ML20064K220: Difference between revisions
StriderTol (talk | contribs) (StriderTol Bot change) |
StriderTol (talk | contribs) (StriderTol Bot change) |
||
| Line 2: | Line 2: | ||
| number = ML20064K220 | | number = ML20064K220 | ||
| issue date = 01/06/1983 | | issue date = 01/06/1983 | ||
| title = Affidavit of B Norton.Amount of Excess Radioactivity Requested by | | title = Affidavit of B Norton.Amount of Excess Radioactivity Requested by UCLA Too High,Safety Margins Too Small & Potential for Destructive Power Excursion Unacceptable Due to Nearby Population Density.Prof Qualifications Encl | ||
| author name = Norton B | | author name = Norton B | ||
| author affiliation = COMMITTEE TO BRIDGE THE GAP | | author affiliation = COMMITTEE TO BRIDGE THE GAP | ||
Latest revision as of 22:41, 31 May 2023
| ML20064K220 | |
| Person / Time | |
|---|---|
| Site: | 05000142 |
| Issue date: | 01/06/1983 |
| From: | Norton B COMMITTEE TO BRIDGE THE GAP |
| To: | |
| Shared Package | |
| ML20064K001 | List: |
| References | |
| NUDOCS 8301180331 | |
| Download: ML20064K220 (94) | |
Text
{{#Wiki_filter:p D UNITED STATES OF AMERICA [" N NUCLEAR REGUIATORY COMMISSION ([.[Tg #p, BEFORE THE ATCMIC SAFETY AND LICliNSING BOARD 9 d,g /p y c, ip - In the Matter of 7% 1 lI Docket No. 50-142 [ m f?- ~ THE REGENTS & THE UNIVERSITY \ G'f f3
& CARm (Proposed Renewal FacilityLicense)
Q' (UCLAResearchReactor) DECLARATION OF BOYD NORTON I, Boyd Norton, do declare as follows: 1 From 1960 to 1969 I was employed in reactor safety studies at the National Reactor Testing Station in Idaho. Five of those years ' vere spent as Group Leader of the Nuclear Test Section of SPERT (Special Power Excursion Reactor Test). I was in charge of operation of both the SPERT I and SPERT III reactors. A more detailed description of my professional qualifications is attached.
- 2. The SPERT program centered on investigating the behavior of various reactors during power excursions in an attempt to better understand the potential vulnerability of certain induced by large reactivity insertions. A power excursion is an accident unique to nuclear reactors in which power can rise from zero to billions of watts in very much less than a second. This can result in nelting of the fuel and explosive disassembly of the reactor core.
3 The SPERT I reactor's "D" core was eventually intentionally destroyed in such a power excursion. I was at the reactor controls 8301100331 830112 PDR ADOCK 05000142 C PDR
.' . on November 5, 1962, the date of the final test, and was responsible for ejecting the neutron-absorbing rod from the reactor which made it supercritical and, milliseconds later, resulted in SPERT I-D's destruction, through extensive melting of the fuel accompied by a violent explosion.
1 5 Perhaps the most significant conclusion of the SPERT destructive tests was the unpredictability of destructive threshhold during power eycursions initiated by larae reactivity insertions, even with i 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 i reactivity insertions), the violent explosion which demolished the reactor came as a surprise. Although the BORAX and SL-1 reactors had suffered similar explosions, there had been no prior indication at SPERT that going to a period slightly smaller than that of previous tests represented crossing a threshhold for SPERT which made possible the violent pressure pulse which would demolish the core.
- 6. Thus, 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 considerably more the case with regards the potential behavior of reactors substantially different from the ones on which we performed our tests--for example ,
the UCLA Argonaut. 7 Several analyses, relying heavily on the SP RT 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 ey. cess reactivity. I have reviewed these analyses and a number of other documents relative to the application of the University of California to obtain relicensing of the reactor situated on the
l UCLA campus.1! 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. I 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.
- 8. Based on my review, it is my conclusion 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, given the population density nearby.
The analyses done to date do not, in my opinion, 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 UCLA reactor, or even an answer merely providing reasonable assurance of its being right, would require further research. Because of the substantial differences between the Argonaut and the reactor types previously investigated, that research would likely necessitate SPERT-type tests on actual Argonaut cores. In the absence of such definitive research, very substantial margins of safety are essential at operating Argonaut-type reactors. 1/ Among the documents reviewed were those sections of the following which address excess reactivity issues: UCLA's 1960 Hazards Analysis and its 1980 Safety Analysis Reports the critique thereof contained in the August 25, 1980, Supplemental Contentions of the Committee to Bridge the Gaps the Contentions as admitted into the relicensing proceeding by the Atomic Safety and Licensing Board: " Analysis of Credible Accidents for Argonaut Reactors" by Hawley, Kathren, and Robking and a November 23, 1981, memorandum from P. Neogy of Brookhaven National Laboratory to J.F. Carew, subject:
" Transient Analysis of the UCLA Argonaut." Certain other related documents were also reviewed and are identified in the text of this declaration.
4-9 For these and other reasons identified herein. I conclude that the original restrictions imposed on the UCLA reactor by the AEC in the initial license, particularly the condition limiting excess reactivity to less than that necessary for prompt criticality, were appropriate, given the operation of the device by students. The subsequent changes to the facility have, in my opinion, resulted in a gradual but significant erosion of important safety margins, increasing both the probability and consequences of a potentially serious accident at the facility. The bases for my conclusions are as follows.
- 10. UCLA stated in its original Hazards Analysis in 19602/,
A reactor which is to be 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. Given the operation of the reactor by students, who can be expected to make mistakes, the limitation to 10 kwth and approximately
.6% 6 k/k excess reactivity was quite prudent.3!
- 11. The Intervenor points to a number of developments over the f
I years at the UCLA reactor which have considerably altered the situations a quadrupling of excess reactivity, a ten-fold increase in reactor power, discovery of smaller than expected negative 2/ page 19
]/ The power limitation was important, the Hazards Analysis indicated, because it limits the consequences of an accident, the radioactivity available for should one release to the environment: occur, by limiting /~T_/he
" amount of contained fission products will be relatively small since it is limited i
to a maximum steady state power of ten kilowatts." ibid. 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. l
reactivity coefficients and the unexpected discovery of several positive reactivity effects, plus the addition of a pneumatic tube " rabbit" system which makes possible new mechanisms for rapid insertion and removal of reactivity. The Intervenor also points to a series of violations of AEC/NRC regulations and license conditions, including several relating to rules and procedures designed to control excess reactivity, as well as calibration errors, bypassed interlocks. maintenance problems, and lapses in administrative controls--all of which it alleges increase the chances of accident at the UCLA facility. I concur tha t safety margins have been substantially reduced by these developments, which will be detailed below.
- 12. As I understand it. UCLA argues that none of the alterations or problems that may have occurred during the reactor's operating history to date are of consequence because there are no credible mechanisms for significant fission product release. In particular.
UCLA 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 rest moct of its case in that regard on the assertion that the BORAX I b! and SPERT I tests conducted at the National Reactor Testine Station in Idaho in the nineteen fifties and eixties " proved" i that the requested level of excess reactivity is safe in the UCLA Argonaut. As UCLA put it in its 1980 Application 5[
"SPERT and BORAX tests showed that plate type fuel elements survived step reactivity insertions of $3 54."
13 That simply is not the case. In fact, the SPERT I reactor core was completely destroyed by a $3.50 insertion, which resulted in extensive melting and explosive disassembly of the core. 4/ The BORAX reactor was a reactor similar to SPERT used for a related series of excursion tests at NRTS in the early 1950s. j/ p. V/3-6, cited in Contention I . 3.f.
Non-explosive melting of fuel was observed with even smaller reactivity insertions.
- 14. It is also worth recalling that the grisly SL-1 reactor accident, which occurred at the Idaho Testing Station not far from SPERT, was initiated by about the same reactivity insertion (in the SL-1 case, 2.4% i .3% 6 k/k)h! 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. One of the workers was impaled on the ceiling by a control rod because of the intense radiation field, it took six days to remove the body by use of a remotely operated crane and closed-circuit television. Staying in the building for mere seconds resulted in a i i the rescue workers.E! year's allowable dose of rad at on for 15 Even were it true that plate-type fuel elements survived step insertions of $3 54 at SPERT--which they most certainly did not, as is clearly demonstrated in the attached photos of melted plates from the $3 50 excursion--that would by itself say nothing about whether plate-type fuel would survive the same insertion in the UCLA Argonaut, a different reactor design with significantly different operating characteristics. There is no magical relationship, as the UCLA statement cited in 12 above implies, between reactivity insertion and fuel plate response, independent i
of the reactor in which the excursion is occurring. A reactivity insertion of $3 50 will melt one core, while leaving another 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 6/ $3 54 would be the equivalent of between approximately 2 3% and 2 7%6 k/k, depending upon the value used for the delayed neutron i fraction. 2/ The SL-1 was a low-power experimental and training reactor utilizing highly enriched aluminum-uranium flat plate fuel, cooled and moderated by water, similar to BORAX and SPERT. I
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.
- 16. Even had SPERT not been destroyed by a $3 50 insertion, the UCLA statement quoted above could not be true, because it implies that SPERT and BORAX tests proved that plate fuel could not be damaged by reactivity insertions of $3 54, no matter in what reactor and under what conditions it was placed. And if we found anything through the SPERT 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. The SPERT and BORAX tests could not, by any stretch of the imagination, "show" that a certain general kind of reactor fuel (e.g., flat plate) could survive a $3 50 insertion in any imaginable reactor.
17 The important question, then, is not what reactivity insertion destroyed SPERT or BORAX or SL-1, or even what insertion could be expected to be the minimum necessary to l induce melting in those reactors, but rather, what level is a safe level for the UCLA Argonaut, with sufficient margins of safety consonant with student operation in a densely populated location. Af ter all, SPERT, BORAX, and SL-1 were all destroyed I in the Idaho desert far from any populated center. And the UCLA Argonaut-type reactor is a substantially different reactor than the three Idaho reactors mentioned above.
- 18. The differences are significant. Plate and meat thicknesses are different, as are coolant channel widths. We used essentially fission-product free cores, with fresh cladding--UCLA's fuel has been l
-8 irradiated for two decades, can be irradiated for another two decades if relicensed, and has been sitting in water, corroding the cladding, for many years. Each of those factors might affect the heat transfer time to the water, potentially elongatinz the transient and increasing the energy release, factors not analyzed in the eristing reports, Furthermore, SPERT and BORAX were entirely water-moderated and -reflected, as was SL-1. UCLA 's reactor is moderated by both water and graphite, and reflected by eraphite.
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. Furthermore , the reported void coefficient is smaller for UCLA than SPERT or BORAX, 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 is increased, as by oscillation, are other important differences. UCLA's fuel meat composition is in the eutectic range, thus melting at a significantly lower temperature than would the BORAX or SPIRT fuel meat. There are a number of other differences as well. 19 These differences can be very significant in determininE the energy release from any particular excursion and whether fuel melting will result. Even different reactors of the same general type produced widely different energy releases for the same period, as is shown in the attached plot of energy versus reactor period, taken from Thompson and Beckerly's Technology of Nuclear Reactor Safety, inculded in the Intervenor's Supplemental Contentions at p. V-23 As is shown there BORAX produced substantially more energy than SPERT, and SL-1 more than either, given the same initial reactor period. Seemingly minuto differences in metal-water ratios, temperature and void coefficients, etc. had marked effects on total energy released.
irradiated for two decades, can be irradinted for another two decades if relicensed, and has been sitting in water, corroding the cladding, for many years. Each of those factors may substantially affect the heat transfer time to the water, substantially elongating the transient and increasing the energy release. Furthermore, SPERT and BORAX 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. Furthermore, the reported void coefficient is smaller for UCLA than SPERT or BORAX, 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 i.s increased, as by oscillation, are other important differences. UCLA's fuel meat composition is in the eutectic range, thus melting at a significantly lower temperature than would the BORAX or SPERT fuel meat. There are a number of other differences as well. l 19 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 energy releases for the same period, as is shown in the attached plot of energy versus reactor period, taken from Thompson and Beckerly's Technology of Nuclear Reactor Safety, included in the Intervenor's Supplemental l Contentions at p. V-23 As is shown there BORAX produced substantially more energy than SPERT, and SL-1 more than either, given the same initial reactor period. Seemingly minute differences in metal-water ratios, temperature and void coefficients, etc. had marked effects on total power released. l
- . _ - =
l l _9_
- 20. This is understandable when one realizes that the process of a power excursion is essentially exponential. The nature of the exponential 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 clad and 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.
- 21. 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 UCLA's own 1960 Hazards Analysis, the Hawley et al review, and the Neogy memorandum all rely on the power excursion tests performed at the National Reactor Testing Station in Idaho. UCLA relies largely on the BORAX tests in its orginal analysis: Hawley et al on the SPERT ID series of tests; and Neogy on the SPERT IA series. (Surprisingly, none even touch on the SL-1 accident.) All are based on the fundamental assumption that one can extrapolate with extremely high precision from the SPERT or BORAX tests to the UCLA Argonaut.
- 22. Based on extensive involvement with the SPERT tests. I take substantial issue with such an assumption. First of all, we never intended the SPERT tests to be used in such a fashion.
We were attempting 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, we never intended that a
hazards analyst would simply look at the exponential period at whi ch 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. 23 If the SPERT core was destroyed with a $3 50 insertion, we would have been quite concerned to learn of a reactor operator using that fact as basis for a $3 40 or $3 00 limitation for another reactor, particularly of a different type and in an urban environment. We never intended our SPERT tests to be so used--the uncertainties are just too large. To say, as the Hawley et al review essentially does, that the SPERT I-D core indicating melting beginning around a 7 msec period meant that the UCLA Argonaut could tolerate a 7 2 msec period excursion without any melting or release of fission products goes far beyond the purpose of the SPERT tests and the statistical significance of our data.
- 24. 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 expulsion, fuel plate expansion.
Doppler effect, density changes. " warm neutron" affects, and the final shutdown mechanism, rapid disassembly of the reactor core. 25 'dith the above comments abou t the difficulties inherent in such extrapolations. I will now discQss the three attempts that have been made to extrapolate the BORAX and SPERT data to l the UCLA case. l l i l
The 1960 Hazards Analysis (1980 SAR)
- 26. The UCLA Argonaut-type reactor was designed for a maximum power of 10 kw th and maximum excess reactivity of about .6 %ak/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
.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 .6% safe, but that a sufficient safety margin existed for a training reactor, the Hazards Analysis attempted to estimate, quite roughly, the level at which melting could be expected. This was done to show the magnitude of the safety l margin and to provide further support for the .6% limitation.
27 To make this showing, the Hazards Analysis relied on BORAX data. Obtaining a proportionality from those tests for temperature rise per Mw-sec of energy release, the analyst determined that it would take approximately 41 Mw-seconds of energy release to raise the temperature of the fuel plate from the temperature of boiling water to the melting point of aluminum (not of the fuel meat, which melts at a 36 F lower temperature). Using a chart obtained from the BORAX tests,
-1 it was estimated that an excursion of reciprocal period 150 sec would give an energy release of 41 Mw-seconds plus the energy necessary to raise the plate temperature to the boiling point
-1 of waters i.e., a reciprocal period of 150 sec would prcduce i
enough energy to raise the plate temperature to the melting point - of aluminum, at the center of the hottest plate. l
- 28. 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.
1
Correcting for the different prompt neutron lifetimes, it was stated that that period corresponds to an insertion of 2 3%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 energy release than could BORAX. because of the different characteristics of the reactor - .41 Mw-sec, plus the energy to bring the water to saturation, as the limit for BORAX, and 28 Mw-sec for UCLA. Conversely BORAX 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. ) 29 There is some confusing language in the text of the Analysis which I trace to the fact that it was apparently copied. virtually verbatim, from the University of Florida Hazards Analysis of a few years earlier. (sections attached). The 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 the analyst states that the reactor will tolerate a power excursion of at least that magnitude without melting occurring at the hottest part of the fuel. While this could simply be viewed as asserting that a certain point is the end of the safe zone instead of saying it is the beginning of the danger zone, another interpretation is made clear by reviewing the U of F Hazards Analysis from which UCLA's is copied, essentially word for word, including the language in question.
- 30. UCLA, in its Hazards Analysis, acknowledges the Hazards Report of the University of Florida, asserts that the reactors are similar, with the "only significant difference between the two reactors" being the fuel enrichment--20% for U of F, 90% for UCLA. A comparison of the analyses shows a related difference--the U of F reactor's fuel was 46 w/o U-Al, whereas UCLA's is right at the eutectic point, 13 4. (See page 1 of U of F and UCLA's " Estimation of Effects of Assumed Large Reactivity Additions.") U of P's fuel meat (where the hot spot would be
located) melts considerably above the melting point of aluminum, unlike UCLA's, which melts below the critical temperature for aluminum. Thus, the original Hazards Analysis for U of F was correct in asserting that its limiting period and excess reactivity would not cause melting (because it would be sufficient merely to raise the center of the plate to the melting point of aluminum, whereas the center melts at a considerably higher temperature.)S! UCLA, when it copied the U of F analysis, failed to correct for the different composition of the fuel, thus keeping in language contradicted by the calculations. At UCLA, the same peak temperature predicted by the U of F analysis as not causing melting would cause such melting.
- 31. As the original Hazards Analysis calculations make clear, 2 3%o k/k would be sufficient to cause fuel melting at UCLA, if the assumptions employed are correct. I have made clear above my objection to such extrapolations from one reactor type to another in the absence of empirical evidence from tests like we conducted at SPERT or very significant error bars at each point in the calculation. As I 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 .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 i .6% co there would be a margin of safety for errors in calculation l
or operational errors that might slightly exceed the license limits. As I read that analysis, it shows melting at 2 3%6 k/k, and supports a .6% limitation. It cannot be used, in my view, to justi fy a limit at or close to 2 3%. 8/ Peak temperature will occur in the fuel meat, where the energy is generated. If the meat melts above the melting point of aluminum, the meat won't melt; and the aluminum, which is not where the peak temperature is generated, will not get hot enough ( just barely) to melt. This would not be the situation at UCLA, where the meat melts at a lower temperature than the clad.
- 32. The Intervenor in its Supplemental Contentions points to a number of aspects of the original Hazards Analysis with which it takes issue. For example, the Hazards Analysis uses a void co efficient 2I-for UCLA of .18% k/% coolant void, whereas the current application cites a value of .164% (p. III/6-5).
If UCLA's reactor has a smaller void coefficient than initially thought, its capacity to tolerate certain excess reactivity insertions is substantially reduced, and fuel melting could thus occur at substantially less than a 2 3% a k/k reactivity insertion. Uncertainties in the precise void coefficient (which can vary by region of the core and other variables) adds substantial reason for added margins of safety. 33 In addition, as has been pointed out above, the Hazards Analysis calculations appear to neglect eutectic meltine. The calculations were based 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 Analysis would bring the fuel to melting. Using the figure of 24.4 F/Mw-sec supplied in the Hazards Analysis, about 1 Mw-sec less energy would be required than previously estimated, producing a commensurate reduction in the amount of excess reactivity necessary to produce fuel melting.
- 34. The Intervenor also points out that the Hazards Analysis used a non-conservative delayed neutron fraction ( O ) of .0074, whereas the Application now cites a figure of .0065 4 is important in the conversion from period to excess reactivity through the "inhour equation." Use of the form of the inhour equation cited in the Hawley review (p. 16) shows that use of the smaller O 2/ A void coefficient is a measure of the effectiveness of the primary inherent safety mechanism in light-water-moderated, highly enriched flat-plate reactors--production of voids in the moderator.
In a power excursion, the power rise causes the temperature to rise, which eventually can cause the nearby water to boil. The density reduction reduces the amount of moderation and increases the neutron leakage, slowing the nuclear reaction and eventually stopping the excursion. The larger the negative coefficient, the more reactivity can be compensated by the voids.
results in a shorter exponential period for the same reactivity insertion, and thus more energy release and higher fuel temperature.1E! Conversely, use of the smaller 8 means a smaller reactivity insertion will produce the same result (i.e. fuel melting) than estimated in the Hazards Analysis employing the larger figure. 35 If the Hazards Analysis concludes that a 2 3%d k/k insertion will bring the hottest parts of the fuel to the melting point of aluminum--and it clearly does--then use of the smaller figures for void coefficient andk , 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.
- 36. There are a number of other factors which should further substantially reduce downward the Hazards Analysis estimate of the excess reactivity necessary to induce melting--the effect of cladding corrosion and fuel irradiation (which reduce thermal conductivity and thus delay shutdown), as well as initial moderator
! temperature, to name just a few. Although the Analysis conservatively assumed the 2 3%Ak/k insertion to occur in a subcooled reactor, i the Hawley review at p. 15 rightly points out that excess reactivity I 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 actually much more at lower-than-normal temperature. Conversely, if 2 34 is dangerous on a cold day. far less than that amount must be installed if measurement is under warm moderator conditions. 37 Thus, given the basic assumptions employed in the Hazards l Analysis, and the numerical values utilized, the Analysis' calculations predict fuel melting with insertions in the range of 2 3%. When a few 10/ The version of the inhour equation cited by Hawley is T = [ak/k (1- d eff) -6 eff]
] I 1 of the numerical values are changed to reflect more appropriate values (@ , void coefficient, and eutectic melting point), substantially less than 2 3% &k/k would appear to be sufficient to induce melting--if th'e methodological assumptions employed are correct. If measured warm, even smaller levels are tolerable.
- 38. I have indicated above, however, that I have considerable reservations about methodology. To use such a method of extrapolating from one reactor to a different one--to three significant figures -without error bars, assumes that there
! exists a complete knowledge of all the differencas between the
; reactors and how those differences affect behavior. As has been
! shown, a number of differences were not considered, and to assume s that what differences are considered can be corrected for using i 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. Substantial errors bars are required, or margins of safety, which is why the Hazards Analysis was correct in concluding i that excess reactivity at this facility should be limi.ted to about .6% k/k. The Analysis cannot properly be used to demonstrate that $3 54, or $3 00, or any similar value is " safe": in fact, , if the methodology of the Hazards Analysis were to be accepted, it must be taken to demonstrate that such levels of excess reactivity could lead to fuel melting in the UCLA reactor. The Hawley, Kathren, and Robkin Review 39 The section of the Hawley, et al, report dealing with excess reactivity issues appears to consist almost exclusively of a brief literature review and some extrapolations from the SPERT I
i l tests. Whereas the 1960 Hazards Analysis took into account a number of differences between the UCLA Argonaut and the BORAX reactor, from which it was extrapolating its data, the Hawley review does not account for several of the UCLA-SPERT differences. particularly UCLA's smaller void coefficient, which would tend. ]* if not otherwise compencated, to suggest that an excursion of the same period in SPERT 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 UCLA. a factor which is helpful to UCLA.
- 40. 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 SPERT I-D. and then scale temperature linearly to the peak temperature estimate during the SPERT I-D destruct test. These approximations are very rough, as can be seen l from the plot of energy release to time of peak power vs.
reciprocal period, taken from p. 11 of the Miller. Sola, and McCardell report on the SPERT I Destruct Test, which Hawley et al used to estimate energy release for a 7 2 millisecond period. ! 41. And yet, even without taking into account factors such as void coefficient differences. which would tend to produce higher temperatures, the analysis estimates peak fuel temperatures only i about 50 below the melting temperature. No error bars whatsoever are provided for the extrapolation steps nor the final conclusion. (I note hat there appears to be a subtraction error in that Hawley et al assert on page 19 of their report that a hot spot of 586 C would be 74 c below the melting point of the fuel meat, which they cite on the previous page as being 640 C.)
- 42. I do not consider 50 to be an adequate margin of safety, particularly when so many of the differences between SPERT and the UCLA Argonaut were not taken into account. Furthermore, significant effects may appear just below the melting point. such as volumetric expansion of the fuel resulting in cladding failure.
\ \
+ . l l
or considerably increased diffusion of fission products through the hot metal. We noted, for example, at SPERT that some of the fuel plates were very substantially softened and wa rpe d , 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 the final destruct test 43 So even were Hawley et a_1 correct in their estimate of peak temperatures 50 or so degrees below the melting temperature. I would still have concerns. However, questionable assumptions used by Hawley et al suggest far greater temperatures could be achieved in the UCLA Argonaut than those estimated.
- 44. Perhaps the most questionable assumption is that ,
a 7 2 mnec period would produce a 12 Mw-second energy release in the UCLA Argonaut. Given the linear scaling assumption of ) temperature to energy release employed by Hawley (p. 19-- 1500 C per 30 7 Mw-seconds, or about 49 C/Mws), a 13 Mw-second energy release would cause melting, if Hawley's assumptions are accepted. That is not much of a margin of safety if his 12 Mw-second estimate is correct. 45 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 Mw-sec, plus the energy necessary to raise the fuel to the boiling point of water. How a longer period is estimated to produce 2h times the energy assumed in the Hawley l report is not explained.
- 46. Note also that a 7 msec period in the SPERT I-A core is reported to have released 23 Mw-see of energy, nearly twice that i assumed by Hawley based on SPERT I-D data. (Schroeder, 1957).
I The plot of period versus energy release (Thompson and Beckerley, 1964), mentioned earlier, likewise shows how the choice of 12 Mw-sec for a 7 2 maec period is quite non-conservative. SL-1 extrapolations, for example, would suggest an energy release five times greater than that assumed by Hawley. When one recalls that an energy release of 13 Mw-sec would cauce melting, if Hawley's other l
._ . . - . - . - -. _ _ _ _ . - ~ . - - _ _ - _ _ _ _ - _
assumptions are correct, then data suggesting releases of 23, 28+, and even 60 Mw-sec of energy from a 7 2 msec period excursion indicate an unacceptable probability of a destructive power excursion, one that could release significant amounts of l fission products. '
- 47. I should add once again, however, that the methodology a'
of very simplified extrapolation from SPERT or BORAX data to the UCLA Argonaut case, as done in the Hawley report, seems to me 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 1 it happened. It released several times more energy than ) Hawley's extrapolatior. from SPERT I-D would predict, even I though it was much more similar to SPERT than is the UCLA Argonaut. The Hawley extrapolations cannot be relied upon to prevent an SL-1 type accident at UCLA, one that would occur not in a remote federal testing station but in the midst of tens of thousands of people. The Neory Memorandum
- 48. The Memorandum provides very little information on the methodology employed, primarily reciting the conclusion reached.
The 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 utilization of a compttter code designed to model 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."
49 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 I-A excursion. The two did not match, so a fudge factor " lambda" was added, to adjust the predicted estimates to the actual data. The comparison of predicted versus actual data from SPERT was apparently
- only done for the one 15 8 msee SPERT transient, where adjustment with " lambda" was found to be necessary. No checking of the program, once modified by " lambda," against other SPERT I-A transients is reported, let alone against SPERT I-D, E0RAX I, or SL-1 transients is indicated.
- 50. Certain non-conservative assumptions appear to have been used in addition. The values for UCLA's void coefficient.
prompt neutron lif atime and delayed neutron fraction are all j larger than values reported elsewhere. In addition, the i ! assumption of a relatively slow ramp insertion is unreasonable. A person manually pulling a control rod, as in the SL-1 case, I 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. The assumption, then, that the $3 00 insertion would produce an excursion of relatively long period (15 8 msec) is inappropriate , and the energy release and temperature estimaten that follow therefrom are thus substantially too low.
- 51. Again one must emphasize that extrapolations from SPFRT to the UCLA Argonaut are fraught with peril. But if one is to make such extrapolations, 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 UCLA facility is inherently safe with its present or proposed excess reactivity loading, each suggests, upon careful reading, the opposite.
1 i 52. There are really only two ways to find out for sure l whether fuel melting can occur with the assumed excess reactivity insertions. One is to do a SPERT-type series of excursion tests at a remote location with an Argonaut core very similar to UCLA 's. The other way is for an accidental
; power excursion to occur at UCLA itself. To relicense the UCLA Argonaut as is would be to risk the latter form of f t
uncontrolled research. I i Related Observations 53 There are numerous credible mechanisms for initiating a power excursion accidentally at the UCLA reactor. , For example, it is my understanding that the facility has had repeated problems with control blades becoming stuck, and that the method of trying to free them is to try to torque them free with a hand wrench applied to the drive mechanism, which is, as I understand it, located outside the reactor shield. While norma) inae" tion rate of reactivity with the control blades is limited by the motor, that would not be the case were the blades tc be manipulated manually, 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 maintenance to the control rod drive mechanism and that a history of eticking control rods, necessitating torquing with a hand wrench. had preceeded the accident.)
- 54. Another such mechanism for rapid insertion of reactivity is for a large negative worth sample to be irradiated (either in an irradiation port or through the pneumatic tube " rabbit" system) and for the sample to be removed without the reactor operator remembering to first reinsert the control rods. Havi ng to rely upon the reactor operator to follow correct procedure, particularly with student operators, is precisely the opposite of the basic premise of an educational reactor--inh? rent Fafety such that the worst mistake possible cannot cause injury.
i 55 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 fissionable material directly into the core. I understand that UCLA at one point requested 250 grams of U-235 for irradiation in the reactor's thermal column. If the material were instead placed in an irradiation port, a very sizeable positive insertion would result. That such material could be inserted in an irradiation port instead of the thermal column--as a prank, by mistake, in an act of sabotage, or as a modification of an experiment-- ] could certainly occur, particularly if there had been a history ! of weak administrative controls at the facility.
- 56. Furthermore, the fact that Technical Specifications 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 Tech Specs.
A history of measurement errors or Tech Spec violations at such a facility would substantially increase the probability of excess reactivity limitations being violated. 57 Other mechanisms of insertion involve water level variations. i 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 into the core would result in the equivalent of a substantial positive reactivity insertion. This could occur during experiments which vary core water level, or through partial failure of the dump valve due to loss of full air pressure which holds it in place. The latter would cause some water level drop, which could rapidly be reversed were a surge of airpressure 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.
l i l l
- 58. 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, localized boiling might occur, reducing moderator density and requiring further withdrawal of control 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.
59 I understand u.'uv that as a result of vibration tests on the reactor, it was determined that reactivity oscillations wer; detected, traced to the fact that the reactor is substantially undermoderated with its present plate spacing. If true, this information would be significant because core distortions, for exampl;, those created in an earthquake or an otherwise non-destructive power excursion involving rapid steam formation and water expulsion, could potentially have the equivalent of increasing plate spacing and thus amount to positive reactivity insertion. Furthermore, an undermoderated core presents the possibility of power excursion through increased moderation being introduced. The Hawley review indicated that up to 18 5% ok/k es.tra reactivity could result from catastrophic mechanical rearrangement or flooding of the core (p. 27), but concluded that such perfect rearrangement or complete flooding was not credible. However, with 18 5% available, far less than perfect rearrangement or complete flooding is necessary for a disastrous power excursion, which all the analyses would appear to accept as occurring at least with a 3% insertion, if not considerably less. Thus, flooding from broken pipes, the shield tank, or the failure of a nearby upstream reservoir could result in a substantial positive insertion, as could the use of water to fight a reactor fire.
ll
- 60. There are numerous other possibilities as well. One entails a power excursion not sufficient to cause melting by itself but which does involve expulsion of the water moderator.
We noted with the SPERT reactors that such expulsion would on i 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. We called such behavior " chugging " and found on several occasions incidents in which the initial reactivity insertion was not sufficient to cause damage, but the repeated excursions caused by rep 9ated reintroduction of the moderator after expulsion caused increasingly larger excursions which, had the event not been terminated by us through scramming the reactor, might have essentially i torn the reactor apart. (A history of sticking control blades
- which could make final termination of such a series of excursions impossible would thus have safety significance. Similarly, if the UCLA Argonaut does not have the deflector plates described in the original Hazards Analysis as designed to prevent such repeated excursions by preventing expelled water from returning to the core, then an important safety feature is lacking.)
- 61. A fire in this reactor raises serious reactivity questions as well. If water, or some other moderatina substance, were used to suppress the fire, a power excursion might result.
If the control blades melted out of position, thb equivalent of a large positive reactivity insertion might ensue. Fur the rmo re , the positive temperature coefficient of the graphite means that as the temperature rose in the graphite, reactivity
.could increase as well. All of these factors would necessitate very careful plans for fire response , andcould make a fire at the reactor quit; se ri ou s .
i
1 l t
- 62. The positive temperature coefficient for the graphite is troubling for other reasons as well. A research reactor used j 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 attributed to the positive temperature coefficient for the graphite (+.0067 4 k/k/ F) is larger than the negativa temperature coefficient cited for the water ( .00484 4k/k/ F). 63 During a power excursion the positive temperature coefficient of the graphite could provide a feature which makes the excursion more destructive than would otherwise be the case. A portion of the energy liberated in a power excursion is given off as prompt neutron and gamma radiation, resulting in a prompt temperature rise in the graphite and other surroundine 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. The addition of even relatively small amounts of positive reactivity can produce a slight delay in the shutdown mechanism taking hold because of the exponential nature of the excursion, even a millisecond additional delay can be significant. Given the extremely small ' margins of safety, e.g. Hawley's 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
V ! -?6-l 64. Hawley (p. 15) has pointed out that eycess reactivity in Argonaut-type reactors is usually measured under normal 4 operating conditions and that the negative temperature ! coefficient of the water thus makes it possible that a l reactor with a measured level of excess reactivity of, say,
$3 00, will at times of cold coolant have considerably more than $3 00 of excess reactivity available. The same is true
- in reverse for the positive coefficient for the graphite.
65 Graphite temperatures rise significantly after an extended run of several hours in the Argonaut. Plots of reactivity versus time and temperature, included as Exhibit A in UCLA's November 9,1981, interrocatory answers to CBG, indicate a rise of approximately 100 F in P hours, to a temperature significantly abovo the tomperature of the water coolant, apparently because the water's heat is continually extracted by the reactor's heat removal system for the coolant and because much of the graphite temperature rise is due to the cumulative effect of heating by radiation from the fuel. Coolant temperature levels off rapidly after start-up and then remains constant; graphite temperature is shown to steadily and continually rise.
- 66. Thus, 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 graphite temperature rise not yet anywhere near its marimum level after a long run, that $3 00 could actually be the equivalent of substantially more at the end of such a long run, where the coolant temperaturo would be the same as as at the time of the measurement but the graphite, with its positive coefficient for temperature, substantially warmer.
67 The poritive temperature coefficient has been reported as approximately + .006%6 k/k/ F (AEC inspection report 50-142/ 68-1, p. 6). A temperature rise of 100 F in the graphite ar normally observed after a two hour run, could thus mean an increase in reactivity of .6%dk/k, or nearly a dollar. A reactor, thus, that was thought to be limited to $3 00 could thus at timee have available $4.00 br.cause of the poritive t:mperature coefficient. l
t Conversely, a kaep to a licensed limit of $3 00, it would i thus be necessary to have a measured ma..imum of around 3?.00, if these figures are correct. i 68. There can, furtharmore, 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 acting alone. Eecause heat is extracted so much faster from the water coolant than from the graphite moderator / reflector, temperature c2n drop more i slowly in the graphite than the water af ter shutdown, particularly if thc reactor coolant system remains functioning aftcr tha 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. I This, in fact, may be the normal reactivity situation of the reactor a few hours after shutdown from a few-hour run. These reactivity coefficients 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. I have indicated elsewhere that I think $3 00 itself is dangerously excessive. 69 The UCLA Argonaut, with the levels of reactivity beinF
- requested, is not inherently safe. Because of the large amount of e>cass reactivity, its safety is dependent upon the proper functioning of engineered safeiy 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 1 in good condition, strong managerial and administrative controls,
strict adherence to regulations and technical specifications, and perhaps most importantly, a healthy respect for the danger j to the public that could result from an accident. A belief that no operator error, equipment failur9, or other event could possibly cause an accident such as a des ructive power excursion i would greatly increase the probability oi such an accident happening. So would failures of the Radiati on Use Committee to adequately review proposed experiments or new procedures, i So would a pattern of violation of regulations and technical specifications, as would a pattern of oparational unreliability evidenced by rapeated unintended scrams cauced by equipment malfunctions or operational errors, or, more worrisome, causes that could not be determined. Failure to calibrate adequately devices which activate scram systems, malfunction of such devices, stuck recording pens that lead to reactivity increases--these would all have safety significance.
- 70. Permitting unlicensed operators like high school or junior high school visitors to operate the reactor controls would likewise increase the risks. SPERT III, for example, suffered a serious accidental power excursion when an unlicensed operator was permitted to bring the reactor to critical. A licensed operator was present in the room, but tha unlicensed operator was at the controls and, in violation of our rules at the time, a senior operator was not also present.
- 71. Use of a low-enriched fuel would add some safety margin to the facility, becauce of the incraased Doppler effect.
At SPERT we found a low-enriched, uranium oxide core able to withstand larger reactivity insertions than the high enriched uranium-aluminum plates. Samples of large reactivity worth, negative or positive, i 72. J j could be inserted in the reactor, oither through the pneumatic tube system or into irradiation ports. There are a number of substances, such as cadmium, that are highly absorbing and would represent significant negative reactivity worth. Fissionable
_-. ._._ ~ .__. - . . -
. materials and perhaps some good moderating materials could have substantial positive worth. Rapid removal of the negative materials or rapid insertion of the positive materials would have the same effect--a potentially large reactivity insertion.
If large samples such as these are not anticipated to be used, then there is no reason not to bring the e., cess reactivity level back down to less than .6f,8 k/k, where it used to be, because there would never be large samples to irradiate for which substantial excess reactivity was necessary to componsate for the neutron absorption. 73 Steam explosion and metal-water reaction are possible > if a power excursion of the magnitude that appearn possible at UCLA were to occur. SPERT, BORAX, and SL-1 all had such reactions explosively occur. As indicated above, the onset of such reactions and the initiating conditions are not fully understood. I am not certain, for example,that such reactions could not occur even if maximum temperatures attained in the fuel were slightly below the melting temperature. An incident that approached but did not quite reach the melting temperature could l still release fission products through cladding failure , etc.
- 74. Furthermore, were substantial Wigner energy stored in the graphite, an excursion not producing enough energy to melt the fuel alone may still have enough to trigger the Wigner release, which could add enough energy to bring the fuel to melting or ignition of either the graphite or fuel.
Furthermore, the means of shutdown, rapid expulsion of water, generates substantial pressura pulsos capable of subetantial alterations to the core configuration. This could increase pathways for airflow to feed the fi re . If the es.cursion were ! very severe, shield blocks weighing several tons could be thrown in the airs as mentioned above, the SL-1 reactor vessel itself was lifted nine feet in the air during the excursion. t i 4
CONCLUSION l 75 Based upon the reviews that have been done to date of the potential for power excursions capable of causing fission product in an Argonaut reactor with a $3 00 or $3 54 licensed limit on installed eycess reactivity, such an accident is credible at the UCLA facility.
- 76. The fission product release from such an accident could be substantial. The consequences would be considerably greater than those arising from damagina one of the reactor's 24 fuel bundles during a fuel handling accident. A considerably larger fraction of the core inventory can be released in an accident at this facility than the .2% of the gaseous fission products assumed in the Hawley study (2 7% of the gaseous products in a single bundle containing 7% of the core inventory) at p. 48.
An accident involving dropping a single fuel bundle is not I the mayimum credible accident at the UCLA reactor. 77 Poor managerial controls, violations of regulations and i technical specifications, inadequate maint nance, aging equipment, operation by unlicensed operators, operational unreliability, and an attitude that no accident is possibla could all substantially increase the probability and magnitude of a potentially destructive power excursion.
- 78. The UCLA reactor is not inherently safe, with the levels of excess reactivity being requested. And even were the licensed I level substantially reduced, mechanisms for insertion of larger-than-licensed amounts would still remain (violation of the license,
[ I mistaken insertion of a large positive sample, core flooding, etc.). Its lack of inherent safety makes it escential that the strictest obedience to regulations and procedures be observed, adequate l managerial controls enforced, and excellent maintenance and l calibration be conducted. Because it is used for student instruction, these controls cannot be counted on, which is why inherent safety or added engineered safeguards are necessary in l l training reactors. i
_ y1_ 79 Copying of Hazards Analyses, which turned out to be copied from other analyses for other reactors, is a poor approach which can reproduce and magnify errors.
- 80. The lack of containment structure, emergency core cooling system, emergency radioactivity removal systems, shields for the control blade drive mechanisms located outride the reactor shiuld. HEPA filters , spare control blade motors, emergency radioactivity holding tanks, and most particularly, siting characteristics which provide substantial distance between the reactor and densely populated areas all can exacerbate the consequences of an accident at the facility and/or increase the likelihood thereof. Sticking control blades and the problems discussed above regarding Fraphite add to the potential for trouble.
- 81. A destructive power excursion, of the SPERT/ BORAX /SL-1 type, cannot be ruled out at the UCLA Argonaut-type reactor.
Such an accident can credibly occur there, and the consequences could be severe, given the high population density and lack o f containment. I declare under penalty of perjury that the foregoing is true and correct to the best of my knowledge and belief. 3 Boyd{fortonf) Dated this / day of A , 19 3 at Evergreen. Colorado
. _- . -_. ~. - - _ - . -_-.. - - _. . _. ._. .
' a 4
Professional qualifications BOYD NORTON My name is Boyd Norton. From 1960 to 1969 I was employed in reactor safety studies at the National Reactor Testing Station in Idaho. During that period I was in charge of operation of both the SPERT I and SPERT III reactors. The SPERT program (Special Power Excursion Reactor Test) was designed to investigate ! the vulnerability of certain kinds of reactors to accidents induced by large reactivity insertions and to understand better the behavior of reactors during power excursions. From 1960 to 1962 I was staff physicist at NRTS, assigned to SPERT I. From 1963 to 1968 I was Group Leader of the Nuclear Test Section of SPERT. I was in charge of operation of both SPERT III and the rebuilt SPERT I. In 1968 I became Section Chief. Euperiment and Analysis Section, of the Power Burst Facility at NRTS. I was in charge I of the Safety Analysis Report for the PBF. Prior to my arrival at NRTS, I received a Bachelor of Science degree in physics from Michigan College of Mining and Technology. I worked summers during 1954-59 at M&C Nuclear, later a subsidiary of Tryas Instruments. M&C Nuclear was a metallurgical laboratory where I did research on fuel elements for nuclear submarines. In 1969 I was offered, as a result of extensive involvement i in conservation work in Idaho , a job with the Wilderness Society, where I remained till 1971 I am now a freelance writor and photographer, mainly on conservation themes but also on nuclear s matters.
.A "A
Th5 P 1 e(j!:'
.+
e N4
._ 4
~
4 :- 1 5k 6 i.;
.\
Np. 1:e i1 y;9 y
.w?
'N y.
A s ; g
# ^
, h,
.i _-
c e j.c'9 ;- (}
,,iU
/Mi (~
.- fo-p N'N b-e E*:
L I E8
=
h UN (.
%g z
Q L ry:; :-g
% +-
e m f ? 3.,;. *
,$h@ f9.: m 1A s
\
,m , , ,1 - - - - - *
- 42
se 4 y-l
\ c ll l i
i, ) l r {M 1
?
rt xpg *
.- p i
% ;Y a
^ . !,'
i f -
. g
\ ,
5
'o ,
~
~~
- n
lj l i ..
./
// ,
~
I v> . Q c . EI C. 35 ' I P l ateg , !!Phe ay ##*
- bly.
- MSition 41
/
--- -...___.- - ~. . . - . . _ - . . _ _ _ _ - .
)
i
- ?
J
- * ~= ~
T Y
\' +
4 . e, p _ A% K N T
, l' '?ll. ', j.
WA [ ll l ll M.W W/<,.W:r- -
-!!!;7 l:
1 7 .s i
'
- 4 ', {
l -l % ,
,. *i $-
l
,_ l ,
Fic. 36 Clo>e vn w in vessel af ter riestrw tive test. T r4 A N $ g g r, t P(. DCS
.x e 'L' ,
t f r p . i o r, i
'* '4 A N9 9 E I vu. n case view m msel mer cemucuve test. {
40 ? i wwm "- -- - = w
- e
'4 W 23- .- -----r. _
n. y , J. jdOstPSON ACCIDENTS AND DESTItUCTIVE TEST 5 13 GT3 I 1000 - i i- i i e i 1, i e siI . . -- i l lii.l -
?
- _ i
- r. I t.. 500 ,
1
- . 5 :. 5 - -
n 5.5 13 t' si.-t ExcuR5:CN .
- 8CW 62 *C SUSCCCt.r3 13 13 _DESTRUCTFfE t., ..
TEST Q t
% M 10 0
=(oLEMENT .
-CORE -
4e 1.0 u speiT Q,/ gP -
,0 4g i -
19 cestnuc- 4 Ng 7.5 - j Ttvt TE3T p - es 50 .,,, r)
- eg*c -
a1 19.4 3 130 :ts g IE og 40 Jeg %o -
' W .
/ , g -
$ PERT 10 -cu 12/25 Cepg a h 0 -
U +0 = s tting !avesti-
-ENERGY TO a 74 ~
~
- PEAK . 80*C
/ C rw
- demon- SUSCC CLED AO k 4 # Ade #*
#<(
~
1a .ston was 5 a a 9 gP-6 shown chat the _ ~w A , to hava been a (c - in, holling, and - C en soms insight I g Last, but not - - l cag ment along C g l 1s possible now - 4- ,
=== ta ^t
, , , I , , , ,1 , , , I,,,,1 . , :
1 I 5 10 50 10 0 400 i SL-1 accident mal withdrawal INmAL REACTOR PERIOD ( msec ) e resceor. This ' transtant whoe s F1C. 3-30 Predicted and measwwe sucAser enerTy reseese vs. penas for BOLD-l. SPERT 8 and $t-1. n in Table 3 3 Total enerTy unless otaerwise spec:See, mrties are e0Lu aus crim reforeses I191 squarse iPEAN OU 12/23 dua trem 163al aan crianglas SPERT-4 aestrucnw test data frwe ($3]. l Le withdrawal of reactor critical a single slug. The water level in the tarJC was pressure region. The pressure wave front which
*n calculate that about 2.5 ft (76.2cm)belowthetopot the vessel and developed no doubt spread out in s11 directions.
{ a contt=uing the the slug therefore, had this distance to acquire striking the vessel side walls next to the core
. (th3 porttion in kinetic energy. This slug hit the bottom of the first and bulging them, then striking the bottom idwrecollapsed of the plate area in thocentral16 elements reached head and giving a net downward force on the t!! ave th"t power the vaportzation temperature and this caused vessel. and finally acceleratingu;rrards the entire amately 4 msee more steam production and violent dests :ction mssa of water above the core. It appears likely of this regier Abcut 209; of the entire core shows that the water moved upwards more or less as utionttrmmated 3 0.4 x 10*Ltw. melting proceeding to the clad surfaces. General apa"*mre Ls the Electric ee n, ares that the total nuclear transtant had just reached energy was 133 e1011w-soo and that no more than *Apparently no one has looked into this dowTt-an additional 33 5!w-seo of energy (best estimate ward force and one c an only contacture as to whether C60*C (3767'F).
Lt or 0.339 mm) 24 +10 5tw-sec) was releasedischemicaireactions this downward fors:e was sufficient to sever the tse ' aces had between the molten or vaportred metal and water. pipe connections to the tank. 2 is di!!! cult to Trt the start The formation of the steam vond terminated judge the resistance to such a shock provided r b excurston. 3~e the nuclear transtant, but it also created a high by the vessel supports. l Q a fogy a f St b S' . / 'YL n p + be d.4 i / 96 c,e i i
^
gw 3
$*# 9 c-
% 17 e-IS83g -11 Y
%,,[/j'%e,
!g 4 p
cv m s APPLICATION FOR A CLASS tog ggggggg FOR A RESEARCN REACTOR FACILITY Sased on Code of Federal Regulations, If t1e 10 Part 50 to U.S. Nuclear Regulatory Comission (* s
.R. R. O'Neill, Dean Scaool of Engineering and Aeolig sg. '
University of California Los Angeles Fecruary 1980
l l
.~;
TABLE !!!/6-1(a) COMPARISON TABLE - GENERAL gectu uervWtstTY 7 CAIPGuilA, 'JetVWtstTf 7 PU3t:34 WtvWts!Tf QF 49tIMm 1.QtATIGI 1.Qs AssE.2s GAtMsVt!J.2, PU3t!DA, M SEAfft.L
- S4tWTOn, M 1.cs M. GI.!PWWIIA M UCpm W. M M M S 45 S)-139
- ncer W. 33-1:12 asen se amens a w untvetsrff a cIJ.se y estasite arvststTf a matum CA17G383A opWIATE N anslLEAa WWWF t.AS3tA10tY tJ2.A l'WWIWWff OP *s2. Epa SW. E!SCE, nta2JAll Welsf JEPesmgef U. OP puuttDA C3.L5lE OP essDestina, U F d asAc3t (TWIDee3Js, ?Meut., Ueff CCEPf Ottstng AG. es Z.
TyM 08tlJ6 mfst c2LEL.1sTr2514710, IW't.E"ID, 3taf**tTE 08t!OS Waselin
- 3319 @ 37 GestAi 'Asa,2Aa get. (pit!PClPE GSetA '432.Dit 031P., QtEAc"3t sYT7Wd U3vETT, STitE2:3GIU1H & *R%
- ssLTJatr O atac"3t 3YsTWU 3.0. AA.*:3e, adlOttTElT M M STATE aflotITE 7 313stlUs&s N AdIOttTET 30afe & C3mtz, (3 JIM:tM) n o Quac en)
C3ETIILCTIGE J3Es Eumets C3ET. C3. 8 U3J!s COOPEt C3ETMCT!OI C3. Eutt"tm .Sff!3'? I 854EsN . 37 c. 3Lsee !?C.-sJ11.3tm A.I.-syg. anr A109tcs-*L%""Ut CD90 efts M ManTUI 02. (RS. SET) a.sens t. e 7:m> .cewsm;.- svt.C3t w-.pa-aan. a vers C3r!IELS, 5.5"IE3etcs 6-Carttua.s
- airstam h
- Steen posi mai. W CSED
- CSEE)
- CSET) wemnm ,
noot
- s Du'enn,oci .3evn, 3." low's E sne:ric azeu .
metAnm vantam.L 3-15 rouestr? was vantAsts 15 m 3 ne es ast vtAn vantAa.:
- soeu est vpa Ac?
- vAftse asias.Ysts, etAc 3t stntre me sam *Trae & w sut 4EAc"Ut opsina sintras, l =trctPai. c.Assaeos mr:as,
! uss y areiAn state, no sacanas esi. um se:etnsis. = sac Ni ansc a y w sn eni. & siteens creums, acT:vAnon uus.rsts t ussi :sn::re raczu:nai, wA ce's. aussAmoi. C~ cretAnns 3 set sein, I sein 4
- ant 7:m 2 act sein, I sein enrau.y 2 eaar 2 set sein, I sein sTarr suett peu7ais T:s wsuas sraras:
Dan rTRsf
- xmeat. M :/S WG1
- Art mA;.
N W&M VS, W 4 StT f.m., h , sz.:n sue.: PJE. 52987 * * ! A> swt of via (smi TrM.) 1 ss naa. *
- I cmstT:o U.4 wr.: u-s au.:-
- =;s.
3'fets!38
- 3.7 12.7 1.3C*
WAT) 35* 3 2' X .GT7
- ) cutztna *
- etsttai. 12 4.:.pitan
() *A3:1M * *
,.teness .35 us.
\ us we s *
- saasses.y seasu.s. uns sz. in Jarnet _
v s ce s.n errs met *
- Asses.Y 11 a u rs 3-M x 2-7/3 x .::;"3 :n. * *
- tes:oe ti sa-Asset.Y UntJ) soeuros A tw g mea.:s arm) x *
- au's 3 *
- O musaa- 4 n.eens :1 poe w i smets eft w s.a- wuecc :n * *aaaur =oe Asset.Y * *
. G8T"T :'u ?!nt E I
emw % e xu. ummi
=-m.:= 2x. e s -vi.oi 111 / 6 - 2
m TABLE Ill/6-1(b) PRIMARY COOLANT um y *
- c;lC1 Arts u ::asc tos y suas emes *
- 5J 8U38 nean er o cuo ru n, savrfr nrfusi *
- c> crui.
auze are :S = :C e - 3 m.~
- > S a vs w .
cas L3 Wsm 4 W sm: 2.'. W sac o aveuer ' in.sr m.
- 8) Gst at . g=hN
.c.
g= gY er erse w tst ?tas-se.L s.s. **; us mrmer *
- sssrmr:m war esswami :naa-ms: sEnemH1 w u. Tun stua.:
> ms:
l'CGI sEDesfrr iTtm) ese y mtmer erstset azertacus % sei *
'SD et/ min) strtcar:oi sms: miu cumit:ss atest w w wneen.:: sis us asfusi m
- u, rum sera.s nesseeerr, rtumsl:sw. stest meY
*=suu.a. mucus :::si cumit: mas NUCLEAR DATA u:.cim o ::u I .2 e
- 3.3 e u cent.
satse ui. u:acim 3.5 e lJ25 . 3,q , p arsocass c, u: acts 2.3::vc *
- cx a mw mu =x ex L5 : LW, et2 ,g.
3 1,; , L w e ,,3,e
'istex 1.5 x :.Ae,2 see es -a s.a ; % d se mmm.
~
N sum 1.0 x L4c4 sm 3.3 x :. N c 2 sac me :vtw
;:: arete: errs i.efst s .
O w. -3.4 x T't. ave (-3.We) -1.0 x f avv C C 2rh/ % H ).
-]3 4::vv.Or (mi s)
- ) ets -3"a v Y'::vv2em -3.21 avr4 y wts -vves wr:
1 * (-19 A < m) J uss -
- =*t:: errs P3 1.,Y .ux (18 =) * ,,,,
- ) :me 3.~: vven F scr.s 2.3::vs G.9)
- 3amaLE .
=2:sem we *
- erms.
szmer 5.5 m W w o5cEs-Beneleu % M (2 :: % :P a) til/6-3
I TABLE lil/6-1(c) REACTOR CHARACTERISTICS Sg; to i [4x',= - 2.s r ' = c,. E . ) p.,. I, l 9=' t ASJS Sgq, . M 31G.0SICM. sets.2 CDCETE CBICMft GN3lEM I mi.L 33tm Masam. A Muan. SAR A) St3St Pa'tA.:Jtt 1) MAMjaL i 43cP mo 4 U A)w.spum poet SmaaJus 3)d sort at taaet CJ cr2 m) sCRm o scET ICn < 2 sE
*TWL 85tt3) 4J33 SE:) 3J MISE G 1Artir 3) sept mE= 3J M et pt. ult a ses d QPW Puat U834t
- m) d 0 331E5 3A.75 tM LDE., 2 Puze 3Jf VALyt scIAM. ! MM. *EY OPWI AEht 1 mTT F) 2PP VAL.VE QP9e itJtel 3' q et*H P wtTH 4 3t MstE 3. ACE 5 t.P l
= ata,g ggtm wgTM eTgt 3sp .
l A) 17 34.Y 1 RJCE I.F = "uS C: f045 A) itJte SY CPW 3) Ll3e M4N ' 3eP E133 A) SY *taWEB GW 1) 1 s3m wres aar vefrttartas d. wasDes opOP ope t 84!P4WF CJ U38 C31E Lf4t. RIPupy aw sagt U3E5 3' 83Gta)d. ass t. DES SC W soet 3t vet mtrupy e 3rmix.uTue FAfe 3) t.cE5 84DuRY pu38 . MIPudtT GEufff PU383)d U38 CNEDJ PC 8U38 3EFE *t3 5'4 88EPt 4 El L38 sett.2 3fet unTUt I.DE. . mTut Lau, ej u3e sera.3 TAfst 7sg O ac *Tt4 t.me. seps Aaouf F) perEco an*us 5) stu,T:3t unTWt 1.bg. v wies SEDeaftY Z DOES (STiu. As>t PJi ft.Ait3J PAft utR Coteust)
#3 34su. 3EF = grAE.1 Lbt. SetT24
!/C. IM fM 3) 80 30 TAM. 6 TArpl.1 Asest.?:E3 l 3nP e anmzi i.aste CD e!T 31
- m ta!T5 A) STastT=JP SOLsE2 CRsfr A) SArt AS SE3m SAft 1) 0F 23 A) PWE't'If*10 5)a t t.sss ass s) se!cm i.s5: mist AaPvt 1 .
sc *s,1'@ . 3r . t.ur
- 3!F d MieSt CJ 3EP !M Str143E CT5/SE:
"Hase (35.De 1 0m SE)3 C))t.2M 0.353E ArruFist QPWIoteret Mies *AE, E!5gt 3r 2 W wr un cretATE MxE sseus5 3J cps tm istT e w 3ttvt E) 1.DGIC - C#eCT wi3+*
3uas 3t M3tE 3. ACE 5 StrtA.*AfmJELY , F1 SEllme PEtIG3 ur tvt SJ 'OE MJ utDE 4Apat 1* l STAffipl 0tA004. AtJetY5 IM GP5tATI38. i I DoetatT 17 CAL.3&11 Gt It!P *TTT i SoffD6 PCT IN CFWIATE MIE (M31 tr 3t!Nt) [
- Jufte A) wies P4 DIT *W9= A) $ = Ansta.E AufDE 1) 3' A2 3 N DIT SET C3e 1 astAttsu = arGe AaEA 3 AaEA M3et*ans - AvRt uw 3 sV M r!34 C*s8'3) r aJ ats 43DtJ He = AaJ3tm.3 Aufpt & 3 .W 4ACIATION ytEAS REAC"*ft 4A01Arit3s IIJTW Mise 34Y* 3 Mt(4Mt
- 94/H4 'crfh &
4A0!O* Mt. *be Otafse.5 3 AL33 = 8U334
- 4AEBIT .
8VAC.lATT34 AufD4 SIRWE. QS5 3' 3138 DLT
- d'.3 tit /641 1I/ >81 REAC". ;R 'OP AC* NE STUt.NIE 8AS4tT 8tXP Alft WEE 3M
- 41DIMMI/Htg sagt' (;347 gp ogafeg, sm2.E ST
- 3 MI/Ht M194 AfPIzt 41 8
STAct
- 1.3 w M:/M. IN AIR 852IVtPE 5 Creat, sutyt ANEA rest:3t d Afff scim 3t =33 3CP owest CJ 2 cas sse = Atcts.1 Aufe,efst :
M =c *tt
- M c3ffieL axpe, U3# 3J Auftt t.DB., STAct E'M ,9ff AELAJETAE.1 ACDC176 *D 804t Oif 3891ATIQM AufM A) U30 $= A) t,233 QB SOgt Qt Opgl CIMCJIT lM A) DuftY CDCyC*IVITf
- Oc 3J uzT 34.7 Ttvf?Y
- x s tvas:AT:3e 5tnei = ess e stEse u3# ,
- XTE "Ut IM wie4 M.;at
- w&TT3 1) PC 53 u Ett35E MystT3tB ll/M IMET 80:lt * *i"5 d :rt =tse1.4el3tT3 l CJ Affr lottalT AfC 3tft.17 4E5157 VITY Auf AusaE.:an des.rer Ct9U36 wt l "lllIe!Timet *Ittp SY 5tpes e RE3 C3C
! DJ B 8'WFGIT!GenL 03stfWt l SWISI:E3 Ms se 'DTDCED aff. Auftpuf C w =tr cru 3 CP5 8487 79t!33 !Peetst? U34T 1 ! 38t* DES 81 8AST amet :Poetatt G3# 19st?345 'CT IN CPWtAft P0413!T Ml'E 80t.) efUt 8'.38 i N3)Aufft C4 sa 1C 1p=> m s== Av.m = Scte i AT m AawE pogi Arzt .3 ss= c
- tuMUEDel t. G# SJ 055 3' W34T3
- I satt.;JIE5 !M 4A0 9e 5 ANEY 1AT"TYY POctsect CJ AiJ3tE,1 AuJr? ! 8C i
i U 34 IF Sect'IDt15 QPU C. eED AfC 3150 C3#5 M34!*3t GJT51 AC255 )
- xDes m *AC:uTV ,
l l 111 / 6 - 4 l l l
TABLE III/6-2 TRAINING REACTOR CHARACTERISTICS i ran 3ED LS fr=t e sime scus, w ow. asci 3 et L'D or cx i.re. (a 3 ed 1 x 10lI Vefsse L! x 2D lI 2 Vcm sse ocess ectivirr (:soi spec unit) 0.2: e at 3:*r 2.3: a a exm v cs:ss acrtyny ustman LE a a =xri e LE a a nom m c.am cas entren.nass 2 D m e- 3 uppecrtw mier ramen ugrus La x 2:r'ssc 2 x F "see wars warst win curricevr 422 84;etS 4233 af wu' a verstu u capric a r 4:8 x Y': :M .C x F 2: s U-3 mass carrtcorr 433 as C-3 rase .E ad C-Z:': sturro sanz 2 cats N 2e 6.5nc 42a muc me mapuru (LU svc:)
- ra=== go m in .
m e r. r=: =ca 0.na .ns Rm. Putts am. 25: emtoen, tHe. au.cr MS :.l:Actra 3, 4 .2 si U- 3 3,5 m U-3 V7t miosass 0.0 tm intet owea. 0. 37 tm auen.n m mates arto (wn.) 0.E mr =recemos 13.:4 wif 'Fr.
=xx.urr .
'd,4 sua: :D s - 2 cam weutuu, tx I!ar ;cre .
wetstuu, x: ITV ;q* ( cz.mus. x xus- e, stratrs v=z. zavt r rau. ruest 3sastrrt Iasaut:m rr artp. tus 0.3:"4 sa: (ca.=.um ' O.5 se (rcAswc) samm. :3 T sa: (mx:uo '0: sac axa om, sastre - I as L5' a - a.E Ias*LC:-af axe om, eaur:rs 1 un 0.C = 0.C a 1as :
*1rr4.
- 5.5 4 'OrN. % 5.3 4 grec :vrir emas mTE, nnx. 0. ': /sse .f5: 2/sse s41:!. (C3 C Eft)
SCEs, 3;:El: ( 87. O Im cast,se e s : t
- s!:Es, DIEJ *NJt SC ( 7T. o Im cast, 3tttwrr
- sI:IE, *1 EPM. 01MW SC 6 nr. 3 lm Cast, PeegrT*t
!*.w. - cast ;3g;pg.g g,;;c 3 a404 C3s 5 st. 20 :r:. esaT:7t si.:rrs
- m.:: F := sretc *ancitx .
DC: 3 8T. a :N ameET! E EXXs DrtPI.1r."X. 'Jc:'#ts M 012.P% ccP!:l:3rr!4. 5 77. x 5 sv. x a 't.11 :m osa C tm x E :n x 43 tm :.:ris au svaatt em. x;rs:. er:ca. >=cvtsrcr4 sen :m:nn::: . Drus Est ust 5 77. x 5 ar. x ir4 er. ! :m = orett.snrn. rces 2 - er=stro C :m :tata .
- 5 - car:rers a :m :turtr a - .cr:=> :s a :x. :tuc==
l 3 ert:u "s .n. :turtm 1 - cr:c4 ;-74 :m :tuc-ceuxuerra, ecus. er *
=s 3 seevam.: aneurn sanaan
=ct:. :. s '.; - or::r 4 'M :n. x It' :n.
2 - cr::x., In :::. x : :n. 111 / 6 - 5
8.0 CESIGN BASIS, ACCIDENTS, AND CONSECUENCES Calculations pertinent to CBA's and Consequences were presented in the 1960 " Hazards Analysis" (Reference 1) that attended the original license application. These are reproduced in the original , form and attached herewith as Attachment A and S. Atta'chment A (titled Appendix 3 in " Hazards Analysis") treats with step changes of reactivity (ak/k .10.006) and limits the maximum licensed excess reactivity to 0.006 to avoid prompt criticality. Amendment 7, approved in 1966, increased the maximum allowable excess reactivity to 0.023 with the restriction that no single exposure cavity would contain an excess reactivity of 0.C06. The excess reactivity limit of 0.023 was justified on the basis of SPERT and
'SORAX experimental results (see Attachment A). The provisions of Amendment 7 remain in effect today.
Attachment 3 (titled Appendix C in " Hazards Analysis") treats with Radiation Doses Resulting frem Release of Fission Products into (the) Atmosphere. The release is not causally related to a - scocific accident, and from the SPERT and 30RAX experiments, one can only state that the calculations attemot to suggest a scale of events that might follow a catastroche of unknown cause. The calculation of fission product inventory is based upon a steady state equilibrium inventory at 10 kwt, and certain assumotions concerning leak rate from the building. (' The consequential dose calculations were apparently unreviewed in the acoroval of Anendment 3 (1963) that increased the maximum licensed power level to 100 kwt. They were reviewed by the Division of Licusing and Regulation in processing the application for Anendment 7 (referred to above). In view of the current restriction of the UCLA Reactor ocerating hours to 5% of the year, the maximum average power is now 5 kwt, a factor of two less than the 10 kwt used in the original calculations. Because the basis of the earlier calculations are not exceeded in l the present acolication, those representations and actions are herewith incorporated in the Technical Specifications. III/ 3-1
APPENDIX I!! ARGONAUT SAFETY ANALYSIS REPORT (ASAR) Attac.Nnent A Esti ation of Effects of Assumed lame Reactivity Additions from UC1A TRAI{I:1G REACTOR HAZARDS AMLYSIS, Final Report, R. D. MacLain UC:.A Depart:nent of Engineering Report #60-18, UCLA-NEL 2 March,1960 (there titled Accendix 3) III/A
w l t ESTIBATION OF EFFECTS OF ASSUMED LARGE REACTIVITY ADDITIONS It has been demonstrated repeatedly in the Boram and SPERT resetors that =eter-cooled, water-coderated reactors of suitable design may have a very suestantial self-protestion against the effects of reactivity accidents, even in the absence of corrective . action by the reactor control system. This self. protection is provided by the negative - steam-void coefficient of reactivity and the negative temperature coefficient of reactivity, both of which can result in important reactivity reductions as the reactor power rises. The UCTR has been desiped with a high degree of self-protection of this type. In this appendix estimates are made of the behavior of the reactor under various hypothetical con = dicions of excess reactivity addition with no corrective action by the control system. The characteristics of the UCTR which determine its behavior daring power transients resulting frae large reactivity additions are quite similar to, but not identical with, those of the Boraz I reactor. Its behavior can be predicted most r4liaoly by utilizing the Boraz I data esth simple correction factors to convert them to the UCTR conditions. The significant quantitative characteristics of the U CTR and the !brax I reactor are compared in Table B-1. TAliLE 0-1 - i
\ C041 PARI 50N OF UCTR AND BORAX ! CHARACTERISTICS l CitAAACTERISTic UCTR SORAI I Fuel slate ' seat' 13.4 w/o U-Al alloy 18 w/o G- Al alloy 90% enriched fully-enriened Fuel plate claading ~ lt s aluminue 1300 aluminue
' W at* thickr.ess 0. 0% inca 0.J20 inch Cladding thickness Q.015 inen 0.020 inch Coolant-enannel th i czness 0.137 inca 0.187 inen Core volute (aooros.) 71 liters 106 liters voie coefficient of =0.18% k/f. cool ant =0.2% h/% coolant reactivity (calculatee) void voia Tesserature coefficient of -0.009% k/*: (estimatee) -0.04% a/*
reactivity (room temaersture) Effective aremst-neutron 1.4 : 10 " sec Q.55 x 10 " sec lifettse (calculatec) Power ratto in core. 1.53 1.12 santeve avarage III/A-1
In addition to the quantitative differences, the UCTR differs from Sorax I in that s s the saximes coolant water level is only a fee inches above the upper ends of the fuel places (instead of about 4 ft) and the coolant water, once it has been ejected forcibly from the core by a power emeursion, cannot fall or flo= back into the core. Effect of 0.8% Exceen nesetivt ty An excess reactivity of 0.6% kett will be available in the reactor if its temperature is abnormally low (nearly freezing). The addition of all this excess reactivity will cause the reactor to operate at a power such that the reactivity losses associated with the tagerature increase and the voids formed til equal the initial excess reactivity. If the reactivity is added slo =ly, after the reactor is critical, the power will aoproach such an equilibrium level slowly as the reactivity is added. If the reactivity is added suddenly, den the reactor is initially saberitical or at very low po=or, the power will at first rise exponentially with a period not shorter than 0.8 sec which is the asyniptotic period corresoonding to the full emessa reactivity of 0.6% k,tt. Waay experiments eich the Borax reactors have demonstrated that for periods of this order of magnitude, the transiston from the exponential power nse to the equilibnum power level (in sich excess reactivity
- i. balanced by tescerature and steem void coefficients) is a smooth one involving little os no power oversnoot. & the basis of this empenence, it can be said that the sagtitude of the po-er excursion nich results from the 0.6% reactivity addition stil not depend greatly on = nether the reactivity is added suddenly or relatrvely slowly and in neither case will it approach a level .hich would cause a fusi place to burn out.
In order to compute the pe=er level at which the reactor will ooerste after the addi-tse of the 0.6% excess reactivity discussed in the fr.,regoing, it is necessary to kiew de water.comperature coefficient of reactivsty. The relative iscertance of the two moders-tors, graphite and wecer, in determung the effective neutron tesoerature introduce un. certainties in the theoretical camoutation of this coefficient. The coefficient cannot, however, have an absolute segnitude less than that of the seter-denstry coefficient of reactivity referred to a tammerature scale, i.e., the coefficient computtd on the assumotion that: d k.tr 3 3 k.tr . - 34 dT 30 ST
.here a refers to the =ecer density and I to the tescerature. G de assumotion that this statmum value is the true value, a rise of =eter tassersture fran sear 0*C to 20*C . auld I reuuce reac tivity by 0.6% k.gf.
l Ihe capacity of the reactor-coolant system is such that if the outside air temperature
. re 0*C and the aversee .eter tencerature in the reactor .ere 30*C. energy would be rer.sved at the rste of 365,000 BTU /hr or 107 k=. Under these conditions the reactor eter-inlet teamersture would be 60*C and the exit tenoersture, coincidentally, .ould be 100*C. It is, therefore, concluded that if the full availaole excess reactivity of 0.6% k.tr .ere added to the reactor on a cold day eith the coolant systen operating, the reactor .ould ocerate at an ecutitorium po er level acout ten times higher (100 k=) enan ses normal maximum etd little or no net steam procaction. En fore rescatng the ecutitbrium po.er, enen the eater in the coolant systen =ould be heated to tae ecutithnum value, the reactor ould coerste at a smewnat higher power level ano some net steam procuction 1stgnt oc cu r. If tne coolant
.ere not flo tnw curing tiae time of excess reactivity amiitton, the .s:utlibrium 'pomer level
*!!/A-2 l
1
1 s would be quite low and equal to the heat losses. In no case sould the power level approach a hi 6h enou6h value to justify any fear,of fuel plate burnout. Easteve Tolerable Sudden Reactivity Addition In order to assess the safety factor which exists between the normal excus reactivity available in the reactor and the excess reactivity necessary for a serious power excursion, it is useful to estimate the value of excess reactivity which, if suddenly inserted and not removed by the control system, would raise the -n==' temperature in the bottat fuel plate to the meltins point. Sach an excursion would hose the reactor core but would not result in any mahacancial release of fission products.
. The first step in the procedure is the estimation of the .6 nential period corruponding to the excus reactivity .nich would have characterized a po or excursion of similar effect in Borax I. The estimate requires that (1) a relationship be established between the saxi-aus temoerature of the fuel place and the energy release of the excursion and (2) the energy release be related to the period of the excursion.
For the case of pe=er excursions of short period, with ructor =eter at saturation tem-perature, it is shown in Reference 1 that the max 2same fuel-plate temperature rise is, to eithin experimental error, proportional to the sazionas energy release of the power excursion. The proportionality was deter.ained to be constant 24.4*F per W-sec.* Meamarm eses of the same type with cold reactor ester (the casa directly applicable to the UCTR) showed a simi-lar relationship but with a proportionality' constant of only about 10*F per MW- see (Bef-
,- erence 3). The differece is not an unressonable one since the succooled vecer represeau
( a more effective heat sinir than the ' saturated water. Ho ever, the esperiments with the saturated water were carried to short periods in the range of interest ehereas the subcooled experiments were limited to longer periods. Therefore, more conservative saturated water
- data will be used.** To raise the enxamus temperature of the fuel place from the tempers-ture of boiling water to the selting point of aluminum, a temperature change of acprcximately 1000*F 1000*F. eauld require a pe==r excursion eith a total energy release of or 41 M % -sec. 24= 4'I/M - see According to the data of Reference 3, replatted in Figure D-la "subcooled" power excur-sian of reciprocal period 150 see 1.ov1d give an energy release of 41 MW-see in addicisa to the energy necessary to raise the fuel plate canoerature to the securation temperature of e s t e r. It is, therefore, concluded taat a power excursion of pened at least as short as 1/150 see (6.7 millisee) could have been tolerated by Borax I eith subcooled eeter without selting at the hattest point in the fuel plates.
Loenments of the Ebrax and SPERT types have not been ude .sch reactors having endely different neutron lifetimes. The gueral evioence of the excenment.s. however. sup-I parts the suoposition that of the three related variables--neutron lifetime, exc esa reactivity, and exponent.tal period-enich characterize the neutron physses of a power ex-curston, it is the exponent 2al period =hica determines the total energy release and sne tesoeratures actarned. The excess reactivity and neutron lifetime have large effects only as they jointly determine the penod. This mapposition is constacent, for exarole. ith
' Actually, tae emer p data of Referees I eere revised in Reference 2 becaese of later and better calthrottees of the t a s t rumen t a t t e e . The aunnera eneve are taaen from tse Later (sere pesassistic) data.
*'If succeeled data re used. tae esse directly soolisente te UC"B. tasa snelvsas eewid teetcase taet stee reactivity edeittoaa 2. 4 tames as large as tasse stacuased here eeule not samag, tne reedter.
- /A-3
that ek total energy transferred to th coolant =eter during a pe=er excursion is many times th enount which = auld reporize enougn water to compensate for the excess reactivity, and that the actual reactivity reduction which occurs during the excursion is ansch larger than the inicial excess remetivity, ne extension of ek Borax results to the UCTR is made as the basis of this evidence. It is convenient, firs t, to treet only ek effecu of the slightly greater fuel-plate specing and the slightly lower void costficient of reactivity of the UCTR relative to th Bors= I. Informatice will also be drawn frem th Borax H experiments. N Borax II re-actor differed from brax 1 in that the coolant-enannel thickness =es srester in the ratio 0.264 in*
- 2.25 and that the calculated void coefficient of reactivity was lo=er in the 0.117 in.
ratio. 0.105 k.tr/% void 1 0.245 kogr/% void
= - 2.4= 0.416 Both of these differences would be expected to cause a higher energy release per fuel place in brax II than in Borax I for a power excursion of given period- h sessuremenu made eith subcooled water at periods down to 23 millisec showed that the energy release per fuel plate in Borax II was between 1.7 and 0.0 timas tnas of Borax I, with the seniler ratio applying to the shortar periods (Beference 2). Therefore, it seems quite conservative to assues, in the case of any two reactors (1) and (2), of the Borax type having a ratio of fuel place spacings, sus 2 , and a ratio of void coefficients of reactivity, CvC2, that the ratio of asergy release per fuel plate for a succooled po er excursion of given period f will be no greater than EvEt 3 SvSt or Evi t: Ct/C2 whichever is the larger. For the
. UCTR'and Borax I the ratios ar* ,
Ser . 0.137 . ,, C 3 , 0.24 _ 53 0.117 Cur 0.13 It is concluded, therefore, tlat a Borax reuctor having a coolant-channel thickness and a void coefficient of reactivity ennal to these of the UCIR would release not more than 1.33 times as mich energy per fuel place as Borax I. The limiting nonmelting period fe; such a ' reactor would be that hich in Borax I gave an energy release of 41/1.33 = 31MW- sec. W period obtained from Figure B-1, corresponding to a total energy release of 31W-sec, is 3.3 millisec. In campering tk behavior of different fuel places, it stat be recognized that the total energy release of de power excursion can no loneer be considered as a definitive variable scause a large fraction of the total energy released is stored in the fuel place during tne important stace of the reactor shut-do-n. For exasole, a reactor composed of fuel placas os high heat capacity undountedly will experience a larger total energy re-lease, but not necessarily a hasher mextamas temperature, during a power excurston of given l period, than a reactor having plates of low heet capacity. l l From examination of the brax results, it seems clear that two distinct phases of the l reactor shut-do n process occur consecutivelv and that both may be important in deter:ntatag ene meninum center temoerature of a fuel place. Le first pnase covers the interval before l an smoortant amoun t o f bo tling occura a t the fuel-plate surface. During this tacerval, tne
- be s t loss to tne ater is sa ll and the
- sortant canaiceration is evtdently the ratio of fuel. place surface tencerature (which determines the start of botlina) to center temoera-l ture. For periods in ene range unoer conalderauen, this tencerature ratto is theoretically not f ar from unity (0.76 statmun for a 10-mt11iuc persod in Bor2x I). E.werimen tally, the
. !!I/.3-4
l i l tempencure ratio was unity for periods down to 5 millisec in the Borax I messunments. Since the total effect is seell and since the temperature ratio for horax and UCTR fuel l plates should not be mach different, the thinner cladding =ill teral to balance the effect
- of the peerer " ment' caductivity. It is concluded, therefore, that there =ill be no in-I portant difference in fuel place performance during this initial phase of the excursion.
De second kase of the power emeursion begins when a significant. rate of boiling is established at the plate surface. 19eactivity and consequently generation are reduced at a rate which saast be a function of the race at which heat can be transferred into the boiling water. Ac & same time, the transfer of heat into the ester removes heat from the fuel ( plate and limits its temperature rise. De important characteristic of the plate during ! this phase of the excursion is the heat flux which it can supply to the seter for a given ! temperature diffenace between the place center and surface. A figun assumed to be roughly i indicative of the relatin performance or merit of fuel places during this phase is the ratio of heat flux to tem.perature difference under steady-state conditions. This ratio (figure of surit) will owremonesize the difference bet = en fuel plates since the tem-perature distrilaation in the plate will be more peaked during a steady-state conduction than during conductimi = hen the general temperature lew! is rising, ne ratio of these figures of merit for borna I and for the UCTR is
' Heat Flux' di e. . <
UCTR
. 0 . 3,,
'H ea t Fl ux '
A T .- . < g, A conservatin procedure sould be to apoly the above factor to the permissible total energy of excursion on the Nrax I curw. At h same time, ho=ever, the difference in gross maxisman to average power ratio for the two reactors should, he taken into account since it is the temperature of the hatter,t point in the hattest fuel place which is being considered. De pe=er ratio for the two reactors is Max A" Borax - 1.32
- 1..,.,
Wax 1.63 l Av gg ( De comoination of these two factors reduces the permissible equivalent energy of the borax. type excursion to -
- 31 x 0.82 x 1.12 = 23.4 un -see
( Le corresponding exponential peried from Fiaure 15-1 is 9.1 s.Allisec. It is, there fere , escluded that tsie UCTH ill tolernte a peer excurston at period at least as snart as 3.1 millisec witacut the meltine of any part of any fuel plate. L e excess reactivity corres-peding to tais period is 2.3% k.f t. successive Power Etcursiong It is typical of the borax and SPERT reactors, tziless the excess reactivtty is renoved by external means, that an inttsal power excurston nich terminates itself oy excelline ester from the reactor core .t11 be followed by sucsequent excurstons as the eter falls and flows cack into : w core. An exception to this benavtor occurs enen the inttial III/A ~
escursion is violent enough to cause a permanent loss of reactivity by throwing a larp amouac of eter completely out of the reactor tank. In the UCTR the total quantity of
.eter.in the core is small, the ashmerpace of the core is small, and baffles above tne
~
core are so arranpd that any ester splash is directed to the outside so that it cannot . return to the core. Consequently, even a relatively mild power excursion (e.g., one hav-ing an uponential period of from 20 to 30 millisec) in the UCTR should result in perma-nent self-inseed shutdown of the reactor. By these same design features, the possibility of larp successive power excursions, such as chose studied in the SPERT project, resul t - ing from the ramp addition of excess reactivity is eliminated. It can be anticipated that the UCTR will be safe against quite larp ramp additions (larpe than 2.2% ketf) provided only that the rsep rate is not. so rapid as to add an excess reactivity of more t.han 2.25 k,gt before the reactor po.ee reaches a high level. To exceed this limit the resp race
=ould need r.o be of the order of 1.0% koff per second or larger.
B en s Tu be Reactivity Effects
% UCTR has two 6-inch diameter been cuces which extend to within 11 inches of the fuel-graphats interfaces. b ==u=.- change on the core reactivity mich enn be effected by these two beam-tube facilities es calculated t.o be 0.185 AKAC or 0.09% aK/K per beam tune. He calculation is based uoon .che effect of a black absorter six inches in diameter placed in the same positiert as the been tubes. De reouction in the reflector savings te to the black nosorber === calculated using the following equation.
S (reflector savings) s U ""M . L (reflector). tan h Tf reflector thickness) [ D(reflector) L(reflector) Beierence: Elements of Nuclaar Reactor Theory, Glasstone and Edlund. De reflector savtags for the 49.5 cm and 28.0 cm of graphite .ere calculated to be 7.33 cm and 5.25 en respectively. Le area of the black absorber is 13% of the adj acent core l face area. Using the reflector savings given above and the area etgnting factors, the reflector savsngs esth and withour, the six-inch diameter black absorcer were calculated to l be 7.33 en and 7.33 cm respectively. 1 The reactivity effect of the single six-inch diameter black absorber .as then deter-mined esiculating the critical buckling .tch and without the black absorber. i Ustng the value of .0.09". AK/K , for a single been tube the shortest period .hich the reactor could go on, &e to the sudden nichdrawal of a black absorcer from one six-tneh been tuhe, .ould be soproximately 30 seconds. Derefore, the reactivity chanp .htch can be effected by tne bens. tuces does not represent a hazard to reactor coeration. In addition to the two 6-inch beam tubes =nich penetrate the outer reflector, there are four 4-inch bears tubes hich terminate outside of the reflector. No calculations eere made for the 4-sneh tunes since their effect on reactivscy .tll be snacn smaller than enat j of the 6-inch euces.
- II/A-6
REFERENCES
- 1. Dietrich, J. R. and D. C. Layman, Trans t an t and Steady State Charseteris-tses of a Boiling Resetor. The Boras Esperiments, 195J, AECD-3840, Argonne National Laboratory February, 1954.
- 2. Dietrich, J. R., Espersmental Determinations of tne Self-Regulation and Safe ty of Operating sater-#oders:ed Reac tors, A/ Conf. 8/P/481, Interna-tional Conference on the Peaceful Uses of Atomic Energy, June 30, 1955.
- 3. Dietrich, J. R., Esperimental Invest.gannon of the Self-Limitatson of Power During Reee tivity Trans sen ts sn a Subcooled. Mater-Moderated i
Reactor. doraz-! Emperiments, 1954, AECD-3668, Argonne National Labora-tory, August 17, 1955.
- 4. Lennox, D. H. and C. N. Kolber, Summary Report on the Hasards of the Argonau t Reac tor, ANL-5647, December 1956.
e ( l l l III/A-7
APPENDIX III ARGCNAUT SAFETY ANALYSIS REPORT (ASAR) Attachment 3 Radiation Ooses Resultine From Release of Fission Products into Atmosehere fr#A UCLA TRAINING. REACTOR HAZARDS ANALYSIS, Final Report. R.D. MacLain, UCLA Cecartment of Engineering Report #60-18, UCLA-NEL 2 ( March,1960 (there titled Appendix C) . III/S
t l i f I l l l BADIATION DOSES RESULTING FROU l l RELEASE OF FISSION PRODUCTS INTG ATN05P!!EEE Estimates have been ande of the radiation dosages which would h received by persons outside the reactor building should there be a release of reactor fission products into the I reactor building and leakage of the building air to th outside. De radiation exposures considered here are those =nich would result fross the passage of the air-borne cloud of { radioactive contaminanu over the growd- Dese include th external beta and gamma radiacion exposures and the internal exposure of critical body organs resulting from inha- l lation of th air-borne contaminanu. He most important, of the inuenal exposures are the ; iodine dose to the thyroid and the strontium dose to the bones. l he radiation exposure received by a person standing at a given distance from the : reactor building obviously depends on such f actors as (a) curies of fission products l stored within th core at the time of release. (b) fract2on of the core fission prodacts escaping into the building air, (c) building out-leskage race, and (d) acaospheric dis- j persive properties. Hence, in the analysis, certain basic assumptions are required as to I the circumstances surrounding the release of the fission pr'odsets, as to atmospheric con-dicians and, as to the tightness of the building at the time of release. Le results ob-trined here are based on assumptions shief., except for the arbitrary one that a release has occurred, are considered reasonable for the reactor and building design. De calcula-tida method is descrihd and illustrated in sufficient detail that additional calaalacions based on other assumptions can.be made if desired. He macenal presented here is divided into three sections. De first section de-scribes the model assumed for the release and spread of radioactivity and gives the nec- { ensary refarences and formlas used in calculating the endiation doses. Le second section l illustruces the calculation procecure. Le third section presents the results ootained l for the radiation exposure hazards with the assumed model. ilethod and 4ssuuntions used tu Oose calculations _ Although such an event is not considered even plaustble because of the li:sitations on available excess reactivity and because of the inherent self-limiting characteristics of the reactor, it as postulated that an acc2 dant has occurred in *hich the reactor poner level has risen to the extent that local melting of the fuel plates has occurred. Le reactor is assumed to have been ooersted continuously at the 10 kw po=er level long enouch to have attained equiliLrste concentrations of the relatively short-lived fission products, i . e. , the iodine, branine, and krypton isotopes. He incident is assumed to result in the transfer of 10i of the volactie fission products frem the reactor fuel plates to the build-tag air. It is assumed further that none of the nonvolactie fission procucts are trans.
, ferred to the building air althmsh they say be released to the reactor coolant eater and retained within the reactor building.
Le ,foreecine set of circumstances is consistent with the reasonable assumption made here that the incident is not violent enougn to blo= off the too and side bioloescal shields so as to cause an Intense sorsy of water steam-rsdioactivity sixture into the Duildine air. He release of 10T of the volatile fission pro.2ucts is procacly too hian for tne assumed
~
incident but is ned to give an upper limit, to the radiation exposure involved. H e vola. tale fission products are oramane, krypton, todine, and xenon. Hence, tne fissten product chatns anten sust be consscered are of at. ante masses 32 to 90 and 131 to 135. Beierence 1 presents tar,les for each chain giving the eoutlihrtum activity of eacn si the fissacn procuct. IM/3-1
]
isotopes in the chain and the decay (or buildup) iallowing reactor shut-down. The informa-uas presented in hierence 1 is used here for the fission prodact activity release into the building air following the asmuend incident. b most likely' pois.c at which radioactive contmaination of'the room air sould be . detected is in the reactor room eshaust & set, since the air is pulled from the reactor room, i and exhausted through the fan room atop the building. L air would not, be considered con-l taminated unt.il the activity exceeds that associated with the A'l normally being discharged. Upon detection of radioactivity the air conditioning unit will be shut off, and the dampers in the inlet and our.let ventilation will be closed. (h major avenues of leakage of the volatile fission prodacts and daughter nongaseous products) frain the reactor room are the two hign hay room entrance doors from the control room and the three emergency esics in the reactor room. Access to the control room is by way of an electrically controlled door from the reception area in Engineering thit III. All access doors will be seether-stripped and emergency doors leading directly to the outside, caulked and sealed for minismum leakage. bse doors will be closed at all times & ring reactor operation and any breeching will be indicated in the control room, on an audiovisual alarm systen. To occain an upper bound for the radiation doses, the outleakage, L., in curies per hour of fission pro <aaet activity, is obtained from Lt = b Y Ci .... (1) s 2 and is assumed castant &rtng the esposure time. In Equation (1), , the latilding leakage rate, has been taken as 2tM of the resuor rous volume, B. per hour for a 30 MPH ( eind. Dis leak-rate value is assumed to be directly proportional to etnd velocity. b
- quantity Ct,... is the m=e= activity in curies of isotope ,i. present outside the fuel platas followsng the assumed relense of 10% of the volatile fission products. For suist of the isotopes in the volatile fiazion product chains, iC ,.., is the activity of the isotope-ar. the time of release from the fuel plates. Le important excepuons are St-a9 and Sr-90 and are formed outside the reactor and reach a nextaman scetvity outside the reactor at some time after the fission product release. De tables in Reference 1 perrite easy calcu-14uen of the activines of Sr-49 and Sr-90 attributable only to the decay of the isolated parent prodacts.
Le concentration of fission product activity in the atmosphere outside the reactor buildina and the resul cant radiation exposure -211 depend on the = tad di,ction and velocity and the degree of atmosoterte turbulence. Le highest done race is cot .aned enen the person exposed is directly downwind from the leak. L camoutation method tm based on O. G. Sutton's formula and utilizes equations and curves gtvan in klerence 3. For calculation of the external beta dome and inhalanon doses from the racioactive iodines and strontiums, the concentration o'( acttvity inn the atmosoberic atr as calculated by the formul. tr. Referer.ce 3, Pe 153. For irround-level conunuous entssion of racio-activity, thas formula reuuces to
*f X - (2) idC0 rC u x -= *nere, 1 : eencent rat ton of settvity, cu rt es per cunic meter at at t L = continuous source stren6th, t.e., tutlatng out-leaxage race in curtes per hr
- !/3-2
i ( x = distance downwind frtm source, meters u = mean wind speed, meters per second C= generalised di ffusion coefficient, seters n/2 a a dimensicaless parameter associated with atmospheric stabil~ity The following representative values of the diffusion parameters for two different at. mospheric conditions are used to calculate the concent. ration of activity, X, for a speci. fied leak rate and atmospieric condition in the outside air for various distances, x, from the leakage source. Atmospheric Candition m C2 u severe inversion 0.5 .004 i elld lasse 0.25 .024
- 3
% external beta dome race during passage of the cloud of radioactive fission products is obtained from the following equation given in Beforence 3, page 100.
I" Dae
- ( 0. 5) (0.64) X,:
e (3) 6.8 x 10 to vev qa where Xg is the' concentration of 4. energy in Mev per see per cubic meter of air and Dg is , the external beta dose rate in roentsens por sec. The relation between Ig and X of Equation (2) is I a 3.7 x 1010 IE - (4) where E is the effecuve beta energy in Mew per disintegration. The activity A deposited per second in clie critical arrans is given by A = J F, I (5) where A = activity deposited in organ, sallicure/sec J z inhalation rate,17/60 liters per see F. = inhalei fraction of activity retained in critical cryan I l The correspondir4 innial internal dose race for a person standing in the fisstas procuct stream is given tv the expression l D = A 2600 t #2 E (6)
.nere, I
U = instzal internal aose rate rep / day t : t2me o f exnosure, hr T = .etant of cratscal organ, ks t 1
- /3-3 1
- The total integrated dose to the critical organ is related to the initial internal dose rate by the equation TID = D 1.44 T (7) dere, TID = total integrated dose, reps T = effective half-life of the radioisotope, days. .
The values of Fe, E, W, and T appearing in Equations (5), (6), and (7) may be obtained from klerence 4 for the various radioisotopes and critical organs involved. k ierence 5 gsves additional information on the various iodine isotopes. For calculation of the external gamma dose race, the J. Z. Holland monogram given as Figure 3.3, hierence 3 is used. The ,zw gives the gasse dosage resulting from sudden discharge into the atmosphere of the contants of a nuclear reactor which has been operating at a steady power level. The dosage read fram the namogram first met be corrected to account for the fact that, none of the nonvolatile and only a fraction of the volatile fission products are assumed to escape from the reactor for the case being considered. Also, since the activity is not temwdiately released into the atmosphere, but leaks out of the builditor at a finite race, the dosase ootained by use of the nemogram mat be converted to dose race. The corrections (actually scaling) applied to the values obtained frae the nommeram were calculated as follows: , . fAe gamma settes ty of the volatile fission products sasumed to be esessing f*on the reactor was in ternine d by use of the curves given in Reference S for 20 msn time sf ter sAmt-dows. E ustson (8.3) of Deference 3 is used to esicu-1ste the game actseity of s!! tk fissson products for the some time sf ter sAut= dcun and the same *esctor power level (10 kW . Fe ce this the frsetion of the total fisssos produc t gamme setsusty streteuted to the assumed 105 escape of the volatile fission producta is de teresnad sad the gsens fase rec.4 from the noso= gros ss scaled tows 4y tass frsetton. fo catssa t?.e dose este ressiting fees the finite este of esdioactiesty !ao4 into the steossbre. !As gasse dose scaled from SAs nosogros eultspised by the (men t i t y f / Tg . Illustrative calcolations Problee 1: Cale:alace the I-131 dose to the thyroid of a person standing at a distance of 61 meters downwind of the lesh for eigt.t hours. Curtne the exposure a severe inversion condition exists in the atmosphere. solution: From hierence 1, ti.e ecuiltbrium curses of I-131 in the reactor fuel places following 10 kw operation is 9.30* = 1018 x .003610
= ,C curses 1.26 x 10" For a 30 MPH .tnd. 3/V, = 0.00. For a severe inversten u = 1 meter /sec l ---e = .015. For ICT reiesse of (2.25 MP5t) so that BW, = 0.20 x 30 1 131 from the reactor, L s .315 x (247 x .10) = 0.37 curten.hr MI/B-!
From Equation (2) X = 2 x 0.37 , t,7 x 10-3 curi'* 3600w x .008 x 1 x 61'
- meter d b thyroid dou from I-131 is calculated from Equations (5), (6), and (7)
A s- 17/60 x .15 x 1.7 x 10 5 = 7.2 x 10*I millicurie/see D = 7.2 x 10-7 x 3600 x 8 x 62 x .22 = 14 rep / day
.020 TID = 14 x 1.44 x 7.7 = 155 rep The values of F,, E, 4, and T used in the calculations were obtained from hierence 4.
Problee 2: Calculate the external Adame from the I-131' isotope for a person standing at a distance of 61 meters doenwind of the leak for eight hours. solotton: For X = 1.7 x 10-5 curies /m2 obtained in Problem 1, 10 If = 3.7 x 10 x 1.7 x 10'5 x 0.22 = 1.4 x 105 y,,f,,, 3 3, From F.quation (3) Dp = 0.5 x 0.64 x 1.4 x 103 , g , g7 g 6.3 x 1010 For eight. hour exposure, TID = 6.6 x 10-7 x 3600 x 3 = .019 r = 19 nr attriktad to only th I-131 isotope. To obtain th total external Adose, the seen procedare must be followed for all of the fission produets amamed to be escaping from the reactor. For the conditions of this proolem, the air concentration of all of the fission products assumed to be escaping fran the reactor is 1.2 x 107 Mov/see 33, in which came the total' external heta dose for eight-hour exposure is 1.6. Because th decay of the fission products in th Wilding and en-route to the person cutside the building was neglected, the dose value calculated is higner than the actual value which .ould be ootained for th assumed conditions. Probles t Calculate the total gamma done to a person stancans il meters do.Ewind of the leek for eigns hours. solutton: Cirect use of the aceogram (kference 3) gives 12r for the total germa done caused by sudden release of the total contents of the core into the at:nospnere. From Nierence 6, ene gamme actinty of 10': of the vol ctie fission products is 6.9 x IOC Wy/sec at 20 nnutes af ter shat-down. For all of the fission procuets at 20 minutes after shut-do.n, Ecuation (3.5) of hierence 3 rives a total gama activity of 5.2 x 10" Weisec. Hence, on the aversee, 10': of uie volatile procucts given a pma accavsty ecual to .01324 of sne activity of all the fis - ston prooucta. Thus, ce dose oue to sucden release of IC* of sne I!I/E-5 1
I
~
' volatile products into the atmosphere is 12 x .01325 = .16 e : 160 er Inasmuch as the cone-t==cs leak out from the building at a finite rate (3/V, = .015 hr- 1), i the genus dose rate is simply 160 x .015 = 2.4 er/hr and for an eight. hour exposure, the accumnalated gemas dose is 2.4 x 8 = 19.2 at Results of t' adle tion Escoeure calculations b results for eight-hour *=pa=re at four different distances douewtad of the point of release under two different acnospheric conditions, calculated as illustrated above, are talma-laced belos. In all cases,105 of only the volatile fission products are assumed to be re-leased frw the r9 actor fuel plates. L building leakage rates are B/V, = . 015 hr- 1 f.or the severe-inversion condition and S/V, = .045 hr*1 for the mild-lapse condition (caused by difference in vind speeds assumed to be prevailing for the different acaospheric conditions).
TOTAL INTEGRATED DOSE (res) FR05 AN EIGET ROUR EXPOSURE 4 7 YARIGUS DISTANCES DOWMRIND FRON REACTOR BUILDING LEA,K ( .. Severe Inverstos a, asters ga te rna l Gassa Thyroid lone Beta Dose Oo se Dese
- Dose 15 61 14.0 .080 1800 .006 152 1.s .019 220 .0007 305 0.4 .010 59 .0002 0.I5 .006 20 -
Mild Lapse 15 2.2 .040 290 .001 63 0.19 .007 25 .0001 152 0.04 .004 6 --- 305 0.012 .002 2 --
*:!/C-6
REFERENCES s
- 1. Faller, I. L. , T. S. Chapman and J. M. Wes t, Calculations on U-235 Fission Product Deesy Chains, ANL-4807, Argonne National Laboratory.
- 2. #esting -Yentila ting Air Condi tioning Guide, American Society of Heating and Vencilating Engineers, 1958.
- 3. Me teorology and Atomic Energy, U. S. Department of Commerce, Weather Bureau, July, 1955
- 4. Masiana Permissible Anounts of Radioiso topes in the Human Body and Maziana
. Permissible Concentrations in A ir and Fa c e r, National Bureau of Scandards, Handbook 52.
- 5. Dunning, G. M. , Thyroid Dos e from Radiciodine in Fallout. Nucleanies, Vol. ,14, p. 40. February, 1956.
- 6. Clark, F. H. . Decay of Fission Product Commes, NCA-27-39, Nuclear Develop-sene Associaces,.Inc., Dec embe r, 1954.
%. e e
I!I/3-7
. U CL A - NEL 2
* ' ~..
D a s s...- il -. d. so -/w, f64
." m .g ;
I - o JMi 17 'SU3,' 1 a ,/ dsp.As 3 . 6 45 jg.,o,) g
- e. . 1
.,L
r, - .- ? g
.A,\ '
,77 .
.- ' \ \
~ ,
> 4 j *D E P A R T M E N T O F E N G I N E E R I N G !
=
5 t. 5
= -
,. - ~1
> = ,
UCIG # #~~- % s% i
? j l
trGining T80Ct0F f[9%fA[L
#jfp j
. h\* f, p s - m hGZGTd5 GRGlJSIS -
t&
# ' :!;?p A <!!
Q
. . . -Ln, L, .. .
FINAL REPORT 1-f g d'j i - R.D. MAC LAIN % ! a .
. O !
b @ ; 3
) d P E t
i E k E LO5 ANGELES i .= U N I V E R 51 T Y OF CALIFORNIA,
.i t
i
?
2 99 h s
- .. .s .--e= . --u--- .
Report. No. 60-18 March 1, 1960
,1 UCLA TRAINING REACTOR HAZARDS ANALYSIS FINAL REPORT l
80 - AB - 478 - 9,7// X7/O/ - R. D. MacLala ( Department of Engineering University of California Los Angeles
INTRODLETION This report has been prepared for submission to the U. S. Atomic Energy Commission as part of the facility license application for a nuclear training reactor which is being constructed by the University of California at Los Angeles. The UCLA reactor is similar to the training reactor at the University of Florida at Gainesville, Elorida, and reported in University of Florida Training Reactor Hazards Summary Report. J . M. Duncan, Bulletin Series, N o. 99, Vol. XII, No. 10, Florida Engineering and Industrial Experiment Station. The only significant difference between the two reactors is the fuel enrichment. The UCLA reactor uses 90% enriched fuel na opposed to the 20% fuel used by the UFTR. I O t se t4I l
e SECTION I A. REACTOR SITE
%e reactor is located in a newly-constructed, perzament, reinforced concrete Imild-ing on the campus of the thiversity of California at Los Anples. He location of the building and its relationship to its surroundings is indicated in Figure I-1.
He normal building populations during a school day are givec in Figure I-2. At times other than school days, campus building populations are re&ced to small fractions of the figures shown. De 400-acre campus is located on coastal plain approxir.ately five miles east of the Pacific Ocean and 15 miles west of the los Angeles civic center. To the south of the campus is a business and shopping district, and to the north, wes: and east are residen- I tial areas. A map of this general area is givai in Figure I-3. Geology he UCLA campus is situated on a coastal plain, and is apprimately 400 feet above sea level. De coastal plain consists of a terraced alluvial fill, 200 feet deep at the reactor site, overlying sedimentary rock of rather recent origin. He coastal plain lies at the base of the Santa boica buntains which are 2000 feet higi He most important __ forination in these cumntains is Santa Monica siste, an old sedimettary layer 2000 feet thick. Overlying this slate stratum are several more recent sedimentary layers. A cross section of the coastal plain near the campus is given in Figure I-1. "Ihis section is at right angles to the anticlinal folding of the Santa Monica huntains. Hy rf ro l o gy l No deep wells have been drilled on the campus of UCLA or in the vidnity of the l campus. He water table is estimated to lie 200 feet below the surface in this area. l A log of a typical test mell made by a foundation engineer near tie site of the reactor building is shown in Figure I-5. Surface runoff water is collected in concrue-lined storm drains which anpty into the ocean. His drainage system has been adequate to preven: any flooding of the esapus by heavy winter rains. De maxismun rainfall in any 24-hour period during the last 75 years was ten inches, as indicated in Figure I-6. It is barely cenceivable that runoff from the watershed area north of the campus could flood Westwood &nlevard and the area l to the west of the reactor site. However, the reactor core lies about ten feet above I this level, and a rainfall equal to the largest ever recorded would not flood the re-actor. In the unlikely event that such floo, ding should occur, it would pose an extreme operational inconvenience, but would not create any radiatics hazard. Se i s mol o ry Southern California is seismically active. He locatices of known active faults are indicated in Figure I-7. He nearest of these to the reactor site is the Inglewood fault running in a north-westerly direction about two miles east :f the can. pus. In Southern California, the region fron. the Mojave Desert to beyerd :he off-shore islands is traversed by a series of active faults. Dese faults extend f can 20 to 50 to many hundreds of miles in length, and the trend is generally betweer. north and west. How-ever, they are only roughly parallel, and in certain instancss a cajor fault zone l divided into two or more well defined faults. In general, t:ese faults are from five to twenty miles apart and apsarently extend to depths of 15 or nore niles below the surface. e
SECTION II TRAINING REACTOR DESCRIPTION _I_n t r o d u c t i o n A reactor which is to be used for student instruction must be designed so that safety is insured without exercising greater restraint on the activities of students than is nor-mally advisable in a university laboratory. Eis necessitates: (1) that the total avail-able excess reactivity be limited to something less than that needed for prompt criti-cality; (2) that the reactor have a high degree of denonstrated inherent safety, and (3) that it be limited to Ice-power operation. Rese requirements are met in this reactor by combining a water-moderated, plate-type fuel section with a graphite system for maintaining a fixed geometrical arrangement. Here is no credible way in which the fission products of this reactor can be made to escape, and the amount of contained fission products will be relatively small since it is limited to a maximum steady state power of ten kilowatts. Nevertheless, because, of the reactor location on the campus environs, it is housed in a structure with a minisaan num-ber of penetrations sealed against gas leakage, a A. BASIC DESIGN FEATURES his reactor is of the same general type as the Argonaut Beactor and similar to the University of Florida Training flenctor.' He basic element of the reactor is a rectan-gular prism ( 5 x 5 x 9 feet) constructed of graphite bars. Le fissionable material is introduced into the graphite prism in the form of aluminum-uranium alloy plates in six aluminum boxes, each of which contains a small amount of water. Se object of this con-struccion is to have the convenience of a solid moderator-reflector and the safety of the water plate arrangenent. De convenience of the structure is illustrated by the fact that three large, plane neutron sources are obtained. %ese plane neutron sources can be con- ' verted to thermal columns merely by adding sufficient graphite, or one of them can be used in shielding and odier experiments merely by placing the structure to be tested against the plane source. He upward direction can be used conveniently as a neutron source for exponential experiments. De distribution of weight on the reactor is a minor problem because the reflector structure is solid graphite. Le 90% enriched uranium-bearing plates are innersed in sufficient water so that a power excursion which would eject the water from the aluminum boxes certainly would reduce i the reactor to below criticality. Since th'e reactor is to be operated at a maxioaun of I 10 kw, only a small amount of reactivity is required for the temperature coefficient and xenon poisoning. Thus, it is possible to operate the reactor with an amount of excess reactivity which is well below that required for prompt criticality. Under these condi-tions, the reactor treets the safety requirements of a training reactor and can tolerate considerable operational error without damage. De fuel is contained in MTil type plates assembled in hundles. %ese fuel bundles are contained in six watertight aluminum boxes set in a two-slab arra) in a 5-ft prism of graphite bars. Le control rods are the swinging-arm type similar to those u: bed on CP-3 and CP-5, and University of Florida reactors. Four cadmiuc vanes protected by magnesium shrouds operate within the spaces between the fuel boxes. These are enved in Cnsterss ty of Florsda trainsq Reactor Hazards Suzmary Repc-: . ' V. Duncan. Bulletin Series No. 99 Vol. Xil. No. 10. Tlorsde Enraneering and Ir.dustria: Experimerit Station.
l B-1 1 l 1 APPENDIX B ESTIMATION OF EFFECTS OF ASSUMED LA2GE REACTIVITY ADDITIONS It has been demonstrated repeatedly in the Borax and SPERT reactors that water-cooled, water-moderated reactors of suitable design may have a very substantial self-protection against the effects of reactivity accidents, even in the absence of corrective action by the reactor control system. His self-protection is provided by the negative steam-void coefficient of reactivity and the negative temperature coefficient of reactivity, both of which can result in important reactivity reductions as the reactor power rises. The UCITl has been designed with a high degree of self-protection of this type. In this appendix estimates are made of the behavior of the reactor under various hypothetical con-ditions of excess reactivity addition with no corrective action by the control system. The characteristics of the UC1TI which determine its behavior kring power transients resulting from large reactivity additions are quite similar to, but not identical with, those of the Borax I reactor. Its behavior can be predicted most reliably by utilizing the Borax I data with simple correction factors to convert them to the UCT11 conditions. The significant quantitative characteristics of the UC111 and the Borax I reactor are compared in Table B-1. s TABLE B-1 COMPARISON OF UCTR AND BORAX I CHiRACTERISTICS CitARACTERISTIC UCTR BORAK I Fuel plate " meat
- 13.4 w/o U-Al alloy 18 w/o U-Al alloy -
90% enriched fully-enriched Fuel plate cladding 1100 aluminua 1100 aluminue
" Heat" thickness 0.040 inch ? 0.020 inch Cladding thickness 0.015, inch 0.020 inch 0.137 inch Coolant-channel thickness 0.ll7 inch T-Core voluce (approx.) 71 liters , 106 liters Void coefficient of -0.88% k/% coolant -0.24% k/% coolant reactivity (calculated) void void Teeperature coef ficient of -0.009% k/ C (estimated) -0.01% k/*C reactivity (room temperature)
Effective prompt-neutron 1.4 x 10~4 sec 0.65 x 10-4 sec lifetime (calculated) Power ratio in core, 1.63 1.82 manimum average
. B-2 In addition to the quantitative differences, the UC'ITl differs from Borax I in that the maxi.ar. coolant water level is only a few inches above the upper ends of the fuel plates (instead of about 4 ft) and the coolant water, once it has been ejected forcibly fraa the - core by a power excursion, cannot fall or flow back into the core. . Effect of 0.6% Excess Reactivity _ An excess reactivity of 0.6% kert will be available in the reactor if its temperature is almomally low (nearly freezing). The addition of all this excess reactivity will cause the reactor to operate at a power su::h that the reactivity losses associated with the tarperature increase and the voids formed mill equal the initial excess reactivity. If the reactivity is added slowly, after the reactor is critical, the power will approach such asi equilibrium level slowly as the reactivity is added. If the reactivity is added-suddenly, when the reactor is initially suberitical or at very low power, the power will at first rise exponentially with a period not shorter than 0.8 see which is the asynytotic period corresponding to the full excess reactivity of 0.6% k ort. Many experiments with the Borax reactors have demonstrated that for periods of this order of magnitude, the transition froc the exponential power rise to the equilibrium power level (in which excesa reactivity is balanced by temperature and steam void coefficients) is a smooth one involving little or no power overshoot. & the basis of this experience, it can be said that the mapiitude " of the power excursion which results from the 0.6% reactivity addition will not depaid greatly on whether the reactivity is added suddenly or relatively slowly and in neither case will it approach a level shich would cause a fuel plate to burn out. In order to compute the po=er level at which the reactor will operate after the addi-tica of the 0.6% excess reactirity discussed in the foregoing, it is necessary to know the water-te=perature coefficient of reactivity. The relative inportance of the two modera- - tors, graphite and water, in determining the effective neutron temperature incrochace un-certainties in the theoretical computation of this coefficient. The coefficient cannot, however, have an absolute magnitude less than that of the water-density coefficist of reactivity referred to a tanperature scale, i.e., the coefficient computed on the assumption - that: d kerr _ 3 kort So dT Sp ST shere o refers to the water density and T to clie tenperature, & the assumption that this n.inimu:n salue is the true value, a rise of water terperature from near 0 C to 80*C would reduce reactivity by 0.6% ke rr. The capacity of the reactor-coolant system is such that if the outside' air temperature were C*C and the average mater ten:perature in the reactor were 80 C, energy would be removed at the rste of 365,000' BTU /hr or 107 kw. thder these conditions the reactor water-inlet terperature sould be 60*C and the exit temperature, coincidentally, would be 100*C. It is, there fore, ecncluded that if tLe full available excess reactivity of 0.6% kerr were added to the reactor on a cold day with the coolant system operating, the reactor would operate at an ecnilibrium power level about ten times higher (100 kw) than its normal maximum with little or no net steam production. be fore reaching the equilibrium pomer, when the water
- in the coolant system would te heated to the equilibrium value, the reactor ould operate at a scre-hat higher power level and some net steam production might occur. If the coolant were r.ot f.o.ing during the tire of excess reactivity addition, the equilibrium power level
B-3 would be quite low and equal to the heat losses. In no case would the power level approach a high enough value to justify any fear of fuel-plate bumout. Eax t erum Tolerable _Sud_ den Reactiv_ity Addition In order to assess the safety factor which exists between the normal excess reactivity available in the reactor and the excess reactivity necessary for a serious power excursion, it is useful to estimate the value of excess reactivity which, if suddenly inserted and not removed by the control system, would raise the exutaatn tenperature in the h_o31 git fuel pl_a_tr to the melting point. Such an excursion would damate the reactor core but would not result in any substantial release of fission products. i Le 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 Borax I. %e estimat.e requires that (1) a relationship be established between the maxi-usam tenperature of the fuel plate and the energy release of the excursion and (2) the energy release be related to the period of the excursion. For the case of power excursions of short period, with reactor water at saturation tem-peratu re, it is shown in Beierence 1 that the maxiansn fuel-plate temperature rise is, to within experimental error, proportional to the maxinum energy release of the power excursion. -* He proportionality was determined to be constant 24.4*F per MW-sec.* Measurements of the same type with cold reactor water (the case directly applicable to the UCTR) showed a simi-lar relationship but with a proportionality constant of only about 10*F per MW- sec (Bef-erence 3). Le difference is not an unreasonable one since the subcooled water represents a more effective heat sink than the saturated water. However, the experiments with the saturated water were carried to short periods in the range of interest whereas the subcnoled experiments were limited to longer periods. Herefore, more conservative saturated water data will be used.** To raise the maxianun tenperature of the fuel plate from the tempera-ture of boiling water to the melting point of aluminum, a tenperature change of approxunately 1000*F, would require a power excursion with a total energy release of 1000*F or 41 M W -sec. 24.4*F/M W -see -
- ; ~ ..:
According to the data of Beference 3, replotted in Figure D-la "subcooled" power excur-I sion of reciprocal period 150 sec-1 would give an energy release of 41 M W-see in ' addition to the energy necessary to raise the fuel plate temperature to the saturation temperature of water. It is, therefore, concluded that a power excursion o(period at least as short as 1/150 see (6.7 millisec) could have been tolerated bgBorax I with subcooled water withour melting at the hottest point in the fuel plates. E.xperiments of the Borax and SPERT types have not been made with reactors having widely different neutron lifetimes. Le general evidence of the experiments, however, sup-ports the supposition that of the three related variables--neutron lifetime, excess reactivity, and exponential period--which characterize the neutron physics of a power ex-cursion, it is the exponential period which determines the total ,.:nergy release and the tesnperatures attained. He excess reactivity and neutron lifetime have large effects only as they jointly determine the period. his supposition is consistent, for example, with l *Actually, the energy data of Reference 1 were revised in Reference 2 because of later and i better calibrations of the instrumentation. The numbers above are taken from the later (more pessimistic) data.
' If subcooled data were used, the case directly applicable to UCIB. this analysis would i
indicate that step reactivity additions 2.4 times sa large as those discussed here would ! not d am e tre the reactor. =
4
" /
0 f B-4 . 'Q 5s
. -W~ ;M ~
r, "[* . d%, ,:, '
. that the total energy transferred to the ' coolant water during a power excursion is many times the amount which would vaporize enough water to compensate for the excess reactivity,
- and that the actual reactivity reduction which occurs during the excursion is much larger l than the initial excess reactivity. 'The ext 2nsion of the Borax results to the UCllt is made on the hasis of this evidence.
It is convenient, firs t, to treet only the effects of the slightly greater fuel-plate spacing and the slightly lower void coefficient of reactivity of the UCTTI relative to the Borax I. Information will also be drawn from the Borax II experiments. The Borax II re-
~
actor differed from Borax I in that the coolant-channel thickness was greater in the retic 0.264 in* :'2.26l and that the calculated void coefficient of reactivity was lower in the 0.117 in. ratio. . e' O.105 k.tr/% void : 1
* * ' ,4' . = 0*416
~
0.24% k.gr/% void \2.4 Both of these differences would be expected to cause a higher energy release per fuel plate
' in borax II than in Borax I for a power excursion of given period. 'lhe measurements madg y ,.
with subcooled water at periods down to 23 millisec showed that the energy release per" fuel
.pjats in Borax II was between{l.7 and 2.0] times that of Borax I, with the smaller ratio s
' . applying to the shorter periods Uleference 2). Berefore, it seems quite conservative to asstane, in the case of any two reactors, (1) and (2), of the Borax type having a ratio of
! fuel plate spacings, S1/S 2 , and a ratio of void coefficients of reactivity, C1/C2, that the ratio of energy release per fuel plat.e for a subcooled power excursion of given period will be no greater than E2/E1 = S2/S1 or E2/E1 = C1/C2 whichtver is the larner. For the .
UCTil and Borax I the ratios are
'D.- SUF Cao 0.24 _
.- 0.137 - 1*17 1.33 0.18
' 'j ] s Sg. 0.1 17 CUF It is concluded, therefore, that a Borax reactor having a coolant-channel thickness and a void coefficient of reactivity equal to those of the UCIll would release not more than 1.33 tim-s as much energy per fuel plate as Borax I. ne limiting nonmelting period for such a reactor would be that which in Borax I gave an energy release of 41/1.33 = 31MW-sec. The period obtained from Figure B-1, corresponding to a total energy release of 31MW-sec, is
- ' 8.3 millisec.
In comparing the behavior of different fuel plates, it must be recognized that the total energy release of the power excursion can no longer be considered as a definitive variable because a large fraction of the total energy released is stored in the fuel plate during the important stage of the reactor shut-down. For example, a reactor composed of fuel plates of high heat capacity undoubtedly will experience a larger total energy re-lease, but not necessarily a higher maximum temperature, during a po.er excursion of given period, than a reactor having plates of low heat capacity. From examination of the Borax results, at seems clear that two distinct phases of the reactor shut-down process occur consecutively and that both may be important in determining the er.axinum center temperature of a fuel plate. The first phase cosers the interval before an important amount of boiling occurs at the fuel-plate surface. During this interval, the , hea t loss to the water is sn.all and the important consideration is evidently the ratio of fu-1-plate surface temperature (which determines the start of toili w) to center tempera-ture. For periods in the rang- under consideration, this terrperature ratio is theoretically Experimentally, the
- .o : far from unity (0.76 minin um for :.10 *cillisec period in Borax I).
A B-5 temperature ratio was unity for periods down to 5 millisec in the Borax I measurements. Since the total effect is small and since the temperature ratio for Borax and UCIR fuel plates should not be much different, the thinner cladding will tend to balance the effece u of the poorer " ment" cm ductivity. It is concluded, therefore, that there will be no in- 'b portant difference in fuel plate performance during this initial phase of the excursion. g
%e second phase of the peer excursicm begins when a significant rate of boiling is established at the plate surface. Reactivity and consequently generation are reduced at a rate which must be a fur.ction of the rate at which heat can be transferred into the boiling water. At the same tine, the transfer of heat into the water removes heat from the fuel plate and limits its tesperature rise. De impo:-tant characteristic of the plate during this phase of the excursica is the heat flux which it can supply to the water for a given temperature difference between the plate center and surface. A figure assumed to be roughly indicative of the relative performance or merit of fuel plates during this phase is the ratio of heat flux to temperature difference under steady-state conditions. Bis ratio (figure of merit) will overemphasize the dif ference betmeen fuel plates since the tem-perature distribution in the plate will be more peaked during a steady-state conduction than during conduction when the general temperature level is rising. Le ratio of these figures of merit for Borax I and for the UCIR is
' Heat Flux' AT e.. -
UCTR - - *
= 0.82
' Heat Flux' A T e.' '
Borex A conservative procedure would be to apply the above factor to the permissible total energy of excursion on the borax I curve. At the same time, however, the difference in gross maxinaso to average power ratio for the two reactors should be taken into account since it in the temperature of the bottest point in the hottest fuel plate which is being considered. Le power ratio for the two reactors is Mu -
^" Borax = 1.82 = 1.12 Max 1.63 Ave UCTR l
He combination of these two factors reduces the permissible equivalent energy of the Borax-type excursion to 31 x 0.82 x 1.12 = 28.4 MW -see
%e corresponding exponential period from Figure B-1 is 9.1 millisec. It is, there fore ,
l concluded that the UCTR will tolerate a peer excursion of period at least as short as 9.1 millisec without the melting of any part of any fuel plate. The excess reactivity corres-pending to this period is 2.3% kegg. Successive Power Excursions It is typical of the Borax and SPERT reactors, unless the excess reactivity is removed by external means, that an initial power excursion which terminates itself by expelling water from the reactor core will be followed by subsequent excursions as the water falls and flows back into the core. An exception to this behavior occurs when the initial
~
l l
l \ B.6 excursion is violent enough to cause a permanent loss of reactinty by throwing a large amount of water completely out of the reactor tank. In the UC.3 the total quantity of water in the core ia small the Argence of the core is maall, and baffles above the core are ao arranged that any water splash is directed to the outside so that it cannot return to the core. Consequent.ly, even a relatively mild power excursion (e.g., one hav-ing an exponential period of from 20 to 30 millisec) in the UC1'R should result in perma-nent self-in&ced shutdown of the reactor. By these same design features, the possibility of large successive power excursions, such as those studied in the SPERT project, result-ing from the ramp addition of excesa reactivity is eliminated. It can be anticipated that the UCTR will be safe against quite large ramp additiens (larger than 2.~S karr) provided only that the ramp rate is not so rapid as to add an excess reactivity of more than 2.3% kort before the reactor power reaches a high level. To exceed this limit the reaip' rate - would need to be of the order of 1.0% ka rt per second or larger. Ream Tube Reactivity Ef fecta
%e UCTR has two 6-inch diameter beam tubes which extend to within 11 inches of the fuel-graphite interfaces. 'lhe mm== change on the core reactivity which can be effected by these two beam-tube facilities oss calculated to be 0.18% AK/K or 0.0R .SK/K per beam tube. He calculation is based upm the effect of a black absorber six inches in diameter placed in the same position as the beam tubes. Le reduction in the reflector s savings Me to the black absorber was calculated using the following equation.
8 (reflector savings) : D(cord . L (reflector) tan h T(reflector thickness) , D(reflector) L(reflector) Beference: Elements of Nuclear Reseror Theory, Glasstone and FAlund. , Le reflector savings for the 49.5 cum and 28.0 an of graphite were calculated to be 7.83 cm and 5.26 cm respectively. He area of the black absorber is 13% of the adjacent core fr.ce area, bsing the reflector savings given abcve and the area weighting factors, the . reflector savings with and without the six-inch diame.ter black absorter were calculated to be 7.33 cm and 7.83 am respectively. The reactivity effect of the single six-inch diameter black absorber was then deter-mined calculating the critical bu& ling with and without the black atsorber. Using the value of -0.09% 6K/K , for a sin 6 1e beam tube the shortest period which the reactor could go on, due to the sudden withdra al of a black absorber froc the six-inch beam tube, would be approximately 80 seconds. Therefore, the reactivity enange which can be effected by the beam tubes does not represent a hazard to reactor operation. in addition to the two 6-inch beam tubes ,hich penetrate the outer reflector, there are four 4-inch beam tubes which tenninate outside of the reflector. No calculations were made for the 4-inch tubes sines their ef fect on reactinty will be auch smaller than that of the 6-inch tubes.
)
9
Eh? REFERENCES
- 1. Diet rich, J. R. and D. C. Layman, Transient and Steady State Characteris-tics of a Boiling Beactor. The Boraz Experiments, 1953, AECD-3840, Argonne National Laboratory, February, 1954,
- 2. Dietrich, J. R. , Experimental De terminat nans of the Self-Regulation and Safe ty of Operating Water-Moderated Reac tors, A/ Coni. 8/P/481, Interne-tional Conference on the Peaceful Uses of Atomic Energy, June 30, 1955.
- 3. Dietrich, J. R. , Experimen tal Inves tigation of the Self-Limitation of Power During Reactivity Transients in a Subcooled, Water-Moderated R eac tor. Borax-I Experiments, 1954, AECD-3668, Argonne National Labora-tory, August 17, 1955.
- 4. Lennox, D. H. and C. N. Kelber, Summary Report on the Hazards of the Argonau t Reac tor, ANL-5647, December 1956.
e f
- y. ~ . .+ n .d ~ -
,:93. wy w.in -
,. .~..
u;..
. ;w. ,~ -
."t.
2
.:, 5:. g.
4. e t p h::: - Y+ hs
- n. .. ~ ,
,. ,p .
,'^ ,#
2y 99 E . 4 s. UCLA g .c,s N/ D' TRAINING REACTOR HAZARDS
SUMMARY
REPORT C s { l-eg- g ,. Prepared by
~,,.. .
AMr ATOMICS
,- a division of AMERICAN MACHINE #4 FOUNDRY COMPANY Greenwich, Connecticut October, 1959 o ,
a ;ice: . x g 7~. , k * '* sh,w. - .
- da . .
** q
- i
_ __ __ ---a-.- ,r a
. r 7-
. ; ,- - ' cv - q , *., . ,, 7. ,.Q '4 '. " , ,{ '- *,- .! . ,.. '. g e
..v, -__,m,d 4 { . .y._ ~ . . i'g M. ,f. 7
.M,.7..-
w L-a.
.d 9
e . s-
\
a
. w o 9
I SECTION 111 5 REACTOR HAZARDS EVALUATION . 3 e w-- l n l l I I I i 4,.=e
-. = -- %= 2_1-
." 1* ~ wa s = ,, .m _m u - _
- * ~ ' . ._ h h-TIh..[___
.-.--.-eo-.L-a-M-: % -M' "*MM-M tit- *T" ~i + ;_ -- ;=5 f /i Kh lg 2-
= n ; ~4 _: . . _.. n '" -- _ _ _ .
- - - . , - 4 s.-tI..a :v;5 E 'p. .
. - r' , - : .
r.- ~s 2 *
~ -- ~ - T.? cya ' ._L +v- - 5 '. ci?g.- ?'4 - - ~ w " %- 3. :7r f
[
.;A .r
/A.
GENERAL SAFETY CONSIDERATIONS '
- A
-> The inherent safety of this reactor is based on four g -
points. First, the amount of excess reactivity in the reactor is lirnited to about 0.6 per cent. Second, the reactor has negative therrr},al,and void coefficients. In addition, the reactor is provided with sufficient interlocks and safety tripo to make a hazardous incident extremely improbable. Third, tbe 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. Although the reactor is designed to operate at a maximum steady power of 10 kw, it is not planned to operate it at this power level Q - continuously. Much of the dperation for the training program will be at considerably lower power levels and will be intermittent. It is estimated, that the reactor will be in operation (about 1800 hours each year or about
' ;.r ..
20 per cent of the total time).* With this type of operational program, no w, very large amount of new fission products will ever exist in the core. The excess reactivity will be limited to 0.6 per cent k by . adjustrnent of the original core loading. (Any additional in ! p.' des will be kept in a locked storage cabinet or fuel storage holes as described in Section II-E).* Loading or rearrangement of fuci in the reactor will be
- To be verified by University.
45
%. 3 1
^
f ^
, f ' '- = l~= =~=W%WM:m. -
_~ _=_h _ ^ ^^j - - -. . . . .:.:} ^ ~ ' ' ' ~~ ~7. y._____ ^ :-. T, - e&w . -~ ==--_
if" .{ y r , ( a3 p:- N E. INITIATION OF REACTOR HAZARD BY EXTERNAL MEANS r/, j. .
- 1. Fire -
I Since none of the materials of construction of the. reactor are , inflammable, and since the reactor core is very well protected fro any external fire by the concrete shield, damnge from fire resulting in the re- . Icase of fission products is extremely remote. (The reactor building is .# fireproof construction and will not be used for storage of quantities.u of : w-
~y m- .
inflammable materialo. ) * ' uc = l Z.. Phenomena of Nature ~- None of the phenomena of nature such as hurricanes, . tornadoes, ' ) . se. a. lightning or floods offers a credible mccus of initiating a hazard related to c;.y ' the presence ' of the reactor. Adequate production has been incorporated into arq. the reactor design to prevent Interal movement of the sbielding blocks due to carthquake forces, and to SCRAM the reactor in the event of an[bisplace-
, 4", e
ment of the shield blocks due to theso forces. (Hurricanes and tornadoes -2 -
Icve been rare and not severe over Los Angeles. Lightning is frequent but, with proper construction and protective devices, should offer no kn=rd to s l the reactor. SM5" The heaviest rainfall for a 24-hr period recorded during the
.. . . q .
last 20 years occurred in (
) ** _with no evidence of flooding in the area of the reactor site. ) *
'fo be verified by UCLA.
f To be specified by UCLA ' s 56
,.-y . _
.R- . 1.u
~
'~ '..a.
m-
- _ - .-- _a .ma.- '
q-
. 3: , . . , ___.. ~
' ~ -
.~;.n., . :-54..
. 2- .-s s ,
o
.V ' APPENDIX A
( REACTOR CALCULATIONS ,
. Two-group diffusion-theory calculations have been made to estimate the reactor-physics For the calculations, the actual configuration of the training-reattor core and cha rac te ri s tic s .
reflector assemblies was transformed into an equivalent one-dimensional model uning the con-
' cept of reficctor savings to account approximately for the neutron leakage in the other two di-rections. Figure 14 is a sketch of the one-dimensional model showi,ng,the three separate I regions considered in the calculations. Region I is the central graphite zone coupling the two core slabs, Region II are the cot e slabs, and Region III is the external graphite reficctor.
For the training-reactor configuration, a reasonably accurate value of over-all reflector savings is required to account property for the neutron leakage, principally from the graphite regiono, in the two directions transverse to that considered in the one-dimensional model (two directions perpendicular to the X-direction shown in Fig. 15 ). In addition to the diffusion leakage of neutrons (accounted for by use of the reflector savings), streaming of neutronsThe oc-curs through the four narrow air slots (3/4-in. air gap) provided for the control blades. leakage current of streaming neutrons cannot be determined accurately by any practical calcu-lational technique and so can be estimated only roughly. For the initial calculation, the Icakage of streaming neuirons through the air slots is neglected. Its effect on the critical-mass re-quirement is roughly estimated later. To obtain a reliable value of over-all reflector savings, advantage was taken of the reactiv-ity measurements made at the Argonne National Laboratory and reported in ANL-3647 for the A rgonaut reactor. There are no air slots in the Argonaut design. One of the Argonaut config-urations for which reactivity values were experimentally determined consisted of two groups of s fuel-box clusters arranged so as to approximate closely two core slabs facing each other and ) sepa rated by two it of graphite. The reflector-savings value used is based on obtaining agree-ment between calculated and measured critical-mass requi r em ents of the Argonaut two-slab , configuration. In the calculations, the two-group nuclear constants used for the core regions were determined by standard methods generally found satisfactory for water-moderated cores. The resonance absorption of the U-238 in the 90 -per-cent-enriched fuel used was included in the calculations. The nuclear constants for graphite of average density equalto 1. 6 gm/ce were obtained from Appendix 13 of ANL-5647. The critical-mass requirement for the training reactor, neglecting the effect of neutron utreaming through the control-blade air slots, was determined in the same manner as that used to check the A rgonaut critical-mass value. In addition, calculations were made of (1) the uni- ! form water-void coefficient; (2) the temperature coefficient; (3) the effective prompt-neutron lifetime; (4) the reactivity change attributed to loss of water f rom the core slabs; (5) the r e - the ef-activity change resulting from consumption of the U 235 content in the core slabs; (6) feet on the critical-mass requirement of bringing the two core slabs tog-ther. The core aluminum-to-water volume ratio is0.51 and the separation distance between core slabs i s 1. 0 f t, for the UCTR design . The fuel is 90-per-cent-enriched U-235. The critical mass required for the training reactor in the cold-clean condition, neglecting the effect of neu-tron streaming throur.n the air slots, was calculated as 2.6 kg of U-235. Additional calculated results obtained are tabulated as follows: Uniform wate r -void coefficient ------------- -0.18 % k/% void T em p e r atu r e co e f fi cient - -- -- -- -- -- - -- - -- - 9.9 x 10 -4 k/oC P r om pt -ne ut r o n li fe tim e --- - -- -- -- - ---- -- 1. 4 x 10 -4 sec ,I U -2 3 5 m a s s coe f ficient ------- --- ------- 0,31 % k /% U -2 3 5 ma s s
- A-1
.a. . =.:_ - L y .g = _
- - -. ; =.----=-- _
,m;y.
2
/
,ompleta removal of water from the core slabs gives a k rt e of 0. 56 which representa a frecctivity lose from the system.
pCalculations made for the two core slabs placed side by side against each other with gra-jte completely surrounding them give a critical-mass requirement for the cold-clean condi-n$n of 1.9 kg In Figure 16 , the critical masses calculated for the two-slab separations (zero and t' In addition, the critical-mass data [1.obtained 0-ft separation) are plotted versus separation distance. for several Argonaut configurations are shown in the plot. The points plotted are design; (2) 1.9 kg. calculated
/ (1) 2.6 kg. calculated for 1.0-ft separation used in the.UGTit p for zero separation between the two core slabs of the UCTR (3) 3.,71,8,kg. measured for 2-ft separation in the Argonaut experiments: and (4) 2.2 kg. measured for a slab loading on one side caly in the Argonaut experiments. The slab dimensions are indicated on the plate. The core compositions for the fuct boxes used in the UCTR design and in the Argonaut experiments are closely comparable in neutron slowing-down and thermal-neutron absorption properties.
It is evident from Figure 16 that as the core slabs are brought together, the critical mass requirement decreases. This is undoubtedly due to increased coupling between the slabs (i.e. a greater fraction of the neutrons born in and leaking from one core slab causes fissions in the other slab). The points in Figure ,16 show generally good agreement between calcu- ~ lated and experimentally measured critical-mass requirements. The calculated-void coefficient, temperature coefficient, neutron lifetime, and U-235 mass coefficient for the UCTR are of the same order of magnitude as measured for the Argonaut two -stab configuration. The small differences are generally in the direction to be expected from the better coupling between core slabs in the IK~TR design. . Data reported in ANL-5647 for the two-slab configuration show that a 4 x 6-in. void, 36 in high, in the outer graphite reflector zone, next to a fuel niab, gives a reactivity decrease of -2. 2 per cent k. This reactivity decrease corresponds to about an 8-per-cent increasein required critical mass. On this basis, it is estimated that the increase in critical mass for Q the UGTIl , caused by the air slots could be as high as 20 per cent of that calculated for a struc
- ture with no air slots. This gives a cold-clean critical-mass requirement of 10 about 3 2 kg. II The core average thermal and fast fluxes are calculated as 8.2 x 10 and 1.9 x 10 n/cm2-sec, respectively, for 10-kw operation. Figure 17 is aThe plot of the thermal and fast abscissa in Figure 1*/
flux distributions along the X-axis as defin'ed in Figure 16 . is measured from the line, of symmetry of the reactor configuration (center line) of internal for graphite zone. Also plotted in Figure 17 are the fast and thermal adjoint fluxes useful determining the important functions for materials placed or changes made, at various positions in the reactor system. Calculations have been made to establish a lower limit to the control effectiveness of the control blades used in the design. The calculations indicate that a minimum of 1. 5 per cent k per blade (three blades) can be expected. The remaining blade used for regulation purposes will be adjusted to give 0. 6 per cent k. l l l l l l l l i A-2 l l
- --m.=-.3 .. _
__3
#' APPENDIX D ESTIMATION OF EFFECTS OF ASSUMED LARCE REACTIVITY ADDfTIONS
/
.c ,
)
It has been demonstrated' repeatedly in the Borax and SPERT reactors that water-cooled, (( water-moderated reactors of suitabic design may have a very substantia the effects of reactivity accidents, even irt the absenco of corrective action by the reactor control system. This scif-protection is provided by the negative steam-void coefficient of reactivity and the negative temperature coefficient of reactivity, both of which can result in important reactivity reductions as the reactor power rises. The MCTR has.beentle' sip,ned with a high degree of self-protection of this typo. In this appendix estimates are made of the behavior of the re-- actor under various hypotheticalconditions of excess reactivity additionwith no corrective action by the control system. , The characteristics of the UCTR which determine its behavior during power transients re-sulting from large reactivity additions are quite similar to, but not identical with, those of the Borax ! reactor. Its behavior can be predicted most reliably by utilizing the Borax I data with simple correction factors to convert them to the UCTR conditions. The significant quantitative characteristics of the UCTRI.and the Borax I reactor are com-pared in Table'. 73. l. TABLE B'-l COMPARISON OF UCTR ' AND BORAX 1 CHARACTERISTICS UCTR' Borax I Characteristic , ithw/o U-Al alloy 18 w/o tf-Al alloy Fuel plate dm eat" (fully -enriched) (20 per cent enriched) 1100 aluminum 1100 aluminum Fuel plate cladding
- 0. 040 in. O. 020 in.
" Meat" thickness Cladding thickness 0. 015 in 0. 020 in.
0.137 in. O.117 in. Coolant-channel thickness Core volume (approx.) 71 liters 106 liters
-0.24 per cent k/
Void coefficient of 0.16.Per cent k/ per cent coolant void reactivity (calculated) per cent coolant void
-0.009 per cent k/oC -0. 01 per cent k/oC Temperature coefficient of reactivity (room (estimated) temperature)
- 1. 4 x 10 -4 sec 0. 65 x 10 -4 see Effective prompt-neutron lifetime (calculated)
Power ratio in core, 1.63 1.82 maximum average D-1 c , - 4 ,,. -- w Q
a* s.
. ' .[. ;- ' -
f. In addition to the quantitative differences, the YAR dtffers from Borax 1in that the maxi-
$ . mum coolant water level is only a few inches above the upper ends of the fuel plates (instead of /f about 4 ft) and the coolant water, once it has been ejected forcibly from the core by a power I encursion, cannot fall or flow back into the core.
Effect of 0.6 Per Cent Excess Reactivity . An excess reactivity of 0.6 per cent k erg will be available in the reactor if its temperature is obnormally low (nearly freezing). The addition of all this excess reactivity will cause the reactor to operate at a power such that the reactivity losses associated with the temperature increase and the voids formed will equal the excess reactivity. If the reactivity is added slowly, after the reactor is critical, the power will approach such an equilibrium level slowly as the reactivity is added. If the reactivity is added suddenly when the reactoris initially subcriticator at verylow power, the power willat first rise exponentially with a period not shorter than 0.8 see which is the asymptotic period corresponding to the full excess reactivity of 0. 6 per cent ke f t. Many experiments with the Borax reactors havedemon-otrated that for periods of this order of magnitude the transition from the exponential power rise to the equilibrium power level (in which excess reactivity is balanced by temperature and steam void coefficients) is a smooth one involving little or no power overshoot. On the basis of this experience. it can be said that the magnitude of the power excursion which results from
. the 0. 6 per cent reactivity addition will r.ot depend greatly on whether the reactivity is added suddenly or rel'atively slowly and in neither case will it a'pproach a level which would cause a fuel plate to burn out.
In order to compute the power level at which the reactor will operate after the addition of the 0.6 per centexcess reactivity discussed in the foregoing it is net.essary to know the water-temperature coefficient o,f reactivity., The relative importance of the two moderators, graphite and water, in determining the effective neutron temperature introduce uncertainties in the theo-retical computation of this coefficient. The coefficient cannot, however, have an absolute mag-nitude less than that of the water-density coefficient of reactivity referred to a temperature scale, i. e. , the coefficient computed on the assumption that: i d k eff _ 6 k eff 69
.dT 6P 6T 69 where p refers to the water density and T to the temperature. Note that p is just the negative i
of the void. coefficient of reactivity. On the assumption that this minimum value is the true value, l a rise of water temperature from near 00C to' 800C would reduce reactivity by O. 6 per cent k gg. e l The capacity of the reactor-coolant system is such that if the outside air temperature were O C and the average water temperature in the icactor were 800C, energy would be removed at the rate of 365,000 BTU /hr or 107 kw. Under these conditions the reactor water-inlet tempera-ture would be 600C and the exit temperature, coincidentally, would be 1000C. It is, ther efore, concluded thatif the full availabic excess reactivity of 0. 6 per cent keff were added to the reactor on a cold day with the coolant system operating, the reactor would operate at an equilibrium j power level about 10 times higher (100 kw) than its normal maximum with little or no net steam I 0
. B2 1
i=-- , - . . . .- -. - -- y.
WU
-L v.
1 p W 7 Before reaching the equilibrium power, when the waterin tha e lent system w:uld - G me net , 'to the equilibrium value, the reactor would operate at a somewh:t' higher power Icvs1*# (g steam production might occur. If the coolant were not flowing during the time of .. p s ,s reactivity addition, the equilibrium power leve1 would be quite low and equal to the, heat _ ~ -j - ses. In no case would the power icvel approach a value high enough to justify any fearof r 1-plate burnout. . $
~~
Maximum Tolerable Sudden Reactivity Addition I In order to assess the safety factor which exists between the normal excess reactivity avail-abic in the reactor and the excess reactivity necessary for a serious,,,p,ower excursion, it is useful to estimate the value of excess reactivity which, if suddchly inserted and not removed by, the control system, would raise the maximum temperature in the hottest fuel plate to the melting ' point. Such an excursion would damage the reactor core but would not result ' in any substantial release of fission products. , 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 stmilar effect in-Borax I. The estimate requires that (1) a relationship be established between the maximum - temperature of the fuel plate ar L
,g rc alance the effect of the poorer " meat" conductivity. It is concluded, therefore, that n Y JI!! be no important difference in fuel plate performance during this initial phase of the
's lon .
M;E -The second phase of the power excursion begins when a significant rate af boiling is estab-unt be a function of the rate at which heat can be transferred into the boiling, water. At the (/"['shed at the plate surface. Reactivity and
" came time the transfer of heat into the water removes heat from the fuct plate and limits its temperature rise. The important characteristic of the plate during this phase of the excursion / ic the heat flux which it can supply to the water for a given temperature difference between the, plate center and surface. A figure assumed to be roughlyindicative of.the relative performance or merit of fuel plates during this phase is the ratio of heat flux to temperature difference under steady-state conditions. This ratio (figure of merit) will overemphasize the difference between fuel plates since the temperature distribution in the plate will be more peaked during a steady-state conduction than during conduction when the general temperature level is rising.
The ratio of these figures of merit fcr Borax 1 and for the UCTR i s' Heat Flux AT c- s - UCTR
= 0.82
. Heat Flux
_Mc- s
- Borax A conservative procedure would be to apply the above factor to the permissible tot'al energy of excursion on the Borax I curve. At the same time, however, the difference in gross maxi-mum to average powe r ratio for the two reactors should be taken into account since it is the h tempe rat'ure of tile hotte st point in the hottest fuel plate which is being considered. The powe r ratio for the two reactors is ,
M ax Ave Borax _ l. 82 = 1.12 Max 1.63
^""UCTR The combination of these two factors reduces the pe rmi s sible equivalent energy of the Borax-type excursion to 31 x 0.82 x 1. 12 = ?8.h hiW- sec.
The corresponding cyponential period from Figure 18 is 9,1 millisec. It is, therefore , concluded that. the UCTR. will tolerate a power excursion of period at least as short as 9,1 millisec without the melting of any part of any fuct plate. The excess reactivity corre spond-ing to this period is.2,3 per cent ke ff' I Succe s sive Power Excur sions It is typical of the Bo r ax and SPERT reactors, unless the exce ss reactivity is removed by external means, an initial power excursion which terminate s itself by expellirig water from D5 l l
r
.h c' tor core will be followod by subsoquent oxcurcions as the water falls cnd T.
I ~ dnck into the core. An exception to this behavior occurs when the initial y / i sion is violent enough to cauco a permanont loss of reactivity by throwing
%; forgo amount of water completely out of the rea'ctor tank. - In the UCTR the total
" antity of water in the core is small, tho submergence of the core is small, an4
,J.k@affles abovo the core aro ao arranged that any water splash is directed to tho -
lo so that it cannot . return to the core. Consequently, even a relatively
/[&ou-mild poser excursion (e.g., one having an exponential period of fmm'20'to 30 g millinoc) in the UCTR should result in permanont self-induced shutdown of the p reactor. By theso samo design features the possibility of large succossiv,o power excursions, such as thoso studied in the SPERT project, resulting 'from the ra, p addition of excess reactivity is eliminated. It can be antici' pated that the UCTR will be care against quito largo ramp additions (larger than 2.3 por cent koff) provided only that the ra p rato is not 00 rapid as to add an excess reactivity of nnro than 2.3 por cent koff before the reactor power reaches a high level.
To exceed this limit the ramp rato would need to be of tho ordor of 1.0 per cent kggg por second or largor. Fenn Tuho Reactivity Effects The UCTR has two 6-inch diameter bean tubs The maximum 3 which chango extend on the cora to within reac'tivity 11 inches of which the fuel-crachite intorfaces. can bo offceted hv these two beam tubo facilities was calculated to be 0.18% or 0.09% AK/K per bean tube. The calculation is based upon the effect of a black iho obsorber 6 inches in diamator placed in the sano position as the beam tubes. reduction in the reflector savings due to the black absorber was calculated using the following equation. Q [(reflector savings) = D(coro) D(reflector) . L(reflector). (tan h [L(renector)T
- j
Reference:
"Elementaof Nuclear Roactor Theory" 01asstono and Edlund.
i The reflector savings for the h9.5 cm and 28.0 en of graphite were calculated to bo 7.83 cm and 5.26 cm'respectivoly. The area of the black absorber is 13T of i the adjacent coro face area. Using the reflector savings given above and the I area weighting factors, the reflector savings with and without the 6" diameter black absorber woro calculated to be 7 33 cm and 7.83 cm respectively. l l
'iho reactivity offect of the singlo 6" diameter black absorber was then determined
) cdcolating the critical buckling with and without the black absorber. I Using the value of -0.09T AK/K, for a single beam tube the shortest teriod which the reactor could no on, due to the sudden withdrawal of a black absorber from the 6" heam tubo, would be apnmrimately 80 seconds. Therefore, the reactivity chantro which can be offected bv the heam tubos does not represent a hazard to reactor operation, k - 3-6 l
./ jy
[' :. . -
;p c
+. -s'
,;; /
W REFERFRCES
,. . ; l 7
~
.v3.1 - nietrich, J.R., and D.C. Lreman, "Transidnt and Steady State Characteristics
. :: of a m iling Reactor., The Borax Experimente,1953,* AECD-38h0,,.Argonne --
National Dhoratory, February,195h. B.2 Dietrich, J.R., '" Experimental Determinations of the Self-Regulation and Safety f~i' of Operating Water-Moderated Reactortt'," A/ Conf. 8/P/ lib 1, International g Conference on the Peaceful Uses of Atomic Energy, June 30;'1966. B.3 Dietrich, J.R., " Experimental Investigation of the Self-Limitation'or Power During Reactivity Transienta in a Subcooled, Fater-Wderated Reactor. Borax-I Experiments,* 195h," AECD-3668, Argonne National Laboratory, August 17,1955.
.* * ** S f ' .
esF9-l' e . 4 0*' { l
=
l e l l 1 B-7 S
=s 9
E . Q'. Q. 1 4 ..s..a. . . . _ . .a.=*<. - - - - - - -
. L- .
- 1 I
's- s_ -+---
r
" ** - ~ * " - '~ ' " *
- l ii '_ i I
[f 7 __._.__ - gl lg f, j i
!;!! i ,,, ; , ,, ;
i
._ ! ! ' !lil 1:li
.,y l j j j. {-j { i i;ii ll '
i tF lM MI f!P q.g g.gg 3.g H.n pr: w W mW a-Am i.,, i o i i .w n ; o m ui t4_ [I~. IJ ;[ jj l llb i *
!!! !l'i 'it i. I l ! ! !l II j'I ll:!
- l:j I i
,4.
.t i
l- - ji I __.
.__ ,-[ ,
, a. . _ . _ - _ . . _ .
,, i.
l.g... . .
,.l._.
Qf gj
- . - . gi, t.;. . . - - - - - -
.3 *
~
~
j il i jjl ,I 'l I! 't ! ' t e , 4 ! hji ( 64 ji j n' t I t'l t .* 9 8, I U
, 9s ; e e
r' l. i'
, ,i .4l f, [ l t 1- j- l -l I l I llI i i*I I'ii 'I i m l f i[. l - '
bh ', I !ltl !!!' Ii!I '
!I i a 5
l
, i,i,v.I
.x, ,
I i
' i; i ,
I 8 4 .i
' :D : - '
i
' . . ___ p
' eli'
, ,,;,gl -
. -} I . tii I -8 ,
j , . j j- !, j' i -- l- l'i t-
! 'I ' l I 8 I '
O j i M:ji% i -E : L - _ E :(-L -f- 'I t -1 M-i -d h I hl! IN I LI ' O
@ isi- Wh hi4 Ndf ~lMEE YF N:I' i NI I -
" E W
aDi Ef ME sz_ : u: = - F-- ==sTT pr. L -ju : 9.i'g -- 1 a
.-he
@y:: : -=r::_::r Tr : :.:
I'-IlT .l:Id~ J - 4 4T : 1 : M i -' 1 4.i l l . 7H3 g j
-7_ .p __ .-- .
.g_ gg) 3
-; __ 7 _ . . _
8
, ,, ,, , ,i ..
- ; , , ii # i I 4
,I
.+ -{ t }1 -l f
' i t * !i f i 1 ' I h i,i
,9 .
;. g trF,N -_ i - t-b i - I- Ii- ' 'd ' I Ii'i i ' ' l '
- M
,i t - p.' .L .j.f-ft- -#-t .
-k=l: rFid-f #1- -
P I-E ;- Eli' hi ! f 4 / -
!!.: .bi t.t t.hhb <.gP dkhE :t Mh = dhh 4 I lib- 1:hES I! I I h ![ '3
' o 8
,, , ,Li ,pi- . i 8
- i t - i- i t- I . 'ti M
'I i I
'i'/
' - il H
, ., , 9 , 9- t 6 -1 -
1- i ..] i 'r' ! ! l Q
,. ., ... j ;i 4 .e 4 1 -t 44.i r. - . , i ' '/ i
'T7',.7:.; :t .
-Q L .7.- i!- L N-1
- i. -j- -t ~: tl ti ;M I' 8 i/' W 1 4-
- IN-b-bb @s-h- M-dA!T DD- -" N i D ' -b "r U 'b 'h II i8
' d-!f $sEi ~4i:h5 $$EEFilS~E $EM5EDI25 -NMD b' 5'd:: sE5 Fifi .N 2MNE= TEND 45M5555 E5MI5 9" ",.7 ~
. .M. i t----p -_; j g- j _;' ,-
$_4.2 :~_i 12 _' ' _ Z _. ---
-- ----'-~~-'-
t-
-t --
, _ r:' '
T'O-
-H-i
, , ,, , ,. , . , , , ii,i . 1 #
/:
! : ,er ** *
' . , 6i li,t eti
- i ! r. -! t 13 j4 l - } I ^8 M' ' 'I ' '
, , l. l* j f 4-i- ii6-l 1. -l i L- j . ri. 7-j S l -i j Li "- l {-84 ", I lb $ l- I e Mi !! I! " I'
! t 1:l l:(-N l -E : MEh Sd M 1d-l I dN NrE l1 IN ! i NM .i i i ,
5
, ;, o hi. . ,,, i4. v Iiii *, .
6li li* Ie ' l ;i Il [l }l i e 8 y' 8 p 1 -l ! l' I t
, I ?
i) i v i-i ! j.) s F M II i 1f tII I l '-l i
'. il,' [ F( i h 'j ! -lf il ! Ii'l +! I i i-l '
i ; i p!. $ t .;L M illt itl1 -?{ j -N 4-l i I! !*I iI Ji'
!Ii , ';'!
I! '4 Md 1hh l# hi . .i N ih M- il !!!' 1- .,_ . . ..4
- I.M qYd: b,.
e, .. , ldli I !5 ;i ; p! ;3 Ii'b y=,-
' v i'r I-
- C 3 "~lI r,
a b;- -- w -
,i wl He g6
-M i 3- i .o .
r% G ,'
,+t1 -
Qh m' kl!.~O, i . l. M}
-H. r +1'r e -o 4
Y r eb -F i, i d AliAliovaa
.. .. ., . ~ - . . ._ . . . . ._ .
- qx:n ,
= _== ~= m.nm__m,_ .. . :=.,+cmmyg; UNIVERSITY OF FLORIDA TRAINING REACTOR HAZARDS
SUMMARY
REPORT . O S
' i'e[$ %
9- NN 1 2 l? g9g V
$?h[&g Prepared by 4 ,
N J. M. Duncan
$5. 00 per copy A Report to The United States Atomic Energy Commission Division of Civilian Application
[ I I l l Frocn the Depamnent of Nuclear Engineering I College of Engineering Univenity of Florida Gainesville, Florida l October,1958 i l l
#^^^- -
l - - . . . - - - . . . . . . . . _ . . ..
l l
. "***~i !
APPENDIX F . l j ESTIMATION OF EFFECTS OF ASSUMED LARGE REACTIVITY ADDITIONS f i It has been demonstrated repeatedly in the Borax and SPERT reactors that water-cooled, I water-moderated reactors of suitable design may have avery substantial self-protection against 2he effects of reactivity accidents, even in the absence of corrective action by the reactor control 'l sy s tem. This self-protection is provided by the negative steam-void coefficient of reactivity and the negative temperature coefficient of reactivity, both of which can result in important '[.. reactivity reductions as the reactor power rises. The UFTRhas beendesigned with a high degree f:. of self-protection of this type. In this appendix estimates are made of the behavior of the re- hh actor under various hypothetical conditions of excess reactivity addition with no corrective action i;0. by the control system. l$1 The characteristics of the UFTR which determine its behavior during power transients re- f;2:.1 sulting from large reactivity additions are quite similar to, but not identical with, those of the {~ " l Borax I reactor. Its behavior can be predicted most reliably by utilizing the Borax I data with (Mil simple correction factors to convert them to the UFTR
- conditions. $~j The significant quantitative characteristics of the UFTR and the Borax I reactor are com- Y.2 .j pared in Table F-1. f!j.- ,
m::
,k u,
TABLE F-1 3.' . y; = . 1.y . COMPARISON OF UFTR AND BORAX I CHARACTERISTICS Characteristic UFTR Borax I
'h:.
f' "5;
-I Fuel plate " meat" 46 w/o U-Al alloy 18 w/o U-Al alloy :4{-
{-{ (20 per cent (fully-enriched) y enriched) p;;. . ; I4. Fuel plate cladding 1100 aluminum 1100 aluminum &g i8
" Meat" thickness 0. 040 in. 0. 020 in. {W.%c,
.9.Ih
- Cladding thickness 0. 015 in 0. 020 in. I':p:l:
[ lI Coolant-channel thickness 0,137 in. O.117 in. . y-
.. s 's!
.c -
Core volume (approx.) 71 liters 106 liters -- Void coefficient of -0. 21 per cent k/ ' O.24 per cent k/ g@. ' reactivity (calculated per cent coolant void per cent coolant void ['-p
.\- t..
Temperature coefficient -0. 01 per cent k/0C i -0. 01 per cent k/oC I? ' of reactivity (room (estimated) temperature) ,. ," : ' l u, Effective prompt-neutron 1. 4 x 10 -4 sec 0. 65 x 10 -4 sec [ li fetim e (calculated) e
- i. .: ,
I Power ratio in core, 1.63 1.82 . maximum average r .
- 107 - . i
. i I h
> i i-.
E -- w
The U-238 in the 20 per cent enriched fuel of the UFTR introduces a negative Doppier
., coefficient of reactivity estimated to be of the order of 4 x 10-7 k/oC equivalent to . 004 per cent reduction in k per 1000C rise in fuel temperature. Although the Doppler coefficient acts instantaneously and would cause the shutdown of the reactor in case of a reactivity accident, its
.,, effect is not expected to be important because expulsion of the water mode ator will terminate j an excursion before the fuel temperature has risen appreciably.
] In addition to the quantitative differences, the UFTR differs from Borax Iin that the maxi-
;j mum coolant water level is only a few inches above the upper ends of the fuel plates (instead of about 4 ft) and the coolant water, once it has been ejected forcibly from the core by a power j excursion, cannot fall or flow back into the core.
. q.
i
- Effect of 0. 6 Per Cent Excess Reactivity
{ An excess reactivity of O. 6 per cent keff will be available in the reactor if its temperature 1 is abnormally low (nearly freezing). I The addition of all this excess reactivity will cause the reactor to operate at a power such
, that the reactivity losses associated with the temperature increase and the voids formed will d equal the excess reactivity. ,- j If the reactivity is added slowly, after the reactor is critical, the power will approach such Ei an equilibrium level slowly as the reactivity is added. If the reactivity is added suddenly when the reactoris initially auberiticalor at verylow power, the power willat first rise exponentially ,]-
with a period not shorter that. 9. 3 see which is the asymptotic period corresponding to the full
.j excess reactivity of 0. 6 per cent keff. Many experiments with the Borax reactors havedemon- *j strated that for periods of this order of magnitude the transition from the exponential power i t' rise to the equilibrium power level (in which excess reactivity is balanced by temperature and .A steam void coefficients) is a smooth one involving little or no power overshoot. On the basis d of this experience, it can be said that the magnitude of the power excursion whicli results from } ;# the 0. 6 per cent reactivity addition will not depend greatly on whether the reactivity is added suddenly or relatively slowly and in n'either case will it approach a level which would cause a
[y fuel plate to burn out. [. In order to compute the power level at which the reactor will operate after the addition of
.l the 0. 6 per cent excess reactivity discussed in the foregoing it is necessaryto know the water-F7 ternperature coefficient of r eactivity. The relative importance of the two moderators, graphite d ~
and water, in determining the effective neutron temperature introduce uncertainties in the theo-
.i retical computation of this coefficient. The coefficient cannot, however, have an absolute mag-Ii r.itud e less than that of the water-densit/ coeffL:ient of reactivi;y referred to a tempt.rature 'j s cale, i. e. , the coefficient computed on the assumption that:
I d k,ff _ - 6 k,ff . 6A
} dT 6A 6T
; 6P where p refers to the water density and T to the temperature. Note that p is just the negative i of the void coefficient of reactivity. On the assumption that this minimum value is the true value, j a rise of water temperature from near 00C to 800C would reduce reactivity by 0. 6 per centk,ff.
-} The capacity of the reactor-coolant system is such that if the outside air temperature were O C and the average water temperature in the i eactor were 800C, energy would be removed at
! j the rate of 365,000 BTU /hr or 107 kw. Under these conditions the reactor water-inlet tempera-ture would be 600C and the exit temperature, coincidentally, would be 1000C. It is, ther efore, I concluded that if the full available excess reactivity of 0. 6 per cent keft were added to the reactor i on a cold day with the coolant system operating, the reactor would operate at an equilibrium I power level about 10 times higher (100 kw) than its normal maximum with little or no net steam
- 108 -
- 2. ,
d' I
_. . n p roduction. Before reaching the equilibrium power, when the waterin the coolant system would be heated to the equilibrium value, the reactor would operate at a somewhat higher power level and some net steam production might occur. If the coolant were not flowing during the time of excess reactivity addition, the equilibrium power level would be quite low and equal to the heat losses. In no case would the power level approach a value high enough to justify any fearof fuel-plate burnout. Maximum Tolerable Sudden Reactivity Addition
~.
In order to assess the safety factor which exists between the normal excess reactivity avail- ,5 able in the reactor and the excess reactivity necessary for a serious power excursion, it is 75 l I useful to estimate the value of excess reactivity which, if suddenly inserted and not removed by 'gj ; the control system, would raise the maximumtemperature in the hottest fuel plate to the melting g'd point. Such an excursion would damage the reactor core but would not result in any substantial i f. , release of fission products. The first step in the procedure is the estimation of the exponential period corresponding to l17'.G jyf . ;l l the excess reactivity which would have characterized a power excursion of similar effect in fM !
- Borax I. The estimate requires that (1) a relationship be established between the maximum
%,b !
tempnrature of the fuel plate .tnd the enarg) release of the excursion and (2) the energy . elease ae relatea to the period of the excursion. ([:yi For the case of power excursions of short period, with reactor water at saturation tempera- [E '" i tur e, it is shown in Reference 1 that the maximum fuel-plate temperature rise is, to within experimental error, proportional to the maximum energy release of the power excursion. The Q:$. p9 c 1 proportionality was determined to be constant 24.40F per MW-sec.* Measurements of the same type with cold reactor water (the case directly applicable to the UFTR) showed a similar relation- i ship but with a proportionality constant of only about 100F per MW-sec (Reference 3). The g I e, ' difference is not an unreasonable one since the subcooled water represents a more effective heat sink than the saturated wate'r. However, the experiments with the saturated water were carried yl,$ ,7 ,i to short periods in the range of interest whereas the subcooled experiments were limited to [g-longer periods. Therefore, more conservative saturated water data will be used. To raise the maximum temperature of the fuel plate from the temperature of boiling water to the melting [-Q point of alurr.inum, a temperature change of approximately 10000F, would requireta power ex- C,. F I cursion with a total energy release of 10000F or 41 MW-sec. #-
- 24. 40F /MW -sec lE i
According to the data of Reference 3, replotted in Figure 6. IA, a "subcooled" power excur-sion of reciprocal period 150 sec-1 wo61d give an energy release of 41 MW-sec in addition to the energy necessary to raise the fuel plate temperature to the saturation temperature of water. It g ;} is therefore concluded thata power excursion of period at leastas short as 1/150 sec (6 7 millisec) could have been tolerated by Sort , I with subcooled water without melting at the hottest point in
' ::H {
] the f act plates. h" [r ( f Experiments of the Borax and SPERT types have not been 'made with reactors having widely [, f different neutron lifetimes. The general evidence of the experiments however supports the sup- ,T$.' position that of the three related variables--neutron lifetime, ex' cess reactivity, and exponential period--which characterize the neutron physics of a power excursion, it is the exponential period 'M LJ ' which determines the total energy release and the temperatures attained. The excess reactivit'f __ g I and neutron lifetime have large effects only as they jointly determine the period. This supposi- '
*Actually, the energy dataof Reference I were revised in Reference 2 because of later and bettec
- calibrations of the instrumentation. The numbers above are taken from the later (more pessi-mistic) data, l j
r E i l .! I. I a
- 109 - '
l
, l 1
1
~f
; tion is consistent, for example, with the observations that the total energy transferred to the l j coolant water during a power excursion is many times the amount which would vaporize enough l j - water to compensate for the excess reactivity, and that the actual reactivity reduction which I
occurs during the excursion is much larger than the initial excess reactivity. The extension of-4 the Borax results to the UFTR.is made on the basis of this evidence.
, ]y It is convenient, first, to treat only the effects of the slightly greater fuel-plate spacing and N the slightly lower void coefficient of reactivity of the UFTR relative to the Borax I. In ~o rmation p will also be drawn from the Borax II experiments. The Borax II reactor differed from Boraxt U
in that the coolant-channel thickness was greater in the ratio 0. 264 in. =, 2. 2 6 and that the cal . 0.117 in. culated void coefficient o'f reactivity was lower in the ratio 0,10% ketf/% void ,1 ,0.416
, 0. 24% k,((/% void 2.4 i
Both of these differences would be expected to cause a higher energy release per fuel plate in l borax II than in Borax I for a power excursion of given period. The measurements made with 1 subcooled water at periods down to 23 millisec showed that the energy release per fuel plate in Borax II was between 1. 7 and 2. 0 times that of Borax I, with the smaller ratio applying to the
, shorter periods (Reference 2). Therefore, it seems quite conservative to assume, in the case I
of any two reactors, (1) and (2), of the Borax type having a ratio of fuel plate spacings, St /S2 . j and a ratio of void coefficients of reactivity, C /C2, 1 that the ratio of energy release per fuel 4 plate for a subcooled power excursion of given period will be no greater than E 2 /El . S 2/S1 or j EZ /El C 1/CZ whichever is the larger. For the UFTR and Borax I the ratios are: I l SUF 0.137 CBo 0.24
. = 1,' 1.14 i SBo 0.117 = i CUF = 0. 21 =
It is concluded, therefore, that a Borax reactor having a coolant-channel thickness and a void 1 J coefficient of reactivity equal to those of the UFTR would release not more than 1.17 times as I much energy per fuel plate as Borax I. The limiting non-melting period for such a reactor woul'd be that which in Borax I gave an energy release of 41/1.17 = 35 MW-sec. The period obtained from Figure 6. IA, corresponding to a total energy release of 35 MW-sec, is 7. 7 millisec. The remaining difference between Borax I and the UFTR is in the composition.of the fuel ! plates. The UFTR plates are thicker; their uranium-aluminum alloy has a somewhat lower con-
- ductivity because of the higher uranium concentration, and their aluminum cladding is thinner.
In corr paring the behavior of different fuel plates, it must be reccpiz :d that the total energy release of the power excursion can no longer be considered as a definitive variable because a large fraction of the total energy released is stored in the fuel plate during the important stage I of the reactor shut-down. For example, a reactor composed of fuel plates of high heat capacity undoubtedly will experience a larger total energy release, but not necessarily a higher maximum
; temperature, during a power excursion of given period, than a reactor having plates of lowheat capacity.
From examinationof the Borax results it seems clear that two distinct phases of the reactor shutdown process occur consecutively and that both may be important in determining the maxi-mum center temperature of a fuel plate. The first phase covers the interval before animportant amount of boiling occurs at the fuel-plate surface. During this interval the heat loss to the water is small and the important consideration is evidently the ratio of fuel-plate surface temperature (which determines the start of boiling) to center temperature. For periods in the range under consideration this temperature ratio is theoretically not far from unity (0.76 minimum for a 10-millisec period in Borax I). Experimentally the temperature ratio was unity for periods down to 5 millisec in the B;/ax I measurements. Since the total effect is smalland since the tempera-ture ratio for Borax and UFTR fuel plates should not be much different, the thinner cladding will I l l l - 110 - 4 _ _ , _ , . _ _ , y_ - _ _ _ _ _ - -~g, _ _ _ , _
--v w- -- - , - - - - - _ , . , - _ . _ _ _ , ,. , , _ , _ , . - _ , _ , ,,,,,_._____,,__g,,3,_ _ , , _ , . , .
tend to balance the effect of the poorer " meat" conductivity. It is concluded, therefore, that there will be no important difference in fuel plate performance during this initial phase of the excur sion. , The second phase of the powe r excursionbegins when a significant rate of boiling is estab-lished at the plate surf ace. R activity and consequently generation are reduced at a rate which t- ust be a function of the rate at which heat can be transferred into the boiling water. At the same time the transfer of heat into the, water removes heat from the fuel plate and limits its . temperature rise. The important characteristic of the plate during this phase of the excursion '. is the heat flux which it can supply to the water for a given temperature difference between the #.- plate center and surface. A figure assumed to be roughlyindicative of the relative performance ^[ _ or merit of fuel plates during this phase is the ratio of heat flux to temperature difference under steady-state conditions. This ratio (figure o' merit) will overemphasize the difference g:5.. between fuel plates since the temperature distribution in the plate will be more peaked during ;jg.'ll a steady-state conduction than during conduction when the general temperature level is rising. p
;r;y The ratio of these figures of merit for Borax 1 and for the UFTR is
- h. r Heat Flux i,.,c y Mr:- s -
UFTR i!M.T
= 0.82 !E.
i:*k.i
" Heat Flux ~ lN
. D'al-
-AT"~* kM
- Borax !b.". y A conservative procedure would be to apply the above f actor to the permissible total energy hc. , '
of excursion on the Borax 1 curve. At the same time, however, the difference in gross maxi- ;L,M mum to average power ratio for the two reactors should be taken into account since it is the EW. tempe rature of the hotte st point in the hotte st fuel plate which is b'eing conside red. The powe r ratio for the two reactors is b
'p , u t.;:2 M.6 :
M ax p 7.. Ave 8h2 + I B o r ax = 1. 82 = 1.12 @;n Max 1.63 !T.a :- l-f.L .
^**UFTR IN[:' '
The combination of these two factors reduces the permissible equivalent energy of the h5
$T' Borax-type excursion to ( [
35 x 0.82 x 1. 12 = 32 M W- se c.
\ ir
, a s
n
;m The corresponding exponential period from Figure 6. l A is 8. 3 millisec. It is, therefore , !b.,
concluded that the UFTR will tolerate a power excursion of period at least as short as S. 3 l. millisec without the melting of any part of any fuel plate. The exces s reactivity co rre spond- , 7.~ ing to this period is 2.4 pe r cent keff. :: Succe s sive Power Excur sions It is typical of the Borax and SPERT reactors, unless the exce ss reactivity is removed , by external means, an initial power excursion which terminate s itself by expelling water f rom .
/
- 111 -
e
y- -2 6
+ t
?
the reactor core will be followed by subsequent excursions as the water f alls and flows back i into the core. An exception to this behavior eccurs when the initial excursion is violent enough j to cause a permanent loss of reactivity by th owing a large amount of water completely out of j the reactor tank. In the UETR the total queenty of water in the core is small, the submergence 3 of the core is small,and baffics above he core are so arranged that any water splashis directed to the outside so that it cannot return to the core. Consequently, even a relatively mild power r .; excursion (e.g., one having an exponential pe riod of from 20 to 30 millisec) in the UETR should
' .$ result in permanent self-inluced shutdown of the reactor. By the se same design features the '[: possibility of large successive power excursions, such as those studied in the SPERT project, ;' resulting from the ramp addition of excess reactivity is eliminated. It can be anticipated that j the UFTR will be safe against quite large ramp additions (larger than 2.4 per cent k,ff) pro-vided or.ly that the ramp rate is not so rapid as to add an excess reactivity of more than 2.4 per cent k,ff before the reactor power reaches a high level. To exceed this limit the ramp l rate would need to be of the order of 1. 0 per cent keff per second or larger.
4 q a
'.1 REFERENCES 4
9 L: gi F.I Die trich, J. R. , and D. C. Layman, " Transient and Steady State Characteristics of
. a Boiling Reactor. The Borax Experiments, 195 3, " AECD- 3840, Argonne National
(: Laboratory, February, 1954
,U ,
F. F.2 Dietrich, J.R. , "Expe rimental Determinations of the Self-Regulation and Safety of
.;; Ope rating Water-Moderated Reactors," A/ Conf 8/P/481, Inte rnational Confe rence } on the Peaceful Uses of Atomic Energy, June 30, 1955.
y F.3 Dietrich, J. R., "Expe rimental Investigation of the Self-Limitation of Power During fi Reactivity Transients in a Subcooled Water-Moderated Reactor. Borax-1 r xpe rt-ij
, ments, 1954," AECD-3668, Argonne National Laboratory, August 17, 1955, a '.i A s J
4 l l I i u I J
}
- 112 - .._
b ' c
, CCNTENTICN VI y .L 0 d A RESFCNSE TO URC STAFF'S ASSERTED MATERIAL FACTS s Qff g 1 5 04N fp 5 "The f~ di for g1 Ar is 410xC.F. 10 g.pCi/ml." Part 20 Appendix 3 release limit for unrestric'id/3r'i
\:;4 4l W pk-NOT DISPUTED r., f ,
l s ,y ' I
- 2. "The Ar releases from the UCLA reactor into unrestricted area:s are 3.8 x 10-9 pCi/ml."
DISFUTED (Foster declaration for VI, 24-8; Lyon declaration.for VI, 2 20 Pulido declaration for XV, F 4) 3 "The UCLA radiation monitoring system data has been verified by an environmental monitoring program." DISPLTED i (Fosterdeclaration,13-26 Lyondeclaration,23-20) 4 "The' most conservative interpretation of the UCLA environmental menitoring program is 30 mrer/yr from reactor radiological releases into unrestricted areas."
- DISFUTED (Fester declaration,I 10-11,13-15,24-26) 5 "A dose of 30 mrer/yr. is 6% of the permissible level in 10 CFR 20.105(a)"
i NOT DISPUTED i Counterfact: A dose of 30 mrer/yr in unrestricted areas of UCLA ! would violate 10 CFR 20.l(c) and would be the equivalent of members of the public, without their permission or'snowledge, receiving the equivalent of an additional chest X-ray per year without any medical need therefor (Monosson declaration I le; Lyon declaration, I 16-17 10 CFR 20.l(c);
- ICRP Publication 22)
- 6. "The radioactive emissions from the UCLA research reactor could not be significantly reduced by additional stack height."
DISFUTED (Pulido declaration for XV, I3-5,7-8,34; Earch 13,1975, Response by ' UCLA to Notice of Violation, second-te-last page, indicating raising stack the 17 feet to the ori6 1 nally required height would increase dispersion and decrease radioactive concentrations in public areas by a factor of five).
_ _ _ _- - . _ . - , _ . .~ - - . _ _ _ _ _ - _ _ _ _ _ _ _ _ - - _ _ _ . . -
] VI-2 1
EESPCNSE TO UCLA'S ASSERTED MATERIAL FACTS ,
- 27. " Eased en conservative assumptions the maximum argon-41 concentratien i seen in the past five years at the Fatheratical Sciences air intake is
- 2. 6 x 10-'> pCi cr). "
DISFUTED. l (Fosterdeclaration,18 SAR to Amendment 10 to UCLA license; Reg Guide 1.111)
- 28. "The highest radiation level on the unrestricted rooftop areas adjacent to the reactor building exhaust stack does not exceed 22 mrem per year above background."
i DISFUTED (Foster declaration, I 10-21; plus the TLD raw data) 29 "The radioactive emissions from the UCLA reactor have been reduced to a
, level that is as low as reasenably achievable."
DISPUTED (Foster declaration,13-26; Pulido declaration for XV, I 4-11,34 Lyon declaration, 14-5,8,16-17,20) ) ) a r s i
, , , , - - - - - - - - , - - , , , - - --,~.-n--- - - - - - - . - --- - , - - - - - . - - - - - - , - - -
-- - - ,}}