ###+## ( RR ( Td @ ###+## ( +442>s ( #+2%s @ '#+% <
100
5C @
PB
100
@
61.50
@ #+'2'
CALCLATION OF OPTI P AN" I OF NON 9ELF REGLATING PROCE99 9ING TANGENT ET8O" ,LIC##7
P:i @ %+% <
Θ = 30o
P:$ @ None+ Le0e is non se$re1uatin1 &rocess
a = 10%/17mm
=: @ #<
b = 15s/12mm
tan θ a
RR @
∆ MV b
/ 15 s / 12 mm
tan tan 30 10 17 mm
@
10
Td @ 2)) ( ,#%s>#2))
@ +24!>s
@ 2+%s
I @ 3+33,Td@ 3+33 ,2+%- @ +32% s
PB @ ###+## ( RR ( Td @ ###+## ###+## ( +24!>s ( +32%s @ '+' <
100
5C @
PB
100
@
6.86
@ #4+%
8
CALCLATION OF OPTI P AN" I OF 9ELF REGLATING PROCE99 9ING NERICAL ET8O" ET8O" ,FIC3#-
P:. )3 >h +!23D2' +!23D2' +!D32 +!4%#2 +D#'#' +D3!!D3 +D42# +D42#
Ti)e. h?)?s #?33?3D #?33?4 #?33?4# #?33?42 #?33?43 #?33?44 #?33?4% #?33?4'
Ti)e. s # 2 3 4 % ' !
P:. < 24+#3 24+#3 2'+4' 2D+#% 3+%' 3#+2' 3#+'# 3#+'#
:< 3% 4 4 4 4 4 4 4
RR. #>s +233 +%2 +4# +2## +#%
−24.13 ( 2)( 1 )( 5 ) @ +233
24.46
RR # @
−24.13 ( 2)( 1 )( 5 ) @ +%2
29.15
RR 2 @
−26.46 ( 2)( 1 )( 5 ) @ +4#
30.56
RR 3 @
RR A A @ RR 2 @ +%2>s
9
Td . s
Tc . s
#
2+D
Td @ 3 2,#-
Tc @ 2,#-
[
[
29.15 −23.13 29.15 −24.13
− 24.13 29.15 −24.13 31.61
]
]
@ #s
@ 2+Ds
I @ 3+33 Td @ 3+33 ( #s @ 3+33 s
PB @ ###+## ( RR ( Td @ ###+## ( +%2>s ( #s @ %%+!<
5C @ #>PB @ #>%%+! @ #+'2'
#- FIC2# 9ELFREGLA 9ELFREGLAT TING GRAP8ICAL GRAP8ICAL ANAL9 ANAL9I9 I9 2- LIC## LIC## NON 9ELFREGLA 9ELFREGLAT TING GRAP8ICAL GRAP8ICAL ANAL ANAL9 9I9 I9 10
3- FIC3# 9ELFREGLA 9ELFREGLAT TING NERICAL NERICAL ANAL ANAL9I9 9I9 Tabe Tabe su))ary o$ &ara)eters and o&ti)u) controer settin1s $or PI controer based on ie1er Nichos 99TE
PROCE9
"eta :
9 Fo*
"eta : Fo(boro
LOO
PROCE99
PB
P BE8A:IOR FIC2# 9e$
5C
I ,s-
Td
TC
RR
,s2+%
,#>s+442
+24
2 +D
! +%2
,<'#+%
#+'2
4+#'2
,s#+2
'+'
' #4+%
% +32%
% 2 +%
#+!D
Le0e
LIC#
Re1uatin1 Non 9e$
Fo*
# FIC3#
Re1uatin1 9e$
%%+!
Re1uatin1
3+33
#
Tabe o$ Tunin1 Test. 9et Point Test and "isturbance Test HControer Gain ,5c- . Inte1ra Ti)e ,I- and Pro&ortiona Band ,PB99TE
PROCE9
LOO
9
P
"eta : "eta : Fo(boro
Fo*
FIC2
Le0e
# LIC#
Fo*
# F IC 3 #
TNING TE9T
9ET POINT
"I9TRBANC
TE9T
E TE9T
PB
5C
I ,s-
OL"
NE ;
,:%
,<
+4
#'+'
#+#)3>
#+%)3>
3 4+!
44+!
! #4+%
% +32
h h %)) %4)) 3+
4+
233+#
% #3+2
#)3>h
4#+%
#
#+%)3>
3#+%
h
DISCUSSION
The e(&eri)ent is to deter)ine the &ara)eters o$ RR. Td and Tc $ro) the cur0e usin1 Tan1ent Tan1ent ethod+ The o&en and cosed oo& test *as conducted $or FIC2#. LIC## and FIC3#+
11
Then. the ie1erNichos tunin1 rues *as cacuated by usin1 &ara)eters RR and Td cacuated be$ore+ For a $ast &rocess. PI $or)ua is used in this cacuation+
The 0aues o$ P. I and 5c are used $or cosed oo&+ The 0aues is set u& and &rocess run in Auto )ode+ The 0aues o$ 5c. P and I is enter into the syste) contro to 1et the resut+ For e(a)&e. the cur0e o$ FIC2# be1an to osciate acti0ey so the 0aue o$ 5c is di0ided by 4 and 0aue o$ I is ti)es by 4+ The &rocess turn into stabe a$ter the 0aues chan1es and is cosed to set &oint+ 5c and I are considered as the contro 0ariabes to 1ain the o&ti)u) controer settin1+ There$ore. the 0aue o$ &rocess to achie0e o&ti)u) 0aue is unde$ined+ 8ence. the 1ra&h ine set &oint and P: are the o&ti)u) 0aue under steady state condition+
For FIC2#. $ro) o&en oo& test. &ara)eters such as RR. Td and Tc Tc is cacuated by usin1 Tan1ent Tan1ent ethod+ The 0aues is +442>s . #+2%s and 2+%s res&ecti0ey+ Then. by usin1 tunin1 rues. the 0aues o$ PB is '#+%<. 5c is #+'2' chan1e into +4! and I is 4+#'2%s into #'+'%s $or tunin1 test in cosed oo& test+ The cur0e is $uctuation and not stabe+ 8ence. by di0idin1 and ti)es the 0aues o$ 5c and I turn the &rocess to stabe and so*er+ For set &oint test. the 0aues o$ set &oint is increases $ro) #+#)3>h to #+%)3>h+ For oad disturbance test. the 0aue o$ : increases $ro) 34+!< to 44+!< *hich is #< di$$erences+
For LIC##. sa)e )ethod is used by usin1 Tan1ent )ethod and ie1erNichos tunin1 rues to obtain the &ara)eters as abo0e+ 8o*e0er. $or the cosed oo& test the &rocess is osciate and stabe+ 8ence. no chan1e in the 0aue o$ PB. I and 5c+ The 0aues o$ RR and Td is +24!>s and 2+%s+ There is no 0aue o$ Tc because the &rocess in non se$re1uatin1+ The 0aues o$ PB. 5c and I is '+'<. #4+% and +32%s res&ecti0ey+ The set &oint increase in set &oint test $ro) %)) to %4))+ The 0aue o$ : aso increase by #< in oad disturbance test+
And asty. $or FIC3# the Nu)erica Tan1ent ethod *as used to obtain the resut o$ &ara)eters+ The 0aues o$ RR. Td and Tc Tc is +%2>s. #s and 2+Ds res&ecti0ey+ res&ecti0ey+ A$ter A$ter that. $or 12
cosed oo& the &rocess is osciate and not stabe+ Then. the 0aues o$ PB is ti)es to 4 $ro) %%+!< to 223+##< and I is ti)es *ith 4 too $ro) +%')in to +222)in+ Then the &rocess beco)e stabe and cosed to set &oint+ The 0aues $or) set &oint test is #)3>h to #+%)3>h+ The resut o$ oad disturbance test is 3#+%< to 4#+%<+ Throu1h the &rocess. *e can see i$ the &rocess need to chan1e the contro 0ariabe is the &rocess is not stabe or too so*+ so*+
CONCLUSION
The ai)s o$ this e(&eri)ent is to anay6e o&en oo& &rocess res&onse *hich consists o$ 0aue o$ res&onse rate ,RR-. dead ti)e ,Td- and ti)e constant ,Tc-. to deter)ine the e$$ect o$ 13
tunin1 rues ie1arNichos and obser0e the e$$ect chan1e o$ oad disturbance and set &oint durin1 the &rocess+ For o&en oo& test. the 0aue o$ Res&onse Rate,RR-. dead ti)e ,Td- and ti)e constant ,Tc- is deter)ine by Tan1ent Tan1ent ethod throu1h 1ra&hica and nu)erica anaysis+ Gra&hica anaysis is used $or FIC2# and LIC## and nu)erica anaysis is $or FIC3#+ These 0aues o$ RR and Td are then substituted into the Tunin1 Rues by ie1erNichos to be used $or the cosed oo& test in order to achie0e ach ie0e o&ti)u) contro 0aue in the &rocess+ Once the &rocess characteristics ha0e been identi$ied. the PI )ode is seected because it is used $or $ast &rocess res&onse+
For cosed oo& test. the 0aues o$ 5c and I then are inserted into the Gain and Reset section res&ecti0ey+ I$ the &rocess osciates and not stabe. the 0aue o$ 5c *i reduce by di0ide to 4 and increase 0aue o$ I by ti)es to 4+ These 0aue are a&&ied $or oo& FIC2# and FIC3# *hich is not stabe+ A$ter that. the &rocess beco)e stabe and near to set &oint+ Ne(t )o0e on to set &oint test+ The set &oint test chan1es $or set &oint abou t #< $ro) the ori1ina set &oint+ Then the oad disturbance test+ The test is chan1es o$ : about #< $ro) the current : unti the &rocess beco)e stabe+
REFERENCES
#+ Abdu Abdu A6i6 A6i6 Isha7. Isha7. J Anuar Anuar Ah)ad Ah)ad++ ,n+d+-+ ,n+d+-+ Application Application of Numerical Technique Technique in Tangent Tangent Analysis of a Discrete Step Response Data+ Data+ Retrie0ed $ro)
14
htt&?>>aabi+uit)+edu+)y>contronotes>PI"<2Tunin1>nu)erica<2anaysis<2o$ <2tan1ent<2)ethod+&d$ 2+ 5in1 9aud ni0ersity+ ,22-+ INTRODUCTION ,22-+ INTRODUCTION TO PROCESS CONTRO+ CONTRO+ Retrie0ed $ro) htt&?>>ho)e+)anhattan+edu>K1ennaro+)a$$ia>"ay<2#<25in1 <29aud+&d$ 3+ Abdu A6i6 A6i6 Isha7+ Isha7+ ,2#3-+ PI" PI" TNING TNING ,FN"AENT ,FN"AENTAL AL CONCEPT9 CONCEPT9 AN" AN" APPLICATION APPLICATION99- ,3rd ed+-+ ed +-+ 5ALA LPR. %'? IT PRE99+
APPENDICES
15
OPEN LOOP TE9T
16
CLO9E" LOOP TE9T ,TNING TE9T. LOA" "I9TRBANCE TE9T AN" AN" 9ET 9 ET POINT TE9T-
17