ML20024A510
| ML20024A510 | |
| Person / Time | |
|---|---|
| Site: | 05000142 |
| Issue date: | 06/14/1983 |
| From: | Aftergood S, Dupont D, Hirsch D, Kaku M, Kohn R, Norton B, Pulido M COMMITTEE TO BRIDGE THE GAP |
| To: | |
| Shared Package | |
| ML20024A493 | List: |
| References | |
| NUDOCS 8306170390 | |
| Download: ML20024A510 (78) | |
Text
{{#Wiki_filter:i 1 PANEL i
&lltllT3 INTRODUCTION -
CONCEPTS OF INHERENT SAFETY AS APPLIED 'IO THE UCLA ARGONAUT REACTOR
- 1. The review reported in the testimony which follows focuses on the question: Is the UCLA nuclear reactor inherently safe?
- 2. A detailed examination of whether the Argonaut-type reactor on the UCLA campus meets that standard is best begun by a brief discussion of what the basic concepts of inherent safety are.*
- 3. Nero (A Guidebook to Nuclear Reactors, DC Press, 1979) points out that reactor safety features may be divided into two categories: intrinsic safety features, "those that are inherent in the physical nature of the reactor concept being considered," and enaineered safety features, " systems that are added to the basic reactor concept." Nero gives the following examples for the two kinds of safety features :
An example of intrinsic mechanisms is that, in a light-water reactor, overheating of the coolant caused by an abnormal rise in the reaction rate tends to reduce the water density and thereby shut down the chain reaction due to insufficient moderation of neutrons. Alternatively, the emergency shutdown-control-rod system may be regarded as an engineered safety feature. (p. 13) Thus, engineered safety relates to devices which, if they operate correctly, can prevent hazard; inherent safety relates to intrinsic design features of the reactor which operate automatically, due to the laws of physics, rather than relying on correct response by devices or human operators.
- 4. The history and underlying philosophy of the development of research and training reactors was an attempt to create reactors that were indeed inherently or intrinsicly safe, i.e.
- We understand that the Atomic Safety and Licensing Board has requested that technical terms and concepts be explained so as to aid the general public in understanding the issues involved.
The brevity of this explanation necessitates considerable simpli-fication of rather complex concepts. 8306170390 830614 PDR ADOCK 05000142 T PDR a
2-that did not need to rely on proper operation of engineered safety features or proper maintenance and operation by human beings. This was because research reactors, unlike power reactors, were to be used for training students, who must be expected to learn by making mistakes. A safe training reactor, thus, must be of an intrinsicly " forgiving" design, so that the worst possible errors could not possibly cause harm, either to the student or to the public.
- 5. A clear indication of the roots of this philosophy of the necessity of inherent safety for research reactors can be found by reviewing the origins of the development of the TRIGA reactor, probably both t he most successful reactor design and the one that most closely meets the inherent safety standard. The TRIGA was designed in 1956 by Edward Teller, Theodore Taylor, Freeman Dyson and others. As Dysan has described it (Disturbing the Universe, Harper & Row,1979),*their task was to " design a reactor so safe that it could be given to a bunch of high school students to play with, without any fear that they would get hurt."
- 6. Meeting-that standard is difficult because of the unique nature of nuclear reactors. As will be described in detail during the panel on power excursions, nuclear reactors are different than virtually any other machine in that they can undergo astronomical rises in power in extraordinarily short periods of time, from zero to billions of watts in a very small portion of a second. Such an event can lead to melting of the fuel and explosive disassembly of the reactor core. Preventing such occurrences is thus the goal of the designers of reactors.
As Dyson puts it: The f irst rule in operating a reactor is that you do not suddenly yank the control rods out of a shut-down reactor. The result of suddenly pulling out the control rods would in most cases be a catastrophic accident, including as one of its minor consequences the death of the idiot who pulled the rods. l l The TRIGA design team knew that, for training reactors to be put on university campuses for student instructional purposes, the reactor must be " idiot-proof." Other kinds of reactors could rely on engineered safety, which Dyson defines as meaning that a catastrophic accident is " theoretically possible but is prevented by the way the control system is designed." t
- 7. For the designers of the TRIGA, Dyson says, " engineered safety was not good enough." The reactor must be designed with inherent safety, l
meaning that its safety must be guaranteed by the laws of nature and not merely by the details of its engineering. !
- excerpts attached as Exhibit C-i-1
. 1 l l
-,,n
=. . . . . .- .-. -
~
l t It must be safe even in the hands of an idiot clever enough to by-pass the entire control system and blow out the control rods with dynamite. Dyson said that stated more: precisely, the ground rule for a safe research reactor was "that if it was started from its shut-down condition and all its control rods:~ instantaneously removed, it would settle down to a steady level of operation without melting any of its fuel." This was the standard for measuring the safety of research reactors with regards one of their principal hazards--potential for power excursion. The same standard--protection by the laws of physics, without need for proper operation of devices or people, either of which can fail--was applied to other hazards (e.g. fire) as well. The inherent design was to be failsafe and foolproof. THE ARGONAUT REACTOR 4
- 8. The original Argonaut, though not nearly as inherently safe as the TRIGA, was far more safe than the Argonaut-type reactors that were a few years later sold commercially by AMF and American Radiator and Standard Sanitary Corporation, before they went out of the reactor business. And as shall be described
' in detail later, the modifications to the UCLA Argonaut-type reactor further. over the years have significantly reduced safety margins
- 9. The original Argonaut did not have the inherent safety feature that is the primary characteristic of the TRIGA--exceedingly prompt, almost instantaneous self-shutdown features, due to part of the moderator being built right into the fuel. But it used low enriched oxide fuel, had an exclusion area around it, was physically restricted to very-low power operation (normal l operating power was 100 watts, one thousandth that of UCLA's; because of the inherent negative temperature coefficient, sustained operation above about 1 kw would result in shutdown), had a series of shutdown and interlock systems not found in the commercial model, as well as other safety features. The low power restriction and exclusion zone features alone serve to reduce the consequences of an accident at the original Argonaut many orders of magnitude j below the consequences of an accident at a contemporary Argonaut- 1 type reactor. These matters will be discussed in more detail in other panels.
- 10. AMF's Argonaut-type reactor--the " Educator"--reduced the safety margins but still maintained some important ones, particularly the excess reactivity restriction (to be explained later) of less than that necessary for prompt criticality.
The reactor's maximum design power was 10 kw, and the requirement for an exclusion zone was removed as were other of the s afety features. l
i 1 .. 11. A few years after the UCLA reactor was built, safety margins were further reduced. Power (and thus fission product inventory available for release in an accident) were increased ten-fold, and excess reactivity limits four-fold, to precisely the level the Hazards Analysis for the reactor calculated would render what inherent shutdown mechanisms the reactor had ineffective in preventing fuel melting. Other modifications, described in other panels, occurred as well, further reducing safety margins.
- 12. Before the UCLA reactor lost these safety margins, the University's Regional Advisory Committee On Radiological Safety a comparison of the Argona_ut and TRIGA reactors in terms of safety.
Their conclusion: "/_ in_/almost every respect, the safety margin seemed to favor the TRIGA reactor, in spite of its higher possible operating level. " Exhibit C-i-2. It should be remembered that after this assessment, the Argonaut power went up by a factor of ten and the excess reactivity level far over the prompt critical threshhold that had previously been its primary inherent safety feature. (see also Exhibit C-1-3). BRIEF EXPLANATION OF RELATED TERMS AND CONCEPTS
- 13. The chief hasard associated with nuclear reactors arises from the ebetremely' toxic nature of the products of nuclear fission. These fission products are intensely radioactive and can pose very considerable danger to the public if ever exposed. In order to prevent that, most reactors engage in multin1. e, varh=%t barriers to fission product release in an effort to ensure that, if ever an accident were to occur, little if any of the radioactive material would reach the environment. These barriers include the fuel cladding, the pressure vessel; the conta4= ant structure, and a series of engineered features to anhance the effectiveness of the contm4==nt, such as containment sprays, ice systems, radioactivity removal systems, filters and the like.
Much of the debate over nuclear safety has focused on the effectiveness of these features.
- 14. In addition to multiple barriers to fission product release, most reactors have substantial exclusion zones and low-population zones surrounding them so as to permit significant dispersion of the radioactive material, if released, before it reaches members of the public. Concentrations drop by several orders of magnitude within the first quarter mile or so, so keeping the
l
~
l _p nearest person at least that distance away, and providing that densely populated areas be considerably further sway provides a measure of protection.
- 15. Some research reactors have containment structures and exclusion son.es, though.not as large or complex as power reactors. N UCLA reactor has neither.
- 16. In the case of the UCLA Argonaut-type research reactor, the primary barrier to fission product release is the cladding on the fuel, aluminum fifteen thousandths of an inch thick. (Application, page V/1-4). If the cladding were damaged, fission products would be released with essentially no other barrier preventing thms from reaching the publio. The small size of the fission product inventory relative to that of large power reactors is essentially compensated for by the lack of exclusion sono and *.2igh population density immediately adjaoent to the reactor room and, the lack of multiple barriers or other means of mitigating fission product release should it occur.
h amount of fission product release would depend on the degree of damage and the temperature of the fuel. At or near the molting point, fission product release would be very substantial. _
- 17. m UCLA fuel meat and Am- (as well as the control blades) j are among the lowest-melting such materials used'in reactors. .
It thus becomes very important to ensure that no accident can occur which can elevate the temperature of these materials to close to their setting point. l
- 18. h primary reactor materiale paz+1m6-17 the graphite, uranium metal, and magnesium- are all potentially combustible. Should they catch fire, fission products, including large quantities of particulate material, would be driven out of the fuel and into the enviramment. h a any con-dition which might result in fire would present a serious threat.
- 19. Certain events can also initiate violent chemical reactions or explosions which can disassemble the core and release significant portions of the fission product inventory. Steam explosions, metal-water reactions, or placement of explosive materials in an irradiation port within the core or merely nearby the reactor can induce such disassembly, perhaps initiating an incendiary reaction as well. Furthermore, numerous chemicals
- - - - - - -l attack =1t=4 === oorrosion of the al-Ad4== can thus penetrate the primary barrier to release of the radioactive materials contained inside.
20 Mechanical damage to b fuel, as initiated in an earthquake which crushes the core or through a fuel h--d14== accident, can damage the oladding and release some of the gaseous fission products inside. 21 Finally, there is a kind of accident peculiar to reactors which is essentially a reactor runaway, where the power escalates by many orders of magnitude in entreely short periods of time. This can result in molting of the fuel, violent steam and ehemioal explosion, and destrucha of the reactor oore. A less severe version of this accident, called a "oriticality accident?, can result idien uraniisa or plutonium is aoaidentally brought into the right configurawn to go "oritical" outside the reactor, remiting in
~
' e intense radia W n burst.
- 22. Because there aren't unitiple barriers to fission product release, and because of the dense popula w n with no emoluzion sone, UCLA must, to be licensed, demonstrate that no credible sooident could occur at its facility that would result in release of more than a very mall fracWn of one percent of the reactor's radioactive inventory. Release of even a very mall fraction of the core inventory would be devastating given the site characteristics.
~
- 23. UCLi has attempted to make that showing by asserting that the immri-na credible sooident at its facility is =mahanical damage to ogg of the 262+
fuel plates of the reactor, after three weeks of " cooling down" has pemitted the radioactivity in that plate to decay quite considerahly, and assuming only 2.?jE of the . -4a4ag Kr-85 Ie-133, I-131 and I-132 are released, all other isotopes remaining in the fuel. (Amended Application, p. III/8-12). UCLA has thus taken as its maximum credible accident the release of 0.03M ouries total. (M., Table III/8-2, collaan 13 0.002/+ + 0.017 + 0.013 + 0.0020 = 0.03 W Ci). This is less than one ten-millionth of the core inventory. (Computer run, portions attached, performed by UCLA indicate inventory goes to 34,000 curies after a single 8-hour rung maximum core inventory is considerably higher).
- - - - - s__-----___ - . , . - _ _ - -
l l 24 Sus, the questica to be examined is whether there are any credible accident scenarios that result in fission product releases substantially in excess of 1 x 10~7 of the core inventory. A release af 10% nf the inventary from the fuel--as shall be aboon, a not-unr===anable estimate for several classes of accident at this facility-would produce consequences a_t_ least a million tians greater than those assumed by UCIA for its asserted maximum credible accident. (It should be noted that, in addition to assuming an extraordinarily small fission product release for its design basis accident, II:IA has used a number of extremely unrealistic assumptions far estimating diW4n. For example, although Technical Specification 3 4.1.2. mandates that the ventilation system be shut down autoastically in an emergency involving excessive radiation, the UCEA analysis =====s - continued operation of the ventilation system and thus substantial dilution
; of effluent which will not exist in an emergency. Sose ina ywydate r calculational methods, when corrected, would increase consequences an additional order of angnitude or two.)
- 25. N issue, then, is not whether the inventory of the UCLA reactor is en small as to be inconsequential in case of accident. It appears established that.rinlasse of a substantial fraction at that inventory would produce unacceptable conseguences, given the lack of other barriers and the site characteristics. He isome appears to be whether any inherent self-limiting features preclude aslease of more than a small fraction af a percent of the core inventory. Se answer, as shall be demonstrates, is in the negative.
- 26. One additional prefatory comment is in order. UCLA, in its amended application, at page III/8-1, appears to assert that only power reactors can have "1kximan Credible Accidents," and that by virtue of being a non-power reactor, the UCIA Argonaut-type reactor is autoastically immune from any accident which could result in release of a substantial fraction of its inventory. Dat is simply not true. In fact, if one examines the history of reactor accidents that have involved fission product release, the vast majority of those to date have occurred in non-power reactors. Of U.S.
commercial power reactors, the Ferni reactor ami the Three Mile Island Unit 2 reactor are the primary instances. (It is interesting to note that the official estimate of iodine release to the environment is within the same order of
angnitude at TMI as the Hawley study assumes for a fuel-handling incident at USA involving one cold fuel. elements the inhalation doses, moreover, , I would be considerably greater at UGA because of the lack of exclusion sone.)
- 27. S e history of non-power reactor accidents involving severe fuel damage is far more extensive that that for power remotors to date, largely I i
because of the lack of engineered safety features, the lack of standardised design, the experimental nature of the program involving such reactors, and the far smaller degree of safety research focused upon them. Se list of non-power reactors suffering major accidents involving substantial fuel damage includes the SL-1, the SIS, the NRK and NRU, the HTR5-3, EBR-1, and Vindeoale, to name a few. Furthermore, the SPERT and BORAI tests clearly demonstrated (as can be seen from the films of their final destructive excursions) that such resetors can suffer substantial fuel molting and even ocaplete core disassembly. It is interesting to notethat all of the above remotors are of the sans vintage as the UGA Argonaut-type reactor. Se lack of a remotor vender still inbu=1mmes to provide updating, spare parts, ami expertise exacerbates the problems regarding the UGA remeter. Se lack of a system for identifying problems with similar reactors and passing on lessons learned and requiring a ,,.;,. late backfitting askes the facility, in addition to being relatively primitive from a safety standpoint due to its design prior to anjer advances and insights in the field, something of an operating relio. Most of the changes that have been made have been in the ! non-safe directions ten-fold incrosse in power and fission-product inventery l very substantiel increase in excess reactivity, to a dangerous levels pneumatic tube system for rapidly inserting and withdrawing reactivityg and others.
- 28. Se panels that follow will address how such changes have affected the potential for serious accident. Se categories of socidents semained includes severe power excursion, fire, chemical reactions such as metal-water reaction or severe a1=ddian corrosion, and fuel-bandung incidents. Attention has been paid to saltiple or cosmon mode failure sequences that could result in accidents involving more than one of the above categories (e.g., seismically- l induced event resulting in both core disruption and fire).
l l l
}
INTRCDUCTIUN Exhibit List Exhibit Number Descri Mien C-i-1 Freeman Dyson, Disturbing the Universe, excerpts C-i-2 Report and Minutes of the Regional Advisory Committee on Radiological Safety UCIA, 5/12/60 C-i-3 " Advantages of TRIGA Fuel for Research Reactors", General Atomies, Cetober 1981 C-i-4 Fission product computer estimate l l l l l
>-e
..f .
y
....c.( ...
g,y,qf.p i;' %.;-(f,,',_
; AL.._
g i %, '1. 9 .-2}f[,,h,.;j 'j.-; ,,# {,'gg [ 9.? # ,,i; M i.e.s 4g.d Ay,g-e 9 i M.. - f Exhibit C-i-1 page 1 of 3 FREEM AN DYSON
- S UR3" ~G pg., , , , , or ,r ,m, ,,
- 4 H H
., . ,,,a -- su aasu sd HARPER & ROW, PUBLISHERS NEW YORK P U 8 t.l s if E D AS PART Cambridge i jj Undon U N D A TION PROCRAM Hagerstown Igf i Meuco City Philadelphia i g Sio Paulo San Francisco L_L _1 Sydney
\
]817 -
l s e
~ w , m> _ __ ['% ,,, g , --, .
._ wy _
. "% . = , e ' C" .
~
** EMS - * * * /*'f -
E ~- ,
l 9 S , .. ... .. u,,,sa,w ~ ~~..~ ~ , U. , staf. Freddy rinted a little red schoolhouse th:t had been rben- tors. Our primaryjob was to End out whether thera was any speciSc doned as obulete by the San Diego public school system. He pro- type of reactor that looked promising as a commercial venture for posed to move into the schoolhouse and begin designing reactors General Asomic to build and sell. there in June. % lectures were excellent. %ey were especially good for me, Freddy had been at IAs Alamos with Edward Teller in 1951 and coming into the reactor business from a position of total ignorance. j had made some of the crucial calculations leading to the invention But even the estabhshed experts learned s lot from each other. no j of the hydrogen bomb. He invited Teller to join him in the school- physicists who knew everything that was to be known about the j house for the summer of 1956. Teller accepted with enthusiasm. He physics of reactors learned about the details of the chemistry and 4 knew that he and Freddy could work well together, and he shared engineering. %e chemists and engmeers learned about the physics. l' Freddy's strong desire to get away from bombs for a while and do Within a few weeks we were all able to understand each other's . something constructive with nuclear energy. problems. Freddy also invited thirty or forty other people to spend the ne afternoon sessions quickly crynealli=d into three working i summer in the schoolhouse, most of them people who had been , , groups, with the titles " Safe Reactor,"
- Test Reactor" and " Ship involved with nuclear energy in one way or another, as physicists, .,,. Reactor." %ese were conndered to be the three main areas where chemists or engineers. Robert Charpie, even younger than Freddy, . . . an inunarliana market for civilian reactors might exist. In retrospect had been the other American in the group of scientiac secretaries of ,
it seems strange that electncity-producing power reactors were act l the Geneva meeting. Ted Taylor came directly from Ims Alamos, on our list. Freddy knew that General Atomic must ultimately at t l where he had been the pioneer of a new art form, the design of small . into the power reactor business, but he wanted the company to beg in i cdEcient bombs that could be squeezed into tight spaces. For some , i. with something smauer and simpler to gain experience. %e s' nip i ) reason, although I had never had anything to do with nuclear energy reactor was intended to be a nuclear engine for a merchant ship, and j and was not even an American citizen, I was also on Freddy's list. - l the test reactor was intended to be a small reactor with a very high i Probably this was a result of my encounter with Teller the previous neutron Sun which could be used for the testing of component parts ! summer. Freddy pronused me a chance to work with Teller. I ac- ', of power reactors. Both these reactors wosld be competing directly I cepted the invitation gladly. I had no idea whether I would be sue- with existing reactors that had already been developed for the Navy I cessful as a reactor designer, but at least I would give it a try. For , and the Atomic Energy Commission. Both of them were designed i nineteen years I had been waiting for this opportunity to === Ira Ed- during the summer and then abandoned when Freddy concluded j l dington's dream come true. that they had no conimercial future. The safe reactor was the only Freddy de Holfmann was my Srst encounter with the world of Big product of our little red schoolhouse which actually got built. Business. I had never before met anybody with the authority to make I- %e safe reactor was Teller's . des, and he took charge of it from decisions so quickly and with so little fuss. I found it remarkable that the beginning. He saw clearly that the problem of safety would be this authority was given to somebody so young. Freddy handled his decisive for the long-range future of civilian reactors. If reactors ! power lightly. He was good. humored, and wdhng to listen and learn. , were unsafe, npbody in the long run would want to use them. He told l He always seemed to have time to spare. l Freddy that the best way for General Atomic to break quickly into l We assembled in June in the schoolhouse, and Freddy told us his
- the reactor market was to build a reactor that was demonstrably safer I
plan of work. Every morning there would be three hours oflectures. " l than anybody else's. He deSned the task of the safe reactor group in
%e people who were already expert in some area of reactor technol- the following way: ne group was to design a reactor so safe that it 4 i ogy would lecture and the others would learn. So at the end of the could be given to a bunch of high school children to play with, u!
! summer we would all be experts. Meanwhile we would spend the without any fear that they would get hurt. %is objective seemed to j afternoons divided into working groups to invent new kinds of reac- i me to make a great deal of sense. Ijomed the safe reactor group and ! ~; . t 4
96 spent b next twa months with Teller Eghting our wry through is . st:ady levtl of oper tion without meltirig any of its luei. N. . One of the Erst st:ps tow:rd the design of the safs reactor was ta i e satisfactory solution of his problem. Working with Teller was as exciting as I had imagined it would introduce an idea called the " warm neutron prmeiple," which says lg be. Almost every day he came to the schoolhouse with some hare- that warm neutrons are less easily captured than cold neutrons and
! are less eS'ective in causing uranium atoms to Ession. ne neutrons !
brained new idea. Some of his ideas were brilliant, some were practi- ' in a water-cooled reactor are slowed down by colksions with hydro-cal, and a few were brilliant and practical. I used his ideas as starting i Points for a more systematic analysis of the problem. His intuition i sen atoms and end up with roughly the same temperature as the and my mathematics Stted together in the design of the safe reactor hydrogen in whatever place they happen to be. In an ordinary water-just as Dick Feynman's intuition and my mathematics had Atted cooled reactor, after the postulated idiot has blown out the control together in the understanding of the electron. I fought with Teller rods, the fuel will be growing rapidly hot but the water will still be , as I had fought with Feynman, demolishirig his wilder schemes and cold, with the result that the neutrons remain cold and their effec- ! squeezing his intuitions down into e.quations. Out of our Serce dis- tiveness in causing Ession is undimimshed, and therefore the fuel ' agreements the shape of the safe reactor gradually emerged. Of continues to grow hotter until it Snally mehs or vaporizes. But sup- ,
. course I was not alone with Teller as I had been with Feynman.no . pose instead that the reactor was designed with only half of the ,
safe reactor group was a team of ten people. Teller and I did most hydrogen in the cooling water and the other half of the hydrogen mixed into the so!'d str.ncture of the fuel rods. In this case, when the of the shouting, while the chemists and engineers in the group did , idiot yanks out fhe control rods, the fuel will grow hot and with it the l most of the real work. hydrogen in the fuel rods, while the hydrogen in the water remains Reactors are controlled by long metal rods containing substances such as boron and cadmium, which absorb neutrons strongly. When cold. De result is then that the neutrons inside the fuel rods are ; you want to make the reactor run faster, you pull the control rods a warmer than the neutrons in the water. %e warm neutrons cause , less Assion and escape more easily into the water to be cooled and ; little way out of the reactor core. When you want to shut the reactor ' down, you push the control rods all the way in. De Erst rule in captured, and the reactor automatically stabilizes itself within a few . thousandths of a second, much faster than any mechanical safety oPer ting a reactor is that you do not suddenly yank the control rods out of a shut-down reactor. He result of suddenly pulling out the ', switch could hope to operate. So the reactor carrying half of its - j . control rods would in most cases be a catastrophic accident, including , s hydrogen in its fuel rods is inherently safe. as one of its minor consequences the death of the idiot who pulled ,
! nere were many practical disculties to be overcome before ;
the rods. Alllarge reactors are therefore built with automatic control
.' these ideas could be embodied in functioning hardware. De great- '
' est contribution to overcoming the practical disculties was made by syst:ms which make it impossible to pull the rods out suddenly. '
nese reactors possess " engineered safety," which means that a Massoud Simned, an Iranian metallurgist who discovered how to ' cctastrophic accident is theoretically possible but is prevented by the make sud rods containing high concentrations of hydrogen. He w:y the control system is designed. For Teller, engineered safety was , made the rods out of an alloy of uranium hydride with zirconium , not good enough. He asked us to design a reactor with " inherent hydride. He found the right proportions of these ingredients to mir 4 safrty," meaning that its safety must be guaranteed by the laws together and the right way to cook them. When the fuel rods y emerged from Massoud's oven, they Iaal I like black, hard, shiny g. ' of nature and not merely by the details of its engineering. It must be safe even in the hands of an idiot elever enough to by. pass the metal, as tough and as corrosion-resistant as good stainless steel. *i cntira control system and blow out the control rods %ith dynamite. After we had understood the physics of the safe reactor and the . 4l St:ted more precisely. Teller's ground rule for the safe reactor chemistry of its fuel rods, many questions still remained to be an-swered. Who would want to buy such a reactor? What would they use was that if it was started from its shut-down condition and all its ~ uli ; it for? How powerful should it be? How much should it cost? Teller
; control rods instantaneously removed, it would settle down to a I
i . l ! e
' i i
?
- ---c. - ,. _ , . , - . - - _ _ _ , _ , - _, _ y
g..
\ Exhibit C-i-2 page 1 of 3
, ner.nr.t :'f th: P. n;. ro, fvSr. t .. e , .i ' : ;. . . e .e , . 1e: . y, :oa S a b.ic .i s' p ?.c . tor M2etia; held Thutcday, !:r.y 12,1%0, Myuits radidin , Uc.2, 3:00 p.n. Frescut: C3.ra '2::cc Roberts, TPc .n S. Hi. chi. Cctried t '.Pi, aq:-or ! i.. I f t by
#n.- (L. R. 3 :nnett') , Kc:.ne th P., 'one%cn7.ie (.: bait r cd* .
Invited pens: Jcha C. Craets, G2reld fi. h:Donell lG1% The ::::: ting was c:.11cd to rc71c; tha :. 'ety :.spects cf a. 3;rcyct::e Trip Auk II r m. t t - Zus :r.11-Wion en the UCIA .;n:.I.atn. The det. tile c f thi:, p.: spes;>l :.re c:n: t .inaJ in a docnnent dated i.tay l' F.'50 c ati<. led "/. proces:5 rad :c y.ne:.t for rt:4ti s t ance for a MULTI-19F101*F RW.CM 10.1 DMI' .'.;"D r.P?u.Ti RIU M.RCil AD TRAIiinM I:; NU"1.On hPACTOR TIK.!!!iTQUilS' to '.he I' .ticns.1 Scie:' .e rct ndrt'on frcm the
!:egcar s of the University of Cali:'oraf t.
Comittfqe Views and Recer.rrendatio :s Thi; ennutes are sum :ctizac r.s folh.;;m 1 rs N c..wi>* 1.ccc..li:n c:7 the Tci:;n r*..)c tor wa3 determined as junf et e t r. - ,rij:;.cre ic ;hs casentlc.;.ly campieted Argonsut rec.ctor, which ic housd as an t cf :%lre? ring un .i IMY. Both r.tactors would tt on the some J eve.7, i.c the i : :n muld he .<e.'. ) hel.n ,;round on two sides. Both ree.crors wou.1d be er.; h d b. .;he e;we pc wm.:1 ani supervisivn. This comittee had previously 0 ay,195f) cut.r%d r.::d approved the Argonaut installation. A comparison wcs mde betue.en the t.:o renetora rad in no esp.*ct was the Triga fcund to be more he.cdct.5 Un.n t!:e 8.gonac t- in ?act. -in al:.nst every respect, the safet/ rear;.in secn.M to fro th.? Tris re:.c -::>r, in spite of its higher possible operating Jevel. A problem distinct from hea1>.h and cafc ty in !;be effect on low level courtin:: In research areas. In this respect bachgroun.1 radi.2;; inn and active gaseous effluent were considered and found to be well belart the axcunt t;hich uculd bother low 1cvel counting in adjacent taildin.gs. Since lott Icrel counting requires a background considerably Icwcr than hec.ith considerations, it is automatic that there is no danger to personnel on the surrounding e t.mpus. The comittee therefore finds that the proposed Triga instr.11ation is perfectly safe from an overall campu.s standpoint, and for personnel direct 1j concerned it appears safer than the present Ar<;cmut reactor. Minutes of the Meeting Dr. Hicks described the propos Ed .lucati<n er.d hew 11 ru.11 1 c int er; tai:ed wi th the present reactor and the futurz baild..ag pr oarr.>6. Dr. McDonell provided the conaittee wi Gi de.wir,tiw 11M: ature on the Fcrk u Triga, pointing out that the ittrk G, F:.h is n e.hr 2 3;:ou: d dtsiga, is nuh more flexible and useful than the M : h rm 4 a ndu :.e ca k.s som: improves i I
page 2 of 3
' s u. n:. .
Stlfety featurce in ta.: bcv t m:&.;;;cc i~c J cc > tc.' d:d iP;: cc 95tiac Vith ~ copies of a "Uc:crds 2crcrt" prc A.u a5 oy Ger.cr cl .' ion 2 ..n f:r, N .?ce.ril rerhr.i".y gave the cornittee infct::htion CJ.:ut c .:dely inv;.r. ;! a,c 5..,t ;t the Unir:::it'. of Illinois wherc a .d:ciltx T: i ;a t 1c .c.tur ber bren ir MMid. '?he study d'c. c6 that the reactor w::n nafe to everyon 4r, t,atieft .' ion. l. u-l i. :c.) report oi ih::e investigation : w311 ha avni1rMe .".n ti.e n.%r frvure. Dr. M:.Denell govided the ca:mittce tii th c: pics of a iciter Jrc"1 H. R. i::irlin of G. nerr.1 !.tctics to the I.a .Jo18.n adecr.:, der rib!.ng t* e acfeta test procru on its prototype renctcr. Pertine r.t pr.r ts of this letter are quoted belcut "Encicaci is a dcrcriptien of the . LIGA Wri: H rccctor cf ..terest ';c the Univercity. Also enclosed in c. retort cf the first phase W an s:peric 2a nl procram conducted las t year by GA en its proto t'/Je vtacter. In this program the inhr. rent so.fety cf the TiUG.'s >:cc"Itirg fro:,1 the use of a unique ' fuel-rroderator clercent devalcred t / G1 brci been c:gerir: cat?.lly '.icrified. In thi'. pro,grn:r. we purposely rapidit/ ejec<.cd a control rod crpioying a special 1::echanisn to do this in en c.tter.tpt tu ci ttir.te en accici >nt. The reault, as predicted theoretical 3c/, tier:cn .trm:ed conclucively ihnt no hannxdc to personnel or equipc:ent Jere pendu.ed. d:dic.r esperi:.ats conducted at the Idaho National RcPctor Te.:tJn .P.; ti na en another tyce of reactor did, in fact, result in hanerdou.c ;crd! tic.as. Our experi nate ve.:e conducted uith persoratel in the in. red:icte d cinity of ihe .:c2: tor at cil times, whereas, those on the oth tr orpe of relcio:: were cotAnc.ed cith percomet approxir?.tely one ha1/ nile o . ,t .t iro:u the rc. .: tor bif 3 ding. 3 "Since the con:pletinu of the firat p'es a of thia pro 6rt:m Inst June, we have subjected the reactor to cr:re inten ce "cul.icf, ngain Uithcut produ;ing r+ny environ: rental har.ards. .
A sumnary of this r.ddi cien:;.1 enterlerco r.c . 6.1 as a E. t cf the 'rrf.1A reactors in operation or bei!vi ira bilad, rre c.Isc snac. sed. Of the domestic installations the r er.ctor nt th. Uni tcraity of Aris.cna is in daily operation in a fermer &afti.y : Irs,r:rcost in the ac3n engineecing building on the Univcesity crngr:s. The .:cactor at the tieterant. ledninistra--
tion Hospital in Onaha. is located in the bssencai. of an eleven story hospital and is indeed a unique ir.:strJ.intion with regard to safety. The reactor at Cornell as well cs ths.t at th : Uniicrcity of Illinois and KancaJ State are both being built on the respective enapuses." Dr. Clara Szego-Roberts raised the question about the active argon 41 preduced in the air around samples placed 8.u the reactor, and Asked herr this vo:11d affect low level counting in adjacent nrms. Pror the data in the Gerecci Atomics hazards report, the ccanittee detertuiurd that t raarer.ahle estinate of the rarica.m value that such background interferen.:e conid attain, would be Icas than .I'7, of natural cosmic ray background at a dir: tom of 150 feet fm.n $.he reactor, and hence would be almost impossible ta detect. There was no reason to re. nest thr'; Hic core of th reces be scaled to prevent escape of argon 41 (this was suggestcJ for fA.: t.rgont.nt rm.: tor.1 since the Triga core is inherently scaled by nature of in.: he.cac danitn.
pase 3 W 3
'I . Agenda P c ; .i
-u Dr. Hicks injected further evid : .:e of Icw expect:d ti.:hground b/ ntating that low level nature.1 craban 14 countin;; c.:.d ecntint:cus w:itoring of natural.
radicactivity rnd fallout is currenti/ proceedits neee 1.50 feet from the rocctor sit:: and is expected to continue with no interference after the reactor is in opera tion.
- Dr. McDonell suggested that safcty experts from General Ate:aics be im.:diately invited to visit the ccmpus and inspect the overall pj ens. The cc:mittec approved the suggestion and recommended the visit.
The chairman advised the comittee that cenbars from other capuses who cculd not attend the meeting had been adviced that the safety of the Tri 6a reactor 11as to be the subject of discuccion. The menbcra at I.a Jolls were the caly ones in any way faciliar teith the Triga. *ihe chairman determinoi that in general they regarded the Triga as very enfo, and asked that if they found any infor-mation to the contrary, that it be fo: Warded in tine for this meeting. No such information has been re.:eived.
/h [ hl'q.c./(N K. R.14ac;;ennie, C!.e. ira,tn Rc;p on.71.1 ivisar Cecnaittac on T:ndioic31 cal Safety Southern Section KRM:bp .
6 5-60 _r n. *
.e .
t ExhibLt 0-i-3 page 1 of 2
,-n
. .s 9e . ,
- fa8 .
g )o.(. 1
'sm .
rmmx::-= :=wc:c:n va a:.wcuamamm, . ,. .,
, CON-11 (Rev. 1). . 4 ADVANTAGES OF TRIGA FUEL FOR RESEARCH REACTORS October 1981
- aYI .sD.DEC& 6 FMy f g***gg%,y ,
+-~> . .m.a., - -.e w, . ,,, . , ._ , . . , ,
_,% , m. g ,. , ,
~
*, L r. *' f f' _ f*, ,
?)
4
..~.>.a..;. - :. -..- ~w_-,.
lw,l^ ..< . , . .... ~;
.~ .,
{ .. . m ;. w + ,m y3.a,,..,u:. v v,
, . - >: w
, m u... m
- . v. . e % .
C;gk %l9 s _ . ' . ,m_,..'s_, u_ _b.p
;t<tgltu'O_k_ ' *::.',% : .:e > .? .Y 'i ?'t ll . _._ ~ _ ,
- .m, <
s cg _ _
.X. .l;%;
, '.7. ,' + ? ..~ :__.,g~
. as;k;q._ _ _ ,'%
kA s'eLee4AMR bygMQM"Qg M_Qpg3upg3y_Q O
,e - r . -, ,
e-- . , . . ,. , ,,r- . e -- > . - - , - - - - >
9.
.. page 2 of 2 ..,. CON-11 (Rav. I)
ADVANTAGES OF TRIGA FUEL FOR RESEARCH REACTORS All TRIGA fuel is made by GA with a uranium enrichment of Just under 20% and thus is classified as Low Enrichment Uranium (LEU) fuel. The following discussion of advantages applies to TRIGA fuel clusters which can be Inserted in existing grid plates to convert plate-type fueled reactors, and some pin-type fueled reactors, to TRIGA. TRIGA's advantages have motivated the owners of twelve plate-type fueled reactors to convert to the use of TRIGA fuel. TRIGA LEU fuel is available now for use in existing reactors operating at steady state power levels to 50 MW.
- 1. UNIQUE SAFETY All of GA's research and test reactors are fueled with UZrH. This unique
. fuel provides the highest degree of safety available in any type of nuclear reactor. In these days of Increasing public c,oncern with perceived hazards of nuclear facilities, these safety advantages alone should justify use of UZrH fuel. A. The UZrH fuel has a prompt negative temperature coefficient of . reactivity, vs. a delayed coefficient in aluminum-clad plate-type fuel. _This allon UZrH cores to safely withstand accidental reactivity insertions that have completely destroyed plate-fueled cores. B. UZrH is chemically stable. I t can be safely quenched at 1200'C in water, while exothermic metal-water reactions take place with aluminum at 650*C. C. High-temperature strength and ductility of TRIGA's incoloy-800 fuel cladding provide a yield ; strength greater than 10,000 psi at 900*C. The aluminum cladding on plate-type fuels melts at about 650*C.
- 0. The UZrH fuel material has very superior fission product retention. The aluminum-clad plate-type fuels melt at 650*C, releasing 100% of the
- volatile fission products. Whereas, at this same temperature UZrH retains
( about 99.9% of these fission products even wi th the cladding removed. l 1
* . ~ ,
I.
/ / r e l l 'a i l s JC0 ' G = FC K . ' v .10 K ! .U = 0 610. ! C= O G IGE N .F . S = L ' . N S C C L A S i= U
// EXEC PGM=OPIGEN.NEGION=390K
, D //STEFLIE CC D I SP= sbr e C SN= JG t 013.MC K .CR I GE b .L C 2 ,
....4**._
**6 E X E C .. F UR T CL G . C O MI L 5 = F Cf(T H . R C= 2S O K . RL = 3 C OK R (. = 3 S.C K . P_G s.Lt100.1001.8.._ . . . _ . .
404 F ARN.L K CD= 'l I S T .O VL Y .M AP .S I ZE = ( 3 0 0 K .10 0 F ) ' .
- es40FT.5)SIA L[0H='fYSt.FORTLID'.LIDA='JGIOl3.MCK.CRIGEb.Lle' DD O'SO=SPR.DSA=JG1013.FCK. CRC.Duf
*ttKED.!YSLMCD CD O I SP= ( CL C. C A TL G ) .DSN= J G I C l 3. h C K . CR t.GEh..LC2 (CP I GEN J .
_ _ _ .4ee 448 SFACE=tTRK.(10.lc.S).ALSE). UNI 1=DAll. .. . _.. C C e = (LRFCL=0. REC F F=U . CSORG=PU.13LK S I ZE = 614 4 )
$$tkFnelIF CC D ISP =SHR . CS h=JG I O l 3.MC K .LR I G E A .L t e e t 4 kCD.! N flh 4e4 00 . D I S82 = S F.F . D Sh.=.J G I O 13. F C K . OR I G E A . C V LY- . , _ . . . _ . . . _ . .
//CC.FTC5FCCI CD DISP =SFF.CSN=JG1013.MCK.UCLA2
//CD.FTCCF(01 DD SY SOU T= A .OC B = (R ECFM=FD A .L RE CL= 133. E LK S I Z E=3 45 8 ) .
. ./ / . .. . . _UN I T = S Y.53 A . SP A C E.= LT P K , ( 200 200)) . . . . . . .
//CC.FT(FFCCI 00 DISP =SHF.DSN=JGIOl3.NCK.CRICEN.FILCl ld //CC.c1C7FCO2 DD D I SP= SH R .D SN= JG 1013.MC K.UR I CEN.F IL C 2 '
// CC .FT C 7FCO3 DJ DISP =SHS.DSN=JGt013.MCK.CRICEN.FILC4
. / / C C .F T C 7 F. C C4. DO DISP =SHE.0Sh=JGIO!3.NCK.CRICEN.FILC"
,' r/CC.FT(7FCCS DO DISP =SHR.DSN=JG1013.4CK.CRICEN.FILCC e/ CC .F T C 7F CCC C9 D I SP= SH R .DS Nv JG I O l 3.MCK .04 t C LN . F IL C 7 e/CC.F1(EFCCI CD DUMFY 4**C.rTCEFC0t DD 9 I S8)= ( N E h . P AS 3 ) . L A D E L = ( 01. 0 L P . , C L T )_ U h.1.T =_T AP E 6 E S O .
4+4 %CL=SER=CCNOS7. 4*4 CC8=(9ECF4=FO.LRECL=lJ0.ULKSIZF=39Co.DEh=4) l 4 8 9 C .F T C.e F C 01 00 D IS p= (N S W . C AT L G ) . DSN= J G 1013 .MCK .H ZP LT 1. bh 1T =Pu2L I C . 40* 444 C Ce = (Q E CF M =FO . L G E CL= 13 0..eL KS I Z E= 39 00 . D S C F C =P E).. SoACE=(TRK.(SelC).RLSE) .
//CC.FTSEFCCI CD D U M .V Y
//
5 .IEF22(1. #LLCC. =09 H w ! ! 4 2 T A . _ _ ,_ ... _ .. g. 81._ .. g IEF2271 E26 ALLCCA TED TC STEPLIO _ __ _ . f_ , ,.
*1CF227! 114 ALLOC A TE3 1C FTOSFOO! g
, IFF227I 116 ALLOCATED TO FT06 FOOL A /
IEF227I E26 ALLCCATEO,fC FTO7F001 ),[h I I?F2271 (2C
!=r2271 E26 ALLUC A TED 1C FT07F002 ALLCCA TED 10 Ft07F003
._ _ U, _ . V P
(
, DV( ff(A) + ff k IEF227[ (Ec ALLCCATCO TO FT0/F004 AI 8" i f),.3LbMd 15F227I E2d ICF2371 E26 ALLCCATE) TO FT07F005 AL L CC A TE 9 TC FT07F006 #
.__.b Ml4 - )h- 1 hw( -
f Qk g 9 w f
. . . . . . - _ . ~
n
* .i- o
.+
UCLA Rt REACTOR 20 YEAR FUEL CYCLE 93 2 EhEICHEC URAh!UM
- ECnER=. 0 3 0 NW .-.U UR h u P_= 18. MW D e.. FLU )= 2.10E+10h/CF**2.5EC .. ... . . ;
NUCLICE RA01CACTIVITY, CURIES BASIS = TCTAL HEAVY FETAL th%EATORY OF CCRE
' ~ ~ '
CbARGE" DISCH$RGE ~ " O '. C' O. 0 1. O ~f. D" ' " '3 0 . D SMt53 C.0 1 92E+01 1.92E+01 1 92E+C1 1.37E+01 1.63E+00 4.76E-04 Els t E 3 C.C 0.0 0.0 0.0 0.0 0.0 00 CC 15 3 __ C.. C ~._ , _. 7 . 3 0 E- 0 6 _ 7. 3 0E- 06. .7. 3 0E - C 6 . 7 2 E E - 0 6. 7.1.6 E .06. 6. 7 0 E- 06 .. .. . ~. _ . . . F F 15 4 C.0 6.49E+0! 4.83E+01 3.39E-C6 0.0 0.0 0.0 SF IE 4 C.C 00 0.0 00 0.0 0.0 0.0 EU154 C.C 9. 89E- C 3 9.89E-0J 9.89E-C3 S.8SE-C3 9.89E-03 9.86E-03 CD.1E4_.. C.. C _ _ _ . _ . . 0 . 0 . . . _. . . . _ .0 . 0_. . . - _ _ O . 0. .. . 00 . .._ _0 . 0 .. . . .,. . ... 0 0 _ . . - SMISE C.C 2 78E + C l 2.70E+0! 4 5 0E + C 0 2 97E-ta 0.0 0.0 EU 155 C.0 6.92E-Ot 6. 92 E- 01 6.93E-Cl 6.92E-01 6.88E-01 6.72E-01 GD155 C.C 0.0 0.0 00 0.0 00 0.C
. ._ _.5.F t 5 6 C.0 5. 4 t E + CO . 3 4 0EA00 5. 02E + C0 9.22E-Cl 2.25E-05 4 74E-23 EUIS6 C.C 3. 5 7E- 01 3.57E-01 3.b6E-C1 4.55E-C1 3.63E-01 1.25E-C1 GDISC C.C 0.0 00 0.0 0.0 0.0 00 S F 15 7 C.C 6 53EtCC 1.50E+CO 0.0 C.0 0.0 0.0
, __.EU157 C.C 2.12Et00 2.12E+00 2.02E + C0 7.CSE-C1 1.01E-03. 1 18E-14 GD 15 7 C.C 0.0 0.0 0.0 0.0 0.0 0.0 EUISE C.C 1.69E+00 1.66E+00 6. 79E - C l 6. 3 7 E- 10 0.0 0.C GD15e C.C 0.0 0.0 0.0 0.0 0.0 0.0
.. 3U 15 S C.C 9. 03E-- C 1 _S. 6 7E- 01 8 80E-C2 C.0 00 .. .0.0 CD 159 C.0 2.435-01 2.43E-01 2.47E-Cl 1 02E-01 4.COE-04 2.35E-13 Te t!S C.C 0.0 0.0 0.0 0.0 0.0 0.0 EU16C C.0 2.53E-Cl 1.a9E-01 1.32E-C8 C.0 0.0 00 GDtcc .C.C 00 _ ._. _ . 0 0 00 0 0.... .0.0.
8.02E-05 . 0.0 -{ l 1016C C.C 8.02E-05 S.02E-C5 7.94E-CS 7.50E-C5 6.0tE-05 CYt6C C.C 0.0 0. 0 0.0 0.0 0.0 0.0 -
, GClet C.0 6.43E-02 5.26E-02 7. 72 E- C 7 C.0 0.0 0.0 T D.161 C.0 3. 64E- 0 3 _;.3. 64 E 03 . 3.65E-C3 3 3LE-C3 1. 81 E- 0 3.. ... ! . 8 0 E- 0 4 CY101 C.C 0.0 0.0 0.0 0.0 0.0 0.C CC162 C.C 3.29E-02 3.07E-02 5.85E-C4 C.0 0.0 0.0
, THIc2M C.C 3.29E-C2 3.2aF-02 1. 78E - C3 C.0 0.0 0.0 Je162 C.C 0.0 00 0.0 C.0 00 0.0 CYlc2 C.C 0.0 0.0 0.0 0.0 0.0 0.0 Te163F C.0 1.52E-C2 1.37E-02 3. S!E- C5 C.0 0.0 0.0
' T0163 C.C 4.09E-06 4. 0 7 E- 06 3.660-C6 2 15E-07 6.76E-14 0.0
.._C_Y.16.2__.....C.C 3.0 .. . . . 0 . 0 _ .. . _ 00 . 00 00 .. ....... . 0. 0 TD164 C.0 1.19E- C3 1.18E-03 1.14E-C3 5.72E-C4 7.47E-06 4.45E-13 CYt64 C.C 0.0 I 0.0 0.0 00 0.0 0.0 CYICSF C.C 9 35E- 05 5.20E-05 3.33E-19 C.0 00 0.0
__.. C Y 16.5_ . ... . C. . O 2. 92E- 0 5.__.2. 9 t E 05.. 2.16E- C5 2.27E-Ce 4.42E.27...0. 0 F0 tt! C.C 0.0 00 0.0 C.0 0.0 0.C CY166 C.C o . 39C- 0 8 9 39E-06 9.3tE-Ce 7.6ed-Ce 2.25E-08 Pot 06M C.0 3.94E-13 3.94E-13 3 94E- 13 2.9 4 E- 13 3.94E-13 2.06E-10 3.94E-13 g P.O.t 6.6 .._..C . . C 0 34 C.- 0 0 ._0 8.4 E. .08 8.86E-C0
. 0 .6 4 E - C b .., _3. 2.9.17.- 0.Q _3.,. 0 7 F, .,13. _ , , , _ N ERL66 C.C 00 0.0 0.0 C.C 0.0 0.0 ER167 C.C 0.0 0.0 .O.0 0.0 } C.0 0.0 TUTAL C.C 3.44E+C5 2 29E+05 6. 73E + C 4 9,.43E+03 2.49E403 1 27E403
...._.-_ .. .. . *~ - = = ~ * - -. . - ~ . - ~ = * * - = = ~ _ - _ ~ . = . -
.M
? OJE L 1 REACTCR RUNAWAY CR SUPERCRITICAL NUCIIAR EXCURSIGN ACCIDENTS "An outstanding characteristic of nuclear reactors is their potential ability to achieve extremely high power levels in a short time if adequate control of the machine is lost. l A typical nuclear runaway accident may start and be over in times appreciably less than a second. In this respect they are different from any other large-scale machines..."
- "The Safety of Nuclear Reactors" 1955 Geneva conference Paper by INTRCDUCTICN McCullou6 h , Mills, and Teller 1 What is a Feuer Excursion Accident?
- 1. A supercritical power excursion or " reactor runaway" is an accident unique to nuclear reactors- in which power can rise out of control, from zero to billions of watts, in very much less than a second. This can result in melting, and even vaporization, of the fuel and explosive disassembly of the reactor.
- 2. In such an accident, the power rises exponentially in periods measured in milliseconds, fuel temperatures shoot past the melting temperature,
, and steam explosions and often metal-water chemical explosive reactions can I destroy the core. Fuel fragments may be sent high into the air, as in the photo below, and may be scattered hundreds of feet around where the core had been. l Pellet of spongy aluminum-vranium mixture
- s
.s.
$ 175, }
v* s . s . l ._ Disassembly of the BCRAX reactor Frogments of fuel plates BORAX reactor fuel fragments 2
r-
*' _ . . . ~ _
3 Eecause of the speed (periods measured in milliseconds) with which these accidents can occur, there can be no time for operator response or automatic engineered features like control rod insertion to prevent the runaway. Protection must rest on inherent design features which, nearly instantaneously, must dampen, stop, and reverse the rapid power rise before the fuel is damaged. The effectiveness of these inherent design features, and the accuracy with which they can be predicted, are extremely important in safety analysis of reactors. Despite detailed safety analysis prior to operation of some reactors, there have teen some very unfortunate surprises. A Case Study: The SL-1 Reactor Accident 4 One of the most grisly power excursion accidents occurred on January 3, 1961, at a small reactor built at the National Reactor Testin6 Station (NRTS) at Idaho Falls, Idaho. The reactor, known as the Stationary Low-Fower Reactor Number One (SL-1 for short), had been built for experimental and training purposes. It had many features similar to the UCLA reactors fuel that was highly enriched, made of uranium-aluminum alloy and clad in aluminum, formed into flat plates bolted together as bundles, water-cooled and
-moderated. (There were differences as well-- the effects of similarities and differences between the UCLA Argonaut-type reactor and the SL-1, the 30?AX, and the SFERT reactors will be discussed in more detail belou.)
5 At the time of the accident at the SL-1, it had been shut down for routine maintenance, and three men were completing certain preparations for nuclear startup. Apparently, in the process of attaching control rods to drive motors, one of the men raised the central control rod too far and too fast, triggering a supercritical power excursion in the reactor core. In a fraction of a second the power reached a magnitude of an estimated several billion watts, melting and perhaps even vaporizing a large part of the core. The water in the core region was vaporized, creating a devastating steam explosion. The remaining water in the reactor vessel was hurled upward at high velocity, striking the underside of the reactor's lid and lifting the whole nine-ton vessel upward, shearing cooling pipes in the process and crushing the men who had been on top of the reactor vessel against the ceiling of the building before the vessel dropped back into place. One of the men remained impaled on the ceiling by a piece of control rod rammed through his body. It all happened in a second or so. (Even had the worker, as he pulled the control rod out, realized he had pulled it out too far, there would not have been time to push it back in l before the reactor exploded, so fast is the exponential power rise in a l runaway reactor.)
- 6. It was a terrible accident, made even more grisly because the intensely radioactive fission products scattered inside the building by the accident hampered the work of recovering the bodies. Staying in the building for mere seconds resulted in a year's allowable dose of radiation for the rescue workers. And it took several days to remove the body that was impaled on the ceiling, requiring use of a remotely operated crane and closed-circuit television because of the intense radiation fields. The bodies were so badly contaminated, the heads and hands of the victims had to be severed and buried with other radioactive wastes at the NRTS.
s.~ .- . :. - . . _ . - _ . . .
. The Lessons of the SL-1 Accident
- 7. The accident at SL-1g and other similar accidents, point to the inherent (and often unanticipated) dangers in certain reactor designs, and the extreme importance of assuring that safety analyses have not overlooked design flaws that could lead to or contribute to serious accidents such as the one at SL-1.
- 8. This is especially true when considering a reactor such as UCLA's, located without a containment structure or exclusion zone in the midst of a densely populated area, where the consequences of a serious accident
~
could be far greater than the accidents which have occurred to date. Virtually all of these occurred, fortunately, in remote locations with considerable distance between them and any. populated centers.
- 9. Therefore, review of potential accidents must ensure, to a very high degree of certainty and with very large safety margLns, that the UCLA 4
reactor is inherently safe from destructive power ' excursions--i.e., that intrinsic self-limiting features would prevent a power rise of sufficient , magnitude to do to the UCLA reactor what the incident at SL-1 did to that - reactor. Varying Effectiveness of Inherent Safety Feature _s 10 Certain research reactors, particularly the TRIGA, seem to meet, to a significant degree, the standard of inherent safety as to power excursion potential. In the case of the TRIGA, this is because part of the moderator is built right into the fuel, so that there is no time delay in the shutdown mechanism. (This will be explained in more detail later.) With the TRIGA, if an excursion occurs and power starts increasing, fuel temperatures increase as well, which promptly cuts off the power rise because the shutdown mechanism
- is built int'o the fuel itself. There is no time delay. (Reactors that use low enriched fuel have a similar prompt component to the shutdown mechanism, Doppler broadening involving neutron-capture in the Uranium-238 which is 4
likewise part of the fuel itself. This too will be discussed in more detail later.)
- 11. The shutdown mechanisms in the case of the UCLA Argonaut-type reactor, on the other hand, are far slower in operation.* The heat that is generated in the fuel meat by the power rise must be transferred through the meat to
. the cladding, through the cladding to the water coolant, and then that heat nust vaporize sufficient water to produce enough voids (steam bubbles) to i reduce moderation and stop the reaction. For a range of power excursichs, this shutdown mechanism is too slow to prevent the runaway from reaching power levels (and thus temperatures) sufficient to melt the fuel. And,
- Furthermore, a number of safety features with which the Argonaut was originally designed have been altered or have eroded over the years, thus further reducing safety margins at UCLA. For example, the UCLA reactor was originally designed with an excess reactivity limit less than that necessary to go prompt critical: a few years later the excess reactivity limit was increased substantially, to precisely the level that the calculations in the UCLA Hazards Analysis indicate could cause fuel melting. More on this below.
r- .
*n -u .,_s-,a..- _ . . _ . . ,_
_4 as indicated earlier, such power excursions can occur far faster than either human operators or automatic engineered safety features can respond.
' Unpredictability of Destructive Threshholds
- 12. For reactors that don't have the inherent safety of extremely prompt shutdown features, such as those of the TRIGA, there is consideIable uncertainty where the danger point lies. A wide variety of factors (such as fuel meat and cladding thickness, void and temperature coefficients, among others) will permit one reactor to tolerate safely an excursion of a particular magnitude while causing another to be totally destroyed. Further-more, predictability for even the same reactor is not terribly good.
- 13. To obtain a better understanding of the mechanisms of self-shutdown, a series of tests were performed on certain reactor designs at the NRTS in the 1950s and 1960s. Several of these reactors were tested to destruction--
i.e., permitted to run away at rates sufficient to cause fuel melting and core explosion. One such was the ELRAX: another was the SPERT. These tests taught us much about the mechanisms of power excursions in certain kinds of reactors, although they revealed large uncertainties as well. 14 Perhaps the most significant conclusion of the SFERT destructive tests was the unpredictability of destructive threshhold during power excursions, even with a reactor that had been as thoroughly studied as SPERT had been. While fuel melting.had been expected in the final test (because melting had been observed in previous tests with somewhat smaller reactivity insertions), the violent explosion which demolished the reactor came as a surprise. Although the ECRAX and SL-1 reactors had suffered similar explosions, there had been no prior indication at SFERT that going to a period slightly smaller than that of previous tests represented crossing a threshhold for SFERT which made possible the violent pressure pulse which would demolish the core.* 15 Even after an extensive series of actual tests with the SPERT reactors, there is much about the behavior of those reactors during power excursions that remained poorly understood and difficult to predict. This is consider-ably more the case with regards the potential behavior of reacto;s substantial-ly different from those tested-- for example, the UCLA Argonaut. Dangers of Extrapolation from One Reactor Type to Another
- 16. The uncertainties are vastly greater when comparing an excursion of one magnitude in a particular reactor, not against an excursion of a different magnitude in the same reactor, but against a completely different reactor,
- As T.J. Thompson put it, describing the SFERT I destructive test, "The sudden onset of total core destruction for only a factor of two increase in
. total energy deposited was a surprise. It emphasizes the need to carry on such extrapolation tests at remote sites such as NRTS." (Technology of Nuclear Reactor Safety, " Accidents and Destructive Tests," p. 685).
~5-with different design features and nuclear characteristics.- It is those uncertainties, the extremely large error margins that must be considered in applying data from the 3CRAX and SFERT tests to a reactor such as the UCLA Argonaut, whose fundamental design and key variables are quite different, which represent one of the key matters to be addressed.
17 Several analyses, relying heavily on the SPERT I tests, have been performed purporting to predict the potential behavior of the UCLA Argonaut-type research reactor during power excursions that might be initiated by insertion of that reactor's available excess reactivity. Because of the large amount of excess reactivity requested (far more than that capable of producing supercriticality on prompt neutrons alone), and because of the highly populated site and lack of a containment structure, we have paid special attention to those portions of the documents which attempt to analyze the capacity of the UCLA Argonaut reactor to undergo a destructive power excursion, one that could result in release of fission products to the environment.
- 18. The three key attempts to apply 3CRAX and SPERT experience to the UCIA Argonaut case have been examined. In each case, correction of just a few erroneous or non-conservative assumptions or values demonstrates the potential for a seriously destructive power excursion at the UCLA reactor, assuming the correctness of the basic method applied in the original analyses.
- SFERT-ID Destroyed by $3.50 Insertion
- 19. For example, as will be discussed in more detail below, UCLA has asserted that its Argonaut-type reactor can safely tolerate a far larger excess reactivity insertion ( a measure of the initiating event for a power excursion, often measured in " dollars") than its original design limit.
UCLA appears to rest most of its case in that regard on the assertion that the 3CRAX I and SPERT I tests " proved" that that level of excess reactivity is safe in the Argonaut. As UCLA put it in its 1980 Application (p. V/3-6), with similar statements in the 1982 versions "SFERT and 3CRAX tests showed that plate type fuel elements survived step reactivity insertions of $3 54."
- 20. That simply isn't correct. In fact. the SFERT-ID reactor core was completely destroyed by a $3.50 insertion, which resulted in explosive disassembly of the core and extensive melting of the fuel, as can be seen from the attached photos from that SPERT excursion.3
- 21. However, even were it correct that SFERT survived a $3 54 insertion, which it did not, that would in no way mean that the UCLA Argonaut, a different reactor, would likewise survive such an excursion. Extrapolations I from 5FERT or 2ORAX experience to the UCIA Argonaut require numerous corrections for different reactor characteristics, each correction increasing the magnitude of the margins of error that must be assumed.
v -- . . - - -
. _ . . m._ _ . _ _ . - . . . .__ - _ . _ _ -
Sumnary of Conclusions
- 22. Based on our review, it is concluded that the amount of excess reactivity requested by UCLA is too high, the safety margins too small, and the potential for a destructive power excursion unacceptable, especially given the local population density, lack of containment structure and exclusion zone, and student operation, among other factors. The analyses done to date do not, we believe, demonstrate that such an accident is not credible. In fact, because of errors made in each, the analyses indicate, when the errors are corrected, that such an accident is indeed credible.
Questionable methodological assumptions employed by the analysts suggest that a definitive answer as to the maximum " safe" reactivity insertion for the UCIA reactor, or even an answer merely providing reasonable assurance of its being right, would require further research.
- 23. Because of the substantial differences between the Argonaut and the reactor types previously investigated, that research would likely necessitate SFERT-type tests on actual Argonaut cores. In the absence of such definitive research, very substantial margins of safety are essential. Furthermore, the changes made since the UCLA facility began operation have, in our opinion, resulted in a gradual but significant erosion of important safety margins, making a potentially serious accident at the UCIA facility both more likely and of greater potential consequence.
24 In uhat follows, we discuss the tasic nuclear concepts related to power excursions, review the analyses done to date, discuss some of the ill ' advised alterations to the UCIA Argonaut over the years, and indicate some of the many potential mechanisms for initiating such an event. 25 We conclude that a series of unfortunate design features and post-design modifications conspire to make the UCIA Argonaut, with its current characteristics, not inherently protected against destructive power excursions of the kind that destroyed the SL-1. Adequate margins of safety are lacking. i
l . _ -.
. . ~ . . . . -
l i i l i
; hh L2 8 t 2 3_es k .f. g; 3 1
- t. 1 E 'i Y ..
?
La
. 5 c_ :2 _ .
t* 5 $'
\ vd l J $
3
?~. .hY c
, t -e l
-. y $n
.a .
, j . . . ^ ' , .'
5,l ' LS:
> .,5# h -
?!s
.+
fu-", h fi 3 e E-- i . rs lA- __ 4 Fuel plate end fragments from assembly D5. (fromSFERT-Idestructtest)3 __ . ,0. , . .. _ .L'. . ._- - - - - - - -;
l - - - - - . . - ._ _ . \
- l
\
- P4 -
- 4 s
{ t t 4, r@> l 5- ?$ f-E5'
^
it h
. fi f-f! l
/ l 9-
- k. ,
l Li l l ' )
,# f 1 t t
s
. R r- -
5 EJ Pi \ R r, i l 1
.'{
~
.. .-g.
- x?
,I
~;
ll
\
5 j7 W v c :. ., e
,_s Typical plates from peripheral assembly-edge triew, position G6.
(from SPERT-I destruct test)3
- .. .'a-
-- . . _ _ _ = -
. 3RIEF EXFLANATICN CF 3ASIC NUCLEAR PhT3ICS C.,NCEFTS RELATED TO FCWER EXCURSICNS
- 26. This section constitutes a brief explanation of some of the basic concepts of nuclear physics related to power excursion phenomena.
3revity will necessitate considerable simplification. Criticality and Delayed Neutrons
- 27. A reactor such as the UCLA Argonaut-type reactor operates by fissioning atoms of uranium-235 uhich, when hit by neutrons of particular energies, will split apart. In the process of fissioning, uranium-235 atoms release additional neutrons which can then cause other uranium-235 atoms to fission, creating a chain reaction, which can be self-sustaining.
- 28. A reactor is said to be critical when the number of neutrons in one generation is equal to the number in the next. In such a situation, k, the criticality or effective multiplication factor, is said to be equal to unity. When the chain reaction is increasing with time, so that the number of neutrons in one generation is larger than the number in the previous, the reactor is said to be supercritical. A supercritical reactor is thus defined by k >l. k(1 means the reactor is subcritical.
29 A nuclear chain reaction runs on two different kinds of neutrons: prompt and delayed. By far the majority, approximately 99 35% for UCLA, are prompt, being produced virtually instantaneously at the moment of fission. A small fraction, approximately 0.65% are delayed neutrons, i.e. neutrons produced by decay of the fission fragments created by the nuclear fission process. These delayed neutrons are produced in periods of time
- ranging from microseconds to hours. If it were not for the delayed neutrons, mechanical control of a reactor would be impossible.
- 30. Neutron generation times are measured in milliseconds; if the only neutrons upon which the chain reaction is based were prompt neutrons, there
- would simply not be enough time for either human or mechanical intervention to prevent a runaway condition. Growth in reactor power (a function of growth in the neutron population) is essentially exponential. Even a small rate of growth from one generation of prompt neutrons to the next could cause reactor power to increase to such a level that fuel melting could occur long before human or mechanical intervention (e.g., insertion of control rods) could be completed in order to prevent such a runaway condition.
Supercriticality and Prompt Neutrons
- 31. Such a runaway condition, where the neutron population grows uncontrol-ably, is called a prompt supercritical power excursion. If the power excursion is severe enough (i.e., if mechanism can be activated) , power gets high fuel melting can enough before possibly occuraas shutdown well as a steam explosion or explosive metal-water reaction. It is very important, l
- a. __, _ . . . . _ _ . , . . .- . . ~ . _ _ _ _ _ . . _ --
l therefore, that a reactor not be able to run away at a rate faster than its shutdown mechanisms can responds i.e., that it not become
" prompt supercritical," or supercritical on prompt neutrons alone.
- 32. When a reactor is running on delayed neutrons-- i.e., when the reaction needs both the 99 35% of neutrons that are prompt and the 0.65% or so that are delayed in origin-- the delayed neutrons provide a margin of safety that permits intervention of control rods or other shutdown features. in time to prevent an increase in power that is so rapid that melting can occur. The delay required for generation of these neutrons provides time for electronic indicators to report an abnormal growth in neutron population, inform a control operator who can take appropriate actions or activate an autocontrol which can mechanically do likewise.
33 However, when a reactor is running on prompt neutrons alone, it has lost that protection. An increase in neutron population can occur so suddenly, and continue to increase exponentially so rapidly, that inter-vention by human response or engineered safety feature is not possible. Thus, this situation is strongly to be avoided. Exponential Increase in power in Milliseconds 34 When a reactor is supercritical, i.e. k >l, meaning that each generation of neutrons is larger than the previous, the power rise is exponential. The exponential period (T) is that amount of time it takes the power to increase by a factor e, or approximatel 2.718. Thus, in five exponential periods, the power would rise by e or about 150 times,for example. The ability of reactor power to rise astronomically on a very short period is thus evident, and explains uhy supercriticality on prompt neutrons can be so dangerous. 35 The effect of the delayed neutrons is to elongate the exponential period T quite substantially, giving time for human or engineered features to come into play before the exponential imperative brings a dangerous power level. But as a reactor approahes prompt supercriticality, the exponential period becomes exceedingly short, making possible massive power rises in very small fractions of a second, given by the following general equations t P = e7 o where P is the initial power and P is the power after the lapse of time t. o For very short reactor periods T, then, very large power rises can occur in very short time intervals t. And when a reactor is supercritical on prompt neutrons, the period T becomes exceedingly short.
- 36. Thus, it would be quite incorrect to assert that prompt' critical is just another point on the curve. Near prompt critical, the exponential period jumps from a manageable range measured in seconds or hours to periods measured in milliseconds, making engineered safety features such as control blades and dump valves potentially useless should the excursion go unchecked. (Any remaining intrinsic safety features will be discussed shortly.)
L .- -. .- ,- _ . l Excess Reactivity 37 A reactor can become supercritical on prompt neutrons-- or go " prompt critical"-- if sufficient " excess reactivity" is inserted in the reactor core (e.g. through addition of extra fuel or moderator or through removal of neutron absorbing materials such as control blades or samples that have been inserted into the core for experimental irradiation). The effect is the same whether positive reactivity is added (by dropping, for example, a sample of uranium-235 into an irradiation port) or removing a negative worth sample by pulling it out of the core-- the reactor " sees" the same thing either way, a flood of extra neutrons, which cause more fissions, which produce more neutrons in the form of the expanding chain reaction.
- 38. If sufficient excess reactivity is added (or negative reactivity removed) so that the delayed neutrons are no longer needed to get the reactor critical, the reactor is then critical on prompt neutrons alone and the exponential period, or e-folding time, becomes very short. Power is thus increasing on such a short period that there would potentially be no time to stop the reaction by mechanical means such as operator response or scramming of control rods automatically.
" Dollars" and f dk/k 39 The capacity of a reactor to go supercritical on prompt neutrons is measured in terms of the available excess reactivity. For a reactor j to be just critical requires, as we indicated above, k = 1. How much reactivity is available to push the reaction beyond just critical is the excess reactivity. Because the delayed neutrons represent approximately 0.65% of the neutrons in the UCLA reactor, if one adds excess reactivity.
of 0.65% or more, the reactor will be supercritical on prompt neutrons alone. The delayed neutron fraction is called # (beta) and that amount of excess reactivity is sometimes measured in units called dollars, with
# = $1. If the percent notation is used, the units are in percent of l S/k.
I
- 40. In short, if a reactor such as UCLA's has available more than
$1 or 0.65% ak/k excess reactivity, and that excess reactivity were for some reason inserted into the core, the reactor's normal engineered safety features such as control blades might be unable to shut the reactor down before power levels sufficient to melt the fuel were attained and the contained fission products released. It is for this reason that the original design of the UCLA Argonaut prohibited the reactor from ever having available more than 0.6%21k/k excess reactivity.
Inherent Shutdown Features Affecting_ Termination of Super-prompt Power Excursions
- 41. If a reactor were to go supercritical on prompt neutrons, power would keep increasing essentially exponentially until the excess reactivity were somehow removed. As indicated above, engineered safety ' features are too slow-acting to be of use in stopping very fast excursions. The only remaining means of shutdown, then, are inherent features-- fast acting, non-mechanical automatic responses not dependent upon operator action, but rather intrinsic design and the laws of physics.
" Doppler" Effect 42 For example, with reactors utilizing low enriched fuel, the neutron capture rate will undergo " Doppler broadening" as temperature in the fuel rises. With increased neutron capture, the number of neutrons available for fissioning is reduced, dampening the power rise. Unfortunately, this effect is virtually nonexistent in highly enriched fuel such as that used currently at UCLA, and is a strong argument for converting the fuel at UCLA from highly enriched uranium to low enriched uranium, for safety in addition to non-proliferation reasons.
- 43. Use of a low-enriched fuel would add some safety margin to the facility, because of the increased Doppler effect. At SFERT, it was found that a low-enriched, uranium oxide core was able to withstand larger reactivity insertions than the highly enriched uranium-aluminum plates. A reduction in enrichment of the fuel can thus decrease the possibihty or consequences of a destructive power excursion or criticality accident. Low enriched fuel has a contribution to limiting an excursion, an abating effect, that is the cancelling out of reactivity because of the increased capturing of neutrons by uranium-238. Highly enriched fuel, of course, is highly enriched in uranium-235, with very little uranium-238 present, to the Doppler effect is essentially not present in such fuel.
Moderator Heating. Void Formation 44 Cther factors which can help terminate a power excursion are heat transfer to the moderator, which reduces the effectiveness of the moderator (for certain moderators), or void formation in the moderator (creation of steam, for example, in a water moderator), or expulsion of the moderator (as from a steam explosion). Thermal reactors, like most power and research reactors, require a moderator to function-- some substance that slows down the neutrons to increase the probability of their causing fissioning in adjoining U-235 atoms. Without the moderator, or with less of it available, the reaction can't keep going. Final Shutdown Mechanism: Disassembly of the Core 45 The final shutdown mechanism is disassembly of the core. (A destructive power excursion, in fact, is sometimes referred to as an "RDA" or Rapid Disassembly Accident.) Essentially the energy rise is so rapid and so large that the core explodes. This happened in the final SPERT and ECRAX tests and in the tragic SL-1 accident.*
- 46. Until one of these factors comes into play, however, the power will continue to rise, the fuel temperature will likewise continue to rise, and substantial release of fission products and energy is possible.
The core destruction and fuel nelting that can ocur from such an excursion is indicated by the photos, films, and videotapes of the 3CRAX I and
*In an extreme case, e.g. the atomic bomb, negative reactivity is introduced because of the rapid expansion of the plasca, governed by the equation of state, which self-terminates the chain reaction.
13-1 SFERT-ID final power excursions. Inherent Safety: Very Large and Very Prompt Negative Coefficients of Reactivity
- 47. The goal of a research reactor designer, particularly one whose reactor might be operated by students, was to design a reactor with a very high degree of inherent safety. A reactor with inherent safety is one in which features involving the very nature of the reactor itself j can limit a power excursion without the necessity of appropriate response by the reactor operator or appropriate function of the reactor's engineered safety features. To have a high degree of inherent safety, 3 the reactor needs very large and very prompt negative temperature coefficients, so that when the power rises, and the temperature accordingly, the temperature rise automatically shuts the reactor down before damage can occur. Inherent safety features are the very last line of defense in a reactor which can go prompt critical, the only defense in fact, other than administrative controls (which can't be counted en at a trainingreacter).
- 48. Degree of inherent protection, i.e. the magnitude and promptness of self-shutdown mechanisms, varies widely, reactor to reactor. In some reactors, these reactivity coefficients are occasionally positive, creating potentially dangerous situations where the ret.ction feeds on itself rather than providing a measure of self-control.
49 These positive reactivity coefficients create positive feedback loops, so that as the power rises, so does the temperature which, rather than force the power back down, pushes it even higher. This can obviously te potentially quite dangerous. Often a reactor will have several reactivity responses, some of which will be negative and some positive. Even if the positive coefficient does not predominate, it can severely limit the effectiveness of the negative coefficient, permitting a far longer power rise before the negative coefficient succeeds in producing shutdown, and thus far greater energy release and higher temperature for the fuel. The UCLA reactor appears to have several such positive reactivity coefficients, as will be discussed later. Some Reactors' Inherent Shutdown Features Eore Effective Than Others': Case in Point, the TRIGA l
- 50. Some reactors' inherent shutdown mechanisms are vastly more effective and reliable than others'. The TRIGA reactor, for example, has part of its moderator built right into the fuels thus there is virtually no time delay in the negative temperature coefficient (which is very large in the TRIGA) taking hold, because there is no delay in transferring the heat to the moderator.
- 51. The main feature of TRIGA research reactors and the TRIGA fuel-moderator elements (even when used in reactors originally designed for flat plate fuel) is a very strong and very prompt negative temperature coefficient, much more prompt than that of other research reactors, which effectively controls large prompt positive reactivity insertions. Any sudden increase in power heats both the fuel and the moderator simultaneously,
-+ --
--.,,-,,-,r, , , , v - . , -
---,my,m---.g .y n - r , - - - --- , - - se -e
^ - l l
causing the moderator to become less effective immediately and to return the reactor automatically and instantaneously to normal operating levels. Such control is intrinsic to the TRIGA reactor fuel and does not rely on mechanical or electrical control devices. This most important property is due to the fact that the fuel elements are constructed of a solid homogenous alloy of uranium fuel and zirconium hydride moderator, making them " fuel-moderator" elements.
- 52. There is, of course, a level of excess reactivity above which that safety feature of not being able to damage any of the fuel with an accidental excursion is no longer true, but that level would be much, much higher for a TRIGA reactor or a reactor with TRIGA fuel than for the UCIA Argonaut with its current flat plate ETR-type fuel. At comparable levels of excess reactivity, the TRIGA fuel would definitely have significant inherent safety advantages.
- 53. In the TRIGA fuel, when the ratio of captures (in the water and other materials) to fissions (in the fuel) goes up, the reactivity goes down.
That effect is produced by a change in temperature in the fuel itself, relative to the cooling water, and thus requires no heat conduction. It happens instantaneously because the heat is liberated by the fission reaction right in the fuel. In the UCLA Argonaut-type reactor, the heat has to be transferred to the water, which takes a while, to make it expand. It is the expansion of that water plus some other effects that have to do with the water having to heat up that makes the reactivity go down. Because of this time-delay involved with the transfer of heat from the fuel meat to the water in the UCIA reactor, the shutdown mech-anism is slower. his allows for greater energy release before shutdcun for an excursion of the same exponential period, and a greater opportunity for fuel melting to occur before the excursion terminates than is true l with the TRIGA reactor or reactors converted to TRIGA fuel. I
- 54 The shutdown mechanism in the UCLA Argonaut, which requires transfer i of the heat to the water to cancel the reactivity, can produce effects in the water like -boiling or a sudden expansion of the coolant which can, in effect, do some damage even if fuel melting does not occur. The j likelihood of changes in the fuel arrangement or other core rearrangement l- or damage is less for the TRIGA than for the UCLA Argonaut, for the same l reasons that the TRIGA is considerably more protected against excursions l leading to melting than is the Argonaut.
The UCLA Argonauts Less Effective Shutdown Features 55 Cther reactors, such as the Argonaut, ha've inherent self-limiting features far less prompt and effective than research reactors like the TRIGA. De original Argonaut reactor, and some of the university Argonaut-type reactors that followed it, used 20% enriched fuel -providing some prompt Doppler contribution to reactor shutdown in an excursion. But for the UCIA Argonaut, utilizing as it does highly enriched plate-type fuel, with little U-238 and no moderator in the fuel meat itself, essentially the only inherent self-shutdown mechanism short of full core disassembly that can limit an excursion is: transfer of the rising heat from the fuel meat to the cladding to the water moderator and then formation of steam and expulsion of the remaining water. This void formation /waterexpulsionreducesmoderation,increasesneutronleakage, and eventually stops the reaction.
- _ _ _ ' _ _ _ _ . . _ _ . _ _ _ . _ . _ . . _.a._ , _ .. - _ - .
]
56 In such a reactor, this last remaining shutdown feature is far slower than that of the TRIGA reactor, in which there is practically no time delay necessary for heat transfer. The delay involved in transferring heat from the fuel meat to the water, for excursions of short exponential period, can prevent self-shutdown occurring before the reactor has reached a dangerous level, resulting in fuel melting and possible explosion. It is very difficult to estimate with a high degree of certainty for different reactor designs what the precise lim-iting period would be, i.e., at how short a period the formation of voids would cease to occur in time to be effective in preventing fuel damage. The effect depends upon a wide variety of variables (plate thickness and conductivity, surface area-to-moderator volume, coolant channel thickness, size and sign of void and temperature coefficients, and so on).
- 57. The UCLA Argonaut-type reactor is now left with this one, delayed inherent shutdown feature (void formation in the water), lacking the Doppler effect of low-enriched fuel of the original Argonaut, the very large, very prompt negative coefficients of the TRICA, and the protection provided by its own original design limitation to less
, excess reactivity than that necessary for prompt criticality. Further-more, compared with the ECRAX and SPERT reactors, upon which the primary reactivity tests have been performed and against which it has been com-pared in safety analyses, UCLA's sole remaining shutdown mechanism of temperature effects in the water coolant / moderator is far less effective in limiting a power excursion than those effects were for 3CRAX and SPERT.
- 58. Reported void and temperature coefficients for the water moderator are considerably smaller for UCLA than for BORAX or SFERT. A principal design difference between the UCLA Argonaut and the BCRAX and SPERT reactors is that the moderator and reflector for the latter was solely water, whereas the UCLA Argonaut has, in addition to a water moderator, large volumes of graphite as moderator and reflector. While the sole moderator at BORAX and SFERT can be readily expelled through steam formation, only part of the moderator can be expelled at UCIA. The solid graphite will remain, reducing the effectiveness of a shutdown mechanism depending upon void formation.
Positive Feedback Features
- 59. In addition to relatively less effective negative feedback features, the Argonaut appears to have certain unfortunate positive feedback features.
60 For example, the graphite has a positive temperature coefficient of reactivity. As it is heated, instead of dampening a power rise as is the case with water moderators and reflectors, the power rise is enhanced. The positive reactivity coefficient for graphite has been known for years, at least in fact since the Manhattan Project era, and it is something of asurprisethatUCLAgidn'tknowofitandhadtolearnofitsexistence from another reactor. This is a good example of the importance of reviewing reactor experience from similar reactors for potential generic problems.
- 61. Another example: If the reactor water is dumped and the reactor is not scrammed, power rises at first and only thereafter decreases.
This is due to over-moderation of reflected neutrons in a portien of the. core region. As water level drops, the over moderation decreases, and reactivity increases, until ideal moderation is reached; if the water level drops further, the reactivity trend reverses and finally starts to drop.5 (In other areas of the core, the reactor is severely undermoderated so that if fuel bundle spacing, or plate spacing, are altered, or other forms of core disruption take place, including flooding, positive reactivity effects may occur.) 62 Given the relatively ineffective nature of the sole shutdown mechanism in the UCLA Argonaut-- production of voids in one of the reactor's two moderators-- and the additional problem of several potential positive reactivity effects, the reactor is far less inhorently safe than other reactors of the TRICA type or of the SFERT/30RAX variety. CRIGINAL INFERENT DESIGN FEATURES CF THE UCIA ARGCNAUT, AND HCW THEY HAVE BEEN ALTERED CVER THE YEARS "A reactor which is to te used for student instruction must be designed so that safety is insured without exercising greater restraint on the activities of students than is normally advisable in a university laboratory. This necessitates: (1) that the total available excess reactivity be limited to something less than that needed for prompt criticality, (2) that the reactor have a high degree of demonstrated inherent safety, and (3) that it be limited to low-power operation."
--original UCLA Reactor hazards Analysis, p. 19 Original Safety Premise _s
- 63. The original Hazards Analysis for the UCIA reactor, the one that formed the basis for granting the original license and the basis for twenty-two years of operation thereafter, examined in some detail the amount of excess reactivity that should be permitted at that reactor, consistent with student operation and urban siting and lack of contain-ment features. It should be recalled that virtually all of the traditional safety features (exclusion zone, containment, radioactivity removal systems for emergency, low population zone, emergency core cooling system, and the like) are lacking at UCLA. There is only one barrier to fission product release-- the fuel cladding, made of low melting aluminum. And there is only one method that might be able to limit the consequences of a reactivity insertion greater than. $1.00, for which control blades and dump valves would be ineffective, and that is a relatively weak and slow voiding effect about which there are numerous uncertainties as to how large a reactivity insertion can be compensated prior to fuel melting occurring. One thin, lew-melting barrier to fission product releases and one uncertain self-shutdown mechanism to prevent penetration of that barrier.
__ ; _a .a u- ,. _ . . _ _ _ _ . - 64 The Hazards Analysis wisely concluded that the fission product inventory should be kept low, by limiting operation to 10 ku, so as to reduce consequences if the fission product barrier were breached; (even so, it estimated thyroid doses as high as 1800 rem for a release of only 10% of the radiciodines): and it concluded also that excess reactivity should be limited to less than that necessary for prompt criticality, for which engineered safety features could still be effective. It demonstrated that this was, in its view, a sufficient margin of safety by estimating that the fuel would reach th'e melting point of aluminum with an excursion of roughly 2 3%Ak/k, based on rough extrapolations from the BCRAX data, corrected for a few of the differences tetween the reactors (and assuming linear corrections were possible). Given operation by students, who can be expected to make mistakes, and given the uncertainties in the calculations (which meant that melting might occur far below 2.3%), restricting the reactor to 10 kwth* and to less than that necessary for prompt criticality was determined necessary. 65 The original Hazards Analysis for this reactor begins a discussion of the general safety premises applied to its design with the following statement:
"The inherent saf,ety of the reactor is based on four points.
First, the amount of excess reactivity in the reactor is limited to about 0.6%. Second, the reactor has negative thermal and void coefficients. In addition, the reactor is provided with sufficient interlocks and safety trips to make a hazardous incident extremely improtable. Third, the amount of contained fission products will be relatively small since the reactor is to be limited to a maximum power of 10 kw. Fourth there is no credible way in which the fission products can be made to escape." (p.59) Intrinsic Safety Features of the Reactor Have Been Substantially Mitigated or Removed
- 66. In the years since the original design and analysis were completed, each of the above four bases for the supposed inherent safety of the reactor has been substantially mitigated. First, the licensed limit of excess reactivity is no longer restricted to 0.6%, less than that necessary to go prompt critical, but is now nearly four-fold larger, j at precisely the level the calculations in the Hazards Analysis indicate could cause melting. . Second, the reactor has teen found to have unexpected positive coefficients and feedback mechanisms, as discussed in the previous section. In addition, the reactor's staff has over the years found ways to disconnect interlocks and safety trips, and the value of the latter
- The power limitation was important, the Hazards Analysis indicated, because it limits the consequences of an accident, should one occur, by l limiting the radioactivity available for release to the environment:
"[T]he amount of contained fission products will be relatively small j since it is limited to a maximum steady state power of ten kilowatts." ibid.
i The increase in power to one hundred kilowatts thus brought with it a concomitant increase in fission product inventory and in possible consequences should an accident result in release of that inventory.
, _ _. . _. .._ _ _ _ _ - _ ..- - m_ -
has been brought into serious question by lack of accurate calibration. Third, the amount of contained fission products is no longer small relative to twenty years ago since the power limit has increased tenfold. And fourth, there are a number of credible ways in which fission products can be made to escape, including power excursions made possible by the increase in excess reactivity available and other factors. 67 In addition to the quadrupling of excess reactivity, far beyond the prompt critical limit prescribed by the Hazards Analysis and up to the precise level which its calculations indicate could cause fuel melting; and in addition to the tenfold increase in reactor power, a number cf other developments over the years at the UCLA reactor have considerably reduced the safety margins presumed initially. These include discovery of smaller-than-expected negative reactivity coefficients (in addition to the unexpected discovery of several positive reactivity effects); apparent lack of deflecters, as designed, to prevent repeated criticality enlargement of irradiation ports, making possible insertion of larger samples and the addition of a pneumatic tube " rabbit" system, which makes possible new mechanisms for rapid insertion and removal of reactivity. (Cther oversights in the original design review, such as errors about combusti-bility of the reactor constituent materials and Wigner energy storage, are discussed elsewhere; they too can impact upon effects of reactivity accidentsatthisreactor.)
- 68. The original design for the UCLA reactor called for substantial inherent safety features as well as large margins of safety: 10 kWth power limitation, large prompt negative temperature and void coefficients, and excess reactivity below that necessary for prompt critical. As for the restriction on excess reactivity to below 0.6%dk/k, the Hazards
, Analysis saids t
"it is possible to operate the reactor with an amount of excess reactivity which is well below that required for prompt criticality.
Under these condition, the reactor meets the safety requirements of a training reactor and can tolerate considerable operational error without damage." (p. 19) If the reactor ever met those safety requirements, it .o longer does. THE DANGER CF EXTRAFulATING, WITEUT VERY LARGE ERRLR 3ARS LR SAFETY FARGINS, FRCM SFERT AND BCRAX TESTS TO THE UCIA ARGCNAUT
- 69. UCIA argues that none of the alterations or problems that may have occurred during the reactor's operating history to date are of consequerce because the reactor is protected by inherent design against significant fission product release. In particular. UCIA argues in its license renewal request that its reactor can safely tolerate a far larger excess reactivity insertion than the reactor's original design limit. UCLA appears to rely heavily on an assertion that the BCRAX I and SPERT I tests conducted at the NRTS in Idaho in the 1950s and 1960s " proved" that the requested level l
l l l l
~
of excess reactivity is safe in the UCLA Argonaut. As UCIA put it in its 1980 Application:
"SFERT and 3CRAX tests showed that plate typa fuel elements survived step reactivity insertions of $3 54." (p. V/3-6)
- 70. As indicated in the introduction, that simply is not the case.
De SFERT I reactor core was completely destroyed by a $3 50 insertion. In fact, non-explosive melting of fuel was observed with even smaller reactivity insertions.
- 71. It is also worth recalling that the grisly SL-1 reactor accident, which occurred at the Idaho Testing Station not far from SFERT, was initiated by about the same reactivity insertion (in the SL-1 case, 2.4% + 0.3%dk/k *). This resulted in an energy release several times greater than that which destroyed SPERT I, sufficient, in fact, to not merely melt the fuel but vaporize parts of it. The resulting steam explosion was so intense that the whole nine ton reactor vessel was lifted nine feet in the air.
- 72. Even were it true that plate type fuel elements survived step insertions of $3 54 at SFERT-- which they most certainly did not (as is plainly demonstrated in the photos of melted plates from the $3 50 excursion found on pages 7 and 8)-- that would by itself say nothing about whether plate fype fuel would survive the same insertion in the UCIA Argonaut, a different reactor design with significantly different operating characteristics. There is no magical relationship, as the UCIA statement cited in paragraph 69 above implies, between reactivity insertion and fuel plate response, independent of the reactor in which ~
the excursion is occurring. A reactivity insertion of $3 50 will melt one core, while leaving abother virtually untouched, depending upon a whole litany of varying characteristics-- plate thickness, coolant channel width, void coefficient, moderator temperature coefficients, the presence of a non-expellable moderator such as graphite, the metal-water ratio in the core, plate surface area, degree of burnup and corrosion, prompt neutron lifetime, fuel enrichment and uranium weight %, starting moderator temperature, and many other factors. 73 Even had SFERT not been destroyed by a $3 50 insertion, the UCIA statement quoted above could not be true, because it implies that STERT and 3CRAX tests proved that plate fuel could not be damaged by reactivity in-sertions of $3 54, no matter in what reactor and under what conditions it was placed. And if anything was learned through the SFERT tests, it was that seemingly minor variations, even within the same reactor (e.g. , degree of subcooling), could significantly affect the total energy release and thus, whether fuel melting occurred. Differences between different reactor types were even more pronounced, affecting the very nature of the shutdown mechanism that terminates, and thus limits, the excursion itself. Se SFERT and ECRAX tests could not, by any stretch of the imagination, "show" that a certain general kind l $3 54 would be the equivalent of between approximately 2 3% and 2.7%6k/k, depending upon the value used for the delayed neutron fraction. (2e I SL-1 was a low-power experimental and training reactor utilizing highly l l enriched aluminum-uranium flat plate fuel, cooled and moderated by water, similar to ECRAX and SPERT. ) i I
~'
,__i_.._.. . _ . - _ . . _ . , . _ _ _ . , _ _ _ , , _
of reactor fuel (e.gt, flat plate) could survive a $3 50 insertion in any imaginable reactor. 74 The important question, then, is not what reactivity insertion destroyed SPERT or BCRAX or SL-1, or even what insertion could be expected to be the minimum necessary to induce melting in those reactors, but rather, what level is a safe level for the UCIA Argonaut, with sufficient margins of safety consonant with student operation in a densely populated location. After all. SFERT, BCRAX and SL-1 were all destroyed in the Idaho desert far from any populated center. Ard the UCLA Argonaut-type reactor is a substantially different reactor than the three Idaho reactors mentioned above.
- 75. The differences are significant, Plate and meat thicknesses are different, as are coolant channel widths. The SFERT tests used essentially fission-product free cores, with fresh cladding. UCLA's fuel has been irradiated for two decades, can be irradiated for another two decades if relicensed, and has been sitting intermittently in water, allowing for some degree of corrosion, for many years. Each of those factors might affect the heat transfer time to the water *, potentially elongating the transient and increasing the energy release, factors not analyzed in the existing reports.
- 76. Furthermore, SFERT and BCRAX were entirely water-moderated and
-reflected, as was SL-1. UCLA's reactor is moderated by both water and graphite, and reflected by graphite. This lengthens the neutron lifetime, producing a longer period for any given reactivity insertion, but it also significantly reduces the value of the shutdown feature caused by expulsion of the water portion of the moderator. In the UCLA case, part of the moderator and reflector, i.e. the graphite, cannot be expelled from the core during the normal course of an excursion, thus reducing the effectiveness of moderator voids in limiting the peak power reached.
And further, the reported void coefficient is smaller for UCLA than SFERT or 3CRAX, as is the temperature coefficient for the water portion of the moderator. The positive coefficient for the graphite further weakens the size of the shutdown mechanism for UCLA, and the positive reactivity effects noted when water level initially drops in the core and when fuel plate spacing (and/or bundle spacing) is altered, as by oscillation, are other important differences.
- 77. These differences can be very significant in determining the energy release from any particular excursion and whether fuel melting will result.
Even different reactors of the same general type produced widely different l ener6y releases for the same period, as is shown in the plot of energy l versus reactor period on the next page, taken from Thompson and Beckerley's
- A potentially significant factor not considered in the analyses to date is the reduction of thermal conductivity in the fuel due to irradiation.
A relatively small degree of burnup can result in reduction of thermal l conductivity to half its initial value. (see J.A.L. Robertson, Irradiation ! Effects in Nuclear Fuels, p. 261: Tipton (ed. ), Reactor _ Handbook, vol.1, l Faterials, p. 192; Report TID-7515, part 2, p.13) l l i l l 1
._.m . - . _ _ . _ _ . _ _ . _ _ _ ._ . . _ _ _ . . _ _ _ _ .. _. . _ . _ . . _ . . _ _ _
n ,.. v ,%. i J. TifCMPSON .; ACCIDENTS AND DESTitUCTIVE TESTS 13 673 1000 -- 6- i i 6 g i Ii og e i g i ii g i i _
! 'J
_ i r.t i E.! 500 , I a :. . -
!.V S.V 2.3 t.4 SL-l EXCURSION ,
- 9CW 62 *C SUSCCCLED -
t 1:3 13 t.e .. -
, TEST Q 30 10 0 CELEMENT -
tem :o* -CCRE u - sc ,,/ - 4.s 1.0 3 - spoir - 7.5 :.9 - ( 00 _ orstnuc. rrve (%% - 3 9) 64 30 v a.: n.4 . 3 ge - 130 2 15 (7
~
Jug ~
, Op go
/ Co -
SPERT 10 -Cu 12/25 CCRg A O e - a bring Investi- - -ENERGY TO U 2
~P EAX , 80* C A CN 4 ~
m 5 demon . 5 St.'8CC CLED -00 pk s, 'a ~ 1a .ston was i shess that the a a c f(e.,b - -
- o havs been a
._ =y a .,
(o
>n bo11hg. and -
C - en sonna insight g Last, but not - -
=agement aloeg C g is possible now -
4-m eats in at , , , l , , , ,l , , ,l,,,,l , , I 5 10 50 10 0 400 i 5I.-1 accident 2.ual withdrswal INITML REACTOR PERIOD ( msec) e reactor. This ' transtanz whose r in Table 3-9 F10. 3-30 Pruns:e6 ami atueaired amiser eenTy reteese vs. portas for scR st t.sPERT.I. and SL-t, Total energy wdess omerwa a s-N c rcias are SoRAX data trem entsensca (19}, squares sPEAN tXJ L2/2s data frwm (63al ara crsassaae sPERT-4 desaracave test saca true (63]. L3 wtthdrawal of reactor critical a single slug. The water levtl in the tanic was pressure region. The pressure ware front wh.ich en calculate that about 2.5 ft (76.2cmlbelo vthetcpof the vessel and developed no doubt spread out in all directions, a continuing the the slug, therefore, hnd this distance to acquire striking the vessel side walls next to the core n (the porition in kinetic energy. This du, hit the bottom cf the first and bulgtag them, then striking the bottom idwirc ullapsed of the plate area tn theceni.al16 elementa reacned head and giving a net downward force on the dieve that power the vaporization temperatM
- and this caused vessel." and fhally acceleratingupward.s the entire
- ims.tely 4 msee more steem production and violent destr=ction mssa of water above the core. It appears likely c.tionterminated of this region. About 20% of the entire core shows that the water moved upwards more or less as 9 d.4 w 10 4 Mw. melting proceeding to the clad surfaces. General
=p;rature in the , Electric estimates that the total nuclear transient had just reached energy was 133 +10 Mw-seo and that no more than *Apparently no one has looked hto this dotm-an additional 33 Mw-sea of energy (best estimate ward force andonecanonlycenjectare asto whether C80'C (3767'T). in. or 0.839 mm) 24+10 Mw-sec) was releasedischemicaireactions this downward force was sufficient to sever the ter ' aces had between the molten or vaporized metal and water. pipe connections to the tank. 2 is difficult to Trt the start The formation of the steam void terminated judge the resistance to such a shock provided h) Cxcursion. 3% the nuclear transient, but it also created a high by the vessel supports.
&$h a Cl Y w4 Lb4 $ f hf tf
. - .. .. . I .
. _ ._ __ _ . _ _ _ _ .,____ . _ 2 _ _. n . . _ _ _ _ _
Technology of Nuclear Reactor Safety, p. 675 As is shown there, BCRAX produced substantially more energy than SPERT, and SL-1 more than either, given the same initial reactor period. ('Ihis is an important reason why estimating the energy release for an excursion of a particular period at UCIA directly from the release for SPERT at the same period is so non-conservative-- the same period produced far higher excursions in other reactors. ) Seemingly minute differences in metal-water ratios, temperatures and void coefficients, etc., had marked effects on total energy released.
- 78. This is understandable when one realizes that the process of a power excursion is essentially exponential. The nature of the exponen-tial rise is that very minor decreases in exponential period (the "e-folding time") or increases in total time of the excursion (by delay in the shutdown
, mechanism) can cause -the. power to increase by large amounts. Thus a delay of a few milliseconds in the transfer of heat from the fuel meat to the
- i. clad a'nd then to the coolant- (caused, for example, by thicker fuel plate or lowered thermal conductivity because of corrosion or irradiation) can mean the difference between an excursion terminated safely and one resulting in melted fuel and substantial fission product release. Thus, minor errors in calculation or extrapolation can have potentially disastrous results.
I 79. In the absence of actual SPERT-type excursion tests with an Argonaut-type reactor, it is understandable perhaps that hazards analysts would attempt to extrapolate from the excursion tests that have been performed, albeit on reactors of different type. Thus UCIA's own 1960 Hazards Analysis, the Hawley et al review, and the Neogy memorandum all rely i_ on the power excursion tests performed at the NRTS in Idaho. UCIA relies largely on the BCRAX tests in its original analysis: Eawley et al on I the SPERT ID series of tests; and Neogy on the SPERT IA series. (Surprisingly, none even touches on the SL-l' accident.) .All are h sed l on the fundamental assumption that one can extrapolate with extremely-l high precision from the SPERT or BCRAX tests to the UCLA Argonaut.
- 80. We take substantial issue with such an asumption. First of all, the SPERT tests were not intended to be used in such a fashion. SFERT was an attempt to understand the mechanisms of shutdown in power excursions, not to produce an absolute number that could be plugged into reactor analyses for significantly different kinds of reactors. In particular, it was never intended that a hazards analyst would simply look at the exponential period at which some melting was expected to begin at SPERT and say that therefore substantially different reactors could safely handle precisely the same period. The SPERT tests simply do not permit such extrapolation to different reactors without an extremely detailed accounting for differences between the reactors, which is very difficult to do, and very significant error bars to take into account the significant uncertainties in such extrapolation.
- 81. If the SPERT core was destroyed with a $3 50 insertion, it would have been of considerable concern if a reactor operator used that fact as tasis for a $3.40, or $3.00 limitation for another reactor, particularly of a different type and in an urban environment. The SPERT tests were never intended to be so used-- the uncertainties are just too large.
To say, as the Hawley et al review essentially does, that the SPERT ID core indicating melting beginning around a 7 msec period meant that the UCLA Argonaut could tolerate a 7.2 maec period excursion without any melting
- l. ,_ ___ _.____ . . _ _ _ ___ _. _
--~ -.
23 or release of fission products goes far beyond the purpose of the SFERT tests and the statistical significance of the data. i
- 82. The primary value .of the SPERT tests was a significant advance in the qualitative understanding of reactor behavior during power excursions and, in particular, the various components of shutdown mechanisms in differing cores-- radiolytic gas production, water-moderator expusion, fuel plate expansion Doppler effect, density changes, " warm neutron" effects, and the final shutdown mechanism, rapid disassembly of.the reactor core.
It provided qualitative understandings about the nature of the phenomena, not direct quantitative data universally applicable, particularly to other reactor types.
- 83. With the above prefatory comments about the difficulties inherent in such extrapolations, an analysis foolows of the three attempts that have been made to extrapolate the 3CRAX and SPERT data to the UCLA case.
THE 1960 UCLA HAZARDS ANALYSIS AND 1980 SAFETY ANALYSIS REFCET Eelting Estimated to Occur Around 2.3%dk/k (The Current Licensed Limit) 84 The UCLA Argonaut-type reactor was designed for a maximum power of 10 kuth and maximum excess reactivity of about 0.6%6k/k. As indicated above, these limitations were considered prudent in light of student operators, lack of containment and dense population immediately next to - the facility. The Analysis supporting the proposed license argued in particular that the 0.6% reactivity limitation was prudent because it was below that necessary for prompt criticality, above which level engineered safety features such as scram systems tend to be too slow to compensate for the rapid power growth. To demonstrate that not only was 0.6% safe, but that a sufficient safety margin existed for a training reactor, the Hazards Analysis attempted to estimate, quite roughly, the level at uhich melting could be expected. This was done, the Analysis indicates, to show the magnitude of the safety margin and to provide further support for the 0.6% limitation. 85 To make this showing, the Hazards Analysis relied on BCRAX data. Cbtaining a proportionality from those tests for temperature rise per IN-sec of energy-release, the analyst determined that it would take approximately 41 Ew-seconds of energy release to raise the temperature of the fuel plate from the temperature of' boiling water to the melting point of aluninum (not of the fuel meat, which melts at a 360F lower temperature). Using a chart obtained from the BCRAX test it was estimated that an excursion of reciprocal period 150 sec-g,would give an energy release of 41 Ew-seconds plus the energy necessary to raise the plate temperature to the boiling point of waters i.e., a reciprocal period of 150 sec-1 would produce enough energy to raise the plate temperature to the melting point of aluminum, at the center of the hottest plate.
- 86. As stated in the Hazards Analysis:
"It is useful to estimate the value of excess reactivity which, if suddenly inserted and not removed by the control system, would raise the maximum temperature in the hottest fuel plate to the
~24- '
~
melting point...
"The first step in the procedure is the estimation of the exponential period corresponding to the excess reactivity which would have characterized a power excursion of similar effect in 3CP.AX I."
(p.III/A-3,emphasisadded) s
- 87. The Analysis then attempted to correct for the different void-coefficients, coolant channel width, figure of merit for. fuel performance, and peak to average power ratio, concluding that the limiting excursion for UCLA is 9.1 milliseconds. Correcting for the different prompt neutron lifetimes, it was stated that that period corresponds to an insertion.
of23%Ak/k. (It is interesting to note that the Hazards Analysis ' estimated that the UCLA reactor could tolerate a considerably smaller power excursion in terms of ener6y release than could 3CRAX, because of the different characteristics of the reactor-- 41 Ew-sec, plus the energy to brirg the water to saturation, as the limit for BCRAX, and 28 EM-sec for UCiA. Conversely, 3CRAX was stated to reach its limit with a 6.7 msec period, UCLA with a 9 1. This shows the problems with assuming that if SPERT, for example, could tolerate a 7 msec period, so too would UCLA.
- 88. As the original Hazards Analysis calculations make elecr, 2 3%A k/k would be sufficient to cause fuel melting at UCLA, if the assumptions employed are correct.
- We have made clear above our objections to such There is some confusing language in the text of the Analysis on this point. me' calculations make' perfectly clear that, if the Analysis is correct, a 2.3% reactivity insertion will bring the hottest part of the fuel meat to the melting point of aluminum. Yet it is ctated at one point that the reactor will tolerate a power excursion of at least that magnitude '
without melting occurring at the hottest part of the fuel. This is primarily a semantic difference, asserting that a certain estimated point is the end of the safety zone instead of~saying it is the beginning of the danger zone. Some of the confusion can be traced to the fact that UCIA copied its 1980 Safety Anal raised to 2.3%) fromysis its Report 1960(by which Analysis Hazards time the (at reactivity limit the which time hadlimit beenwas 0.6%), which in turn was copied from a 1959 AMF analysis, which in turn was copied from a 1958 analysis for the University of Florida reactor. (seeattachments). A comparison of the analyses indicates that while the language was copied virtually verbatim, there was a significant difference between the University of Florida reacter, upon which the original analysis was based, and the UCLA reactor. De fuel at the former was 20% enriched, 90% enriched for UCLA: the U of F fuel'was 46 w/o U-A1, whereas UCIA's is right at the 'eutectic point,13 4 w/o. (See page 1 of ' U of F aul UCLA's " Estimation of Effects of Assumed large Reactivity Additions.") The uranium-aluminum alley in the U of F fuel meat melts considerably above the melting point of aluminum, unlike the alloy in UCLA's, which melts below the critical temperature of aluminum. Furthermore, the U of F fuel had a Doppler contribution to shutdown, since it was LEU, whereas UCLA practically does not. The 1960 UCLA
extrapolations from one reactor type to another in the absence of empirical evidence from tests like we conducted at SFERT or very significant error bars at each point in the calculation. As we read the Hazards Analysis, this was recognized by its author, who recognized the approximations he was making required substantial margins for error. These margins were provided by the fact that the analyst was not trying to show that 2.3%, or 2.2% or some similar number was safe, but rather that 0.6% was prudent and had a sufficient margin of safety for a training reactor. He did this by estimating, through some rather crude extrapolations, that danger might be found in the 2,3% range, and therefore limited the facility to 0.6% so there would be a margin of safety for errors in calculation or operational errors that might slightly exceed the license limits. As we read that analysis, it shows melting at around 2 3%dk/k, and supports a 0.6% limitation. It cannot be used to justify a limit at or close to 2 35 In fact, as will be discussed in the next section, correcting for some non-conservative values in the Hazards Analysis calculations indicates melting substantially below that level. The Hazards Analysis. When Corrected. Indicates Risks 3elow 2.3%6 k/k 89 The Hazards Analysis makes clear that the fuel meat could reach the melting point of aluminum with a 2 3% insertion. Corrected for more conservative void coefficients and delayed neutron fraction, plus consideration-of eutectic melting, indicates danger with considerably smaller insertions. Furthermore, proper consideration of error tars at each step in the calculation, as well as consideration of UCLA's positive feedback features, would reduce considerably further the estimated reactivity insertion that could be tolerated without melting, i.e. that could te successfully terminated by steam formation.
- 90. The Hazards Analysis uses a void coefficient for UCLA of -0.16A/
% coolant void, uhereas the current application cites a value of -0.164%
(p. III/6-6). If UCLA's reactor has a smaller void coefficient than , initially thought, its capacity to tolerate certain excess reactivity l insertions is substantially reduced, and fuel melting could thus occur at substantially less than a 2.3%dk/k reactivity insertion. Uncertainties in the precise void coefficient (7hich can vary by region of the core and other variables) add substantial reason for added margins of safety.
- 91. Using the Hazards Analysis calculations and merely substituting l the more conservative-- although perhaps not sufficiently conservative--
i void coefficient for the value used initially, before measurements had teen made, results in the excursion that could cause melting being 3 Eus smaller than that assumed in the Hazards Analysis: Analysis merely removed the sections of the U of F analysis dealing with the Doppler effect and other fuel characteristics, failing to correct for i the differences, and keeping in language contradicted by the calculations. l And thereafter UCLA used the Analysis, which had concluded 0.6% was safe and 2.3% dangerous, to support modification of reactor limits to 2,3%Ak/k. ! This is but one example of the problems that can occur when copying analyses performed by others or for other reactors. l i
using 1960 estimate ) usir4 1980 estimate for void coefficient for void coefficient C Borax 0.24 Borax 0.24
- C UCIA " 0.18 " '33 CIA
" 0.164 "
41 W sec sec 1 33
=
31 W see 1.46
= 28 W sec 31 W sec x .82 x 1.12 = 28.4 W sec 28 W sec x .82 x 1.12 = 25.7 W see
- 92. In addition, as has been pointed out above, the Hazards Aralysis calculations appear to neglect.eutectic melting. The calculations were insed on the melting point of aluminum, whereas the UCLA fuel meat is in the eutectic range and melts at 20 C lower temperature than aluminum. (Hawley, p.18).
Thus, a smaller excursion than estimated in the Hazards Aralysis would bring the fuel to melting. Using the figure of 24.40 F/W-sec supplied in the Hazards Analysis, about 1 W-sec less er.ergy would be required than previously estirated, producing a commensurste reduction in the amount of excess reactivity necessary to produce fuel melting.
- 93. The Hazards Analysis used a non-conservative delayed neutron fraction (p ) of 0.0074, whereas the Application now cites a figure of 0.0065. #
is important in the conversion from period to excess reactivity through the "inhour equation." Use of the form of the inheur equation cited in the Hawley review (p.16) shows that use of the smaller a results in - a shorter exponential period for the same reactivity insertion, and thus more energy release and hi 6 her fuel temperature.
- Conversely, use of the smaller A means a smaller reactivity insertion will produce the same result (i.e. fuel melting) than estimated in the Hazards Analysis employing the larger figure.
94 If the Hazards Analysis concludes that a 2 35dk/k insertion will bring the hottest parts of the fuel to the melting point of aluninum-- and it clearly does-- then use of the smaller figures for vcid coefficient and 4 , as well as consideration of the eutectic melting point of the meat (below that of aluminum), would indicate fuel melting occurring with a substantially smaller reactivity insertion. 95 There are a number of other factors which should further substantially reduce the Hazards Analysis estimate of the excess reactivity necessary to induce melting-- the effect of fuel irradiation or cladding corrosion (which can reduce thermal conductivity and thus delay shutdown), as well The version of the inheur equation cited by Hawley is T . 1 S k/k (1- /3,ff) - A,ff J
~ . . . _ ._. _ ._ ._. - ._. -- .-- - ~L ' -
usit a few. Although the asinitialmoderatortemperature,tonamej/kinsertiontooccurina Analysis conservatively assumed the 2 3% 6k sutcooled reactor, the Hawley review at p.15 rightly points out that excess reactivity is normally measured at normal operating temperatures of the reactor and that negative temperature coefficients for the water would make, for example, 2 3% at operating temperature ac+aa13y much more at lower-than-normal temperature. Conversely, if 2 37, .r dangerous on a cold day, far less than that amount must be install. Jf measurement is urder warm moderator conditions. . 96. : And, as discussed in more detail later the positive coefficient for the graphite can likewise mean that 2 3%4k/,k measured when the graphite is' cool can result in more than 2 3%Ak/k being available after its temperature has risen. That factor, plus positive feedback effects in an excursion
- (such as the positive coefficient for the graphite, the positive. void coefficient in a portion of the water moderator, and the positive effects from changes in plate and bundle spacing that might accompany the initial stages of the excursion) further dramatically reduce the " safe" level.
Proper inclusion of adequate error tars for the various steps in the calculation, pushes the level even further down.
- 97. Thus, given the basic assumptions employed in the Hazards Analysis, and the numerical values utilized, the Analysis' calculations predict fuel melting with insertions in the range of 2 3%. When a few of the numerical values are changed to reflect more appropriate values (e.g.,
/9 , void coefficient, and eutectic melting point), substantially less than 2 35 dk/k would appear to be sufficient to induce melting-- if the method-ological assumptions. employed are correct. If other factors are included, even smaller levels are tolerable.
- 98. There are problems, as indicated at the outset, with extrapolating from one reactor to a different one-- to three significant figures--
without error bars. This assumes that there exists a complete knowledge of all the differences between the reactors and how those differences precisely affect behavior. As has been shown, a number of differences were not considered, and to assume that what differences are considered can be corrected for using simple linear relationships is inappropriate for the level of !- precision assumed. For example, the Hazards Analysis assumes a linear relationship between void coefficients and total energy release, which is unlikely to be correct, given the exponential nature of energy release in a power excursion.
- 99. The Hazards Analysis merely decines that the 0.6% k/k limit has a reasonable safety margin to compensate for the potential errors in extrapolating from the BCRAX data. It is filled with terms describing the calculations clearly as estimates and extrapolations, based on unverified assumptions:
_On the assumption that this minimum value is the true value, a rise of water temperature from near 0 C to 800C would reduce reactivity by 0.6% keff. III/A-2emphasisadded
The characteristics of the UCTR which determine its behavior during power' transients resulting from large reactivity additions are quite similar to. but not identical with, those of the 3CRAX I reactor. III/A-1emphasisadded Experiments of the ECRAX and SPERT types have not been nade with reactors havird widely different neutron lifetimes. The general evidence of the experiments, however, supports the supposition that... III/A-3 emphasis added In comparing the behavior of different fuel plates, it must be recognized that the total energy releasc of the power excursion can no longer be considered as a definitive variable... III/A-4 emphasis added that the maximum fuel-plate temperature rise is, to within experimental error, proportional to the raximum energy release of the power excursion. The proportionality was determined to be constant 24.40F por W-sec. III/A-3emphasisadded
, The relative importance of the two moderators, graphite and water, in determining the effective neutron temperature introduce uncertainties in the theoretical computation of this computation.
III/A-2, emphasis added 100 The text is replete with phrases about estimation, assumption, uncertainties, suppositions, and so on. Ihe Hazards Analysis was designed to merely estirate hou large a safet margin the reactor would have at the then-licensed limit of 0.65 6 . 101. Substantial error tars, or margins of se.fety, are required in such analysis, which is why the Hazards Analysis concluded that excess reactivity at this facility should be limited to about 0.65 6k/k. The Hazards Analysis demonstrates that currently requested levels of excess reactivity provide no margin of safety and could lead to fuel melting in the UCIA reactor. THE HAVI.EY ot al REVIEW A Very Small Margin of Safety 102. The Fawley study attempts to address the same issue as the Hazards Analysis, except that feuer corrections are made for the difference in char-acteristics between the UCIA reactor and the SPERT reactor, the original data source. Eawley concludes that temperatures about 540 below the melting point of the fuel could te attained; given the NRC Staff assumption of a 75 C starting temperature for the fuel instead of 60 00, for the same energy release (SER p. h-3 and 14-4), there would be only a 39 00mar 61n - of safety. Given the extremely crude approximations used, the numerous factors not considered (e.g. lower void coefficient) that would markedly j increase the estirated energy release and temperature, and the lack of l error inrs, just a few of these corrections could push the temperature l above the melting temperature of the fuel.
- -. -= - -. - . - - .-. .- --- --- .
- 103. The'section of the Hawley, el g, report dealing.with excess reactivity issues appears to consist almost exclusively of a brief literature review and some extrapolations from the SPERT I tests. Whereas the 1960 Hazards Analysis took into account a number of differences between the UCIA Argonaut and the ECRAX reactor, from which it was extrapolating its data, the Hawley review does not account for several of the UCLA-SPERT differences, partic-ularly UCLA's smaller void coefficient, which would tend, if not otherwise com- !
pensated, to suggest that an excursion of the same period in SFERT and the . Argonaut would produce greater energy release at UCLA. The Hawley report's primary consideration of differences between the two reactors consists
- of correcting for the longer neutron lifetime at UCIA, a factor which is helpful to UCIA, .
104. The Hawley approach was extremely simple-- calculate the period produced by an insertion of available excess reactivity, estimate the energy release an excursion of similar period would have produced at SPIRT ID, and then scale temperature linearly to the peak temperature estimate during the SPERT ID destruct test. i 105 And yet, even without taking into account factors such as void coefficient differences, which would tend to produce higher temperatures, l- the analysis estimates peak fuel temperatures only about 500 below the molting temperature. No error bars whatsoever are provided for the extrapolation ; steps nor for the final conclusion. (There appears to be a subtraction
~
1 ! error in that Hawley et al assert on page 19 of their report that a hot
- spot of 586 C would be 7 PC below the melting point of the fuel meat, which they cite on the previous page as being 6400C.)
106. 50 0is not an adequate margin of safety, particularly when so many 1 of the differences between SPERT and the UCIA Argonaut were not taken into l account. Furthermore, significant effects may appear just below the melting point, such as volumetric expansion of the fuel resulting in cladding failure, or considerably increased diffusion of fission products through the hot metal. It was noted at SPERT, fer example, that some of the s fuel plates were very substantially softened and warped, even though not ! truly melted, and that they would stay in that softened form for several days thereafter, behaving something like a-wet noodle. This was prior to j the final destruct test. 107. So even if Hawley el d were correct in their estimate of peak temperatures 50 or so degrees below the melting temperature, there would still be concerns. However, questionable assumptions used by Hawley ,e_t,t i al suggest far greater temperatures could be achieved in the UCIA Argonaut than those estimated. i Questionable Assumptions 108. Perhaps the most questionable assumption is that a 7 2 msec period would produce a 12 W-sec energy release in the UCLA Argonaut. Given the linearscalingassumgtionoftemperaturetoenergyreleaseemployedby Hawley (p. 19: - 1500 C Per 30.7 W-sec, or about 49eC/W-sec), a 13 W-sec energy release would cause melting, if Hawley's assumptions are accepted. That is not much of a margin of safety if his 12 W-sec estimate is correct. I
109. The non-conservative nature of the Hawley analysis can be demonstrated by comparing his results with those of the other analyses he cites at pages 4-7. Hawley assumes an excursion of period 7.2 msec at UCIA will only release 12 FM-sec of energye getting within a few degrees of melting, from a 2.6% insertion (supposedly the equivalent of 2 3% on a cold day). However, it is noted that the 1960 Hazards Analysis estimates a considerably longer period than the one assumed by Hawley (91 instead of 7.2 msec) will produce an energy release of 28.4 FN-sec, plus the energy necessary to raise the fuel to the boiling point of water. How a longer period is estimated toproduce2}timestheenergyassumedintheHawleyreportisnotexplained. The 1961 ATL analysis is reported as indicating a far smaller insertion than that assumed by Hawley,15%Ak/k, will produce an energy release of 24 LM-sec, double that assumed by Hawley. The GNEC report assumes the same period, around 7 msec, producing 32 FN-sec, plus about 4 }M-see to raise the tenperature to the saturation point of water (i.e. , about 36 PM-sec total). The Jason reactors are referred to as estimating 10 FM-sec, nearly that estirated by Hawley for UCIA for a 2.6% insertion, occurring from only 0 5% insertion at Jason. Hawley notes that the variations "are not resolved" in the available documentation. (p. 7). With such uide variation, and lack of documentation to explain the variation, it is most non-conservative of Hawley to utilize a 12 PM-sec estirate of energy release whereas other estirates several times higher exist, noting that less than 1 MW-sec additional energy release would eliminate Hawley's 390 margin of safety (even ignoring the lack of error tars, which would obliterate margins of safety far larger). 110. Thus, were Hawley right that a 13 FM-sec excursion could cause fuel melting, and were the estirates of energy release from any of these other studies correct, fuel temperatures would be considerably over the melting temperature, not, as Hawley asserts, just under. The ATL estirate of 24 PM-sec for a 15% insertion would te over the melting threshhold. The 1960 Hazards Analysis estimate of 28.4+ PM-see would be over. So would the CNEC estimate. So would the Jason estirate. 111 Empirical data from actual excursions also underscore the non-conservative nature of the Hawley assumptions. For example, a 7 msec period in the SFERT IA core is reported to have released 23 MW-sec of energy, nearly twice that assumed by Hawley tased on STERT ID data (Schroeder. 1957). The plot of period versus energy release (Thompson and 3eckerley, 1964), mentioned earlier, likewise shows how the choice of 12 FM-see for a 7.2 msec period is quite non-conservative. SL-1 extra-polations, for example, would suggest an energy release five times greater than that assumed by Fawley. When one recalls that an energy release of 13 FM-see would cause melting, if Hawley's other assumptions are correct, then data suggesting releases of 23, 28+, and even 60 FM-sec of ' energy from a 7.2 msec period excursion, not the 12 PM-sec assumed by Hawley, indicate an unacceptable protability of a destructive power ex-cursion, one that could release significant amounts of fission products. 112 Hawley'schieferrorisinequation(2)onpage17,whereheassumes that the total enerdy release from an excursion in the UCLA reactor can be precisely determined by doubling the ratio of the reciprocal period to a reactivity coefficient found from excursions during the SPERT ID series of tests. He assumes that he can apply, without any modification, that reactivity coefficient (which was substantially different than the reactivity coefficient found in the SPERT IA tests, the 30RAX tests, or from the SL-1 accident) directly to the UCIA case.
. . - - . -- ~ - . . _- _. . . .
_30 113. Hawley commits this error by focusing on' those factors affecting
. the first part of a power excursion, the power rise, but ignoring the second part of the excursion, its termination by self-shutdown features.
i He assumes that total energy release for an excursion at the UCIA reactor is entirely controlled by the exponential period of the rise. However,
. total energy release (which determines the severity of the incident) is tightly controlled, not just by how fast the power rises, but by how quickly the power rise is terminated. A slow power rise in one reactor may cause far more damage than a fast rise in another, if the shutdown mechanisms in the former are slow as well.
114.~ As-discussed earlier, power rise in a power excursion is exponential, essentially increasin6 by a factor of 2.718 every few milliseconds. The amount. of energy released is thus a function of essentially two features: the exponential period (the e-folding time) ard the . length of the excursion before shutdown (i.e., the number of e-folding periods). From the equation given below (the same as in 35 above), we see immediately that very small changes in either t (the time that elapses before shutdown mechanisms take hold and. terminate the power rise)'or T (the exponential period, or e-folding times the time it takes the power to increase by 2 718) can have very large effects on the power reached: e o Hawley essentially ignores the fact that any linear delay in the shutdown mechanism can cause a nonline_ar (i.e., exponential) increase in the total j power.
- 115 Because of the reported longer neutrort lifetime at UCIA, the same reactivity insertion will produce a longer exponential period T than it would at STERT. Hawley takes into account this difference between UCIA and SPERT (which helps UCIA), but ignores the differences between
. the reactors which will mean a longer total excursion becausa of slower .
' or smaller shutdown mechanisms.. Thus, T may be longer, which Hawley f
considers, but t may also be longer, which he does not. Since the power rise is exponential.. ignoring even a few millisecond delay in shutdown mechanisms can be devastating. 116. Assume an exponential period T of .7 msec and a time interval of rising power before the shutdown mechanism acts of .07 sec (i.e., the time it takes the heat to transfer from fuel to clad to coolant and cause voiding of the goderator). 'Ihe power would thus rise by e raised to the 07/.007, . or e , a very large number (about a 22,000-fold rise in power). If initial power was 100 kw, seven hundredths of a second later the power would be over 2000 IN. If. the shutdown mechanism at UCIA is even a few percent slower or less effective than that of SPERT (e.g., because of the 50% smaller void coefficient, the thicker plate dimensions, a little bit of added corrosion on the clad, or the positive effect of the initial coolant drop or the graphite temperature coefficient), the difference in peak power can be very substantial. o m.--ew-e-r = ...-ew.-,, ww-3-ysew-,,e,.p --- , ,u*--w p--- --<-.4y,.r----a----w- m -W e w T as-+-wv.--reg .~ei---v.r--ri+W--'T-'---r1 *e'? W - '-*w *T-'T *
, _ . _ _ _ _ _ _ _ _ . _ . _ . _ _ _ _ _ _ _ _ =_ -_ ._. _ - _.
~31-117. Taking.the example given above, and assuming a very modest difference of 10% in speed of shutdown, representing a few milliseconds, one additional
- e-folding period would occur at UCLA before shutdown than at SFERT, from which Hawley obtained his 12 W-sec estimate. This could mean, thus, peak power 2 7' times higher, just because of a delay of a few thousandths of a second in transferring heat to the coolant, voiding the coolant, or the reactivity worth of voiding the coolant. In other words, a few percent less prompt or less effective shutdown mechanism does not mean a few percent higher peak power, but because of the exponential nature of the rise, would mean several times higher peak power.
118. All indications are that the shutdown mechanisms for UCLA could be substantially slower and smaller in,effect than those of the SPERT or 3CRAX reactors with which they are being compared. The 1960 Hazards Analysis made clear that just correcting for a few of the differences between
, UCLA and 3CRAX, the minimum period Uf".A was expected to be able to tolerate was considerably longer than that estimated for 3CRAX. The void coefficient is smaller. which is quite important, and simple effects like the 50% reduction in thermal conductivity in the Al-U fuel meat caused by irradiation can substantially elongate the time interval for the heat generated in the excursion to be transferred to the moderator for eventual shutdown.
Given the exponential nature of the rise, and the exponential period measured in milliseconds', delays of a millisecond or two in transferring the heat,~ and differences of a few percent in the effectiveness of the voids once formed in the coolant, mean melting can occur substantially below the reactivity insertions assumed by Hawley or the original Hazards Analysis. 3ased on the analyses done to date, insertion of either $3 00 or $3 54 must be censidered a credible cause of fuel melting. 119 It should be noted once again, however, that the methodology of very simplified extrapolation from SFERT or 3CRAX data to the UCIA Argonaut case, as done in the Hawley report, seems most inappropriate given the differences in the reactors and the difficulties in correcting for those differences. The SL-1 accident, which took the lives of the only people nearby at the _ time, was "non-credible" in Hawley's terms, yet it happened. It released several times more energy than Hawley's extrapolations from SPERT ID would predict, even though it was much more similar to SFERT than is 1 " the UCLA Argonaut.
- The Hawley extrapolations cannot be ' relied upon to prevent an SL-1 type accident at UCIA, one that would occur not in a remote t
federal testing station but in the midst of tens of thousands of people.
* ' Er. Cstrander, in his September 1,1982, declaration, at page lo, asserts that the reason why 3CRAX data suggest a so much larger power excursion for the came period- than.does SFERT (and why he believes it appropriate to
- ignore the more conservative ECRAX data) is because of different active core height producing hydrostatic pressure and inertia forces which impede
' toiling more in the 30RAX case. This is an interesting hypothesist un-fortunately, its validity has not been demonstrated. However, assuming for the moment that it is correct, such an effect , may well be very unfavorable for UCIA, because among the many differences i between SFERT, 3CRAX and UCIA, a clear one is that the former were open tank reactors at atmospheric pressure. There was nothing to impede the expulsion of the moderator out of the core., In the UCLA case, the moderator i is in a closed systems in order for the coolant to be expelled, a pressure ! pulse must be generated in the core region, transmitted through the coolant t
1 THE NEOGY IG &RANDUM 120 Se Memorandum provides very little information on the methodology employed, primarily reciting the conclusion reached. he following points can be readily made from what information is provided: the choice of a relatively slow ramp insertion is most unrealistic, the use of clad temperature instead of peak meat temperature is non-conservative, the utilisation of a computer code designed to model Loss of Coolant Accidents (LOCAs) and other transients for BWRs and PWRs for analysis of ) reactivity accidents in small research reactors seems of unproven validity, ) and the use of an adjusted " lambda" seems little more than a " fudge factor." 121. Neogy is said to have " qualified" the RETRAN program for assessing Argonaut research reactor power excursions, a purpose apparently not intended in the original program, by comparing the predicted power trace with an actual measured power trace from a single SPERT IA excursion. Se two did not match, so a fudge factor " lambda" was added, to adjust the predicted estimates to the actual data. Se comparison of predicted versus actual data from SPERT was apparently only done for the one 15 8 meec SPERT transient, where adjustment with " lambda" was found to be necessary. No checking of the program, once modified by " lambda " against other SPERT IA transients is reported, let alone against SPIRT ID, BORAX I, or SL-1 transients. 122. Certain non-conservative assumptions appear to have been used in addition. For example, values such as UCLA's void coefficient, prompt neutron lifetime and delayed neutron fraction are all larger than values reported elsewhere. 123. Furthermore, the very premise of the analysis-- the assumption of a relatively slow ramp insertion- is unreasonable. A person manually pulling a control rod, as in the SL-1 case, or withdrawing a neutron-absorbing sample from an irradiation port, could insert reactivity very much faster than the ramp rate assumed in the Neogy memorandum. Se assumption, then, that the $3.00 insertion would produce an excursion of relatively longperiod(158asec)isinappropriate. Se energy release and temperature estimates that follow therefrom are thus substantially too low. Correction of these assumptions, and consideration of the positive feedback features, void the conclusion that melting will not occur. 124 Again one must emphasise that extrapolations from SPERT to the UCLA Argonaut are fraught with peril. But if one is to make such extra-polations, they should be done with a significant element of conservatism. The analyses done to date have lacked sufficient conservatism and have made a number of other errors. Rather than indicating that the UCIA facility through relatively narrow piping and several junctions to a rupture disk, where sufficient pressure must be generated to cause it to break, and the coolant then to drain out. All of this can take considerable time. Under normal conditions, it takes approximately 20 seconds for 20% of the core water to drain out of the dump valves under pressure it would be faster, but the central question is whether this rather complicated sequence of events for water to be removed from the core would result in a delay over the SPERT/3CRAX shutdown mechanism of simple expulsion out of the reactor tank open top. As indicated earlier, a delay of even milliseconds can mean substantially higher power released. Thus, if Mr Catrander were correct in his explanation of the SFERT/ECRAX differences, the situation for UCLA might be even less favorable than either.
l . . -_ - - . ! is inherently safe with its present or proposed excess reactivity loading, each suggests, upon careful reading, the opposite. 125. Bere are really only two ways to find out for sure whether fuel molting can occur with the assumed excess reactivity insertions. One ' is to do a SPRT-type series of excursion tests at a remote location with an Argonaut core very similar to UCLA's. He other way is for an accidental power excursion to occur at UCIA itself. 2 relicense the UCIA Argonaut
- as is would be to risk the latter form of uncontrolled research.
l REIATED OBSDVATIONS 126. Bere are numerous muchanisms for initiating a power excursion at the UCIA Argonaut-type reactor. Here are basically two categories or ways of initiating the events insertion of positive reactivity and removal of negative reactivity. 127. The original Hasards Analysis recognised one such scenarkos one procedure to achieve anximum excess reactivity in the reactor would be to insert into the reactor a sample with sufficient absorption to prevent start-up. When the controls were fully withdrawn and crit-icality was not achieved, the anximum reactivity would be added if the sample were removed without reinserting the control blades. (p.60) 128. Sus, if a large negative worth sample were inserted for irradiation (either in the enlarged irradiation ports or through the added-on pneuantic
" rabbit" system) and the sample was removed without the reactor operator remembering to first reinsert the control blades, a power excursion could result. Having to rely upon the reactor operator to follow correct procedure, particularly with student operators learning s.t the controls, is precisely the oposite of the basic promise of an educational reactor--
inherent safety, a " forgiving" fail-safe machine, such that the worst mistake possible cannot cause injury. (A potential precursor of such an accident is suggested in the attached November 16, 1981, notice of violation from UCIA.) 129. Substances of large reactivity worth, negative or positive, can be inserted in the reactor, through the pneumatic tube system, the irradiation ports, or through other means. Bare are a number of substances that are highly absorbing and would represent significant negative reactivity worth. If large negative worth samples were neither possible nor anticipated to be needed, then UCIA would have had no need to increase its excess reactivity level from 0.6% to 2 3% if such insertion is impossible or not anticipated, then there should be no reason not to reduce the level back to the design value, at which other Argonaut reactors operate and at which this one did for some time. 130 Just as removal of a large negative sample from the core region, without a compensating prior insertion of control blades, can result in the equivalent of a large positive reactivity insertion, initiating a power excursion, so too can insertion of anterial of positive reactivity. Fission-able materials and perhaps some good moderating materials could have substantial positive worth. Rapid removal of the negative materials or rapid insertion of the positive anterials would have the same effect- a potentially large
reactivity insertion. For example, UCLA at one point requested 250 grams of U-235 for irradiation in the reactor's thermal column. If such material were instead placed in an irradiation port, a very sizeable positive insertion couM result. Datsuchnatorial(orperhapsan unexpectedly good moderator) could be inserted in an irradiation port-as a prank, by mistake, or as an inadequately reviewed experiment- ' couM certainly occur, particularly if there had been a history of weak administrative controls at the facility. ' 131. Mere are numerous other mechanisms for accidentally initiating a power excursion. For example, the facility has had repeated problems with control blades becoming stuck. Se method of trying to free them is to try to torque them free with a hand wrench applied to the drive nochanism, which is located outside the reactor shield and readily accessible (See photos to be filed at a later date in compliance with protective order). While normal insertion rate of reactivity with the control blades should be limited by the motor (if proper speed actors are used), that would not be the case were the blades to be manipulated annually, as in an effort to free them or otherwise to do maintenance on them. (It should be noted that the SL-1 accident occurred during such amintenance to the control red drive mechanism and that a history of sticking control rods, necessitating torquing with a hand wrench, had proceeded the accident.) 132 Other mechanisms of insertion involve water level variations. Should the water level in the core drop for one reason or another, and the reactor be kept critical by further withdrawal of control blades, a sudden rush of water (particularly, cold water) into the core couM result in the equivalent of a substantial positive reactivity insertion. his couM occur during experiments which vary core water level, or through partial failure of the dump valve due to loss of full air pressure which hoMs it in place. The latter would cause some water level drop, which could rapidly be reversed were a surge of air pressure to fully close again the dump valve. Violations of excess reactivity restrictions during core water level experiments, or problems with air pressure to the dump valve, could thus have serious safety implications. (See 1978 Annual 22, 1977, p.4,Report, attached. p.33) and Radiation Use Committee Minutes December 133. Some event which induces some coolant boiling could also result in positive reactivity insertion. If coolant channels were partially restricted, or coolant flow or heat removal slowed, or power slightly overshot, localised boiling might occur, reducing moderator density and requiring further withdrawal of contro1 blades to keep the reactor critical. A sudden fluctuation altering the amount of boiling could result in an insertion of positive reactivity because of the negative void coefficient in the central region or the positive coefficient in the higher ro6 ion. Multiple Failure Modes 134 A number of unfortunate features of the UCLA Argonaut-type reactor create potentials for multiple and common mode failures. For' example, the reactivity change occasioned by e=11 shifting of reactor kmMes within the fuel boxes (changing slightly the gap between the bundles)
6 35-could result in positive reactivity being added in the midst of an excursion which might not, of itself, be sufficient to cause molting. Similarly, expansion or bowing effects that increased the plate spacing could push an excursion "over the top," as could the ini+1=11y positive effect noted upon dropping the water. 135 There are numerous other possibilities as well. One entails a power excursion not sufficient to cause molting by itself but which does involve expulsion of the water moderator. It was noted with the S*ERT reactors that such expulsion would on occasion lead to repeated criticality as the expelled water condensed and dropped back into the core. An excursion limited by moderator expulsion, as at SPERT or BORAX, can send a plume of water and steam high in the air. When that water returns, it does so at a significant velocity, which amounts to a very rapid insertion of substantial excess reactivity. Such behavior.is called " chugging", and on several occasions incidents occurred in which the initial reactivity insertion was not sufficient to cause damage, l but the repos.ted excursions caused by repeated reintroduction of the moderator after expulsion caused increasingly larger excursions which, had the event not been terminated through scr===ing the reactor, might have essentially torn the reactor apart. (A history of sticking control blades which could aske final termination of such a series of excursions impossible would thus have safety significance. Similarly, the lack of deflector plates descrived in the original Hazards Analysis as designed to prevent such repeated excursions by preventing expelled water from returning to the core, means that an important safety feature is missing.) 136. Se positive temperature coefficient for the graphite is troubling as well. A research reactor used by students needs to be inherently safe. Inherent safety necessitates large negative temperature and void coefficients. Any positive coefficients (which are thereby autocatalytic) are to be strongly avoided. This is especially true when the value at-tributed to the positive temperature coefficient for the graphite (+0.006%Ak/k/0)islargerthanthenegativetemperaturecoefficient F citedforthewater(-0.0048%Ak/k/'F). 137. During a power excursion the positive temperature coefficient j of the graphite could provide a feature which makes the excursion more destructive than would otherwise be the case. A portion of the energy lib-ersted in a power excursion is given off as prompt neutron and ganan radiation, resulting in a prcmpt temperature rise in the graphite and other surrounding materials bombarded by that radiation. Even a few degree rise in the graphite temperature would mean the addition of positive reactivity at a time when negative reactivity is needed to limit the power excursion. Se addition of even relatively small amounts of positive , reactivity can produce a slight delay in the shutdown mechanism taking l holdt because of the exponential nature of the excursion, even a milli-second l additional delay can be significant. Given the extremely small margins of safety, e.g., Hawley's 40-50 , even assuming all the assumptions made are correct and the absence of other uncertainties, a slight addition of positive reactivity during the excursion can cause a small margin of safety to become far smaller. I 138. Hawley (p. 15) has Pointed out that excess reactivity in Argonaut-type reactors is usually measured under normal operating conditions and that the negative temperature coefficient of the water thus makes it
- possible that a reactor with a measured level of excess reactivity of, say, ;
--y,ry,,- r-- i,-----m , e- ---,..y -
$3.00, will at times of cold coolant have considerably more than $3.00
- of excess reactivity available. De same is true in reverse for the , positive coefficient for the graphite. l 139. Graphite temperatures rise significantly after an extended run , of several hours in the Argonaut. Flots of reactivity versus time and temperature 6 indicate a rise of approximately 100 0F in 2 hours, to a temperature significantly above the temperature of the water coolant, 1 spparently because the water's heat is continually extracted by the reactor's l heat removal system for the coolant and because auch of the graphite , temperature rise is due to the cumulative effect of heating by radiation l from the fuel. Coolant temperature levels off rapidly after start-up and then remains constants graphite temperature is shown to steadily and continually rise. 140. Sus, if excess reactivity of, say, $3 00, was measured near the beginning of a run, or during a short run, when the water was warm but the , l graphite temperature rise not yet anywhere near its maximum level after a long ' I run, that $3 00 could actually be the equivalent .of substantially more at the end of such a long run, where the coolant temperature would be the same as at the time of the measurement but the graphite, with its positive coefficient for temperature, would be substantially warmer. 141. The positive temperature coefficient has been reported as approx-instely+0.006%4k/k/0F(AECinspectionreport 50-142/68-1, p.6, attached). A temperature rise of 1000F in the graphite, as normally observed after a two hour run, could thus mean an increase in reactivity of 0.6%d k/k, or nearly a dollar. A reactor, thus, that was thought to be limited to
$3.00 could at times have available $4.00 because of the positive ten- ,
perature coefficient. Conversely, to keep to a licensed limit of $3 00, it would be necessary to have a measured anximum of around $2.00, if these figures are correct. l 142. Here can, furth .&., be occasions when the positive graphite coefficient and negative water coefficient interact in such a fashion as to produce a greater reactivity addition than can either coefficient i acting alone. Because heat is extracted so much faster from the water coolantthanfromthegraphitemoderator/ reflector,temperaturecan drop more slowly in the graphite than the water after shutdown, partic-ularly if the reactor coolant system remains functioning after the control blades are reinserted. Thus, after an hour shutdown or so, the reactor might have water substantially cooler and graphite still substantially hotter than the conditions at which the $3 00 limiting value of excess reactivity was measured. One could then have far more than $3 00 available because of the hotter-than-normal. graphite, and additional reactivity on top of that because of the cooler-than-normal water. Sis, in fact, any be the normal reactivity situation of the reactor a few hours after shutdown from a few-hour run. Rose reactivity coef-ficients then would necessitate limiting the measured value of reactivity to less than $2.00 in order to ensure that no more than $3 00 is ever available. We have indicated elsewhere that $3 00 itself is dangerously excessive. l
_37 143. Thus, factors such as those discussed above could mean that the reactor, at the time when reactivity was measured, was below the licensed limit, but at other times, due to positive reactivity effects, was above. Furthernore, the existence of an excess reactivity limit in the license does not mean that ht limit will not otherwise be exceeded, unless the reactor's inherent design does not permit any more reactivity than that level, which is not the case with the UCLA Argonaut. Reactivity is controlled by the amount of the fuel and the effectiveness 'of the moderator, both of which are easily modified in the Argonaut. (Note the large quantity of heavy water stored next to the UCIA reactor, for example.) UCLA has a substantial quantity of spare fuel on siter reactivity is readily added by removing dummy fuel plates from the core and replacing with actual ones. Mis is how refueling is done to compensate for burnup and other factors. 144 Therefore, the fact that the Technical Specificsitions may contain a limitation of $3 54, or $3 00, on excess reactivity does not mean that that limit will not be overshot from time to time, given errors in measurement or violations of Technical Specifications. A history of measuresent errors or Tech Spec violations at such a facility would substantially increase the probability of exooss reactivity limitations. 145 Internetions may potentially occur between power excursion accidents and accidents of other types. For example, various core disruption events could cause or contribute to positive reactivity insertions. Flooding, he it hy pipe break or other event, could add moderation (because of the dangerously undermodersted nature of the reactor) and thus cause a positive excursion. Core-crushing could move the core to more of an optimal arrangement for moderation. Seismic jolt could cause a negative l sample, or a control bisde, to move out of the core region. An event which - i caused the fuel bundle spacing or fuel plate spacing to alter could like-wise contribute positive reactivity. A small seismically-induced excursiss, not sufficient in itself to cause molting, could increase the maximum fuel temperature reached in a crushed core. more detail in the panel on core disruption.)(These will be discussed in 146. Fire could likewise cause some positive reactivity effects. Were the low-melting cadmium control blades to melt out of the core region, a positive reactivity effect could be observed. Vere the graphite to heat up substantially, the positive coefficient could add reactivity to the core. Vere firefighters to use water (or perhaps other moderating substances) to fight the fire, again a positive insertion might result. A power excursion could provide the initial heat necessary to start such a fire. These matters will be disen===d in more detail in the chemical reaction panel, as will be the stema explosions and explosive metal-water reactions which have accompanied several destructive power excursions such as SL-1, BORAK and SPERT. And a power excursion of insufficient ang-nitude to melt the fuel by itself any be sufficient to trigger Wigner energy release, which could add sufficient energy to either melt the fuel or ignite parts of the core. CONCLUSION 147. The UCIA Argonaut, in its current configuration, is not inherently safe. Because of the large amount of excess reactivity, and features by
_~ _ _ . _ _ . . _ _ _ _ . l
't which reactivity can be added, its safety is dependent upon the proper functioning of engineered safety features, strict adherence to proper procedures, absence of operator errors, thorough and careful calibration and maintenance of the equipment, adequate funds and attention devoted to keeping the facility in good condition, strong annagerial and administrative i controls, adherence to regulations and technical specifications, and perhaps most importantly, a healthy respect for the danger to the public that could result from an accident. A belief that no operator error, equipment failure, or other event could possibly cause an accident such as a destructive power excursion would greatly increase the probability of such an accident occuring. So would failures of the Radiation Use Committee to adequately review prop-M experiments or new procedures.
So would a pattern of violation of regulations and technical specifications, as would a pattern of operational unreliability evidenced by repeated unintended scrans caused by equipment malfunctions or operational errors, or, more worrisone, causes that could not be determined. Failure to calibrate adequa.tely devices which activate scram systems, malfunction of such devices, stuck recording pees that lead' to reactivity increases, permitting un-licensed operators like high school visitors to operate the reactor controls-these would all have safety significance. l 148. Se UCLA Argonaut-type reactor is not inherently protected against destructive power excursions of the type which destroyed the SPERT, 3ORAX, and SL-1 cores. De primary inherent self-limiting feature of the UCIA reactor (a relatively slow and small voiding effect in one of the reactor's
- two moderators) is far less effective than that of other university reactors and insufficient to prevent a serious power excursion from d==-4y fuel.
149. Safety margins, which have eroded over the years, are unacceptably small. Se current and proposed licensed limits on excess reactivity appear carable of causing core melting in a power excursion. Even were the licensed limit of core reactivity substantially reduced, such as back to the design level, mechanisms for insertion of larger-than-licensed amounts would still remain (insertion of large positive worth samples, core flooding or crushing, positive acderator effects, violation oftheTechnicalSpecifications,etc.). 150. h e effect of a serious power excursion on fuel elements would obviously be quite severe. One might expect substantial fission product release, perhaps 25% of the radiciodines and a significant fraction of
- other isotopes. The consequences of damage to the fuel from a power excursion would be considerably greater than those arising from d==-15 one of the roastor's 24 fuel bundles during a fuel handling accident, the Hawley muinam credible accident, which assumes a 2 7% release of the gaseous fission products from the dropped bundle. An accident involving dropping a single fuel bundle is not the anximum credible accident at the UCIA reactor.
151. Se type of accident which destroyed the SL-1 reactor could happen on the UCIA campus. Se UCIA Argonaut is not only not inherently safe, but it is not so by a wide margin.
m .., _
)
l l Footnotes in the text
- 1. frea ProceaMngre of the International Conference on the Peaceful Uses of Atonio Energy, United Nations, New York,1956 Vol. 13, p. 79-87
- 2. Source: J.R4 Dietrich, " Experimental Determinations of the Self-Regulation and Safety of Operating Water-Moderated Reactors", in the same volume as proceeding source, p. 89, 99
- 3. Sotree: Miller, Sola, McCardell, " Report of the SPERT I Destructive het Program on an Aluminum, Plate-type, Water-Moderated Reactor",
IDo-16883, June 1964 4 See inspection report 68-01, attached as Exhibit C-I-6, page 6. 5 See AEC Inspection Report dated March 1,1962, p. 4, attached as Exhibit C-I-l See also Radiation Use Committee Minutes, 9/2/81, p. 6-7, attached as Exhibit C-I-2
- 6. see Exhibit A in UCIA's November 9,1981, interrogatory answers to CBG l
REACTOR RUNAWAY Exhibit List Exhibit Number Description C-I-l , AECInspectionReport,3/1/62 C-I-2 Minutes of Radiation Use Committee, 9/2/81 C-I-3 Notice of Violation, UCLA to NRC, 11/16/81 C-I-4 1978 UCLA Annual Report C-I-5 Minutes of andiation Use Cenaittee, 12/22/77 C-I-6 AEC Inspection Report No. 68-01 C-I-7 kcorpts from 1980 UCIA SAR i C-I-8 Excerpts from 1960 UCLA Hasards Analysis C-I-9 Excerpts from 1959 AW Hasards Summary Report-C-I-10 Excerpts from 1958 University of Florida Hasards Su==mv7 C-I-11 Full text, 1960 UCLA Hazazds Analysis-C-I-12 fi1Myideotape,BORAIIdestructtest C-I-13 filMvideotape,SPERTIdestructtest C-I-14 photos taken within the Nuclear Energy Isb l 1
%I Exhibit C-I-1 'page 1 of 4 r
P. A. Norris, Assistaat Director for Rasetors. Divistoa of Compliance, Seedguartere u.
** g ig',
TBE0 : R. 5. Regelhas, Inspection Specialist (teacters) E. T. Dodds. Inspectime specialist (temeters) N Region T. Divistom of Compliance ' g.; , ,. g., g g y '.. "* 3 UNIVERSITT OF CALIFORNIA AT IDS ANCILES (UCIA) TRAINING REACTOR DOCIET ID. 50-142 31Ma3L CD T:RfD
- Attached to our report describias a recent visit to'the subject g facility. The principal reenite of the visit were discussed g with Mr. 5. Klug of SEAR em Jamary 30,1H1.
The report describes two items that appear to constitute 8 violations of the reactor operating license. Specifically, we refer to the following: 1
- 1. During the Spring of 1961, the licensee conducted esperiments
,e in the reactor facility that at that time were not specifically authorised by the license. Until ths. iseusace by DIAR of g
- Ama dment No. 2 on June 28, 1961, theticensee was caly authorised to perfom expertamats outlined la section III of the Basards Analysia - Final Report. These arperiments are very restrictive la that they consist of: Freoperational Testing; Initial Loading precedures Calibrettoo and Shaka-devo; touttaa Start-up and Operation and feel Sandling.
The licensee is now authorised by the broad scope of A-a%t No. 2 to conduct additional esperismeto factuding those that were performed prior to the issuance of the
" nt.
- 2. During a reactor shielding survey in April 1961, the reactor was operated between 20 and 25 Eu. Licensed power to 10 Ew.
Apparently, this wee done without the knowledge or consent of the Easctor Supervising Engineer. The overpower operation appeared *co be an error in judgement on the part of the reactor operator. It is our opinion that this could have been prevented had the scram settings for the safety emplifier careutts been set to prevent operator errors in judgement of this ungnitude. In the UCIA Basards Analyett, Section III-C. Calibration and Shakedown, it states that, "The safety circuits will be recalibrated and their trip points adjusted to 1$0% of normal power." As as additional requirement, we { s W C( * . .h mP- . I i n MO A *.M
~
Exhibit C-I-l page 2 of 4 r
- p. A. Morrie =2* pg i udc feel them the 11eensee should set these tripe so the reactor will he scrammed at a ension of 1501 et licensed, neuer (i.e.15 Ew) and me just 150% (full scale) of the amplifier instrumen: reading. The everpower operation
, occurred one acoth after the reactor reached maxima licenaad power of 10 Ew. We would like to aghanise the fact that the only high flus tripe on the reactor ere those in the two safety amplifiers. ,
we were favorehty impressed with tbs superimental prograno that
- were heias eenducted at the reactor facility. It tas gratifying to see a truly active reactor progree at a university. The preeant reactor staff appears to be very competent.
E
Enclosure:
1aspection Report Origiant and 1 copy y
/
L e P f I
'k I
l i k l I .
kuhibit C-I-l page 3 of 4 D3 AFT f U. S. ATOMIC ENtaCY Cupet18810N REGION V DIVISION OF CEBqpl1ANCE To p. A. Morris, Assistaat Director for Beactors, DATE: a Divietoe of Compliance Readquartere Y I-FaoM : 1. T. Dodde. Inspection Specialist (taattore) Basion V Division of Compliance SURJECT: CWIVERSITY OF CALIFORNIA AT 145 ANCI!.53 DOCERI NO. 50 142 (UCIA)TTRAINING REACTOR , S0letART The Daiversity of California at Los Angeles, California, wee viatted on January "
! 25 and 26, 1962, for the purpose of inspectag the Dniversity: Argonaut-type 10 Ev
' training reactor.
/ 1. A recirculating system bee been installed for the obield tank yater.
- 2. Reactor room structure and usage i,64' observed to be different than that
(* , described in the original application.
- 5. Start-up and operation of ths reactor at 10 Es was observed.
- 4. Water level variation experimente indicate that when the reactor water is dumped, the reactos power initially increases before descreasing.
- 5. Reactor experiments were conducted in the Spring of 1961 prior to receipt of authorisation for these experiments. Amendment No. 2, which authorisee these and other experimente, une teaued by DIAR on June 28, 1961.
- 6. During a reactor obielding survey la April 1961, the reactor was operated at apparent powere of up to 25 Kw. Licensed power levels are limited to 10 Kv.
- 7. The Reactor Basards, Committee appears to be active.
- 8. Results of a abielding survey during 10 Kw operation indicate the existance of game plus neutron radiatica levels of up po approximately 235 mram/br on the reactor top.
P
- 1. The reactor shield tank walls were cleaned by lowering a " frogman" into the tank. Personnel radiation exposure was below the sensitivity of the film badges worn by the individual who performed the operation.
f.
\
p l l i n l Reviewed by:
- 1. B. E p . Region V, Divietoew of Compliance b
l
\'
o
+
Exhibit C-I-1 page 4 or 4 II. Results of Visit (Continued) temperature. 1.e. : appremimetely 70*F. MacLata said that so January 13. 1962, alumiam red spacers were tastalled between the fuel elemmate to beep them from obifting. He sold that the spacers were placed so the fuel elements would be in their meet reactive posittees.
- 3. At the regneet of the taspector, the reactor power wee raised to 10 Ew. After the system had stabilised. the reactor water la and out temperature differential wee moted to be 7'F at a flow rate of 10 kpm. MacLata said that undeg equilibrium canditions best belance ,
studies and demonstrated that 7 F was equivalent to a thermal power of 10 Ew. It was ebeerved that the 1.og N recorder read 100, the #1
. level recorder 971, and the #2 level recorder 881. The. power was g
held at this level until couplettom of a radiacios survey of the facility. The results of this survey are reported under report I - section F. Radietten Moottoriaa. B
- 4. MacLain said that emperimente conducted under the licones amendment for Water Level Variatione had demonstrated that if the reactor water to dumped and the seactor mot acromed initially the reactor power will rise approaimately 15% and then decrease. up said this wee due to the over-moderation of reflected neutrone above the
/
( core. The normal water beight above the core la approutmately 12 inches. . A four tach lead obield to located above the water and on top of the lead is a graphite reflector. The neutrone, traveling through 12 inches of water, become over moderated. As the water level to lowered, the over moderation decreases atti ideal moderetton is reached when the water beight le between 2 and 3 inches above the core. MacLain said that the met effect to worth approximately 0.05% delte E/E. MacLain demonstrated this effect to the inspector. The reactor power wee lowered from 10 Ew to 8 untte. As permitted by the amendment authorising the Veter Level Variatione experiment, the core water level, primmry coolant pump, ehd water flow safety interlocke were bypassed. The dump valve was opened and the water level rapidly dropped. The Log-N recorder initially rose from 0.08 c*o 0.1 and the linear level recorder rose from 8% to 121 before power rapidly decrossed.
- The water level variation experiment was discussed during a recent phone conversationdth Dr. Babb of the University of Waebington.
Be said that they bad conducted a sta11ar experimmat en their Argonaut. type reactor and had determined, by holding ete reactor power constant with the regulating blade as the water level was stouly lowered, that the loss of water level from 8 inches above the core to the top of the fuel was worth an estimated +0.12% delta t/K. Be stated that dropping the water level the next 4 inches {. below the top of the fuel decreased reactivity by 0.53% delta E/K.
*. Mactata stad that the minumum magnet holding current use 90 - 110 m1111mmes. The normal magnet evrrent during operstice uma set at '
150 milliamps. He ante that the blade drop time mesaurammate have demonstrated a total drop time for each blade of approrimetely 300 milliseconde. The time was asseured from the time of secuationd a scram signal untti the blade seat light wee activated. Be said that the blade drive system has had trouble-free operation.
- 6. Reactor taatrumentation. Borner said the lastr uentation had been m relatively trouble-f ree. poor power regulation for the componested too chambers had resulted is neue spurious scrams. To correct this situotton, a John Flube regulated power surtty use taita11ed in April, 1961. Occasionally the period meter had indicated spurious
{g periode. This eoadition has been corrected by cleantes the asumatator of the generator for the teatrument AC supply. Borner pointed out to the inspector that the Leg.N amplifier le bapt in the '11 Calibrate" position when the reactor to not in use. Trouble had been experienced with the instrumente drifttag and getting out of ca11bretton during periods of low power operation., se sold that 4 9
~
g . . Exhibit 0-I-2 page 1 of 3 sheets 111NUTES OF THE R U C MEETING
. OF 2 SEPTEi2ER 1981 TO: Members of the. Radiation Use Committee & Guests FROM: Anthony Zane, Secretary Radiation Use Committee 2567 doelter Hall, Nuclear Energy Lab MEMBERS PRESENT: GUESTS:
I. CATTON J. ALBERT / INTERNAL A'J DIT/UCLA R. CONN C. ASHBAUGH/NEL J. GARRICK N. OSTRANDER/NEL J. KAUFMANN R. REYES/HP/ROS W. WEGST K. SIME/NEL A. ZANE G. SMITH / INTERNAL AUDIT /UCLA Mr. Zane noted that a quorum was present and called the meeting to order. No old business was discussed as the general status of unfinished business had not changed appreciably.
- - The annual report was the first item of business. Mr. Zane mentioned that one q'f scram was not recorded in the annual report. The auditors found the recording error during the audit af ter the report was written. That scram will be added to the report.
Dr. Wegst commented that the statement regarding prudent scheduling for an increase in port operating hours with a concurrent decrease in actual operating time provided a good example of ALARA if anyone wanted to know what kind of an ALARA program we practice. Dr. Kaufmann noted that Table 3 on page.16, film badges 203 and 265, should have the beta symbol appended to their values under the 4th quarter and total columns to be consistent with the rest of their readings as it was beta that was read. Mr. Ostrander commented that as far as we know, gamma exposure of those two badges has never been definitively observed. Dr. Garrick inquired about unscheduled shutdowns (scrams) for which the cause is unknown, and asked what we do about those. In particular, he inquired about the scram where rods number 1 and number 4 dropped without apparent reason. Mr. Zane stated that we looked into the causes but couldn't find any cause. Later, during an N.R.C. inspection, one of the inspectors suggested we put a megger on the magnets to test for short circuits. This was done and all magnets were clear of shorts up to 600 volts. The cause remains unknown. Dr. Garrick noted that this was a 1980 report, and wondered if any new develop-
<Ls ments had shed additional light on the cause. 11r. Zane replied that nothing e new had evolved.
.-,- - - - . - - n. . o
o
. =
Brhibit C-I-2 page 6 hMW3M l Analyses. Mr. Zane replied that the Technical Specification applied to all new
, experiments.
The first experiment was that submitted by Malcom W. Ewell of the California Institute of Technology. ESA number 81-17 Mr. Zane said that the novelty of this experiment was the use of an electrical heater in the thermal column which entails a shielding re-arrangement to bring the heater wires from the thermal column cavity. The main apparent question is the slight streaming of the radiation through the crack left in the thermal column entrance as the inner block will have to go where the outer block fits and the outer block will have to back the other on the outside of the shield. He said that the power had been changed from the indicated 10 kw to 500 watts to produce the desired fluence of ten to the eleventh. Professor Conn remarked that the results of this kind of experi-ment are quite insensitive to the rate at which the fluence is accumulated. Mr. Ostrander. asked if the chamber was pressurized. "Under vacuum," said Mr. Zane. Mr. Zane described the chamber as aluminum with a top cap of aluminum and a bottom cap of lucite. He stated that the chamber contained a minute quantity of U-235 in the form of fission foils and that the amount of fission heat given off would be in the order of 10 microwatts and that the heater power was 3 amps at some low voltage so that the electric power would probably be in the order of 15 watts. He suggested that consideration be given to the partially open thermal column. Professor Catton wanted to know in which direction the streaming would be. "Toward the east and Tokomak facility," responded Mr. Zane. Professor Catton remarked that this was toward the thick concrete wall, c) Mr. Zane then described a 1980 event when some film badges were irradiated in the open thermal column at various reactor power levels. At 700 watts, the reactor scrammed on a secondary radiation alarm, possibly by a neutron inter-action with the scintillator. The area monitors were reading approximately 0.5 mR/ hour at the time. He said that with the partial shielding there really should be no similar problem, but the health physicist should monitor the operation, attend the student in the reactor room, and that all personnel in the high bay will be badged. The committee felt that this experiment represented no hazard to the reactor, personnel, or the general public, and signed the ESA. The next experiment was 81-18. Zane explained that there really was no evident hazard associated with this experiment, but it was new, and committee review was required. Mr. Ostrander described the experiment as trying to create a positive reactivity sample by inserting a good moderator into a void space and therefore chose to insert pol.yethylene into the center vertical hole. He stated that since no one can remember ever seeing a sample exhibiting positive reactivity, this is an item of curiosity. Ms. Sime actually drew up the ESA and the reactivity of the polyehtylene is to be measured at 1 watt using 2-inch increments of polyethylene. Thecommitteeapprovedtheexperiment[Butseetheamendedversionsuggested below by Mr. Ashbaugh and approved by the committee. AZ]. Dr. Garrick asked if there was any way that one could get a positive void coefficient, as some of the old systems using aluminum plate fuel with w'ider n, channel spacing had demonstrated that effect. Dr. Garrick was curious because
- 1 of Mr. Ostrander's mention that we had never seen a positive reactivity experi-ment. Mr. Ostrander said that he had been referring to conventional samples rather than to. experiments. Mr. Zane stated that we can see that kind of effect when we dump the water as we have more than optimum water above the fuel. Dr. Garrick
^p BehihLt 0-I-2 page 7, g3
,, responded that he was hoping that we were both talking about the same reactor
.. [Dr. Garrick is quite familiar with the UCLA reactor].
Mr. Ashbaugh, with an after-thought, asked if.the ESA was already signed, lir . Zane said yes, but if there were any suggestions to state them now. Mr. Ash-baugh suggested that we insert the whole slug of polyethylene at once, do an approach to critical and in that way,- only handle the sample once. His sug-gestion was added to the ESA and initialed by four members of the committee. The last experiment 18-19 was described by Mr. Ostrander. He noted that the interveners had brought up the positive graphite temperature coefficient, and that the University of Washington had actually measured a value for this co-efficient. Our reactor does exhibit a positive coefficient of some sort after an hour or so of operation. Tests show a linear relationship composed of a negative coefficient associated with the water temperature and a positive co-efficient associated with the graphite temperature. If the reactor is run . long enough so that equilibrium is achieved, we get-the kind of coefficient that people normally report for Argonaut reactors. The positive value from our transient tests is approximately the same value as reported by the U. of W. and might be equivalent to the offset of the negative coefficient in the graphite of the Brookhaven reactor as reported to the Geneva convention in 1955, attributed to the expulsion of nitrogen in the graphite or the decrease in the air density. This experiment will follow a different kind of a trans-ient, plus it would allow us to deviate from our normal procedures and also it would allow the primary to get hotter than our normal operating range. itr.
<- Zane put a limit on the primary flow at 40 gallons per minute. Dr. Uegst y suggested ~ a limit on the primary outlet temperature of 180 degrees F. After
, some discussion, Dr. Wegst pointed out that running with the secondary cooling off constituted the unique part of the experiment. The committee approved the ESA. The meeting was adjourned at 11:40 A.M. A. Zane, Secretary Radiation-Use Committee
<w
'*-W.'.
--__m
4.,. .. . . BahiMt 0-I-3 page 1 of 2 r . . . UNIVERSITY OF CALIFORNIA LOS 'ANCELES UCL/ twakte.tQ utses entszet . Lus4Ntxtes e ets r.nstM . sgs onr.o . sis ensscass.cs I ! 94%74 stthmsa s . s.t.s t.s c ..
.. i
- y. <n .
.t 5 -
. COMMUNITY SA FFTY DEPA R'Df E.5
__t *. ' CD OFFICF.OF BESEMICII f OCCUPAT!Ont.SAFET c,: , 1.05 ANCF.M5, CAUFOH.m Mc m . a 25 November 16,.1981 . Director - U.S. Nuclear Regulatory Commission , ,' " Washington, D. C. 20555 o
.a Docket 50-142
~ License R-71
Dear Sir:
~
.:, , Two possible violations of UCLA's Technical Specifications were r.worted by tele-
. .. phone to the USilRC Region V on October 29, 1981. The action cons: m: ting the first possible violation was committed on October 23. On October i, it occurred to us that we might have violated the Technical Specifications, and we subsequentiy reported the action on October 29. We believ'e the reporting delay may also be classified as a violatien. - The first apparent violation was a failure to insert all control blades prior to re-moval of a sample of large negative reactivity (UCLA Technical Specification VII.B.2) No physical consequences ensued as a negative shut down margin of 70 'o 80 cents re-mained after removal of the sample. 1 i
'If Technical Specification VII.B.2.was indeed violated then the' reporting delay is also a violation under Technical Specification VIII.M.lA.
- The precise nature of the violation is uncertain and depends upon the interpretation
-of Technical Specification VII.B.2. That specification clearly applies to a critical
^
- reactor; but does it also apply to a sub-critical reactor containing a' sample of kno..
negative reactivity and known shutdown margin? UCLA requests !!RC clarification of r_. . . .. . _ . Tf[{.thisq st, ion. ,.. _. _ .g ,,
~
'UCLA's Radiation use Co.Tsit't ce was convened on flavembei- 2',1981 t'o review the circum-i stances of the apparent violations. (A copy of the meeting minutes is.available if desired). The following is a management summary of those minutes:
First Violation - Findings .
- 1. A new and novel experiment was run on October 23, 1981. The experiment was in-tended to identify a possible sample of positive reactivity.
- 2. The written procedure for the conduct.of the experiment assumed that the sample t _ would disp 1ay a positive reactivity._
;, . 3-I- - -
. Exhibit C-I-3 page 2 of 2
- 3. The Reactor Supervisor reviewed the written procedure and assumed that standard procedures would govern in the case of unforseen developments. .. .
- 4. The written procedure was not submitted to the Radiation Use Committee.
- 5. The Senior Reactor Operator running the reactor followed 'the pro'cedure as written, but failed to implement the standard procedure when it was found .that the sample reactivity was in fact negative'. -
- 6. The Senior Reactor Operator acted with unnecessary haste, in prosecuting what may have been a deficient procedure. Although he knew his procedure was safe, he did not consider the possibility of a technical violation.
7.. The principal cause of this apparent violation was a failure to anticipate, and
. correctly respond to an unexpected development. , . , . .
First Violation - Recommendation / Approvals . 21 . The Committee approved a recommendation that the procedural aspects of new and novel ESA's.must be reviewed by the Radiation Use Committee prior to implementatio y Second Violation - Findings and Recommendations. -
- 2. All personnel normally expected to notify the flRC were reminded of their res-ponsibilities in that regard.
Very truly yours,
,o j .f_, '
,[, g46- . LL-} f -
I Walter F. Wegst Director . Research & Occupational Safety . WW/jr - cc: Ualnut Creek USilRC Reg. V -
. -9:-
~
. --E : : A . . ,
- a. * ....t- ' w...u : ' V. - . . 2 --* ;-:.-J ?" . - = .* r; n:-/ '
G
* . e4 .
l
l . likhibit C-I-4 page 1 of 4 ANNUAL REPORT UCLA NUCLEAR REACTOR 1 January 1978 through 31 December 1978 i .
.\
r p. }
; g. Reactor Operating Experience f The operations of the UCLA R-1 Reactor totaled 340 port operating
..k hours (a port operating hour is the number of irradation ports used times l T the irradiation time) for the year 1978, and expended 20.3 megawatt hours i' T h,'
',,-of thermal energy. Of the 340 port hours of operation, 71.8% were devoted ;
c 4
. "' to research, 15.8% to class instruction and demonstrations (includes the 2 l- 6, i
' ' training of new reactor operators), 10.5% to reactor maintenance which '
l
, . .0' i , includes calibrations and test runs, and 1.9% to demonstrations for'high '
)
. school groups and other miscellaneous tours.
- g. ;I The total operating time was up 17% over that of 1977; an increase t
.\
attr:butable to a combination of increased demand and reduced down time. '"I 4L
, Table 1 shows the overall comparison witti the four previous years.
! h '
I' " TABLE I .<e-
'9 1.
l[ ' E 1974 1975 1976 1977 1978 I - Research Hours 177 146 158 188 244
, h ._
3 [;;, Class Instruction 28 39 27 88 60 'y:
-[i;
, .t. .~
~ . Maintenance 52 31 23 14 36 I n
TOTAL 257 216 208 290 340 ?
- ^!!
14.8 Mwh(th) 11.9 13.1 15.9 20.3 i 4' f f g' 1 ,-
'c - j -
L. - s; .$.. [7
l
][
B. Unscheduled Shutdowns and Abnormal Occurrences Exhibit C-I-4 page 2 of 4
~i
- 1. Unscheduled Shutdowns j
'C
.a . 30 January 1978. A drop rods scram was initiated by the loss of air pressure which caused the dump valve to partially open. This was reported as an abnormal occurrence and will be described in that section.
- b. 10 February 1978. A full scram occurred approximately one and a half hours into the run when the UCLA campus suffered a mometary
))-
1 power failure. Since the reactor instrumentation is designed to 4 shut the reactor down under these. conditions, the reactor operator : continued the run upon restoration of power. "
- c. 17 March 1978. A High Flux scram occurred approximately 5 seconds af ter a rabbit sample worth $.10 was fired out of the reactor ($ 10 -
worth of reactivity will cause the reactor to go on a stable period of 99 seconds). The scram was initiated by channel 1 of the safety amplifier which under test showed a trip level of 118% rather than - 125%. Since the At 41 recorder which has a slow response time also showed a transient spike at the time of scram, it was concluded that a line transient when the reactor was on a positive transient period above 100 KW could have caused the safety amplifier to trip.
- d. 24 March 1978. A Period Scram occurred as the operator was making his approach to critical. Since the Log N and Period Amplifier was recently replaced and not as yet optimized to the reactor operating system, the operator, being unfamiliar with the new unit, made a normal approach to critical as he had done in the past. When the j Log N and Period Amplifier came on scale, the transient jump re- ,
sulted in a period scram. The operator was cautioned to proceed more slowly until the instrument came on scale then continue in a normal manner. All other operators were again verbally forewarned about the problem.
- e. 31 March 1978. A Period Scram similar to that of 24 March 1978 occurred. All operators were verbally cautioned to hold power at ,'~'
.02W until the Log N and Period Amplifier came on scale and then ,, i proceed to the designated power in a normal manner. In the mean time, continued communication with the manufacturer finally pro-duced an explanation of the behavior. Since the step junction i
through 0 is normal according to the manufacturer, a possible solution would be to decrease the high impedance damping while increasing the low impedance damping. The suggestion was adopted and the action corrected the problem.
- f. 18 May 1978. A High Flux scram was initiated by the safety ampli-fier when the student trainee, unfamiliar with the controls attempted to level off at 100KW and over shot the mark and caused ,pp the safety amplifier to initiate shut down when the power reached -s 125KW. It was recommended that the trainees be more closely super- }j vised while approaching full power until they develope the proper feel and technique for leveling off at an assigned power.
2 e a=-w___ _ __
7:l Exhibit C-I-4 pa6e 3 of 4 e
- 2. Abnormal Occurrances -
y_--
- a. '0:
During a run on 30 January 1978, the reactor experienced a drop 'i.C rods scram. The scram was caused by the drop in pressure of the W Engineering air supply. This allowed the dump value to open E enough to activate the " dump value closed" sensing micro switch. .y. The loss of air pressure was due to the operation of an air tur- 1 . !, bine in another building of the Engineering complex. 1 # The unanticipated reactivity change came about when the dump value, ! being partially opened, bypassed part of the primary flow through the core, thereby causing an increase in moderator temperature. ' Since the reactor has a negative temperature coefficient, the rod under automatic control began driving out to compensate for this loss in reactivity.. The reg rod reached 80% before it was noticed 'E by the operator and the reactor shutdown before the operator could ;, take further action. A check of the hourly readings showed the reg " rod at the 50% position before the drop in air pressure occured. j The problem of the loss of air pressure h'd a been noted previously : and had been brought before the Radiation Use Committee which re-commended the purchase and installation of a back-up air compressor system to the reactor air supply. The purchase order for that com-pressor was initiated on 1 January 1978. - In view of the scram of 30 January 1978, the Radiation Use Committee met on 31 January 1978 and recommended that until the air compressor back-up system was operational, the reactor could continue to run if I the dump value was powered by a high pressure nitrogen cylinder and i an observer stationed at the turbine laboratory. This was immediately implemented by an engineering change order. j
,g l '
The back-up compressor ~ system was installed and became operational on 8 March 1978; at which time the operating procedures were returned to normal.
- b. There were no other abnormal occurrances during 1978.
I p C; Preventive and Corrective Maintenance
- The required annual tests and calibrations were completed and recorded in early 1978. Advantage was taken of a scheduled shut-down in December, 1978, to perform the tests and calibrations for 1979. The various radiation measuring systems were also calibrated on a semi-annual basis'as required l by the Technical Specifications.
l , The corrective maintenance having any safety significance is divided into two catagories, electrical and mechanical. A brief summary of t.he maintenance follows: .. l l 3 ' J R_ #-
1 Electrical: . Exhibit C-I-4 ps6e 4 of 4 5 January 1978. The " Dump Value Clossd" sensing switch checked and
. readjusted for proper operation. ~
17 March 1978. Safety amplifier checked for proper operation af ter a high flux transient scram. Channel 1 tripped at 118% and was re- ; adjusted to 125%. Amplifier checked OK and all circuits functioned g properly. 21 March 1978. The log CIC was removed and checked for proper electrical connections when it appeared to be driving negative even with the chamber i gamma compensation voltage removed. All connections checked out OK al-though there may have been moisture in the connectors.. They were dried ) with a hot air gun before reinstallation. j _ 4 May 1978. The new log N and period amplifier. function switch was 1 d installed and the unit calibration checked. $ 1 10 May 1978. The linear recorder was cleaned and serviced due to irratic operation. 11 May 1978. The Solu Bridge was serviced and checked because of
- 1 dimenished brightness of the eye tube.
18 May 1978. When the integrity of the auto-controller was questioned, a systems check was performed on the unit. The system was found to be . well within specifications. 23 May 1978. The inhibit bypass switch on the Log N recorder required ' readjustment as it appears the trainees were using this disc to turn the recorde.r during the check ~out procedure. , 24 May 1978. The start-up channel was responding to switching tran-l sients. While the problem was probably due to a faulty ground connection l in the preamp, all power connections were bypassed, tube shields left off during the previous maintenance were replaced, and the system cleaned l and checked. The unit now functions properly. 19 June 1978. Fuse replaced in area radiation monitor power supply. 26 June 1978. The voltage reference source in the linear recorder l failed. The unit was red ref and the system checked using the Keithley l calibration current sor ce. ' Mechanical: 11 May 1978. The filter in the shield tank demineralizer system was replaced. g 2 June 1978. A pint of 30% H2O was added to shield tank to remove i the algae growth. 4 x
A- :_...,.
-(
f [L , ~~ ~ , ., . .-..a. ._m. ._.m ~ _ . . __ . . . .
,.c.
( ExhiMt C-I-5 pase 1 of 4 x Minutes of the Radiation Use Committee 1sk s December 22, 1977 Members present: 1. Catton, C. K. Chan, J. Hornor, G. Pomraning, A. Zane Members absent: V. J. Dhir Guests: C. Ashbaugh, N. Ostrander Dr. Catton called the meeting to order. The first order of business was Mr. Hornor's report on his attempts to effect an exchange of in-depth reviews of the reactor operations and procedures with other similar facilities. Since cost was an important factor, Mr. Hornor limited the area of interest to a 200 mile radius which encompassed the following reactors: Cal Poly, UCSB. Northrup, General Atomic, and UCI. The commercial reactors, Northrup.and G.A. were immediately eliminated since much of their work was proprietary. The Cal Poly and UCSB reactors are of low power and are in no way similar to'ours, which then narrows the field down to UCI. The reactor HP at UC1 informed Mr. Hornor that the in-depth review at UCI is done quarterly by a different faculty member and that while this does take one man day, they like the program as it dEk allows each faculty member to become familiar with the reactor program.
'It was suggested that our. director contact UCI's director if this approach is to be pursued further.
Dr. Pomraning suggested that perhaps the nuclear faculty members should take on the task on a rotating basis, to which Dr. Catton named . the available members as being Drs.'Pomraning, Dhir, Chan, and Kastenberg. Dr. Catton being the director would have a vested interest in the operation and therefore would abstain from serving. Dr. Pomraning then inquired as to what was involved in this review. i Mr. Hornor replied that the review mainly consisted of reading the opera-tions log, the reactor supervisors' logs and the procedures for any incon-
. sistencies and possible violations of the technical specifications.
Dr. Pomraning then suggested that Mr. Hornor do the review for 1977 4 and that the four aforementioned faculty members conduct a quarterly h~ review 'for 1978 on a rotating basis under the guidance of Mr. Hornor and possibly Mr.~ Zane provided there is no conflict of interest. The second order of business introduced by Mr. Zane was Mr. Ashbaugh's . and Mr. Ostrander's plan to train personnel from nuclear oriented electri-cal utilities in hot fuel handling. Dr. Catton stated that the reason to consider this type of operation was strictly economic and that monies earned could go the the support of graduate student research. One of the d"k ways to earn this money was through the use of the reactor as a training i tool. In conversations with the utility companies in the area, one of
~ - - -. -. _
Ekhibit C-I-5. pase 2 of 4 Minutes of the Radiation Use Committee December 22, 1977 Page 2 4 their desires is to give their own people hands-on experience in handling hot' fuel. Furthermore, the characteristics of our reactor are quite similar to combustion Engineering's, PNR as well as GE's PNR.
~
Dr. Pomraning inquired as to what would be required in order to
- satisfy the NRC. Dr. Catton replied that the NRC would like to send an observer the next time we did any fuel handling and that they would be invited.
Mr. Ostrander then stated that there may be some complications as it is not quite clear as to what extent the hazards analysis has been superceded by the technical specifications and that the hazard analysis specifics that fuel handling will be done by the reactor operating m ir. t.is say te que:;tiened by the NRC. Mr. Hornor then stated that in checking with EHSS they see no reason why this fuel handling operation can't be handled. However they want a wuitten plan that is approved by this committee since this is the single most hazardous operation that is conducted in.this laboratory. A person handling the fuel cask would require just 4 seconds of exposure for a reportable incident should the. shielding of the cask suddenly be removed and a few minutes to a lethal dose. The approved written plan along with the invitation to the NRC should
~4 be submitted to EH5S at least 90 days prior to the actual operation to
- allow the NRC time to review the operation for compliance with ALARA.
! Another point brought out by EHSS as related by Mr. Hornor was the fact that their trainees who will be handling the fuel are considered students and consequently since the campus community now becomes involved, we must also obtain approval by the Campus Radiation Safety Committee because the liability for student injury falls back on the University. Mr. Zane then brought up the alternative of having the reactor staff remove and store the hot fuel then allow the trainees to load and unload
- the reactor using the new fuel which for all practical purposes can be l handled by hand.
Dr. Catton agreed that this alternative is a good one but we should let them believe that they are handling the hot stuff as part of this training is performance under stress. l Dr. Pomraning inquired as to what the problems are in unloading and inspecting the fuel and about how long would it take? Mr. Ashbaugh replied l that in 1974 when he conducted the operation, 16 fuel bundles were removed l- and inspected in one day and the rest was completed the following day. To l . this, Mr. Hornor remarked that the reactor should be shutdown for at least [- three weeks as the fission decay curve doesn't start to linearize until after the third week. This would then entail a realistic down time of O at least six weeks which includes the training and the reloading of the original fuel. i c
,- azhiMt C-I-5 page 3 or 4 Minutes of the Radiation Use Committee December 22, 1977 Page 3 Dr. Catton and Mr. Zane agreed in response.to Dr. Pomraning's question
. that the six week down time would not adversely affect the normal reactor operation as our average use factor is approximately 6 hr/ week and' by advanced notification to our users, they could adjust their. schedules accordingly.
Mr. Ashbaugh stated that this was really only a feeler to see is SCE would actually go for a proposal such as this whereby Dr. Catton emphasized that while we only expected somewhere in the order of $3000 for this initial operation, it was the visibility that was important as we would like to develop an extensive training program for.the nucicar oriented utilities. Once they were aware of our ability to perform there services, we would expand the training program to satisfy these needs. Dr. Pomraning suggested that a detailed plan be implemented, Dr. Catton commissioned Mr. Ashbaugh to prepare the detailed plan which would be brought before the committee for discussion and approval, after which Dr. Catton and Mr. Ashbaugh would present the plan to the Campus Radiation Safety Committee, and report back to this committee. The next order of business concerned an experiment proposed by Applied Nucleonics Corporation as explained by Dr. Catton. A shield wall cable penetration is to be tested both for neutron and gamma leakage. Special cast concrete blocks will be fabricated to accommodate the cable penetra-
~ q- tion. These blocks with a combined thickness of 6 feet will replace the
-two original two blocks above the reactor core.
! Mr. Hornor stated that EH6S. wished to know how we intended to dispose of 20' tons'of garbage and would we require an amendment to do the experi- ' ment? Mr. Hornor futher stated that he would try to get a reading from
~
the state as to the level of activity an item could have and still be considered non-radioactive. Since there may be follow-on experiments and other multi-uses for these blocks as well as disposal problems it was decided by the committee to give considerable throught to their design such as casting in layers to minimize the disposal problem and to accommodate the origital S foot thick-
- ness should the need arise.
The committee agreed that the project can be executed with no increased j . probability or consequences of any previously analyzed accident nor did it create the possibility of a new type accident provided the radiation levels above the shield were tightly monitored so as not to exceed the current l
- levels above the reactor at full power. The NRC would be advised of the actual equipment once it becomes firmed up so that they may review it before hand.
Other non-agenda business included Mr. Ashbaugh's report on the reactor air supply. - The reactor air is supplied from the main supply of Engineering i 7 Unit I. At times there are experiments conducted in other engineering labs that utilize this supply and can drag the supply pressure down to a point where it could cause the dump value to release which would scram the reactor.
Exhi M t C-I-5 page 4 of 4-Minutes of the Radiation Use Committee a December 22, 1977 p Page 4 Some serious implications of this problem are that the dump, valve
. . closed sensing switch could. be misadjusted which would allow the dump valve to partially open before ini tiating a scram. This would cause reduced flow through the core which would -cause a greater temperature rise and consequently a reduction in activity which would be made up by the automatic withdrawal of'the reg rod. Should the air pressure sudden 1v-revert to normal, it could cause the dump value_to fully close
'thus restoring.the normal flow through the core which would result in a positive reactivity excursion.
The shops and maintenance personnel suggested as a solution to the
-problem that we install a backup air supply tank and compressor with the appropriate check valves that would isolate our supply from the main supply should that supply fail to maintain pressure.
Upon discussion of the problem, the committee recommende'd that 'the laboratory purchase an air compressor system as a back up and install it in the reactor control air supply. . The final order of business brought forth by Mr. Ashbaugh was a request for a power increase to the reactor. The advantages would be that we could increase the sensitivity of our activation analysis as well as possibly enter the isotope production field. Dr. Catton suggested that we do this in stages. The-firs.t stage being to install a proper heat exchanger and clean up the process pit plumbing. The heat exchanger should be at least I megawatt and preferably 2 megawatts. A cooling tower should also be investigated.
-The meeting was adjourned at this point.
A. Zane, Secretary Radiation Use Committee j AZ/li
~ n, , , - --
._ - . . _ . ~ ~
i (1 t (
. . . I Exhibit C-I-6 page 1 of 2 sheets U. S. ATOMIC ENERCY COMMISSION
- f. ,
% DIVISION OF COMPLIANCE REGION V Report of Inspection F' CO Report No. 50-142/68-1 f f
tj University of California h Licensee: at Los Angeles (UCLA) h License No. R-71 ( Category E May 22-23, 1968- - Date of Inspection: . October 11-12, 1967 Date of Previous Inspection: ti)ci0f G~1-G3 Inspected by: W. E. Vetter 7 Reactor Inspector l
?bb. .h. :.. ,. s < ~
i Reviewed by: G. S. Spencer, Senior Reactor Inspector t None Proprietary Information: SCOPE l I Argonaut Type of Facility: 100 kut (500 kwt for short periods of time)
' Power Level _:
Los Angeles, California f I, Location: g Routine, announced S i~ Type of Inspection _: 4 None > Accompanyinz Personnel:
, I:
i .3 l CL S_UMMARY i Safety Items. - No significant safety items were noted.
, h visit.
Noncompliance Items _ - No items of noncompliance were noted during F' g *## 4 - g
, ,,. ;.y :
.,,z.-..-
. ~~* * /,w,'
( ( .. Exhibit C-I-6 page 6 -
~6~ the 2nd of two sheets stabilize nuclear channel long-term operation so that the need for detector relocation can be kept to a minimum.
According to the facility records, during conduct of the 500 kw operation experiment (see Section S. of this report) the nuclear a h power level channels were set to trip at 1157. of 500 kw. In addition, a special linear power level channel was provided and ( set to initiate a reactor scram at 115% of each test power level, q 1.e. ,115, 230, 345, 460 and 575 kw. H 1 . b
- 4. Graphite Temperature Coef ficient A report to the Comnission by the University of Washington (letter to D. J. Skovholt from A. L. Babb, dated January 4,1968) deals with a positive graphite temperature coefficient which had been noted during operation of the Univ ersity of Washington Argonaut reactor. As a result of the subject report, an effort was made during the current visit to identify possible similar effects relative to operation of the UCLA Argonaut reactor.
Dr. Smith informed the inspector that he had received a copy of A. L. Babb's letter and that he had attempted, unsuccessfully, to measure the af fect of graphite heating in the UCLA reactor. He said that preparations for the test had involved the fabrication
' of, a graphite log, which was to be inserted adjacent to a fuct can
( and heated, incrementally, to determine possible reactivity effects. Smith said the experiment had never been performed because the heater wires around the graphite log persistently " burned out" during out-of-core tests. He said the problem was one of inadequate heater wire insulation. However, during the review of the console logbook, the inspector noted that several, three to four hour, reactor operating periods l at 100 kw had been performed. By reference to the console logbook l data concerned with core reactivity changes as a function of time and the temperature of the water moderator, it appears that a positive graphite temperature of 0.0067.dk/k/ F exists. This is about one-half of the coefficient measured during the University y of Washington experiment. Dr. Smith said that in spite of the fore- ; going, he intended to experimentally determine the graphite ( tengerature coefficient as soon as promising test equipment could be l l developed. l C. Core and Internals Control Rod Drop Times a l. Although control rod drop times are not required by the facility license, Mr. Horner agreed during the previous inspection visit
% that drop times should be measured on a routine basis (see CO l
4* :
, ':,. [;i,L 3}}