Controlling fired heaters
P. Eng., 2000 May 20. CONTROLLING FIRED HEATERS HEATERS © Walter Driedger, P. walter(at)driedger(dot)ca First First !"lished in Hydrocarbon Processing , #ril $%%&. 'his #do"e le is a*aila"le for download.
INTRODUCTION. 'he !rose of a
red heater is *ery *er y si+le 'o 'o add heat to a rocess -!id. ts reresentation on a rocess -ow diagra+ is also *ery si+le. /!t, of co!rse, red heaters are a re a+ong the +ost co+le ieces of rocess control e1!i+ent. Each f!rnace is, after all, at least two ieces of e1!i+ent in one. Firstly, it is a secial *ariant of the shell and t!"e heat echanger since its !rose is to echange heat. econdly, it is a che+ical reactor in which f!el and air !ndergo etre+ely eother+ic eother+ic reactions to rod!ce the re1!ired heat. n re*io!s articles of this series $, 2, 3, 4, the rocess asects of controlling a iece of e1!i+ent were resented "efore dealing with rotection and safety. safety. 'his ti+e the toics will "e re*ersed n the case of red heaters, it +!st "e safety rst5
SAFETY. f red heaters had not "een in*ented and were "eing roosed for the rst ti+e, wo!ld ro"a"ly say, 67o!8*e got to to "e 9idding. 'hat thing will "low ! in yo!r face the rst
ti+e yo! throw a +atch in it.6 :owe*er, at least a half a "illion gas red heaters are in ser*ice aro!nd the world (according to the #+erican ;as #ssociation). Most of the+ are oerated "y eole with no technical eerience whatsoe*er< few heaters "low !. till, the a*erage do+estic water heater is not in the sa+e leag!e as a hydrogen refor+er f!rnace. 'he fact that accidents and disasters are as few as they are, is d!e to the long eerience the h!+an race has in dealing with re. # +illion years, 8+ told. For the last cent!ry, this eerience has "een e+"odied in *ario!s codes a nd standards that ha*e "een written into law and are en=forced "y insectors aro!nd the world. THE CODE. 'he +ost o!lar, or notorio!s, of these codes in >orth #+erica is >FP# ?@@ and ?AA iss!ed "y the >ational Fire Protection #ssociation. 'hese ha*e "een considera"ly !dated in recent years, esecially in ter+s of clarity. clarity. >e*ertheless, there is still the ro"le+ of interretation. 'he code is not at all easy to read as it co+"ines +any facets of constr!ction, instr!+entation and oeration in a single doc!+ent. >ot only that, "!t the code@ contains the following disclai+ers It is not possible for these stanars to en!o"pass spe!i#! har$are appli!ations% nor sho&l these be !onsiere a '!oo(boo(' for the esi)n of safet* s*ste"s.
and This stanar applies to boilers $ith a f&el inp&t of +,%-% /t&0hr 12332 (45 or )reater. )reater. This stanar applies onl* to boiler6f&rna!es &sin) sin)le b&rners #rin)7 a5 Nat&ral )as onl* as e#ne in Chapter 2. b5 Other )as $ith a /TU 8al&e an !hara!teristi!s si"ilar to nat&ral )as. !5 F&el oil of No. ,....
and F&rna!es s&!h as those of pro!ess heaters &se in !he"i!al an petrole&" "an&fa!t&re% $herein stea" )eneration is in!iental to the operation of a pro!essin) s*ste"% are not !o8ere in this stanar.
What is an engineer to !se for a g!ide when the f!rnace is not a "oiler, "!t a feed heater< does not eceed $2B +illion /t!hr, "!t is only fo!r +illion< does not "!rn nat!ral gas as dened in Chater 3< "!t renery o=gas with a high hydrogen content Desite the disclai+ers, the >FP# series is still an ecellent g!ide to the instr!+entation and control of any f!rnace. standard deals with a *ariety FUEL GAS FIRED% SINGLE /URNER FURNACES. 'he >FP# standard of f!els, "oth oil and gas. 'he disc!ssion that follows restricts itself to f!el gas red, single "!rners. >FP# standards ha*e "een followed as +!ch as ossi"le and ha*e so+eti+es "een eceeded "y adding co+onents and control f!nctions w here the secial re1!ire+ents of rocess control +a9e it ad*isa"le. Fig!re Fig!re @=2, shows the in=line instr!+ents tyically installed on a "!rner f!el gas train. Dia+ond sy+"ols with an 66 in the+ refer to of the /!rner Manage+ent yste+ (/M).
FUEL GAS SU99LY INSTRU:ENTS 9C;6+ 'he f!el gas s!ly reg!lator is only re1!ired when the f!el gas ress!re +!st "e
red!ced in two stages. 'his is often the case in renery ser*ice. ee Fig!re @=3 for tyical reg!lator settings. 9I6+ E*ery reg!lator sho!ld ha*e a ga!ge so that the oerator can set the reg!lator roerly and so that he can 9now that it is doing its Go". 9S;6+ Many standard f!el gas train co+onents ha*e an !er ress!re li+it of aroi+ately $00 sig (&00 9Pag). Fail!re of "oth PCH=$ and PCH=2$ wo!ld o*erress!re the f!el gas train if the s!ly ress!re eceeds the rating of any downstrea+ co+onent. n s!ch cases ro*ision for ress!re relief is re1!ired. >ote that it is not !n!s!al to consider do!"le Geoardy in "!rner safety analysis. FE62 'his is the lace to !t f!el gas +etering, if re1!ired. >FP# !ts the f!el gas +eter on the +ain gas line down=strea+ of the +ain gas reg!lator. /!t when !sing acc!rate f!elair ratio control it +a9es sense to incl!de the ilot as art of the total f!el s!ly as it +ay s!ly as +!ch as $0I of the total f!el. 9T62 Jarge "!rners with a *aria"le f!el gas s!ly ress!re +ay re1!ire ress!re, and ossi"ly e*en te+erat!re, co+ensation for the f!el gas +eas!re+ent.
9ILOT GAS LINE INSTRU:ENTS 9C;6++ 'he ilot gas reg!lator is set for the act!al re1!ired ilot gas ress!re. 9I6++ f co!rse the reg!lator has a ga!ge5 /;6+< n s+all "!rners, the rst ilot sh!to *al*e +ay "e a standard ind!strial, two ort,
solenoid *al*e. Ji+it switches can "e considered otional on solenoid *al*es. n large "!rners, a "!rner safety sh!to *al*e is generally the referred choice. 'hese *al*es ha*e an internal sring to force the *al*e sh!t, a solenoid to hold the+ oen, and a s+all +otor to re=oen the *al*e. te+ /H=24, following, ro*ides +ore detail concerning this tye of *al*e. o+e installations !se a standard, ind!strial, diahrag+ oerated control *al*e. n e ither case, the *al*e +!st "e fail=closed. =SC0O6+< #t least one li+it switch is re1!ired on /H=$4 if it is anything +ore than a si+le solenoid *al*e. 'his is needed to ro*e that the *al*e is sh!t d!ring the !rge hase. For failsafe oeration it is "est to ha*e a li+it switch at each end of tra*el. 'he !er li+it switch ro*es that the *al*e is f!lly oen at all ti+es when the ilot -a+e is s!osed to "e on. 'he article, 6Ji+it witches Key to Hal*e Lelia"ility6? , elains eactly how to connect do!"le, failsafe, li+it switches. /;6+- 'he ilot *ent *al*e +a9es certain that there is ne*er any gas ress!re on the second
sh!to *al*e desite any lea9age thro!gh the rst one. ince its only !rose is to *ent lea9age, the *ent line has a s+aller "ore than the s!ly line. t +!st "e fail=oen. 'he Canadian ;as #ssociation code C#>C;#=3.%=M?& % reeats the following ta"le fro+ >FP# ?A# & for deter+ining *ent line sies ;as !ly Jine ie >P ++ $B ( 40 ) 2 ( @0 ) 2B ( A@ ) 3 ( ?0 ) 3B ( %0 ) 4 ($00) @ ($20) A ($@0) ? (200) ? (200)
;as Hent Jine ie
>P ++ N ( 20 ) $ ( 2@ ) $O ( 32 ) $O ( 32 ) $B ( 40 ) 2 ( @0 ) 2 ( @0 ) 2B ( 40 ) 3B ( %0 ) $@ I line cross= sectional area For low +olec!lar weight f!els s!ch as hydrogen (Mol. Wt. Q 2) or +ethane (Mol. Wt. Q $A) it is s!Rcient to *ent the *al*e to at+oshere o!tside the "!ilding. For f!els hea*ier than air, s!ch as ethane, roane and "!tane (Mol. Wt. Q 30, 44, @? resecti*ely), the *ent sho!ld "e ied to a -are header. Ssing the ta"le a"o*e, it +ay "e deter+ined that a N6 or $6, si+le solenoid *al*e is s!Rcient. f not, a fail=oen "!rner safety *al*e sho!ld "e !sed. t oerates in a si+ilar, "!t oosite, +anner to /H=$4. =SC0O6+- Ji+it switches sho!ld "e incl!ded and incororated in the /M logic, if a f!ll=sied *ent *al*e is re1!ired. /;6+3 'he second ilot sh!to *al*e is identical to the rst, /H=$4. =SC0O6+3 Ji+it switches sho!ld "e incl!ded and incororated in the /M logic, if a f!ll= sied, second ilot *al*e is re1!ired. 9I6+> 'he nal ress!re ga!ge conr+s that all *al*es are in their correct osition and that the aroriate ress!re is a*aila"le for the ilot -a+e. /Y6+> 'he igniter itself is essentially a sar9 l!g owered "y a high *oltage transfor+er. t is caa"le of sar9ing contin!o!sly as re1!ired "y the /M. /SLL6+? 'he ilot -a+e detector is !sed to conr+ that the ilot light has ignited and is "!rning in a sta"le +anner. # *ariety of tyes eist deending on the sie of the "!rner and the tye of f!el. # -a+e rod is a si+le electrode that roGects into the -a+e. #n electric c!rrent asses thro!gh the -a+e to the ilot gas nole and energies a sensiti*e relay or electronic circ!it. ince it only senses -a+e at a oint, it will not detect the +ain -a+e. # disad*antage is that the tis "!rn o after a eriod of ti+e and a n!isance tri of the f!rnace will res!lt. #n !ltra=*iolet (SH) -a+e detector is ro"a"ly the +ost o!lar on rocess heaters. t is, howe*er, a rather co+le de*ice that re1!ires certain reca!tions for relia"le oeration. SH detectors are otical de*ices. # lens in front +!st "e ai+ed directly at the -a+e. Deending on the artic!lar arrange+ent !sed, it +ay or +ay not "e desira"le for the detector to 6see6 only the ilot -a+e or the +ain -a+e as well. ne or +ore *iewing windows, /;=4@# to T, +!st "e ro*ided "y the f!rnace fa"ricator to assist in aligning the detector. SH detectors are +o!nted on "all swi*els to er+it acc!rate align+ent. #n instr!+ent air !rge co+lete with a rota+eter and a needle *al*e sho!ld "e connected to the t!"e "etween the lens and the -a+e in order to re*ent d!st fro+ acc!+!lating on the lens and to cool it. o+e !nits also re1!ire a s!ly of cooling water. FUEL GAS LINE INSTRU:ENTS 9C;6,+ 'he +ain gas reg!lator is set to the +ai+!+ allowa"le f!el gas ress!re for the
+ain "!rner. 9I6,+ E*ery reg!lator re1!ires a ress!re ga!ge. 'his ga!ge is also !sed to adG!st the setoints of PJJ=22 and P::=2&. 9SLL6,, 'he f!el gas low ress!re switch re*ents the oerator fro+ atte+ting ignition when there is ins!Rcient ress!re to co+lete it. 'he !strea+ location is so that it is not necessary to "yass its f!nction d!ring the ignition se1!ence. /;6,< 'he rst f!el gas safety sh!to *al*e is one of the 6safety sh!to *al*es6. t +!st "e esecially certied for red heater !se. 'he rst of the three *al*es in the +ain gas train has a +an!al reset and a solenoid that f!nction as follows = When the solenoid is de=energied the *al*e is sh!t and the +an!al reset has no f!nction whatsoe*er. = When the solenoid is energied, the *al*e re+ains sh!t !ntil the +an!al reset is lifted. = Jifting the +an!al reset oens the *al*e. 'he *al*e then re+ains oen as long as the solenoid is energied. # *al*e with this f!nction is so+eti+es ter+ed a 6free handle6 *al*e. nce a tri has occ!rred, the +an!al reset on the rst safety sh !to *al*e will re*ent the "!rner fro+ reigniting. 'his is an etre+ely i+ortant safety feat!re. E*en if there is ins!Rcient ress!re to +aintain a -a+e, !n"!rned gas +ay still collect in the re"o and ignite with elosi*e force once a large 1!antity has acc!+!lated =SC0O6,< n addition, /H=24 re1!ires a li+it switch to ro*e that it is sh!t d!ring the !rge and ignition hases. te+ UC=$4, a"o*e, details the re1!ire+ents of fail=safe li+it switch arrange+ents. /;6,- 'he f!el gas *ent *al*e has si+ilar criteria to those of /H=$@. 'he sa+e siing ta"le alies. =SC0O6,- 'hese li+it switches are otional, deending on the sie of the *ent *al*e. /;6,3 'he second +ain gas safety sh!to *al*e is a little dierent fro+ the rst in that it does not re1!ire a +an!al reset. 'he a!to+atic reset fe at!re is acco+lished "y a s+all electric +otor that oens the *al*e as soon as it is energied. 'he *al*e fails closed !on de= energiation. #!to+atic reset *al*es are identied "y the little 6M6 instead of the 66 in the sy+"ols on Fig!re @=2. 'he *al*e +ay incororate a slow oening feat!re. Power to the reset +otor sho!ld "e disconnected as soon as the *al*e is oen so that a +o+entary ower loss to the solenoid does not ca!se the *al*e to cycle. =SC0O6,3 /;6,3 sho!ld also ha*e li+it switches whose stat!s is incororated into the /M logic. 9SHH6,@ 'he f!el gas high ress!re switch sh!ts down the +ain f!rnace when the f!el gas ress!re eceeds the +ai+!+ allowa"le for a sta"le -a+e. Ecessi*ely high ress!re can "low the -a+e o!t only to ha*e it reignite, erhas elosi*ely, higher ! in the f!rnace. 'here is a trend toward the !se of analog trans+itters instead of switches for sensing rocess *al!es, e*en for sh!tdown !roses. 'he tri *al!e is then rogra++ed into the control syste+ instead of "eing adG!sted at the switch. 'rans+itters ha*e "eco+e +ore relia"le than switches and also ro*ide +!ch +ore infor+ation$0 . f this is the olicy at yo!r installation, the f!nctions of PJJ=22 and P::=2& can "e co+"ined in a single trans+itter located at PJJ=22. # second trans+itter can then "e !sed to ro*ide re=alar+s or an a!to+atic red!ndancy sche+e. F;62 Most large ind!strial f!rnaces ha*e a gas -ow control *al*e to +od!late the heat in!t. Details of the control syste+ are disc!ssed in the sections on f!el control. 'he gas -ow control *al*e +!st "e fail=closed "!t +!st also ha*e so+e +eans of ens!ring that it does not sh!t too closely to the seat or an !nsta"le -a+e will res!lt. #roi+ately 3@I of f!ll -ow is a tyical +ini+!+. o+e engineers rogra+ a +ini+!+ o!t!t into the control syste+. thers weld a s+all sto onto the shaft. oftware "loc9s can "e e asily altered and welded stos +ay not "e re=installed when the *al*e is relaced. !se "oth. 'he software li+it has the ad*antage that it can "e raised if the original setting is too low. t sho!ld "e set slightly higher than the welded sto in order to inhi"it reset wind! "efore the 0I controller o!t!t is reached. F!rnaces with cyclical ser*ice, s!ch as those !sed in dryer regeneration ser*ice, +ay ha*e
the +ain -a+e t!rned on and o on a reg!lar "asis. n s!ch cases, there cannot "e a +ini+!+ sto on the control *al*e. nstead, logic sho!ld "e ro*ided that sh!ts the *al*e f!lly whene*er the controller tries to osition it "elow the +ini+!+ -a+e setting. 'he controller +!st also "e switched to +an!al to re*ent reset wind! whene*er this is done. =SC62 'he gas -ow control *al*e re1!ires a li+it switch to conr+ that it is at the +ini+!+ -ow osition "efore lighto. 'he switch sho!ld "e set slightly a"o*e the +ini+!+ sto. 9I6,> # ress!re ga!ge sho!ld "e laced last "efore the f!el nole itself to ro*ide assistance in diagnosing +ysterio!s ro"le+s. Dirt does collect in *al*es and other laces5 /SLL6,? 'he +ain -a+e detector is generally of the co+"ination !ltra=*iolet and infra=red (SHL) tye. 'he sa+e re1!ire+ents aly as to the SH detector for the ilot. /!rners in cyclical ser*ice do not re1!ire a +ain -a+e detector. 'hey rely entirely on the ilot -a+e detector, which +ay "e adG!sted so that it can see "oth -a+es. AIR SU99LY INSTRU:ENTS. 'here is considera"le *ariation in the air s!ly syste+ a+ong f!rnaces of dierent sies. +all heaters, and so+e not so s+all, +ay "e entirely nat!ral draft and are controlled only with a +an!ally set air da+er. For f!rnaces with forced draft (FD) fans so+e, or all, of the following instr!+ents are re1!ired. 9DSH62+ f the air inta9e has a lter, it +!st "e tted with a dierential ress!re switch or
trans+itter connected to an alar+ to warn of l!gging. FE62, f the FD fan is large, a -ow +eas!re+ent de*ice is needed at the inta9e to +eas!re -ow thro!gh the fan so that s!rge can "e re*ented. # si+le a*eraging itot t!"e (#nn!"ar) or a ther+al de*ice ro*ides s!Rcient acc!racy witho!t signicant loss of head. F;62, Jarge fans re1!ire a +ini+!+ -ow "low=o *al*e to re*ent s!rge. FE622 'he act!al air -ow to the "!rner +!st "e +eas!red if an acc!rate f!elair ratio is to "e +aintained. 'his is generally done for large f!rnaces. +aller ac9aged !nits ha*e the f!el and the air lin9ed with ca+s and no air +eas!re+ent is done. # re*io!s article$, elains why it is not ossi"le to !se sa+e -ow ele+ent for "oth +ini+!+ -ow and thro!gh!t -ow control. t +ay "e re1!ired to !se a *ent!ri to achie*e "oth high acc!racy and low head loss. n *ery large "!rners, arrays of itot t!"es or ther+al -ow +eters distri"!ted across the d!ct, are the referred +ethod. :owe*er, these can "e so+ewhat ti+e=cons!+ing to install and +aintain. 'he associated trans+itter is also !sed to ro*ide the re=alar+ for FJJ=3A. F;622 # fail=oen co+"!stion air control *al*e is laced at the inlet to the wind "o for those "!rners re1!iring eternal air control. # "!tter-y *al*e is a o!lar choice. Hery large fans +ay ha*e ro*ision for inlet g!ide *anes which ro*ide a +ore eRcient +e ans of doing the sa+e thing. 'he *al*e re1!ires a +ini+!+ sto to re*ent co+lete clos!re. =SO622 F;622 has a li+it switch to conr+ that it is f!lly oen d!ring !rging. ;SHH62< Jarge fans sho!ld "e ro*ided with a seis+ic tye *i"ration switch. # re*io!s article3 disc!sses these de*ices. 'he switch sho!ld "e connected to sh!t down the entire syste+ as the heater cannot "e oerated witho!t the fan. Hery large fans +ay e*en incl!de an entire "earing *i"ration +onitoring ac9age. YS62< 'he FD fan +otor stat!s contact is !sed to ro*ide infor+ation to the /M. FSH62- 'he !rge air switch is !sed to indicate ade1!ate !rge -ow. t is !s!ally a dierential ress!re switch across the wind "o, if any, of the f!rnace or it +ay "e a signal fro+ the -ow trans+itter, F'=33. #s can "e seen in 'a"le @=$, there is a *ariety of oinions concerning +ini+!+ !rge re1!ire+ents. ecication #ME ection H $$ #P(LP @@0 V Part 4) $2 >FP# ?@0$@ C;# /$4%.3%
'i+e @ +in!tes
Hol!+e Flow @ *ol!+es 2@I f!ll -ow @ +in!tes @ *ol!+es 2@I f!ll -ow not secied ? *ol!+es &0I f!ll -ow not secied 4 *ol!+es A0I f!ll
-ow Table -6+ Mini+!+ P!rge Le1!ire+ents FSLL623 'he co+"!stion air low -ow sh!tdown switch sh!ts down the "!rner if there is ins!Rcient -ow for safe co+"!stion. FURNACE INSTRU:ENTS TE6<+ ne or +ore s9in ther+oco!les sho!ld "e ro*ided on the t!"es of the f!rnace. 'he
ther+oco!le ele+ent is welded directly to the t!"es. >ote that the de*ice that recei*es the 'C signal +!st "e caa"le of acceting gro!nded signals. TSHH6<, :igh stac9 gas te+erat!re is ca!se for sh!tdown. 'he f!rnace sho!ld also "e isolated fro+ the rocess feed if the feed is -a++a"le. '!"e r!t!re +ay "e the ca!se for the high stac9 te+erat!re. AE6<2 For a "!rner to oerate at ea9 eRciency, stac9 oygen and co+"!sti"les +!st "e +eas!red and controlled. :owe*er, the additional cost and +aintenance of stac9 analyers li+it their !se to large "!rners. :igh and low alar+s s ho!ld "e rogra++ed. 9I6<< # draft ga!ge +!st "e ro*ided to gi*e a grade le*el indication of the ress!re in the f!rnace. 'his is a secial tye of ress!re ga!ge sensiti*e to inches of water ress!re. /G6<- A to FG6<- A to Hiewing windows were +entioned re*io!sly with resect to aligning the
-a+e detectors. ther windows +ay "e needed as well. 'hese are !s!ally +ade of glass and re1!ire a s+all air !rge to 9ee the+ cool and clean. Leg!lators are often !sed control the air -ow. 'his is rather ointless since it is -ow control that is re1!ired. 'he resence of ress!re +ay only +ean that the line is l!gged. 'he a"sence of ress!re +ay +ean that the !rge is -owing freely at a rate "eyond the caacity of the reg!lator. Press!re alone ro*es nothing. 'he aroriate instr!+ent is a rota+eter with a needle *al*e. $@ CF: is a co++on -ow setting. 'he rota+eter ro*ides roof that !rge is act!ally occ!rring. # n!+"er of !rge rota+eters can "e +o!nted on a late and re=t!"ed to a s+all instr!+ent air header. 'his greatly red!ces eld la"o!r and ro*ides a con*enient way of chec9ing all !rge rates fro+ a single location. Lecent 'hreads "y Esa+ • •
Crude oil specifcation
•
Article: ISA Instrumentation Codes in...
•
Article: X YZY[Y\ - ]^_` b_`
•
Lely
Article: Crude oil specifcation
Y_` _` jk m Degassing Stations
Lely With p!ote
/log this Post 'han9s
0?=0?=20$$ 034? #M B ADS
Spons Cir!&it a8ertise"ent qoin Date #lways Posts Many
0?=0?=20$$, 03@2 #M 2
Esa" •
Hiew Prole
•
Hiew For!+ Posts Pri*ate Message Hiew /log Entries Hiew #rticles
• • •
#dd as Contact
•
q!nior Me+"er qoin Date May 200& Posts 232 /log Entries $% 'han9s 4 'han9ed 32 'i+es in 2$ Posts 9ROCESS RELATED SAFETY INSTRU:ENTS. 'he details of rocess related safety
instr!+entation deend *ery +!ch on the indi*id!al rocess. t is hard to generalie. 'he +aGority of f!rnace incidents occ!r d!ring light=o. 'herefore, it is not necessarily a good idea to add e*ery a!to+atic sh!tdown f!nction that +ight s!ggest itself. 'he safety *al!e of each interloc9 +!st "e weighed against the ris9 of !nnecessary sh!tdown and relight cycles. #t least one n!isance tri is "o!nd to occ!r d!ring the life of the e1!i+ent for e*ery sh!tdown de*ice installed. F!rnace isolation *al*es are esecially ro"le+atical.
;6-+ t may so!nd desira"le to a!to+atically sh!t isolation *al*es if it a ears that a t!"e
r!t!re has occ!rred. :owe*er, re+e+"er that any clos!re of an isolation *al*e d!ring oeration res!lts in loss of -ow. F!rther interloc9s +!st "e ro*ided to sh!t down the "!rner if this occ!rs. E*en after the sh!tdown, radiation fro+ the refractory contin!es to heat the traed -!id. /oiling +ay res!lt in o*er=heated t!"es. f the -!id is s!sceti"le to co9ing, t!"ing da+age +ay acc!+!late e*en after the i++ediate incident is ast. f caref!l analysis shows that an isolation *al*e wo!ld ro*ide a net contri"!tion to safety, the *al*e +!st "e installed correctly. Details of TH=@$ are shown in Fig!re @=4. 'he *al*e +!st "e a re rated *al*e. 'hat is, it +!st "e caa"le of +aintaining a seal in the resence of re. #P PEC AF#$3 and AFC$4 o!tline the re1!ire+ents. 'he wiring for "oth solenoids and for the associated li+it switches sho!ld also "e reroof. Mineral ins!lated (M) ca"le is an aroriate choice. ;6-+ sho!ld "e located at least @0 feet ($@ +eters) fro+ the f!rnace. Pro*ision +!st "e +ade to +an!ally initiate *al*e clos!re fro+ a location at least another @0 feet fro+ the *al*e itself. 'he DC console in the +ain control roo+ is an ideal location. YO0C6-+ f the rocess -!id is s!Rciently haardo!s that isolation *al*es are considered necessary, a diRc!lt iss!e +!st "e addressed What sho!ld the fail!re +ode of the *al*es "e f the *al*e fails to sh!t d!ring an e+ergency, a haardo!s sit!ation eists. f the *al*e
fails closed d!e to a ower or instr!+ent fail!re, other haards occ!r. ne sol!tion to this ro"le+ is to !se do!"le=acting, fail=last (=loc9ed) *al*e act!ators. 'hese sho!ld "e installed with two searate solenoids T7 to oen the *al*e and T7C to sh!t it. f this is done, two fail!res are re1!ired to ca!se an !ndesired *al*e action. For ea+le, if the *al*e is re1!ired to "e in the oen state, T7 will "e energied. # signal fail!re res!lts in "oth solenoids "eing de=energied. :owe*er, the *al*e will not +o*e since air ress!re on the iston is still "alanced. t re1!ires energiation of T7C "efore anything haens. # *al*e that has "een sh!t "eca!se of a re will not re=oen d!e to "!rned o!t wiring. ncidentally, do!"le=acting act!ators are considera"ly less eensi*e than sring=ret!rn act!ators. =SC0O6-+ Ji+it switches ? are ro*ided on the *al*e to ens!re that it is f!lly oen "efore f!rnace ignition is er+itted, and f!lly sh!t when necessary. ;6-, f the f!rnace contains a large in*entory of -a++a"le -!id, a deress!riation *al*e +ay "e re1!ired. t sho!ld "e caa"le of d!+ing the entire contents of the f!rnace to the -are. # chec9 *al*e sho!ld "e added at the o!tlet of the f!rnace so that the "lowdown *al*e does not atte+t to deress!re the entire downstrea+ rocess. YO0C6-, 'he ro"le+s associated with a fail=oen deress!riation *al*e +a9e it *ery !nli9ely that it wo!ld "e acceta"le to oerations. #ny ower or air fail!re wo!ld res!lt in +assi*e -aring. 'h!s a fail=last *al*e act!ator arranged in s a+e way as TH=@$ is the +ost reasona"le choice. =SC0O6-, Ji+it switches on TH=@2 sho!ld "e interloc9ed to re*ent oening TH=@$ when the deress!riation *al*e is oen. FSLL6-2 Meas!ring the rocess o!tlet -ow is an ecellent safety feat!re. Le+e+"er that a rocess f!rnace is essentially a heat echanger. f there is no rocess -ow, the f!rnace t!"es will o*erheat and +ay co9e or e*en r!t!re. 'he -!e gas will o*erheat !er section of the f!rnace and the stac9. >o f!rnace can "e allowed to oerate !nder these conditions. f the inlet rocess -ow, FE=33, is high and the o!tlet is low, it can only +ean that a t!"e has r!t!red. ++ediate sh!tdown is i+erati*e. TSHH6-2 :igh rocess o!tlet te+erat!re +eans that the te+erat!re controls are not f!nctioning roerly 'he f!rnace +!st "e sh!t down. 'he note !nder P::=2& concerning the !se of trans+itters instead of switches also alies here. Modern rogra++a"le logic controllers (PJCs) and distri"!ted control syste+s (DCes) are 1!ite caa"le of interreting ther+oco!le and L'D signals directly. Consideration sho!ld "e gi*en to !sing the+ instead of the less relia"le te+erat!re switches or the co+"ination of ther+oco!le and +illi*olt tri relay $0 . 9S;6-< n all ro"a"ility there will "e "loc9 *al*es so+ewhere on either side of the heater so that it can "e isolated +an!ally. f these eist, a relief *al*e +!st "e ro*ided. SU9ER;ISED :ANUAL. 'he >FP# standards @, A secify se*eral tyes of "!rner +anage+ent syste+s • o o o
#!to+atic (Lecycling) #!to+atic (>onrecycling) !er*ised Man!al.
'he tye +ost co++only !sed in ind!strial rocess f!rnaces is s!er*ised +an!al, often called se+ia!to+atic. >FP# descri"es it "est S&per8ise :an&al. A s*ste" b* $hi!h the f&rna!e is p&r)e an a b&rner starte% i)nite% an stoppe "an&all*% $ith interlo!(s to ens&re that the operation follo$s proper pro!e&res.
'hat this +eans is that the oerator +!st "e resent at the "!rner to control the ignition roced!re. #t *ario!s 9ey oints he !shes a "!tton or ta9es so+e other action to ad*ance to the net ste. 'he /M s!er*ises the oerator and conr+s that all conditions are aroriate for that ste "efore allowing it to roceed. f conditions are wrong, the /M sh!ts down the entire se1!ence and the oerator +!st rectify the ro"le+ and start all o*er
again. t8s a little li9e laying na9es and Jadders ecet that there is no 1!ic9 way to the to. # detailed ea+le of the ignition se1!ence for a f!rnace with an FD fan is gi*en "elow. 'he indented aragrahs +ar9ed 6• 6 indicate a!to+atic action. STE9 + Close the gas -ow control *al*e, FH=3, to its +ini+!+ sto. en the co+"!stion air *al*e, FH=33, f!ll. P!sh the tart Fan "!tton, :=34. ('he fan +ay "e started fro+ the +ain control anel or "y so+e other +eans.) • 'he green Fan L!nning light, 7J=34, co+es on. 9AUSE 'he se1!ence +ay wait indenitely at this oint. STE9 , witch the Ena"le switch, :=2, to >. ('his can also "e done "efore the fan is
started.) #ny other sh!tdowns, s!ch as fro+ the Main Control Panel or a lant e+ergency sh!tdown (ED) +!st "e cleared "efore this has any eect. • 'he green Leady light, TJ=2, co+es on if there are no acti*e sh!tdowns. • #!to+atic f!nctions are ena"led once the switch is >. Sntil that oint, no lights are on ecet the Fan L!nning light, and no !sh "!ttons f!nction ecet the Ja+ 'est. #ll "loc9 and *ent *al*es are de=energied.
9AUSE 'he se1!ence +ay wait indenitely at this oint. STE9 2 P!sh the Ja+ 'est "!tton, :=%.
#ll lights sho!ld co+e on. STE9 < P!sh the tart P!rge, :=3@, "!tton. • 'he following conditions +!st "e +et "efore the !rge se1!ence can "egin = /oth safety sh!to *al*es are conr+ed sh!t. = 'he gas -ow control *al*e is conr+ed on +ini+!+ sto. = 'he co+"!stion air control *al*e is conr+ed wide oen. = 'he fan is conr+ed r!nning (7=34). = 'he !rge -ow signal, F:=3@, is on. • nce these conditions are +et, a ti+er, KC=3@, is started. 'he ti+er is set so that a +ini+!+ of eight air changes of the f!rnace is ass!red. (ee 'a"le @=$, Mini+!+ P!rge Le1!ire+ents.) 'he green P!rge Co+lete light, FJ:=3@, -ashes while the ti+er is r!nning. •
#ny ti+e any of the conditions are not +et, the ti+er is reset and the light goes o. • nce the ti+er has co+leted the !rge inter*al, the green P!rge Co+lete light is on steady. 9AUSE 'he se1!ence +ay wait indenitely at this oint as long as the fan is r!nning. STE9 - P!sh the tart gnition "!tton, :=$4.
# twel*e second ti+er, KC=$4, is started. • 'he ignition transfor+er, /7=$?, is t!rned on. • 'he green Pilot Fla+e n light, /J:=$%, is -ashing. • 'he ilot *ent *al*e, /H=2@ is sh!t. • 'he rst ilot sh!to *al*e, /H=24, is oened. • 'he second ilot sh!to *al*e, /H=2A, is oened. • Pilot -a+e, /JJ=$%, +!st "e conr+ed for two seconds d!ring the twel*e second inter*al or the /M re*erts to te 4. nce the ilot -a+e, /JJ=$%, is conr+ed. • 'he ignition transfor+er, /7=$?, is t!rned o. • 'he green Pilot Fla+e n light, /J:=$%, is on steady. 9AUSE 'he se1!ence +ay wait indenitely at this oint as long as ilot -a+e is detected. •
STE9 3 P!sh the en Main F!el "!tton, :=24.
# second twel*e second ti+er, KC=24, is started and the green Main Fla+e n light, /J:= 2%, is -ashing. • 'he f!el gas *ent *al*e, /H=2@, is sh!t. • 'he rst f!el gas safety sh!to *al*e, /H=24, is energied. STE9 @ Jift the le*er on the rst f!el gas safety sh!to *al*e, /H=24, within ten seconds. 'he *al*e will then stay latched in the oen osition. t will not oen at all if conditions are not safe. • f this is not conr+ed "y U=24 within the allowed ten se conds, the se1!ence re*erts to te 4. f it is, the oen li+it switch will trigger the /M to oen the second f!el gas safety sh!to *al*e, /H=2A. • Main -a+e +!st "e conr+ed "y /JJ=2% for two seconds d!ring the twel*e second inter*al or the /M re*erts to te 4. • nce the Main -a+e is conr+ed, the green Main Fla+e n light, /J:=2%, is on steady. 9AUSE 'he f!rnace is now r!nning on +ini+!+ -a+e (also 9nown as 6low re6). 'he se1!ence +ay wait indenitely at this oint. •
STE9 > 'he oerator +ay oen the gas -ow control *al*e at any ti+e. E:ERGENCY SHUTDO4N. Hario!s tro!"le indications will trigger a artial or total f!rnace
sh!tdown. 'hey are ta"!lated in 'a"le @=2. 'hese sh!tdowns, ecet the two -a+e fail!res, are acti*e at all ti+es, incl!ding d!ring the ignition se1!ence. 'he -a+e detectors are "yassed d!ring the trial for ignition eriod. /!rner E D ca!se H::=34 fan :=2 +an!al FJJ=3A air /JJ=$% ilot /JJ=2% +ain PJJ=22 f!el P::=2& f!el '::=4$ t!"e
Main ;as /H=242@2A De=energie De=energie De=energie De=energie De=energie De=energie De=energie De=energie
Pilot ;as FD Man!al restart /H=$4$@$A Fan at De=energie to te $ De=energie De=energie te 4 De=energie
te A
'::=43 stac9 De=energie FJJ=@3 rocess De=energie '::=@3 rocess De=energie Table -6, F!rnace ED Key
ther, rocess related ED f!nctions s!ch as isolation and deress!riation of the f!rnace, +ay also "e i+le+ented. # thoro!gh analysis of the entire rocess is re1!ired to deter+ine what else, if anything, +ay "e re1!ired. LOCAL CONTROL 9ANELS. Fig!re @=@, 'he /asic Jocal Control Panel, shows the +ini+!+ stat!s lights, !sh "!ttons and switches re1!ired at the "!rner. 'his +ini+!+ ass!+es that all alar+s and sh!tdowns are dislayed on a +ain control console, erhas a DC, so+ewhere else in the lant. 'he local oerator and the console oerator +!st "e in close co++!nication d!ring the light=o se1!ence. #t other ti+es the "!rner is !nattended. Many "!rners are s!lied as self=contained ac9ages. For s!ch installations, a +ore co+lete local anel is re1!ired. E*ery re=alar+, sh!tdown alar+ and *al*e stat!s is dislayed as a searate indicating light on the anel. n general, a good colo!r code is Pre=#lar+ range h!tdown #lar+ Led Leady Jight ;reen Hal*e en ;reen Hal*e h!t Led #*oid the !se of "l!e indicating lights as they are ractically in*isi"le in "right daylight. #lar+ logic +!st "e ro*ided to dri*e the alar+ indicators and a horn according to the standard # #!to+atic Leset se1!ence #. 'he First !t and Ja+ 'est feat!res (# se1!ence$@ F3#=$4) are often re1!ested 'hese can "e acco+lished "y dedicated alar+ ann!nciator logic or "y a PJC. Jogic and control hardware +ay "e ho!sed within the local anel or elsewhere in the lant. FEED/AC FUEL CONTROL. 'he si+lest co+"!stion control arrange+ent is to ha*e a te+erat!re controller ('C) on the rocess o!tlet controlling the f!el gas ring rate. 'his syste+ is essentially li9e the +edi!+ side, inlet throttling, arrange+ent descri"ed 6Controlling hell and '!"e Echangers6 3. 'he 'C +ay dri*e the f!el *al*e directly, in which case the *al*e is tagged 6'H6, or it +ay cascade to a -ow controller (FC) on the f!el gas. FEEDFOR4ARD FUEL CONTROL. n theory, feed"ac9 es e*erything. n ractice, there are ro"le+s. Fired heaters ha*e a considera"le resonse lag to changes in either the ring rate or rocess inlet conditions. 'here will "e transient dis or "!+s whene*er a signicant change occ!rs on any rocess, f!el, or air *aria"le. f de*iation fro+ setoint is the disease, then feed="ac9 is the c!re. /!t an o!nce (2? gra+s) of re*ention is worth a o!nd (4@4 gra+s) of c!re. Feedforward is the techni1!e of re*enting the ro"le+ "efore it occ!rs. f there is signicant *ariation in feed rate, feed forward can "e alied as is descri"ed in 3 6Controlling hell and '!"e Echangers6 . i+ilar +ethods can "e !sed to co+ensate for raid *ariations in feed te+erat!re.
Direct control of the f!el *al*e is not eno!gh for any "!t the s+allest "!rners. 'he f!el +!st "e controlled to ro*ide the eact a+o!nt of energy that the rocess re1!ires. 'his i+lies that the heat content of the f!el as well as its -ow rate +!st "e 9nown. f the f!el ress!re, te+erat!re and co+osition are constant, a si+le -ow +eter is s!Rcient. f signicant *ariations are eected, +ore co+le feedforward techni1!es can "e alied. Fig!re @=A shows the 6f!ll +eal deal6. t is seldo+ necessary to incl!de all this instr!+entation. # chro+atograh +eas!res the co+osition. Fro+ this the density, heating *al!e, and oygen de+and can "e calc!lated. Co+"ining these calc!lations with the P', 'E and F' readings res!lts in an eact g!re for heat in!t er !nit ti+e. 'he res!lt of this arrange+ent is that any changes in f!el gas condition will "e corrected "efore any change in the rocess te+erat!re occ!rs. n the ast, a calori+eter has "een !sed instead of a chro+atograh. t was secially designed to rod!ce a single g!re called the 6Wo""e nde6. ( don8t 9now if this rhy+es with 6lo""y6 or 6'o"y6.) 'his inde is dened as /t!scf(secic gra*ity) $2 . 'he s1!are root in the deno+inator a!to+atically acco!nts for the eects of +olec!lar weight changes on an orice late. # calori+eter with s!ch an o!t!t is fre1!ently called a 6Wo""e +eter6. Wo""e +eters f!nction "y act!ally "!rning a s+all a+o!nt of f!el that has "een +etered thro!gh a ed restriction orice. n this +anner the eect of changing density on the -ow +eter and the control *al*e is +i+ic9ed. 'hey ser*e *ery well with +ied hydrocar"on f!el gas. 'he ass!+tion is that there is a roortional relationshi "etween heat content and oygen de+and. 'his ass!+tion is reasona"ly acc!rate for hydrocar"ons "!t is not relia"le if there are large, *aria"le roortions of hydrogen in the f!el. n s!ch cases, +ore co+le analyers are needed to ro*ide an eact oygen de+and g!re which can "e !sed to a dG!st the ratio of the air controller, FFC=33. AIR CONTROL. #ir +!st "e s!lied to the "!rner in roortion to the f!el. n s+all, si+le !nits this is acco+lished "y ha*ing the two *al*es lin9ed "y a ca+. 'he ratio "etween the two -ow rates is set "y +echanically adG!sting the ca+. e*eral signicant ass!+tions are +ade when this arrange+ent is !sed • 'he ress!re of "oth the f!el and the air is constant. • 'he te+erat!re of "oth the f!el and the air is constant. >ote that a %0F° (@0C° ) swing in te+erat!re res!lts in an $?I *ariation in density. /eca!se of the s1!are root relationshi "etween density and the -ow rate, this will res!lt in an error of %I in the +eas!re+ent and hence a $.?I change in ecess oygen. • 'he co+osition of the f!el, and hence its air de+and, is constant. f these conditions are not +et, the f!elair ratio will *ary. # f!el=rich +it!re res!lts in serio!s waste of f!el and ossi"ly a haardo!s condition. #n air=rich +it!re also res!lts in ineRciency d!e the *ol!+e of air that +!st "e heated and "lown o!t the stac9 while red!cing heat transfer.
>at!ral draft "!rners +ay ha*e a set of act!ator=dri*en lo!*res controlling co+"!stion air. #cc!rate -ow +eas!re+ent +ay not "e ossi"le for s!ch installations. 'he correct lo!*re osition can "e controlled "y sending it a signal that is roortional to the signal sent to the f!el gas *al*e. Eact adG!st+ent of the ratio can only "e done on=line with the assistance of the stac9 2 analyer. LEAD0LAG CO:/USTION CONTROL. n order to control the f!elair ratio on a contin!o!s "asis thro!gh all load changes, a -ow ratio controller is !sed. 'his in*ol*es +eas!ring "oth the f!el and the air. Fig!re @=& shows the co+lete arrange+ent< it has a n!+"er of secialied feat!res. Firstly, the total heat de+and is deter+ined "y the load controller, TC= %%, resonding to the de+ands of the rocess. Most ro"a"ly it is a te+erat!re controller on the o!tlet. 'he TC is cascaded to "oth the f!el and air controllers. ince these two controllers are +eas!ring dierent -ows, the signal fro+ the TC +!st "e in !nits that "oth controllers can !nderstand. 'he +ost con*enient set of !nits is to ha*e the 0 to $00I signal e1!al to the san of the f!el gas controller, either in CFM, /'S+in or so+e other aroriate !nit. o+e DCs ha*e the feat!res to do this a!to+atically. gnoring, for a +o+ent, the two selectors F7=3/ and F7=33/, the TC=%% signal "eco+es the setoint of "oth the f!el controller, FC=3, and the air controller, FFC=33. 'he latter is cong!red as a ratio controller which has the aroriate airf!el ratio "!ilt in. /e caref!l here o+eti+es, for reasons of tradition, f!el is +eas!red in CFM and air is +eas!red in l"+in. Ens!re that the ratio rogra++ed into the control syste+ ta9es s!ch ec!liarities into acco!nt. nce the f!rnace is in oeration, the ratio +ay "e adG!sted with the hel of the -!e gas analyer.
o far, so good. /!t this arrange+ent does not allow for either raid transients in the f!el de+and nor the ossi"ility that either the f!el or the air *al*es +ay not "e resonding correctly. 'his is where the two li+iters co+e in. 'hey are arranged in what is co++only called a 6leadlag6 arrange+ent. (>ote that the eression 6leadlag6 has at least three totally !nrelated +eanings in rocess control 'he rst refers to a secic for+ of dyna+ic signal conditioner< the second to a two=stage !+ or co+ressor control< the third alies to "!rners and is elained here.) o+e call this arrange+ent a 6cross li+iter6. 'he !rose of the high selector on the setoint of the air -ow is to ass the greater of the rocess dri*en air de+and and the act!al f!el r ate. n a rising de+and signal, the air will lead the f!el since it resonds i++ediately and does not wait for the f!el to rise. n a falling de+and signal, the air setoint will not fall !ntil the f!el, as +eas!red "y F'=3, has fallen. 'h!s the air will lag the f!el on the way down. /oth the f!el and the air controllers ha*e their setoints "ased on the sa+e f!el rate. n the e*ent that the f!el *al*e oens witho!t the setoint to the air controller rising a corresonding a+o!nt, the increased f!el rate will ass thro!gh the high selector to increase the setoint to the air controller to agree with the act!al f!el inGected into the "!rner. 'he !rose of the low selector on the setoint of the f!el controller is to re*ent the f!el -ow fro+ rising "efore the air -ow has risen. 'he f!el lags the air on the way ! and leads on the way down. f the air s!ly sho!ld fail, the f!el setoint will fall with the air, regardless of the de+ands of TC=%%. ne detail re+ains FC=3 is not a ratio controller. ts setoint is dened in ter+s of the f!el -ow rate. /oth in!ts to the selector F7=3/ +!st ha*e identical !nits and range. F7=33C is a si+le scaling f!nction with the scale constant 6K6 eactly e1!al to the recirocal of the airf!el ratio of FFC=33. n a DC it is ossi"le to cong!re this f!nction "loc9 to a!to+atically read the *al!e of the r atio fro+ FFC=33 and aly it directly in the for+!la. 'his is i+ortant "eca!se the *al!e of the ratio +ay "e changed in resonse to changes in the f!el gas co+osition or the co+"!stion eRciency. t is easy to o*erloo9 the K factor when +a9ing this adG!st+ent. FLUE GAS :EASURE:ENT. Most +odern rocess f!rnaces, e*en relati*ely s+all ones, ha*e so+e for+ of contin!o!s ecess air +eas!re+ent in the -!e. 'his is i+ortant for "oth eRciency and safety. n the ast, s!ch instr!+ents ha*e had a well deser*ed re!tation for high +aintenance and low relia"ility. 'his is no longer the case. 'oday8s instr!+ents are "ased on the electroche+ical resonse of irconi!+ oide (Ur 2), also 9nown as irconia. # ro"e is inserted directly into the -!e and does not re1!ire any sa+ling syste+. #n elanation of the +eaning of ecess air is in order. Ecess air is co+"!stion air in ecess of the theoretical, or stoichio+etric, air. 'his is aroi+ately, "!t not eactly, the sa+e as what is +eas!red "y the stac9 2 analyer. 'he dierence is that there will "e at least so+e oygen in the stac9 that co!ld ha*e "!rned "!t did not "eca!se of oor co+"!stion conditions. 'his i+lies that it is ossi"le to ha*e so+e 2 as well as !n"!rned f!el going ! the -!e. 'his condition cannot "e c!red "y instr!+entation. #ir is a"o!t 2$ *ol!+e ercent oygen. 'his +eans that a $0I ecess 2 reading i+lies that only half the oygen is "eing cons!+ed and that twice as +!ch air is "eing "lown into the f!rnace, heated !, and sent ! the stac9, as is necessary for co+"!stion. 'hat wastes a lot of energy and red!ces the caacity of the f!rnace. ince the cost of !n"!rned f!el is higher that than of ecess air, a tyical oti+!+ is aroi+ately 2I ecess 2 reresenting aroi+ately $0I ecess air. >at!ral draft f!rnaces witho!t contin!o!s air control sho!ld "e oerated at aroi+ately AI ecess 2. 'his ens!res co+lete co+"!stion !nder !set conditions. Jarge f!rnaces s!ch as ower "oilers, where the cost of ineRciency is high, often incl!de a co+"!sti"les analyer in the sa+e !nit as the ecess 2 analyer. 'hese de*ices can also +a9e in=sit! +eas!re+ents witho!t sa+ling. 'hey !se an infrared "ea+ directed across the stac9 to +eas!re the car"on +onoide (C) content. Ecess air le*els "elow $I gi*e no !sef!l indication of co+"!stion eRciency as there is no longer any correlation "etween the a+o!nt of resid!al oygen and the a+o!nt of !n"!rned f!el going ! the -!e. # co+"!sti"les analyer can ro*ide that infor+ation. 'hey are eensi*e, howe*er, and +!st
"e G!stied on a cost"enet "asis. o+eti+es the -!e gas analyer is !sed to a!to+atically adG!st the airf!el ratio of F FC=33. ;reat care sho!ld "e eercised in ta9ing this o"*io!s=loo9ing ste. Desite their considera"le i+ro*e+ent in ter+s of acc!racy and relia"ility, -!e gas analyers are still co+le instr!+ents and do not aroach the relia"ility of -ow trans+itters. 'his ste sho!ld only "e ta9en if the "enet G!sties the necessary co++it+ent to +aintenance. E*en then, high and low li+its +!st "e laced on the signal going to the ratio controller so that fail!res or inacc!racies cannot dri*e the airf!el ratio to dangero!s etre+es. /URNER CONTROL an :ANAGE:ENT SYSTE:S. >FP#@, A denes three distinct control syste+s /oiler Control S*ste". The )ro&p of !ontrol s*ste"s that re)&lates the boiler pro!ess in!l&in) the !o"b&stion !ontrol s*ste" b&t not the b&rner "ana)e"ent s*ste". Co"b&stion Control S*ste". The !ontrol s*ste" that re )&lates the f&rna!e f&el an air inp&ts to "aintain air0f&el ratio $ithin the li"its re&ire for !ontin&o&s !o"b&stion an stable a"e thro&)ho&t the operatin) ran)e of the boiler in a!!oran!e $ith e"an. This !ontrol s*ste" in!l&es the f&rna!e raft !ontrol $here appli!able. /&rner :ana)e"ent S*ste". The !ontrol s*ste" ei!ate t o boiler6f&rna!e safet*% operator assistan!e in the startin) an stoppin) of f&el preparation an b&rnin) e&ip"ent% an pre8ention of a"a)e to f&el preparation an b&rnin) e&ip"ent.
For a rocess heater, the *ario!s rocess related controls wo!ld "e e1!i*alent to the "oiler control syste+. Contrary to co++on "elief, neither PJCs nor DCs are +entioned in any of the >FP# ?@00 series. 'hey are not +entioned in the C;# Code either. What >FP# says is The lo)i! s*ste" perfor"in) the safet* f&n!tions for b&rner "ana)e"ent an boiler !ontrol s*ste"s shall not be !o"bine $ith an* other lo)i! or !ontrol s*ste".
Snfort!nately it is not at all clear what the word 6co+"ined6 +eans, nor is there any elanation of the word 6syste+6. 'here see+ to "e +any interretations, so+e of the+ 1!ite etre+e. ne ha*e heard is that 6>FP# re1!ires that the co+"!stion controls shall "e in the DC, the se1!enced start shall "e in a dedicated PJC and there shall "e hardwired sh!tdown relays.6 'he ro"le+ with this aroach is that there are too +any lin9s "etween these three syste+s. Each lin9 adds co+leity that red!ces the relia"ility of the whole. For ea+le, the se1!enced start re1!ires that the -a+e detectors +!st "e "yassed d!ring the trial for ignition eriod. nce the relays ha*e "yasses fro+ the PJC, they no longer ha*e any indeendent relia"ility. #nd "esides, who says that relays are safer than PJCs$0 'here are also lin9s "etween the PJC and the DC. 'he start se1!ence t!rns the f!el gas control *al*e down to its +ini+!+ sto. t then +ani!lates the air -ow for !rging. Whether these lin9s ha*e the for+ of signals assing "etween the PJC and the DC or they ta9e the for+ of solenoid *al*es interr!ting the air signals to the act!ators, co+leity is added that co+ro+ises the integrity of "oth the /M a nd the co+"!stion control syste+. Firstly, the start se1!ence and the safety interloc9s are one inseara"le control syste+. >or+al oeration is si+ly the nal stage of the start se1!ence. 'he +ain -a+e is now on, and all safety interloc9s are f!lly oerational. Progra++a"le control syste+s co++only !sed in today8s rocess ind!stries are "!ilt of large and owerf!l !nits that can handle logic and control f!nctions far +ore co+le than control of a single "!rner. # DC, in artic!lar, consists of +any indeendent f!nction "loc9s that are either lin9ed together or 9et searate according to the re1!ire+ents of the control strategy. 'here is no reason why all asects of a "!rner control cannot "e carried o!t within a single hysical controller. Certain reca!tions +!st "e ta9en to ens!re that the >FP# re1!ire+ent for searation is +et $. 'he /M +!st reside in a clearly identied "loc9 of logic f!nctions. 'his "loc9 +!st ha*e all in!ts and o!t!ts clearly identied within the "loc9. # DC that re1!ires this
identication as art of its rogra++ing aroach is acceta"le. # DC that re1!ires 6hardware6 addressing is not. For ea+le #ss!+e that Jogic /loc9 $3 is the /M. 'he state+ent 6et Legister 2$ in Jogic /loc9 $3 >6 aearing elsewhere in the DC is an ea+le of ris9y rogra++ing ractice. #n insection of /loc9 $3 wo!ld not re*eal the connection. 'his tye of state+ent is e*en +ore dangero!s if it is +ade in error in a totally !nrelated f!nction. # safe +ethod of acco+lishing the desired f!nction is to incl!de the state+ent 6f :=$2 is >, then set /H= 24 >6 within the "ody of /loc9 $3. #nyone ea+ining the /M can readily see the state+ent and G!dge its *alidity. !ch state+ents cannot accidentally "e incl!ded in /loc9 $3 "y so+eone wor9ing on a dierent "loc9. 2. !tgoing lin9s "etween the /M and other control f!nctions +!st always "e 6!shed6 "y the /M. For ea+le 'he /M forces the f!el gas *al*e sh!t and then !ts it "ac9 on a!to+atic at the aroriate ti+es. 'he f!el gas controller does not chec9 the /M to see what it sho!ld do. 3. nco+ing lin9s "etween the /M and other control f!nctions +!st always "e 6!lled6 "y the /M. For ea+le 'he /M chec9s the stat!s of a "yass switch to see if it sho!ld ignore a high discharge te+erat!re. 'he "yass switch does not force the /M. 4. >o !nrelated logic f!nction +ay reside in the sa+e "loc9 as a /M. t +!st always "e etre+ely clear to anyone +a9ing +odications whether or not they are in a /M "loc9. ince there +ay "e ti+es when it is necessary to disa"le so+e !nrelated f!nction, it +!st always "e ossi"le to do this witho!t disa"ling any art of the /M. @. n!t and o!t!t +od!les for /M related signals sho!ld not also handle other signals. A. # hardwired 69ill6 switch sho!ld "e incl!ded on the oerator8s console. 'his switch de= energies all gas safety sh!to *al*es. 'his can easily "e acco+lished "y c!tting the ower to the o!t!t +od!les. f r!le @ is followed this can easily "e acco+lished witho!t dist!r"ing !nrelated f!nctions of the DC. 'he 9ill switch sho!ld also ha*e a second contact wired to ro*ide in!t to the /M logic in the DC. 'he /M will then carry o!t an indeendent sh!tdown. n this way f!ll red!ndancy is acco+lished. n addition, the DC logs the !se of the 9ill switch. &. # DC control !nit that carries o!t any f!nctions "eyond the control of a single f!rnace +!st "e f!lly d!al red!ndant with a!to+atic transfer to an identical, on=line, "ac9! !nit. #ll co++!nications "etween the control !nit and the oerator8s console +!st "e f!lly d!al red!ndant with a!to transfer. 'he consoles the+sel*es +!st "e at least d!al red!ndant with no shared co+onents. t is +y own oinion that a /M can "e integrated into DC control !nit safely, relia"ly, and legally if the a"o*e r!les are followed. /&t... realie that this a co+le s!"Gect and that +any !sers ha*e had dierent eerience. wo!ld *ery +!ch welco+e feed"ac9 on this contro*ersial s!"Gect. /eca!se of the wide range of interretations of the f!nctional searation r!le, it is *ery wise to o!tline any roosed control syste+ in detail and ha*e it aro*ed "y the 6a!thority ha*ing G!risdiction6 "efore roceeding with i+le+entation. !nderstand it is not *ery leasant to ha*e a design ref!sed d!ring co++issioning. CO::ISSIONING. >o /!rner Manage+ent yste+ can "e certied safe !ntil it has "een eld tested. >FP# is 1!ite clear on this oint The safet* interlo!( s*ste" an prote!ti8e e8i!es shall be ointl* teste b* the or)aniation $ith the s*ste" esi)n responsibilit* an those $ho operate an "aintain s&!h s*ste"s &rin) the nor"al operatin) life of the plant. These tests shall be a!!o"plishe before initial operation. 'he roof of the !dding is in the eating. >o safety syste+ is safe !ntil it is e"onstrate
to "e safe. REFERENCES
$. Driedger, W. C., 6Controlling Centrif!gal P!+s6< Hydrocarbon Processing, q!ly $%%@. httwww.driedger.cace$cCE$CP.ht+l 2. Driedger, W. C., 6Controlling Positi*e Dislace+ent P!+s6< Hydrocarbon Processing, May
$%%A. httwww.driedger.cace2dCE2PDP.ht+l 3. Driedger, W. C., 6Controlling hell and '!"e Echangers6< Hydrocarbon Processing, March $%%?. httwww.driedger.cace3stCE3'T.ht+l 4. Driedger, W. C., 6Controlling tea+ :eaters6< Hydrocarbon Processing, >o*e+"er $%%A. httwww.driegger.caCE4:CE4:.ht+l @. >FP# ?@, /oiler and Co+"!stion yste+s :aards Code (For+erly ?@0$, ?@02, ?@#, ?@C). httcatalog.nfa.org
A. >FP# ?A, tandard for *ens and F!rnaces. httcatalog.nfa.org &. Driedger, W., Ji+it witches Key to Hal*e Lelia"ility< Intech, qan!ary $%%3. httwww.driedger.cali+itswJi+itw.ht+l ?. C# /$4%.3, Code for the Field #ro*al of F!el=Lelated Co+onents on #liances and E1!i+ent. httwww.csa=intl.orgonlinestorewelco+e.as %. ;r!hn, P. and Cheddie, :.J., afety h!tdown yste+s Design, #nalysis and q!stication, #. httwww.isa.org $0. #ME /oiler and Press!re Hessel Code, ection H, Care of Power /oilers. httwww.as+e.orgcatalog $$. #P LP @@A, Fired :eaters and tea+ ;enerators. httwww.cssinfo.co+aigate.ht+l $2. #P PEC AF#, ecication of Fire 'est for Hal*es. httwww.cssinfo.co+aigate.ht+l $3. #P PEC AFC, ecication of Fire 'est for Hal*es with electi*e /ac9seats. httwww.cssinfo.co+aigate.ht+l $4. # $?.$, #nn!nciator e1!ences and ecications. httwww.isa.org $@. # &&.4$, Fossil F!el Power Plant /oiler Co+"!stion Controls. httwww.isa.org $A. C# 3.%, #!to+atic afety h!t= ;as Hal*es. httwww.csa=intl.orgonlinestorewelco+e.as
Lead +ore httwww.eget.net*"threads4%%2?=Controlling=red=heatersi$1dGFdrS
