Sei sulla pagina 1di 92

Distributed in A~.,:- :·".

&Sia by:
INSTITUTE OF INSTRUMENTATION AND CONTROL AUSTRALIA, INC.

CONTROLLER TUNING
AND
CONTROL LOOP
PERFORMANCE

Is E c o No E o I TI o NI

A Primer
By David W. St. Clair

PUBLISHED BY:
STRAIGHT-LINE CONTROL COMPANY, INCORPORATED

,,
r • '""1!t;m s.;0.2. .. c
'
··.•..
....
-·.>r.··,J(','
··f~.7::~~r,~-· ·""·
.•. ··~~<~:::,:, . : . . .::... . .
-~..._~,_
>
. ..,. . ,.. . .

,' .
, :•
. : ".'
. . . ". '
··*-·~..
.·.... .
,'i;'• .
·>,._
...,.

BACKGROUND
..
..
•,,
f

THIS BOOKLET WAS ORIGINALLY ISSUED IN 1983 AS AN INTERNAL REPORT IN
THE DUPONT COMPANY TO HELP ENGINEERS AND TECHNICIANS, WHO HA VE
NO SPECIAL TRAINING IN FEEDBACK CONTROL, UNDERSTAND THE BASIC
',

.
.
;
CONSIDERATIONS AND LIMITATIONS. IT HANDILY BROKE ALL RECORDS AT

.
.
. THE DUPONT COMPANY FOR NUMBER OF REQUESTED COPIES (OVER 1200)
WHEN ISSUED. THAT REPORT WAS RELEASED TO THE PUBLIC AND
.

PUBLISHED IN 1990. IT SUBSEQUENTLY SOLD OVER 16,000 COPIES. IN 1992


DUPONT REQUESTED A MANUAL TO BE WRITTEN SPECIFICALLY FOR
TRAINING, EXPANDING ON THE ORIGINAL REPORT. THIS SECOND EDITION IS
BASED PERHAPS 80% ON THAT TRAINING MANUAL.

ABOUT THE AUTHOR

THE AUTHOR RETIRED · AFTER 40 YEARS OF PRACTICE IN THE FIELD OF


INSTRUMENTATION AND CONTROL IN THE PROCESS INDUSTRIES (8 YEARS
WITH EASTMAN KODAK AND 32 YEARS WITH DUPONT.) HE TOOK IN 1947
WHAT HE UNDERSTOOD TO BE THE FIRST COLLEGE COURSE OFFERED IN THE
THEORY OF FEEDBACK CONTROL, A CHANCE EVENT AT MIT THAT STARTED
HIS CAREER IN THE FIELD. HE ARGUABLY HAS APPLIED THE SCIENTIFIC
METHOD TO SOLVING CONTROL PROBLEMS IN THE PROCESS INDUSTRIES
LONGER T ANYONE, OR AT LEAST THAT WAS PROBABLY TRUE WHEN
HE RETIRED IN 1987. HE HAS BEEN EXPLAINING THE CONCEPTS TO THE NON-
SPECIALIST FOR MOST OF THAT TIME. HE RELISHES THIS OPPORTUNITY TO
SPREAD THE WORD TO A LARGER AUDIENCE.

HENCE THIS PUBLICATION

......... -· .. .

... .

.. .
.,
.... . . .. . . .

. . .. . . ,
:
:.

. . .: ,, •
..
. !· . ..
- . • • . ..

. .
.
. . . . . \

I I .,
!",.
.

.. . . ,. ...... . ... . . .. . . .
. .
:
.
·, . . .
,
. . •/

. ' . ' . .

.
.A'Prlmer
. ,
.
. ~
.
.
-~ . . ~

.. •:·. .
.. . .

. '
. . ,....

PUBLISHED BY:
ST IGHT-LINE CONTROL CO., INC.

. '·.
·'
..

- ..

··•.': ·.'.,,:>:.~·,·· . ~.
·......,. : .... : , ·: .. ·, . ;,;,·

·,
., ·,

Copyright© 1989, 1993 by David W. St. Clair. . . . ..

Straight .. Line Control Company, Inc,


. . . .. . .... '~· . < ·~·-"
::::::. .
....... ..- ,>. ;
',•.£
,: ., . '··:·,.. :.:,
'
../ ..
•• ~ .
•'::> ·.. ..,,... ·.·:.._.r
.... : .
• •• ~<.
.........
;.
...........
·.,
........
. ·.:.~:~..:·
.,
• ··! .,. ~:-:: '·

... ,: ,.
. ', ·.
.
.. ,. .· . :,·:, .
: ::, ;.
' . . .. , ', .'·
,, ·.. : ·: ....... ; ...

. '
.. .

,....
: . ·.

.' .

·'

. ••' . ·;,

. .•:
. : ....;
;. ...... ,.:·,, . <
,, '·
·:1,.

.... : .:.,,,~., . t.
. , :·. ·.. :,., ·:.;
.,
'

.' ·..
:,
.-,
'. ;. ··"·
-s-

without . :,~

.'
·,...
~ ..
. ,.
.. .
.. .,· ·'.
'
; , ..· ,. .,..
/.
,: .
..... "···.. ,
·~t ..
s:

• >
·~.
·:.:..... :::.:
• • ••

" .. /
,·..
.:.,,
. ;,..
:~
r
,:•

Library of Congress Catalog Card: Nilinber: 93-93731 . . .. ' . .. . .


..,
'

·.. ,

.. -:
.'

FIRST EDITION . ~-
,·.:.. ,
~..
'j;;

First

·,,•, :: . . Tfiird
Forth t-

.' ...

SECOND EDITION
First Printing: December, ·1993
Second Printing: January, 1995, Chapter 8 expa
. rr .
C&l'I
. .
·. .. . ..•


.


. \~·· ..• ·. ·····tt•'. •• , ...• ,_:·. ·.

..
. · .. :: ·:·,:'. >.
.",... :/· -, f,.~.·.':,·

Third Printing: August, 1995 ··. ~ '.

Fourth Printing: June, 1996


,...

....... ,: .
··,.~· ·,\t'

..·~, .
·.y;i;: .
..; .....
-: ·,
'• ....
. .
.. .. ·. , '·
·,

..
. . ·.~ . .
,:;::: ..
-. '. ~ .:/.... ·\·:;:·~~. :.f;, ..... /·i: \.::-,~·:: . ··.,· .'
.,
·'· . •'

.....
. '
:
'·,:
"· .
'. ~;:J<, ·:-:,;:.:(: .: '?5 ,.
... ,. . .,,·
: :-: . .;.,.
. .:.,; ",
.. .. . .. ,
'·.
J
. ·. :, .• ·;.·i:?

;, .
.. ·.·, . ' .. ·...,, ;. '. ...... ·:. . ·:...... ,·: ~·

;,·

·,
'. ·,. ·,. ,.

-,
.
., ... ,'
·., ,....;...
.. . :.· .
. . .'·
. .
.
.... , ...
: .'

... .......
.... . .<:.
• ,. ·t... .,
.
. ,..·
. ·.·····.:}}~~
. :.. ·~· :.
-:

......-.. ~. . . ..
.
.. ->'.
. ._
..
. •
···., . ~ , ·.,.
... ..>
-·.. . ." .
-. .. .,.

..;',·. .,
....· .. .... . . ' ., .
..
. _.. . ..
.·. . .. ., ·.
>
. ···:. . .. .

.·. ·.....··
..
. ·,·
... . .·: ".
. .. ,
.. '· . ' ...
::~·.A:•
.:,,
:·.. ::
·,. .
... ,,
• '· ..•.
.., : ,.. :•, "
. ..
.. . :, ..•. ·· . ··' . :· ......:..... .. . .. ,
, ..
.
. •... . '.·
. .,.. ... . ·...
.
. t...•• •:
. ·.. ·~
·.~ .
r
'-·l'BR:··1, ·OETTING; ST AR TED ·1 ~ :··..
\·:
.: ,
....·· :
..
·:. .:·,
. \: :·
••
-
..

• ~·
·'·.· ;.~ ,..;· : ....•
., ·-·· t .,
. ....'.•
. .... .. .. '· ,.
·,

..~ . i,-. r-. :._· ·.... :.. . ·. -~·:· . :,:!'•,,•·· :-... ·· ;_., ·:. ··: ..
.
Science. or Art. . ·. '·.· . . .
,.

.. 2 ,. ,,.:.
·.. ::
f:,,··-,
..
· :. . .·.
..
:· , .
:.' .,
., ..
.:
;\• : ..
~: ....... .. : ,.. •·.. . ·-.· ·'· ..
·;:. ·.: . .,, ·. . .' .. "
'·•... ·..•... , ·:: ..:.... ~ -'. ,..{
·'.·
, .,.. ... :. .·
••
.. ..·,
'···..·
,
. •.
.•····~... .,.. . . •.. :··
History .· 3:•·.·
. . . .
..
i ".
. :·· ·., ;.;...
.. .
...
.
·: ...·., ...·. i• ..
.
·,
/:
.
·'·

. 4· .... ·.
The :4ogl.lg.,::·of Q<mt,ol, ':· .· •· .· ..
... .
.·. ·~ . ·.
•..
• .1 ·••

.. :: .....·.:'' ·. . . . · ··>.· ·i:.{ .


.lf.,..
< .

. .
pf
.

.. . .
Tenn.'-µ.ol9gy~~f;.~4-.;P~sqripiiqn
. ·. . .
Controller
..
Settings ·. ·.. . . . . . . . ·. . . . .
.
·.
..6.
...
..
'

.,
'; ...
·',, :
. ··..r; .:: r- . . •..... '·
..·. :.·. .: .,

. Proportionel J\~;o.,j ;_ S · ..
••
. .. ..
. . :, ' . ..

Integfal Action . . -8 . . . . ...


... :,
.
......
~ .. '·.
. ..
. ,.
..
.:·· ~ ! . ..•. •··
."(..
,. •:. .......
:, .
_ ....
,
..


: . •• •
- ...~
.· •·;:. '>: . -: .·,:
.. . ..·... ., ..
..: ~..... · ..:f.~;.·....
, . . ..
· ,:r~

-;
. . . ::
.. • .(·

· ;Reset Win4~P 9 ...• ,.


,.... -: >.. ..~
••, •• ,v~· •'
-

. .. .'
.·.
.. .
:· ••• t • • • • ••• • • . :. . ,.

Derivative Action··. · 11,. . ... ,. ·.


; /' >) {;,\•.
. . . .. ~.. .
.<

. ·..~:·.:.·:. :.• ..
·' .
...
: ·,.. .
,,.:
·. ·,.

Filter·, Time 13 · ;
:
.,
., . . . ..•.
. .· ·.: .

•<
·, ·...·· ·~·(·· ... ~. .:;.. .
·~ .,
·- ..
. . . . .

··F·ilter· Time .attd Derivative Action 14


•.. ..:, . .. ;· : ..·' ,.. ....
: .• -c... ., . .. . .. · . ":. ,;
,~· ':"·~·. . : .: .....
".
·.·
.... ...,. .. . .., .. .. ,...'..: . .. :.~
..... $
:.. . .
. . ...• ,.•'
'
·~··

.TER·. ··--2 '


.. •

c .
.
. . ..
.
.
, :·.
.
.
.

.
G RULES AND PROCED is . ·.
: ..
t
·:.~,'
~
.• <
·.,. ::
.
.....
,

. .~
.... ,.
. ,· . ·.

PreparatiOll ;.16 ..···. .


to ?
·.•.. ~··.. .
. ., ..
. .....
..
. .
Closed-loop Tuning: What Do 17 1

. ..'
.~
Closed-loop ~Tua~g:··Howto•-Do It 18 ·. .' ·: ~ : .
, ...
:'. . ; ...
··~ ., ··> •:. ·,; .. : ..
.
....
: ..
•• ;,;~ •••• <·
~:
.
,,~ ,;. .,

·OpeD·loop Tuning: What to-Do · 21 .:· ( :··

- .• ...·..
Open~k,o.p.·Tuning:;::llow·toDo It . ·23 . . .. . . . . .
. .

When.t~ l)seOpen~J.<>o.p.Tunin9:: · .2~


\ . .
• ..

Wh~n to Not;Tuneby the Rules 26


Tuning Rules Overview 28
·..,>

c G OBJECTIVES AND EXPECTED LOOP . .. . ..

31
..
Response to-:Cy~l.ic Upsets 34
;.

Factors Affecting·:~e.Natural Period


. . . .. . ~-."::. '::(:· .<: :: ...
36 . .
. . ,. .
•• ! . .
; . ,..

c TER 4··.LAGS
' .
AND GAINS·' BUT MOSTLY . . ..
. ';. ..
;
LAGS 39 .
. ·,
',

Dead Time 42 .
. ,.,._
. ~·
.
;.o
...·.... . . .
.....
.. . . '-, .: ~·. . ·'. ·... ,
.
Integrator 43 . ....
. ..
·.' ... :.
,.
.. ..: . . , .... ., \•.
. ... ·'
.. ,,· . .. :.. ·": ..

..,·. . :
.: ·:-· ".
...
..·. ,.,':',,.,,
First Order Lag 43 ' . ,:.., .
·.~
,. ' .. ·..-.
:.-' ,, ..
.,4·.,: ·.s. ;.-· · ·
•. . ... ,·· .
•• ....· ·.·
·... :: ·'
.. .• ..· . ·. : , .. , ·.·· r.. s' :,.. ,.: ...., ...,.: .:
. ...... : :.: ..• ·.< . .:"·.::·-'. ·.... '..
-': ,· ...

Combining Building Blocks· . . .


r...
, ~·:
; .... ·: . ,.,, .

Gains 47 . .

.. -. .
I

,., . -.
.,
.. :, .
..
·.;
. . .. ..
- ,·

. ,.,
•• . ..
.., . . . ...
•.

,;: . .,

. ,...
..
.- .

• •••
111
~· ..
.·.' ...~ .
. :,;
. . ;
•• . : ..
. ··.
. . ~-
.,. ... .
:, :
.· .
... ' ,:.

.' . . :.~ :·.::_

' -~~ ·. ::

.. '·: ~:.
'
..
.
. . ' :.
. ' .·:
:.. ' . . ~ '· .
: ·:
.. :-
·.. . ..
c ; - LES OF ACTUAL LAGS · . , .51
•••• •• ·,. • • ••• '! •• •••
>·.: -r-'.:, .. · · ... ·.. >.' .

.....··· ,

Dead Time '


. . '· . .
. ... .. . . . . .
. : ,· . ":
;
·:, ..
. ·. . . .:
.. ":.
~.

Controllers . .. . . ., .
. . ';:.' t,
.
. ·.-~ '• .

Pneumatic Transmission Lags 53 .. ::· ...


:•
. ·..
..
.
.
: _:..
.".-
. «' •.
.:

:\.
. ·.. t' . ·~. ~;
Valves 53
Transmitters 54 . ·. .. . .... ,
. . . .
Temperature Measurement 54 .. .
-. - . . . :?· . ···.-~?'.~·\: . . _:: ·-~~}-:}:: i ·?·;. . . ~: ·: \: ' . .:_:. . : :· .
Tanks, Liquid Flow Lag · . SS ! -. :, '

. :.:.. .

Tanks, Compositional Lag · 55 '

Thermal Process 4gs· · · .56 :.·.· ·. . . '


:

.
'
' ... '
'
'
'
' ..
,'
. · i ,>\;~: >{ ·> : ; ~-trl :~~ · ., .

. pica I Na· tural .P.er·


', ' '

Ty ·1·0...1•;. · - , _i'J:,:£.
U. •
· •• . \ : : : •
• • .·~ I • • ' • • • -~ • .... :... .,
= · .. · • .-

•• : :J.8':1.::r~~<.
• ~· :-. •'. ·. :::::·
:· • ,._ •- •., • • r . ·•··· ·. .'
.
'', -: ::, : : ..· '<. .. :./
.
', ' · · =:-:.i'\ ·.\ ;7 ! :''.-.. ·,,:\-:!-;! } ,
. .
. . . . . . . . . . .

..
,.; '
'
,
'
·:
,,
:. ' ' , ' . ".
. . . . . :. ..': ,. ; :: ·.... ;t:,r
' ' ' ,, ' ' ' . ·, ~·

C ·. •· . · · ·. · TeR. '* C}\SCAl)E CO · . OJ., 51 ' ' . ' '

... . .. • ·. ,.'°'.}~/\:~'.i :·): i::i\ /t:_: ;'.r::.··: . ·; ·: .. ,_:. ~: : :·:: . : .:::: ..•
' ' ' ' ' ' ' ' ' ' ' . .
'
t'·;:. . : .·:~: , ...·...
c .· . '. . ·. · 1'ER 7 DBRIVATIVE ACTioNt,/, ":.· :;61•.•. ., ) . .. · :.:.·,.· : .
·. . . .·.: · ... · · . , ... : ·:··>:.. ··•·· •. ·.: .. ·: .: : .:·, ' ·..•. . ·.' .• ::
,'

'
'' "

·~;
'

. '
: ' .' t» '::·'

'' ''

•.' :·:
• ' .;'

... : : .,:.: ·. ·.,. \:· .'::,' :


',, '·:· ... · .... :· ... ·.. ,, '' ,' :,
'' .:'
'•,,,
.. ··
. '
<:
'
'''
' . '
'',
:'
,"••,
,,, .
. ~ /~\·:· ..
' .. ·~ .. :·.-~\::.. : .·.~ ... :". · s. ... ·.~.·. .:.: .·z ·:::?/;'·'
'
~ ..,,t. ... '
' ' ,,
'.,
'

.· '.:',:,:::·. ..'.
..,: .... ,,~,:'.
'\' ,>
' :·
'
'.

'
·-··
\'' ' '
'
'
.
/ "J . \ •' : ,r-;. . ; ' ":f . . ': ' ~. ; ) <
c - . · TER 8, INTERACTIONS ·. · NONL'.INJ'~ARITIES 65 . ~. · . ·
Interactions 65. ·. :;,;~:.·1·.>·t\. r· . ·•r::J'._:',:.f''7;>•:t,.{ ,.:::·,s~\,· {.
N on 1·mear.11es
•t· . .·66 . · . · · . · . . . . . .
' ' ' ,'

.
- •:: · .. . .
·<- :. :· •.:' i·:·::·. <: : · ·. · .. :• .: ., •• ··.
.
. ·.
.
. .
·,:-~' .
Process · 67 . . . .
:';
-: :·'.::
.:·, ',:'
'
''

: ,.
'

'
'

, ·:.,' ', '


' . ..
.
.,
'
', ' ,' ''

''\ .
.. ' ' .: ',
.,.
,, ,, '.' '
' ' "
' _: ' ' '
' "

... ~· :. :·

. ' •,:.\ t -, ' ', ' », ' . .: ,


· Hardware.ContinuousNonlinearities .· 67 ,,
'·.·.'.,:..
. !,·. '
".::,
', ·:,
•!••
',,: '

'
'
''
'

,:
..
,':'
'
'',
·, ,' ·:.
' .
.,
',
.
' ''
' '
• ' .. . ~· :: / ', . ''
.
:. '
.
' ' . '.
. .
:·:,? ~·· ': '': ',: .:~
. ·.,.,.·.,·. . .. ' ' ' '

Hardware, Discontinuous Nonlinearities 67 :·


'
' ,
' ':. ":.:: '•,
~: '•, ':,: '' '''
',/·>·,· .. :
.': ·.· :' '.~
:.·''
'' .'· " ' : ' ;
"
'
,, '
,, '
' , ..~
' '
,·,,,'
! ' ,
·~. ·. :·
' ' •
':··.'
' :; ,, ',., '
,'.:·.( ':·~
..
't '

:·•. ·.,.'="., ' ' ' ,. '' '' '


Velocity Limiting . ·. 67 :.: . d·. . . . ' \'\. !~ '\ { .. : . ,.· '', ,' ,; ', ·. ' .:.'' ·:·.
,
" .' . .< .•. ·.
' ', ·:.·
:

. .

DeadBand 68 ,·''
:
'
'
'
.,
.
':
.
' '
' ' '
' '
..... · .· . ,:r ' ,. '
.. ' '
'
', ''

'' : : ' ':'


'. ,, -
' ' . ·:f ;:
''
"
·:.· ,,

.: :\''
'
,, .
' '

Valves at Limits 71 . ·. · '{ :- 5 ··_ \''. · .· ; -' ·. · . · · · ·. _:;!. : ' .:. . :,:::'' '·::· ? ·.-· ·: ,,
···r·, ,.~.. ·; .. ·,·
., ', ',' ...... ,.
.· ..
' ':
. .
' ·. .....
. ' ' -:
:
.
',
''

·.
'

,
' ' ,, '

Integral (Reset) Windup · 72 .· ··.· ·.. ,, ., .;


', ,,,,
'

' ' ·. ,, ~ ' ' . .


: ·..
'

' .
,,'

'
', ,, ''" ', .,;· ?' ,1:· ·.
' . ~ : : ' ,, .. '

·; "'\,' '
,, " ' .· .... ·: ', ,, ~ ,, '

' ·.:' '', ' •;,:·;t·:. ,. '.· ', '


' . · ·· . · . . · ···. · ....T ..
··· .. ER
·. · .. · . . 9., POTPO
•·.. · .. · .'. .·. .. .. ,:';.·. , .·. ......
. _···7
,'.:-·. ,:..3·,. · · .:. ,·.,~>,:··, . .··,.·-. ··..--:·::t·'.
,~·. .·. .·:. ·..,.:.'(.··-':·... ,\·. . ..> :. ·.;'·... . .?·~·.,.=:; >· · ·
: ,·. ...." .... · ......
,, .' ,·. ,· ,.·• · . ··.·: .; <· :, ~;/··
. · . • . • ·.·.' • · .. ··:°" ·;,·
. CHA
p· · ·
·.+. . .,~. r -:..... · ~. . : . .·, . . . ... ·:... ,. . . ·> :. . ..··_·.... :,·.. i·.,. ·.;, . :,.' ,· .· . . ·. ·. ....::=-. .·;·.··::.:· ''..·...
.
.. ,.,::,:"·. ·
, ' ' '

· Digital Control Algorithms • ·. · 73 :·· .·.:. ;. -. ; . . . .·'·. . · .. •: ,'. :·~ :, ··, · •- < ; ·i - .•
Sampling Frequency and Loop :perfonn.ance ;· >74 . · · .· ;: ,: . , : . . : . ~: ,:. . ., .
Load Changes /Upsets I Disturbances '.· . '. 76. ·.. , . ;):/:'.\;. ·{r;.. _; '· . ) ·. ·. ·. C: ; .: 0. ; . : ...• :' . :, ', ) ··. , - -
Da.mpening Noisy Measurements, . · 77 · . . ·- -:; , : :·:-. . . . . . . . . -: . : - . · .
' ' '' ' ' ', ' ' ' ' ',

,.'' .,
.'
. ' '' ~·'
'' ,,,: .

' ,,, .'


' . ·:, - '.'·. '
,,
'
, .

.. ' '' ',


' ', ' '
' . . : ' . ,: ,: ': ,:
.
' ' ' ' ' ' ' ; '' ' ,' ' ' "''
. :: ·:.,
" ' ' . ,. ' ' ' : ' ','
.
' ' ': ,: ' ' ' •':'•,• ,, ' '
- ,' ·.. : : ,, " '' ' ' ' '

81 .·: ·.:. ·.·:,


.
· ·..:··.".· ~
. ·'

. ..

GLOSSARY · 83
' '' ; '
't :;
·•· .'· '. ': ,

' '
. ',.'

APPENDIX A-1
Pure Dead Time Process .· A-2 . .

Process with Dead Time and Integration .· · A~3 .· . . .

Derivative Frequency Response. A-4·.·


Order Forni A~S

..
'' ,, '' '' ' ' ' . ''


IV
..

PREFACE
. .

This second edition of Controller Tuning and Control Loop Performance has . . ,. . . . .

been extended in both directions :fmtn ,the first. Sections have been added for the
very beginner and for the somewhat more experienced. It is about twice the size. . .

Sections have been added on the what-to-do and how-to-do-it of tuning, to help
. . . . . . . . .

the person who may have never done it before. Then interspersed throughout are
. . . . . . •.

paragraphs that extend some of the non-math concepts to the realm of math, or at
least algebra. These sections explaining concepts in math (sometimes frequency
response terms) are clearly identified to make them easy to skip. This second
.
printing of the second edition · also has expanded part of chapter 2 and has added
. . . . . . .

two pages to chapter 8, as compared with the first printing. It still stands on its
. . own, of explaining the essence of feedback control, without referring to math. I
hope these new references will help any reader who wants to bridge the gap from
the nonmath to the math.
....,
. .

'
· The first edition was essentially a verbatim copy of a report written for DuPont in . .. . . . .

...
,
;

1983, which I was allowed to make public in 1990. This second edition is perhaps
'.. '.
80% based on a 1992 update of that original report, written for a training course
..
....
for DuPont instrument technicians and engineers. The new version was to have
. . . . . . .
specific references to the DuPont situation, and was co-authored by Paul S.
Fruehauf (of DuPont) and myself (DuPont retired). I am very appreciative of the
permission from William X. Alzos (of DuPont) to use what I wished from those
course notes.·

I am particularly grateful to Paul S. Fruehauf who has worn two hats ·in. the
preparation of this second edition, first as co-author. of the DuPont report, and . . .

second· as critical reviewer of my modifications and additions to that report. Most


of the material in chapter 2 is his. The first draft of much of that material was his,
. . .

and he persuaded me to include it in this booklet. He is currently an employee of


. .

Applied Control Engineering, Inc., a consulting firm in Delaware.

I have tried to make this second edition appeal to readers whose background may . ..

not be the chemical processing industries. I know I can only partly succeed in this
broadened scope, for all of my 40 years in the automatic control business were in
that industry.
I hope this booklet meets what I perceive as a need for more information on the
beginning end of training on the subject of controller tuning and control loop
performance.

ENJOY

·'
·-'·,·· •
...
This booklet on controller tuning and con- scapegoat, being blamed for problems that
trot loop performance stops where most are not related to tuning, with the result that
books and courses on the subject begin. Too time and energy are spent needlessly.

often the subject IS introduced with math Meanwhile a proper solution goes unsought.
unfamiliar to-the reader. That does not have
to be there are simple concepts to help While I will give ruJes · for tuning, the rules
those unschooled in the math to know and themselves are only' part of ;the picture. The
understand the basics, to appreciate the ''tuner'' needs to know what the desired per-
limitations and to know· what can be formance is and what to expect-when the
expected. system is responding as well as can be ex-

pected, and when is it not. If it ts not, then
the rules may not apply, or should be modi-
fied. This booklet teaches not only the rules,
but· what can and cannot be expected
from tuning. It is also to teach some of the
common pitfalls. Why do the tuning rules
not seem to work sometimes? In addition,
My field for 40 years was industrial process
tuning is often done to fix some problem.
control. The basic concepts of control are
You· cannot use or fix anything unless you
the same, regardless of the field. The exam-
know how it should work, and that includes
ples will change, but the concepts, princi-
control loops.
ples, and much of thevocabulary won't. For
readers whose field. is different from mine I
hope you will gain some useful insight
. . .
into
your situation.

Not everyone needs to know about control-


ler tuning. Many businesses, like banking
and insurance, probably need no one. Other
businesses, like the automotive business, Tuning rules presume that the desired result
probably need only a few. But that still is a "tight" system, one that does the best
job of reducing the effects of disturbances,
leaves many businesses that do need to .

know, and you wouldn't be reading this if and/or one that responds quickly to setpoint

you didn't feel a need to know! In many in- changes. This may not always be what IS
dustries proper tuning is vital to quality, and desired. Many level controls are often delib-
often decisions are made to take expensive erately detuned (made more sluggish than
steps when better tuning might do the job. the tuning rules would make it), a condition
On other occasions controller tuning is the referred to as averaging level control. Many

loops Ill a plant do not have a very vital
r -,....
.,....
·.. , .;iiit;
'.

2 Chapter 1, Getting Started

bearing on quality or other business consid- made? Not many. Quite possibly not any.
erations, so whether they are tuned tightly Usually there are at least a few loops that
or not is not all that important. How many stay on manual for some time, sometimes

new operations are started up and have all even years. It IS hard to argue that these
the loops on automatic for the' first product loops need tight tuning.
·'
'.

Controller tuning is mostly science, Tuning mathematically pure and simple models are
rules are based on mathematically clean and used to represent the ''typical'' process.
simple models that approximate the real Don't worry about that, certainly not at this
stage, The differences are relatively small
compared with what I consider realistic
goals in tuning. We will not be concerned
about determining settings to within 1%
world. If the real world were and generally not within 10 or 20%. For '
mathematically clean and. simple.. then instance, if the tuning rules determine that a

controller .tuning would . . be all science controller setting should be 1.00, it doesn't
. . .

(provided of course, therewas agreement on


. . '
really matter · if it is set for 1.01 or 1.10.
what was desired from the tuning). Happily,
. . . ..
.. .. . . .

experience (and higher math) has shown


that the real world can be simplified without
sacrificing accuracy enoughto worry about,

It IS known, with .a reasonable degree of

certainty, when this simplification lS
. . .

invalid, and therefore when the rules for Even if· set for 1.20 you would be hard

tuning will break down, pressed to see the difference ID most •

practical cases. Determining settings within


. . . ·.. . . .

30 to 50% is a more realistic expectation. ' .

Two specialists in tuning will almost surely


come up with different settings in any given
situation. They are far more likely to, indeed
will .almost . certainly,
~ . -
come up with the . . .

There are ··numerous publications givmg • • same analysis of what m.ay be wrong with a
tuning rules, and, as you might expect, they loop. They are less likely to agree on what
. .: .

don't all. give exactly the same rules. This IS the best solution is. It is rather like politics
: •·. ';. . . . . ' . . . .

Ill that regard .
because different .criteria are used for what . . .

constitutes "proper" tuning. Different

'.
••

Chapter 1, Getting Started 3


.... :.. . . :.· ·· . \
,,:·
. . . ····.,,.,
.. . .. ,.·. ··.., : ~··:··· ·. ,. ··.. ·.. .•·· . . ... · .. ·.:·.';. .
....
. ..
.. . . '
;, . '· . . .


. .,. . . . . . : ~: . ,
. . .
. . . .. . . . . ... .
..
..
. . . .
.
.... .
..
. ... .. . . .: .. . .. . .· .. ·.. . ·. ... . . .. :
'
··.·:,;,. . . . ' .
·. ' ,:: .. .' ., ......

. ,. . ... ' .

No reasonably thorough writing -on control- loop· method. · Nichols . then verified . the
ler tuning would be complete without pay- mathematical validity of; the ·open-loop ap-
ing tribute to J . .G. Ziegler and N. B .. Nichols proach.
(Optimum Settings .. <for · Automatic
. .

Controllers,•·Transactions··;of the ASME,· v For history buffs there is a book you should

. 64, Nov. 1942; · p7.S.:.?)., :_ Their contribution
...... .
know about: Automatic Control, aassical
·..I
.
., was a quantum leap. forward in the science
.· . . ···~
Linear Theory, edited by George J.· Thaler,
and/or art of tuning industrial controllers. It Naval Postgraduate School. It was published
..
. - .
'
·. took perhaps /10 years or more after that be- by Dowden, Hutchinson and Ross .Inc.,
fore subsequent authors started to hone and
. .
Stroudsbµrg,11-~A./.:in_ ~l'.9Jfl~,
. . :
It
.... is qui of ·:ptint
. ·. . . . . . ,.. .. ~.. ; _ ·-~:':.~:.. . .,·. ':;~· ,.. , . : . ;. . . .. . ·.'· , ~ .. . . ..• .. . .

refine . their: recommendations, .. but . the es- now, :biit~can·;t,e·-9i,ttfn.ed frqm ·major techni-
sence of their approach has remained un-
. .
cal· ··Iibraties. )-/The :iibook "is .one ··.of the
.. ~ . . .

.. . ·' ...
. .
"Benchmark Papers in Electrical. Engineer-
..,. .mg. anc d .Computer
. 'Science,
. . '' v. 7, wit. h
. .
. . .. •, .
.· .
.


,. . ..
·.
'
Library of Congress Catalog .Number: 74-
f ..

2469, ., and . ISBN: · 0-87933-083-X,._ ·It ·• is a


r
f
photographic· reproduction .of .milestone
..

f ,.

. .
papers. on the math of· the feedback control .
(

.. scathed to this day! . Ziegler and Nicholsnot loop, with-editorial ·comments on .the contri-
'

. only brought. order out .of chaos; but· they bution. each made from-a historical: view-
'

presented . it · in · a simple, understandable point.··•·. The British· papers by Callender,


rr,. way. They· presented two ways· 'of· de- Hartree.. Porter (and Stevenson); 1936· and
i
•·

'' ,. termining controller settings. One was based 1937, from which Nichols was able to con-
t
r
.'
,,.
on closed-loop tests, the other on open-loop firm the formulas presented by himself and
I
'
:'·
i'
tests. . · They were both based on sound Ziegler, are also containedin it, as well as
i
r-

'
'~ . '· mathematics, though their peers did not rec- . the original Ziegler and Nichols paper.

' ' .
;'
•t ognize oraccept it at the time. A 1991 con- .
~
i

ii. versation wi.th each of . them revealed· that The Ziegler and Nichols paper is also
included in a collection of papers on PIO
(
• ,.
'~-
',.. Nichols, with a mathematical bent, was pri-
marily responsible for verifying the math of
'f
"
;
f. tuning: Reference Guide to PID Tuning,
.... ·
;

..'
i.
the closed-loop formulas, while: Ziegler; of ·a published by Control Engineering. · .· . .
''
,.

•''
i
more empirical bent, conceived the open-
'~
r,.
~. ..
1
'
t
!.
.' .
'~ .
.
'

;~
. ·.. ··- .
' ::· .
. ~.,.'

'
'

.
4 Chapter 1, Getting Started '

The task of tuning a controller can run from The science of control is based on math that
fairly simple to quite complex. It is rather is formidable to most persons. Happily it is
like. income taxes. Many.·.·•cases· ate· .quite not necessary to understand or even use the
simple, ·but<then:·,there:ate a few that need a supporting math to absorb · the governing
specialist. Also like taxes, it has to bedone. principles. The math will be largely, if not '

For the simpler cases, which constitute totally, omitted. No proofs will be provided. .
'
'..
..

. '
'
possibly 80% (somewhere between half and There are however concepts, which may be ""

"
',.

all) of the loops typically encountered, the new, which should be mastered. These .... ..
procedure can .be reduced to a set· of easy- relate primarily to understanding the
to-follow rules. · importance TIME has in the control loop. .
··,., .
. .

AMOUNT is also important, but not as


much as time. By far the most important
concept to master in understanding control
loops- is the concept of LAGS. An effect
happens AFTER some cause. A control '

..' .
valve moves AFTER the controller output ... . ,'

changes. The measurement of a temperature


While these rules are based on sound sci- in a well LAGS the actual temperature .. .. .
ence, applying them without knowing. what outside the well . The coldest day of the year
is · expected leaves little · understanding of occurs AF*l'ER the shortest day of the year.
what is· going on .. Each tuning experience . ' ·.
. .
.

becomes an isolated event. There . is no


..
.
' .
·:
·. ,·
.: :

framework on which to build understanding,


no adequate way to transfer experience from
'

one. time to the next, or from one person to '


',

.••
another. One goal of this booklet, is to pro-
vide that framework, that way of defining Not all lags are the.same, orhave the same
. .

the experience so it is both. understandable importance in a control loop. It will be a


. . .

to the person doing the tuning, and transfer- major part of this training material to de- '
', .

able to others. It· might be called· THE velop. an understanding of where lags come .
. . .

LANGUAGE OF CONTROL. Inciden- from, the different types that are used to ap-
tally,· while I assume .the reader has ·some proximate the real world, and what their
familiarity with· many of the terms used in · relative importance· is. A few words, mean-
. . . . . . . '
." .
.

automatic feedback control, I have provided ing specific things in a control loop, will be
a glossary for the terms most likely to need added to your vocabulary. Again, it is THE
defining. LANGUAGE OF CONTROL.

;
'

..
,

Chapter 1, Getting Started s

As with, I suppose, all fields, the science are sometimes · available if the tuning· is
and art of feedback control grew before the done with a digital· control system. This
committees on standard terminology were constancy of the PID function in an era of
formed. The predicable result is that several
different terms are used to name the same
thing .. The controllers we are going to talk
about have . three adjustments. They · are
troportional, Integral and Derivative (PID).
Computer based systems· often have the . .' . .

fourth, which is an adjustment for the filter phenomenal techni~al progress is a sobering
. .

time. Computer based controllers may also thought. THE P.RINCIPLES HAVE
. , . ! .

have a decision to be made about the cycle REMAINED UNCHANGED. I understand


time (how often the controller looks at the that some really high powered · math has
shown that the PIO function is the best gen-
·.

process), but this is not considered a con-


troller setting. It is, however, quite impor- eral purpose function to use to do the job.
tant, but it will-be discussed in chapter 9 More sophisticated control algorithms will
r
i:
t..
'i·
produce better performance when fitted to a
. .
I
r
1. PID controllers have been around since specific process, but poorer performance

!.
l•· about 1940. Modem controllers perform the results if the process changes. This sensitiv-
f
f same functions as those, perhaps with a few ity to process changes is called robustness,
"
Ii
r
embellishments and certainly more accu- with more robust being less sensitive. The
,.
! rately, but the same functions nonetheless. PID algorithm is an excellent trade-off
f
i
So the tuning rules have remained essen- between robustness and performance.
'e . .
'
tially the same over the years, though aids

We talk ·about proportional action . but we . .

tend to refer to the adjustment itself as gain Figure .1.1 . shows what gain does to the
. .

or proportional band. The action means that controller output in response to the .error.
. . .

the controller output moves in proportion For a gain of one, the output . changes the
to the error between setpoint and controlled same amount as the controlled ·variable (or
variable. Many terms have been used by the setpoint). Higher and lower gains cause
different manufacturers to designate this greater or smaller changes in the output for ·
. . .

action. It has been called proportional band, the same change in _the error. If the. output
proportional gain, gain, throttling band, increases . as . the controlled variable in-
sensitivity and surely others. Some are creases, then the controller is said to be
reciprocals of others. For instance gain is
I 00 divided by proportional band. I will use

gazn .

.;

...
-, ,·.,~: .. · ...:.,··:, . ··.,, . . . .
·.·.:·,.,,·, ..
.; -': : . ." : ... ·,·.:.. . ... ..
.
.··
.
.•

' . ·.. ·.. :··'./i/< .. ·.. ·~· . . . . ',

..
'-· ,, ".. .
','

6 · · Chapter.I, Getting Started .,

c
.·,
.
',

- .
. ·.· :...: . . ... . ·,
"
. .
, ... .
.. . '

. . . ·.: be an offset
"

if the coll~Jlerhas no Qltegrat ·. . . ·. . ·. . . '


....
'
·· : · : Gain :::;: 2 2_ • · .
~ ..
. · · actio ,,n· • · ·T· he · · offset
. .
11-~
.m· ay·,· '. n· A•': ,-:L~:.~:
.. . . .
{lg:
~:~:·./,·rta:":;
.-·1111no . . ·
r ~::. ·. ·. R·t·· .'',:· '. ·:· ·. · ··
. . . . . . . .. . .. ,. .
.\;%1, ...
:_.· ,·1.~.~::.·· . :··· , ..: ·.. . . . . . . ..
.,.
,.:·
•.

. · enough to worry. about, >but:-it·~/Wltt!-'.~'--thet.e> . . ·:. · · . ,•

c
.
. . . . . . .... :: .:· . . .. ·.·.: . : ; . . . . . -: ·. . ·..:·~:· .<.::··i:· .';:~·~ .. . . ·;.· ', ··. _:•. ·:· . : .. :. . . .,
···'·.,
: It has to be; e:xcept;;,-~~~;tgµe precise point::<i~,/- :; . · -. l
'>

'5 G» .
the output-versus-errorcurve. . . - , . ...

a.-
.. ca
.
. ·. . . ..

Figure 1.1 is not totally accurate. Many


_- ..c· digital controllers · can 'be.:· configured to
. e:.. .
_.., CJ
: . · have the proportional action occur on·ly on . ·
c: .. .
o·• ~ .:; . . .
oo.. the controlled variable.~·notontheJerror, to
. .. . . . . .· , . . . . . .....
avoid potentially undesirable . actionion ·
. · .. I · · .· See:Text for setpoint changes.· This· is a 'desirable op- ·
. . ·- . . :~I · : · aua·11t1cat1ons
. . . . · •. . . . . : :_ . ·.· I
. .· ' . . .
.
·.
....
..
.
".. . ; .. . . :·.. .' tion, but will not be discussed here. Also, ..
·•.

if · · a controller . · has . no · automatic reset : "


,·'

. . .. o .· :·. . . . .
: . ·j ·.... ; ,. . . . ·.
" ,,. '
;.
. (integral) · action, · to. be described • soon, ·
. ···· ... ·:,··.·. ·\· .. : ",
. . . . . . . . .
..
·,
',

· . Error, Percent of Scale then it will usually .have a manual reset .· ,


,

','

Figure 1.1. A proportion~l~onlycontroller has a fixed


. . .
· (integral) adjustment, 'This· is an-adjust- .· ,·..
'
..

relationship betwe_en error and output •.


· ment that allows . some manual compen-
.
. . .
sation· for the. offset. ·
direct: acting. If it decreases as the con- . . . .. .
.
·~
., .
. .·.
·.
.
~ ...
.
. . .

trolled variable increases, then it is· called·


. . .
.
. . ..
~.. ··
. .... ·. .. . . .·'.
.....
·: ..·:: ..::.
:
. ·.· .
reverse acting. The controller action is set .

. · -· . Controller
(or checked) initially, when a controller is
. ~ . ..
Setpoint ·_· . ·:- ·. · , .· ·., ; ·. · · · · · ·.. · . . · · · . :• . · Output·.
first put into :seryice,. and is not changed. . . p ........-
after that .. The action has to:. be right to get '. ..

the . controller . output. to. go .in _ the ·right.


. .
.

. ~ . . . . ..
- . .
. . .
.

..

direction when.. the . controlled .. · variable ·. Controlled


. . . . . . ' ·: . : ·.... ··,. ., . ..! . ·, ·.. .

...................... ·. Manual
changes, or else ..the controlled variable
. .

. . . ·. . . . .. . . . : . . .Variable . Reset
will avalanche away from the setpoint, ..... 0: : · .
'"
... . ,· .
.
. .

Figure 1.2. · A signal-flow·diagram to show how manual


reset may·
. be
.. . .
used to shift·. the output· for any· given
error. · · .: ·:· . .. .

· Figure 1.2 shows how this is represented


.
"
in a ·sigiia-l;'flow diagram, -: and • Figure · : 1.3
. ... .· . . ·.. ... . ·.· .. ~· ·. .. , .... ; ... :.. . ..

shows · how· this· might be ·represented··


\ :·

With any controller that is proportional-only 'graphically. It allows


. .
for adjusting what
(no integrating action; yet to be discussed), the controller - output is'. to be when the
there has . to be. an . error between . the
.. .

. .
_ error .is zeto.· It: may·;-t,e,.:thought · of _as
setpoint · and the controlled
. .
variable. · This .
sliding: the gain curve up· or down on . the
error is· frequently called offset. The · easiest graph. · Manual reset permits reducing the ·
way to understand this is to look again - at offset · at the- normal operating conditions,
Figure 1.1. · The only time the error is zero is but : it does not change .· the · basic
when the controller output is at· 50%. If the characteristic of proportional-only control,
controller output· is at any other value, then that there will· always be an offset, except : ·
there has to be an error to produce ·that out- at-one exact point ..
put. Simply remember that there will 'always
,.

. .r!J· .

Chapter.I, (jetting Started .


.
.ll/
,::.~·
..
...
.

. .

... ·. ·.· . . · ·• · .... ·• ·


;:- ,.;, · .:.: . The.:::.~ertRn.-iJ-·,lhe setP9int:·:r.mnus the . ·
. .
. · .· . · .· ·. iJ·.·.
. . .. . ··..· ·,:·. ·· :· · .. ,
'i'V'idBiit~\~li'atJJe'
i · ·..· . ·-.'·.··. •. ·.:·.··i<;·•··.··;:. "..· ·. •.·. ·. • . ·. · . .. •. .· .:·.·· ..·:,• ;·.:.'·i< ·• ·. · ·::-.. ·. :..·:. ·, ..·.~ ,;:·. ;:,;,.:··.·.~··,;. · · ···. :., .·.·· :·. ······· .,·:·. ·· .. :· . ~. ·,·Ti,iJ.':g.:'°a1bx!,~:
. .
,~ .. >> ·::~··.
. ,,. ·.. ·.:· '"1 .. ·1..9· .·. ·. . .

r: , . : · , · , : • ·
.:_<.·· .. • ".·' . ,. •·. ':·:. '., ,,

. -= 70 . .· ... , . . · :. ,. . . .• .· ·
. . . .. ,.,. . •. : . . .. .f.·r,.• ;., ..•. ,,, . • :.·. . . ·.,,·,: .. ··• .. , , ~.·.. ·•·· , ·,.:··)'""·'

i .. · : · - · .
1,, •• , ,., •• ••••••• ••• •• • ···'· ,... • • •• ••••• ,•· • • .

..- W'
·.·I::,_
, ., . . . . . . . . . ·c;uently
. .. nFJ(Tled
. .. . Kc,
. . for , ga1n:Ot-the-
. . . .. • . · .
S- i . ·. · ••;. . . ·. ,,.;I ·· '. 01,, ,, ••.~~ . .: · . ·• ga.~t';'lle(. The at,c,\{6 equation then '. ·.:
. .~ 0
8 :! -: . . . . . .·
.. . . ' . . ·. .. .: . . tt . . . ; L. .-.. . . ,,; ~~ ''? . .. . ;
·• . . . could btf writtena~: . · · . · ·. · · · •· · .: · ·. • · .· ·. · ·. ·
. .. . . .
. .

--e·· .,c:. ··}. '.: . , •· , . . . . ..· · · ~. · · : .. ·.·._ · ·. ·:·,. ·__ · .....


·. · . -:. _..,.·. _
• • • •
o· ·u•·..··ip· ' u··-··t·
..... ~· ·e·; ·Kic ·~::·a··· .:,-:.. . · . . · · · · · . · ·
' ' • ,,~ , • . :·•••f•' • • ••• , ,.. • • • ; •

. . .. .
. :,•:.;
. ..
.•...
.: .'
-· i ....
. . .
: .
..... . . : . . .
·I' . .. . . . . . .. .
. . . . . . ,,
.
, .....
.. :
.. .. . .. .. . . ,: . :.:;·,;- . . . . ·. - . _·.,
. : . .. . .. .. . , .. . . . :~. . . . . . :"'

· . ~ .a. 30 .... • . .. · ·,,. . '. · .. . .. · ··


.,
.·. . ··~fJ . . ...·.·e.:.. =, eno' · t
wr., . ·e:re'.. . •' . . ·. . ·. ....J.
·. . . . . .,
! • · ..: · - <·: - • • • •• •
···.· ·,· .. ~· . . ... .. . . . .. . '.. . . . .. . .. : . . .... ..,~ . . . . . : . ·.
.· . .

I · .· · · . . · .: . ·.· · . ·. . · •.. · • . Kc·7' C9.lltf9ller gain , . 1


... -~.. ·.~· : · .· : ~· : . . . ·: .,: . . . -, ·.... \· . :. : . . : ·. . . . . : ,·:.~· \ .. · . ·.. ' . . . ; . . . . .
· · B-b,as
. . . :. . . . . : . \ < . : .· ; . : ; : ; ...
· ··
. •. . .· .. : : .•. ·.· .' . · . . . . . . .. ·. . ..
· ..
. . . . . . . ~ . . .
. : . .·
. . . . .. .

i. >· · • ' i o . . .. 'II~ 't,fuV( Veiy ffni<I :gf ~rittf19 81(thJt ,


.. Error, Percent of Scale ,r.,ath . "(a.P.l{!BIIY~ fA/gel:>~fJ) ;· : and :-:'Jejy. ·. . .
. i

... .
. . quickly forget ~aboiit· tfiti· ~bias,: 'B~ ·as· it . · · : · .
Figure: .. 1,~3~·<· Ma.11ual _ reset ·. in . a proportional-only : is; rarely: of:· since it does . riot;: .• /. . coneem
. contr9.ller : ctiar,ge.~ ;the : fixed relationship between affect · dynamic - . {)fliformance. -.: The . . . · '
error and contr9ller·:O\ltput : . - : . .. ·. .· .. proportional~nly: controller :then , ,has: · : . .
.:. . ,· ·:·... . . .·•·· : ... . -. : . . ... ·: . . . ·. ·.'':··.·::;;:·I:./::.~,~.. .: . . .. . .

Any· Ioop Will Cycltt.~~- UllStabte} if a ·~ :La~lace' traniJ.onn : . · or·. transfer.· : . ·


the gain is increased far enough; The task function of: . · · :. , . · .. , ...
of setting the. gain ·is on~ . -~f.. g~tting the . .

effect · you want wi~out causing Output =Kc


instability. . · :i · · · · Error
. . .. '··.··· ..'.·.,.::·.. ·. . . '. . . . ..·, ,·. ·. . . . ,: ... · : ... :··:. ···: ..... : ·:·, ,.-,,:;,~;ti.!:':...
. .

or more simply:

K·c
. . . . .

· · This
is· ·the first ·section dealing with
the· math I algepra of control. If you· · since the left hand side of the
... want . to. ·skip:;. :,t, ,: ~:sny subse,quent or equation is ·understood. to-,;be. .- the···.· .
section-· so devoted, --1 encourage you output divid,d . by:" :t.fJs inp1li#t!t·:;:[hf>.:_:. ,., .
.· to . ao; so~ :!/:1.€11."' i$ no need tq get transfer· function· :Of ,any . tle:ment. '.:in;,.: . : . . . ·
. bogged qqwn:~/ il1::. this and frightened .... th~. a,olJtrpfJopp, is,:..f!iii/:O:WJ?Ut, diy~· <) '.·.
. off what l_afTI t.,ying-.to say. lv1y intent by the , ilJPP.t•.1. l:ri, tt,i~ case it: is what .~ "., ·
. is to make the booklet stand alone
. .
. · you : m~ltiply·"lrie: ei#,r\ bY to· get the ·:. · ·
• '·without- reference:··::.to math. ·These··· ·
:

. . . .
. .
>·. dutpLlti\ · _ ' · \ ,.(,i .ti·. • · ii < i · . •·.·· · · . ·. · · ·. • · · . . .. · .
. . .. ,. . : .~. : .. ·'.r .'). '· ·.,:·. .. . . :. :: ;: .,: ·.~' \~:·. . .. . .:, . . . . .. ; :. , . ,: : ..

· .· brief : :. sections · ar:e.: .. · presented-·;;· to .· ·. ·. ·. . .


. .

· .· . introcJuce . the math td. Lthose~ of you. · . . .


.
· · : Once yea :g$t ;nto: ·. the · algebra you · · ; · · .·
. ·. who might .be -interested.,:: . with the : ·•> >
. . ..
·.
. · :/·,-can · :satisfy •yourself. that·· these· . are·;; ;,; . • ·
. hopt;1 that you won't be intimidatf}d if . ,.::;. ; -legitlltJate · simplifications~. You ·should · . · ;i :
. you decide.to ·,ead other materiflt.on ·. -'.t·,::·· .. . . n.Qt:., play ,round: : V:f'-1}(.: much. with . the .·. :/ .
· th~ subject. . . . . . ·. . ·.· ., ; .. . , ;;.,.,// . . ·· · algebra ;.wit~o.ut. ::lea.rnif!g·, a. great deal : ,
. . . .. . . .. .. . . .
. .
. ;
. . · more than this lxJoklet will teach you. ·
• • • • • •• • •• • •• • : • •• • • :·.,,:'~ ..,·,, • • .' •• ·:. • ,. , • • ~· 6,';;):,.\v, , • ' •

.. . : ·the .. math . ··:of : : a,. . proporlional-dnly . . ... . I ha_ye IJO intention of teaching you .
. controller is quite: $.lmple: .. . .. . the. rigors of the underlying math.·
>
.. ,. . ..
. .. .
··.: ·: .. :
.
.. ·.··: )' .. ~·. . :... If you decide to learn more, you need
•.. : to use another source. . : . . . : ..
: · Output = (Ertot x :Gain) + Bias ·
. . ' . . , . . .

.. ..

.
. '

' .
·.


·,,
. .. .. ,.... ' .
: ..
:....

8 ChapterL'Getting Started

In the earlier days of industrial automatic was due to proportional· action alone.
control the integral function was- almost Within the physical · constraints of the
universally called reset. Now the more sci- controller, the output will continue to
entifically correct term integral is gaining change at the same rate. This change comes . .

widespread use. I tend· to use them · inter- from integrating the error.
changeably, • especially when talking as
compared with writing. When referring to So, the integral action causes the: controller
the adjustment the terms reset time, .and output to change at a RATE proportional
reset rate are both in common use. One is .
. . .

the reciprocal of the other, so of course it is


vital to know which one you are talking
about. To say to: ''tum the reset up'' .is. an ..Cl) ...
. . .

.
I· ... -. i
ambiguous statement, because you· .don't =ea. :s
know whether the speaker is talking . about .......
c :, - - - __.. _I A- -
reset time or reset rate. It usually means to (JoO I .
A
decrease the integral time, ·but the phrase .__, I + __
still leaves uncertainty. It is rather like
"
I • Integral_ I
saying to tum the air conditioner up. Does I Time I
that mean to get more cooling or to tum the
I
thermostat higher? I will use reset time· or . .: . . . .

integral time when· referring to the setting


I . . . : .

... I
itself, discouraging the use of reset rate. e...
w o.__..,
Integral action .is not as easy to understand
',

as proportional action. The graph ·that.· is Time


often used to explain it is· given as Figure Figure 1.4. A proportional .. plus integral controller will "

1.4, which is really for proportional-plus- integrate the error. to··add an amount to the ·output equal
integral action. Imagine· a controller just by to the proportional change in one integral time. ·
itself, not connected to · a process. Then . . to the error. The longer the integral time
imagine that from an initial condition for the slower it changes. A controller with
which the error is .zero, that an error is integral. action will eventually reduce the
suddenly introduced, called a step change. error to zero, as the output will continue to
The controller output will then change to a change until there ·is no error. That is, this
new value, and the amount of the change is · will happen if there . are no continuing
arbitrarily called ''A'' in Figure 1.4.: · · disturbances to · require the output to
continue to change, and if the manipulated
After that the controller output continues to variable has enough ''muscle'' to achieve
move in the same direction it went initially.
. . that. Manufacturers build their integrating
It will move an amount equal to the initial
. . function to be as close to mathematically
amount ''A'' in a time that is the integral pure as they can, and they do a good job of
time or the reset time. The units of reset it, whether it be one of the very first
time or integral time are minutes per repeat. pneumatic controllers, or one of the latest
The reason for this terminology is illustrated digital controllers.
in Figure 1.4, which shows that the integral
time is the time to repeat the change that
'


Chapter 1, Getting Started


.

Before the advent of digital controllers there Not much more will be. said about the reset
were integral-only c(jntrollers, but they were windup problem -at this point, . except to· say
. . . .

not in widespread use. The . · function is the twothings. One isthatit isaphenomenon
same as in . a proportional-plus-integral that does exist; and two is that the measures
controller,··. except of course' there is .no taken to combat the problem work with only
change in controller output .. · due to varying degrees of success. These measures .,, .

proportional · action. The . change in con- seldom totally eliminate the problem. It is
. . . . . .

troller output is all from integrating the er- far better to take steps to see that the
ror. With essentially all digital controllers controller does not windup in the first place,
there is the · option to have integral-only than to ·. expect · the . anti-windup features . to
action. When · this · might be used will be keep you out of trouble, Forbatchprocesses
discussed later. reset. windup · can be an especially severe
problem on start .up ... Specially
. .
. configured
. .. ... . . .

controllers exist to combat this problem, but . . . .

they will not be discussed in this booklet.



.
• •
.
• >
.

.

. .

. ' .
. . . . . . . .

Any loop will cycle · if yoµ · reduce the The math (algebra) for a propor-
. . .

integral .· time . far enough. This .. is. true tional-only controller had nothing in it
relative to time. The proportional-
whether the controller is proportional-plus-
plus-integral controller does. · This
integral . or only integral ...The task. of. setting introduces a new symbol, which is
the integral time is one of setting it low used in essentially all of the literature
enough but not too low. today, and that is the lower case ''s. ''

d
s=-
dt

..
Any control loop with integral action is If you didlT't know what it was before,
••
subject to having a problem called reset you still don't! The d is the derivative
windup, or more recently, integral windup. . dt
This refers to. the condition when the con- relative to time. If you see :!_, this is the
troller output does not have enough muscle . s
reciprocal of derivative, which . is
to reduce the error to zero. Since the con- integral. Please simply accept that. The
troller integrates this error, the output will . Laplace transform (transfer function) for
continue to change until it reaches some the proportional-plus-integral controller
limit, which may or may not be the limit of is written like this:
.
'
' the manipulated variable. In digital con-
trollers this is a limit set in the menu· for that
.
•.

controller, or it may be set in the software .


Kc 1+ 1
1
,

'
-,
T:s
I
For electronic controllers it might be set
with a manual adjustment. For pneumatic where T; = integral time
controllers the normal situation is that no
;
provision is made to avoid windup, but that Sometimes it is written this way:
.
,.. extra instrument items can be installed to
'
combat the problem. 1js+1 ·
\
'
Kc T;s
'


..

,.
,· ...·..
.. · :;,

...
10 . . ;,

-:

. .: ·; .... . ., . · ..,'·•.. :..,.., ..···,· ··.'


• •• :·. ····;.:·,
.
.·:
<,
·:

,,,••·~
-. ···.· ,.
:~·.
. ....
·,, ~:·,·· .
': ~ ....... ·~·-·
.
:. ··;.~ -. :.,;": : ..
. ... . ...
:; ••• ';,: .. :: • -, -,
. . • <· -, ••••

. ··.··K
,.

.•
.. . \.,....
.. ::::
~,.;:; . c.
,:t:
f~', .'·.'
. ~·

. . . . . ..
~ .. . . . ·... •, ...
.··,··:. . . .
• .. r-: . •
; .. ·,(._.: . . ;,
"s,
.. . :· ;' ·..,, ... . . . .. ··.:
:',;.
...·.··..
, . . . :..~... .
' ·,
.
.. .
. .
. . ...,. . . · ..
'•

' .......... . ...... . .. '. . .. ,..


,.:.,.... ·"' .. .. . . . .· ,:· ·'. .•... :. . . .: (·~·::.... , .·
.. ';.:..
.. t~,.. .:. ..,~..:. ' '
"'.·K··.'. ·.. :
. ... ''i .~

. ·.• .. ... . .. :,
..
.
. :' ·,
~.
.:· .· ·.. ::' . ::.. .K
. +:, .,..·.
'

c·-· . ,..... :· . •..


-. '":"' .... . ,....
·:c" ··r:s'.
: '· ; ·:
. . . ·,,:·· . . .. . .
. . >·:·:··,.,: .
. .•.'
<.
.
'
.:'· ..,
. • ..,> . •••.
·.·; ··.· ·,.... .. . . ·... , :~

. . . . i ;...
·.., ·, . :
. :;. : . ,, ···.; .
.
. ,:,.: ,, :. . '. ·..:· . ·...
,, .. . :. . ; .
. :·· ·.. . . : ~·
.,-,:::·:··· ·~1· ;.: . ··-. •. -:.
i.(':_ ·.: ,{/. . •,• ::> . i/( .,. ·-·~> >): . . :_;. ·:'. . . '''.\/?;t\({" . ..· .·· ·:.~. :· .: ..
··"> ...., ·:,\•

The . firlit. ·umn i$ \tb~ . · i,r.<ipQtfipft."'/


.,
.. :
t~m;J. J
• .....
..·. .· -s
•.. . . '•
·..... · ..:

and. th<t:s~·tli,;,.inti,iiijf t~rm. th~ ..


Ci ;. );~Ile i;s,m·~'."1 : :J:lhat the ~ .
.. . ..

c: ·1 . .
..• :· t'
•.. ..
:~

::. '
•' .: .: · :;:>~_)'· ·\:-: . ::;\'
~lge.br., . ,t;i;gw.~. ·:qi!.:: ·: . /P r . . . . .: : : -.' ·
i; ;;
< ; •
. ..

-···:.>·'....
-. 41"'·,.. .
... : .. . . ......... : .
.•· :·kc;:+
.·.. ·.·.. ·Ts· .·. :
,. ,.
. .:.·
·. . . .
:· . ... '
. :~·
.·•:
·~.,
.. >

corttniiul;Ofl ' Pf · the . tntegrar ection . ....


:, . . . ~ : ·:.fl,.,:l.·1;· , . .... ·.>.•··
·.. ·.;>.······{(·."':
..... . . ,... :· <· -~
:. : . . /:.. ~
• ··. :·:: . . :, . •. :,

·'
.

..
••

. · iribieil~IJ~·-·:1i<i

t~ Cf::Onttoller gain .· Is·:··


. .;

.

..
. '< : lncreased·aift!J:·rleicreases es. the integra·, .•
. ·.
~:

z
..
; tim.e is incrffs.fic/1;:/
..
• • ,
.. ~ .
•·...

,,::·· ...
.. ···•· . ,. :··
.......f.:. ..·. •.-· ,, '. :. . :' .,,,,
t , :· .
..
:· .. ·...,., . ' . r: ,. . e, •
..:,

-»,
. .::·~· '··
. ,, ·..:·· .· .
,·. :· .
. . .... ,.
. , ... ..,:·.

·,
..... . .

..,. . ..~

...
1 . ::~
. .· ....,
:..
-..
?:
.'

or ·.. ,... :

. ' :·
..
..... .
. .. .
. . . . .,. .
. . ...
.. . ..! : •
. '. •: '>·· . ··,
.. 1...;
.
,.
...,. : ... -· :..
! ',.
., . ...
. . ,.
i .. . ...
.. .
. ..•

.... ..

. . .. ... ,: : _:· .. . . . . .. .·, ..


....... . ·•. ~·: . . . '·.· ..., .. •.. ,...·.. ·, .-: :!; ...'... ,·. .
,•: ....... :·. . ., . . ·•. . ,:'··\: ·.: ,., .... . . ': ;.. .
.:.. .. ... . . .
'.

.. .. ... ..
. . ..
..
.
. ·. : .
:. : ..... . /.: ::·: ,' ~··.:.. . ' ....... ; : . .· .
'::.
'
<>t· ···;·:.::.:,. :, ..
.
-,. •·.
... .· .
.· ..
. ::: : . ··. •:, :s ... , ·..
..... . .
>
·:

. •·,
. ..
·. ,· .• . ... .. ,.
,...·· . .. ... .. ..:: \•: .- ·:. ., .
·.·
. . ·. '
..··. ,.... :..·.. •·..·:.. . . . ,· ' .
.., ..: ... :.'· .....
,
' . . . . . . . ....·...
·:· ,·,~~ . :
. :, .. .. . ....
~-·

.· ,. ..
·' .•. . ·. ! ., • . .. . ..:, : . ·.. ·. .... . .
. ·....
·:.

, ...
... ..·..>·
. .:..... ·.\
',
..
<
.
.. ... . ·,,. .. •. . -. , .. .· ..;,·,·...•,·-·. ....,.
.,. . ..
<,''·,,, ..·.f,. ' ". .. . ·. ... ; ..
.' . ..... •
···'· ·~ .. ::
•' •• " >
. :. ·"
·'
.,
:
':
. : ·~ .("
·.:~· ··.. . ·. • >
. .. ••• • • :
..
•·· :
,: :. . :t ·<.: . :

. ..' ..
•. ·,',.: ., •f :~·: ·~ · ::.. .,.
~.·.:,.,.,~.
.. , •
(•
..... : :":> .·.·. ; ..· .. \
. . , .. . ..,.~ ~ ..
...
,... . .. .
.. ,• • •< .,....,.' ··,:..
..
. :,. •. ·.;. ,. ,... ·. . . . .. : . .,~ ·:·. . >
....-.... . :•· . : ·~·-: ...
·:::... :-: ··::. ·..
: .
·':.
.;. (
.: . .• .. ,...
.......
.. . . .

.·: .. ·.: ·... : .
·:: . . .
,··· ··, ..., .•.
., ·::· . . ...
·.•., ·. ..
.,., :;·· . , ..... ·.. ··..... :·,.. ..... ........ "' . .,
J .. ~ ·• ·.
..> . ·,:., ~· . :·, •.. .
. ·~::
··· ..:- ... ' •',.
··.
. ..
... ·~.
. :·
·.·
·,

,• . :•· -: :· .. '),.... ,· ... ·.


:···,..
·.. ·.... ... ·.......
.. ... , ; .. ·. .; .. ..·.. ·?·.. , ·: ... . •.
." ... .;,
.. ,
·, .. .
... ..... . ..... .....' . ·. •
·. >,.
.. ·.·
.:
,... ..
,>.~ ·:. .·
.. • . ... ~.. ..
..... . ,.
. ~· . :: . . .
. '

...·
,·· ..... ·:',
.. , .: • j.
·,,
.. < .
..
. ·'
>

·. . ··, ,·,.... . ·: .:~.•.: '



. ... . :
.~ .. '\ ;:
..
·: .. • •>
, ,, .. .
..
;
... ~
. . .. ..
'
: .
.
~ ., .
. .
. '· ;
.• •.
.
.: , ,...· ..,..:. ...
.• , , .. .
• '· . .. :.·.. . .. ....•>
< • . ,.
.. .:.i,·

.
. . ·· . ...
:.~.
. ...
:.
' .. . .,,.. ··, :. .... . .,. ~·';'.~.
.•. . . . . . .•.
'· ., .·• .
·., • ..t '· • •• ' ·.
"
..
·- ·t s.,.
... . ·;.,)• • ·: ·, . . '· . .,
'· .•
. : .: . ,·
.>·
..
·,

·.
. '··
'

...,

,.
' .•
,.
.. . ......
.. ·.
..~
..•. ;.: . .' ·. .. ...
'·(· :. '; ..·'.\.
· ...... !

,,,

. . . •... ·. ;. .

• >
·.' .·.·
f, •• :·. )

...


•..

.Chapterl, Getting Started 11

·. ·.,. .. ' . .. ... ,: .... : . . . .... . . :- .. . .


.· '. ~ · .. ·. :~( ·., . .. . ·_::.. >", .: : ... '· . . . · . . .> ..
, •• ..!
.

. . .
..', .:· '~ . . . . .. . . . ·.
. . .

Fortunately, only a few terms haveevolved valueand thendecay back to .some lower
over the years to refer to the function of de- steady state value. The amount of the steady
rivative .. action, · and· the scientific· term de- state change 'is that .due ·to the proportional
action only. You might ask · why . the
output is changing between . ·its peak· and
its final steady-state value, when the
. .
error is not changing, and therefore there . . .· '• .·

.-.-.. Derivat,ve Compo.nent should be no output component due to-


1- the derivative action. This imperfect
Q)..,
=::,
O ca.- derivative action is a practical matter on
c:, . -·
a..., -
. ... . . two counts .. One . is that it .is physically
8 O. . . .: - . ....-. Proportional _Component
impossible to build ~ . mathematically
. . . I , · ..
. . ..
- perfect 'derivatl~.· funct16n,.tarid two . is
c-,

I that 'you,_ want to - even if you ·Wa.rtldh't


could. -:At this po.ittt~ pleasesimply accept
.... .

I .

. . both. points. Derivative action is .. de lib~


. .

.. . . I
2
~ --------------------------
...,.._........
. . erately imperfect but. achieves most of ..

w
~
. the . desirable ·results sought
. . . .
. when using .. .

._ ......... . ·... · . the derivative function ..


··Time
. . . .
. . . . .

Figure 1.5. A proportional-plus-derivative controller will . . . .


. -. Figure . 1.6.. illustrates .: another way of
respond to a step change. in error by adding to the . conveying what . the derivative. function- .

proportional component that decays with.· time. The does. This time, instead of introducing a . . . . . . .
· longer the derivative. the longer the decay time.
. .
. \
.
. .
. . step change in error,. a ramp change is
rivative seems to have held sway. . ..

Rate · (and Pre-Act, Taylor, starting .: . ..

about 1940) have been used. I will use · J -s


. -
. · Proportional-plus- .
· · Oerivative. Response, · · .·
derivative time and· derivative action. e.., .&::, ... ,.. -- -
,..
It is mathematically the opposite of g o - . ~ . .

O
integral action, but while we might r I __ - - - "' · I- Lproportlonal-only
_ _. - .. Response
have an. integral-only controller, we
. ...,____,,, - , . ---1 Derivative I
. .

would never have a· derivative-only . l .... ,..... Time· .. ·.'1·. . .

controller (though we could have a


proportional-plus-derivative control- ·. I
ler, with .: no integral action). The ·
I
, .. . .

reason for this is 'that derivative action


e
~

only knows that the error is changing. I ~.


w I
It doesn't know what the setpoint
actually is, so by itself it cannot
control to a setpoint. Time
Figure 1.6. A proportional-plus derivative controller will
Figure 1.5 shows the step response of respond to a ramp change in error by adding to the propor-
a proportional-plus-derivative control- tional-only response. The amount added will increase as the
ler. The output will peak at some derivative time is increased.
·.
.,,:,
. f:·
..
,:·
-~,
•!,<
,.
">
, . . -.: :.-:·:.:· ... : :. . .... .;
"
\,

.. .,.
i

12 Chapter-I, Getting Started . ·.


' .
f;.
..'~.. '
..·'.,i
..·'.
'•
-·.·~
used. This is simply a change that continues derivative would have a transfer function '.
~

of: ·. · · -:
'
at a fixed rate, rather than all at once, as for
. .
. . :.,~!.}, ,; :JJ: . .;: ft' ? ·.... :., . . . ·•
.,'

,.

'
a .step change. The derivative· function· adds .'•,
'
.,

to the output that would normally occur, in -,


i
·,

·:
.'}
effect advancing the response by an amount . . . . -:
\

in time . equal to the derivative . time. ... The transfer function for .~ proportional-
.. . . . .
"•;
..

Actually the advance in time is notquiteas plus-derivative controller·would then be ..


·.
.,

· this: •·
,
...
large. as the derivative time, . which is a
result of the· deliberate . imperfection in the
function.
where Td is the derivativ~ tim.e and s, as -:
·;

· noted before, is the d(!riVative function. ,,.


,.

.. :,. .
.
You
..
already know that the; derivative .

.,
·.

function is not mathematically perfect.


Actually the way the algebra is written is
to. write it as a proportiBn,al-plus-deriva-
<

tive function, with the proportional part . .


:>.;

. :;

having a gain of one. Here is the trans- ..;

Derivative action has the potential to fer function typically used to describe
improve performance but is unlike pro- the (proportional-plus-) derivative
. ..\
. ·.:,.
..
.,
'' . . . ... . . . .. . . ..
portional or integral action in one important function:
'
'.
..
-~
aspect. Withthose, it is· mostly amatter of '·.
·'
,'

..
using enough but not too much. If you did .,
·,.
·:,

not use enough there would still be '


'
..
.
··:;
.
" ..
beneficial action, and performance would be ·'·
;

'
better than if you did not use them at· all. ,.
'•
·.
. . \ . . . .
f,

With derivative the problem is ·. still one of The numerator is the ideal part and the · .

·1
. .;

-;

using enough but not too much, ·but -if you denominator is · the practical necessity. ...."
. ...i
..

":-:
do not use enough, there is no benefit at all The-denominator tstne transfer function ".l

-~
··1

and there could be some harm. If you use of a lag, which will be discussed more in ·.· !
'
·'·•i

just a little bit too . much the troubles the next secion · on. filter time. · The new.
. . . .. .. . . . ...
:;:

::
'
....
. . .
parameter, Kd, is known as .th,e _deriva~ . '
''

increase a lot faster. than the benefits. IF '


.,

tive gain .. It determines the .height of the . :~

.'

USED AT ALL,.. IT. HAS TO BE SET peak in.Figure 1.5. If the derivative gain
•.I
.,.;,
.
INTELLIGENTLY. .. .

is 10, a typical figure, then the maximum


. .
'.i

·'

the derivative function can magnify any


. .

rate of change is 10.·· l·:


...,
~


.,'
·'.J
·,

The algebra for the derivative function It should be remembered that j\vsually .:{'
•L
'
••

gets more involved than what has been the derivative function on a digital ~
~
.,

controller is set up to act only on the


.':,
..;.

presented up until now. The ideal ...'~'


,
controlled variable, not on the error. .;
.
•'
..

.~
.'
,..

'
,.:
'.
..'
.,.••
·,

'..
<

·'

1

·'
,
. ·.
...
.:.

..
.~.
·;:
:·:
.,.,
.'
.."
;

....,
.. ,
(

.;
,.
.,
·.
!
··:

.
<•
k
..

Chapter 1, Getting Started 13


,.
',

'3 ' ..
. .· .
,>

. . .
.
.
.. .'.

With many : digital control systems, die . . . . . .


$.0you . do not see the waves - generated . by
menu for controller settings includes a set- the motion ofthe.ear.: ·
• < : •• • ,· •• • .: •

ting iot.filter time. It is not normally in-


cluded 'in most published rules ·for· tuning, · i . The task ~f sett~g ,th~ f11ter time is one of
because when they were written there were .' . : us:ing:as,;mu0:h .asyeu dare. without degrad- . ' .

only analog (non-digital) controllers around. · ing the P'-'J.'fe>.rmaqct.'. of the. loop. Too long.a
The filter is a digital controller phenomenon filter time will.affect controller settings and
and . helps compensate for the small also •. make. ,tlie cOll~oller •. slower to : respond . ~ . ··~ .

. . . . · . to disturbances. ) The use of it· at all is


' .
. . .. . .· . . . .

·. · "; · · . · ~ctuat · Change in . . . . likely .•. ·. to . $~a1.:t - .: ~ . • _-lively · discussion


· . ·. · · · · · Conttolled Variable .
. . .. ·,.:·
.
. ;
. ~-
~. .
·. •·. . . . ·:.. ·.
.
.. : . .
·.•· · between. those Who grew up without its
,, .. .
. ·..
.
,. ·. I· ~< • evailabilityr- ~d .those who grew up
CD CD
-- - - .
... after its :availalJility. . . · ·.
=15 I .......
ec ca...
.,

... ·- I
.Jll"I
oJ
0 I . I 0.63A
. ..

. . . I . .
. . . . :
. .

..__ --------- The filter is just one name tor a


.
· . • I . · · Filter Time
. .. :. . '· \ : . . : ,..· .
very simple and important. element
. -. : ·.·... ,· ,. .• . .. •. ··rime
.
~
..· ·..
·.: ...
~·. . ~.
·.
":;'·
.. -~· :.
... ··::.
:·: .
. .
.
: ..
.
. .
:. ~
.
··:
,
.·. ..
.. · . .· ..
.
in control loops. It is really too early
Flgure:·1.7.· Thif.eff~-;ef:.:the1.filter time in a digital controller. in the development of the subject
is to slow dovln the :cmH·,ge-~the ·controller sees. . .
to get into it here, but I will simply
give you the transfer function:
vari~tions ·.· m'' re;dfu~ the . process variable . . . . . . .

because of sampling and because . of round-


. . . 1
off errors. r,s+1

Figure I. 7 shows what


. .
it does to the mea- .
It will get discussed more later .. No-
surement; it slows.it down a bit, or averages tice that it .. has
. the same form. as the . . .

it. The gas gauge in a car is heavily filtered, . denqminator in . the proportional-plus-
. . .

derivative function.· . . .

. ·,
e.

14

... .
'

The filter action and the derivative action . '

are· opposites for all 'practical purposes. . '.


. .. .
. .

They can cancel each other. A filter time of ":


.
..·. :.. : . . .
.
. _;.·. ,/ . '
. . ,
.
..
..
·'
.
:" .
..

one minute will cancel a derivative time of • : •


. rile lilter
that hi:#$ just been discussed, •· .
•1, :

one minute, with the· result being essentially . has the transfer function:
the same as if you· had used neitlierJ This :i·s
still true. if the filter time> is set toenly half 1 . .:

or a third of the deti'Vflti~ ~m.ef land · it .


• •
r,s+1 .
.. . .
.

makes ·no sense at .allto.,S~t',~ higher.


Frequently the <feci~i~l!t:'.~S E¢llfh~ t~ llSe When you neve . :,a, ; filter and the de-
rivative . function in a : controller the
. one or the other, ' bot{, ll0t1'both. :f>eriVative .·. -. , -, · , . . . .

action bounces the--:~ntrollet·oUtl>ttt -rnore . .• . . . resulting transfef.function·is obtained by


·. ritultlp/ylng the two together, which
than · . When .' it · is • not : Used, Filter •· action ._· . . . · be.comes: . . ·.. · . ·-. · :'.:,., ·
· dampens this f><>tuifing, but if tOO .much is · . ' : : .
.::.
~
:' .
. .

usediiiitl~~perforfu'ariceand cancel .·, . . . Tds+1 .. ~ .

the benefit of derivative action. . . ..


.:... . . ..
. . : r,s+1
. : ., . . .. . ... '· .
. . ,.,· .
. .
. . . ·.... . :,,,· .~ . ·. ·.. · .. ,.,. . .. . . .

· ..... ····• FfOfn thiS you,CBnsee that if the two


times are;;·! ;~et the same, then the
This is a place Where the algebra gets · . : . , . numerator:; and:-> denominator are · th~ •.
quite neat. . The prop()rtional-plus-<leriva- . · same; and the whole · transfer · function .· ·
tive function has this· simplified transfer .. reduces to one, which is no dynami9 ·
function: ·
. . . . . . <effect at all. . . ' .
. .
. . . .
.... ·. '·

..
.
. ,.. .
. . .

. . . .., .
, . . . . .·. . . ' ' . ....
. .•. . . . •. .
.
.
. • • >
' : ••
,. . . . ..
. .
... . . ·:.. ..·. ": ,: .. .
.. . .
~ ~~ .. ...
,: ' . . ·.
.. ..
..... :. -. :· :1 . .t. .
., .··. . .
.'
.·. ··,;:;.:~· ": . . .
. : : "·, .
··
. . . . ,. . .
. . ..
. .
'·.

. ·.. .

.. . . . . . . ,
.. .. . . . . . . . ~ . . . .
'··
.
. . ,: . . .. . . . . "
. . . . .

.,

:. .. ,


·.
' rs . .

,., . . : . :' . ..
. ' :....
;,
,. "\ .

.,.
·'
·.

::,
...... ·c.·f.'.. .:
.. J

'!

,.
..

. . .

I want you to understand not only 'the me- combination at times. The open-loop
chanics of tuning (tuning by-the-numbers), methodis a bit harderlo use but yields more
but what you can expect from tuning. What fundamental TIMF! and ·AMOUNT. infor-
should you be looking for? There is ·a basic
. . . . . . .
mation about the process. It is essential that .
. .
.
. .



dilemma 1n explaining tuning rules, •if part the persons involved feel confident in per- . .

of that explanation is committed


.
to teaching . .
forming the necessary procedures. With
what the tuning is, doing. To understand either approach, open- or closed-loop, there
~ -
. . .

what the tuning can do, you have to under .. is a section on what to do, and another sec-
·.

stand the importance 'of lags in the process. tion on how to do it (procedures and
To understand the importance of the lags, techniques).The key is to upset the process
you have to understand what the tuning can enough to get the information you need,
. .

do with lags present. It IS like the control without getting into trouble.
loop itself. . where do we start? 1·· have de-
cided to · start with the · tuning rules. If you Tuning settings can also be calculated be-
are new to the subject this will almost surely fore or after· a:loo,p 'exists. 'This writing deals
only with tuning-: in the field(thoughtuning
'
••
require that you cycle back. and forth m your
reading between the tuning rules and the for level· loops might be calculated at any

lags. time).

The basic principle of tuning is to set the •

TIME and AMOUNT parameters of the


controller to fit the TIME and AMOUNT
parameters ( called dynamics) of the process.
Tuning procedures are procedures for
learning the necessary dynamic characteris-
tics of the process. To do this you need to The tuning rules assume the controller algo-

upset the process. While there are more so- rithm IS the interacting type. The signifi-

phisticated, computer-based, ways to learn cance of this distinction IS discussed else-
the process dynamics, there are two simple, where. You may safely assume equipment
• •
time-honored, ways. One IS done ID auto- of a major manufacturer is of the interacting
• •
matic (closed-loop), the other IS done ID type. Digital systems often provide a choice,
··- manual (open-loop). You should understand so choose the interacting type.

each of these approaches and use them Ill
m· . ·,,:, -,,.,.' ·.. ·. :. "." . . . . .' .
:. ·.:.... ':: - . :. . ~~.,;).._ ·.·:..;. .}· ·.:.: ·.>: ...
rr
.r. !
.. : '"

...
t, .

,.
. . :· . '(ff!Y.

16 Chapter 2, Tuning Rules and Procedures


.• •
.. ..
: ,
,·. 's..
. . -~..

Now I know you're are eager to get into the


.'
.' rules themselves (you may have even 0 You should know and agree on who is . .
skipped chapter 1 in your hurry). Be not. \\;~,.';{Yr goiqg to .do tJie ~Wal adjusting of the
·.'<; : •

hasty. Would you drive a car .without,~itst.,·;~* •· -:::I(" :: .::.:' · con{t91tepl,8Qttings,. and be assured that
learning the rules of the. road? You;·,coiild .':!. · .•. · tmit l)ersOn knows how to do it physi-
crash yourself, whether learning to . drive .or . . . "· cally and won't· become confused if . . . .

to tune. This section .on. pre.fa~iQO may · · something has... to be done in a hurry.
seem like overkill; but.;:<it:\is:;.y_ifltl;iiYou :\Viti
. . . . . .
. . . · : .. •. ·.
. :. ."· . ·~
.
..·
. .
:·. <. .
.\ .
. ..

live to tell about it, .both · literally .and . ' : 0 You should kn~w and agree on who is
Po litically · : ., : · : · · · :·/~; /:·· : · · : . · · '. , · . ~- · .· · going, to · switch between manual and
~ .' . . ,, .: .
.
;. .
.
. ... . .
,. .... ..
..: .
..
. . ,·
.
.
.
. ·. .::. ~ .. .. . . . .
..
'. ·. . . . .·

.
. . ,. ..
.
' . .
.
·::..
.
. . :: ::,... . .
.
.
< · · automatic, and who is going to make
Before ·you· tune a controller there are · · setpoint changes if needed, again
several . items _ you should have reviewed, looking for 0confidence that proper · ac-
... . . .
certainly with . yourself, and possibly . with tion can be ~n in a hurry if needed.
. . . .
supervisi~n if that is appropriate for Your
(

. . .

location, If you . 81'e experienced . at . tuning, 0 YoU should have •. knowledge of the
some of these items may have · become . . . safety interlocks . and any other safety
, . . .
second nature, yet it would be prudent to . con~ems for the process.
use the following as a check Iist.. : . . .
.
. .
0 You should pick a time to tune such that
0 · You Should have some id~a of what you
. . . . : . . . . you will be available for a reasonable
are trying to accomplish, what you can. . time . afterwards, to be . contacted if
. . . . ~ .

expect· to · accomplish,
.
I hope . · this . : . . . . . . . problems arise.
booklet will give you that understand- . .

mg,
~ You should record the existing settings
. ; ·. . . · . and· the controller output, in the. event
You should have some idea of how fast
. . . . . . . . . . . · · · .: you want to return to them, either in a
. . . . . . . . . . .. .
and how far the process is going to . · hurry or .simply to leave the· system as ·
. . . . . .
respond to the controller _output.·_ Will
:

. . .· : . you found it..


you be able to restore stability with
confidenceifneeded? ~ You should make sure that changes ate
communicated to all operating people.
0 You should know and agree on how . If a . log book is kept for this purpose,
much change you · will : allow in the .

use it..
~. . . .. . ..

process and. in the controller. output. . . .

Get all the information you canhere,


Chapter 2, Tuning Rules and Procedures .17


. . : . : . . . . ...
.
. .. .'· : . . . . . .,
. , . : . . .. . . ·. .. ' .
.

' .. . .. .

The Ziegler-and Nichols methodconsistsSf


. . .
Ideally the filter time· should be · set to zero.
these steps: · · · Its purpose is to . reduce the activity the
valve· sees because· of noise, In .practice you
1. Turn··. the integral time to as high a
. . .
may increase the filter· time until the valve
number as; possible (often called "off"). . .

. . . . . .
motion -is·· as • quiet as you would . like, but
With digital controllers this is usually ·.

generally no more than . that recommended


to set the integraltime tozero, which is , : . . .. ·:. . .

not zero time but a convention


. to. indi-
. . . .

cate no integral· action. Tum the de-


rivative time. (if used) to as small a
number as possible ( often called "off").
2. Increase the gain in steps . untilthe loop
. cycles (there is technique to this). It is
a fundamental ·truth that any 'control above. Actually, if the above setting is used
loop · will cycle if the contro lier gain is with derivative action, the two tend to and
.
made high enough. may cancel each other. You will understand
3. Observe the period of cycling, P0, to be more about the potentially· detrimental
called the. natural . period. Implant this effect of the filter on performance later, and
securely in your · mind. Nothing so .
a. special: section on the subj.e.ct Is. included -~. .

· simple is more important. Note also the toward the end of the ibooklet4' , , . .
gain at which it · cycled, Kcu, to be . . .·.: .. " . r ·::: ' :.·:::'v;,:;_::: ·_:_:_; , : :~:'.:'~ .:- .... ·: •. . ; ...
called the ultimate.gain.
4. Setthe controller settings to:
Kc = Kcu / 2 (aggressive)
Kc = Kc0-/ 4 · (conservative)
Most· references, ·including Ziegler and
Tj=l.2Pn Nichols, · recommend . slight variations de-
Td=Pn/8 pending on whether the controller is ·P; Pl,
PD, or PID . In principle the absence of inte-
Tr<= Pn/ 8 gral action would call· for a slight ( 10%) in-
.where: crease in gain, and the use of derivative ac-
Kc = Controller Gain, % output I % input tion. would allow· a slight increase· in gain
.. . .
. .
.
. ( 10 to .· 2.0%) as well. as a reduced integral
Kcu = Controller gain that produced the time (30% ?). I feel these relatively minor
sustained cycle, % output I % input. modifications are within the normal toler-
Ti = Integral time, minutes ances for setting the adjustments in the first
place, so no special emphasis will beplaced
T d = Derivative time, minutes on whether the controller is P, PI, PD or
P0 = Natural period, minutes (from step 3, PID. The directions these. setting may be
discussed below). changed are worth noting, and perhaps as
you become more familiar with the whole
T f ,=:= Filter time
concept, you will understand why.
........,.. -.. . .
.
.· :
. . .
. . .:.:~:-··. :.··: .
. .
.. .f .
. ..
'; .. · .
. . : ··..· .· \;
. ~: ...... ·. ,:... ·.':'·~ :. ·. .
,.. ,. .

18 Chapter 2, Tuning Rules and Procedures

'
The basic idea Inthe closed-loop method of change, most likely in the opposite direction
-,

tuning is to. get the loop to. cycle .without to be the safest. Once again, you are looking
getting into . trouble, ob.serv.e·tht:1'f#l.uralpe- for any signs of a cycle in the process. Con-
riod and .• the U,tlmtitega/1' atF;1n3.t::.point, . .
tinue this procedure of increasing the gain
and . .thenrback . off... several ·. ·.preparation
. . and testingfor stability, until you see a sug-
items. related .t~ people · and safety have gestion of -a cycle .. At that point you should
already been listed to be observed before changegain Iess than a factor of two, per-
~ . . ·. . .

any tuning activity should be performed. haps · by .· only · · 50%, · and continue the
Once . those·.··· . preparations . · have .. · been
. . .
. .

procedure as before. As the loop becomes


. . .

completed· .the next . · step is to ·. tum the


• • • N • •

more · oscillatory, make smaller changes in


integral and . derivative functions off; or in gain. If the controller design permits it, you
the case of some controllers, as far toward may often safely change the gain while the
off as possible, This is to have the integral
. .. ·. . .. .. . . . . .. . . .. .. : .. ,, : . . . controlled variable is moving. This saves
time as long as possible and the derivative making a new setpeint change to disturb the
time as short as 'possible.
.
· · . . process, since. the process. is already being
disturbed, and-that isthe purpose of the step
in the first place.
'
. . .

Be· alert
. .
for any. variations that seem to be
. ·. ~~

the start of a cycle. It is useful to watch the


controller • output, as well as the controlled ' .

. : . .. . :. . . . . . . variable, when looking for signs of cycling . . .

The next. step, if you: have no idea "What the Often these · signs can be seen first in the
stability with the present gain is,'; would be controller output. If the controller . output
to make a· small setpoint change. Make it in saturates, that is, goes to a high or low .limit, . .

the· direction judged to be the safest and ob- beware. The results may be invalid for use
serve the response · for · some . signs of in this procedure. Try reducing the size of . ..

cycling. This. is. the time. when it is very the setpoint change, or try to· introduce an
helpful. to have at least some idea of what upset to perturb the process, rather than us-
the period . of the cycle is likely to be .. · How ing a setpoint change to do it. Still be cau-
-
you can estimate. this , will be 'discussed in tious if the controller output saturates.' If it
chapter 4. If .you expect, for some reason or saturates in this test, it might saturate in •

other, that the present. gain is· fartoo low, normal running, and .that could (not neces-
then you might save time by not performing sarily would) create stability problems.
the setpoint change but simply increasing
the gain from its present setting· by a factor Be prepared to· return to the original gain
of, typically, two. It is generally a waste of setting if the cycle starts . to increase in .
time to make small · changes. in gain, like amplitude, or even to manual if that is ·
10% or 20%, on this first effort to learn how judged the safest thing to do. Time the
close to cycling the loop might be. · .· •· · switch ·. to occur when the . output is
approaching its original value. Some
If this first set point change does not pro- controllers cannot be switched to manual,
duce anything that looks like the beginning and some bump the output when a gain
of a cycle, increase the gain by a factor of change is made, so take these potential char-
two, and make another small setpoint acteristics into account as you adjust the
,

Chapter 2, Tuning Rules and Procedures 19


. ·. . . . . : . . ·. .·. . . . . . . • · ...•. · • . ·.· ... · . enough 'to the true natural pe-
... ·. .
· ·. '"-: ... · _.· ..·. · · . · · Kc· = 1_.· .·:'·,_.,:·_.·.:.·• _:.:_·Setpolnt
. ·. ·. . . · •.·..• · · :-· ·• ·· · · r. io· ·d· · . · . fi. ·or
.' · . determin
· · ing
· · controller
·
.· . · . · · ·. tc ' . ·· .·.· · • .: i · .• · •· .. · ·. • · · · ,/ · · •. ·• · .· •· · . · ·. . settings; The . · · reward · . . for
_-.a
i .. . .

;'(_'· . > _> ,· ,; >6< ·. · .. ·. · determining the exact gain that


e ..
...... ,-

I
:· 'v .~

. ·
. . . '.;

..
.. . .

·.· · · _ . '.\ i-, · · . =


• . ... . ·:'"'>· ·:: . . . ,

. . will produce a steady. cycle is


·.. . .

c •. 6.
0:>
. .
CJ .
.. ' ·..
.

. · .·. . .· ·. ·. . _ . risk that thecycle'witl continue


·. I . . . · · · · · . 1
·. Kc - to increase in amplitude and

- I . .. ·. . . . . . Create a safety or quality


. ~·. . . · · ·
. .
problem, ', · • · ii. ·. . . ...
-- ......... ~ ...... - .ir.:
O . .·. 20 : · . : 40 · ;60 80 100 120
. . ~~:

...
· · ·.· · · - · -.· · . Time·. . . Having· . then .determlned · ·· the
. · . · · . . . . ·. . . . . . · . . . gain that will slls~in? or . almost
Figure 2.1. With a . proportional-only co~troller t~e error (offset) is sustain . a . continuous cycle,
reduced as gain- iS increased, but at the pnce of an increased tendency OBSERVE . THE PERIOD
to cycle. . . . .: · . . . .. . . OF THE OSCILL\JlON.
gain, · If the controlled varia~Ie is already in . Then promptly . cut ( the . • gain in . h~lf to
motion, it niay not be necessary to upset the · achieve stability. This' period is called the
system with a . · setpoint change, sin~e the natural .. period. Some writers, including
only purpose ofthat is to see some action. Ziegler and Nichols, call it the ultimate
· · · · period, and the gain that caused . it. the ul~
Fig .. 2.1 shows the typical set of response timate gain, This gain and period are then
curves to be expected from this procedure, .· plugged into the formulas given, to establish
If the gain is · significantly lower than the ul- •. · • the controller Sfflitlgs. Notice that gain may
timate gain, the process will respond _witlJ., .· ·• .•· be; set for tt range ()f' values. Once 'the
no hint of cycling, arid' also the process 'Y!lJ . ·. ; < · · settings hive· been' made, confirm that they
change only a small fraction of the aipotinf are acceptable With a small .setpoint change,
requested by the change in S(:tpoint,: One or observe the behavior tinder operating
point to be made .about Fig. f.·I·. is tlJ.~t the conditions to cOrifihrt that all is acceptable.
gains and offsets (errors) ~e sa'm_p_fe~. only. : . . . . . . -~ .
· .
T h ey d O no.t DeCeSSarl Y re_preS . · . · · · · 1 · · · e · · n · . .
t · n u un ·e
. ..
·r1 -- ~~-~~~~-~~- · · · · · · : ; : , · : , : , : : , : : ; , . : , . : , : : : . ; : , : : : : : : ,
, .. ,.:··
. . . . . .
.. :. :gm1~1~grrirm.mtn}1m11rn~1mtmrnHmH1{E\t\t/ttm~ft:~1~1~1i11~~}:::1{tHlH.Il!·1j\\jjjj\l1111ll;
·,.·:. ..... ,.,.,··sf.·....... ,.,: ...... · . .... ,
~~~~~~~~

:ml1ll·llllUl•~:=:Jf/JillH11i~:~[~:;;:l·lrlll!ll11.l

Ca 11Y. Wh.at . You.. wi.: .' 11 experience . : . : · : ; : : : · : :


.::,., : ~ =· ~ :: ::: ::::.::::
.:::
. :
~r:::::.:: :: .
· . · ; : : ; ·:·:·:~
.
·=~:
.
::~:
.
:,
::
· ~ : .
·····": ; :
::::.:::::==~·=·: ;.:; ;·:1;:·::,
..,.,, ; . : . .
.. ,,.. : : ..
.........::::::::;:::.:::::::::=:
.... .. ·
: .,. · ·: -,,. ..: ,.,.::::
..
: ; : · : : ; :
==;: :: =· =·= = = = ::::
;
,.::,,: :
= ::;:,:::::: ;::
. .==· . :;D::.
==========
·.:: .. =. ··:::::::.. : ; ; : : : ..: :; ::;:

•.: . . . . . . but o
;;::;:;:::::,·

-. , '. . . th. .ey : .. . d . . ; : ,


11111
; : ;
m
: ,
................ : :
1111
: :
m•!i: , : • ·:1, :... r,,. ·. ,; • "1,..·;· :::,i.
. • • :
•.......: .. , .,!: ; : ..... :[I/: ;:. ,rj·· ;· •·., ,. , -,,--=mi;
•t
. ·- . ; , : : t. ..:; ,:: ..,;ir• ?!i
. ..,. · , . , . 11·: ..; ·.: ,.: :: . •fiiii::t
::.......... ..
: · : : • : - : : , : . ..

rep. r'esent en_ d· s_ . · t 1·s· · ,·. no_.·.·.·t :.· _e· .-~_n. t·· to_ itnpl_y·_ ' ...............
gi iii ?i:::::: :·::::: i i;; ~~~?\ ~ ??~
~1 :11111111:::r:
·::: .:: ..
n ;;: : :
it 111111.ti ni 1 ~111·1 ~
..
;:]mi ii !·:::
f
. ;:;:
1~11~;~~11~;:: ::: : : ~~ ~; 1~ 111~~ 1111;~ i m 1~ t·l l; ;.u·i 111
'" ,·
:::";'::·::::.:.;:::
1 n 11~ /i 111i11.11111 i 1 I 11~J 1~:~~~~~f
tr.· 1· m_,·._.. , . , . , . . . , " '·························;···:::
:::·:::::.::;.::::::::.:.:::" ··:.. . ·..
, ,
··1··.. , . .
. . . \ .
·..· . . s··· .. ····::
: , . , .
th a· t ·the ·· ·100
.
· . . p
. ··.you· . .
. .
a r · e tun1·
. .. · . . ng , .
·w1 · 11 b e con· . , .
ser-. . . . .
· .
, , . , . . . . .
....................... .
. . . : .
. : ,. .
. ., .
· ,. "' . ..;.;. . . . ,. . .. .
: · , , . , . , · , , · · : ·
) ~·; ' :·,
::::::::::·::::::.:::::
. . , · . · ·
.
. •;;;.;·

. .
::
. ·:.
.
:::
", , . .
.;.,.....
·:~·:::: . .":·
;·:: . .:: . . .
..
·:.;::··
~·:·
.. .
.
. , . .. ...
. ..
·:;:;;:;•. · . ·::·
:::::: "::· :
.. f• : ·..
.. · •· · · · · · · · · · · ·..
. ::·•·;:;'
;
.. ::
..
;:
"
..:: ::·
~:
"
.,
:::,.
"
·..
.
.. :.
::
•.· :::
··m·
:
:· . ·:·::,.
·:
:. " :
:..
::
:.: ·
·::·::::~:::::.:::::::;;";:
.
.
:~:;:~:::·:::::::::::::::::·:
:

. ,. . . . ~. . . . · · . · , ; . . . , , . ,.,,,,,.,,, .
i'. ···:, .. ·· • :,. .· ·, : , ,., , , ..
. . ..
vatively · stilble at• a gliin · of 1 and . cycle at . a · . · · ·
. •. , , ,., ··.•.•.•.• ,..·.,.;. : : · ·.·.· :: :.; ::;: .
.·· .·. . . ·. . · · . . . . . .. . . . . · . : :}: : :: : n::::::::: . l l; l n l rill r:::::::: :.: : :..:.: : :·::
.. :~.::: :::::: : : :: : : : : :.:·;:
l.l·l.:::: :.: :.: : : : : :.: :: ::.:: ::: :.:: :: :: :: :.:: : :: :: :: :: :: :: :: :.:.:::: ;: : : : l l rl+l.l+l l l l l·l:: l: n:•:• ·.·. '·' 1~~ ;~~ 11· 1 ~· 1 · 1 11 :·:: :: ::: :: : : : : : : : : : : : : : : : :·::::::::::::::
, , ,. . . . . . . . · · · · · · ·· · · · · · · ::::::::::::::::::: •: : : : : : : : : : : ,.,., '·'. ,., : :: ::: ::::::: :: : : : : :;: : : : : : .

gain of 6. It simply shows ·that as you i~- . ·. Fig. 2.2 shows what adding integral (reset)
.
>
crease gain, two things will happen. One ts • action to the process used for Fig. 2.1· will
''·
't
••
,.
that the offset (error) will. be reduced. The · . do. Again use the numerical values to · un-

'l
f,.
other is that the response will become more ·. derstand · trends and . . approximate
f.

r •
oscillatory (unstable). relationships only. The effect of adding in-.
.'
"

tegral action is to reduce the error to zero


"

-
,,
It is generally of little value to determine the (eventually), and to increase the tendency to
-.
?'
,
" exact gain that will produce a steady cycle. cycle. The period of this cycle will be
The difference between that gain and one longer than that when only proportional
that produces a slightly decaying cycle is action was used.
small, and the period of cycling you will ob-
serve with a slightly decaying cycle is close .
! ' , . .·.
[
\;
.
'
.
·..,

-·~1
-;
~
20 Chapter 2, Tuning Rules and Procedures
'·. .
:~
.; .
.
. .
. ·.'.·
. .
.
.
.
.

modify the .: t:eSpQP$e:~:; :_:)ltld11cing


. .

'.:·,:; :~{!~
.. .

,......_~T1 = 5 gain increasesj


ing integralr-time'?b:lso. inc~s ·
T1 = 10 All Curves
for Kc = 6 stability-if t.he dampened period·
'
'
T1 = 20
is signifi¢~tly Ionger than the ·
I,,r
.. '
••
.
..
.
naturalperiod, If . the derivative
.
Setpolnt . .: . ..

,:i, .
.
• . •
.
. : .·, .. ··.··· \, .· .. ·~:: ,' ...
.
.
.. .·
. time is~:too long, it will increase
,:~··''•' :..
.
. ,\,

... . .. . '

, ' ' .
. the itendency tocycle, but: the
r . ' •
.

. .
period · will · be shorter than the
I .
. .
.. .
..
·.
. .... : . . .:
..
.. . ..
.

natural period, More will be


.
.. .·:!'
;... ....
-~.f~
I Tl = 40_ · . · · · ·. . .·. .. .
,•

.
'

.
. . .-·~. ....,•' ...·..,,t,,,•. • ..··:.1:'

.

·yI,
.

. >,.
1..• :
'

= .100
.
• • .• '•'· • • •• ••

•;·,.; .·
' • •

said about thislater in the .sec- . . ' .


., . '
. .. . . -~ .
I
. ., . . . .' . . ..
'l$•
•• ·(
• • • t

..
'

tion. on open-loop testing., · · · ·


~
. ,. ....
1
.. . }.
.. . .
. . ~· ,: .
\
.· .
.· .. · .·~··,
'
.
~ .

.l, .

'I1 · . . · · . . . . ,,
.. . -,' . . . . ·.. .
. ·. · • ., .; ·. · · '; · . ·• , · provide programmed aids. for
• •

l
1.
)
I:
· · · · .,. . . . . .. : ·. :'. . : · · . . . .·~ tuning. · I • do n~t know enough
about "these. to. comment. There_
(

· · ·: : •., ij .: ., •·20 . ·. 40 . ·.60 .; ·, . 80 ·100 . 120


'


.. . . .• · · -. Time · -; . · · · · .· · is i distiribt ch8.Dce that you will
. · ·· have to knOW quite a bit . about
. oect&asing the integiat titn(t increases the tendency to cycle, and at . use them effectively.
a ·longer period than for preporti~_al-~ly. _ '_ · · · · · · . . .. . .·
. . . . . ... ,.. .. .. . . .
.
;
:
.
;
t

. . ·. •· .· . . . ( . . .· •. · . · . ,. . · • . . . . . •. ' . ' • Firid some way to record the ultimate gain


If the performance With 'the selecte(f ~f;ti,ngs and the natural . period, so . that a person
is judged not acceptable it. is· ~ot generally foll~Willg i~"your tuning· footsteps willhave
necessary, or ev.en desirable, tc, r~peat.~e that· information, · You might want to keep
full testing procedure. That is, it is notnee- . your riWII records as well. It would ,also be ·
.. ' .
. essM>: . to tum the . inte~al 3:11d · derivat!ve, · appropriate · to record something · . about · the ·
functions off. The natural period, on which . operating conditions, such ~· th~ production
these settings are based, has already . been . . rate, setpoint and controller output. . .
established. At ·this point · · it is well to · · · · , , · · , · ·· · ·· · · · ·
remember that the tuning rules · are for A finlll steJ) tiefore leaving the scene is to
typical loops. To build . experience· and advise 8.11 who might be concerned, what the
the ref ore confidence in tuning, you may settings were before the tuning effort, and
then alter the settings · to modify the per- what they are now.
f ormance in the direction desired, using· the
. . .,.
concepts presented in figures 2.1 and 2.2 to . .: . ., .. .
....
. .
.
,' . . . . ·... . . .

.. . .' . .

. ,...... . .·: :. ·.. .


. . .. ." . ., . .· .:
.. . ,, ..., . .
. . ..
. . . ... .
' '

.. . ..

..

..

'
'

Chapter 2, Tuning Rules and Procedures 21

·' .

The method proposed in 1942 by Ziegler


and Nichols was the first procedure ever T·1 -- SL
proposed for determining controller settings
from an open .. loop test, though other authors - L/2
have honed it a bit in the meantime. It was
a stroke of empirical genius by Ziegler. Tr<=L/2
The procedure is to place . the controller in
manual, and when the process is sufficiently R -- Rate. It is the change per minute in the
stable, make a step change in the. controller process variable, (expressed as a % of
output. The process response is expected to the transmitter span), divided by the
look something like Fig. 2.3. Ziegler and step change ( expressed as a% of the
Nichols called this the process reaction controller output span).
curve. More recently it is called the process
step response curve. If the process levels L -- The apparent dead time

out, it IS called self-regulating. If it
would not, it is called· non-self-regulat-

tng or integrating (much like the Slope, R /
integrating function of a controller, See Text.tor<-.../.

which keeps on going if the error Definition
Controlled Variable
remains). The open-loop rules i May or May Not
generally apply whether the process IS • --
-.a • Level Out
e...c -ca..
self-regulating or not, though modifica- oj
tions to the rules are in order if the 0
..._R L
process is self-regulating and the time V-
to reach equilibrium is short relative to I· • 1-- L, Apparent Dead Time
L. This is part of the "honing" that has I
taken place. More about this later. I
..
=ea. G> ...
::, I
From the step response. of Figure 2.3 . . .., ... I
c:1
the tuning settings are determined as oO
0
follows:
Time
-- 1 I RL (aggressive) Figure 2.3. Theopen-loop step response yields parameters
-- I I 2RL (conservative) R and L, from which controller setting may be determined.


:~~,~J~·:}f;~::--··· ~ ~~~!:~1~=~tiffijl\@ti.f~t~ii~~~~~rt~tlli~~r;f~iti;t=?~Jitii111ir11=ltittf~~~,m1.1ftimimt111111i1111111111m
;:;:;::::f:::::::::::::::::::::::::::;:::::::::::::::::::::::f:::~~:::::::::~:::~~:~;·;~.;:::::::::::::::::::;:f.~::~:::::::::::;:;:;:;:;:::;:~;::t::~:;;.::::::::::~~:::;::::::::~:
.·.·.·.·.·.·.·.·.·.·.···.·.·.·.·.· ·.·.·.·.·.·.··········"····I',................... . , · ·························.·..,. ,. _ ·.·.·.·.·.·.·.·.-, · :~;;:;:;. .. . .~;~~~::..~
~·· ···:~:~L~ . . . · ::::::::::::
.. .. ~
, . , . _._. -·:~:~~~:~:~:::..
.· .·.-......... ·=~·. .·:~=~;:;:::::;:;:::;:::;::::~: :·..:::::::::;:::;:
.. ···········. ·::~*. ffl·
m . ~.:,· · . . .··.·. -*-~
..e«·.· x?~~= . ;-.,. :~~;::··
. . . . "'" '~·-·. :~~~~~~~:~:::::
·:.-C~..••• •
....... x ·•·• ···.·~" • · • • • :::~i:J::::::::::::~~~~ ·:~:::::::::;;~:- :~· ·. •~:::.~~::::~:~~*::::::::::::::::::::::~::~:::
•••• ••••••••••••• ••••• ••••• .,. •••••• ••••• • • ••

[f,I!lI!; i!:!i1!1i!i1! !1![jl~!l[l!i! l[i!:~1Ii: 1:i : 1111111111111r:~~~!1/!!i!!!!!!!!!!i!~lf;w!t)J!~\!!i!!lli \j \i::i!:i:!:mi~fiil itl!!~!j• ·:·!·>.'•••'.


: .···:· :·:·· · :.;.;.;.;.;.;.;.;.;.;.; :·:·:.;.; ;...;.;·,<.; ·:,,·.; •• .;.;.;,;.,,,.;,;,,.,.;.;.;,,.;v;.•. , •,•,••,,,,:,:·:,:,,,:,,...• ·.••••,),,.•:,;.:,:,.,:,:,:·.,•. • • ·.·.·.·~• • •

·:::..:. ·. ', ::· :::~;.;:;:::;:::::;:•:::


....•.•...•.•.. ,.• ,,,::.-: .·,:-::::::::~::~;::.:~:.:t~*::::::::;::::::::·;::&
...•.••.• ;.., •• ;,t':,o:~... '·' '·' .,.. ·••.·.•:-~:::::::::·:,:::::::::::::~·
·. ·.· '•' ••,.. ·.,.· :··.·.·.-.·.·.·.·.·.·.·.·,·.·.·.·,·.
:-:-;:::: :~:.::;:::::;:::;~·:,;.- Z%
• ·%·.·
:!:1::1::1:i i1:1:1::111:i1:~11111,ii~~~}i• ~'1!!!,:)i
•••,·.·,·,•,•,•,•,•,•,•• ' • .. ' ~ • •

·. ·.· ...:,;,:,:.:,:,:,:,:,:,:,:,:,:
•..,,,,,,,,,,,,,,,,,,,,,,,,,,..
'• ,'I',
!i!i[![\i' i'•'•'~•'·'•'•
!i i!!i]!i)i!,,_.!!!ii!·~\i' !!)i;'ji~' !i' [•,:,••
' ,·,:~::·,::::,
• • '· .,..
[!Ii!i;:•~~1~i
'•t:;,:,;
•••• '·
~~J;:~!j)I]i~f 11:~11111:1r1111~11,!!i!i!i!i)~i!i!!iliii!i~i!i~~~1;~11111~IIt;i1~~\lll~i!~liill1,l!il\fi,*!r<l1,i~1!1f~~ll~[it~~k~..
~ · · · · • ' •,··'•'·······'.:,','•'1'•'•' ' ' ,,,,.,,,•,• ' ' ' ' ·, '• •' ' '

,,:'i.·:·•• :,~:::;.•
.,

·~ ' ' ' '

• ·)..:,;,:,:,:,:,:,:,,,::,:~·:
'•'•'•'•'•'•'•'•'• ,,,,,,,,,
· ,,.., .....X • ·' ' • • • • • ·.:,·. •••·•• · · • • • · • • ;.;-.~·· ..•:-:;;· · ' •;:,•·•.;,~,·· ·.. · • '•'•

,:,:>.,:,;
,,,,,,,,,,
''
,,
.:;,:,
.,.,,,,
~·· .}6 ..,
·'-!y •••••• ·~:«<>,.·%:
:-..,-· 1,.'"":.t ~
~:v.-- .~.
:.-.."t'i• ,,• :$:• ~ •,•"
·~~~~~,,~.1, ~1 '-

:d.:
•,:
!:l !:l! i!(i!i!i~~!i1:i!il! ~i:!1!1[!(!1/[i!i:!i :!i! :l! i1! ~\!1]1!1:!
1 1

~"'.'l:,j.:X.:.:·:~·:,:,:,~:·:·~:·:·:,:•:,:~,:::::::::~.;::::.;,::::;::,:,::::;.:·:.::,!.:
·.·.·,
11 i ttffirifl lililiI~!jl=.~li t1fil
X •'•.;o,.,;.••.;.•,•,:••'• .;,;, • •'•'•'\O•••··,,,,_.•'•·•'•'·'·'··•·····'··•_.:,_.··•'•·•······ ····•.'!·•• ••••••• •••••.·.·.:,·.·.·•••• ••• •• •••••••••••••••••••••••
·:·:·:-:-:·:·;-:-:·;-:·::;-:·:•:;: ·:i,:·:·:,:-:·-~:-:·
·.·.·.·.·.·.,.· .. ·.·.··• .., •. · ..,.· .. ·.·.·.·.·.·.·.·.·.·.·.,.·.,.·.· :-:·:·:·:·:·:··~:.:·N·:·:
.• ••••·•·•••••••·.•.• ·:;:.9,.,;e:··,,&
•••••• '.<···.-.-.·.···;s•

::·::;:: :· • : .~:::::::::::•:: .c-;,:,.·::. ::;::::::::::::::: ::;:::::::::: :::::::: ::::::,;::::::: :::::::::·::;:<::::;:::;:::;:::::::::::~~:i:·t :::~:;::::::::~:::::::::: ,,, . ,. .;~::::·::;::::::::::::: ::::::. .::: : :::. ::;:;:;;;~:: .. . . ..;:;::~:::::::. ;:~:::::::, :::::~: ·:::::::::::::::::::~:~:. ::::~:::: ::;. .::: ;!;.~':5; :;;:::;:::;:::::::;:::::!'.::;,. .,..; ,-·.<x ~· •."..: ;: ' ,_.:;ci.,S:' • :.,-,:::;;:. • •~::;. •-:=.::::.. :~;.:!:.::::::~: ::::·:;::::;; :;:: r::;~;:'i·~:::::;:::::;:: :::::::::::::::;:;:::::::;:::::::::::::::;::: ::::::::;:::::::::::::;:::;:::::::,:::~::;:::::::::::::;:::::::::::;:::;~:::;~:;::::~::,::-:,;:~::::i.~

·,·.·.·.. ·••••••....,.·.·.·.·.·.;.·.·.·.·.·.·••••.·•·.·.;.·.·,;,:,• ..,.,,..•• ...,,,,,.,,,,,,,,,,,,,,,,,,,..,,_. ••···•,:•,,,,,,,,.:,:··· ..;. :-.·:·:· ·.; •.;.;·:<.;.;.;.;.; .;.:.;~.;.; -:-: .;.:-:·:·:·:-:,-:-;-: .;.;.:·:.;,;,;.;,;,:·: .;.;,:,;,;,;,:,;,•,•,•,•,·,•,·,·,.• ,,,.... · ~ ..:,• •·.,:.:,:·:·:,:.;,:,:·:,:.;,•,:· ·.··:·:·:<.;.;.;.;.;. ·.;.;.;·:·;.;.:.;.;.;.;.;.;.;.;.;.;.;.;.; .;.;.;.;.;.;.;.;,;.;,;,;. :,;.;.;.;,;.;.;.·,:,:,;,,,:,:.;.;,:,;,;,:. ·,; .~; ,:,·.. ,: .-.,,.: •,. .·: . :·.·. • ·-~ .. ~-·••-:~·=·-~'!'·>: :-:·:·:·:·:·:·:·.·: •..;<·:-: -: -:-:-;.:.; -: s.;.;.;.;.;.; ...;.;.; .;,;.;. :.;.;·N•..:.;<~·> ·~.;,,,: .: :·:•:•":s .."i s .:.. ,,:, . · , ·· , . A . s .·,,:,:,:,:,;. ;,,...,:·:,;.:,•,:,•,:,•,. • •,•,: <•l-:.:·:.:,: .;,:,;.:.;,:,:.; ·:·:·:,:.;.; .;.;.:.;,:.;.:.;.; ·:<.;. ··:<.;·:<.;.;.;.;.;.;.;.;.;.;.;.; .;,;.;.;.·.;.;-:·:·

:·:·:< -. «. .;.;.;.;.;.;.;.;.;.;.·-: . ·.. :-:.;.·.;.;.;.;.;.:···:-:.;.;.;.;.;.;.:,:.;


:::~::::: .:;:; ;:;:<;>:;:;:::.:;:;:;:: ;.:;::::;:;:;:;:;:;:::::·.::·:.;:;:;:;:;:::;:;:;:;
-:·:,,~-:-:-;.:
:;:;.::i:t
.
:::;::::
,:~.
:~;f . ,,:,:,:,:,:,:,:,:,:,:,
.~(:~:::::::::::::::::~:::: ~~:.
:~:W.~·
.· ·:...'-,;r,.; .;.;.;,:,,•.·,:.;·:.;~.;.:.;
~~;:: •.;,;.;.· .~:;:;;::;:;·
..~:,,:~·-.
~:::::~~
$.:-:,:,
·.;:f,•;•
:1,::·:·
:::,~:: :::,,
:,:,:,:,:,:,,,,.
'•::::::::: •::::,:
,:,;,:,:·:·:·:·:·>>:·:
:::::::::::::::::::::;:;:.
-:
.:;.
·.·:r.·:·
.:;:::;:::; :;;:'.::
·-:~·'.·:
;::::~;
«:
:~••:;:
. ...; -~"
.·::.· • •
~·~~~ ~<l
··--:_i < .S.·.•{~1·'·1? :·~·:·
'•'•'' ,:;.~:;;:
.;;;;~-:~.:·:·
::::·:;:·:•.;,:i,::~:;;;,
:J,>:·:·.·:·
~~i,~::::;::
.. . xx.-:
::;:;:::;:;:;:;:;::·. :;:;.
.·:-~·,·
·~:~:;;;: • •
~= ...
:::r
1~·. . .·:;"~ ·: ·: ·:,.:·
~· ;... • ~:::(:: .,;:,,:•:;::.;:;:::::::;.
.J·~·-;. ···:·:···:·:· . ;:-:·:·:·:·: . :·:·:·:·:·:·:·:·.·:·:·:·.·: ·.···:·:·:·:·.·:·:·:·:·:·:•:•;
:,,, :,Y,;~::::;::.;:.:;:•i.':•:;::~:::;:;:::;:::;::;;:;:;:::;:;:;:;:;:;:::::;:;:;:;:;:;:::;:;'.;:::;

i~:=::::=:::::::::::::::::::::::::::::::::::.:;· ·:: ; 1:~:,,:~;<1- ~?.:::,:,,,,,,,:,: ·:,:,::;;:;,: :·..,.;;,,,,,,,:,:, .....,.. ;,m -·=·"==·.;=,:,:,:,:,:,:,:,:,:,,,,,;\l'~,,~:1,~,- .·. . . ' ' '*' .,., ,. .: : : : , , , , , , , , :, . ,. . . :,:,:,,,,,,,,:,,,,,,j,~'' ' [ll~'' '.,,:;-:,:r1~· ~. :(~: : : : : : : : ;.;.;. ·: : : >.<<·.·.:::::::::::::;: :::;:;;:=::::::;:: ::::::::;:;:;:,:::::"-.~: *::::::;:;:;:;:::::::::::::::::: :::::::::·~:::::~2:: ·:::::::·:::::·:·:· :·:·:::::::::: ::·:·:::::,:::::::::: :;.:::.· :;:;;;:::::::::::::::::;::::::::::::::.::·: ··....... ::···:::;:.::::
,.

22 Chapter 2, Tuning Rules and Procedures


. .

You. may ''eyeball'' the· 95%. re,,~p0,~~0iifJle -, : ' .- ·_·;


.: ·. ·Comments .. have · already beenmade about
. . .. ' . . . ..
· .•. · and then divideby J to es·~~M·~~.~
. . . :··;>/
:

. ·the use ·of the filter .. ·They have also been sponse time, T 63• If your:::syi;tem/is·· r"-~PO.ild.. · : .· : .
made about whether the controller is P, PI,.··· ing according to typically .assumed 'Idealre- -: ; ., '.
PD orPID. sponses this relationship of T95 to T63 is .es-
sentially exact. Since most-processes arenot
There is a type of response that deserves ·. that cooperative, · 1 often use the 900/o : re~
special mention. This is when the lag fol- ... · sponse time. If you are judging it by eye it is
lowing the apparent dead time is short rela- . hard to tell the difference between· 90% and
tive to the apparent dead time, which: im- 95% . · .. . : ·
plies also that the process isself-regulating,
.
Figure 2.4 is provided to help discuss this This· special situation that I am discussingis
situation. · · : · · · , · . : · · . . ·.
worth 'separating from the rest of the-field
· ': onty··if T63 IL is 3 or Iess (perhaps. even 2 ..
1
. .

"CJ.
--
. CD CD
-.a
e...
. .. ·. '... . .. .

c -
ca . . .
... ·
.
·.. . ·-o •.9A
": .
.
. A ·. .
. .• . . . ·.· .- . . . . . . . . -.·. . . . . ·...
o~
.. . . , · ._ , '. . . . : . . .
Kc::: 1 I 3RL -. . .·
(J . .

:.·. ; ;:·..··.. ·.· · , .. · •. . . . . '. ... . T •:


. 1 ·. = T
..... 6·3. . .. . .'.

• L
. 'lhese modifications to the rules have you
setting' . the ·. gain · lower than you: would
..Cl) .... .
with the closed Ioop rule's and the: inte-
=::,
e a. _.,_____ · · · -: .•. : ~,-r-
B · . ·
. .
. gral ::time. . .. .
'much shorter than with :"~ither
................ the :,:,:closea:· ·. ·_ or open-loop .rules. ·•- Thfs
. . .
..• .

i~
(J
T ·
means, of course, that if you choose. to
Time use .: these ·. ltlodifications,. y~1l ·. lllUsl · Use
Figure 2.4. The step response of a self~regulating ;p~ . them ; tog~t~er, . you cannot Use ;th~ 'te
ess can yield parameters helpful to guide tun'ing' ·;1 <the set just the. gain or just the integ;.a.l time.
response time is short relative to the apparent dead time; . · Esserl.tiall}' youare going·toward inte~alli
.. , . · · ·. only control as r· . I L becomes~~--sfflall an
. ·. . . .· . . . · . . . . , . . 6~i _ . '., . . _ . . _ . ·" . . ~- .
A few more, parameters are defined. The- option· that . was ; 'rarely exercised before >

step size is B and the final change in the . digital control': The closed-loop 'rules witl
process is A. A parameter called. the 95% , · .·.· . still .give.you ·stability and reasonably· good
response time is as defined in the figure. performance, but they will give you differ-
This response time does not have to · be
' . .
· ent settings. .. .
· .
·
defined very precisely, as will become '· . . a ••

apparent as you learn what is done with it. On rare occasions the. slope,· R, will con-
Ideally· what we .·want, is the 63% response · · tinue to increase . •. 'ffiis_ '.:sitilatiott:·:·is often
time, but that may be hard to determine at called open-loop unstable or conditionally
the time, when you don't know even what A . . stable. or runaway. On other occasions the
is· .going to be .. If· yo~\.j,l:~~~g:.;i~:/ · . • ,:'\ ''~po~~,}i!~·soJn·.the otherdirection first,
response, then it. is possible to detert.ri!in.e, :!be:·' j . • ; ,;i · . called :·, inverse ,: response. . The open-loop '.
63% response time from the reC()rditi.8, ijj . ; rules for tuningdo not apply in either of
still it doesn't have . to be detenDined. ,;:·;, : these cases, and indeed, the closed-loop ·.
'precisely. ·. ·. •· >..· i .··· . ·. · . · . i\ procedure is-subject to pitfalls then too. ·
. --

. . ··,·. .
. .,. . . :.
. .. . . .·.··. . :
. .
·;. . . .
...
' . . .. . .
. ···, . ·. ... .

.. . .
.; ,.
-v.. ''. . . . .
.
. ·. .. .' .... . . . . . .
•.
••• . .•• .,.. • . .• . . •••• ••••• • '·,:: . ·'·••• ••• : >.

. . . . . . . . . ·.· .
' . '·. ·.. . . . ·,:···,:.: i:..::f .:: ,:·~.'. -::'. ·. " . . . . :... ,, ... ·. . .. : ... ·. . . . .
. . .
: . ': . . . . ·' . . . ·.. • ... ·.'.. ·. . . .. '•
••

Chapter 2, Tuning Rules and Procedures 23


.. .
.. · . .
.
·. : .,....:::. :... ·,.,: .....::\-~.: ~ . ·: . . . . ·.. : : ·.: .

. . . . . ' , . .. . ... ,,-, ..·....


. ·.. : . ..·<>::·:: ::$:: :·~: .... ,.... ,::· -:: ··:·;·:,::.. ·t·-,:J....-: ,·,\.f:~t~t·· ..
· . ::. .,. · · . . -· .~ ·;: :·.;._:i;}'L ·.. . . ~'. · · -~:·,:: ·-.-~:~·-~-·i:.<~::·::.=i£:..:~]:~~1-~·;\'.-:·

The. open-loop approach. to ~u~~g'.:~~r,quues You Wi.11 have decided. beforehand~;hat size
more care than the. closed-loop approach, step to.introduce first, based onsafety . ~-d
but yields ..more .· fundamental ·<information. quality concems.You may.have.chosen a
More care is requiredbecause thetuner may . . .
value .considered 'very '. conservative.' With
not feel comfortable controlling the process digital and. other modem controllers there is
in manual, or may feel nervous about what likely ··to be no 'problem. in making ·exactly
size· step · to .. inject, . The. basic .idea pf an the change desired. With older equipment it
open-loop test is to learn certain TIME and is often difficult to ·. make ·small. changes
AMOUNT characteristics of the process by precisely.Doyourbest. The . important thing ·
putting in a step change in controller output. is to not continue fussing with the controller
. . . output ifthedesiredchangeis not achieved
For .the open-loop test, to be useful, the exactly.If you continue.to adj:ust the output,·
process must be 11:lllPing· fairly smoothly. If trying· to get. exactly the desired. change,·
it isn't, then it may be too hard to differenti- then the upset is riot a step, but tis something
ate between what the step input caused, and else. If all this happens fast relative to the
what might have been •. going to happen· any- apparent dead ·time, then the "sin" is not· so
way .. So achieve reasonable stability of the serious. If it happens too slowly, relative to
process .· before putting the 'controller . on the· dead time, then the results may not be
manual ..··Most modern 'controllers · have a validly interpreted as described.
bumpless procedure for transferring from
automatic · to 'manual.: If. the controller you To perform an open-loop test you must be
are working with does not, do the best you confident you can control the process in
· can. It is very . important •that any · bump to manual. It is very desirable to have some
the controller output at the time of transfer- idea of what R and L , and especially L, will
ring to manual not· be superimposed on. the be. Information given in Chaptervl will help
planned .step change in controller output. If you estimate L~ .
there is a bump from the transferring proce- ... .'

.
dure, or from your efforts to stabilize the . ·.
-
,.
. . .

.
..
process in .· manual, let · the effects of it .. ."
.
'
. . .
.
. ... · .· ~ '
. .
. . . ·..· .. ::... ·· ·. '. ·.:
. ~
.. .
.
. . ...... ·
. ·.
.
settle out before trying a step test. . . . .
.: ·.. : ·:·. . . :: . ~ '.:_·· .' i" ...: ·.. . ... . \ . . . .

_ ...............
,
....
. . : ; . ·....
.. . . ; ... . .

Before you. make the . step. change, decide · . . . ~ . . . . .·

how. you 'are going to: observe· the results. . . .•.


~· . . ..

The · results might -be · simply. observed;


mentally · noting .: .the - · apparent:' 'dead
. t ime.. ·. ·TI~
They might berecorded on a test-recorder.- : .. ·. . . . .
~fyou hav~ a DCS with a CRT di~play, set_.: Figure as, A useful technique ,for impres$ing -a step
it up :to- display the process ~ariable and change in the controller output. is to· impose a balanced
the controller . output. .Occasionally i:.;;you disturbance. as shown above. ·. · · •. . ·. . . · ·
have to have semeone call out and record . . · . . :. .
theresults manually, When using a DCS, Make the step change. When the results of
youwill have to select the. scales. This is the change are seen, reverse the step change
somewhat .· of . a guess-and-test procedure, direction and do it again, returning the out-
and you will get better at it the more you do. put to its starting value. If the process does
It is desirable to see the step response well, not return to the same.value within the Iim- . .

without it exceeding the scales· set· for the its · expected, .bring the · process . back to
CRT, and without it being too sm~ll. to see where you want It manually .. This is likely
either. A time span about IO times the to. happen if the process is an integrating
apparent dead time is a good choice. type, or if you could not wait long enough to
reach equilibrium. If this is the case then
... .

24 Chapter 2, Tuning Rules and Procedures

you will likely want to not return the output. which could present problems. A :8~·;,:way. -. , . .
to its original value, but rather double its to check for dead band is to .tna.ke.l/fwo: small .· . ·• . ·. . ·
size in the opposite direction, such as shown
. . ... ~· .,.,,. r ... ··.'· '.·•·--: ...
: . . . . . .·· ,

changes in the same directi6il/}nt1,


, en. re ~- • . .•
. .. ....
·th····.·.·.
. .

in Fig. 2.5. This format gives a balanced verse direction with the same two ·. small .
..

;
..
.
disturbance to the process. The intent is to changes. If the process does not repeat itself
end up where you started. This approach is for the same outputs, this is a good indica-
also useful when you don't know what size tion that significant dead band is present.
step to start with. You can start the step size Problems with dead band are discussed in
and duration at very conservative values and Chapter 8 .. ·.
then increase both until you start to get · the <

results desired. Once •the open-loop testing is done, calcu-



.;, .
. . .
.
. -~·. \ '
. late· and install the settings and return the
controller to . automatic. At this time. you
should · .; consider · · making · small setpoint
changes 'and observing the response; much
as if you were performing a closed-loop
test, basicallytc confirmthatthe resultsare
close to what you :Would expect. If the re·
. . .. .•
. . . .· .
. . . . .. sponse is judged' too. sluggish, increase the
. . .
. .
. : :. . . . gain. Tty a factol of 1.5 · or two. If it is
It is very desirable. to perform the test. more judged . too oscillatory;' decrease the gain.
than once,. especially. if there are a lot ·of Again try a factor of 1.5 or two. If it Is oscil-
other variations going on, which casts some latory and the period is much 'longer . than
•:
. .. doubt on what you caused and what would ·the natural period, . say· more than···.30%
have happened anyway. Sometimes these Ionge~,.: increase the integral time, possibly
other variations .can be identified and by 50%. Thesesuggested amounts are to get
stopped, by putting. ·another controller in you started. The· point. isto not make small
manual. Other times the step size. must be changes, like 10% or 20%. You are looking
increased so ·the dominant: effect observed is for a response . which is ·essentially.complete
due to the step change. If you .are observing in three to fivedead times, not 20 to 50.
the results on a monitor and· have the ability
to make a hard copy, do it. This will· facili- YOU should plan. to monitor the performance
tate extracting the value for R. From these _ . . over a period of time to confirm. that. there
collective results, estimate the best value for -,· are no problems. .
the apparent dead time, L, and for· the rate,
R. If these vary a lot, be conservative, use If the response is radically different from
the longest L and the largest R to calculate what I. have. shown, then it· is likely there is
the controller settings. something abnormal about the loop, or that
the loop is being constantly· hit by distur-
If the response is self-regulating, and if the bances. If making these tuning changes does
response time is short relative to the appar- not produce the expected results, and . no
ent dead time, then use the modified rules disturbance that could produce the behavior
given earlier, essentially using a shorter in- is identified, then the· judgment of .a more
tegral time than you might when using the experienced person should be sought .. This
general rules, and a somewhat smaller gain. of course 'assumes that the problem justifies
. .. '· . ·.;

.
.
. a deeper scrutiny and that other efforts to re-
If circumstances permit, it is a good idea to solve the observed behavior come up short.
make step changes of different .sizes as'well . .

as in . different directions. If the response Observe the responsibility chores after· tun-
curves are not in proportion to the step size, ing, related to staying available and letting
this is a good indicator of nonlinearities, people know what you have done.
r~-
'

'i
'

'··
'

Chapter 2, Tuning Rules and Procedares 25

.. ~. . .. . . . . ·.
. . ·, . ', ..
:. ' . . : . ,·· .

The open-loop and closed-loop methods of The open-loop approach to tuning might be
testing· for process dynamics· wil:f; not··· give preferred as · a precursor· to the closed-loop
exactly the same · controller · settings, but approach. when L isvlong, .Iike -minutes
they should not be very different. The open- rather than seconds. This could· be the case
loop test is probably done more often when for· composition· and ··temperature loops, as
a monitor is available to display the com with pressure, level and flow
controller output and the controlled variable loops. Indeed for pressure, level and flow
systems, the L .and R are likely to be so fast
. . " . . . .. . . ... ·. . . .. . . . . .

on an expanded - scale; than. when· it is not.


This · ability to examine the __ - open-loop . . . . . .
that typically av-atlable recording equipment . . . .
response . "under a microscope" has . greatly might _ not be able to establish them
. . . .

facilitated the use of the open-loop approach accurately, whichcould leave you using ·the
to tuning. The open-loop test is also very closed-loop approach to tuning.
useful for troubleshooting, as there is usu-
ally some idea ··what the step ·re~ponse
should look like, and· if it· doesn't, then there
are clues as to what the problem might be.
. .
. . .

The contribution of John Ziegler to the


open-loop tuning method is especially
noteworthy. Realize that in the late ·1930's
there was very little known in the process
. .

The open-loop response can give more use- _ industries about control principles, _ th~ math
..

ful information if the process is self-regulat- of control, if you will. -Very little. In the.
. .

ing. Assume the open-loop step response is


.
process
. . industries.
there were .
perhaps

four
something like Figure 2.4, for which the fi-
. ' .
·to six "experts", and· they .could not agree
nal steady state change is A for a step among themselveshow to tune a controller.
change of size B. The process gain, K , is · Ziegler was not considered one of them.
then said to be A I B. This ratio will ave About 1940 Taylor Instrument Co. (now
units of% -process I% controller output.. ·-1f .-ABB-Kent-Taylor) introduced the first PID
the controller .gain, · Kc, is then set to less
. ·. . . ..
controller all in one· case, and the company
than I I Kp (l~ss than BI A), the loop has had to help its salesmen and customers learn
to be stable. Understanding this can be very how to tune it. Ziegler took on the task.
helpful at times, especially for - those Fromtalking with virtually everyone in the
situations for which it may · be difficult to field worth. talking. with .he learned that, in
establish a slope, R, for the open-loop his words, "lags were bad and capacitance
' response curve. This can happen when the was good.'' His use of · the term
. . .

''
l
~ response is too fast to see on the equipment ''capacitance'' will need more interpretation
. . .
l

! readily available. Sometimes A I B can be


~

than will be presented right now, but a fuller


calculated, in which case a value can be understanding of it should evolve · as you
placed on gain that would be assured to read further. By "lags" he meant those· fac-
·.

'
'
' give stability. tors that contribute to L, and · by
"capacitance" he meant those factors that
contributed to R .
.,
..

.,.

'
'
"::
. ·. . . .

26 Chapter 2, Tuning Rules and Procedures


. . . . • • :: • • ,. • > • • • ••
. .

From that understanding he :conceived the .reverse. Ifwehave thusand so.tunijig;:alld ·. ·.·. •
graphical open-loop method for determining R and/or L changes, what would .liappen. to · ·. · ·
controller settings. Many complex situations stability? If you are a novioe<·.on?control at ·
can be much better understood if time is this stage in your reading, I simply ask .you
·,.
taken to ask what would. the open-loop to keep this open-loop · concept in .: mind.
step response look like? What. would happen You will find it very useful later, especially
to R and L, and therefore, what would that . . . in understanding the effects . of interactions
do to the desired settings? Would that make and non-linearities. These are the· principal
the system more stable or less?·It is .rather conditions that cause a control loop to
like using the open-loop tuning . rules in misbehave.

. .. ' . . . . .
. . :,. . ·.. . : .:·: :· .. _:.: . . . . ,·.. ":. : :~.:·.·:~ .,-,: ;.
. . . , ,·. .. . . .
.. ~ . ., .
..
. .. . .
. : '.~ . .. . .. .. .,
.
' .. .
.. ':'·
: .. . .
, . . . - . .

There are . two . common times when going


. . . . .· . . · storage capacity the vessel to absorb of
through the µm_ing': procedure by-the-num- ·
.. . . . . upsets.··. This . is e~ed averaging level con- . .

bers is either not desirable or not necessary.· trol. This. is illustrated . in Figure 2~6. The
. These are levelloops and flowloops, ·•· ·
. . flow in can be quite "wild", but the capacity
. .. . ., .
. . .. . . .. . . . . . . . ... . .

of the tank is used to . smooth


. ... · .• out the fluct1lations·~6they are
. not passed on to the next step
of the process. .·.· ·. . ...

·time·. . . . ..' . ... .

. ·: . .
Formulas will · be . given. for ·
..............____.~-......_. . . determining the settings. First
..: . . .

. . determine how much the 'level . . .


. . .. . .

··•-time
•I · may vary. If it is only a surge •

1 · tank, . you -; might . use .. 80%. ·,·

. Other cases wil! . vary. The . . .

level in. : a distillation · column .. . .· . .

.
· · will affect heat transfer in the .
.
. ' ...

Figure 2~6. Averaging level control can smooth out flow fluctuatio~s. · · reboiler if it is too low, and
keeping them. from being passed on from one part of the process to may . entrain liquid up · the
another. . .

column if too high. . :' .

First determine the following parameters: . ~ .


": ·.

. . .

Level is different from most loops because


UL% :::: Allowed upper limit, % of scale
it rarely is a factor in quality control; as . .. : . . . .

flows, temperatures, pressures and . compo-


LL%= Allowed lower limit,% of scale. ·
sitions are. Indeed, it is almost the opposite, · ..

.
if you can conceive of an opposite. Often it .

SP%= Setpoint,.% of scale.


is desirable not to have tight level control,
because this passes on upsets in onepart of
V = Volume between UL% and LL%, ft3
a process to another part. Rather . it is desir-
able to have loose level control, using the
Q = Maximum flow rate, ft3 I min
..

%., .
. . . Chapter 2, Tuning Rules and Procedures . .

..
.
..
.:

. .

Then calculate the controller settings: Ori rare-occasions level. falls ;in ·the qual.ity~
.... -;: '1 . ·.". ·· , . ... .. . . . . . . ·: • . . . • . allecting; category'. and .should be controlled
·

Kc· 1· = I 00% I (UL.r«,:{,~;; SP%) .r> ~ : · :.,,\-. ·. . ; . tig· .htlv.-


',,} ;O.· :ne... .·· ex. amp le .of this is , a· ·po. lymeri-
. ·;~~- . .
·. ·.::. ·...~ t~
:' .. ·
: >.'..rf'.. ·. . : : ... ; ...· ·. :._ . ..
· -• . zation vessel in- which "the- shorelinebuilds
~· . .

Ka = 100% / (SPolc> - LL%). • · • .: · . · , , , ,> -: : ' iii' · •· · • . up •· a deposit ,of> degraded . polymer . that
·'. . . .
.
. I ...
;· . .
·..
.
.
. . ~ . · ·· : · should aet.be.disturbed.Jestit break off and .
Kc= Smaller ofKc1.and Kc2 contaminate the product Y ou.may think -of
:· -: .. '.· .. :. -. . :·
...
', · .' other examples. So again, you are back to
the tuning rules .. · · ·. · . . _

These .. rules . are;.:..J<.J..9qµa!e. for ·-.;ih~-; vast . .. : ;'· . '


. . ..
..,

.'
mi.J~rity of level ••~ Ifie . • needto .use .. .
.:, . :·: ..
... ·. ' . . .
. . .. ~ '.

i
formal tuning procedures for level is tare. If · By now you understand that the tuning of a r • •• ••••••••• • .•

th~. :.$.pan.···.for maximuni .•t~·.tpi~:i!P~lll .~OWS loop depends on .. its . tittle :. or dynamic .... ·....
does: not require the full travel of the valve, response. Since flow; loops tend· to be
then a lower gain may be used. This . will designed the s~~;~a.).{._:~~y. ~F~~::~P.~:Jt~v~;,~e~
allow in ore· of the volume ·to· be · used for
same. dynamic ';!~·" ; e ···~··. · #Iere(~. the
surge averaging. If thisis done, recalculate same tuning se,ttif\gs~. . . ¢-~,typteaI procedure
the integral time. · .. : . .. ·.. . • ..·•· . ·.·.·. ·. ·· for tuning . a., . loop may: usually be. bypassed
. . •. . . .
. . .. . (., .

for . tiqw )OOps. ·. A . gain. qr. Q,1 . to 0.7 .•. suits .


. . . . . . . . . ·. · ..
· .. ;-• • .
• ..
• ··: .. • . . •• < •• •• • ·.• • , •••••• :.: > .: .... · • . . . . ·. : ,. .:.
.· : . . . . · ..

The above. recommendations apply If . the most ·fl9w · loops, Th~ir- natura] periods
val~e and level ar~f closely coupled. That is,. dCpen<l on f~io111 not. yet discussed but will
if the apparent de8.d time between · valve fall in the.one to 10 second range..and most
movement and level starting to respond is : . fall in the ~o to five second category, Since
. . . . . . . .. . . . . : . . . .

short (say. one. to IQ .secondsj.Tf it is Jong ; · fl9W··.:.Joo.ps •.. ar~:·· fast. relative. to· . most.
(and there is no hard and fast rule about the. . accompanyingprocesses they .don't, have to
one to 10 seconds) thenthe considerations be tuned to within an .inch of their lives, so.
are more complex and . you are .advised ..to an integral time <>.f 0.1 · to .. 0.3 · .minutes is
apply the open-loop concepts to the task of ; · usually, :a.d~q~a~:.~· Thes~-. settings will get
t tuning .... · ·. you in the ballpark, : : ·; . . ; . . . .
. . . .
. ·' .. .
' .
.. : . . . . .. . . . . .

A case can be made for notusingintegralat · · If y~u need to be Jnore precise than the
all for. level control. This_:is. a more. viable above generalities yield, then continue with
concept now, with digital control, than· .it the closed-loop approach to tuning, trying
' was for; the days . of predominantly analog'. . small setpoint changes to test for stability.
·.
..
?. control. For standardization on aplant, very. ~ The open . . loop, response Is usually too ,fast .
few proportional only controllers were ' to ~APWI'~ a9~urately on typical . co~t,()J . i
specified; with the result that PI controllers : roolll' :llionitoring equi,pment, so ~~the. open- .• :
were ·. used . for level .. At•· issue _is, something
.· . . : . . . . . . .· loop ···approach:,. is··usually not•·. u·sablei'··ae . . ·•
that is hard to .explain without getting more • aware · though.·· of one factor when; tuning ·
scientific. . An integrating process, as level ·• flow loops. T,hat is tha~ ~he· ope.11~,~9.p gain ..
usually is; coupled with .a controller .with · . is likely to be .higher at high flows than at. · ·
integral action, tends to. look like a runaway low flows. The open-loop gain is how much
system. . '{his . presents . \ the . potential for the ('low moves (in percent ()f scale), . . ill
. .

stability problems that I will try· to, explain


:

response. to a· change in t~e .con~roller output


later. A cycle may develop- whose period is (in p~rcent of seal~)~ If this is higher at higµ
long relative to the natural period. -: . flows,. as it often is, then tµ.n~g done a.t low
flows may become unstable at high flows.

'
;

'
28 . Chapter 2, Tuning Rules and Procedures
Occasionally a flow will be controlled in- Level and flow are examples of classes of
directly. That is; the valve .and the flow loops that only rarely justify being tuned to
meter are not in the same line. In these cases be tight. There will be many other indi-
·,•
the typical flow-loop dynamics do not exist vidual loops which need only to be stable
so the typical tuning settings do not exist and perform reasonably well. They do not
either. Tune these loops as you would any need to run the four-minute mile. But sup-
other, choosing between the open and posing you do want to, how fast can · you
closed- loop methods . · ·. based · on reasonably be expected to run the mile?
considerations discussed elsewhere.

Pause a moment to think about it.Processes could be worse than if a less customized ap- .
come in a virtually limitless variety, while proach had been used.
controllers come with only two adjustments
to fit them . Oh yes, in truth there are three,
~roportional, Integral and Derivative (PID),
but integral and derivative are determined
by the· same process parameter, so having
."
·.;.
set one, the setting for the other is also
determined. Broadly conceived, the task of \

tuning is concerned with fitting the time and The reset time, the derivative time and the
amount parameters of the controller to the filter time · are all keyed into the . time pa-
. .

time and amount. parameters of the process. rameter of the process. Unfortunately the
This is similar to fitting clothing (shoes are controller gain is not uniquely tied to an
. .

a good example) to a human being (also amount parameter of the process; it is tied
coming in a virtually limitless variety), by to both time and amount parameters of the
specifying only two parameters, like width process. Actually it is tied to a time-depend-
and length. In both cases it turns out not all ent amount, but that is getting· too compli-
that badly. cated at this stage. Just · remember that the
. . .

time· settings · of a controller are tied to a


time parameter in the process, so the time
! .
'
settings all have a relationship to each other.

The reason a· controller, having essentially


only two adjustments, can fit a process that


might be temperature, pressure, flow, com-
position or an almost limitless variety of
controlled variables (when given a name
such as these), is that the controller doesn't
With computers it is possible, of course, to care what physically is being controlled. It
build a controller to "custom fit" the proc- cares only about the time and amount fea-
ess, much the same as a tailor would custom tures of the process. It doesn't care about
fit clothes to a particular person. Let the units at all. It is true though, that processes
process, or person, then change. The result by name often have time and amount
.
. .



. . . .
.. ·
. . . . . . . .....~·. . ·. : . ' .

Chapter 2, Tuning Rules and Proeedures ,·. ~


·:· .
.. · . .
.

,
. ..

.. .. .

. .
. . ~ . . . .

reasonably be ~xpeeted • to time . and . . . . ..· ··.. . •. .·. actions with other loops. have
amount charaotlristics much like another . ·'. .: : ., .· . · · · .· • · • .
flow loop. · . · ·. . . , ''"''ii\• ". ,f;;:,' . · Iii The loop should be a typical loop in
· < ,,,,,.,. " ·. .·. terms of its · lags. It should not have an . . . . .
. .

Part of an overview on these tuning rules is . · -. inverse .response ,Pr be. :open-loop un-
to realize that notonly are .: th~y designed to stable. •.· :··.~· . .·. . ...
give tight control, ;but that they :lite .predi-
. .
. .

cated on a few assumptions. Remember that the rules· are· approximate~-


. .
They will get. you nicely in :tho; ballpark for·
&:1 There·• sh~ld Im ~;iMfi1es&rile noise. tight control. The. calibrattOJl.,f~ll · . 'e . knobs
. Noise can:1,c,.~(j~jlt of ils bwdt$irfli>Ie for analog controllers .· is · :freqlently poor: ·
· variations .· in the· measurement; either . Not only is the calibration at the dial mark- ..
not · . meaningful ·. · · variations · and/or ing inaccurate, but frequently; . there is a
. . variations . too fast for. the: controller to large gap •tw~~;.tn~rlcings., Sg if,yput.g9al
· . do inyth~gabQllt. > · · i . . . · •· . ::::~t11e .
· is tight::,;~ontrlli·~.;",. - .. rules, :tit\~ou\J·-udge
..;· ·.·· .. ·. . . ·· , , , . . . . ,. ,,. .. ,·.·

.... ,
.

.··:
.
~

·,·,.
.

·.
.

..
.

.
.
. .
.·,:.
.

· ...
.
.
.
.
..,. .

,:·
.
.

...
.

· . the··~~rotindiiY.,abf<,i1, te expeCtaU6DS, and


lit There . should·. · be . no· troublesome non- .
. : . .
· if you: knQ.w ihat yo.u are 'deing, you have a
linearities, no . dead· band, no velocity licenseto' disbeliev~\'fhekttiatie.lJibrtttions. ·.•.
: ..•..: ' ...· . .• .· . .. ' ~• ; • . . . , ; . :-:: ( :,i . . . ,': ',; - <:< : . • . . . ·. . .
·. limiting.a.ndotherstobe·discussed later. . .: -:· ,... . .... . . . .
.

. .. . :
. .
. ·. . : ·.
.
,
.
.
.
-:
.
·.
.
.
~-
. . ·. .
. : .. :: ' . .. ..
·:
·. . . ·· ... }: -~· .·: '· ... ·.:. ', . .. . :··
. .. . ...
.. ·::. ~
. ... . .. . .. . . . ,... ~· . .
.· . . ·.. .
. ' :
.
. .
. . ·. ·: . .. . . .. : . . .: .. .. . . . .. ; . .

.,,, .
.. . ·. ·. . . . . •.. •. .
.. .·. r .
: .:.;···.\ .. · · ... ,• .. . . . , ;· . .. . ....
. . . . ~ ·. .
:·~·
..;:- • • h ••

"
. (
.
. : . .. ...
. .
..
~.
. . .
. .
.
. . . . .. . :
, . . . ,.

. . .. . ·. . . .

~. . ~1 . .
·." .. · . .": ... :. ,·:·;.·::.. ::· . ~i ... ·.:· ·. : .. ',',
. . .

.. . .. ). :.:. r:: :..: .


. -.. . . • • • •••• • <
... : :;.. ·: .
::
. . . ..
~:;·. .:..:::~.. :.~.·. ·. : ..
. . ..

. • . ., .. ...,.. . .. ,
• :

. . ,. .:. . . .
.. . . .
.. .
. :/.. ·. }:··.. . ,..:.. :. :.'...: :• .
. .
.
' •· . . ·.. . . . . ·.. . .1 ...... ,: .. ·.. . . ... . .
l '

. . .. . .
.
:
.
; .
..
:. .
.

t:, ..... . . :: . :;. . . .. . . . . ... .,,.


•<
~·. . .... · ..
·. . . . ·, .· .:·. -. . ·. ,.:.... ', .. .' '. ·.:.: : ·: . .. ... : . ·.:: ,. . . .
·~ . .
.·'
;.·. ..

,· .
·. . . ·. . . . ·. . .: . :. . . ··: . ·.· . . . .. .
.· . . . . . . . ... , . . . ': :·:. ·. . . :. . . . . . . · .~.•.; \ .... :' <· ": <. . .. ., .
' , . . ·: . ·.: •.: . .. . . .
-·...
• . .·. . . ..
..·'·. .
.:
.
. .. . . . ·\· :. ·.
:. .
.
>

·.·.
. . . .· .. .)

>
. .
'
•.·., ... . .
.. . .. ; . . .. . . ·.;
;'
. ~:::\.··:\ . ·.,\. '. :. ~:·:. . .> . . . ... 1 . . . . ...
- . . .. ·. :. . . : ·. . ·. ' . :~. . ..·
. ..
..
:·•
. . . . . . :·
.
·: ..
. . ..
.. . . . ·,· ·.·

. . . . .. . .. . . .
.
. ' .

.. .

. ·. . . .
. .. .
·, . .

., .
·;.·. . . . ..,. .. . ··f... ,. ... . .
. ..

. ·. . .

.
. .... ,. ., . . .

,. .
. . ......
, . ..

'

. ....,
<

31

Tuning rules are designed to give ''tight'' models for the process, and simple, well
control. This means they are set to give as defined disturbances. They present results
fast a response to setpoirtt changes or upsets defined to three significant figures. Real
. .

as can be had without excessive cycling. Its processes and real disturbances are not often
like teaching everyone how to run the four-
. . .
mathematically simple, and for disturbances

minute mile. That may be desirable at times, Ill particular, are often not mathematically
but frequently it is not necessary. Most of defined at all .• Despite this, these studies are
the time it is not necessary. Sometimes tight very helpful for understanding generalities.

control IS not even desirable, as has just So general relationships will be presented.
been stated with level
controls. Imagine a controller
on manual and that an
With the above upset occurs which
qualifiers I will now causes the controlled
concentrate on what variable to respond as

performance you can shown 10 Figure 3.1 a

reasonably expect to The figure also shows


get from a loop that what can be expected
• • •
IS tightly tuned. Performance is judged by after the same upset if the controller IS 10
how closely the controlled variable is held automatic. Again
.
the numerical
.
values for
to the setpoint, both for setpoint changes the settings are to show trends only. Now
and for disturbances. People studying the comes a very important point. If tightly
problem mathematically have used a variety tuned, the controlled variable will deviate
of criteria. These methods take on scientific about as far as it would have without control

language, like integral of the error (IE), for a time a tad longer than P n/2, which IS
integral of the absolute error (JAE), integral about equal to two apparent ( or real) dead

of the error squared (ISE), integral of the times. An oversimplified reason for this lS
absolute error multiplied by time (ITAE), that it takes one dead time for the controller
peak error and surely others. These studies to know about the upset and then another
all assume certain mathematically simple dead time for it to do anything about it.
• ' ..

32 Chapter 3, Tuning Objectives and


Expected Loop Performance

30% variations ,.
from
K0=0 expectations. Even hitting it
(Manual)
within a factor of two is
K0=0.25 immensely better than many do
1
--CD
-~
ec ca..- without these concepts. For a
.... .. ·. . , K·c -o·s
. -
..• . . ... single loop, shorter periods are
oi
o
.. . -~

always better, if minimized


. deviation from the setpoint is the
Kc=2.5 Kc=4 goal.
~ ........... ~--- ··.··~---
. If integral action is added to the •...
' . .. ..

system shown in Figure 3.1, then


~Setpolnt

the results shown in Figure 3.2 · . . ..

0 20 40 60 80 100 120 140 160 obtain. The peak error occurs


. · · · · · Time. · ..· . . slightly later, and the use · of
. "··
. .
:·. . . . . . .

Figure 3.1. For an upset, increasing gain in a pi'oportional-only integral· .action increases the
controUer reduces the error, and the tendency to cycle is · tendency to · cycle, and ·at a
increased. The maximum deviation for a tightly tuned controller slightly longer period.
occurs at a time slightly longer than Pn / 2. ·
. A second and very important
For the stated ·case, it then becomes observation must be made on the above
apparent that · the natural period may· be example. If tightly tuned, the controlled ·'

considered · a way · of rating , -the loop's variable will return to the setpoint in a
performance. · A loop . with · a P n of one time a.bout equal to 2P0• Whether the·
minute will typically be twice· as · good as multiplier is 1.5 or 3 or something else close
one having a· P0 of two minutes, all else is not. as important as the concept that • . the
being equal. This is somewhat . of a time to recover is proportional to P 0. This
simplification, but is '. the · essence of .provides another reason for shortening the
understanding perf ormance. As the . basic . . natural period. This is one of those happy
tuning rules were · stated as approximations, · circumstances in which one good is actually
so is this. In industrial control· ·we rarely· · · two goods! The error size and duration are
concern ourselves with 10%, ·20% or even. both reduced. ·.

1. T1. = ''Off'' •
2. T1 = 100
3. T1 50=
4. T1 = 25
All curves with K0 = 2.5 5. =
T1 13.

)J!
-..o
1
e,ca
.. -
c ..
o.!!
fJ ........ ....-:;._- - -· -4· -:
5

0 20 40 60 80 100 120 140 160 180 200


Time
. . .

Figure 3.2. Integral action restores the controlled variable to the setpoint.
Reducing the integral time increases the tendency to cycle, and at a longer
period than for proportional-only control.
~~~~~~~~~~~~~~~~~--------------------------------------------------------------------- .........
'

Chapter 3, Tuning Objectives and 33


Expected Loop Performance

I had hoped to get your attention earlier by scale, be able to reach the desired gain for
emphasizing the importance of the natural stability.
period. What you may not hay~ noti~. ~~--
the absence of any emphasis on the specific
ultimate gain it took to cause the loop to
cycle. This was deliberate. Whether the loop
cycled with a gain of 0.1 or 10 tells you
very little about the expected performance.

What the gain turns out to be numerically is


largely irrelevant. What the natural period

turns out to be IS very important . It
frequently happens that changes made to
improve the performance will result in using
For instance a particular temperature loop a higher gain, but the gain itself did not
may have a transmitter span of 5 or 500 cause the improvement, the changes did.
degrees. The ultimate gain in the two cases
will differ by a factor of I 00, yet the Before I explain the factors which

performance ts the same. There may be determine I need to convey another
accuracy problems related to the wider span important concept about what a control loop
• •
but this IS a separate subject. All that IS can and cannot do to reduce the effect of

important lS that the controller, within the upsets.
rangeof its adjustments, be able to reach the
ultimate gain, or at the other end of the
34 Chapter 3,, Tuning Objectives and
Expected Loop Performance

..
In the example just given the upset was not able to do a pretty good job of
defined, either numerically or in kind. It was· compensating for it. Of course you may not
implied that
. .
the: upset came and stayed; the
.· . . ..
have time to wash, but that is a separate
concepts presented were 'important, There is problem!
· another concept which is also important, but·
it must be presented in a different way.
. .

Imagine an upset which does not come _ and


stay, but one which cycles sinusoidally. You
. .

do not even have to think of a real situation, This· case is treated last because it is more
. .

just hypothesize. Now let's consider three complicated, less · intuitive. It happens that
situations: · - · · · · control action is ·a·· detriment. Errors are . .

worse · with control than without. This


happens because the controller output is
'

zigging when it should be zagging. An


. . . . .

analogy often
.
given is that of a hand-held . .

mass suspended by a spring. If you ''disturb''


. . . . . .

With a . little thought, you will be able to your hand holding the spring at just the right
. . ·. . .

accept that· if the. disturbance is fast enough period, the mass will move much more than
(short period) the control system will be too your hand. If that analogy helps, fine. If it
slow to do anything about it. Control action does not, .: please accept the conclusion
gives no benefit for this situation. Imagine anyway. 1 · can't think of another example
you are in a shower and the hot water within the range of normal experience. The
pressure changes with . a period of half a shower example breaks down because your
second. You simply will not . be .able to brain would soon tell you you were making
compensate fast enough. things worse and you would quit trying.
. . ' . .

· •·The · overall picture is 'presented in Figure


.: 3 .3. Variations in tuning modify the center .
- · .and right-hand sections of the curve. Higher
. . : gain makes things · worse near the natural
. .
. .. . . . ·.
After contemplating this condition . for a period, with some improvement in
while, you will be able to accept that control attenuating properties for the long periods,
will do some good, as compared to no as shown by the dotted line. Not shown are
control. Also that· as the period of the upset
. .. . that, shorter, integral • times . help with the
.

gets longer and longer, the. integral action of attenuation part, at the sacrifice of more .
the controller has more time to work, so ,the···. _ · resonance near the natural period. It is much
benefit of control is proportional to the. · ' · the same as increased gain but with
period. Again . the shower example.' If the '.
. ~ . . differences too subtle to discuss here.
. .

hot water pressure changes sinusoidally


with a period of one minute, you will be

Chapter 3, Tuning Objectives and 35


Expected Loop Performance
. .. . "
Control. Helps ..... ·• ..
...:. in Proportion .• ·
· ·.... · ·. 1:0· _. Per·,o· ··d · · · · · · " ; -, ·. ·. ·,
..·.... . .... . . .
• • • , • •• , ,.. >

oti. Disturbance · . ·
r ...
-
-..... ..ec
·O
c 0
0
••
..
·-.. . .c.. :,0
0 0
0
..
;-
I
\ 1
-,
',

! ·-;:· .. .&:. I 1
--0 ,:s ·-3 Natural
I
en :,
..
Cl)

,::,
Cl)
PeriodL
• I ' <, 1
.9 - ..:, -a. -a. I -,
'
E
I -,
<( E
c(
I
I
-r: Higher Gain

Log of Period of Disturbance

Figure 3.3. Control action can either help, hurt or do nothing to. reduce the .
effect of a disturbance, depending on whether the disturbance is slq~er,
".
at or
faster than the· natural period. ·
.. . .
" .

The part to. remember is that if the


. . .
natural period, the untrained person is likely
disturbance· is 'periodic, the control system to conclude that the controller is. 11.Qt doing
its job. You. need to recognize this· situation,
. . ~~ ' .

can either be -belpless; -. aggravating, or


helpful, depending on the relationship of the inform tae appropriate: persons. that more .•
disturbance period to the natural period. · efforts on tuning will'. not do the. job, and .
. . . . .

. .. ·., ;. -". ··.... . . . . .. . ·.. . : ' ..

You could be thinking that disturbances


don't come in nice cycles. This is true,
though sometimes a disturbance to one loop
is the result of another loop cycling. Even if
that never happens to . you, the concept is
important. If the controlled variable is not
. .

steady, observe whether the variations are . '. . .

fast or slow relative to the natural period. advise them that · another ·. approach is in
You will . · · then know whether · the order. Loops with long· dead times (say
controller might have a chance of doing . longer. than .10 minutes) .are . frequently the
something about it or not. subjects of complaints on their performance,
because it. takes so . long to .recover from
. . . . .

It is very important to understand that in disturbances. In this case it is advisable to


some situations a well tuned controller will do · something about the: disturbance in the
. . . .. .

not produce the desired results. If the first place, if possible.


disturbances are too fast relative to the

. ' .
36 Chapter 3, Tuning Objectives and
Expected Loop Performance

Since the natural period is so Important in Where: - Integral time


assessing expected performance and tn •
-- Natural period

setting the reset and derivative times, it IS
important to know what factors· determine T·1 -
- SL for the open-loop testing

the natural period. A great revelation IS
about to descend on you. Where: L = Apparent dead time

Look again at the two equations for tuning: Combine these two equations and you have
the most important simple concept ID •
T·1 -- 1.2 Pn for the closed-loop testing control:

When you get a little further along, I hope


you can look back and see this for the
,,_,,
Pn - 2L
great insight it provides. The symbol -
- '
meaning approximate, is used because the Time
tuning rulesare approximate. Actually, the • L •
a. Pure Dead Time Only
use of the tuning rules to ''prove'' this
relationship is not valid. The tuning rules
are based on more scientific principles and
the proof is in those scientific principles. Pn -- 4L
• Time
The above relationship of Pn to L JS an

L •

approximation. It IS possible to be more b. Pure Dead Time Plus Integrator
precise. In Figure 3.4 are shown two
extremes and something in between. If the i--Cl)
-..a
step response of the process is as shown, ec CG~
... ·-
then the relationship given to the right 2L < Pn < 4L
o-J
(.)
applies. It is possible to confirm the first
Time
example (Figure 3.4a), by logically L
drawing out what happens as a function of c. Mixture

time around the loop when it IS cycling.
• Figure 3.4. The natural period varies between 2L
Refer to appendix page A-2. This one ts
and 4L, depending on. the other lags in the
easy, and the equation Pn=2L is exact. system.
,.

Chapter 3,-Tuning Objectives and 37


Expected Loop Performance

Before tackling the next example, accept on
:,,
control ;Joops~ that lS where I started by
...... · ,,,,. ,· .

faith that th~~'.' step . tespql).se shown can


. .-~ ·. .~ . '. , .
combining · the tuning rules for open and
legitimately· be divided cftifb.,~;?~-4~~~ ~~!Be closed-loep.. but- now, .hopefully you have
some :.iasight-::·;clQ.$.0/ ~t,Jae. ,,ason whr, . where
'> .... ,,...~.,,, ··~~-·.. :;·"'·. .;>·- \, .....

followed by an integrator. Tlie1·1fntegfator


has the characteristic that the output lags the before it was just.the :result of·.some algebra.
• .. : '·
input by I/4 ·of" a cycle when the loop ts
,,.·:
:·· ..

cycling. This will . take some. explaining.


I

Refer-to Figure 3.5. When a loop Is cycling

I .'
I
I
I Now you may feel as if · you've learned
. . .

I something, and you41:ve, but as soon as you


=a.c try to apply this knowledge you will have
- more questions. A few processes have a true
dead time, but most do not. So what
Tlme contributes to the apparent dead time? Well,
first the answer, and, then a .long explanation
Figure 3.5. When a loop is cycling the output of
an integrator lags .the input by 90°. of terms and examples. The apparent dead
• •
time rs any pure dead time plus all the little
the integrator output is moving at its fastest •
lags in a control loop. The paradox ts that
. .

rate when the input is at. its maximum. The


. . . . . . . the · largest lag in a control loop, sometimes
result is that- the output of the integrator lags an integrating element, has little or nothing
••
the input by 1/4 of the. period of the sine to do with the natural period.
wave, or 900 as 'shown. It will always lag
the input by 90°. The . amplitude of· the
output wave will not match the amplitude of
the· input wave, as · it did for a pure dead
time, but the phase lag will always be 90°.
This second example, of dead time plus
integrator, is shown on appendix page A-3.
It is admittedly harder to follow the logic of
the signal around the loop than it was for the ,..

pure dead time case, but try it. The Thisis a difficultconcept to accept. To the
relationship P -- 4L is exact.
~ . .

untraineditseemsalmost contradictory. It is
reasonable to think that the speed of a loop
The last example, Figure 3.4c, is more like will depend on the lags, but it IS

real life, but no illustration comparable to unreasonable to think that the largest lag
pages ·A-2 and A-3 is provided. I appeal to has nothing to do with speed ( of the loop).
your sense of reasonableness that P n will Some may be asking about now, what a lag
fall between 2L and 4L. Place permanently is. Let's just leave the above paradox, go off
in your memory that Pn = 4L and you will
. .

to understand some · lags, and then come


cover a preponderant majority of industrial back to try to tie everything together.
...... ~- , •. ,. ,.,, ....,, . ._., _,..,, ;·,·..... .. . . : ,:~ ··~ ,,_ ·-,··";i."•·. "• . . . . .
. ,~:~:-·::.: ,.,.•.·.;., ··.. ·'.\·:::' · ,·: ' . . . ': ··· .. ··> /7:::~\{.:.:;.;~/:::·.?\'*"i.·:·'.: .
. . . .. . : :.. ,, :;' ·.' ~;' ;::-~:~:~;(;;?- . . . .. . ..
· · · . ·· ., .· ·:. .:-.~::.:·:.·>:)·f?f~r:.~~·., . ·: ·. . . . ·.. .: .
. . ·~ .· . ~. .
\
... . . ., . . .
. .
, . ': ', .
.. ·
..

38 Chapter 3,,Ttm;ing Objectives and


Expected Lo.op Performance . · ·•. ·. .: .·

. outptn. amplitude · to inf)#~-~~~/jtriplltude)~ ...


. . . . . . . . . . . . .
.
.

.
'
.
.
· When the loop is cy~Jir,g, if you. _multiply ..
I
:i
· . . · · : I will. venture at this point to introduce a all the · amplitude . ratios together, the .
· · : · · concept you· must understand if ·you are result is unity, and dimensionless. · . ·. · ·
to get into the math I algebra of control · ..·. .

loops. The key is to think of wne: is , ·_ No,w, . lets


. look the . phase . angles at
.. . happe(iin.n . WfJ•n · .:~e:i,:· '.:loofi::· _ .,; . ,;,;;Nin• .g·· ·- .· .·.
· · contributed by each element in tne ·
. . ' . ·.. .: : ....... -~•-, ,,,?f;. ..; :. . - -~f:l:.. :. :~>,, ... 'If.;,:;,,,;,:~-""-: : .,,,,,,;r.;J.iilti •·.
.Rete.r'aga;ii-to t,ig.ei';A-2 iiifci)A~3 : > :::> · • · ·; example· on page· ·A-2. · The controller.·,
ii . . '
. .

.....•..... · . . . ·. ·. ·.·.·.· . . ;, : < ; i. f • : .< \ • ' .. <i. ; idealized; has no phase lag. A pure dead · ·
_·. When· .tlJ~ ;IO()p !.f ~y~l{IJg,::.~v.ety_ppint in . ·'. · time has a phase lag of:
.• the loop is cy~lirig.· wn·en. the icyc/e gets . '.
· back to -: any point ir,. : the loop · it: started . • 360L
Phase lag=
· from it matches the,. cycle · there, so the · p
. . . ~. .. . : ..

· . cycle·sustainsitse/l ·If it did,·notmatch,··in · ·


. amplitude_ and phase,. it would not keep Where: L = dead time
. . cycling in a steady msnner: Either the ·. . . . P=period
· · cycle·would grow or decay. · ·· ·
·- · - : · . :_ . ·.. · ....· :- . . -._. . -. . . .. ·. · . . ,· ._.. . . . . . . _ -·. !his· -, · should . be' easy to understand. ·
. · There are methods for determining . . .· ·. · When the dead time and th~ period are
. · whether the cycle will g~t worse .0, •. the s~me, the phase lag is 3600. So
. better, but tl:,ey are too compex.to ta~ . ·. . ··•·.. wha! ts P ~hen the phase lag is 1800?
. . here. The phastJ lag around the lcx;,p is . . . . YQu ve got it!
. . 36C1'. wh_,n it is :cycling, or the phase is · · · ··
· minus 3~()0'. H,a!f Qf this comes from the · .··· . • .· · .. · ·.· p =.. 360L = 360L = 2L r · .. · ..
. cont~l!er :its~/~-· Notiee in the signal-flow .·. . ;c . . · · .· ._·• Ptlase lag 180 ·
. . diag~rn' for thfl. co.n~toller (page A-2) that · ··· · · ·
the controlled \tar/able ls subtractecl . All h,u have to do is adjust the cbl1fro11er
·. · from the setpoint.: · · · .. · · · · · · · - · ·:gain . to make the amplitude right, and a
·· - ,, . .. . . · · , .: , ·. . · . · · cycle at P = 2 L will result. -· · · -· · · ·.
. . .. ·.
..
This minus sign has to. be , · · else . the
.. . . .. . .•. . . . . . . . .
: ,.
.

controller . woufd· ':.SC.t_ . -·.in . · · . the wrong · .: Now lets take the example on·page A-3. · ·
direction.· · The · minus sign · inverls, .· the ·· · The . integrator has e phase·lag .of 900- all .
.· sine wave, which ends up being :a phase · ' · :· the time; at any period. So the dead time .
· lag_ of · 1800, .. whtJn the loop is . · cyg(ing~ · has . to. _contribute only 9<P:. ·What· ·is P .•
This leaves 1800. to be contributed by · .· wnen the phase lag is 900? .
the lags. · . ·. · · · ; . ·_ · ·. '.. ·. , _ · · · ·_-. . · ·
. . . . P= _ 360L = 360L =4L
When the loop is· cycling the ·contribution Phase Ja.g . . 90
of each element to the 1800 lag is a· · .. . . '

specific value, as well as another Marvelous! Was that hard? : ·


parameter, the amplitude ratio (ratio of
.·.. '

. .. ... . . . ; . ·.
.. . .. ....; . .
··, .

,.. .
..·
. .
. ·, '
. . . ·,

. ·'
39

...,,

.·~
...

..,

...
.'

The word lag


••
ISgeneral. It means any are close enough. There are other building-

relationship 1n which some result happens block lags which the more experienced

after some cause. There IS an associated practitioner may use, but the need for them

word, lead, which in some. respects ts the is rare and they will not be discussed here.
.'

• opposite (though no· result can happen


before some cause). It will · be discussed

more later 1n the special section on
derivative action. You cannot really In the typical control loop, lags act in series,
understand control without understanding· the output'.of one being the input to another.
lags, but there are different depths of For instance, .the lags around a simple
understanding, and it is possible to get a temperature-loop might be:
useful amount · of understanding with a
modest amount of training. No proofs will
be given. Most and possibly all of the facts
• The output of a controller is the input to
a pneumatic transmission lag.
used are consistent with reasoning powers.

• ., The output of this transmission lag IS
the input to a valve lag.

The output of the valve lag is the input


FIRST A FEW POIN'l'S. • to a process heat lag.
. .

•• The output ofthe process heatlag is the


input to a measurement lag.
In real life, mathematically exact

descriptions of lags are very complex, and The output of the measurement lag
often beyond definition. Small matter. Their
• the input to a transmission lag.
IS


effect 1n a control loop can be represented
by combining three relatively simple The output of the transmission lag is the
building-block lags. While these lags are not
• input to a controller lag.
exactly the same as the real system, they
-e ,'. , ·.. ~-" ~,.·
. ,. .. ,',,"'-·~·.,,
':
.... ·i ... .·.
. ,,..
•.. ·, ..
._;,; -
,,

40 Chapter 4, Lags and Gains, but Mostly Lags

Pneumatic Valve Pneumatic


Controller Transmission Lag Process Measurement Transmission
{ .;._ :·~-- .
' '

a. Lags In logical order .. ·......


' '
·,:·.'·. ~ c

.. . . : - Pneumatic) Pneumatic Process Valve


Controller Measurement Transmission Transmission (Heat) Lag

,
....,.,
i ,.
._ __ -,1_,.
•;
.._,..
,, . .....
-~ .. _

b. ~ags-•ln scrambled order ,,, ;,.:


,',

Figure 4.1. For the overall lag, it doesn't matter whether the component lags are in their proper
sequence or scrambled.
·~::

were noninteracting lags. Figure


illustrates -the difference. In the
In this case six lags have been enumerated, noninteracting case the level in the second
each in itself quite complex if you were to tank does not affect the flow out of the first,
study it closely. Their net series effect- can in the interacting case it does. The point
be represented by two or three building- made about being able to scramble the order
block lags. - That was the last point. This of the lags .applies only to noninteracting

point IS that, as long as you are not lags, interacting lags must be treated as a

interested 1n some intermediate ; point, it single dynamic element.
doesn't matter what order the lags are in or •

what - order the building blocks are tn. A This ts more of an academic point than

• •
signal-flow diagram IS shown IR Figure real one at this stage ID the development,
4.la. The order of the lags .may be . . . . .

scrambled, as shown in figure 4.1 b, which .


. • • . .

makes no physical sense, but the lag from


beginning to end will be the same,
scrambled or not scrambled.
l

' A

a. Interacting -lags

___.. k,.__ _

-. . .
II\\ .

------"A\\\' . . . . .

Perhaps here is the place to make a ._ ~A~


distinction. It is necessary to understand that b. Noninteracting lags

10 real life there are interacting and
noninteracting lags. All of the above lags Figure 4.2. Interacting lags must be treated as a
set of inseparable lags.
L·:
L.... d G' · . • . . b . M· · , : . ; • . : ..
C~·P·• ~, · ags an ·. · -· a1ns, : . ut, . • qttlJf-:i :.'.·:ags
<;f.. , · . ; .. i"'.k.A·
,,,:· 41..
. .

,,

. ".'

••
Ill a.
• •

to .
:t

:,<\: . , •.... ,..


.. ., . -s
..• '• . .~:·:·:.. :; .. . . ,,.... :,.7;. :···.: ..
. •" .. :
..
,'
• .. "
"
.., .,.. .-. ·f .~ ·:

." }: :·
. ·.,~....:
,· .. .: . ·.:·;
'•
··' .- .,, ··< ·.:..
"" -. ," ·• ':t ~·..• ·:·,':'.·
..
•:.• ·:
,...
•••••
"
• •• o'. •
.
. '.. ...., ,.
",.
• •
' ..•. • •
. .. ~
.
. . . . ·.: ·,. •
.:.:'
,.
,• •
..
,.•"
,. '
"•J ~
..
'···

.· :··~ ...··,
. ··.· .. ,.· «,..
,;
,r,
•·r .•

. ·.,: .'
, ... ,
•'

The following. names are givento the three· In 14~:· ~a.tblt1~8~~~"·~i9.ns1l,,::will' ii,tt~c.e
types ''i)f. lags that may . bei· combined to their; ~lap,e.~q-~ns{()fffis:···4iai:ve . :YOU · -, a_:c.w
represent most physical -systems~ form11JA,···_\~'?.fr~,P@~~-:i::::~d ir~;ffli~~.j··f~!PS:~ . : .• I
. .:. ~·..
. ". '. ..
.. . {"\: ·· .
. .:· . : ,.: . . .· . ·.: .. . . ... ... :... . :: .... · .. . .: ~ ·.·.: ··:. ·.' :: .. - ,:::··;
... ....~:.,
·<); ... • . .,., : .....
will •.. use. ~~~·rt9 .::~l.luttr~l~ . . a. ;'.f~W,:;:: ,ll.9utts~
... ·.,', ·. ' .
Most people-. :\Ii~· :~~//<>Jl-lY·rt.98.: . a.;; ty~. pf
.
.. .....···.. .. : ·. ..
:,,.
algebraic shorthand : .jn. . ~i~g,. wi.thqyf
. \:

... worrying about ~e more, . fundamental


• l9ttgrator . . ,,:
•• ;. •• > •• -'
... ·•. ; -, .. ..... i~
~ ~ ...,·
.... , ..
,... :.. ;· .
. ,..., ......... :. ,,.. . . .
.. ;
.. derivations, and without, using , th~ . \to .
••••.1. ·},, •• "; ..

actually solve for anything. . ·.


... . . .·' i .~·, :-.
.. . :, .;,...;. :.

• First order lag ,,

.·.... .·.
."· .... '., ..;
·,
•• ."

. .,.. :.,." .
.,..
""
...·:· .• :· .:..
"
.. ~· .· ., . .... .,:·
' .: ' :
; ..· ..... ,,·
:, ....
........
.: , :, " . ..' . '·
;: . .
-.. ..-:' ·. ·~~ -:: ...
·. ..
·". ......
·.,:.. .. J·",,.:·~.:·
t { :.; . . .. .., ,

. ,,. . ··~ ~· . . . '.
'"
. . .'· . .:/··
•,
··?t'::.:.:..:,
.;. .. ?
.·.;:'.: . . »..:
. ., ,, ,,
-;
... .:.
·.. ,,:· ! f

~ . ' . ' .' . . "· ,.


: . ·. ,• . . . .
.
. . ·: ; ... /·
. .
: ,: .
.
. ·:.,,.. ,.; : ·'. ."
,.
· ':t : ·. ·?.r·::?. · ,.
·,,
"'·
..:: ::;.
.
"

...·:.,
: : ..'.,..
. . . ·: \.;
"
., ..
·... ··:·~··,~ :•: ;:: I) .f. :. '·
-,'
. . ~ .. : . : : ·:· ::.' e<: : './
... . . . ..· . • . '·. •, le-
..• "

'
"
..
.. ·..' ..;.
!' ·.' / .. . ·,

. '
. ··.,

.. ." ,. ,.
: .•. ?,.
. . .
...
. ':: ...

"··.. ,. ..·~
·~'
,. :
<
',.. . .,
...;· ~
:. ' ,· .,:
. ......

,.. ' " .


.; '
.?:-; ;
" . . .·
. : ' '

..
. ·. . '·. ." -· " .
.. : ·. . '.. -·..
.

. .; ..

·, .
. . ··. ; .:':':.·,,:. . . .. ·'
',·.····.··
." ••.....· ··.·,<. • ·< ! .
! .: .< • • •• • 'O ' ;

•• . .. ··" .....
-...
•·: ..
' :...

. ·,

:· ...
" .. i : •.
,.. ': ...
-;

'.~ ..,
.
. ·.:···....' ;.' ;;: .
,·, .
.-::,.•·.
~.:.. ,.:,. ··.
. .:. ~: ..'
··;.:.· .., ,
. ..,: ...; .... ..~ . .. .·.: ..
' .. ;, ·. .
..
.. ,. ,,, : .. >'•• ' •
l, .

...... ..~.,. ...


.,, .

,. . ..• ;
.... '
·.," ..
. ... , .
: ;; .
.,. ·~· .
~. . ·...,..>,. ·.. ,. ...
·~ ·.:.
.

' .,
,

..,- •.. ' .


.,
.;•• <" • .' "
",•.
,.
,· ., . c. ·"-, ...... -:
': ,: .... ;'. . '·

".

. ..... :,
" ... j.
~· •i· • • ~
l· • . .
...

. .
.f " . •. . : ..
• •

. . .,.. .,:,' :' . .. .. ,' .,·:.:·:: / .


.. '
' ..
•.·
.'. ' ..,·.· '~·
. ' ', ,.:. ...." • .: . ;, ,:

.... . ..
'
'
. ··-·
' : .. '.

.· • _:' :?;:/{Jif,\fi .
.
. ...

42 Chapter 4, · Lags and Gains, but Mostly Lags .

The dead time step response is shown . ~ Time of Input Step Change. . ., .

graphically in Figure 4.3. A cause or


. .

input occurs, and nothing happens until· · · 'S • L -•


after a time called the dead time. Then · · · f L = Dead Time · . . . .,
0
everything comes out the same as it Laplace Transform: e·Ls
went in. This phenomenon almost - ·. ·
always and possibly always involves the · _ · _ . _ Time
transport of material and/or energy from ': ·.: · - · . . .
one physical location to another: Some · · ~igure,4.3. With a dead time lag th~ output follows the
· · lit t · · ~- . - t - it ·. - . , . . , _ . . ·
1 era ure re1ers o -· 1 as a ,1 ans o, ta.tton · · ·
p·i ·;....A.~ ; input exactly, but delayed by the dead time. ·.
.
lag.·oradi~~c·e~~~c1·fy·:ja·g;·.·Ane~~mp·Ie..
. •.

. g. to · .a . fl1 1·m
~IErn~~~~~~~~~~r~r-m•
• • ... . ..
.. ,., . . ..... .. . ......
•.••••• ··,;.,.·.;,;.,,,,,,,.,,;.·.·,·:i:.•·>i·,•,•,· .' ,(~. .
,,,l*.;; ~~ .,,.,.• ,. ,,,,,,,,,,, ,,,,,.,.,,,,,, ,,,,, ,,,,,,,,,,,,,,, ,,,, ,,, . ,,.. . . ·,;. . ·...,.;. •.· .·.·.,.·.·.·.·.·.·.·.·.·.·.·.·.·.,.·.·.·.·.·.·.·.·.·.·.·.
.

e app . ..m. g a coatin . ~:.·. ,·_. :,.•:,.:,.,:·. ,:. ,:.-·:,._·:., .; , ._., ., ., ., ., ., ;,. , ., ., , ., ., ., ., _., ._.;,_- . tmr. .
,:, : ·~; ,,,,,,,,,, ,,,,,,,,,,,,,, ,,, •.•.·•·.·.·.·.·.·.·.·.·.·.·.·.·.·.

wou . . · ·. . . /~~t{:111111111111111111itt::.=:.:i ·.t·.·. :l..: : :·:·:·:·:;: .....:''·: ', f ..


it .../ .....:·:.d L...::1:1:1t..Jf:1:... ·:~···:i~:: rrif~~~~?tt rrrrr
..... ..:i1ft :.::::::Jii::..:::::·:::.•.....,:,,·:,:,:,., .•, .,:·:,:,:,, ••·•·•·· .,,
..::1111mmr ...)1:i:;:;. ..... ·.·.·.·.·.·.,.·.·.·.·:·:·,,:·.,:,.·,, :1111\ ~~·: ..:::::::::...tr

Id b I .

. ' . . . . ·. ..
:-::·:·:·:·:·:,:·:·:·:,:·:·:·::X·:···:,:···:,:,:·:·:·:·:·:· .·:·.·····:·:·:,:,:•:,···,,.·., .•.·.·.·.·.,.·.,,,.·.,,·.·.·.·.·.,.,,· .. :,:,:,;,:,:,:,:, ..... :·•,:,:,•,•,;,·,;.·,.·.;.;,:,•,•,·.·.;.;. ,.·.·.·.·.·.;.;.·.·.·.;.;.·.·.·.;,·
.,.,.,.,.,.,.,.,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,.,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,_:,_,,,,,.,,,,,,,,,,,,,:c,· _.,.,,,,,·,,,,,,,,:,,,,,.,,,,,,,,,,,,,,,,,._.,::=·,.,·,,,,,,,,,,,,,,,,,,,,,,,,,

., . '.

where the: thickness · of the coating . is The Laplace transform for a dead time,
measured downstream. . . . •
,s:
'

Frequently, adjectives · are· used to


differentiate between this lag and others that
have essentially the same effect on control. where e is 2. 72, the base of the natural
This one is a ''pure'' dead time, while others logarithms. It is a characteristic · of the
Laplace transform that you may
are called ''apparent'' or ''effective'' dead
substitute for s as follows:
time. The distinction is important, not only
to • be mathematically · and grammatically s = Jro •

correct, but to . convey the · facts. Soon you


will· learn that several of the last type of lag ro = 21t f = 6.28 IP
to be discussed may combine to act like a
dead time in the loop, but they are an to get the frequency or p~riod
apparent dead time. ·response, where

If the input to a dead time lag is a sine j = imaginary notation


wave, the output will be a sine wave of the =
co (omega) angular frequency =
1t =pi= 3. 14
same amplitude, for all input periods. The
.
..

dead time does not filter or dampen any


f = frequency 1 IP =
L = dead time
input, but simply delays it. The phase of the P =period
output wave lags that of the input wave by:
You don't · really want to do that
360L . substitution, especially for the dead time.
p Computer solutions involving dead time
are done by simulation, rather than by
solving the Laplace transform. The
degrees. This makes reasoned sense. When
transform is still ve,y useful as a form of
L and P match, the phase lag is one full notation, to convey what is being
cycle, or 360°. discussed. The.· substitution is more
useful for other lags than for deadtime.
'
~..
r
·;

'.
'
'

Chapters, Lags and Gains, but> Mostly Lags 43

. . ...,.,,
. . . ...
. . .. . .. . ·. .. .. ....;- .

The step response of an integt~t.~ or . .


. .,. ...
.

integrating lag is shown in Figure 4.4.


'
>,z; ~Time of Step Input Change. . . .

A level control . with a . positive . . ~


...:::,
displacement pump on either inflow or ...:::,
Q. .
. .
. . . 1 . .
. . .

outflow would represent a loop with 0 Laplace Transforrn: Ts .


an integrator.
Time
In . some respects it is the opposite of a
dead time lag. The integrator . always. Figure 4.4. An integrator lag responds to a step input by
has a phase lag of 900, regardless of a ramp output.
the period of. oscillation. This was
illustrated in figure 3.5. Remember the. dead .
1
time had a phase lag proportional to the t»
..
period.
After substituting ·. for s . the ·amplitude
While · the output of the dead time matched ratio is:
at
the input . all periods, the output of an
p
integrator. does not. It . increases as the . '
..
6.28T
period gets longer, Which makes reasoned
. .. . : .
.

sense, as . there . is - more · time for the As just stated, the longer the period the i ·
integration to take place. · . · more time there is to integrate the input · .
wave, so the amplitude is proportional to ·
. .

period .. The longer·· the integrator's time


parameter the slower it is to respond, so
. . . the amplitude ratio is inversely
The Laplace transform for the integrator proportional to it~ ',
. . . . . . .
• . . .

IS:
. ...; ..
:

. .

This next building-block lag goes by many Its slope at any point is such that it would
names. It has . been called a time constant '
reach the final value in a time called the
lag, an RC lag, an exponential lag, a Shuple . time constant. It .actually goes 63% of the
lag, a single linear Iag, arid surely many . .
way to the final value in one time constant,
others. I will call it a first order lag. This is 95% in three time constants, 99% in five
a somewhat arbitrary choice, based on what time constants. This holds at any point on
seems to be gaining in popular usage. The the curve. If it had gone 23% of the way, it
differential equation that describes it is first would go 63% of the rest of the way in one
order. This is still a term that will mean time constant. Much more could be said
nothing to you unless you are schooled in about this one's characteristics. It results
differential equations. Its step response is from a resistance to energy or material flow
shown in Figure 4.5. and a storage for that energy or material. An
example would be the temperature in a
••4.
.,·

:,

Chapter 4, Lags and ·Gains, but Mostly Lags


. . . ~ . . .... .
stirred tank, responding to ·
.. . ~ . . .·

.#1---Time of Step Input Change


a · change in wall tempera- / . )!~j})ii '{ ' ' .·• . ·
ture. ..::, 7' T = Time Constant ·· ·
a.
....,
::,
0 I t
63°k
100°,4 Laplace Transform:

! 1
f

The first order lag is a


Ts+ 1
I
cross between the pure 1- .. T I I
dead time and the Time
integrator. The way it
responds to a sinusoidal Figure 4.5. A first order lag· will respond exponentially to a step input
input depends on the change.
period of the oscillation
relative to its time constant. If the period The phase Jag is:
..
:
<

is short relative to the time constant, then ·


the phase lag and the amplitude ratio will

The notation ''tan-1,, is read as ''the angle .


whose tangent is. ''

These afe complicated and people don't


. .
solve them by hand. If you decide to
..

follow the formula for the integrator. If the . learn more about . frequency response
period is long relative to the time constant,
. .
.: ·.·techniques,. you will soon memorize .
then it will approximate the phase lag and these characteristics. They are
amplitude ratio of a dead time. This might commonly presented in plots called
be accepted by reasoning, but· it is not . as Bode plots, which will not be presented
in this booklet. ·
easy as the dead time and integrator cases.
Loo« at the Laplace transform. Wnen Ts
. . . ..

in the denominator is large relative to


one, then the denominator is essentially
The frequency response of the first order Ts, which makes the transform the same
. . lag is the most complicated of the·. three. . · as that . ·for · the integrator.··.· Now again,
simple building blocks.· Both the
. .
look at the denominator·. of the transfer ·
. . ..

amplitude.· ratio . and the · phase lag function. When Ts is · small relative to
. · depend on the period. .: The Laplace , one, then the transform approaches one .
transform is: . .

When Ts is one. (when 21tTIP is one), . . . .

1 the amplitude ratio · is · 0. 10·1 and the·


Ts+1 phase Jag is 45°, . a condition commonly .
called the corner frequency, because in ·
· The amplitude ratio is.· . '
· a Bode plot it is where the asymptotes to
. . .

· · the low·. ·frequency and high frequency ·


. . 1 ·. ·1 .. - ...._ . ...· · · characteristics meet.
= --========-
1 + (co T) 2 6.28T 2 . > '
·.. 1+ .

p
,

Chapter 4, Lags and Gains, but Mostly Lags 45

•• • ·, ••• • ; •••'·; $' •• : • • ~- ' • •

-.;. . . . ·.;.•; .. .,.·.··;• .. ,,.,. ... ••vt;·\:··..~.~ ·~:'-·t ... .


. .:.•.
. f .. ~-·r . ~:~.}.7~~~:f·..
·'·.·;.;~'¢.f~i:~> ..... ;.,·~,
. .,....• ~·~·1<f;tP''':.

.
_{;: t.,.. -~.:~/~· ;.,. <: ...

Now lets· combine these building-block lags first order· lags. are .summed · and drawn first,
'
to approximate a real . system ... Remember to delay theintegrator or Jargest first order
that I said the order of the lags in a loop lag by that amount.
could be scrambledvwith no change in the
overall . result. This · is the same as saying So what· happens when there is · no
that the step response of the system will integrator, no real dead time, and no first
look the same regardless of the order of the order lag that is predominantly longer than
. . ··,..
. .
t.
.· '. . . .
...,.. . :., .
. the rest? In that case, ·
there is no really
. .
simple rule · : to
..
• ·1 · Integrator· Drawn Last
determine. an apparent
, . .
'
. .
dead time, that· is, the
'parameter · that would
. .. . be used to predict the
, •.•
. .

First order 'lags su~,med to d?,lay the _integrating element "natural period. But all
by that- amount, an apparent dead time, drawn next, . ·... : . ; ... is net lost. It is known
.' . P~re Dead Tim~Dr~wn.First. . ·. · . · < that -the step response
'. · · ·. .
·.; . · '. ·will ·be· 50 to 63% . . \

Fig~re.~.6. The step response of a series of lags may be drawn by combining ·. complete . in . a time
the 1nd1vtdual tags as shown. . · . . · . , ·. ·. •. . •. . . equal to the . sum of
. ..

lags. While· this is still true it is not true that . . . · ·•· · · ·. ·. ·. .: · .· the; time. constants, as
the .step response may be drawn in any
. . .
shown in Figure 4.7.
- order. Exactly the opposite is true, as will . . ~

become evident as . you assimilate · the . The .· 50% . point is the · limit for an infinite
method. series ~- of infinitely small first . order lags
(pure dead time). and the 63% point is the
Suppose a process has pure dead time limit for a single. . first order lag .. With that
"

anywhere in a· series of Iags. When drawing knowledge, it is possible 'to put some limit
the step response of the entire process from on the apparent dead time. Actually, in this
left to right; the dead time is drawn first,
.. '
case, the natural period is not equal to four
..

even though it may not occur physically


first in the process. If the lag set 'has in jt. · . .. . .. . . .~

an integrator, then I think . it is appare11,t 100°/o


"

this . effect would be drawn


. : last. 1'r. the
. ·. . . ·,,-: ' .. . ; ' . . .

process bas first · order lags, · these are ....::,


.
summed, and the integrating line is A
'
,.
'·,.
.'
l•
delayed by this amount. All of this is
....Q.
:::, t
50°/o to 63°/o of A
. .
t 0
f.;,
"
shown in Figure 4.6 .. ~

Time
If the lag set does not have an integrator,
but rather a first order lag which is long
v' relative to the rest, it is drawn last and the
' Figure 4. 7. If very little is known about the component lags
·'
',

,'
treatment is the same. If there is no pure in a loop, something about the sum of all the lags can be
·,.

.,"

'
,
dead time in the system, then the smaller deduced from knowing the response time as shown above.
"
;
·-.
'
.
,

46 Chapter 4, Lags and Gains, but Mostly Lags - ·

times the apparent "dead time, so a further . . . .


· · So now I hope you nave some insight . ,
. .. . .

• approximation exists
.
·_-_- in predicting loop . .
leading to acceptance of the p~~do,i:::·stated ....
perf onnance. · That is ·. the penalty paid · previously: it is the sma.ll Iiig~: arid , dead : ·

:i..
without· getting more complicated. · time, which determine· L · and, therefore the
. natural period. The largest lag or the
A useful analogy for understanding the integrator have little or · no bearing on
way lags.combine is the way vehicles start - .. . . ·. . .'
establishing the natural period.
out after a traffic light turns from.red.to ·
green. Each vehicle responds after the one
preceeding it, similar to· the way elements · A paradox!
in a control loop. respond to the · one
preceeding it in the:" loop. Some. drivers
respond slowly, some more quickly. If you
are at the end of a long line it doesn't
matter whether the slowest driver was first There is not much more I want to say ·
in line, in the middle somewhere or just in . . . about the . math I algebra at this point
front of you, the rate at which you can except to point out the similarity of the ...

accelerate is the same. first order lag to the dead· time wht!Jn the
.· . . . . .

period is long relative to the time


Since you now. know how to estimate the . constant or dead time. - The following
apparent dead time from the individual. lags · table shows these similarities:
in _ the system, you also know how to
estimate the natural period. Actually the Phase lag, Amplitude ratio
step response· does not even need to be - .
·.
'
degrees
..
. . '...
.

drawn; you can calculate the apparent dead TIP First Dead First .
Dead
. .

time directly from: the .estimated lags. or order time order time.
LIP la g
.
lag
······ : ......................................................•.······································-····· . . . .. . . . . . . . . . . . . . . . . . .. . . . .·.~.·.·.. . . . .... . . . . . . . . . . . . . . . . . .. . 0.01 3.59 3.6 0.998 1.0
.. . . . . . . .. ,. . . ,.. . . . .. . . ..
sm
· .,
..·u· .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 11·················
. . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . .. . .. . . . . . . .
..
. , ,,, .. , ~ ,,,,,,,,,.,,,, : : ...•...... ,.,,, , ,,.,
........... .. . . .. ..
•. ,, , .. ,, _,. ..., . .. .·1·. • """
···s···-,, .. ,· ·; ····,,e,····, ..
"
. .. ... . "" . .,e, ,, ,,.;;
"" ·a,·····;;g------·,s·----",,.a·,---···;-u----··,;;··--
. . .. . . .. "" · · · = ,,.a
. ,,, • . .. ... . .. .... . ":.,..e' '""'" .. ;-
.......
........... .. ,. ••••. "" ,... .•
. .. ".. .. . .. • """" .. ...... ... d·•· .. .. ..
.• . .... .
. . . . . . . .. .. . . . . . . . . . . . . . . . . . . . . .. . . .. . . ... . . . . . . . . . . . . . . . . . . . .
.... ,......,.,. ,. . .. . .. .. .
. . . ,. . . . . .. ... . . . .. ". """'. . . . .. . . . .. . ' . ······ . . . .
.
... ..:: .::. ,.,..
.: :.~:::::::;:::
. . . . . . :~;::
,. .
. ..: .: .: ; : .; : ;:. .::. : : :.:::;:::::
.
" '
. .. .
. . . . ,...
. ::..; ; :;..: : : :: . ; ; : : : :; : ; : ; ; ; ; : :.; :
"
. "
, :: : :: : : : : : ;;, : :·:·~:: :: : : : : : : : : : : : :::~: :: ::.;
. . 0.02 7~ 16 7.2 0.992 1.0
. ; ;: : : : : : . : : :; : ; : : : : : :;: "l-.;; : ; : ;: :; : .: ; .: . :-::t. ; : : : :::::,::::::::
.. ................................................................
: ..; ... .:». : : : : : '"................
: : : : : : : : : : : : :·::: :.: : :~:·.::::;':';;::: . : :·.·:(::~:·::: ::::;: ::::: : : : : :::
.... .. .. . . . . ·1·. . .. . .
···································· ..·1··· ' ·:.~ ················································································································································· .
. . . . . . . . .
..··········•············ . . . . . . . . . .
, ......•.................................... . . . . . . . . . . . , , .. , , , •:
"11··· ·.. .
: -: : :- : . : : -: : : : : : : : : : : : : :; : :-: :- : :; :;
. ,.,
o~os
04•••

·, ···· ············• ········a·· ····.···· ····· m······· ·············· ·· ······a·


,•'I•/• ••••

·. 0.954
.

,1:i . . ····, -tir , ·;e·, ::DI- uc :; :,

:n·· .;::;;::
··············· ..
::::;:: :::

:::::(;:::
.. . •·. . :e· = ·:-
.. :
•-
. .. ;;1;, .. . . . . , e · · :;:: ::
.. . . . ·,: . ····· .. . " , ,
. . . . -~....... .
r·, ':DJ(/ !\.!i:ii'ii'ii\ . 17.4 · 18.0 1.0
r~ ~; In;
"~"

!. i. .~}t?: ~~~fl i~;:::: i j:::; ;\~ i::: ~l 1:;::: :I}:: i 1:::; ~ :::::.::;:;:\ t:;) i·i; 1::; 1;: :;·~~·::. . .f::. . .:1::. . . d~::~}H~;:;:::~\~;; J:;: l 1:: ! : 1 l:::; tl : : : ; :; : ; : : : ; ; : : : ;~:::;:) l :: :: 1~~~\:::~;;:: 1 I
., ,.. ,. ...... ..-. ...= .=. .... . ...... ...= .-.. =... =.. .-. .=. ..- .=.. =.. =... =... ..-. ..=. ..:.. .... .=. =.... :::. .. ... .=. =... =... .-. ...= .=.. =... ...:. ... =... =... =... =.. =. .:;::;. . .. ... =. =.. =.. =.. =.. ..:. .. =.. =. =======,,,,::,:I::,,·
.. .. ... . . .. .. . . . . . . ...... .. ... .... .. ... ... ... .. ... .. .. .. .. .. .. . .'..........
. . . . . . . . . . .. .... . ... .... .. ... ... ... ... .... ... .. . ... . .. .. ... ... ... ... ... .. .. ... . .......... .. .. .. ... . .. .. .. .. .. ... .. .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...
r- - :.
:, ,:~::::::::::::::::::::;:::::::::::::;:::::::::::::::::::;;:
, , • • • • ,::::::::" = • ·, • • • • • • • • ,, "' • • • ·; • • • • • • ,= = :::
: :::::::::::::·:::::;:;:::::::::·;;:::.::~:;:::::::::;::;::::.
······=a
.. ::, ..:.=n=::,::,"";
. :· .
:. . : :
,,;;,1-
::::. · ..
, , ; = ="::·.,, = = :;;,~===
~:::: :: : :·
: ,= ''e" .==::··
: .:.
·•
.. :=====,=,, ,u·. - , , , , , , , =,,,,,, .: = .: = =; :,, ;1·:. - ·::::::::::,,,,,,,,,

:
.:
•·.
·o='-:"11t'
. . . u: , '''e·,
;~
. .,, , •::.::.;:::·~=~::::;:::::::::::::::::::::o;!:::::;::::::::
:
, , • ·, • • • ,, ,, ·" "' • • •:
;.: ::::;::::;:::::::;:::::::::::::::::::·:::;;::::::;:::-:
, , , :: , , =:: = = - ,, :: =,, = :; :: -: ;· -: , , , , , , , , == - = :::: = = = = =,,, =,, =
;;;;;,"""=""" • • • ·, • • • ·,,,,, ,:::::
=-·-·, :::..:
0.10 32.1 36.0 0.846 · 1.0
:. ;::::::::::::·::::::::::::::::::;~;;;:;;::::::::::·;;:::~;·
·········· .. -: .. .:. ~ :;:;; .." . :. ".: :.;;. ··.: ::-. ; ;: :::. . ··.:::::;;:;:;;;;:;::::;:·:;;.;::::::::I;:::::::::;:::::::;:.:::
, .
:. :. :. :. :. :. :. :. :. :. :. :. :. ::.. .: : ::. . :. :. :. :. :. :. :. :. :. ;. ;. ;. ~:. . ::. ..: ·::::::::::::::::: ;:;~: ;: :: :i:::::::::::::;:; :: : ~-: ;; : :~::;:::::::;:::::: :~::::::::::::::::::::: :: : : :: : : : : : : : : : : :: ::: : ::::: ~=: ::·:::::::: ~=.:::;;::::.::;::
.. . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . .. . . .. . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. . . . . . . . .. . . . . .. . . . . .. . . . . .. .. .. . . . . . . . . :·::::::::::: :·:::.:;::.
.
-: iH:Hn/::n:::;•
- - . - - - - - . - - - ,11:n:::..,it:=·- .,..",·11·..::::;e,=::t<n'
. .. =::==::,a"'
- . - . - . - - - - -'=- -:;:·- - - . ,. - - - . - . - . ·1· - . - - .
: : : : : : : : : : : : :: :: :=:~:~::::; ~;::: f;; :::;: :.: : : :.: : : : : : : : : : ~: :;; :; ~:: li~::::::::: :; ; ~::: :: ; : ; ! ;; =~:.:::::::: :·:;:::::::::::::::::::::::;::;:: : : :: :;: : :.: : ::; :: :: : : : ; ;: ;;~; :: ; : ~: ~=::;:::::~:: : : : : : : : : : : : ; : ; : : : ; : : : ;.;,; :; : - - - . - .
=::::'.::tim=:::,,_a, ::, :::-:1: ::::=::e,::",:rr;::;,,,1.
- . .:===:·=,?·:,. . .(tn:i:::n:n: :::):- :] • - .i(){
--- .
.... ... . .... .. . .. . .. . ..
'
............. . .. ....
......... .
: :::::::::::::::::·:.<:
. . . . . . . . . . . . . . . . :::::::::::::::::. ·:
.. .. ..
:. · ::::: : :
.. ....
·: ·:. :; ::: : ::::::
...
tu...................... .
. .
:e::~:::::::::::: ::::::::1:::::~;;::;::::::.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .
..
= .:
'

:. ::: ::
0 :· .:. ·
- .
. Assume a:loop has-a dead time element
'

: ··: ··: ::··········· ·····:·········:·


; : : : : : : : : : : : : : : : : : : : ::: ... ·:::::;::;:::.:; :::;:.: :;i:::::: :: : :: ::::::: :·:::: :: : : : : : : : : : :: : :: :::::::•: ::::::. ::;:::;::::: : : : : : : : : : : : : : : : : : : :·: ::~::::::::::: :.: : : : : : : : : : : : :.: ::.:::::: : ; : ; :: ~: :;:; :; ::::::::::::
···················,·······················:·:::::::::··:·::···:::•:::::··:··· :·····::·:••::::····························. ·":'·:·:···::······:··:·:

. . . . . __ or a first order lag element of 0. 05 times .


The concepts developed here can be used the natural period. The one contributes a
in reverse. If the step response is knoWil, it phase lag of 17.4° while the other
can be dissected into simple lag building contributes 1 So. The first has an
blocks. You will not be able to get as many · amplitude rartio of 0.954' while the . .

lag elements as there actually are. You second has an amplitude ratio · of 1. 0.
should be able to satisfy yourself that there From these . numbers it is· easy to see
are lag elements that could combine to be : that a ~ead time and a first order lag act
verymilchthe.same in a loop when the
the pure dead _time, the effective dead time, _ . . .

phase lag ·. being contributed at the


. . .

_ the long first· order lag, etc., which were natural period is less than 20 or 30 ·
observed. ·Indeed, this reverse -fit --,trying-- degrees.
to take a test result · and determine if its
components are reasonable · is a very
useful trouble-shooting technique.
f.f
'
r

.. ..

. .

Chapter 4, Lags and Gains, but Mostly'.:La:gS · 47

.. ' . .-:,·· .. :\ ,,. · · . .


. ·': .. . :. . ... . ... ·: .L .:-.: : /: . '. , ~

.. . . . . ... ·· -, :.... :· . ;., . .;:·· : . . i:. :, ·.: _·, ...: > :.::·_:· . . . .: .. : . ·. : . . .. }
. . .

I have. stated earlier that what the controller:


. . ~
confusing ;at be$.t,_.I will say a little more
gain. turns · out · to be is . largely .· irrelevant. ·. .
about it in the. math/algebra: paragraphs . . at
There is · another .gain; which· ·I·· call the the end of this section. . . · · . . .·.
process· gain. It bears. directly on what· the , ' ..

controller gain will be; but since· Lhave said A person not trained to separate. time from
the latter is not veij important, why. discuss amount is likely to use what I would
something that .. con~b~(es. to. it .. · It . is consider the . wrong word to describe
important, .for thescompleteness -of your
. . :. : . ..
something. He or she might say something
understanding, to realize.that.processes not responds faster when I · would say it
only have lags (of various,;:types),,. but also ..
responds more, or has more gain. Figure 4.8
gains. It is . an · important : .part . of your is an example of a first order lag. In each
vocabulary, the languag« of control. case the time constant~- is the same,but . one •
~ .· ..

. ~ . . . ·~ ·.
has .twice the. gain of the other. A person.
unschooled in the discipline of our language
of control might say the one withthe higher
gain .responds faster, -becaus~.·. the ; initial
slope is twice as fast as the other .. I would
say it responds . more. ·The· .dif.ference· .is
critical.
I've talked a lot about lags. A full . ..
.
. '
. . . .•

description of a component in a control loop 6 Un· 1·ts ..---


..___ - - -
·- ,. :~:::------
·.·. ....
· -·--·. · .··· . · ·
contains not only information about the time
63°/o -- -
parameter but something about the amount R·esponse to· .: · . ·
....:,
parameter. So· far. I have largely ignored . · A 1 unit ::Input ·step . ..
....a.
this. The amount parameter of any element . :I .
·o· ._ __,._ __...,_
.
.
.
···.
.
'
f
.,
.:
. ~
\
. .. .. : ·: . , ~- ' __
. . -~ . :

in a control loop is called gain, as it is for . .


.

:" ·• . · . ·· ··
. .

Time
the . controller. It will . also . have .: units, · .or . .
-. T •
dimensions, The units for- a valve might be . :
'
. , .

gpm/% valve . stroke. . The. units . for .· a :


..
·'

transmitter . · could· . : be % · . ·. transmitter


.. . .. ·. . . . . · 12 units -. .._ - - ....... - - -~::;,.------
- - ·
output/gpm, The units for a heat exchanger
might . be °C output/°C input. You. get · the .. '
,

idea. . · · · - .
•,

63%•-- - -
. .. ' .
. .
. .

This is easy to understand for elements that ..:, Response to .


are · self- regulating; ;It · is · not as easy to ....a.
:,
fl 1 unit Input step
understand. for an element that is not self-
. . ... . .. . . . . .
0
regulating .•. If the· · outflow · from . a tank. is
flow controlled, . the level responds to an . . Time· · ·
input flow change by integrating the . flow •-· T •
difference. It is not self- regulating. The
. . . . .

tank is stillconsidered to have a gain, but Figure 4.8. Two lags. may .differ only in gain. The
v the gain and what might be called the untrained person may say that the one with the
higher gain responds faster, when it responds ·
integrating time are all mixed together. It is more, not faster.

··,'
'

..>~: •
•.

~.
~:.
~
i
..
i ' ..
·... ·
48 Chapter 4 Lags and Gains, but:Mostly Lags
. . .. . ~ .. : ...

There is one point to be made about gain. It


.. .

..,:,
.

is made up oftwo parts, a steady state gain '


.' .
.
.
·.

and a dynamic gain. The steady-state gain ...:,


a. . .. ...~-

of a temperature transmitter · is easily 0


defined. But if you subject the temperature
sensor · to a sinusoidal change in
.
temperature, the output amplitude divided Time
by the input amplitude (its amplitude ·ratio)
will be· less. This ratio is the dynamic gain. Figure 4.9. If the open-loop· step response has an
Often, and perhaps nearly always, the ever . increasing slope, there are special control
considerations that do not follow the simple rules ..
dynamic gain is thought of in· terms of its· . \ .

fraction of the steady-state gain, · · · . .


1•

that will assure stability (Kc < 1 I Kp ). If it


. . . ..: ·; . ,. . " ..
. .. .
,•
·. . . .
isn't stable at that gain, then hardware
The steady-state gain is/·often. called zero- problems are a likely suspect. The same is
frequency gain and ' the 'dynamic 'gain a true of processes not having an integrator.
frequency-dependent gain. · These terms· are The "period-dependent gain of the first-
used 'because one: of the ways of studying order-type lags can never be-greater than the
control loops 'is called··frequency response period-independent or steady-state gain. If
analysis. Since 1· · am · trying to· teach what· I
. . .
this gain'. is used to calculate . a . controller
might loosely call time response analysis, gain, . it · will be on the conservative side,
and I· talk· about · the natural period, I will usually quite conservative. This is useful in
call · these gains steady-state or period- · trouble-shooting · flow control systems. The
independent, and period-dependent. · · period-independent gain is relatively easy to
calculate.
· Each element in a control loop can be ..

thought of as having. a gain at a particular ·· Another" time· when it is useful to· keep in
period. If these gains are multiplied together
. .
· mind that gain· has· -both . period-dependent
. . .

when a control loop is cycling,,the product


: . .
and period-independent components is when
will be· one, and· ·it will be dimensionless. process conditions change. Maybe a loop
This loop product is called ''loop gain.'' If
. . .
was well · tuned,·· ·but·.·. then· ··something
the loop gain is less than one at the natural changed.·. Only rarely do I end ·.up tuning
period, the loop will be nearly always be differently. for· different 'Conditions.. Usually
stable. The runaway process is an exception I tune for the worst case and let 'the rest go.
to this. rule. You could live your entire life If I did, or ·if I tuned and ·it turned out not to
without running into one of those. be for the worst case, then it is helpful to try
Generally, a runaway . system will have a to reason whether the change might have
step response with ever increasing slope, been ·in a. period-dependent or · period-
. .

such as figure 4.9. independent gain, or both. If it is reasoned,


for instance, that only the valve position has
This next point was briefly mentioned in changed .aad ·that process lags have not
chapter 2. The concept of period-dependent changed, ·. then· it would be necessary to
and period-independent components is not change only the controller's gain, since the
=
nearly as important as P ·4L, but you will period-dependent gain of a valve is not
find the concept useful. A typical example going tochange with position. Well, it won't
is on pure dead time processes for which the change if it has a positioner. If it is
gain does not vary with the period. For diaphragm operated · and ·has no positioner
those, it is usually possible to calculate what then the lag will increase as the volume in
Kp. is, .knowing instrument spans, etc -. Then the topworks increases. This writing will not
it is possible to calculate a controller gain cover that kind of detail. .
,.

. . .

. 49
. >
. . . . ··'.,':J. .. . . .

. . .
.·. . .
.. K . . . . .

". . . .
" . .. .· ·C:·:···.::"·' .... ·:. -: ..:... :.
. . . . .
· .. ·. . . . ; .. · . ..:11;-.·
;: ,
. .
-·m. ~ ·
-.gra t
""<·: ':- ...... .,. ....'. . . .'<: ·..;. .. , .
p:.. ·.. · ",,

. .
:·i· .: ,' · .. · ".•·...

. . . ..
. ... ..
'Of' .
. .
. .
. . ..
. .
,
. . .

. . .

. .
. . . . : \ .. ? .· ·. ..

..
; ·. . ·.. . . . . . . :·: . . . . . .
. . . .

. .
" . .. . . . . . : .
. .. . .. . : . ·:. . : .: . . ~- ...
. . . .

,. , . ..- . .. ~- . . ... . .. : ·:: .


.
. . .. ·.: . ,. .: . ·.;.·..
.
-,
. . .
.. :

.. . .·
. . ..
:. . ··. . .; :. :. ·.: . /· . -~- ; . ~- ~·· ·:> ·. : . : . . . . . ., . . . . ..... ·. . ·:-:· : . :.. ·::.·\.i ., . ..
. . . . . .

. . .. . . .. ..
·... -.' . ·... ·.' . . . . . -, . · .· ·. ··..• .: t"" ; -;;::.}'.:. ·.;. ·.~

. . . . .
. : . . .. .. . . . :; . .
. ~- . .
\;
. . . . . : ·. " ·: . -:~ >· · .. · ·: · ::·/·:· . . '-, :. ;· .; .:·:::·_.:: ·.< ·. :,.><· .. -: . · .. , · .. .- ;_:;:: ~- .
·: .
.·..
' -~
. .
..... : .":

: :,:.
; : .
,. .:
·.:
..
·· . . .
. ·:

. . .. . . . . ;-: ~ . : ...
. .· , .. : ..
. ·. .. ~ .. ::·:·~·.·· ..
.•. . .. . .
. ·.

.. ' , . .

. .· . ., ·' . .

.. . .. : . .. . : . .
. · .. ': :::· ... :·. ·~ ·:'.. ·: ...... :· ....... >: '. : ;,..·:.··. ··~·:;. . .....
" .
....., . .
. . .. . . .· . . . . -
... .· . .: . : . :· .. ··.. :· : ·.. ·: . . ... .. . ,:'· . . ·.
; . .
,.. .. \
. ..... · .

. . . .
. . . . ·: ... . . . . . . . :· . ..'· : . . . ·.·. . . . ·. . .. · . .
. .
:, :
·. . ...· . . . . : . . ·...· .:
:
. :: . . . . . . . . . .
··~·. . . . . . .
. . .

. . ;
. . . .
.. . . , .. . .
.. . . :

.. . ·.· .. ... ,... :.::-::.:


.
. . .. ·
. .
.. .. .
> ·. . ~ "
·= .. : ·-, :, ··...:·~· ~ ... ::... . ': .; .: : . ·:
.
·. . . .
: ~ :·
... ..
. . . . . ..
,
. ;-· : ..... .
"
.
. .
"

..
>
. .. . . . .
. . ,·• : . . .
. ···r . .;: . . ·. ·... ·...: :· .. ··i:/}~:· ·. ·. · ·· ,
. . ... : . :
... . .. ~.
.
... · . . .....
.
- ··:
• >
. .. .
. . .. .
. -; .' :...." _: .:. . .··:: .. -. ·. ,: ~.
. . . .. .. . .
:· . . ,

.: . ; : . . t{,:J:::; '.~f (~:~ ~~ .. \ ..


.
.. .
·: .,
. . .. .
.. . .x . .... . ; . .. . . . . :
. . ·, ...
•.

·. .
. : .. . . ' : . . ·.... . ..

·'

. , .: . .
. . . ·. .::: .

,. .,

. . .
.
.

..

. .
.. , . . . . ·. . '· .

·•. . ·•.., ..
51

.. ' . .
., •. . .. . : .' . .
.· :.-.:: ~~·:.; -~,:·;·::· :· ·., .. .. .., .;.
.
. . .: :. . . . . . . . ' . ·: ..
.. . ·:··.., .. :· . . .
...... • > ..

': ·,. . ·. : , . . ' . ·. . ' . . . ·. . ...· . :


. . . ... .. . . :· : . -~ . . : :. . . . .

. . ·.
. '. .... , .
' '.·' . .
;. . . . ,. ·..
·.
...
·{
. .
\: ...... ,
:.
.
.
:
. . ··;· ~;:· .. : . ::· ·.: ;.. ·. .
: .·.. ·: ...~~- .:: ._:. :- ; . . . ·..
.... ..
. . ... .
.
,
' ·:, .
.
:
..'
. . ..
. .
. . . .. .. . . . . . .. !
. '/t{J.. t:· ·. ·. · ...•.. ; .. •. : ·.:t:;t):[ . !.'.;-;:. .. . ' ...

..
:

'.
. '.· •.

. ' ..
·.: ....
. .
'
·. ...
. . .:· } . .'. _.:
' .
_

. ;....... . . . . . .. .": ;. . . ·.. .


. . .. ,..
.. . . .. .
. . . '
.. .. • .......·.: ·. . :~~ )::::.- :;~ .: -'. . r- ) ·_; .: : -:
. . .... ' .: . . ... ·. . .
.. .
.. ~
.
.
~-. .. . ..
' .·.. ,: .. .:. .. :
. ..
·:. . .
. : , : .: . . . .. .

. . . . ' .' ,,., :· ... ,·'

l ..•

·. .. .

,... :: ·"

'.

. . ..

.. ..
,
:
,.. ·.

. . . . . .. '.
. .
·. : . . . ' . . . . . . .. . .

This. section will present specific examples this dead time is usually only afew seconds;
of lags tohelp transcend the·. .gapfrom the but it is still a pure dead time. · Well almost. . .

general _concept _to the actual syste.m.. The Actually · the lag · from· brine entrance to
organization ., of material is: 'mixed ·between temperature measurement is more complex
type · of .· lag · and · .of •.• equipment . or tY.Pt : , than this. It has that pure dead time element,
process. · :·." · · · but then it also has more delay because of
the need. to change the. pipe temperature.
The pipe absorbs some of the energy which
·,
• ;. > . •, : .. .
,. . . :
would otherwise be physically transported. I
Probably all pure dead time lags . involve a am not aware of any quick estimating
distance and a, velocity. · Generally both technique . to determine how much lag this
might . add; ·. · it· · would · depend·.. · on • the
distance and velocity are well know, so the . . .

parameters .ofthe system. Generally·.. it. is


dead time is well known, The velocity may
small ·• .relative •.:'· to .·· .typieal · . temperature
change>. with process .eonditions, but a· new
deadtime can be calculated. 'One example measurement lags (to be.discussed later) .• · • •·
.. . .·. ·:. . . ·.
..
.
.
·.
, . . . .
-,
,
would be coating weight.on .a moving film.
Samplelines.for analyzers·,._such-as,for.lR or
Another would . be -the actual. weight -, after . .

casting or forming 'Jtlte, · sheet" i• ~ the _ first . chromatography, Introduce .a dead .time. If
place; such as in: polyester sheetmaterialor
the flow is turbulent, then there is ,.Jittle
in paper making. These dead times wndr~
.
·.. .
,::.. ldngifudillal mixing and. the dead time of
run tominutes, because- ofthe distanbes~~d~ . . . . ...
eomJ)Ositioj:t •: to::: ithe:' ,; :analyzer .is . ·Virtu.ally
velocities involved. · , Jt', · · · .· ·.. · · . , . · ·• · ~::
,

,.
·.
pure .. If the:7.flow is. iaminar,;;suchi;;:as:in. ·a
:- . : .. .:
sample line-taken off for a viscometer ' then·
. . . . .

·there ·is substantial longitudinal mixirig and


:.·
Another example is in fluid flow. Imagine a
..
the dead time is not pure. It has a pure dead
''

brine recirculating system around a jacketed


time component which is shorter than that
reactor. Colder brine is introduced in the
computed from average velocity because the
loop and frequently the circulating brine
r-
. velocity in the center of the pipe is higher
temperature is measured. The distance from
;

.. ·
1/
than the average velocity. Then the response
.';'..
•. the point of fresh brine entrance to the
.'
,:.;

.
> has a lag beyond that as the material along
i·.
lt..,
temperature measurement point creates a
~.,. the edge of the pipe comes along, but this is
~
••1,,

rt dead time. With a typical design velocity



t
~

f \
~
'
::.; · .
.··
;

.
'
·~

52 Chapter 5, Examples of Actual Lags

not pure dead time. Generally, in the connection with Figure 3.3, it may tum out
viscosity case, you won't be far wrong by that the lag is too long relative to the upsets
assuming pure dead time, calculated from expected. The point I want to make is that
distance and average velocity. The the dead time, per se, is no problem. It is the
consequences of this simplification depend speed and severity of the upsets relative to
on what other lags are in the loop. Typically .
the dead time that are important.
'. . . .

it would be a conservative assumption. _ . ·


.
. . ..,...

While on the subject of analyzers, what is


the effective lag of a chromatograph if used Most controllers are so fast that their lags
in a control loop? For illustration_:,:.~sume can be igllor~, even pneumatic controllers.
the unit has a cycle tun~ of l'Q,minu:is-. Thi~-. In ·-loops .op8fating · with natural periods of
introduces a minimum · effective dead time one second and less this may not be true,
of half that, or S minutes, But that is not .all, ·- but in· that 'region you will likely need to
To this must be. added the .time .between knowmore thanthesenotes will teach you
when the sample valve opens. and when· the . >''
. . . .

anyway. Most and possibly all instrument


peak being controlled gets measured and manufacturers "oan now give you the
held. If this were 3 minutes· · then the dynamic ··characteristics of a11· _ their
chromatograph should be _ considered as equipment. If a pneumatic controller is
adding aneffective dead time of·s minutes operatinga valve without a positioner, or is
to the loop dynamics. And this excludes any having . to fill a comparably large volume,
lag in the sampling line. The very· act of the controller might have a small lag. In all
sa.mpl-ing introduces other dynamic cases I can think of where there was interest
considerations but these are discussed under in dynamic performance, the valve was
Sampling Frequency and Loop· Performance equipped with. a· positioner (small volume
in chapter 9 . ·_ ·
for the controller and transmission line to
fill) so . the . controller's . lag was made
A quick word· about processes which have a irrelevantly short.
pure dead time as a dominant lag. Many
people who are not in · the field of control When_ I say that the lags in a controller can
think that if a process has "a lot of dead usually be ignored I mean the proportional-
time'' it is uncontrollable, ·Not so. It - is just action part of the controller. 'The integral
as controllable as any other process. Indeed, action is .in .itself a lag, but it is under our
an argument can be · made that it . is more control. . The derivative action· is the
controllable, since the natural period may be opposite, a lead, but it is also under· our
closer . to 2L than 4L. The view that control. In analyzing the lags in a complete
processes with pure dead time are hard to control system .I prefer to. think of the
control derives from the fact that frequently· controller as having no lags ( or leads), since
pure dead times are long . several minutes. their contribution is part of the tuning
Then in applying the concepts explained in considerations.

.' .
,.

'

Chapter 5, Examples of Actual Lags· 53


40.-----....--- ..........................
~------

. .
• •
;;
:

.

. : . .:.:
:
?,:. "' • ••
:-
-, ·:~.···· •• • •
~ . .
.~,..,r.-··· ~':~
..
301-------t---+--+--+--+-~+-1----- --4
~ ,c • • ; » ~~....~'::·:::?~~:. ::!~{1:~~~:;:~;::;:~~~;~~:?~~~::~
,:~· :S: , .,,1»:" O:•,c,I.· :·:·~· ~.;.::e,:,~··._~::_~,:_;~:inY/.1,',:,:<
: -r~ .. . . . ·:=-.~·~ ,..c..~w.!'~ :·~:~~
• ,, ··~ . '•, . ';,•,,: • ,• • s,~;;;P,;:,,.;,;~
;::,:::..l,:•,,:-,(0:;:!.~~~:· :.. ··:i,~
:;,. , -,-:-~•,.;,•,. ,,.; ~ s•
.. s • -.: • •...:, ,s,,,:·~·f,,;~

,,,
• •.
.

.
. :,
··~:·.;,;
~
. ••~,. ,_ s.., ·},c:l",;:
,.;,,, ..~ :;:, t-'·:$;~-i, :,c'~i:
·:;.
.
•,: "J''"
~~··:.•,/.•

:t; ;;'~:· ..:,!.;,


;'C;,;
..,1,,~·)'
::·;e.~-:
20 -----+---+---+--+--+-+-I-.._.._ ----
. .·
'•
.
/·,.,.:,: ·:.r,·,,.:; ··.: .;~.
:~ ~·. : ::...q;:i;:.:
. -. ·.1·:·ff:!!::~:..
. ~<;;
. ... ·., ',
. . '':
.
..
.
·.·.· · ..;~: · ..·.. ..
. fN..'1;;
. •""~···
. .
. : . .
.
..
. : ·.
.
.. .. .
. . . . /
•',.! ·.:.:~;··. ! .. , ... -::,.· .. · :·.
• .·:·.,.(•J,$ ~: ••• ;:::>,:::;··~ ,..~ .. ~: •• .. ,· .. . . . . ,•.•t. ·•·.
;
. '

t.: t >~:. . ~r .: : . ·. ·: '. ·. .·. . ~ ...


: ., . . :· ~ . ·. . ..: .,. ;.;. ~.:. . : . \·:\: ).: ·:, :..>. . .
·:l J.::· . . . : /
~:;~
.
. "~

Lags of transmission lines are ~':·fifMel~~;( : :·'~


· . . ·:·.·.~·1.,,.:,
. . . . ... ·.. . .· ..
. , ~ ·,¢'.''11·:<i••·~..:·,.,··.
~
·. , ; · r:: .., ,
.... "'''·. ·:: ....•, . . ,, ··~ .. ·r .

10
.
.·: :
. .
. .:;:,·
· . ......
' '·....
: · . . ..
.;-~,.,, ..•f,., »;
_ . .
..,....
. . . · .
....- . .....
·-·· .., . . . ..
. - . · . · , .
·,i·,..; ..,'> . .
...
' . . - __ ....
,...\' ....··,./:.·,to·>· .•.... '.·•:(• .. .:'..'.. . ' . ' ·.· . . ... .. ,. ···. •'
.,,... .. ; . . '• .
• .• .. ·.•; ·,...• ····~.. • .•.: . ...•, ·;· ··:·. ,.,,.
·. .. . . ..· ... , ... ' > · .. .
• ". - :
. ::·~ .. : . ·--e·• ~-··· , ..
· .. , · .. ,. ~,. ..•·.•. ·:
?,.'. •• ·.<., ·~ ·, .,; : , •
... ·.·.' ·.• :.;• ··· .. ·
• ;;_ • u ; • •. • • I
.· .. • .. • •

..
matically very_ complex, · depending not . . . . ,. . . . . .
. ·..' ..
·.

---+--·
,

....... . . . . . '. . . . . ,+--. . . .,.;..;.·. . __....


. .
~ ·
... · :.,· .; ·
!\ ,'.
.. , ;. ·~ ..
··. .'· ·..
i
.·. .
. ·j·•• :··:.~
• :"+·:
,:,··:,.
.,,: •
-~~·
.;,,;,.·
.;..c~.
~\1 ·.ii~.
.,:. ,,_·
,;_;-: -.,;
. .. ,
·,·.·.~:···r ... :.;..... •··;: .:.:,.:· . : .
; t··~·,.:· .. • .: .·: • .· • • ....

6
5 __
L...:.' ---·

___..__- ....___..._-+-. . . P1 .. · . . .
. ii,jiii,ii·· •• •. •• •
. •
... " • •••
. • •
... • • •. • :•

only on lengthand inside diameter, but 4 .. . . ~ : '. . . 'V , . · ' .- ·" . :


also . on absolute pressure and · tempera- ,,• 3 . .. . ···v.
. .• . t.• .
··v,· .·

·:•
·

:·• .•, .
···
. 4
.......,. .......... ___

ture· and on termination volume. If the ' .

termination volume -actually increases · 2-------- ........ -- ........ ...-~+++-+-------f


with pressure, as. With .: an unpositioned m V4" TUBING. L.--:-- rt 3/•".TUBING
valve operator, the lag is even longer. !J 1.0 ·' .....
· . . . . _.....
: ....
· . .-··. . . . ,....j. ii-_ ...._. __ ...,.
0.8 ·. - <. , • ·... ··. . / ·, .: ·'· - ·:. - . _·_ .._ ·. , . ,· • ,._ ;', -:·

You need molecules · to increase the


. 0.6 ........- ...•... 7;...,- +-- .........+--tl-+-~~-----1 .

pressure of the existing volume · and to o.s-----


0.4
. . _ . . . ..
__.......,_....._~l-loo+-+-+----
·; :: •• -. :.. • > ':' •

fill . the ... incrementally . new volume,


. 0~3 ..... --- ~--- ,. · ·-·: ....._......................', .......+-....... -.---- .

Realize that· the incrementally new vol-


1-' ..

~ ' ..
0.2 - -··-· · ..,__· --4 ........ ~- · _.__._ ................. _. __
ume needs molecules to· fill it from zero / '
. . '
. : . . :: . ·.
. '
..
..
'· . .

absolute pressure, notjust to change an V. . :


. ·: .
.
' . . . .. . ..

0.1._ ............._....................... _ . . .. . .

incremental .- amount. · . This can be a


100 · · 200 300 400 : reoo · 1,000 · 2,000
significant part of the - total · effective. .
Tubing Length, Feet.· · · ·
volume. Except for unpositioned valves,
. . .
. ':: ·. .

termination , · volumes · · of pneumatic · F.1gure 5• 1 • -·The Iag of : transm1ss10 · · n tu b"1ng · may be


equipment are generally small enough to approximated by a first order lag,,. .
be ignored. · .. · · ·
·.

I tend to simplify the transmission lags and


.
talk about an approximate equivalent first '
. . . .

order lag time· with· negligible termination The unpositioned valve ·will. not be treated
volume even though they are more complex. here. If a loop has questionable performance
Figure 5.1 shows this approximation of the and· 'it has: an- unpositioned valve, the first
transmission lag. · · ' .
step is to equip it with· a positioner and see
iftheperformance problem· disappears. This
If the ··larger tubing is , used, it might be is advice for those relatively isolated from
necessary .to ·use ·a _booster·• ·to· provide
, . .. ·' .. . . .
access·· to ''experts'' on control loop
enough air flow to' _fill • it. That~,.-aepends on performance. Depending on your internal
the air-flowing capabilities of the pneumatic bookkeeping methods, , it will often cost
instrument. This subject .
is beyond ·the . . .· . ·., .
. .
. . much more ..to have a specialist study the
intended scope of this c~verage.i.:lf the problem· than. to install a positioner. A
pneumatic. transmission lag_; is a significant specialist would likely suggest installing a
part· of the· total, lags in the system, and if it positioner anyway, before studying'rthe
is important to reduce. the lags, then simply problem further. -:·
consider using electronic transmission.
As for transmission lags,· lags· of positioned
valves tend to · be · complex. ; i • For · large
changes the system may ''velocitylirnit''; for
very small changes it may exhibit a· dead
band. Both of these are nonlinearities and
will be discussed under Nonlinearities in

.....
,,
C:...
Ii'·:..
54 Chapter 5, Examples of Actual Lags

chapter 8. In the mid range of changes, say


from 1 to I 0%, a valve lag may· vary from
.
.. . . . . ... . . .
.
0.2 to 1 or 2 seconds. The longer lags go . . ..

. . . .
Lags in temperature measurement can vary
with the larger valves normally. Treat it as a
from less than a second. to several minutes.
first order lag, or as two first order lags
Bare thermocouples in a flowing liquid will
whose sum is the chosen value. Positioner . . ..

have
.
time lags down . into the millisecond
manufacturers should . be able . to give ... . .

. range. Most industrial installations act more .


dynamic response information . about· their . . .
like two . first order lags than one. .This is
equipment on a particular valve, or a class ... · . ·~ . .

predominately due to the use . of a . well,


of valves. ..
though systems without a well are · often
-~. . .. '

better represented by two lags . than one,


Positioning systems can be made faster · · · · also. . . .
using a booster .. between (he pos·it:io.he.r
output arid the Valve: . operator. Ne\V-er . .. . .

For general insight, lags increase with the


positioners are 'better 'older. positioners, than diameter: of· the sensor and/or.: well. . The
.: . ·: .. . . ·. . . . .
but even the.. newest can be made faster with ~

: .. .
more 'thermal mass present (usually . metal),
boosters .. This technique is seldom needed, ~ . . . .

the 'slower .they are. Lower fluid velocities


so the details essential for good operation. . . . .. . . . ~

make the measurement slower to respond,


will not be reviewed here. The time lag: can . . . .. : . .

though, .not .. markedly· until. velocities -, are


be, reduced, even. on relatively. large valves,
quite low. Of course, lags in gas service are
to 0 . 2 seconds, and on small valves to 0.1 . . . .. . .

.. . . .
longer than in liquid service. One of the big
seconds. An . additional benefit, . and
concerns when using wells is whether there
sometimes the-principal reason for usinga ·. · · is good metal-to-metalcontact on the inside.
.
booster, is that it also reduces the; "bad . . . . . . . .

Sometimes a heat transfer· oil or paste is . .


effects from the nonlinearities of velocity
used to improve speed (and accuracy).
limiting and dead -~ band. . . . .. . .· ..
. . ~
. . . .. .
A . typical . thermocouple in. a typical well . . .
. '

will have a lag of perhaps half a minute to a


. .
. . . . .
.
. .. . . .
. ..
·.
. .
. .. ,~ ..
. .
few minutes. If the service. is in liquid, then . . . . . . . .

Transmitters, like ·controllers,. tend to be so


.. . . . . . . . .. . . ' . ..
it is often possible to set up an experiment . .: . .

fast that you . can ignore .. their· . lags. in the instrument shop . with identical . or . . . ~

Temperature measurement m~y be slow, but . .


similar equipment. The thermal system may
the transmitter itself is usually amply fast.
. . . . .
be suddenly immersed in a bucket of water
Some . transmitters have dampening .. or (hot or cold relative to room temperature)
adjustable dampening available or built in. and swirled . around. The time to respond . . ~ .

This adjustment is sometimes called a tilter 63% of the .way can . • be.. observed .. Actually . .. . . . .. . .. . . .

and is basically a first order lag. Consult the . observing the time to respond two thirds. of . . ~

instruction manual on the equipment being


. . . . . the way is close .enough, The lag may then
used, · and if jugged · necessary, call the be treated as afirst order lag or as two equal
. ·-~. . . . . . . .

vendor. Smart transmitters almost always . .. . . .. . .. . . . . .


first order. lags, each half the total. Use this . . . . . .

have a filtering function. If a large value _is . . . .· .. . ·.


information
. ... to understand how . much .this
: . .. . .· . . . . .

entered for this . time constant it will measurement lag may be affecting the loop's
adversely affect control performance. More
. '. .. . . .
performance, using the concepts developed
will be said about this in a later section. on
. . . . ~
around figure 4.6.
DampeningNoisy Measurements, in chapter
9. A lag in the measurement, of which tem-
perature is a typical example, has an addi-
tional impact on the dynamic performance
... ·~. .

.5·5.· .
Chapter 5, Examples of Actual bqs J .

. . . .
. ···. . . .

of·a·.lQ.9.p,}\::,;:;"~::~'·ienaJ,1/tbt\t···is. to tb~2expeeted .· . · ·. • ·.:J;~s.-;sucl!~\::lt. ·wo.µI~::~'.~: exactly fir-st;i'.~rder·· .. if


degradation Of p.atufal period. It deceives · ·. · ·. flow through th.e' exit restriction were
b
. e,o ~erver ti\~~;: . +: . · · .• . ... . . th t · · -&.
~; .> , ,~'.... ~: .P~i. tQf~atlc~ · .. ··.
. . 1 · . . • (fl . . 1 · .
~m•nat ,;ow ill~.ar~ , . ea . . . . · .; . ith h d)
th . . . . .... ::•(: . . · , ... ·»tr~·.,=A-..·,1.~~t...
r ,:
·:.·~·. :•·.·~.· ·~.. · · .:e .,... · ·,. · · · · • ·:·,' : · ··, .. : ·· · · ·· · • · · ·;· · · ,,· ·: ·· .. ;· ·.. ·· : · ' · · .:· : ~.: ;.•: · .. '~·- ;:. -: · · . .,.. · ~~.. ; ·. . . .. . . . . . . ,· , . .

is.better than it. is, Tfiit::·o/.; . : ~~~::¥.ai:i3tiQMS . . . · .... ; : ;.,:::;·:'•• ~'-;,j• .... ; :,,:,; .· . . .. . ·.· . . ...
are .· alway!
. . larger
. . ' : thaft
. ' • : .:;_;;;· :'.
? ,f;(;A::;~·;,
·- i · .. Jt.JS'.~it.-'tt~a!Jh~.~tual
,...;,. .
. ·. . : . .
volume. ill . .. . . .. :.~ ,. . . . :. . . . .. . . . . . . . .. . . : .. . .. . . .. . ·. . . . . . . ..
.
Sometimes this 1s not impertant; sometimes · .: · · . '*-he •:~1~:·-•rt~~v~1,11~ Wha.i.-•~;-t~le~~t ts the
.. . . t ·... .. . ·:. .. . . . ... ·.·' .·.·. . : . . .; . .... ": .. ·.,:· ,.. 'r,.. ·.. · ·:' ·.:.;...·t,~>:·:.;t,~ .::io':,· j• ..... ·'!;·__,:, ,·: •. ::-··! ...• ::.:1,, : ... ~ ... : ...:· .· ......... , ;.. ·;.··."'·.·..:· :.:,.,. .:. : ". . .. ·" .· .:. ". .. . ....

ii is~-~ P~a$.i~~lly . prof 011nd ,. iJtipr9.yem~nts · ;· · · · ch@g~_, : i~fS¥6..lit111.~ '.~fctrJ;;1ft.:'/: ~:~_~~ '. :_ >. · . ~-'1~--- f.19\\T.
in · • ·~ntiPf" ··.p¢Qrn\ance . . can ~ .: achieved ThiS ~PP~iich:•ppli~S, ~¥~~ if.the 1eve1 is On
tltr<>~gb.~ faster measurement, •·. ··•· • · propOrti9~i . 110#l)'. ~~trol ~/ ..is much
. • . . . .. . 'f simpler fh@:D : the, i ~Q~e ; : . fundamental
•. ' . . . . .

derivation. : ,·: .·...;, . :·: ..~:·r· :' ·.·.:,:~:s--


·.
: ·.•. '
i \ .. :: 7 ·~:· ,. . . '. . .. . • • --:. •.=.
<,· . . .
·. :·. . .· .. :·.. ·. .
:
·.: .' :'.·
;
.:· . .
:
.
, .·. . . . ....
.. . . .. : '. : .. :. . .. , ,·..
..:··.,
: .' . ·. :· . . ·~· . . . . : ..
. .. ·. -· ~ ..:
. . . .

consider the systems ln fi~.5:;:2;~


. ;(• .•. > :. . •.• . . . . . . . . . . . ..· . ·.. .. .
. . >. . .: . : .(:.. ;: .. ·. t . . . •. · . ~· .: ·: :· .· . : . ,:. ·~ ;: ·. :~ .·,: . . . . ·. . . \ . . . :. ~·
.
.i •

Assume · that each ·tank:·.·~.·is: holdi*1g:: 200 In the system shown in figure 5.3 · the
gallons. and that ::the; 'COtitffow·: depends composition in the tank will follow the feed
·', linearly on head (is not:Ypttmped<out):· It is
.
composition with a first order lag calculated . . . . . . . . .. . ,·. . . ,
·.
.. . ··:
. .
..
. .~ ·::
. . . ·..
'... . .
; , .:
,
..
r: .. :\
. - from the hold-up volume and the flow · ·· • ,. ; : , •• '1

. _100 gpm· ·•. . · , · · · · .. , . 100 gpm::~;ft:'>. :i:.:<- -. · i · rate. Temperature would be a special
..
..
.... :
. .
. ·. \ . . . : . :, .
. .
. .
·
.
.
. .
. .
· · ... ): ; ,\.'_I :II<> ·./ ·:\ ·
. .
'
:
. i'
:.
··.
.
.
case of· composition. If the tank is
. . .

.: · ... ·. · :· • · • .: ·· :_<, ~":,) ·. . ;:


:·~·,·

.
...~·.· perfectly
.. stirred and the level control is . ••
· : .·. ·. • perfect, then the lag will be a perfect first
. :., . .
• · ·• order. Ja,g. To the extent the first two . . : . .. . .

• . · : · ·• .. ·• assumptions are only approximations, the


· • ·'.·. - la.g is; only an approximation. In a
" . i'':. ... . . ,:• ·; . . . . . . . . ;·. . . . . . jacke.tf.c:1;_:-Ve~seL the, temperature lag in
'. . r; r , : ... • • :,:.,:

': : . . . . ·. . the jacket . will be approximately the"


...... , ~-· : ·: .: ·-·...• /'.: , ::. . · . ·; . ·. ·. :(:./·\<: ·~V:plume-iyf~:bythetbro\Tghput, if well
;: :. . . . . . . . . :. .
. · .
. '. . .· . .
: · - : : ' . llliied;::
. . .. . . : . .
:1.t ' 1 1e>t well 'rriixed"} then the total'
: ; . • : . i• < . ·.....·. . . . . . ·~ •· ; . ' .· · •. . . . . . .-.
·· .,,, ,. - · lag:J)needs .to·M.,·be . d1v1ded:·- between an,
· · ·· .. ·· : ,.. · ·• •· •• • . ••• •• 1

. Figure 5.2. The. lag . Of ,a .tank is · not always its . volume . · appat'ent d~ time and a first order lag.
· divided by its throughput.· .· . · •. ·. . . . .. · ·. · ·. · · · · . ·
· -: · · • · · ·· · . _; · . If there is a chemical reaction taking
tempting to · say that the' lag is 200 place, all bets are off for ahy simple answer.
gallonsllOO · gpm or two milllite$> It· is also The differential equations represent the only
. wrong. This concept·applies to composition road I know of to the solution.
but not to volumetric flow. If the flow in is · . ... .. . .

increased to say 120 gpm, the level·~ill rise.•


until eventually the flow. out· is the. same.· ...
While.· the lag can be derived from the.·
. . . . .-,. LC
differential equations, most persons would
..
shy away from that approach. Happily it can
be determined a simpler way. Calculate
what the new level would be, based· on the
new flow. From this, calculate what the
change in inventory is. Assume this
Figure 5.3. The lag of a stirred tank to
calculation yields 10 gallons. The lag is then composition changes is its volume
10 gallons/ 20 gpm or 0.5 minutes. It would divided by its throughput.
be close enough to a first order lag to treat it
,,---

56 Chapter S, Example_,;:'.of Actual Lags

. -.

This is a rather broad-brush category meant While there are exceptions, contrblYloops on
to include everything but measurement lags. the same type of variable tend to have the . . •
Process lags, such as heat exchangers, are same natural periods. Flow loops Wi'II have
not only complex, but the variety of a natural period of 1 or 2 seconds if
equipment is broad. These factors combine transmission lags are short, going up to· 5 to
to discourage trying to condense the fieldto IO seconds for long pneumatic lines. Flow
typical numbers. Lags get longer 'as Ul -l-'Le··
should be back to normal, after an upset,
surface.;.to-volume. (mass) ratio . decreases, within a very short time, say 5 to 10 seconds
They cover the gamut from a few seconds to
- •
for the faster· loops. Pressure ·. Joops fall ID
hours. the same category. They will be fast. The
,;
•• •
":": ....·,.
gain on pressure loops frequently can be
much higher than on flow loops.
.'
Temperature loops tend to be much slower,
...

'· aS,i)'O\l' would expect after this dissertation


·. .,.

on: lags. Their natural periods will typically


fall in the range of, say, half a minute to

several minutes.
. There would be many
. . .. .

exceptions. Composition control loops tend


to be slower still, with natural periods of
. .
·,
several minutes to several hours.
..-:

·, .·. ·. .<
••

-.

--·

·.
••

. :." :': ; :·· /. ··: ":·


'.

. ..

. -:
. ,. ....

.
-
. . ._.,~~-,··.~~-
.... ..
. . . •·.

' .
. . :

. . ... ,,~-· ·.
· :;··:.!.:.~·.,••·i.t.·. . . ·. . . •.: . {-, -~ :: .
... .
... . -s, ·.. : ·· ~:..·.•... ~ r -. -~.:,::: • ". • . i
."

. : .
; ·,
, ; : . ,:...... .
. .
. ..•
.
.
. .
. . ,,

' ·. . . ' .
.. . .

. ..
.~ . " . :. . ,.

.
.. . ...

,- .. ·~:·· ·.,.· ..
. . . -~· ".: · ·.t -~· .. ... .· ··: .
. . ~ ·. .. .

- . .
. .. .; . .
' ·.
. .

Until now,
all~:· material· :, presented· has Secondary Primary
applied : to : a . ;single '. fodp, with no Contro er Controller.
. .

concerns • · about 1: associated , loops · and


TC Tc -, · . . . .

their· • tuning :arid/or performance. One


frequent situation is the· use of the output . . . . .

of one controller to manipulate ·the set


point of another controller. This is called
. . . . . . . .. .:

cascade . control. · Figure . 6.1 . is an Brine . ; .


. .. . :. - ·.
' ..• : ..... :· ( . . ",

example.
.•

Supply.
.
·-. : :i:Reactor ·: •· .. · ,·.
.. . ·. . . . :f . . .. .
~.... .
If .· yon . ·have• .aever . • been. ·. exposed to a . , . :· '

· Brine . . .. :· ·.. ~· . ., .

cascade control. system . . it is confusing at Exhau11··. . .: . . ,

. . . . . . . . . . . \
. . ·: i, .. ~ .... . ,,. ;,·
. ' .:
..• -:.. :· :.:
first. In the example, .· the reactor . ·. •' ~ . . .

temperature is the primary controller and Figu.-. 6.1.: :An example· of a cascade: control: system. The
the brine temperature is the secondary output of a· primary:·:~trotler adjusts the · setpoint of the
secondary·:controJler. : . ; : . . ; ·
controller. These have been called master ~ ~ . .

and· slave and sometimes, are· called outer · · There are· four' main reasons· for using
.

. .

loop "and ; inner·. loop; 1hese .latter 'terms cascade control:


!

derive from . tile concept sit.Pwi, ;in .Figure


. 6.2, which i~ a sig11a\:_flow)' ·aijJ.f3111. of a
cascade control sysiertt: ' . . . ·. \' ; . .; > : ·. . .· ...
..
:,:· .: .•:, .... : : ....
. ::: .. ·.·.: :li . ..- :
>.
' .
. .
. .
Setpoints

Process2· Process1
--... ._ Controller 1
__,+
Controller 2 .......... - (-Jacket) . ..

(Reactor)
+
·- -
Inner Loop

Outer Loop
.. '

Figure 6.2. The signal-flow diagram of a cascade control system shows the inner and outer loops~

..
~:
. . ·. . . . .
'. . ..
. . ..
... ..
..·. . .. . . ·, . ,·.. . . .

. . .

58 Chapter 6, Cascade Control


.: ·:~···:.,. :~
... : .,.,...
. ,.
'
·.
.
1. It will catch certain ·lo.ad changes sooner
. .
Again let'S i~ok at ·. the . example, Without
and correct for them quicker, so the effect
cascade control, the reacto~,~::temperatqre
on the primary variable is less. Control is
controller output will. simplymove the brine
·better.·.
valve. The brine temperature in the jacket
In the example, if the brine header pressure follows with some lag, depending on the
changes, due say to other users. of the brine, f9hltne . of brine ·. in . the cirdUlating . system
then the brine temperature controller Will ': and to some. extent C>9 the atnOunt of metal
know about it sooner and take corredive · closely linked thermally to if~ This can be
action. Without the cascade system, no relatively long. Now consider the cascade
knowledge of the upset due to brine header · · system. The reactor temperature controller
pressure would exist untilit:itad ld'f~tti.~. <,:,,, · Oyn>-ut wants a temperature, ttPt a Valve
reactor · -tem pera ... :~e··· : . : ; · • · · · ··'l"L.· 1•5· fi~;{~1~u::;·l>d/. :··-;';t:b::::·-' :(i. · · · · . oesition
. . . . · :~ :~~, ~:: ·~f:•·.< :f:.~~:·;:Y~ <. '.,.".::·'... @:,< .. C · •. · f:"."'" .
.The,
··" .. ·
·;second$1roc· . •
: . ·.· ... _..,...J-·::,• ··.
o
. : n~~' . 1•··· 1ir·:.;: ;~s
r:,'!"' 1;~~--~

substantially later, SO (ho:.dis~ ~Oltld, ·. ·:. ·• • . · ~r tltaf request by 'doin$ .ill it .ciin to Set
have been present for a much longer time. the ~emperature requested. The result is that
'< • > , · . ·. . .,. ,, , . , . -: • . the ~cket temperature changes much more.
2. It the secondary rapidly than it did ,to a simple valve position
1\#ill'it(ectivel~·1ttieaiiize
. variable- to .a change in setpoint from the cbqe. This result, . of course, . reduces the .
.. . lag the primary controller thinks/is in the
pr~at')' controller, and linearity is generally
..·.... · . . . p~imal')', lRop is reduced. Refet again to

controller output will result in . a defined , ·< change is Iess as the controller is tightly
change in. jacket temperature (because the . . . -; tuned, than it is when it is not.
brine temperature· controller will make it . · · ·. . . ·

/ 111 the •..• primary controller . output, .. goihg .· ' q ·. . . . pl~f ~~ .· Such ·~.· 'incorporating · limits ·in ·. the ·
. directly to . the brine valve, )'Vf lf ~ti in. ~ ii . · · seco~ set :point, · · .' -. · · · . • . i i • .: · · .· · •. • . · . ·• .
s-e ., '.;

undefined change . jn . jacket 'temperai.ire, . ' . . . . . . < ' . . . · ,•: .. ·. ·. · ..· . .. i '; . · ... · ' <: ...

depending on the valve' cbaracfuristie and: r i .End Of l}Oi1lts'.··.·•···.· •· . •· ·. ·,··. · · .· · ;r· '' .: .· ·
relative·
' . • . . . . 1· .1·· A . . simple
brine . pressures~.· .
.. . d single . ·. . . . ... . . · . · . · . · . . · . .....
· . · . · . , : ·. ""I"
__ ./ ' . . ·,;,' ·' ' · . .
oop, no case a e JS rea . y a casca e system·
.-._v,.,.~ 1-·.l'k».li
.;t:;:~;:;:~;:;:;~-:;:;:~::i:>,:::.:;::-i<~:«<=~q;{ .: X:)·>
,::.·'-. ;·..$.·.·':~ .·.;<,.·.• · ~--·:,,:.."'~,"'!: "~·:,;,;.;,;,•,•,i,•,•,•,•o.·,•, ,·.·.·.·.;.··· ;.,:.·,.~: ,•:j::~
. W{~ic.."i=:=i&i);.=.,i;..~5:=:-;_=.::.:~"N>::::::::::::::~::::::::::::;:;~:;~:::·~!:~::Y.··: ~ . ••. . . •:;~>,-~~-·~;.
. ;o:l).::::::;:::::-:::::::•;:::::::;:..:;~:: :::::~~:·::,I( . ••· . •• • • ~~-
~·:.:-.·~·:.. -: x "' ,...
~~tt~t · /~~~:Ji~:f;~·-··-.,~~~=~~~t:!:tr1tmmr11i~i~m~mmr:11:1:1~\m't~![f~~t ·'-'·" :::.~~-~f"!:tlf~ti:1tm~jt1lF.~ll~l~t~:~tr~ .~ · ,.. .. J:~t~~it~1l
.t:~. (~/(!){f(11111:tr~-ij:;J: .·.· ;1:! ·
::~s,~:;::::::::::::::::::;.:.;.;.;.;.;:;:;:;:;: .. ·• ·' . •• • :·:"<::::~:.;~::::::::::::::::::::,:::::::r,:;. ;,• ~;~,~~:,::,, ·' •• •R''}:·;;>~:·;;·~: .. ·.:,:;:<·' .·'.·:<·>>:·'.<·'""·'' •,:,:,:' ·:.·• ' , ,.• · •.-,I, .••,...:::,' :·::;;:::~:;.::;;;:·:·:·:·:·:·~ .
I. f . l t positioner
. d. . . . . ' . . . . .' .. . "t•
·
. . . . .... : . ·. . .
· .; •.
.

..
. . ::··., . .• · ' • . . .- .t.,

:; : : : ;: ;: ;: ;: ;: ,: ;: : : : ;: : : .: •: -:, ~: , .;,.; .; ,: ; . :; .; ., ~:; :; . :;.·.•.;,)'"· · ·: . ,


,. ,., .... ·.·.·... ·.·.·.,.·.·.·.··.·.·,··.·.e
,.,,, •• ,,,.•.·.· •' • • • .·.·.·,·.·.·.·.·.·.·.·•• •.·

:,,,,,,;,:,,,;,,;:.;:;;\i;'t,~J>;:~,;~:
t ~ ~
i,:/•:•:•:•:•:•:•::f:i:,,;.;,:,,;:;f,:;:~;,'1
•,
7,<1
.
·>~ ,. ,.
.: '
,,


1 . ;:·
, •,. ~
• ••


-:; .. : :~ !.*~
'
·
;,
, •·•
·.-.·. ,, •••,, ,,. '','.','•''

..... : •
,, ''

..
·;: ~~- : .
'o,;o,;
,:::,:;:·.
',: ,,,

.
.!,"

. ·;
.
,,,,• .•.·,'
,.;
·~·:o,·.·,·.·.·.•.:.:.·,
I'
-~~ ..•
• o,;~,.·,·,·.·.·/,•t•,u
·:,,,,;,;1··.' ·:,; .
,. :;;,,~'.
'O' .-,:,:,:••,.· • • • • •••••••••••••••
.cn;,•,_,..•oo.·.·.·.-...·.·.·.·.·
·~·-·(;,.,;-_;~,:~J¥J:;:;'
"':,·~i:3!}5:~:::=~;

1 a Va Ve . .
. . ls u S e d • Th e pos '
1 l On er . I\!::I:!i
·. .,. , ,,.. , ,. , ! :\:!:\j ! i ! ]JII111i I :1::!t:I{~i { i ~ Ii ] 1i t )J]Jf i;j(\::~,w,~f
8 , ., , ~: :· . ··~'.1t-,: .,,,.!*,, , ., , , •,: : : : : ,· ·,
m~!:fiilllt]l!i!!\)i!!!:/!!!1\II!\If Jt)ii![)J)):i:)III::::?!Iit)l/' \ [ti}ft.ilttl~~IJltl~tli.lJf{#:' % 1 •.
<•.v,·::••. w,.,.,.,,:, ... :::,, .... , .... : .... ,.:;:~ . =· .·. • ':, . / :l;Sj:l, ,, ::, .. ~ ,y_:·~i~~~


1 Ion con O er re · · · • · . :r ,
..
iii:,. ::::r:f~J .
:•'•,•'•'••••••:••••••,•,..,,•,:•;<:•;,<'~,~<::'>'•
:i:;;,•,::::.i,,:
,.,.,•:'.:>:•:•,•···'••:*::·•':
. , :; ; ;. !\, . ·': :, : \:
,,~·
~,,.,:,:~~,,~
.
,~:· • •
:· · .,
i: :~,?< . . · . . '(AJtt?~i~~~\r,-
:::,:,::::,::,!\ · ,:. ··.·:,. ., ":: :,: _,,: :: : ,: ""
,:
•,:,,::!:!,;;,
•. -~
·: > .. ·.
. 'j;j ,,A~?:<;ii::;J::::::~•:•,:•;,:,::·

:)
. , .,:;:::;·:·
~.,. -··· .. · · ··~::,;:;::,.:::::,:r,,, ·····::::::::::::~::~••f,:•••:•••'•'•:::•·:<·<:••S:;'•J«<::'< · ·.· .-:;,t··=·-~ :::::" s. ~~s,:s:•'M<·s;
::,.,i:~·;;,,••,,,. .
:;<·' .,..... ,: .. ·,~
r: !! [;
,:
:'•,
·<.·:;;;;.<;:•:;r.;.:,y,.·-:
.:;,;~:~:t~{,::::,t~-

ls rea l l a Os t tfi II -1(.;.::::::::::::: .;:::::::::::::::·:·~~:,:,: ;,:,:,:,:·:::::.:::.:·::: ::i,~::,; .;::::::::~ ..;.;:<.;·j·••••; .;••••••·::::;••••••• ~.... : : •·~· .,,:::,,,•• ,,:,:' •,:, .:::::~:::::
• • ....• <~. . . ..·:; . . .;:::::;:: •:. . ·::,, . ,,:::, . ·~~::?,. ·:. .·:=~::::::::••:=:-::::: :::.~Y. :W.:-:

.
:, •, .•, .•,· •,•,: •, : ,:
,:::::::::::::::;::::,:~ · :: : :, ,, , ·~ :,, .:, ,.
•, •. "y."
, <" .,.
,, .,
. 5~v· : ~··
·.•" .•.• ·.·.·.·.;,·.·.·.·.··.·., • ,,,;, ,:, ' ' I, •·•• ·:-':x·>. •r •• •• • lx!-~•
Z · · ,.
:,•°' :,;,·,·,•,•,·,•,•
·· •'·' ·::::::::::~: '' ,
• .·.-.· • .·:·:,:
;~:::~::;::::::~:::

er
:'-!-. •• ••

. . . •
n ..~~::,:::::::::::?:::::
..,..,.,.,.,.,.,.,.,, •. •,. ..., :• .. .;:::·.., ::• :;: ·•·:;: ~:•:• .: ,;.. .... . !::,. ..:,: ,:~: • ••. ..•••• ~·~ . :-. .. ·'..: ,,.
· :~·:(,: :,·: ~:; : :~ ~:(,: (.: -;~. · ·:- :
,:, ,, .,,,,,,,,,,,,.,, ,,: . . . ·..
1>,:::·:;':>,::::::::::>,:<•>'.•:••::<:•' · · · ·:::·•::•:::>;::::::::::::!:',\:::: ,:i.~Y,; ...
.•. . ·'
:·•<~:••<·>•· .
·.·
.'·.·:S·:·•·•·<:,>,«;, .·.
~::::;::::::,::·:·:::::::::::::::::::···:x· .. : ~z:::··:~~:.;::::.!-::~:;:::·:-:~~~=-·-~«z:.
,..

. . .
• . .. ·,
• ....
" .··:·:..· . .
...
~~==-~-= ~-·.· :· .
.
.,
, ; i~·~':~~-· -:-:-':=·<:~:: ••f·.·n,,; ::,. :ff.if...~:..-::,•:;.;,;,,,:,:,:·:·~::~·::: .. •
,>Y,<·. . • :, • '•:::::~.
. . :. '''l'
: : : ;,~ :., ~
-:.,,,
~: : : : : : ~,:.,~:;
··.· ·.·.·.·.-.·.·.·:"
;~:_;;;;,.:.•:•
..:, .:t··::((,:~·~·:::

ase 1• overcomes most y, anyway) the t~Jtt~~tir~~irtritt\lJt1ttffft,Jf/'·',.,:~1trt1!tEtt1~lt: . 1Jtli.. . ,,:·,,. ~~.' J~~;twtii1It1tii1tim~Jt1tttJ1r1f1tfu,:~.,~ ~irtt$t
.,.'"~~-.-,,:-:,-z-:-»·,;"·'
th 1 S C t
• ( I .
<:< .·:s·:·w . ,. ,· .. ,,.· ·.· "· ·.· ..... ,.,,..
\/!.:,f)· •. .;, ;: >,,) :• \
r:;;i~r.:.:::. . :. '\ .... • · t ·
.'d. •
:-:-:-·-·.·,-,·. ,.,
•••
, ,z~ ·:~,,..,.., ..
"··, -.·, ,t: ,· ~.... i;\\ : · .,...,,, . ,
~f . «~=~<-:···=~-:-,:-,:--1,:-~·""...-''k·' .;,:,,,.,.,.,.,,..,..,.,..
, ," ,,

: ·, \: , . : · .
·(''*''':'<::':':i<i ,. :-:-:,;;·:;:;;~~~~;
. r·m.t
w.· <·. ···;,,~,·: ~ . . •

• t • Of dead band du e tO PackIng . it·: :· -:·. :·.:}:/t ,»:;: .··:··';:•:..·.·,·:·)'·,.•.•:•;.;,,.. ;.,y~,:.,,,,,,,,


: i ;:j.~if~i· ·}fr.@l~Ii1:!:\]1~\!i!\)!::::triitlftti~~,- . ~-~;,;~":.!~~~}?1~,r?[,i?1Ji,\,,,}{)):;111Iii1\ll:J£,j)l{:(:r.{W:~:;,;,:J1)1f:\li
::<:'..,:::;:•'•,;:·:,; •:::•:·:,:,:,:,:,:.;.;,;.;.;:;:;:;:::;:;:;:;:;:;:;:;:;:;:;:;'.:'.·'.;:·'.•'.·•~»·.(,~· S:, ,,. , • ,,,, r,::x•:•.. '-. •• ~~· ·.;,. " ·· ~·'!,·.o.·~-·~

nonllnearlty :·. ·: ·: :· ::III!:!:!:::!:!i\Iiii


:· ·: ·: ·: :· :·: •: ·:•.;:·;:·: ·: ! ·::ft::::::::::~====~=:::.:::::~-~~~~
:·:::::::=:::::::::::.::::::::::;:::::::::::::;:::::::;:::::::::::::::::: ;:· .:; . :. ::; ·.~::~::::::~::::~~~~:::::::(~·~~:~.,::f¥f.··~·"'
:11· . ~i· . . ...
'-! •,
. ..
::~ ,;.:,
· ··=:=·· ~,
,, :'X'-"~
• ,•'
· .,....~:': ·... · ;. · · .~.:::·:·>~:.,·:···;i~=Jsi~~========·~:f=:::~·:=:J~~:~~:*:::::::::::*::::::::::::::::::~~";:::t
~
. ... . ·-,~,.-; .....,._.,.,,.,,..;.:,:·:·:·>:·:·:·:·:·:·:·:·:·:·:·:····:·:·.
• ···::t::::::;::~:::;:::::;:::::;:;:::::::::::::::::::::::::::::::::~
• rt r 1 :rn:11 :/:})it .. ::>.,., .,{.. J:m:m:,,. .,::. .,t.. , , . :.. , , , -·. ~ .' - . ~ ~ttJ?i!t: :::::?i:ri::m:m:m:mJr:
g lan d frle. Ion . . .. •. . . .. • . , .... .
=: ....,:; ~ ..
\!! ! : )::!:!:!:!:!:::::::::::::::,,,:,!,]{ii~:}:: :t: •'i(::,:,:::!!:!:!:!'!:!:!:!:!:!:!:!:(!:!:!:~~=*! • . >;,;~- ,,f;),,i:;; ,,,~..,~•..,:·. .. .. -:·.;.. -~,,,,.·.·:::::,;.;.,;;,,} •: :J,li!'\!!l\f!(:!:!'!:!:!:!:!:!:: •
. .

3. It <
'
.·. • / • ·.·. • •..
. . .
..,,
.
.·-,,
.
>;
. .._ . . . . . .. • . .
.t,· ·.·· ::- • • .::;.,
.
·~ • ·. ·. • ··• .·· • •; < ':• (· : ..,,C
· · ··
• ._ ':
·
.::• .... ••
· · ·· ·.• ·.-.·,·· .·.·.· ,
: ·.
·...-..,. v·
•• ••••••"••"""""
•.··>·············· ,•,•:·:·:·:···:·•,· , •,•,•,•,•,•,•,•,<•,•,•:•:•.•>'•'·'
-'•'•'•'•"..'•"•'•'•'•'•'•':"•'•:·•:·:·<•':,
··,···,.e"';· .. ·.·.-..,,.-.... -.-...·,.·.·,;;-' '·" ····· ·
··:..::,·:v~::;,;;:;,";:'.:;.;::«':·.·.... •

circumst~nces, shorten the n~fural [:period


.., . . . ·. ·. ·. . . · · ..
ot .,
· ·
· ·Th·· · · ere
· · · _. . · .-. · : . : : . : ·
ts · an important
;
d
the primary loop, If the seco*1,dary tOC>P hllS . ,.. . · ynamic . . .. . . · •
a lag in if that would also be in the pritrttuY ·.· cons1derat1on when contemplating whether
variable loop without cascade and if this ·. · • > . to ~se cascade control or not. The natural
lag would be of consequence i~ establishing · · · · p~rt~d · of · the secondary · loop should be
the apparent dead . time . : ···.n ·. ...th· e · ....· .primary
•. s1gn1ficantly shorter
. Of th· . . . .... ·1 . .than theI natural period
variable loop, then cascade control will be · · · e prim~' O<>p wou d be without
faster than no-cascade control. cascade. There ts no hard and fast rule on
,.
-·····--· .... - _,

Chapter 6, Cascade Control 59



the desired difference. A factor of ten ts
. , . . . . .. controller, like another lag, or combination
. .. . . ..

certainly acceptable,
.. A '
factor of three of lags. The amount of that lag depends on
becomes questionable and a factor of two tuning. In this sense, the tuning of the
becomes very questiortabl~~fJYfPr the latter secondary loop interacts with the tuning of
situation to be attractive, . , ; . . ,.'if;;;~.,
.:»

atiQhs·. the primary loop, but not vice versa.


:.:< ....

other than dynamic would have to dominate


the decision.
Always tune the secondary loop first. It then
becomes a lag in the primary loop. In
cascade control systems, we nearly always
tune the secondary for tight control. It
would rarely make ~l,ly .sense to have a
There is an argument for not using integral
action in a secondary controller. You have
learned that integral action complicates the
tuning somewhat but: also that there can be a
problem with integral windup. In earlier •
deliberately sluggish secondary loop, as
. . times, virtually all controllers were Pl, so to
then mat).y of the benefits of cascade control simplify maintenance and replacement,
would not accrue. Then. tune the primary these considerations often overrode
loop as you would any other loop. The thing performance considerations. Now, with
to remember is· that the amount of lag seen digital controllers having the capability of
by the primary loop depends on the tuning being either PI or simply P only, the pros
of the· secondary loop. Refer again to Figure and cons of using integral action for the
2.2,.toremind yourself that.the. response to a
. . . . secondary controller should be reviewed.

setpoint change simply looks, to the primary

,.
r,.
.

61

. :'

' ..

• •

.,

• •
Derivative action IS fairly tricky to use The following are the most important

. .

successfully. It IS very· helpful to have at application points about the use of
least some appreciation of how frequency derivative action:
response .analysis applies to closed loop
systems, but most people for whom this It will decrease the natural period of

booklet IS intended will not have that. So almost any loop and, therefore, improve

its performance, The improvement Ill
the natural period normally is modest,
being in the IO to 20% range, reaching
perhaps 30% in rare instances.
Normally better tuning of a PI controller
will yield far more improvement than
that. The effect of derivative can be
what can be said to be helpful, without visualized in the step response shown in
. . .

getting in too deeply? figure 7.1 • It is seen that if the normal


step response has some curvature before
The rule for setting derivative action is to the main response; the derivative can
. . '· . '. . . . .

set it equal to P0/8 (L/2), though some accentuate - this and in effect reduce the
~ .

recommend .the divisor to be six. The point apparent dead time. This in turn reduces
• •
IS that there IS not much room· for error.
••
Use too little derivative and it IS hardly Proportional plus derivative
worth ·the bother. Use too much and you applied to normal response
. ~

can create stability problems. The ratio of /


Ic
too much to too little is relatively small, 0
like perhaps 3 to I, or possibly even 2 to u, a.
G)
- Normal response
1 • It is not unusual with a pneumatic cc without derivative
controller to have the calibration of the
,
·,

derivative adjustment in error by a factor ,,


, Time
of two, so that makes it particularly
difficult to use properly. It can make • •
heroes out of the knowledgeable and bums ..

out of everyone else. That IS why it IS • • L,

seldom specified and often IS turned off Figure 7 .1. The effect of derivative action may be
visualized as reducing the effective dead time.
even if specified.
62 Chapter 7, Derivative Action
. .
the natural period which is not now its . amplified. no more thanthat gain (the typi-
''natural'' period. cally 6 or 10). The step response to the typi-
cal implementation of proportional plus
I.ti Derivative action can make a runaway derivative action is fairly easy to visualize
process stable. Once in _a great while and is shown in figure 7.2.
there will be a process that cannot be
controlled in manual. It can be likened . . :,. It.is a spike whose gain is.K, coupled with a . . .
.. .· . .

to trying to balance a long stick . c)n your 'i<; · time constant decay of gain K-1, and time t • •

finger. If you hold your finger>still .... constant T.


(open the loop of eyes, · · brain and
muscles) the stick ,will.fall. The system ' ...•.... • T~•I
.· .···
I
is called unstable, not S'b4pa~i:{ it cycle~. ·i:~., .
'. ..
·. I . . :~ .
but be~a;ttse· it will 11ot:::s,y 1ia6.le~ lf·tfi~ · : .· . - ~ - - -1- - .. _ . . _ ... . : ·,·: . .

loop is closed using the brain as a .. . . .


.. ..
I · t
.
· · ·. .
. · . . .

I 0.63 (K-1)
c~ntroller between eye. and fmg~r, !11e I . K K-1
'stick can be held upright. Derivative
.
o·····
~-
. . ....

action is not. essential to stabilize : ·. . · ·:;


. .. a: ......
runaway process, but it • can· · help, 1 . ;

. sometimes significantly. +-
,, : . ·.
...
. ..
: .
Time
·llJ··:For a batch process it can help "turn the .. ·;..
. . . ... ·.

corner" without overshoot (undershoot), .


· Figure 7 .2 •. The · step response of a proportional-plus-
: . . .

especially if the integral and· derivative deriv~ti'(~ .·_controller will ··.spike up -at first· and ·then
corners · (these corners· are· frequency
decay•t>ack.to steady state value. a
..

response lingojare exchanged. This will


·require that the norinal controller gain A mathematically ; pure derivative function
· be reduced by · the separation of the . . . .
. would respond · . to: a ··pure step. :as . a pulse,
cornets. There is more to this . than I . . . . . . infinitely · · high, · · infinitely short, · · but
· · want to get into here.. ·. ·· · : · ·: · · · .. enclosing an area equal . to or proportional to
. . .. . .
. ... .
the derivative time. That is a rather abstract
ltf For a batch process, I used it once to concept. Actual implementation of the . .

· compensate · for · a long pneumatic ;·

derivative · function can · be · visualized as


transmission lag. A vesselwas filling or trying to get this area under the response

emptying, so the signal to · close the curve without going to · infinite amplitude .
valve was based on level and its The lower the amplitude, the slower must be
derivative. the· ·decay to get· ·the· same- area. If the

amplitude is· too low, the decay is· too· long


I.ti It has been used to measure flow as the · and the desired benefit. from • derivative is
derivative of level ( or weight).
... .. .· . .
not · achieved. In the .above example the . . .
..•
larger the K'the smaller the T for- the same
··I.ti It · is the lead in a lead/lag computer derivative 'time, · ·
. .. .. ...
element. . ,. . . .
,;

. ., . .. . .,... . ., .....
.. The sampling nature of digital computers
Derivative action is never mathematically presents special problems, especially for
pure, as this produces a very noisy control- getting meaningful derivative · action for
ler output. So practical implementation of derivative · times.' short relative to the
derivative action limits the high frequency . .
sampling. rate. Let's assume for. illustration
gain (short period gain), values of 6 or 10 · · that it is desired to use derivative action on
being typical. This means that the noise is · a loop having a natural period of 16
..
rr-
1 ,..
' '

''
::..

63 ..

.'
;•
,,
.,
Chapter 7, Derivative Action .
..
..
. . ..
· ...... ·'. ·.~_ .... ·:;:.
. . . seco.:· nds • .··. .:iI11.1tta wou ld call for a derivativ e ·
·....
.· tl)J]\]:):I::;J!:~''''''~:;;~,~~~:-~··; . . .... ·,,'.:;~itt'=<:'.J{ .· . ;;[$."'{"i;{)f''lr:,:;\:N'\ ... ::·>·>::::::;:,:i:?:~:~=l('\:J(ifl&@Y.fl@::

·
· · .· .• .·. · -./' ·. · ·, • :< · >...·; · 1•
· .·. ·. . . . .
·. · · .
: .
· · . • .· · :
·
· · .
.
:~fi,,,;,;,,,,;,,~~t,,-t,;:,:;:;,,,,~,:·:,,,,,'
~~'}:i::::~1 1::1:1~1:::r;.. J, ','
1 ;. ,.' .•, .. '.: ~~~
~f .;,;,r_ ,;,,;,;, ::,'i,,,. .r,,Ji· .,*i,,. ,'
:;· .. · /'.:::.· ·:'.' .... ·.·. . , : :·:···: ,,.
./ ,,,:r,. .. ,=$;,.· :;,:;f,=,=~r:::~r,,tr;,,,;;:,;,;,,,,';,:
: ':~: l~m*-~w~i:;t~1:: ~l: i . .
+., 70 sec Oft.OS: . -. ~. If'th e de· 1• vative ga· n S
,;- . .,t:::, J:.. ..-::- ·

.
tim .
e o f
. . : l. VY 1• 1• . . ·. ~. . : ...~·~:· . ·. _..... :
'f.
.. . . . : . . .
. :):~~!:~Jti1r·.1t:(:1:t1:1:1:1:1:1:1:1:mim1mm~~i:f;~;fi1~~1f\:;11~t111ir:1:111111111~11imm~~~~W~it=Jtttim1mi~m11~1:11§;lff;~;~ttit1~ttt11it{illllllllllll:l1llk~1.~1l~~l\1\i;l1(ffli~I1\f:Jt~t[fmlt
.

1 O, then the cortfi.~;.Ji!!!~~ needs to know The Laplace transform for derivative
about variations 10 tiift''.'-]t:::: ;: , :t ~1,!'~ 1!'A ;i . < .· · action was pre$ented in chapter 1. It is ·
seconds (believe it), or 0.2 :·:;s .. ··:, ~;.;l:>:k:;.::~:,iJ1)~!1:\.:.··~ :;.. : ·:'. ·. p~~ttnt-.:flgain here. . . ,,~'.j
.•~ ..

sampling system to not be confused about·:::' ·~:;'; :<j:'~~~·??:.:' . ~' :: .. ·, . ; •··. ·. -:. : • . . . ..
what is going.e _fi;every·.,·o
:: . . 2· s·econ:-d; :s
' • : • . . . . .,: · .· . it should : x;.:/:-~~r' .. ·:
. .•. . . . : . . ·,:. . :.}ti<;t· · · .; c · }::·::\ ·· · td·. ·· s·:+.tr• -:-: ; .. • ·
: > . :·< . : . .· \};;_:,,;;;//. >: ·.· . .

be sampling .<ti:v~ - or : ~-0 ::;(ime.s~~f:-.Et$····· . • fast, or ·: . . , : :- ·. .. : · · · .:,: . /./ ·. . Td . · . ·•1


. ,. . . . . . . . . . . :. ·. ' . . . . ·. . . .· :. . .: . s+ . . . i ,~-~: . .· .· .. · • . ·· .

every 0.04 to 0.02 · seconds (25 to 50 times ·.· ·. . .· · ~ .· .: · .· · · ·


per second), .· · There : . is: more about this . . . .. .
problem i11aJ~~'~tt()U:>daJf~·:~pling ·,Hefe· ~ •>·bit ~·''fl(gi,fil'a < The .
Freqriency · .· aid );\'~!\. Perfofuiifu~, in response'· ~::Jb. ·~. 7 2 ~$y be
. chaptijf .·tiiT:fti~ ~~s U) , ¢61t~~Jii that if considefed a$ '1/ia"fle u't,'t>f ~ pa,ts, firSt . .·
the··•cPhtr<"tller is 5$nipting the f)rocess.10 ·.. · the responseofKd,.al'lrJ.thenthe t;Jecay·
times per second; }'oU should question the • . ·.: · back. The algebra of this ts: · · ·. · · ·
use of deriva~ve:action:fcir:·process<,s having ..
a natura.lpetiQ.d:.~hortertth~about'<a minute .....•
. . ·· .· / ·,'·"·,/;:,:-
.
'.:. '..·.::;:· · .. .. _ ' ·.
~···. . ../ . .
:

· Manufacturers of itigital controllers do not


divulge. how they have · tried· to solve this
problem. . Jt ·is;·very.,:.difficutt·in the .field to If you do the algebra on these two terms
'assess the . benefit. that ·acJnies from the use . to . get one . term, you will· get. the . one
. .. .. . .
given~. Isn't . that. neat! Page A-4. shows
.

of derivative action, whether it is as good as .. . . ·. .. .

the Bode . . ·.plot~.·. (more frequefJCY


would be :expected or not. 'This is not meant response lingo) · for various . values·· of
in any wayto ;suggest you not use it, but ·. derivative!.gain.: Perso.ns .· experienced i.lJ
rather: to say that the results, particularly on that fiel.d.inayfind.these curves· a useful
. · relativelyfast processes; may .not quite Iive . .
reference~ I have. · · ··
.. , . up to expectations.Jt only costs your time to .. · .....
'. ': ' : : try ..~t. < . . : ; ' : . . .·.· . : . . . >' : • •· . : . ·.. .,.. .
. . \. .. .

. ' '.': ' . ·. ·. : _'.; ,/··, ./ ' . ' .' . . <. . . : .. :.


. (. . . , : ,. . . :·.... : .. : ·: .. '. . , . . . . ·. . . . . :
·.
.:
. .
. . .·
. .. ·. ;._,. . .... . .
.. . ,·~
:.
......
· ~ ...
. . ·> . .. :
~· ..·
.
.
.
!
' . .. . .

.. .. . .. · .. ;~· ·.··..~\ .· . . .
. · · .. : .. · ·. : .. . . .. . . .. .: :·.··, . . . ·. . . . ,

..
.. . . . . . .
.

. . .

·...
.
. :.. . . . . ,. : .
. .· : :;,. . . . ...:· ,.. . .. . .
. . ;. ·',{ ·. : ·.:. ·~ .
. '· . ·. . . ... . .
. .'··..
.' . ··•..
. ,,, . .

.......
. . . . ',
.-
. .
. ..
·: ·:·,,:>·~.. :':"· .~: .. '· ..... ·. .
,.
;:. . . '
.-
~ . .:. . . . . . .
., .·
' .
. ·.
. . ··: . . . . .:. . . . ·::· ..:1.. .. .
: • • • • , • • •<(
. •
.
• •• •

. ..
.. • •..
·. :
·:· .. <.... ,.·.:
·. . .
· .. .. .. : .. : .·~.:.: · ;.·
.. ·,. ·. ··.. ··.:,,.;:... .·::·. : ..
. . . . .
. . . . .
.. ·.,;.·_. ··. :·..'·, ..

.. . ·,. . ·: . . . .• . . . . . . .. ·:..

. .
. : .. . . .

··~...
\
.

~·.. . .. ,

.. . ... . .. ,
... . . . . ..
. . : . ·. . .. . . .
. .' ·. : .. : ... · .. ··
.
·: ·. . .. ·~ .: ..
r
•·
!

65

··' ,;,.

. ;·.


e
,.



Interactions and nonlinearities make control particularly helpful to think about what lS
• • •
loops interesting. Loop behavior 1n their happening to the open-loop step response in

presence does not follow the nice clean terms of R and L, and then interpret that in
patterns I have described. You have to know terms of what the interaction or nonlinearity

the nice clean patterns so you know what would mean to stability. I encourage you to
you could expect if it were not for the inter-
. . . .
do the same when you are faced with
actions· and nonlinearities. I have found it something that does not quite fit the norm.

Consider the system shown in figure 8.1 • of the other. The interaction in this example

IS particularly bad because the natural
periods of the two loops are essentially
PC FC identical. One simple helpful approach for

understanding interaction problems 18 to
reason what happens to the open loop step
response of the other loop when one loop is
tuned tightly. Then apply the concepts of
Figure 8.1. Tuning of interacting loops can be difficult. the Ziegler and Nichols open-loop tuning
rules to this open loop step response to see
whether the desired gain or integral might
A movement of the pressure control valve
have changed, based on Rand L (and RL) in
affects flow. A movement of the flow Figure 2.3.
control valve affects pressure. The two
control systems are said to interact. The There are several approaches to combating
tuning of one will affect the open loop interaction problems, and the following list
response of the other and, hence, the tuning
is not all-inclusive.
. .. . ····-··~....~-
J•, ·.--.-...•;,~ryt
. . ,. ·-::~S~); >,
. . . ..... ;;".'•.:· ... ~P. ..
- .
••
·.
,. '·
.
. . ... :., . ..·'. . ...... ),,. . .. . . .. .

66 Chapter 8, Interactions and Nonlinearities


. ·'

li:l' Decide whether one of the loops is su- there. would be .any problems with the two
perfluous. This does happen. For in- loops talking to each other. So once again it
stance in the example given, is the pres- is essential in solving contrcl ·problems to . .

sure loop really needed? know what the natural period is, what
determines it and what can be done to
0 Rearrange what valve is manipulated by change it.
what variable. This could require more . .'.·
·.... ~;. .
~· . .
.
.. .. .. .
.,...
insight than the novice might have. :i To repeat what has· already. been said, but
. . ·.. ..

with a different . emphasis interacting


li:l' Decide which variable is more systems can be tricky to tune. It is possible
important andtunethatloop . to.betight ....•••·. · to arrive atone set of tuningparameters for
In the example 9iv~n '.tpe,t:. flew loop • ·/Olle loop if the other loop is on manual, and
might be made the 'tight oritf i · ·. · · ·..... • · ·'. · · . · · a different set if the other loop is on auto-
.. · . · · · · · · .... · .: · . · ·. · ._ . • . •. · · matic and tightly tuned. It is possible for
0 If a computer-based. control system is . > · ·. one loop. to become unstable if the other
used, it might be practical to program or . . · · ·. loop js switched from automatic to manual,
configure decoupling terms to .reduee • ".· · . or vice versa. This section on interactions is
the · cross talk. Again, this may be too . . more to alert you io the potential problems
complicated for the novice to undertake. , · than to equip you to solve them. Basically,
unless the problem deserves a more elegant
Almost implicit, when · there is · a problem approach, I recommend tuning for the worst
with interactionisthat the natural periods of caseIthose.conditions that are most likely to
the .two loops are 'similar. If one loop has a produce cycling), and accepting the per-
natural period of one second and the other formance ·at other conditions. · . . · · .. ·
of one minute, · it is highly · unlikely that •


. . . ··.~ .

... . .

·,.
:_N •.: • .• • ~-

Until now, I have assumed a linear system. · · ·. . Nonlinearities come in qt~y forms, Some · . · . ·. . ·. · . •.
Mathematically this ·means that I· can de- · · · · can -be troublesome. Most, ,:\generally, are · · ·
scribe the system by aset of ordinary linear not. To the extent they.degrade loop· per-
differential equations. If you didn't know formance, the reduced performance is
what Iinear meant before, likely you still simply accepted. In most cases the reduction . ~ .

don't. It means that if a change is made in in performance would be hard to see.


the up direction, the response is exactly the . . . . . ~ . . .

same as if made down, except reversed of So that leaves a few you cannot ignore. At
course. It means that if you double a · · least if you know their characteristics, ·you
change, you · • get· . exactly twice what can decide whether to ignore them or try to
happened. before, We ate aware .that no -do something-to minimize their effect. I di-
physical. system· is linear; yet we ; live our vide nonlinearities into two categories,
lives. and. conduct our analyses · as if they · process . and hardware. Then for the hard-
. were .Iinear, It's so much easier that way, ware I divide those nonlinearities into .con- . . .

and so often is quite acceptable. tinuous and discontinuous, the distinction . . , .

becoming . more · apparent as you. read


further.
s,
-,
;


..

t
r
r.
Chapter. 8, Interactions and Nonlidearities. 67
'"
!
·.

dard tuning rules and procedures. Basically


you almost .have. to forget . all. you- Iearned
Process nonlinearitiC$ ~und~ Heat trans- about linear.system behavior, or at 'least re-
fer
. ' mass transfer
·. ' .
ch~fi~-;:::.r: ,_;:r!.:('.(These:
. ,,,.
. .·. <ti'~-;...,~,._,,
. ,·-. ~ .
are all . learn exceptions.
. . .
; : . .
· · · ·
nonlinear when examine :+(91~ \ - :~~::\"~t;t~t:};I:~: '.~: -,: -: _:: - . •. . •. . • · · .· .' · · : ... · · ·. · .: .· · · - · · -, · .:
< ' . .. •• . ,, .· • :, • . . . .. ..

:. ·
·.. . • ' . . • . .

changes. Usually they · are .· ~~tihtidilil; . ·. · · • •.· . ·. · . ,l!t>¥tlloeity limiting. · •· ·


though occasionally you may· have to deal <,; . .. • ·: · · .: · · . . C , · • .

with a discontinuous one. Generally; simple Velocity .• limiting usually · occurs in. a valve,
analysis is . done by .·assuming.· linearity at It Would occur . with . an. electrical motor . as
different operating' conditions and drawing the valve: opera1:o.t.~: It could. occur· with: a
conclusions frc,m: that .. The···· pH ,-:titration pneumaticvalvepositioner.vlt 'could also
curve is probablythe most severe continu- occur with a pneumatic. controller 'and no
ous process nonlinearity. 'Yet we use linear valve positioner .. to understand whether
concepts to solve pFI' eontrol problems. For velocity limiting is 'likely. to cause stability
important problems, computer simulation problems it is · necessary to understand the
techniques should be used. -It would be nice role lags play in the loop. · · · ·
to say more about process nonlinearities, but •):
that would make a whole book in itself. · · Recall from Chapter 2 and Figure 2.3 that
· · · ·· the gain is set proportional to 1/RL. · If the
product of R times L increases, then a lower
gain should be used. If RL increases and no . . .

change is made in gain, the loop may cycle. . . .

Remember from Figure 4.6 and the


Hardware. ( and . software) 'nonlinearities in associated discussion in Chapter 4 that L is
the continuous category .are usually that way the. result of. combining all the smaller lags
intentionally, to compensate for some other in the· loop, and the rate, R, is the response
nonlinearity. The goal is usually to make an of the longest lag in the loop.
overall system ·. · 1inear. · A. square root
extractor · on a.- differential pressure Now let's look at 'what velocity limiting
transmitter is a prime example. With the does. A valve positioner may be able to
flexibility . of· computer-based controllers make a 2% change with afirst order lag .of
there is a wide selection of algorithms that 0.3 seconds. For 8;: large 'change, .say 50%, it
Olay be applied to make an overall system may velocity limit;' So that if you tried to
linear, . or at least more nearly linear. make a first approximation to· a first order
Whether a continuous hardware nonlinearity . . .. · lag, the time, constant might 'be closer to S
is deliberate or not, the effect of it. · is .. · seconds: The•. concept is shown· 'in Figure
analyzed in a loop much the same as a ... 8.2 .:': - . . ._ . · ·. .· ·. · · .
process nonlinearity the system . Is . · ·. . . •. ; - ·
. .
·: .
.;- ·
,...

assumed to be linear .over a small. operating: . ,;. • · .. · ·. .· ... · - · . . .:


. . .. . \ . .

range, and extreme· conditions are examined · · ~


~··
~ ;" . .
.
. .

for potential problems. I . Respons• to large :change ' ,


c
0
Q.
- Velqclty limiting
=
a:
f Response to small change
. .

Now · comes the fun part. Discontinuous ·.


Time
nonlinearities can raise havoc. They can re-
Figure.8.2. Velocity limiting ·will make a sho~,lag look
verse what you would expect from the stan- like a long lag, and this can cause instability.
... ~ . .., ·":"· ·•:::;,\~

'

.68 Chapter 8, Interactions an.d Nonlinearities

packing · friction ..•· It can occur in other


This change in effective time constant· from situations . also. Many people . think that a
0.3 to 5 .seconds can make an otherwise valve positioner solves all )J.nlblems .: from
. . .~ . . .

stable loop unstable. If the 0.3 seconds was packing friction. It doesn't, ;'.t.J.iough of course
one of the smaller lags in the loop, it was it helps significantly. _With.·a positioner the
likely contributing to the apparent dead dynamic response. is. altered for .. .small
time, L. If the 5 seconds becomes a changes, The phenomenon usually occurs
significant part of the apparent . dead .time, . . . . : .: ..
for changes of less than 1 %. The controller
then L becomes larger, making . the
. . . _RL . . ,. .. .
output may ask for a 1/4_% or a 1% change,
product larger. If the loop had • been · tightly but the valve doesn't move. Or it· is slower
tuned before, .then it is· likelythe.loop will to move ; _ than normal, looking like. a. few
become . unstable, cycling at. a . ~.onstant · seconds' dead time, Figure 8.3_ shows this
amplitude, And the cycle ''will continue

. •
until . • N
·. •
. .
• •
- • ••
. . •
characterist1c,.·greatly simplified. Or it may
the gain is reduced- or the controller is put not- move and •then jerk, moving too far.
on manual. :_If the controller is put in manual This is not shown. , .. ..

to stop the cycle, and. then placed back in .


.
.
•. . . . .

automatic, then the loop will again be .stable This non-ideal . g_osition~ng usually . does not
until . some upset comes . through • _ large
:. .. .. . . .
create tuning or stlbility problems, though a
enough to
.,. .
· cause.
. .
the
. .
offensive .
velocity· . .
circumstance is described where ·it. can. A
limiting. .. . . . .
.
perfect positioner would move the valve -1 %
for a 1 % output change. The imperfect po-
sitioner might- only move it 1/2% or 1/4% .
. This is in the direction of reducing the gain
and does not normally cause problems, even
though · some small lags are introduced, as
shown in Figure 8.3.
. . .·
.. ..
. . . . .

In summary, velocity limiting may or . .


.;

may . not cause stability problems. If its _........_ . :1% change


lag for small upsets was . contributing to
the apparent dead· time, then. it is
possible -that _large upsets .. will cause . . . . . .
~----. ~..,--....,-... 1/P4_change
instability ... Switching. briefly to manual
will restore stability. If high performance
1/4% change
is not required, the gain may simply be
reduced enough to assure stability for the
worst case. If the velocity limiting comes Time
from a pneumatic valve positioner, then
one possible fix; if high performance is Figure 8~3. A. positioned valve does not- have a linear
required, is the use of a booster between response for small changes.
the positioner and the valve motor.
With digital control systems and magnified
2. Dead Band. 'displays it is possible to perform the tuning
procedure at -: quite small amplitudes, much
. . . .

Dead band- · is arguably the · largest smaller than with analog systems and non-
contributor · ··by an instrument item to magnified displays. If the tuning procedure
. . . .

problems with tuning a loop. It is the type of is done in manual; with very small
nonlinearity that results from valve stem controller output changes, then there is the
,.

Chapter-S, Interactions and Nonlinearities 69

chance that the gain between the controller


. .
small· amount of dead band (and .there can
output and the.cactual valve motion was not
. .. .
be some, i even with a positioned· .valve),
one, as is assumedcbut may be only 1/2 or there will always be a .small error. The
1/4. If this is the case;~th... eparameters R integral action .. keepsmoving the controller
and L (from Figure 2.3) are1:::--~,lf-·., . .,,., ••••• ...... ~:·:;:.~~· • .v-, :· ,· ••• ·!~::-··'.':' -~ . . . . : . :· .
output until; .the valve moves. The valve .
. ,· . .

determined, . and so settings . will · n6t ,·be · · . ~ · . IDPY~~, .too much, eventually reversing . the
accurately determined. ·. If .· the tuning : error, The. integn,;tirig,. ~t!~i-~S to work.
procedure is · done with . the controller in · on that error and eventually moves the valve .
automatic, then the gain established · for back, but again tOO far. Figure 8.4 shows the · . . .

small changes may be too . high for large phenomenon :<~nr aZ- f~st:_; process, such as · a •
changes, and the loop may become unstable. flow control .· loop, : Qn ·. a slower . loop the
Usually this problem, once recognized, is . .
presence of dead band · is not as · easily
simply solved by, . reducing-the gain. If'.high recognized, as the lag between the valve
fidelity is .· required .for ·small · changes, movement andthe process response.is more
consider . a positioner with higher. gain..
. .
obscured. . .
Sometimes a booster between the positioner
and the valve motor helps, it depends on the This problem has a .characteristic finger
air flow characteristics of the positioner. . · print. A small cycle will be seen. Sometimes
".
. : . ., the cycle may be so small it will not be seen
If the valve is not positioned the dead band in normal records, but its effects will be
is likely to occur, at larger values, like 5% or observed elsewhere in the process. Once a
10%··or more.i.The . . same phenomenonjust high-quality pressure transmitter on a boiler
described.: · may. · .· then · occur,. except steam header had a very small dead band.
'

everything happens atIarger amplitudes. If The cycle could barely be. seen on the
the valve is . not positioned; then a booster pressure record but the effects of the cycle
between the controller and · the valve isn't .
. .
permeated through many users of the steam.
going help, , and the first remedial step
would be to instal I· a positioner. . Decreasing the . gain, which is what is
normally done if a cycle is observed, does
There is another effect of dead band that is not . solve the problem. The cycle simply
probably the more important. The problem gets a longer period, If the period is
arises . when using control . with integral observed, normal tuningruleswould call for
action, or if the process integrates. With a increasing the integral time. If this is done
.. . .
. . ,.
Controll.er o~tput· ·· ·

. . . Actuet valve, motion, .


------···· ·,.'.,/·;·: .. .·. : ;: ·.:. . .:.·. . .

..
.
. .. . . . .. .
·. '· ·::.. :.
. ·. '.. . .

·~. . . . .. . . ' .
. · a. Cycle before gain or Integral time change '

Controller output
. . . . .: . .

----- Actual valve motion ~ . .

." .
.
......,.. ... '

b. Cycle after reduced gain or Increased. Integral time

Figure 8.4. Dead band in a valve often results in a small limit cycle, characterized
by a triangular wave in the controller output and a square wave in the actual valve .
motion. Reducing the gain or increasing the integral time increases the period but ·. ··
does not otherwise alter the cycle significantly.
.. . .. . ..... ~ __......

70 Chapter 8, Interactions and Nonlinearities


:,:
..
the period again gets longer. If this behavior . good way to check for dead band is to make :.t
.,"
't
. .
.
. :·

is not recognized . as a dead, band problem, two small changes, in manual, inthe same '
. :,,::'
·.·. ·:.
,.'
the controller settings will · be greatly direction, and then reverse - · direction · with ·.,.
...··'·~
·,:'
:.;:

reduced from optimum, and the tuner will the same two small changes, :1r· the process ~·:·:r
..'
...:·:;

have a high degree of frustration. does not repeat itself for the same outputs, .;
.:~.
. ";-
,
...
this is a good ·. indication that . significant ·.;Z

'·• :"$
:,
',
..'.

dead band·is present. · · ·. ..


:f
.;.
.'


',
"

If the controller is in automatic · then it is


harder to detect the presence of dead band,
but · if when you · make 'small · setpoint
changes, you observe a longer delay before
the . process starts to respond for ·. small
'
··A
j
·.)
··1,,

How do you recognize a problem resulting changes than for large, it is likely significant
;.
'.::-
. i
..
"..,;
'

from dead band? If the process is fast, then dead band is present ·::;
"
~',

..: .
the behavior illustrated in Figure 8.4 · is '
:;.
·,
.
possibly. the easiest way, particularly. if you If a fix is· necessary, perhaps the stem .'

·:
";
""
.'
have a monitor with . the ability to. magnify friction·.... can be reduced, . or a better '..
'
'

the · amplitude . and time scales. .: The positioner chosen, or a booster used · . ',
'

·,
'

controller output will tend to be triangular between · the · · · positioner · · and the valve •
,'


in shape, and the actual valve · motion will operator. The. booster· will not affect· the "

..'
tend to be a square wave. Realize· that this
. .
dead · band but it will improve , the small-
figure is idealized. What you see will not be amplitude. · dynamics, which · often is a
''

that tidy. If the . process· is slower, what you sufficient. improvement to not require any '
"

observe may look very little like Figure 8.4. more effort. Not all. positioners are created
But you will see at least the evidence of a equal in their ability to overcome dead zone.
small, relatively fixed amplitude cycle, Look .· for. one with a high gain. ·. The
whose period lengthens as gain is decreased manufacturer should be able. to supply you
or as integral time is increased, The period . with that information. A gain .of 50 .is too
will· be significantly longer ·than the natural low in my opinion .. The · dead band with a
period, which is approximately 4L, the L positioner -is essentially the ·dead band
coming from Figure 2.3. without a positioner · · ·divided by · the
.'
. .
.
. .,
.
positioner gain. · . ·
If you have a digital system with a monitor,
then the presence of dead band may be .
You cannot really be sure from· the control .

observed when making small step changes · .room whetherthevalve is moving or not. I
in manual. These are the types of changes .have used two methods. One is to simply
made in testing for process dynamics in place my fingers on the . valve stem, next to .

open-loop tests to determine tuning settings. the packing. You will need to set up some
With small changes you may see . no · communication between the control room
response, or the -, slope, - R, ·. will not be and the . valve, to know when the controller
_ proportional to the size of the step change, output has been changed, and by how much. ·
but rather will increase more than the size of With
.
digital systems,
.
you may connect a
the upset as the size is increased. Also you .: · · digital meter next to you in the field. The
may see a longer apparent dead time ·than · ·· · ·. · human senses··•· can· ., detect very small
you might expect. If the apparent dead time changes, even less than 0.001 inches, but
gets shorter as you increase the amplitude of · not . quantitatively .· and not · if they occur
. . . . .

the step, this is. a good . indicator that . you · slowly. So · this method · ma~ be used in a
. . . .. . .

probably have significant dead band. A · . pinch. If you observe dead band using this
,.

Chapter 8, Interactionsand Nonlinearities 71


. .

method; thffl:}'ou are sure it is there. If you limit, the positicner puts full supply .or full
cannot. detectitthis way, then· you are not vent to the valve motor. Thenwhen the
sure whether it'is:pfe~ln.t or. not. . controller signal, . co~es_ back . on scale, . the
. . · . ·,:.·.· '·•1:, Jii~~·'· .: . .·• · . '. t ·. . ). . p()sition~r outp~th.~ to change substantially
A better method is to. .~r·~ .. '· •
/ ! . •· .· ...•. J?emt<1.~ y~tY~ comesoff astop, This can
• ••.. f .,, •• f._ • .:/:" •• ~ •• ~· ••• ···i - ,._. ... ,.. . .... '" "• . . .... ,. . . . / . , : . . . ·. . .

micrometer to monitor the valve movement.··.· ·1··· . . · ... ·t~,;.~tttipg ·~~·•,top)\'()l"~S,. from. supply
A kit of clamps will be required.Jf there is a .·.·· ···. ·pfes$ill-e b@~ itj/ :()™' ....n! . pressure, ·. or
linkage between the positioner and the final ·. filluig .the ·;t9J>\\'.9l'ks' · from • atmospheric
flow · restriction, as : there . often· . is . with a pressure · · to · ()~~ting . pressure. · On · a. fast
butterfly valve, then you will ,.lso have to 'loop, this de,ta}' C!i!I, set UP a: limitcycle. ·. v : ••

consider. the possibility of dead band in the . · · · · ' · ''" · · · · · . · .; . · ', · . · · · · · ·


. linkages. It does happ~n. · .. ·. . · ... , .· . ·.· ·• ·•· · · · · .·. \ •. . · . : ·. · •· · · · . It is easy to gi~~: ~e ~d~ice to simply not let
. . . . .
1:
. , ,,: •., . :• , , .· ... ·. ·. the valve i get ii:i this predicament, That
Insummery, dea<lband.isproba\llythemost advice is easierto give than to follow, At
. common J}\St,ument ..cause why controller least now you Can recognize itas a potential
. · .tuning . efforts. de ·n~i look, like the· nice · problem. The cycle can bestopped by going
.:.. i sffl-Ooth curves p~~rtte<J; in 1;1t,:c:ar(icles on to manual momentarily. Sometimes it helps ·
. · ·. · tuning.Be aware-of the different ways it can ·. to reduce · the sup~ly pressure to the
. · contribute, especially for Small excursions. positioner (or colltroller) if the problem is at
Be particularly awore if you observe a small that end of the scale. Be careful if you are
amplitude · . · cycle whose period · . is limiting the controller output to avoid this
signJfi¢alltly longer than the loop's natural problem. It is not · so · critical if it is . at the
period. Quite possibly the smart valves now open end of the travel, but if it is at the
Coming on..the marketwill minimize the closed end you could end up not being able
problems discussed here, but there will still to close the valve. If you . are using a
be · non ..smart ·. valves . around . for · several program-your-own computer/controller you
years. ·. · · · ·. can . write a · program, using feedback. from
·, · the positioner output, to. · manage this
·.. . : . .

· 3- Valves at limits. problem.


.. Even without a .positioner-saturation
This section .. is not about integral windup. . . problem it is generally bad, news to have a .
· This section · is about problems that arise · . .
valve · hit· · a limit. It can upset all . the ·
when .a valve is operating near its· limit,
. . .
.

premises 011. which tuning was· based. So 'if


. .

usually and perhaps always its. closed limit,


. . you ate 'having performance problems, one
.
.and with apositioner. .v- ••
' . ·...·. .

-, of .. the ·.;' points· . to cover .in analyzing the


.

. . - ~ problem is to determine if the valve is near


If there . is . a valve positioner there is ::•P · its limit, p_a,rtic.ularlyits closed limit.
added lag when the valve hits . a step, $is . . . .. ·: : ··,...-.; ,. :· .. ' .. ':. . : .' . .. . .

can create problems. Once the valve hits a ~


. . . . .

. . .
. . . ·'.
. :: . . .
. :···, .
, . . .
.. . ...
..

·72 Chapter 8, Interactions and Nolllinearities .

· 4. Integral Willdup. . .. . . . .
. .
. .
. ··~
• own · limit, far beyond the valve limit, -The
. . . . . .
• controller has to ''unwind'' and ::.this .,takes
The phenomenon of integral (reset) windup time. Frequently this time~i:{§ff~:exbessive ' .. . . .

has beeri recognized probably since · the giving rise to poor performance, and
function was invented. I believe this was in
. .
potentially to instability .. · .·•,: ·
. . ..
the 1930's. Initially there was little that
;

could be done, and what was done · was .· · :To combat this problem, instrument manu-
awkward. In the:m-iddle period of·pnetlmatic • · facturers have offered . a variety of solutions,
controllers ( early 1:95()''$):::instrument iteftls. .generically called anti-windup protection.
were developed· to coJll~at · the problerii.: •· 'Most and probably all electronic andcom-
When electronic .. controllers came . along, puter controllers now on the . market either
more elegant and simpler methods ' were have the feature as standard, or offer it. as an
developed. · :Th.en . came . digital controllers option. Pneumatic controllers generally re-
and still better ways were
. .
f ound,
.
..
· .: . : ; -
. . . . ..
• quire· a marriage with external components.
.- .. ··. . . . ...
·. A cautionary note is this all anti-windup
, (.. :

Integral . windup
. ·•· is a phenomenon
·....
thati Call ·.·. ,::·;,;· .· ··. · · .·. . ·.. ·... ·· ··: ·.··.
feature~ do- not perform the· same way.
occur with a controller· havirig'i'reset qr . 'Some · may · require ' that the error · signal
integral action, which most 'do. If the valve . .: reverse·. sign. bef ore the valve · will begin to
goes to its limit, either futty closed or fully · move. Others may get ·a· valve "kick" from
open, the valve. is doing all .• it do .. 'The can <
·.
.
proportional : action. Many variations are
controller
. may
. .
. riot know this. If :it::.dfoesn't'. . .:.· . . :· .. ,.. .... .. . ·. . .. : . :. . . .
possible, .· · 1f you· have a repetitive . problem
the controller will continue to change the with · integral windup, it may pay to ·. study
output (windup) based on the integral. of the · the details· ; of ··how a particular ·. vendor
· error. ·.. This isn't the problem yet. .: The implements the protection. It is likely that
problem occurs when the. valve needs to · one vendor would be preferred· to another if
come back . on scale, for; at. that .tillle '.. the.
. . . .. . . . . . .. . .
the· windup circumstances tend to· ·be the
controller output. is likely to have gone to its same.
. . . ,. .

.· :·.. ... .
.... . ..,. .. :

. . .. .

, .
..

. ...
·.' . .. ..
. . .. .- ;...

. . .. . .... .
·.
·. . ..

..
73

.,

: . . . . ..
. . : : ::...,, ·.. .
-'
'.. ,
··.,.·,·····>,·
. ··~.
.
,. .
~ .....·,: •:.
.
..,
.. ..
·. •\...
. .
~.:··.·.
.
. . ..
.
...
.
:."
..
...


I

. . . . . . ..

The Ziegler and Nichols tuning rules were


. .
only the programmer knows for sure. If it is
based on the way most analog controllers PID then you should use the older algorithm
then and now are made. With the advent of to be sure the tuning rules apply. I under-
digital· controllers designers and program- stand there could be certain circumstances
mers werenot constrained by what could be
~ . . . . . .
where the newer algorithm might be pre-
economically built mechanically or elec-
. .
ferred, but I have not studied this possibil-
tronically. The ·result. is that · many, and ity. If the controller is PI (or PD, which is
perhaps most, manufacturers of digital con- not used very often) or if the controller is a
trollers offer more than one algorithm. In PLC, then I can't give you a general rule,
addition, many . · laboratory-type environ- since there are so many possibilities.
ments may be using a non-instrument- Controller
vendor computer, with the control algorithm Setpoint --Output
D .....____
written by their own people. It is not
unusual to find these have non-classical -
algorithms combined with reasoned and Controlled
... .

some intuitive modifications. ·In this ·Case Variable


the classical tuning rules must be used with
a. Series Algorithm
caution, as the programmer may not have
been knowledgeable about control, : loop
dynamics and the ·. algorithms tuning rules· Controller ..... p
have been based on. . Setpoint·
I
Various words get used to define the vari-
-
ations from the older algorithm. Sometimes
Controlled D
the older is called interacting, while vari- Variable
ations on a newer are called non-interacting.
Sometimes the older is called series and a. Parallel Algorithm
others called parallel. If the controller is
Figure 9.1. The difference between a series and a
just P or just I, then it doesn't matter, unless parallel algorithm is shown conceptually in this
the algorithm is home-grown, in which case signal-flow diagram.
- ·;· .- 3.,
...
., ":.
..
...'
,.
.':·..
J

74 . · . · . Chapter 9, Potpourri
. ~·-
.~ .
< ',
..••..
••
'
.
:;
~
'.
.'
.
·'.
-,
. · . called parallel ·. or non-interacting,· the "'
,,'
..
. ·''..
transfer function has this ger,eral form: . ·.

·.·'
·~,

. ..

The conceptual differences between the P+l+D .·=: ·



'
'

two basic forms for the algorithm are


shown in the· equations below. For the The Laplace transform for this becomes:
older or series or interacting. algorithm . :·
.
. .. . :
'
. . .. ,;. . . . ·. ..
.. . . .
the transfer function has the following
~
.
. . '. -·. . . .

: . · Tds · .. 1 .
general form: . · · · .. . . . .. Kc + + --==----
.. .. T;s rd s+1
P (1 + I) (1 + D) Kd
. ':' . . . ... ;;, .. ·. . . . . ..
. . . .' ~ ·. . . . . .

The Laplace trenstormior this is: _: - · ·. · · · .· · · · ·.·.· In prinCif)li it:·~hould be possible to get one
. . . . transfer. _:fu.noliori to essentially match the
\
..... ~ . .
: . . . ,

. . .·.,. .
other, by equating like terms. I have simply . . ~

. . always used the familiar algorithm unless I


. K. , .t> • Tds + 1 · •. · · . . . · am working with software written in-house,

· . ·· < ', · Kd
B'flebra to g~t it tn .ttle familta~ form. With .. . .•.•. . •· .': · • · • · ·• . · .· ·
. . . . . •. . . .· ..•. . ·..
.. :.·.····.··
. . . . -. - . - _ , . . . . · · sottw:are: ...,: written:.. : in-bouse ,t . may be
··.· · :~ ,·. ·. ··:... :.. , .. · .·· . ··.·: ·· ... ,·· '. '. . . . .

.. For the algorithms that beceme available·.·_


. . .
possible to persuade the programmer to .

with digital .. computers, w.hiqh nevebeen make the algorithm the series type. · ·
. .. . . .. . . . ~ ·~ . . . . .

.... .. :· ·~..· ·: . ·.
. . . . ,· : .
~ ~

Digital . controllers do . . . not . measure constraints, discussed elsewhere, .··apply.


continuously,· as do analog .controllers, but
. . . .
Many · ', commercial . controllers sample 10
rather do it periodically .at some frequency
. . . .
times per second, which should. be adequate
called the sampJi11g frequency or a time . . ·.· for. virtually. ·. all loops · in . · the process
called . the cycle · time. This . has
. .
two. . industries, though again it may not provide
potentially harmful effects on. performance, · · adequate derivative response for the faster
a
It introduces potentially undesirable lag in loops. · · ..
the system with the result . that digital . .. .

control is slower than analog control. The Suppose a device. had a. cycle time of one
other potentially harmful effect is that it second, If it measured . and . instantly
throws away information that might be · produced an · output . based ·.on . that
needed.
measurement, the effective dead time would
be. half that, or 0.5 seconds. If however, it
For the lag, . it· introduces an incremental measured, then calculated, then produced an :.
'
-
effective dead time of between . · half the
.
.
. .
.
. . .
:
.
.
.
output one second later (right before the ,
.,..

cycle time ·and· 1.5 times the cycle · time. next measurement) then the effective dead
..
·;
,.,
..

Applying the rules for performance that I time would be 1.5 seconds. The incremental
. . .
,,·'

have discussed before, a sampling period of . . . effective dead time added is thus half the
. ~~
·~
'
r
0.25 seconds (4 times per second) will be cycle time. plus the. full time between· the .',.
...,~-
,.·'
;i
adequate for most loops unless derivative . sample and the output based on that sample. ::·:
,...
·,:.
·....
..
action is to be applied. In this case other .,
• >

:1
..

..·<,
j
·'··:
·'
1
j
..,1-
.,•·'.
·,
·~
-;
..,,,·
. '.'

'
'
..'
.·.··..
-:
·.

'
. ·,~·
; '
• ·,l
. ·~
,

Chapter 9, Potpourri 75
· .. ·.. : -, :c. iiL '•. ', . '. AJ,ia~ed signal ..: . -:.". - . 60-Hz sign.al . . · . · · · .· .. _ . . . · .. ·... · .. · -: ·. ...·, ~ ... :··~.' . ·.. .
: · > · : · . <: ·. ';.;:/ :{; i:ftp·m samples
'". ,
.. ·. . ..
. . . , .. . . ,· : .. " . .

. ~ ..
.. .
~ .·~~
.. . .

. . . . , . : .· : . . . .. . .. .. . .··. ... ···i :. ·:·· .. <.: ... . . ~ ... :. :. . . . . ·.


,. . . . ..... ..
, . . .·. ·.. " . ·. : •. :()_;:> _'' :: ·'. r:· ·.. . •: -'~ .• ~ .. .. · ::.• .
·. .

. . ···:.···,.r,..
'
1.·• .f{~i· ~ , . , ,.. .., :•., .~,). . .. . ·. .. . . .. .. . . . . . . . . . .. . . . : .. . . . .• . ·. . . -~ . . .,, ~ ·. . . . .
. ·.~f~---~·~.:.~·.:1'.;li;~··.::.o~:;'.:.; .,._,._::;~ :' .· ·":.· · . · ·. ·.. · ·.. · .·· .. ·. .:· · . ·. ···.. · .· ·. · ·,: · · : . :·-,.~·· .. · . · ·:. ··. ·. :. ::: · ·:···.. : .:·.: ., · • .·
. . . . ' /

', : 'if}/::~;t:§1;;;:;:/i i:,2'. ;: . (i


: ~·· , . -~. ·.•, ..,. ,,. , ..,,, ,,'l:.. .• .. .•• .. . ., ." • .. •
>~:;r . •
. ·. .: .....· ., :·....
.. . ...
. . . ,. .. . '/" .
'. /
. · · ..·. ·· . '.· · ·., ···:::\··:.:'
·......• ··,.:.. ..· ·:~~:·...~,,,.;,"-.: · . ·:,:·...
~- -. ~~,...
;-1 ·.•,... . ·.. ·:. . x·. . · ; ... .. :,.·.,
:· · ·,·...~~-,,........
: ·· . . · .:. ,. ..
. .. . . . ·. ·.. • . ; '. '' ..' . .. i . ·. . • •. : . / .. t . . . . . . . .
. . --- .....,-+-. -_ -_----i-. __ .... o. .·.o-4--+----+---0--·.o o ....o~a _.__~: -.-. 0. . .1-- _-_.--ti....
1

t-+---_.-_ - _+-_ -o-~0. .2~- . '


....a~· · _ . .\_'. ('\.;·,--+
. . . •-.• :""""'"';'!"P.- · -. -+..... ,rrie(s) . · ··
.
. ,' ....
...
·,
~
...... ·.
~:
/.
\
. .

. . ..
·~·~
. '.,. . . . .· . . ..,, '

..
.· :.. .
.
.
. .·
...
' "'''
·..
. / /
/
. ..
-: .
.
:
.......:
.
. :
..
. : .~ ·: . ·: .

. ., . ..
. ·..... . . . . ·.:·. . . .·.. . ·.

- ' ......... _ .....


. . .. ',' .. · ::.. ·.. ~- ... , . .
. . ...
. ' . .. .,. . \ • t •• • • • '.'
J,;·. •••• •••• • • : :.:·
. , '
.' .~ .

: .

·: . .... ~. . . .
. . : . . • • • •
., . .
. . ~

· .. . ·.. . . , : ':. ••., : -: • i·· .:· ... :·::· •• .:... •• ." ·... •• • • • • • •••••

.
. . ,,. . : :.
..·::.. ". , . ·• > ..-':: .. -.
.

;. ·: ; . . . . 50-.Hz samples .

'. ·... " .·. . . . .. .
'
. .

Figure 9.2. ff a :cycle is. not sampled often enough it will look like a longer cycle, and: create havoc. ·· · .

. . .
: . ·. . . : ·.. ·. ' . . .
. ,·
. . . .

There· -;.is another: consideration if the cycle relatively fast processes, the use of
time is · long relative. . to the potential derivative action in adigital controller is not
variations !'. in . 'the' .process.. .whether these recommended. It probably is not doing what
variations ·be· noise or . -real, Imagine these . . .
it is supposed to do.
variations .: to . -· have · periodicity. . Then the
sampling system , should . sample several These notes are not intended to. .provide
times :(5: or 10,)::(luring-- one of these ·periods. adequate training in. deciding when· - the
If-lt .samples.rfewer than twice, then a sampling frequency is sufficient to· avoid
potentially- very: - serious·_·
-
- numerical-type : . .
aliasing, but rather are intended 011Iy to: alert
problem .called 'aliasing arises. Figure 9.1 you to . the fact . that sampling: · intr<:>d.u~s
. . . ~ .

demonstrates the problem. · problems .beyond - the -.• .· siro.ple .: ·,: lags
introduced. - · . . '. ; ·._ ·. •- , .: :· \ : : · · : : : ,· : · . · . ·, . • •
.. .. .. . ... ,.:... '·........ . ._. : .., . ·. '
• • ••
.. •• '.·· '. 'r .'
.. : ,.. :. ·. .. .. ·. . .
.
' .
.
.

If . a ; digital . controller is configured · to


duplicate the. (U.t.19tioqality of. an· analog. 1

controller, itis probably ·safe to say that it


only: . approaches but never quite reaches the
dynamic;·· : · performance · · . of -an · · analog . .

controller. In many cases, and perhaps most,


If a sine wave is sampled less than twiceper the difference is too small to see. The
cycle, a much slower cycle will appear to
. .

. . . . . . problem arises, if· it does at all, · when the


be present, and· this can really confuse a ._ sampling· rate is slowed down· to allow one
controller. In essence, all of the information . .. microprocessor to handle many loops. This
in the variations that aregoing · on too fast becomes an economic decision. ·. Does the
for the sampling rate to reconstruct properly '

economy achieved by using .one CPU


is · pushed over into longer periods. The ( central processing unit) cost anything in
results can· be disastrous .if there is reduced performance? It is seldom. possible
significant variability that is happening too to assign a dollar value . to · reduced
· fast for the sampling system to reconstruct. performance. My . suspicion . is that the
This is · · a ·. · special consideration when selection of most sampling rates is based on
providing · digital derivative action, as the what is available, and the performance falls
sampling rate has to be significantly higher out as a by-product ...
than the derivative time setting. For
. . ., ... ,-··-.~·~·~?'4(£,WJ.V . .

76 Chapter 9, Potpourri

Sometimes, sometimes. typically slow loops.: Frequently the source


of a disturbance is a mystery or cannot be
Sometimes the disturbance to a control ·. . .. . agreed upon,. and the measurement itself
loop is known, and sometimes it is not. · may be suspect. It would be very difficult to
Sometimes the effect of the disturbance is determine whether a loop is performing as
known and sometimes it is not. All of which well as it can without retuning. If the natural
makes an otherwise almost exact science · ·
. . . period is already known, then perhaps some
into something of an art. How is it possible judgment could be made using the concept
to look at some recording charts and decide presented in figure 3 .3 .
anything about ·anything?· Well, sometimes
it's easy and sometimes it's next to Between these two extremes is a lot of
impossible. . territory; · Experience can be a · . key
.
ingredient. . If. a control system has been
.

Let's take an. easy one · first. Consider a giving . excellent .performance and then it is
simple flow control loop. The· only. process poor, the, . temptation is . to look for a
reasons for flow changing are a change in coincident cause. If a control· system· has
either· .. upstream· or downstream pressure been drawing straight lines and then appears
(unless the flow is "critical," in which· case · · · to be drawing somewhat . Jess straight lines
only· ·.the· · upstream pressure applies). · .. · . · well, it just is hard to make any general
Frequently there is knowledge. aboutthese comment. · You .' need to use · all • the
changes, or the potential for change; so knowledge you have . gained on what affects
there is knowledge about. the upset · or the performance of a loop.
potential for upset. Since the flow loop will
be fast, any effects that last very long can be ; Probably the first consideration Isto put the
judged as unnecessary and fixable by tuning loop Oli IJlanual, to see if things. get better or
or by hardware improvements. ·. worse. . This ; is easier . to do in some
.. .

situations than· others. It· is an alternative to


At the other extreme might · · ··be a . keep in mind .. It is particularly
. . .
helpful in
composition control . system. These are situations where interaction is suspected.
'

Chapter 9, Potpourri 77

, . :: ·'·:·J~~.::,.~:..' ~'.I·.;
~.;: :·:· ;-:···,~ :·~ : ~~ ~ ; . . . . . :· ..
:: .... ""· .i., •., ••••••
• •

Occasionally. the desirdf~;1-iM;.7.1;_f9r:,;:/tb~ in filtering or dampening noisy


controller cannot be set because:·:c,f.~a acij~y.\' '.: - .measurements is to reduce what you. don't . . . . . .

measurement. If the desired gain were used, · ·: · ·_· want andkeep what you.do want, The first
. . . . . . . . .· . . . .

the valve- would move too . much and order lag is calleda "low pass'' filter. It
actually make things worse instead of better. passes (does not dampen) a low frequency
In addition there· is excessive wear - and tear variation; it attenuates high frequencies. So
on the .. valve and excessive . instrument air if a decision .is made to use a first order lag . .

consumption. Sometimes this is an easy for dampening noisy measurements, this


problem to solve .end. sometimes it isn't. involves an. .implied decision . that a
This section will . deal with the use .· of the separation can be made between desired
first order lagand with the Moore Products
·. . .
signal and undesired signal on the basis of
inverse derivative . unit to. _ combat this frequency (period).
problem .. : Elegant options that might be
available with a . computer are omitted. The This separation between what is attenuated
filter time setting is a first order lag, and and what is not is not abrupt. That is, the
therefore Is · not -_• considered elegant. It is filter does not, for instance, ''pass'' one cycle
almost specifically for this purpose. per minute and greatly attenuate 1.1 cpm.
The noise usually is not composed of a
single frequency anyway, but rather has
several components. For the uninitiated, try
to think of filtering sections of the noise
signal with sections of sine waves of
different frequencies. This will usually give
an· adequate feel· to make· the necessary
decisions.
The first order lag is widely used because it
is relatively easy to implement and is well A .· simplified · formula for. the attenuation
understood. Some - instruments, especially characteristic of a first order-lag is: ·
transmitters, are built with adjustable
dampening, and this is almost always a first Output amplitude P
.

order lag. It is very easy to approximate


.
......
.
. ·.. . .. ---
Input amplitude 6.28T
with digital equipment, requiring minimal
storage and operations. To appreciate the · · · ·_
Where: . P = Period, same time units as T
dampening characteristics of a first order
T = Time constant of filter
lag it is desirable apd- almost necessary to . . .

think in terms of frequency or period. While


This is not the whole story, for the ratio
its behavior has been explained in its time
cannot be greater than one. So this formula
response (figure 4.5), it is far more helpful
applies only if P is less than 6.28T. Even
to understand its response to cyclic upsets.
that is not the whole story, but if all you are
interested in is getting some meaningful
In the time domain the observed reading is
attenuation, that is all you need to know.
always changing at a rate such that it
The full story is in the equation for
would get to the true value in one time
amplitude ratio given in chapter 4, in the
constant. That says it all, but it is
math/algebra section for the first order lag.
information that is hard to apply. The trick
It is repeated in this section. The problem
:··1.
..
·:J
.,
...
· ..
. .;
• ->:

..·.·
..
.

78 . .'

.
. ..

. '
·.

· is.·that·'to get tlie.atienuation· . you also······· .. ·· If your loop··is.dfthepn.eumatic vintage you . . .


. · introduce · a lag. that will affect the; ·:: , : .might consider : · the; Moore Products
·. apparent . dead time and : hence the ·. pneumatic ·relay, Model 59~rr~1Whicli they · · . · > • •·

performance efthe loop. . . . . call a reverse action derivative}telay.' Itwill ' .


reduce the high frequencyamplitudes by .a
In summary, it is generally acceptable to use . factor of 6. A time adjustment -on it: is then
a first order lag to dampen noise when the setto-determine what frequencies are .not
periodicity of the noise is much shorter than attenuated. Its advantages are hard to
the · natural· . period. As · these periods explain without· recourse . to·. · .frequency
approach· each other, the disadvantages may response analysis. It is included in: this
outweigh the advantages. Incidentally, there
.
discussion because it is easily implemented.
. ~

is no objection to dampening ·a- recording. It is normally .placed in the outputof the


.

only, to make a chart mote , readable. contro Iler; but occasionally ·on. the input. · It
Dampen it as much . as:· · you "like; but can, under special situations, · minimize the ..

remember that the truthisworse than meets effect the filter has on the natural period, ;If
the eye. · · ,. you. are. in the versatile world ·of· digital
control ·: other · options · are available that· ,are
beyond·the scope ofthis booklet, ·
. .. ~:. .. . .

. '

. . . . . .
. . . ..
· In chapter 4 the amplitude ratio for a first
What · can be done · when the periodicity of · . order lag, Which is what a simple filter is, · ·.
. . .

the noise i's near the . natural period? .. was givt,n · as: . ·.·. .. . . :, . . ' . . . . .
, . . . : . :, . .
.. .
Frequently a good compromise is to reduce . ..
.
...
. . .
. .:

.. --:1. . . ·. .
.: . .

. 1 ·· . . ..·. . .... ' .


. . . . . ' . . .. : . . ·. : . .. ·,. .· ..
. .. . . . . .. . . . :

the gain and integral time together. If the .

--=====- = ~=========-
. . . . . .. . . . . :: , . '

normal tuning rules wouldhave called for a


.. . .. '
gain of 5 and an integral time .·. of IO p
minutes, try a gain of 3 · and 6 minutes of ..
• •
.
• ••• .; < .: •
. .

integral action .. Maybe even a gain of 1 and ... This· quantity· is ~/ways less than ·one. •.
. . . ~ .
an. integral time ·of 2 minutes would.
. ..
be a . . . . . . · . · For the lag to give sigr,ificant attenuation
good compromise. Reduce the gain to make
. . . .
· the roT term· must b~: significantly .larger.
the effect of noise acceptable,
. .
decrease the .. .
than one, in which case tneebov« terms
integral time until stability· prob lems arise. .. approximate these:' . · ... · · ·. · · ·. ·.
Or you could also· use an integral-only . . ' . 1 p
controller. This approach · is not . · · Amplitude ratio= ro T =
recommended for· level loops or other non- . ·.
. .
. :
6_287
·. . , .

self-regulating loops, as it will get you into .. .

The attenuation · is the reciprocal of· tne ·.


cycling troubles in a hurry... amplitude ratio. . ·
: . . . ·.

, . . .

., . .. .

' .
,.

79

-:
. . . ~.· ...
: .:,: . ,.
.' ••
... ...
..
: '. . ··:,
-":..

I know· of no "by-the-numbers" approach to improving control


performance. The following might be a broad general approach. The
first ones are easy to do and frequently cost very little. The · last ones •
'

may be impractical to do or cost very much.


. . .

~ Check the tuning,


·~ Reduce the lags to reduce the natural period.
~ Seek a faster or sooner measurement.
~Use cascade control.
~ Use feedforward,
~ Other?

~ Retune other controllers to help attenuate disturbances.


~ Perhaps use a new control loop to hold disturbances in check.

~ Make the process faster, such as mixing hot and cold to control

temperature; rather than using a heat exchanger, or putting a
recirculating pump on a jacket fluid, rather than having it be just a
one-pass.
~ Improve the inherent self-regulation of the process, such as using .a
·-,

larger tank to absorb the energy of the disturbance better.


,

81

TUNING CONTROLLERS IS MOSTLY SCIENCE. It consists


. . of fittin .. the time and amount arameters of the controller to the
time an .amount parameters of e process. An open-loop test of the
process yields the needed parameters, and sim le tuning rules based .
. on these parameters have roven to apply wel to a large · . ortion of ·
industrial control loops. unin parameters can also be etennined .
from a closed-loop test, thou the test is not as thorough. For a
. large family of loops it is possible to predict what is lik~lyto· happen
to • performance wh~n the process changes or when, the tuning
adjustments are set differently. · .
ALL CONTROL LOOPS WILL CYCLE if the controller gain is
high enough. The period of this cycle is called the natural period, and
it largely determines the potential performance of the loop. The.
shorter the eriod the better. The natural period in tum is closely
linked to t e apparent ( or real) dead time in the loop. It is
aradoxical that the natural period is not determined b the largest
ags in. the loop, but rather by the dead time and the sma ler lags. The
potential performance of a loop is limited by certain lags in the loop,
and trying to eke out better performance through tuning is often an
exercise in futility. ··
TUNINGRULES
. . . ARE DESIGNED TO GIVE REASONABLY
. .

TIGHT CONTROL. This may not always be ~e objective. Many,


and perha s most, loo . s do not need to be tuned tightly, However all
loops nee .to be tune as part of the process of J?Utltng the controller
into operation. Most. loops respond to changes m tuning parameters
much like the response curves given in chapter 2. Consideration
should be given, when contemplating retuning a loo , to what the
justification for the effort is, and whether the desire improvement
can reasonably be expected from tuning.
THERE · . A FEW TYPICAL GREMLINS which cause loops
to not behave in the typical fashion. It is important to recognize ·
these, for failure to do so can result in detuning a loop, not to
mention a loss of faith in the scientific approach. You will not have
to tune many loops before you run into one of these gremlins.
. .

BE PATIENT IN LEARNING THE METHODS which will be


new to you. They do work, and the method of understanding loop
performance will allow you to converse with others on a common
ground, sharing your experiences. Otherwise tuning is just one
isolated hit-or-miss experience after another.
83
. . ··:· ·.. ·.~·
. .
·.: -. . :>::'·
·. . .. . , . ----- ----- . . . r, . ·~ :_ ., ..: :: -~· : .' . . · .: ·.· . , .:'. ·:_,._ ·.'.·. . . . . ·. :· . . . . . . .· ', . ... "· ··.
. .· : ,;· . . . '

' ........ ,.
·····
. .
. . . .. . .. . . . . , ...
....... . . ~ .· . . . ... .. : . . .. . . -~ .. ·:.: : . . _;, ··'. . . .:·:··
; : .•..

. .

.. . ,... .: .. . . . .- . . . .. . . .
. ...

. ' . .. ..

..
,: •

.
,f. ·..

. . . . . \
•.:·

. . .. ·. . . .. . . .'. .
, . . . . ...,·:.. . .
. . : ·. :.\ ......::·.. :_'. : : .. ": .· :_ ... -. :~::.:?. ":·:.: . . · : . . : . ~~.' . :·. : v• ·: •• .::. • • .: ••

. .
. . · . ..··. .....: ,.' .... . ..,· :.. ·. -· ..' .:·. •.· ': ... : .:· ... ~ .. ·. . : .·
•..·... . . . ·., ...
.~· . . .· .. . . . ·· .... . .

The definitions .marked' below with: an asterisk (*) have· been,:·reprinted:·byr·permi·ssien,.:~from


Comprehensive-·<Drcii~naf'y . of 'Measurement and Control,. 2nd· -Editioff~i··by W~ H. · Cubberly,
copyright 1991, Instrument Society of America. Some of the terms have<:broad::,-meanings:, in
which easeon:Jy.tnostfmeani11gs:which most apply to the usein this booklet:have.,:beett·edpied~ ..·.
. ..

. .

Averaging'·· level· control · A description of Dead band * · The . range· through which, aa
the,tnning';:method·•which allows the volume input may be· changed without initiating an
of 'atank tobe 'used ·as a surge capacity, observable change in output. · There are
rather· than·· sending· upsets on to· the other separate and . distinct input-output relation-
parts of a process, as would happen if the ships for increasing and decreasing input
level loop were tuned'. by the typical tuning signals.
rules, · ·. · . · · ·· ·
.. . . . . Derivati"Ve action * · A -type ·of control-sys-
Capacitance· ·The ability of a systemor de- tem : action : in which , a · predetermined
vice to store energy. relationexistsbetween the- position of: the
final control-etemenraad the derivative· . of .
. .
.
. .

Closed . . loop:.(feedback· loop) 'See loop, the controlled variable . with respect
..
to time.
. . ·. . . . .. ... . . . . . . . . . ...
. · . . . .. .
·>:r~~: :·. ·_. ....,_:·. :. . . .: ~-. .~·: {(·. ;. ·:\-,:~:,.· ,._· . · .: ·. :. :· -: . . . .·.
~
. ··:: . . . . :

closed. · ..... ·. ..··:_;r:_~ ·. ··: ·_. ··:.


.

Derivative time · *··iJn ·ptoportioaal-plus~de~


Conditional stability . * 1. A linear · system
. . .

rivative control "action; for •· a .: unit · ramp


is . considered conditionally stable if it · is signal 'input, : the .: advance in time of · the
stable for a certain interval of values of the output · · signal· . · ,(after::: ·.. transients- . ·: have
open-loop gain, and · unstable for . certain subsided).··. caused.oby · derivative .: 'control
lower and higher values ... 2. The property of . . . .
. action, as compared with . the output . signal
a controlled process by which it can func- due to proportional control action only, . •. · . · ·• : !.
..
tion/ -in either . a .stable: .:or ·unstal.,·le: mode,' de- . .. . . . \" ..
. . .·

.Pen··a: ing :o;n/~~d:'4,1;,~~·;-"'1:m· · . ·.:p,~.· ·o· ·•se'.~/~i:5.t,}\\


·'-'UI"l :. ~f.~ '-'•:f.:J· ·. ·. U. · ·.. · · :.·. ·. · ·_ ..•. ::· • .· . ·:.: i, ·· . . ··. . Desired value See value, desired.
. . .. . .. . .
·:\~~:·'..;:}:·.: : : : -~ . ·: ·. : . ·. ·.' ·.· ":: / :~~ : .

Control algorithm* A mathematicalrepre- :· ...· . .• .. ,.·•.. DCS,'.A;digital controlsystem.. ·


sentation of the control action .to ~-&er:pefi · · ·.·•· · ·f <>t/:i \:i::.';.: )/,, . 1.i::) ·>,·>:': : - · \ .• :: __ ·t .< ;,:: .

formed.. . · _. · - ... ··. · -. · ·. • : .. ~;;;;:~}';,<':· · ~·.~ · .; .: • Disturbance * An· undesired change: in a . ·c·~· . ..

· · · · · .· · · .: :::-, , , .. : : · · · . variable applied to a system which tends to


Controller gain See gain, controller. · affect adversely: the value of· a controlled
variable. · • · ... . ·
Cycle time * 1. The time required by a
.· ~
.' . . .. ·: ·.. ·:,_:.:,. -~:' ~ :· :.. · .: ..: .... ··.:..): , -, . ·.. ..

computer to read from or write into the sys- Error *,


S. In: a: single· · automatic control
tem memory. If system memory is core, the loop, . the setpoint · minus . the · controlled
read cycle time includes a write-after-read variable measurement. . . · · ..
(restore) subcycle. 2. Cycle time is· often
used as a measure of computer perf onnance, Filter time The time constant applied to
since this is a measure of the time to fetch the measured variable before it is introduced
an instruction. to the controller.

84 . Glossary

Gain, controller The ratio of a change in . _ Lag· • _ 1. A _ relative . measure of the time ,,

the output to a change in either the con- delay between two events, states, or mecha- -:.,~
. :;.·,.
.·.·:

trolled variable or the setpoint. msms, . ''
·:,
··:;
..
.-,'
. :.
,,

Gain, ultimate The controller gain which Limit cycle A cycle whose amplitude is '

...
·?.

.i....~

will maintain the loop in a continuous cycle limited, typically because of some nonlin- ....~

'
..~.l
,!

.
of constant amplitude, without depending _ . <~arity in the loop. . >·•
. :~.
.'

on nonlinearities to limit the amplitude. i11 :, _, .. ·..'


"
....
. '•
''

other controller functions; such' as integral ·.-. Linear * The type of relationship which :·}
.•..
-~-
··.'..·}

and derivative action must be not it effect. exists between two variables when the ratio i
.:·.
• t,

_ of the value of one variable to the corre- ..fl:


'.'

Gain, .process. · Unless otherwise, defined in · - • · · · sponding value of the other is constant over ;i
:1
context, it is the steady state ratio of the . _· . the entire range of possible values. ·-·~,
. \..(
·-,1

• •....c

output of a process to a change in its .input.·· -- : · . -~


...t; ·~ !'
_.;,q
It will always have units. Load change The same as a disturbance. ..~.:..
.\I!
\•
•.:t:

"' ~
.: .;~
. ,.
'..·,
IAE * Integral absolute error. : A measure Leep, closed (feedback loop) *. A signal
··:~'i
..
..i:
.~-s
....).",
.,
of controller errordefined.by the integral of path which includes a forward path, a feed- :. :~
,j
v
:1

..
·.~

the absolute value of a time-dependent error back path and a summing point, and forms a .,
...,
\~

function; used in tuning automatic control- closed· circuit. \;


. :-...
. ·:-:
••
-·.
ler to respond properly to process transients. .. ..,.

*
. :.

See also ITAE. Loop, open A signal - path without feed- ·.{
·'.

'

back.
..

IE The integral of the error. Itis a measure '

of the performance of· a control loop - to Non-self-regulating The opposite . of self-


disturbances -or setpoint changes, but posi- regulating. ·
tive errors cancel negative errors.
. :· ,. Open-loop unstable The same as .' condi-
ITAE * Integral time absolute error. A tionally stable.
measure of the controller error defined. by
the integral of the product of time and the Period, natural - The period of the cycle
absolute value of a time-dependent error present when the ultimate gain is in effect.
.
function; whereas the -_ absolute value: - pre- . .
'

vents . opposite excursions in · the · process Period, ultimate The same as natural
variable from canceling each other, the mul- period.
tiplication by time places. a - more - severe
penalty on sustained transients. PID action * A mode of control action in
which proportional, integral and derivative
Integral action * A typeof controller func- action are combined. .
tion where the output ( control) signal or ..
action is a time integral of the input (sensor) Proportional action That part of a control-
signal. ler's action which produces a change - in the
.
output in proportion to a change in the con-
Integral time In a proportional-plus-inte- trolled variable.
. . . . .

gral controller it is the time, in an open-loop


test, for the controller, while integrating the Proportional gain See gain, controller.
error, to cause an output change equal to the
proportional change, when introducing. an Proportional band * 1. The change in in-
error from an initial zero-error state. put required to· produce a full range change
in . output due to proportional action. It is
Inverse response * The dynamic character- reciprocally.related to proportional gain.
istic of a process by - which· its output re-
sponds to an input change by moving in- Ramp A word used to describe a parameter
itially in one direction but finally in another. which is changing at a constant rate relative
to time.
. . .

•.

GLOSSARY 85

* _· l.
..

Rate action . Another name for the de-


. . .
Self-regulating •· ·. the· property of a process
rivative control mode. . .
or machine which permits attainment of
"'- .·. .
equilibrium , after a disturbance, without the
Rate
ti time Anoili;r ri.•~1RJlie
. :•. derivative ,., .. :. . ... ,.".~··.•
·)· ., ... ' . ' ...
::::.-:.,:~~"?:;;·:·;·~.:::). . .
intervention of. a controller.
. , .

1me. · · · -·\:;.1<(?:.:::-. . ·. _·. . : . . . ". . . . . . . . . . . .. .


... · ~ · . · . .< · .: . ··• : S~-dy·/~tat~ * · A characteristic of a condi-
Reaction curve Another name for the step tioa, · such as a ·value,· rate, periodicity, or
response curve. · amplitude, . exhibiting . only . negligible
change over an .arbitrarily long period of
Reset action . Another name for integral time. It may describe a condition in which
action. some characteristics are static, others
. .
dynamic.
Reset rate The: reciprocal of reset time or
integral time. . .
Step change *
The change from one value
·: ·. . . . ·. .
. . to another· in a single increment in negligi-
Reset time Another 1UlllJe for integral time. ·...... .
ble time.
. . . . . .. ...
. ' : . . ..... , . .. .

Response, step * The ·total. (transient plus· Step response See response, step.
steady-state) time response resulting from a
sudden change from one constant level of Time constant For a first-order lag, the
input to another. time to reach 63.2% of the final value in re-
sponse to a step change.
Sampling rate* For agiven measurement,
the number of times it is· sampled per sec- Value, desired * In process instrumenta-
ond in. a time-division-multiplexed system. tion, the value of the controlled variable
Typically it is at least five times the highest wanted or chosen. The desired value equals ·
data frequency of the measurement. the ideal value in an idealized system, ·

Velocity limiting * A limit which the · rate


of change of a specified variable may not
exceed

. .
. '· .
: . .
•.

A-1

·,.

.,.
. ::··~·
. ·,, .~
·}'
..• ::/ · ..~ . ;,,..

.' ...,,:. .:-v c:


' .. .•

.
·"
. ,:

.,

i;

., .
. ......;·., .
-._..... · .. . . .
. ', ..
. . .
. . .
.


' .
.. .. .

A-2
-. . .. . . ...
.:''·:-..·~l
••• (<4

... •V
-~ :·!:,
.•. "i

PURE DEAD TIME.PROCESS }':,


......:\..s
·. : .. ..
.. . . . . . . .. . , . ·.:..:.
.. . ... .
..,. . ,· ...
!,;..
:
.
. ·.:

. ·-.:~.
..
Graphical Illustration that P = 2L . .
. ,:,
."
. .::··,.
-: ~
.: ·/
... ....... . : :·
....
.. . . .. ..
..
.. i:. . -:
. :,'

r ... Controller
- , Process Pure

.. '
. ·.'

Dead Time Controlled ,.:.

I Set I Variable

I Point
Kc L
I
' - --- ....... ·-· ..... ..J

.. . , : . . . . . ... . . . :, . . . . ,. . .

..

..'·. .
., .. . . .

.

....
•• \
.
~. ,
..
• •

. ..
·:\~" ·.':•• -', .: ·.:•••••••

-~·, -."'. ,·r .


: • :·: ••
·
• ·< ~ ..',
:.·..
• • •
: ·. . . . .
. .
.
. '- ~
.
. .. ~ . . ..
> . ·, .·
.. .; . .·~.. .... . ,.. .. . . . .. . . - ...'. ,. . .. . :....
. ..,·. ·.. .
··.. . ..
,. ... . . .. . . ..
·. '"·: ·~ ,. ... · . :·
. .. ' .. . . . .. .
. ;·. . .,, . .. . . ': . ..... : , ·.
. :· .. ··'
(
. · · ......:·:·. ~ . --~-~: ....... · :· .:.:. . \:::·.:,xi· · -: · ·:: · ?. · · ·
·...

·. . . . .. .I . .. .. . . .I ..
I
,. . . ~ .

I
Point I I I This curve is
#1
repeated from
bottom of page.
I
I I I
I I
I I I
I I I I
I I I
Point ........, I
~
No. 2 is No. 1
#2
multiplied by
I I I
proportional gain.
(Kc= 1/2)
I I l
I P = 2L
I I
I ' I
Point
I I No. 3 is #2 delayed
#3
--- ---- -- -- -- by L, and multiplied
I I I by process gain.
· (Kp = 2, arbitrarily)
I I I I
.
NoteP=2L
I I
I I
I I
I I
Point I I I I
No. 1 is the
inverse of No. 3,
#4 I for correct control
action. Note this
is -180° phase
related to No. 3
,.

A-3
PROCESS WITH DEAD TIME AND INTEGRATION

Graphical Illustration that P = 4L

r --· Controll~r ·.
···
.. .. .. , . Process
-. -,
I 4
. .. :. . L . :fr-: \::::::;:;;. ./ • . I :.
I
> -~ . • •. . .. ·.
. Kc .... ):'· ,. • r: • .
. .. ....»,

L L--------~-------~

Point I I I I This curve is


#1 ------.....-1~------ ..... _ repeated from
bottom of page.
I I
I I
I I No. 2 is No. 1
#2 .................... 1_. ------- ..... ---- multiplied by
I proportional gain.
I ' I (Kc= 2)
.I I I I
I
I I I
I . I I
L I
I No. 3 is No. ·2
Point I I I I delayed by dead
#3 time, L. Assume
. .

process gain for


I I I I this part = 1. Note
I ·. . I I 90° phase lag.
.. I I . . ... ·. ~· ..

.
~

P = 4L No. 4 is changing
I I I: at ·its maxi·mum
Point ·I I I rate when No. 3 is
#4 ....... ----~~ .......-- at its maximum
: . .. . . . . .. .

.I I and is not · ·
changing at all

~- . , . . . . . . . . . . ·. . ·. - I
. I.
when No. 3 is
zero. -Note the
·.~:::::f,.:.:;·.·.:. :.~~·: .. ·? :-·· •• :·... • ~ ->. • ,· : •• •

. . . .:
• • < • • •• ' •••

. :';·:
90° phase tag and
an· arbitrary
I
'

.
. . ... . 1· ; .·. .

attenuation of 2.·· ·
I I
I I
I I I No. 1 is the ...

:I
I I inverse of No. 4,
Point I for correct control
#1' . •. ' -·.
.
I action . Note this •
. . .. .

is a 180° phase
lag.

,.,
. . . ·'· . .. . . . . .. ..,."
.. .·
. ··•;;.,
.:,
. .· ~ . . : ~:
.
..
.
. '.
. : . :.
... . . ' .
:.,:::
.. ....
··.r:
. .;
. ,. ·.-·:·
.. ...
A-4 . ):
,.
· .. t.
··,·,'
.·:: :{
.. ···.·•.
' ...,
•• >

. . . . ·. :
. :.
.. .. ' .' ., ,. :.
·. .. . . ':'~. . .

..
..... ··,,
•'

. . . .: ,..., ·,.·::.: .. . ;. .:
~ . . . .. · ... ,
..
. . /··
..

·.. , _:::. ·:t .. :·,..:, :~- . .· ... .,. . . ..· ..;. ..... ~~-. .

30 . ·: ./
·.•..
•'
..
. ::
,. . . '
.. . . . ·. .. ..
·.· .•
•• ..
'
. . . .·~·· , .
., ..
.. .;· .
. ... :. - ·.. Derivative·.. ': ..• - · . -• . . ·•·. · .: -, '·· ·
DERIVATIVE .
.
FREQUl;NCY RESPONSE;.: i-·_:: ;; ·. : : ; Gain= 50 ,.;.~ -~ " '··: · '·

. . .. . :" .· ·.. .: . ..: .· .. ·-~· ·.· : ... · ::~.:,: .;. ..;:/~ .. ~?··:~::, ·~:,.~_:..,.· ...:-. :..··.:,;:~> .. ~·. ,. : :·~·.·_.:.·; :,'

24 -------1---· ----:- . . .. . ·.·


.

Td
.· .·· .. ·_T· .d·-;_·_.s·.:. +·. ,1 ::,, .:
s>~ 1. :\.·...
.

.· .. :. _:.,::, . · _·._ · ·_ · · .; , ·. . . _.· ·


. ·\: ;.:~
.

·i :-:· :,;;:.
.

.~ .
. .

.· ...•.. ·
.

:. . . ... ,· ,,

. .: . . . -r. . .. ' ... : . ., .


. . . ·.. : .. . . . .. . . . . ..... . ... . .

. . . ..
., ·. . . .. . ... : ·.. · : . .. , :· ,.·,~~;.··
...
···:.,· .. :. , .

. ..
·'
10
·i ~--·--·~-------_.;.---+----------------I-------~, . .~---+--#--~.~----~------------_.
-

18 -~------ __......___,_
~

- 0
..,· ...
.
, , ,,,, ..
.

/
. .

·S- ·.. . ::.·...... ·: : . :.·


·. .. . ·.:-, ·.. : . ·,. -: . .
~- '.
:,
. , .;.· •.. t:· ·.. •· ~ ~ .. , . . .
..,: ••••• •·. -c •. ·: ••• • •• ···~ • -". : ·,:.
;
-: .
..;J
. -. /.. ·/···· . 6
g ·.'.
. ·.:.
.....:........ . ... ··'. ~':" .... ·.
.
. ·.: .• .
. .
.
. . . .. . '
........ ..
.:··-. .
. :;;,
. . .,. . .' .
. .
. , . ·. . .


I- . ···~·: .. : ·' . . ,
~

m 12
= .. · . ,..:·:.'.:- ..· .
..

Q
c•
·-e
.'

..
cw ...

....
. . ·., ;, . ·.. · . . . . . .·
. ..; ·. ::;···· . . .
..
. .. .,. ·:., : .. ~ ·. ·... . .·
" .. . . ..

.. . . .
.~.
. .
·:

6
~.
l'

.
..,. .

.
·····, .
..... · .
•• •

.,
.
...

' ..
. . ·i,.·. ... . . ~··
..
....
.~
.
•. .
•••
.

.
.. ··
. .·

.
.
:
. .
.
.
'
·.
.. -
..
. ~ .
..

.
. ·.,. :: ,·..
.. . ..
...

0 ,
'
.

.
. ..
.: . .. . ·., . . ..
-,"

' ...,
.... -: .: c-orner Fre~uency . .
: . . . : . ·~ . i ....

..
-. · .. : Set by Rule · . •. - - ; · ,: .·
. .
..
.' . _,.. ..
. ..
. . . .
, ..
..
. ._ ,_ .
/> ;: .,,,,,,,-
. · · . •< . Td-· · ~ ··. Pn/8- :-, . ~ . · .·.. - . . . :·. : . - . · :. •._. .·.. ".............
, ,. '. ~ -· • . - ·. .
..· . '. ..,, . ~·.. . .
14,.Y
. .. .
:
: ~. ::· .. : ..::.:··
• • _ ...--
. . . .. . "':.: . .. -·,: : :· . . . ..
#J
. . .,; :
.·'
' .··:·
.. . , ·. ....~...,·,
.. . '.., . ... .
·. . :>;' .. ..

90 .. . . :: . . .

.
'
·:: ...~·. . : . -~: ·. ·.:. ... . . . . ~. :··.·... • ...... .';.:,t.~1.. . . -', ·... I,.... . . .: : •:.,,:, ... ·.. . . . . ·,.· .·· .... >;

..., ..
.....,.._ Frequency Representing . . .. · '/. Derivative ':
..
P" in ~UJlipn.T, ~ P"/8 ...... _z 't _
'
.
.. ·.
:~: .
.'
.
.
..
.
. .. p ---
.
.•. ---.
.
_ ~

. ..
.
.
.. -~
~
.
Gain = 50 ·
. ~
. . ....:.
.
.
.

' • . . . ···.·.· . . ' !' i, ; !


I
'. .
~ ..
,
.. .
.ff//1!1' . . . . . . :'.' ' '
..·. ·.· . . .
. . . ...· ' . . . !: : : . :~ ~~

'
• • ----. . . --~------1~-----+-- . . . ..----------.,,~
/
. . . ·.. , . ·1
~

~-------------1----------:--+---.-,------
.
.... -----------------
' ..... . ' .'-',.;. -~- ....
.. . --

· · ---- " 1·0·· · .. · . · . ,_·.·-,;. ·:··< ,: _: . . ~ ,


...........
' '
4 . . . ___

• /
-~ ..,. .. -- . .

• ""/
/
6
~.
·,.·
', ""' 4 ------
... - __....... _
-
.....

..,...
,•

.. :;
., .
4

' ... - ...


..,,
.......
·>

. .. .
...
• <

- '.
,.
... . .
. '., ·..
·.
. . . ·•.
.. .\;.:·'
.
;.

·~
,·,.

.
.
·. ·.
. . .... :
'


•• ,, .
..
0 ... . .' . .
... :'

: . . .... ·. ·: . ·,. :. ·.
0.2 - · .• - , 0.5 .,,-; .· · ·-1· , ':- ·
..,
... , ', ·._:.... . . :- .
...
· . ·. ···2· ..... , .. . :
.. . .
:. ·. ... :·

5 .. ; 10 . · •
..

..
. . /.· .
. .. : ·:
20 5
. ....
. -~ .·
.. .
·,

Dimens,onf'ess Frequency Tdm (=2 n: ftd)


•. . .
. ·,

. .

.'
. ·;:,;.
. '.
. ..
·. :
..
. '.·
. ,:.: .
. .
..
..
. .
. . . . . . ' . . .

' .

..
... . ,
,:,. "
... MO
.. :

'' .

. The first edition of this boo · . ~ · , s self-publi,Md;- .:l didn't _ · :. . . t to do'. that but I did. It sold
close to 17,000 copies! This secon . · · ,is ril-~--~~-:-!:~-;;p~d. :J had so much. fun··
publishing the first that I'm looking forward to this explrid~~~~-:J;:~'.:,·~_:;;;~::··~;~'.:,·~?:;~f-~~--,::.~.:~{:·N:!.tl\• ~1.:panded
treatment of the subject useful. With the fourth printing of the second edition: rtlte :booklet. has .
now sold more than 22,000 copies.

Reader Feedback
. .
. .
If you care to make any comments about this booklet, I would be happy to hear ftoln you. And if
there is ever a second printing, I will consider incorporating your suggestions. I have left some . .. . .

white space to help in that regard. I fully realize I have experienced only • part of the total field
of feedback control. et

Marketing
. . .

(
· If you think this booklet has been helpful I would appreciate your passing the word along. I am
. . .

dependent on FRIENDLY AND INTERESTED PEOPLE to:

PASS THIS AMONG YOUR COLLEAGUES


or
. POST IT ON YOUR BULLETIN BOARD · \

or
ASK YOUR BOSS TO ORDER SEVERAL COPIES
. .
.

· BUT ONLY IF IT PLEASES YOUI

. '.
WHEN AFRlEND GREETS YOU WITH ''WHAT'S NEW?''
. . . .. .

YOU CAN MENTION Tms PUBLICATION!


. .. . .

.. . . \
.
. . .. .
. . . . ·. :.. . . : . . .. . .,

Orders will normally be shipped the tl.(.St'.·i· '· . _ •. tf you want faster U. S. delivery, Priority
. : .. . /.:.· .···· . . ~· : ·,.·: .. · . - . . .. . . . . . .·' ; , . ·... . . . .

Mail is an option at extra co.~t, Th.is is a· one.:ttWl sho.\¥. :.ff I:. '0.U~ of :town for more 'than a few
days a friend will fill ord-~feceived. by mail, fax or phone;· but not by e-mail. If your order
~ . . .
..
'

seems late in arriving, please call. Something has gone wrong.


. .

WITH HIGH HOPES YOU WILL FEEL MORE COMFORTABLE


ABOUT AUTOMATIC FEEDBACK CONTROL·. .

.
.
. .
.
,• ·:.

AFTER YOU HAVE READ THIS TUTORIAL THAN BEFORE, .


: ..
......
...'
....'
I WISH YOU HAPPINESS! ;
: : ",

,. . .
. ,.'

·'

David W. St. Clair . ..· ·....


..... . ::::·
:
· ..
. .,
. .. ·. . ·.:...
. .
. . .. . .'
. ..

". . ..
,

DISTRIBUTED IN AUSTRALASIA BY
INSTITUTE OF INSTRUMENTATION AND CONTROL AUSTRALIA, INC.
PO BOX 82
Balwyn, Victoria 3 103
Australia
Ph 03 9816 3333, Fax 03 9857 5057

PUBLISHED BY
STRAIGHT-LINE CONTROL COMPANY, INCORPORATED
3 Bridle Brook Lane
Newark, DE 19711-2003
(302) 731-4699, fax (302) 454-8599
E-mail: dwstclair@aol.com
Home Page: http://members.aol.com/pidcontrol/booklet.html

Potrebbero piacerti anche