ML20197B166

Reviews RESAR-414.Reliability Being Degraded Due to Addition of Safety Sys.Dismayed at Inclusion of Feedwater Isolation Circuitry

ML20197B166
Person / Time
Site:	05000572
Issue date:	08/14/1978
From:	Epler E AFFILIATION NOT ASSIGNED
To:	Fraley R Advisory Committee on Reactor Safeguards
References
ACRS-CT-1030, NUDOCS 7811010251
Download: ML20197B166 (9)

Text

- - - - - - - - - -

oL.ogo I

,c,q uanu mu 3 m jyysxg13wg3ogsfpygg 213!.I AJC h/7 /O O 2

4tg 31oytov yggnn3 YGAll:0d' A 30:'I:'!t1133 ON 4Y Wf003*13NN3!533 C/9(0 3V210E SY1:30fVCF A'S' NT3 11 FSC066t gnil: f >l' fG49

/']' t.30 jaa ' yxa]n( v g]wfaj 4f'R k yentso>l ;oucuyyas 0a yav>popuja6no>pc #g 'G ? iu ons sou>>>sotiou a v i ;14 > tro s w o u piv c. iowwsupop fit 2y,tyl3nN1)CfO fI? dA()00A(faJ>f 1ll Uga)t( Cl1iL, fO OLU f? 0 QA( uaL rogspuvfruj !utd'o^awavf iv r4cfadu sajiv7rpp mttsu N ( /00000'"t '."'I nu't' 'noni'f uaap ro qa qin yyiyap' jarofo /vr l lUL )Q Adlltf A( ( ffldfdrl lb o f f fiD lOlll DAD)OU) dad (d b ( '"P t ti 'oup gjyye t '" a "'" h a v 00'* " 'l' '. 7 f09 fi}V 'f / /Q}(dlil ff50 AE }Cda llil fIf {(J l.ld} fCJ S dAl I O d$$U Ql ?.O Al*f A{ fl A fl}f {UIl ( f 0fjafO ,G ( 11 /C UAAUS fUElA)d' /t ) V T d' n /b " r l'v l' P fo j y.3 udowrusapnyou qmj tiv d vut, suv 90 ada/vfsP.1,,,! V U fV'al tf'"( pe i s' tua odaiv(rev ivf ysi preuu (ria edr!A trmt n ju unp uo,o Qf /llv {O/ U yafif104%

• 3oftt Cf tfl /U~f J4)rfJf Idt, f fL d YOS[I.fI[l.f

)fS)OlG }V (y)fL A(I O)$ ) //tlh)Qif/([1 ff.t y ddn}}M Q O odsuipious-l'f {$ O$$} } l} f-ffV Qnl )c O/lT {n ^ f f L f */ f ul )0 /10 .jq3 >sfr,vqptf ot druujs'ma utoL 9 rotua esdda orjnotiK abuFv fF" of )0('R 7'~(U' LV I S '"tl o"T l /*~ni? 'S*" L -tL-te tsifs> y aupa uo> tid </ 4 rM ut . &//OW f?b

/n -/he review of llESR R 4/4 Jhere appeitY luvlha/ indit.altons -lhaJ reliabili]y is lseing dei)vaded, in enq leflef Jo kerr and Shewinan Av] /4, /978, pages 9,/o nnd II, g Jhe erosion el reliabthlq c>l lhe sirnple redundon) pair',11 y os)I\\ .shown lo ha ve indireclly been brovghJ nbov) b c addilibns made in the inderesd of ornila kltl I am y nol.sv<j 7 es lin g thal lhe 4 rend be rwer. sed,hv) rolhw Jhoi pro <j <e ss i s no kn <, e/ being mode - A mo<e sin <Jhuq renktaJib n howe <e< u 1he inc Ivs a-Ine 1ndec,<aJed P<olevikn 3p em ol leaJv<es whath l in IvansienJx &<e o<e/ Jhe are capable al inibbitnc,l ) ]>y sf velous of eva dvnn.dend wav'lol be in]lia er is.soreynI ordr<s al mo ynrivdo rno<c,,<obo ble Jhon Jhi htiv<e le sc enm urkhk ha s been of concern-This capabilil ,on Jhe en<J ol a g><olechhn syJem,01 inillaling Jransiends is the svid ec0 0) lhe allached pape/ " Dan <j er.s in Safe }y Ji,s-lem3 ' which,9 ou will nole, was published in 19c1 Todn y's reo clors a re subj ec}er) la 4ransienis suc h a s ks.s al leedwolen av los3 oJ food a hnI( do2 on J>ines go/ a,ent. Endar reoclots, nitordinq Jo Ihe popo/ "shovbl --- he ope <aled in s vc h n wa y J ho) Jhe saleh sysdem would be colled upon lo profeci Jhe.. a g rentlo/ r>o more elJon lhon ome or Jwite durincg Jhe life el lhe j)tonl

• Those eoeliu renclov] Side wodr.l on }

Jhrav9 h }hei/ rooline; sysle>ns

en.lorrup ion of toolnni llow caused h loss .611,or Jhan l ol' oil s))e powev, Jhe principal sovrie of lvonsienl> would be reaclivil dtslurhances, of Jhese Jhe rnosL fnleresding were lhose ro* e evenis uherein Jhe dronsieni v was co v.s enl, o< voorsoned, by lhe projetIan syslein ilsell. Two rn echant s ms made nn nypen /a nte, cr, rod s fo lh'n c; w heineg e[ocled f rom Jhe to<e or h,inderacian of proletlan and condrols on13 o n e al ea t h a l Ih ese, de.s c vi h ed in J h e paper, resulled in desivuclion of lhe core, however alhors tuo re nn ero wly n yerden). One inlereslih c example wn s J h e Just 7 Arm 3 Pa tlin e; e Power Ren clor. During a desig n revie w in test 71 was disclosed thal Jhe rod.s would he released by a rolnyt cluJch, The coolnnI flow would be upword n ad Jwo pvinp tuore n <arin ble, on e al lhese a spore., IJ tuo.s ivrJher disclosed Jhal Jne ne] nlownwon>l accelevolkn ol lhe rod 3 too v /d h e

0. 2. cJ a lJ wa s in sla ndlyfosJulaled lhad when lhe.sfarepwnp Luovld be 3}a/}ed,a cold 3lvg would vesu JJ ist a s c ra n1, a nd Ihr ne ) kn e o n JM ro,Js wovld be upwo w],

7J> e rein ed wn s chuia us - povie es uro uld he add ed t< hie A would allow /h e rod s k move on l \\ do wn wn ed n t 4 h ve.sp ei J Lo lhe drives tchen Jht s some d rive wa s apph'ed J<> SL)i lha s Idnt a bove lh e core, ov t fits ) Jh ou g h t tva s o f n. pre.s.s u re Mvced Jronsf~end ahon Jhn popw was w<tMen rl hod no) heen e5Jn blis hed Jhal JJoe lrnnsieni had bnn induied y a de)<),easle o ch b c v---

' Du ring he y ea r.s Sollowincy /961, we saw no more J occidends toused b Jhe projecJibn sy.sdem >lse)L Tiwre y is,oJ couv3e, jhe fnir,5vl 13 Lull Ronl Oroj) probJern uhic h is n.roviodidn c>),lh e rod Jall>W;,Lyom Jhe core. Thh powbrhl has resoIded in ilunts3 opiroIron n n>l loss si y avntlah:1ih,, In tonsidernJi6n oJ Jhe elleda oypended ove/ Nhe yca r.s 's n assorinq Jhal lhe proJetJidn sejslem wcvid n o l in ;-lin de o r worsen o n en onion, ?4 n wilh soin e disma9 lha l we s e e in nesp>R 'ij it l Jhe o elitsornie in clust bn in Jh e J f.S -lh e cin o)4, ,Lc/ in j]a>]inq Ferdwo Je/ Isolaban,which is Jho mosJ severe of Jhe Andicipaled Tra ns> eni.s. The end of lhe roned may e n eo <o/ Jhsri h tu e con hted-t S l-f}An L___________------_-------------

F g as its LA M X M M p E. P. Epler l E strumentation and Controls Diviolon Of.X Rid 6e I:ntiond Laboratory

• Oak Ridge, Tennessec j

At Lon Alcmos, P.ay 21, 19 4, the came plutonium ophere vno being nore= bled,thic timeThe final-htry--Renetor cafety cycterns have failed in 1

such a vay as to dectroy the reactor which they

,vith hemicpherical reflector clemento.hemicphere Sofety roda, when re-1 vere intended to protect. cred into pinec with a r.mnll ceparation being ' 3 cased, enn, throuch calfunction, inerence rather The ele-tronic and rnintoined by inennr, of a hand-held cerevdriver. than decren:.e renetivity. seraing portions of the cafety system are ucually As the cerevdriver van slowly being tvinipulnted, it clipped and allowed the noccably to be com-c,rraneed so that random failurcs ncy on rarc A binst occurred immediately and the occasions result in a sli6ht impairment of pro-pleted. The total locs of protection should not nacembly was dicrupted. The pernon conducting tection. in itself lead to damage to the renetor. the acsembly died ninc dayo Inter. Syste - catic failurcs have occurred however, such that These early occidents occurred becauce of the the safety and control cyctems have become Modern cmplete ]nck of a formal onfety cyctcm.reactoro couplei either deliberately or through oversight 'in such a vay that the loco of protection and reactor runnvny have been enused by a circle event. elaborate cafety devicco but in opite of this, It in evident, in examLn-accidento still occur. thece occidento that remote control of cncem-During opproxirately twenty yenrs of reactor blies has been a large factor in reducing the lcna in6 snd critical as:,cebly eperation, a number of In each instance it of life of the operator and that the absence of cecidents have o-eurred. c:u6t be snid that the accident van rnde posolble nectenulated fiosion producto han been a factor in It is reducing the danger to the cencral public. because of the cbcence of a proper cafety cyatem probably true 'that r.nfety cyotema by their prompt, er the inadequacy of the cafety cyntem either in and correct nction have prevented occidentn but tems of time response, or failure to respond at Understnndably and In several instances the neci-here the record is not clear, dent has been increased in ceverity by action of Icgivttably 'the record of cafety cyctem failure is all vhen needed. the safety cyutem and in come cases the accident core complete. Mc even been initiated by the safety system. Tailurcs of cafety cystems vill be treated y V. R. Stratton $ns reported and described in the following categorics: d . cichteen prompt crit.ical incidents in the Unite 1. Safety redo Since publication of thic vork the Instrumento, ocncorn and replifiers 2. States alone. /snother group of 3. Systematic railures SL-1 nccident has occurred. incidents in the United St.ates has resulted in substantial drennge to tr e core, ac for example, The startup accident has received a great In this accident the control A number of lecser deal of attention. the IIRI rsnd IGIO incidents. excurcions which rerulted in ro dn=nge lic re-rods are withdrawn in nn uncontrolled mnner at corded in log books or have passed unnoticed as naximum cpced until the reculting nuclear excur-sion causco the cafety rystem to be actuated. The pst another false ceram, cafety cystem rust have a respDnte tim at least In order to illustrate the role of the cafety cuf ficiently chort to protect the reactor from the consequences of thin most clerentary accident. 'tystem in relation to accidents, emncples will be dravn from the record, frca recollection of f uite of ten it har. been necessary to reduce the cvents or poctualtes of events which nay reason-maximum rod withirnval speed in order to permit ably be expected to occur.

• the use of a refety rystem having a poor time In the light of exceples to follov it.

must be acknowledged that the startup nccident is response. At Ics /dtcos, August 21, 1945, a critical essembly was being erented by hand stacking re-relatively ninor. flector bricks around n spherical plutonium core. Sticking of cafety or control rodc vhich is The final brick van being enrried over the nssem-a minor failure alco receives a great deal of bly when it becone evident, tAat the addition of It in felt that if a safety rcd be-this briek would unduly inerense the renetivity, attention. At this point the brick clipped and fell into the comen etnck or its response beccraen ir:pnited it center of the ar,sembly enucing a power excursion. v11.1 thereby become ineffective in cane of need. The irdividual who ancembled the experiment died If the situation ic not. remedied it in posolble that all of the roda m16ht become immobilized. twenty-eight dnys later. Concern in of ten exprenned that suljacent fuel ' Operated by Union Carbide Uuclear Company for elemento might becm.e damaged no a connequence of the U. S. Atenic F.ncrcy Commiosion. friction with the control rod, and proponals are

1961, cometinco cmde to limit the tonlue of the drive

,Ecco1vai by t!A Pro; ram Cos:nittoo, July L,.,M motorn in order to mintmize thin drew.e. Althc4gh To bo prosanted au ps;mr ho. the randou stickisyrroh constituten a minor Rrnch Ifd.thn Joint buclmr Imtim.ontat. ion Syuposium, oigh, H. C.; Septaabor 7,1%1. a I F1 L

hazard if promptly repaired, the failure of rods with the' problem of applying rosi drives which are completely contained within the prcosure vencel to operate in the unner intended by the deoigner or which are constructed with the driving rotors .should be* recognized as a varning of impending outo1Je the tank and connected to the rodo throud trouble of much greater consequence, a senl. If the ocal is of the oliding type it lo Another more cerious connequence of sticiing readily opparent thnt n piston la formed Duch that a force of ceveral thoucand pounds may be continu-rods to illustrated in Fig.1. In thic configu-ration reactivity is inerenced by withdroving ously acting in the direction to eject the control poison downward and enfety actie 1 in effected by rodo from the core. It is cuotocary to add de-returning the poison upward against gravity. In a vicco intended to prevent the ejection of rodo, recent incident the control rod became stuck and giving ripe in turn, to endless speculation no to ceparated from its drive end thus fatled to be the conocquence of failure of these added devicco. Withirawn. The danger in this cituation lien in Figure 3 to an example of the rock and pinion the ;ossibility of the subsequent uncontrolled drive with a rotating ocal in which the piston downward movement of the rod, which in come cases night le aided by both gravity and downward vnter ef fect in not no apparent. When the reactor lo scremmed the clutch dicengagen the driving gear l flow. co that the rcx1 in free to move either up or down. Another consequence of eticking rodo lies in Unleso odditional devices are ndded exprecsly for the Jossibility of dislodging adjacent fuel. As the purpooe, any upward forces on the rod which the rod 10 vithdrawn upwardo fuel may be lif ted. may exist during the progreca of an accident, c.ny if, after the reactor becor.es critical, the fuel eject the rods ani vorsen the uccident. becomes diclodged and dropo back into the core nn cxcursion vill result. Again in-the presence of The bell housing which covers the rock downvard water flow the ceverity of the accident usually prevents the rod from becoming a piston, however when the housing is unbolted or recoved vill be increased. the rod raay be lif ted freely by hand or may be If a single enrety rod, on beir#; relenced, ejected from the core by any prescures developed fnllo uninpeded through the bottom of the core a vithin the teutk. The worst possible cituation severe accident can result. The poicon leaven the vould occur when a icoser cecident caused the core having attained a high velocity whereas the reactor to generate steam pressure which in turn purcuing cafety rois must otart from rest. At would enuoe rodr. to be violently ejected, further least one reactor was found in its intter stages increasing the accident. of construction to have no piovinion to prevent w carcty rods from falling out of the bottom chould This configuration exicted in the Sbl cecident which resulted in the loco of three a fnilure occur. lives. Reconctruction of the necident has not Although no accident has been recorded in dicciosed the full extent to which thin copect of which t, cafety rod han fallen through the bottom rod drive der.icn contributed to the accident, of the core, a somewhat similar accident has occur-however it in known thaf, the bell housing vns re-red which enhaneco the credibility. rioved and the operator vno at the time manually lifsing the r 4. It to very probable that he vas At OR!3, May 26,1954, an experiment van in not twnre ad the poccibility of piston action. progreso in which the critical properties of nqueous t.olutione vere being investigated. An TLe abt ve examples chos how cttfety rcds ecn excurolon occurred while liquid van being added initinte or worsen accidento. l'.any other exmples

• slowly to oppronch a critical configuration, are nyniinble, however the /qunrium accident is Tigure 2 illustrates the cituation before nrai af ter the moct convincing case of a critical annembly the t;ccident. The inner cylinder which van essen-being dentroyed by its own safety system.

tinlly a poison rcd, became detached from its cup; ort and fell sidevise to n lesa effective At Los Ainnon, February 1, 1951, a critieni ponition, so that the rystem reactivity exceeded separation experiment van being concluded. The esacably at the tine van suberitical vith a prompt critical. i raultiplication of 65. The ceram button van pusheti Any renetor which 10 pressurized or which cnn nnd inctantaneously a vapor cloud appertred above become precsurized say be cubject to accidento the reactor. Figure 4 chovs the arrangement. TW W vherein reactivity in n function of presoure and r.ansses of U vere suspended in unter. The Icft may be cuddenly increased as precoure in inerenced hand mass van cluipped with a fast acting pneu-or decrenced. During the early operation of the matic cylinder which effected a scram by lif ting Lim at OR:1 the reactor tank became inadvertently the mass upmrd. At the ccme tirto n cadnium sheet pressurized by irproper operation of pinnt vnter van dropped and vnter van drained fran the tank. eyoten valves. The top tank ecetion included n In thin instance the center of renetivity of the bellovo to facilitate alignmento nnd the control lef t httnd nans van lover than that of the other so ral driven vere mounted on the tank cover above that on cerni:: ming, the reactivity van at flrot the bellova. The inerensed preocure enuced the inerenced. At the onme time Bern. (111 forces entire control drive system to be lif ted with caused the two r.acnen to eving toi 'ler thus respect to the core thereby increening renetivity, further increnning the renetivity. The icar of loon of protection throu[h Preocurized vnter renetor designero nre fnoed $2 G 9

.......,--...n ditional shutdown rods of different dcoinn in tceted by the test regime. One should expect the der to 1.nsure that at least one kind vill be reactor to be operated in cuch a vay that the etable in cgcc of need. An examination of safety cyctem would be called on to protect the rattoMr, cichtcen prompt critical accidento reactor no more often than once or twice during wever, discloces not a single case in which the life of,the plant. It is reasonable to + acking rods failed to protect the reactor. On expect,therefore, thnt should the electronic r other hand, at leact two of the accidento system fnil and' the failurc go undetected, tempo-re clearly cauced by calfunctioning chutdown arily, an accident vould not occur during this rices. It vould appear therefore that the tinne. rater danger lico in calfunctions other than kiir.g and that until it is certain that the Although instrwent failure alone cannot j

gerous characteri: tics of these devices have produce an accident, catactrophic failure has i

ta removed, additional chutdown devices r21 ht resulted when infomation channels to the centrol j 5 udly increase the danger. system, to the carcty cystem, and to the operator have been blocked by the came event. In thio The danger in sticking rods should not be cituntion the reactor may appear to be chutting irely discounted however since cticking can be itccif down. The automatic control vill with-leur varning of more cerious trouble. When drav rods in an attempt to caintain the desired . function of rods is co=nonplace t.ny varning of power level, the operator accing the power level ending disaster r.ny be micunderstood and droop will look on in come puzzlcment or vill cred. attempt to maintain power by the withdraval of rods. The safety cystem, if it receivce the same In contrast to ssfety rods which through information vill not respond. lure can increase reactivity and thereby f 2 aroy the reactor, the electronic portion of At 1:IE the 133 was undergoing testo at i. system is pa:,1ve anl can fail only in that progreceively h15 er power levels which vere far ,j h tection cay r.ot be available when needed. below the maxinum rated level. On this day the reactor was being brought up to a new power level r, Ionization chcabers and explifiers used in and the ionization chtnbers were therefore re. ctor safety systems have been considered quired to operate at a higher current than had damentally ur. reliable devices and perhaps for previously been experienced. The reactor had s reason have received cuch more attention almost attained the dcof red power level when the 1 other portions of the safety system. Sophio-safety instruments, which vere the only instru-I ated circuit arrangements have been developed ments on ccde at the time, indicated to the , }[

b. that almost any conceivable failure causes operator that the power Icyc1 vas decreasing. Tbc circuit to trip and is therefore " fail safe."

automatic control channel which chared ionization 1 trnately monitoring systems have been deviced chambers with the safety syctem also received f s that circuit t.nd ccuponent failures are mado infomatien that the powcr level van decreasing 3-ca co that repair or chutdown cay be acecm-and opernted correctly in withdrawing the control tbed as desired. More recently automatic rod. A few seconds later the reactor chut itself ting optems have beco=e available which per-dovn by ec1 ting a cubstantial portion of the core, j? a I tests on the cicctronic portion of tho

ty system at frequent intervalo. Other sys-In this Instance a cyctcmatic failure occur-I have been used which interpose absorbers red in that a noise filter which had been inctal-icen the ionization chambers nnd the core in led in all sonization chamber power supplien was r to include the chambers in the portion of of su.fficiently high resistance to make it impos-

[: system undergoing test. In order to increase sible for the ionization chambers to operate at .her the reliability of thece systems it han the required current. At the some time, the ithe practice to employ redundancy of chana increasing gamma radiation ohunted to ground , always installing at 1 cast two and usue.lly around the chamber an inercasing cmount of currett e identical channels. In order to offset the which caused the indicated power level to decrease Taced nu=ber of falce trips brought about by when in fact the power level was increasing. Two added equipment, coincidence is of ten cm- 'pointo here should be emphasized: i hd such that the ogreement of two or more acic is required to actuate the trip. Al-1. In the interecting cases of cysteentic ( 6h this arrangement is in the direction of failure, the monitoring cystem or testing pro-safety, a large inerence in cafety is poten-cedurco are unable to clicciose the difficulty. ly availnbic in that nny one channel may be ed and tripped at vill without the penalty of 2. The safety cystem, automatic control i tor chutdovn. Some of these systema have byntem nnd manual control system become coupled used repeatedly over a period of years on and therefore all rcccive the some erroneoua Lors in many countries to the point that thce information. In particular, the operator has no k=s, though complex, are vell understood and better information than the automatic control and been ccupletely " debugged." thereforo' cannot be expected to remedy the clifficultien. Irr epite of the elaborato testing regimes l redundancy of channels it is conceivable that There are many reactorn in o;eration today 1 1 l e $3 l

Bn which the safety cycten and autociatic control hardware vere not as well developed no m16ht be systes chare the ccme ionization ctmbero and it expected today. It muot not be ccricluded however e 3ay be c.nticipted that core failures of thin that rulern reactor systems are free of thene 5 Bype vill occur. Even though ceparate chambero dangern. Although it in now locaible to buy a Eay be installed it han been observed that inad-lockaged rod drive oyoten or an electronic packagc 7ertent coupling cay be obtained in ceveral vayo. it ruuut not be ascumed that these vill be free of j dangerous characteristico; indeed it win be 1. Duried in log books vill be found refer-necessary not only to eliminate dangerous char-2nces to ensco wherein a vater leak has vet the acteristics but to be constantly on the alert for thieldits material between the ionization chambers new dangerc Which may be dicclosed as the result ird the core. Since water attenuates the neutron of accidente. flu.x availcble to the chanbers the power level rin eppear to be decrcasing. Even though another In this conocction it 10 unfortunate that instm:1cnt cuch as a therzaoccuple not to affected more cannot be learned from near misses and Isall ay show cn inercase in power, it vould take incidentc. Indeed it 10 not alvuyc true that the

reat presence of mind on the part of an operator most useful conclusions arc drawn fras catastro-o belle vc thic cirgle correct indication in the phic failurec which are thoroughly investigated ace of contrary inforcation from a n ionization and reported. For example the IrITO accident niglt hambers. In the CIO exa ple the accident be dicmicced as instrument failure and corrective ecurred at a power level too lov for the thermo-measures limited to the removal of current 11 nit-suple indication to be on ccale, ing resintors which raight be found in chamber power supplies. A broader viev vould conclude 2.

In come reactors all ionization chonbers that this vac an example of systematic failure re placed in the came chield tank. Thic is in which the safety and control systems were hually a reparate water tank which carrounds the coupled. On the basic of thic varning all exict-rescure veccel and ir, filled with water to pro-ing reactor cyntoms at ORNL have been combed over ide additional chiciding. It is not unlikely in search of such couplings as could renconably i hat boron could be incdvertently added to thin. be removed. It must be conceded that cafety and I hield tank in order to improve the ch*-lding, in control cast chare certain facilities and that the l 11ch case the neutron flux to all 4 era vould possibility of circultaneous and catastrophic a attenuated cLd cystematic fc11tt. a cd occur. failure can therefore never be ruled out. As an i example it vill never be poisible to rule out i A story is told of a case in which the vrong instructions to rervic3 personnel or a tactor phyoicists cdded boron to cuch a chield faulty test instrument as the prime cause of, mk in order to improve the neutron chielding. systematic failure. Ley were at first at a losc to t.nderstand why se cacma radiation through thic chield in-In a citailar manner the SL-1 accident night 'ensed. Further consideration dicelosed that be interpreted as a crisly dc=onstration of the, f te controls were r.aintaining the reactor power loscible consequence of picton action when cafety j n hir,her icvel in order to caintnin the neutmi rods are subjected to differential pressure. It g' hx constant at the ionization chnnbers and should not be necessary to prove that picton j 'stenatic failure was being deronstrated. ection occurred in thic particular accident in g order that thin danger be recognized. Indeed it l 3. It han been the practice to flow can r.ttcht be proved that platon netion did not occur { 3 rough Jonization chambers using bottled nitro-1:st thic thould in no vay prevent stepo being -l 9 for the purpse. A single bottle of vet. taken to recognize and remove cuch dangers from 6 trogen could ground all cables oo as to parn-safety' cysteca in other reactora. l ;3 ke both the ortrety and control inforcation h.nnels. In a recent instance c.t CRNL trouble Acknovicitement i been experienced due to noisture in the coax f bles cnd a cingle drying system vas installed It is a pleasure to acknowledge the interent' prevent moisture from enterirc the cables. and helpful cuggestions of S.11. lianauer in the to it was recognized that a single vnter leak preparation of this paper. tid utsit vater to a.n cables, the system vaa

ediately changed to provide an individual References ring unit for cach channel.

,f It chould be noted that many of the testing

• E,* Stratt n, "A Review of Criticality

^ 6"" N '8 monitoring rystems nov in uce vould not de(cet 8 'se systematic failurec. It 10 rot probable, M* 3, MgaMn MSD DM. i ' example, that any exinting testing cyctem ( 3, g ld detect a vnter leak between the core arx1 'L I*, Ipler, "KITtE-3 Excurolon," Fuclear chambers. t cafety, Vol.1, No. 2, (December 1939). [oncluoton l f>ome of the exampleo illustrating danger in Ety systemn vere drnvn frco early accidenta to lical experinenta where r,daittedly re.ethods and ' A l ~- C l I p w

vacuwento OnmL-La-0wo 600i0 + ~ p CONTIA. A00' / ~. - 911e m a tvf L moef LIP T3 '/100 4 ~~ 'i 4 \\ ll } 'T==d'.:)e/ \\ s t l: j a t I i 0 .f'

• AbwuW CLAD b

/ * / s. E I s ~- ~1

vuN(gn %

t' / l'l* / / >* g l f/ LI(:..I / / 7 cum jj j' [j.L q,,,otync,, A I / I/ q. l ( ,d' y;L[m i, s

=

ry. Dorser in Sticking nod , ry. e f%on Folfny (u of Cors. '1 r WCa ca%- t a.1wer e s=a some UwcLAssrep CRNL.LR. pag 60(Mt ORwt sco, M "' LT E I'q" J"2. {D""'] b = l,h d}: ). w ~* g .. -C ADMiUM SHEE T 1 [SCA4W twfDs,d / DR0f'S TO 14 HAM patigo L. .J a UTTS TO ~ = g.. __.- g ____ma----- SCHAM% \\ , ~~l ust _.--.f-O wn.u: A..-L

-P.

=. - ; w - /~ 70]g*! "fW'"" tM t 9 ss ggg,y 9 RE ACilVITY [ HEDUCilord l)l t rP% mm ,I L. I 1 ry.& Aqwrke connguretbn. \\ l g e I r&e. ) Romerol of SWie $wyrwtg holon fictg l e 0 f* e 's e e s s S 8 1 6 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ - _ _ _ _ _ _ _ _ _. - _ _ _ _ _ _ _ _ _ _ ' _ _ _ _}}