ML19318B823
| ML19318B823 | |
| Person / Time | |
|---|---|
| Issue date: | 03/31/1978 |
| From: | Siegel B Office of Nuclear Reactor Regulation |
| To: | |
| References | |
| NUREG-0303, NUREG-303, NUDOCS 8006300122 | |
| Download: ML19318B823 (26) | |
Text
r i
NU REG-0303 EVALUATION OF THE BEHAVIOR OF WATERLOGGED FUEL ROD FAILURES IN LWRs B. Siegel p
M,.8
\\y(dsyji 9.....-
Office of Nuclear Reactor hegniation U. S. Nuclear Regulatory Commission 8 0063 0 0 l1
I l
I i
)
I Available from National Technical Informu
- n Service Springfield, Virginia 22161 Price: Printed Copy.4.50 ; Microfiche $3.00 The price of this d cument for requesters outside of the North American Continent can be obtained from the National Technical Information Service.
l
..m m
NUREG 0303 EVALUATION OF THE BEHAVIOR OF WATERLOGGED FUEL ROD FAILURES IN LWRs B. Siegel Manuscript Completed: November 1977 Date Published: March 1978 Division of Systems Safety Office of Nuclear Reactor Regulation l
U. S. Nuclear Regulatory Commission Washington, D. C. 20555 1
Abstract A summary of the available information on waterlogged fuel rod failures is presented. The information includes experimental results from waterlogging tests in research reactors, observations of water-logging failures in commercial reactors, and reactor vendor assess-ments.
It is concluded that (a) operating restrictions to reduce pellet / cladding interactions also reduce the potential for waterlogging f.ilures during transients, (b) tests to simulate accident conditions produced the worst waterlogging failures, and (c) there is nc apparent threat from waterlogging failures to the overall coolability of the core t.
to safe reactor shutdown.
Tablo of Contents Pave I.
Introduction...............................................
1 II.
Results From Experiments...................................
2 III.
Reactor Waterlogging Failures..............................
7 A.
Western New York Research Reactor......................
7 B.
Boiling Nuclear Superheater Reactor....................
8 C.
SPERT ( Capsule D:-iver Core) Reactor Failures........... 8 D.
Commercial LWR Fuel Failures...........................
10 E.
Irradiation With Intentionally Defected Fuel Rods......
11 IV.
Vendor Assessment of Waterlogging Failures.................
13 A.
Westinghouse...........................................
13 B.
Combustion Engineering.................................
13 C.
Babcock & Wilcox.......................................
14 D.
General Electric.......................................
14 V.
Discussion.................................................
15 A.
Transients............................................
16 B.
Accidents..............................................
16 VI.
Conclusions................................................
19 VII.
References................................................
20
1 I.
Introduction Experience in operating reactors has shown that a very small number j
l
(<1%) of the fuel rods in a typical core develop leaks. While rods with minor defects can remain in the reactor for sustained periods of time without further degradation, the potential exists for water to seep in and saturate the cladding interior and pellets. When this happens the rod may simply " dry out" withou't complications on return to power and
[
high pellet temperatures. However, if a rapid power rise should occur when fuel rods are waterlogged, the rods may not be able to vent steam 4
l rapidly enough and the resulting pressure could rupture the cladding.
l 4
The NRC guide entitled " Standard Format and Content of Safety Analysis Reports for Nuclear Power Plants" (J), states in Section 4.2.3 t
that the fuel system design evaluation provided in applications for nuclear power plant licenses should address the potential for water-logging rupture. These evaluations usually reference results from l
earlier SPERT reactor experiments, but the evaluations have been rather i
limited. To obtain more information on waterlogged fuel rod failure, a request was made to the Nuclear Safety Information Center (NSIC) for-all references related to fuel element waterlogging. The NSIC listing provided 22 citations on this subject (2.).
This information combined 4
with references from the vendors
- safety analysis reports and the recent experimental results from the Japanese Atomic Energy Research Institute (JAERI) from their Nuclear Safety Research Reactor (NSRR) forms the basis for the present evaluation.. This report thus discusses experi-
. mental results from waterlogging tests in the SPERT and NSRR reactors e
i l
4
- - - ~, -
w
-_---...e,-
7-+-.-. m
1 ;.
observations of waterlogging failures in research and commercial i
reactors, reactor vendor assessments of waterlogging failures and an independent evaluation of the consequences of waterlogging fuel rod failures.
II.
Results From Experiments Reactivity insertion accident 'RIA) tests on waterlogged rods were performed in the test section of the Capsule Driver Core (CDC) as part of the Subassembly Test Program at SPERT during 1969 and 1970 (2,1).
Additional tests on waterlogged fuel rods are currently being performed by the Japanese Atomic Energy Research Institute (JAERI) in their nuclear safety research reactor (NSRR), which is a TRIGA annular core pulse reactor (1,1). A summary of test results from both of these programs is provided in Table 1 for Zircalcy clad fuel rods. A direct comparison of all of the test results was not possible since the rate of energy input (reactor i
period varied between 1.5 to 5.5 msec) and the total energy input were not the same. For comparison the reactor period for the worst reactivity 1
insertion accident under normal operating conditions for a typical LWR power reactor (one which results in an energy insertion of 280 cal /gm) would be greater than 10 msec. However, under certain extreme PWR condit-ions (end of life, zero power, and artificially high rod worths) a reactor
~
period as short as 2 msee could conceivably be reached, i
SPERT tests, unless otherwise noted, were performed on rods that were waterlogged by first evacuating the fuel rods, allowing water to enter, and then sealing the rods. The fuel rods used in these SPERT tests had 1
T J ;
i rod diameters between 0.250 to 0.562 inches, fuel lengths between 5 to 9 incl.es, fuel rod enrichments between 5 to 8 percent, and cladding thickness
]
from 14 to 32 mils. Tests 555, 556, 558 and 559 (3) were performed under l
Very similar test cond'itions (total energy depoci' ion of 225 cal /gm and reactor period of 5.5 msec). The only differences between these tests were the use i
i of annealed cladding in two tests and cold worked cladding in the other two.
The differences in heat treatment of the Zirealoy cladding probably contri-1 j
buted to the small differences in failure threshold energies (160 cal /gm for annealed cladding compared with 150 and 120 cal /gm for cold worked cladding f
tests).
Tests 552 and 547 (3) were run with a shorter reactor period and a 3
i much higher total energy input than the four tests mentioned above. The 1
failure threshold for test 547 was much lower than for test 552, although the maximum internal pressure was about the same. The heat treatment of 1
the cladding was probably responsible for the differences in these results also. An important difference between these two tests and the four tests 4
discussed previously was in the quantity of fuel expelled. For tests 555 j
thru 559, which had a total energy deposition of 225 cal /gm, only 1 gm of i
fuel was expelled whereas for tests 547 and 552, which had a total energy deposition of 430 cal /gm, all of the fuel was expelled. For comparison it is noted that the calculated energy depositions in commercial power i
reactors are limited to 280 cal /gm (1) for the postulated rod ejection accident, which is the worst reactivity insertion accident considered.
i SPERT test 496 (Table 1) was run at a total energy deposition of i
220 cal /gm, which is about the same as for tests 555, 556, 558 and 559, yet the failure threshold energy for this test was only 60 cal /gm, which is similar to the threshold energy for test 547. Test 496 had the same short reactor period of 3.0 msec as test 547.
The JAERI program to evaluate the effects of waterlogged fuel rod failures is still in progress. The results reported here are early results from this program and are lacking some details and post-irradiation examination evaluations. Many more tests under varying conditions are planned.
In Table 1 are listed the results of seven waterlogged fuel rod tests (401-3, 401-3b, 401-4, 401-4b, 401-4c,401-5, and 401-6) performed in the NSRR. The 0.481-inch diameter fuel rods in these tests were nearly completely filled with water and sealed. The failure threshold energy for these tests was between 100 and 130 cal /gm and the reactor period was 1.5 msec. The total energy deposition was less than 147 cal /gm for all but test 401-6, which had an energy deposition of 216 cal /gm. The failure threshold energy determined from test 401-5 was 115 cal /gm.
Although these test v3re performed under different conditions from those in the SPERT program, the results appear to be consistent when the differences are taken into acecount. For example, the failure thres-holds for the JAERI tests were generally lower than for the SPERT tests.
This would be expected because of the faster rate of energy input (1.5 msec.
reactor period for JAERI tests compared with 3.0 and 5.5 msec. periods for the SPERT tests).
Comparison of the data from all the tests discussed indicates that
. ~ _.
~
1 significant variations in the failure threshold energy can occur depending on the reactor period and the total energy input. However, there are several conclusions that can be drawn from these tests that are important j
from a reactor safety standpoint: (1) the total nuclear-to-mechanical energy J
conversion was less than 0.2%, and (2) the maximum caps.!1e pressure was less than 1200 pai..The nuclear-to-mechanical energy conversion provides i
an indication of the potential for failure propagation. The lower the j
_value, the more that is dissipated as thermal energy. Based on SPERT tests,
]
to achieve nuclear-to-mechanical energy conversions of 1 to 2% r. quires a
I severe fragmentation and dispersal of the fuel into the coolant and this only occurs at considerably higher energy inputs (>600 cal /gm).
j Waterlogged tests on fuel rods with stainless steel cladding have l
also been performed (1,1,8).
The results from these tests were consis-tent with the results obtained using Zircaloy cladding. The annealed
. cladding resulted in a higher failure threshold energy (greater than 280 cal /gm) than cold worked stainless steel cladding (110 cal /gm) and j
the nuclear-to-mechanical energy conversions and capsule pressures were j
not appreciably different from the test resrlts using Zircaloy. The j
comparisons in Table 1 were intentionally limited to tests using Zircaloy cladding because of the differences in failure thresholds between Zircaloy and stainless steel and the fact that most commercial reactors use Zircaloy cladding.
The tests previously described were performed on fuel rods com-pletely filled with water and sealed. Tests have also been performed l
1
.L
6-on rods that were intentionally defected prior to testing to simulate actual in-reactor conditions; the defect size, type of defect, defect location and quantity of water within the rods were varied. NSRR test 402-3 (see Table 1), in which the radial gap was filled with water, was tested at similar conditions to tests 401-3 and 3b, but did not fail.
Tests 411-3 and 421-3, which had small initial defects, were also tested at conditions similar to 401-3 and 3b.
The fuel pin in test 411-3, which had a 15-mil hole in the upper part of the cladding, ruptured, whereas the fuel pin in test 421-3, which had a similar defect in the middle part of the cladding, did not. This indicates that the location of the defect ay affect the failure threshold. However, as would be expected, the results were no worse than for sealed rods containing water, and in two out of the three tests the results were less severe.
Tests have also been performed in NSRR with fuel rods at two different energy depositions (approximately i45 and 217 cal /gm) in which the void regions of the fuel rod were filled with different quantities of water ranging from 11 to 100% of the total void volume. Unpublished.lERI results shown in Table I show that the failure thresholds were independent of the quantit/ of water in the rods in the range tested.
Two tests were run in CDC at SPERT (S) on fuel rods that were in-tentionally defected with slits 5 in. long and 30 mils wide with the center of the slit 12 in, from the bottom of a 30-in. fueled section.
One rod was filled with 9 gm of water and the other with 1 gm of water.
At an energy input of 18 cal /gm, both rods opened to a width of 3/16
4
_7_
in. at the center of the defect and expelled practically all the fuel, although the slits did not elongate.
Tests in the CDC have also been performed on stainless steel clad fuel rods with intentional defects (S.).
Two of these tests were per-formed on fuel rods with 0.040 in, holes located 2 in. from the bottom of the rod and with water filling approximately 18 in of the 30-in.
long fueled region. In one of these tests, a companion rod with no defects was tested alongside the defected rod. The energy input to the fuel rods in both tests was approximateli 40 cal /gm at the time of failure and the reactor period was 8 maec. The rupture length of the waterlogged rods was 4 in., starting between 6.5 and 7 in, from the l
bottom of the defected fuel rods in both tests, and the capsule pressure pulses were between 5000 and 6000 psi in both tests. No damage to the dry companion rod was observed.
III.
Reactor Waterlogging Failures Failures that have been attributed to waterlogging have occurred in j
the fuel rods of several reactors. These failures were unintentional, in contrast to those described in the previous section. A brief description of these events and their consequences is given below.
A.
Western New York Research Reactor A fuel rod failure occurred in the Western New York Research Reactor, which'is a pulsed reactor (ja,jj). Following the las? of a series of pulses, high primary coolant activity was detected. The last nine pulses 0
in this series ranged from 1000 to 1400 MW peak power with a maximum energy of 35 MW sec. Examination of the core revealed one of the fuel t
--..--,e g
, rods had failed. The failed rod had a cladding split over an axial length of approximately 1.9 in and the width of the split extended over approximately one third of the circumference. The location of the frac-ture was at the approximate location of the hot spot and was attributed to waterlogging. The three fuel rods adjacent to the failed pin showed no visible damage or distortion of any kind. No other fuel rods in the core were found to be damaged. Approximately three pellets were lost from the failed rod, and metallurgical examination of the failed clad-ding placed an upper limit of 856)C on the cladding temperat'.re; this temperature is well below the melting point of Zircaloy 2.
Examinations were perform +d to locate the original defect, but it could not be found.
It was concluded that the defect was in the vicinity of the burst section.
B.
Boiling Fuclear Superheater Reactor (BONUS)
The BONUS boiling water reactor had a power transient when the steam flow was reduced for about 2.5 min as the result of an operator error (12). This transient resulted in the failure of two fuel rods.
The diameter of one rod increased a maximum of 15% in the unfailed region and the cladding had ballooned along its entire length. An 11.5-in.
long rupture occurred near the bottom of the rod. The other fuel rod had a maximum diameter increase of 8%, but did not rupture. Leak testing of this rod disclosed two small pinholes in the weld between the cladding and support cap. Based on the evidence it was
.uded that both rod failures were due to waterlogging during the transient.
C.
SPERT (Capsule Driver Core) Reactor Failures
h,
fuel rod ruptures occurred in the reactor core (not in tests rods)during a test series to determine the operational range of the Capsule Driver Core i
(13). The fuel was comprised of UO2 powder compacted in stainless steel 2
cladding. One fuel rod ruptured per test in eight of the nine tests that i
resulted in fuel rod failures. In the other test, three fuel rods ruptured and a rod adjacent to a ruptured rod was sheared at an intermediate grid.
1 Two to ten additional rods were bowed 0 3 in, or more during each event.
The fuel rod failures did not result in sufficient damage to prevent continued operation of the reactor and safe shutdown. The energy at the time of rupture was approximately 50 cal /gm in all cases. The i
failures were attributed to waterlogging because of the similarity of j
the rupture to failures known to be caused by waterlogging. The typical rupture resulted in a cladding split 10 to 20 in. long and expulsion of approximately 50% of the fuel. Metallurgical examination indicated that cladding weld defects, large enough to permit water to enter the fuel rod, i
l were present in all of the failed rods.
i Additional instances during which fuel rod failures occurred in the cperation of the CDC were reported in reference 9.
One CDC fuel rod ruptured during each of three transient tests associated with lead rod and water-l logged fuel rod tests in the CDC test space. The energy insertion at the o
time of rupture was between 30 and 50 cal /gm. The description of the condition of the ruptured rods and the adjacent rods as taken from reference 9 is given below:
"Each of the three failed fuel rods in the CDC ejected a large fraction of their fuel through openings generated 4
-in th?ir cladding during failure. In the 5.1-msec-period. test, an 8 inch long split was observed in the 1
ruptured rod cladding that started about 9 inches above the bottom of the fuel rod. This fuel rod failure caused six neighboring fuel rods to bow. The fuel rod that ruptured in the 4.6-msec-period transient had a 21-inch-long split that started in the cladding about 9 inches above the bottom of the fuel rod. Two fuel rods adjacent to the ruptured rod were bowed.
The bottom section of the ruptured fuel rod contained no fuel, and the top section contained about 30 inches of fuel. The third ruptured fuel rod (4.3-maec-period test) had a split about 24 inches long, starting 6 inches from the bottom of the fuel rod, and three adjacent fuel rods were bowed.
The ruptured fuel rod lost essentially the same quantity of fuel as the ruptured fuel rod in the 4.6-msec-period test, retaining only 30 inches of fuel in the top section of the rod."
Metallurgical examination performed on the cladding from one of the fuel rods showed weld defects that could have permitted water to enter the rod and cause the waterlogging failure. The appearance and consequences of these failures were very similar to those failures (in CDC) that were previously described and reported in reference 13.
D.
Commercial LWR Fuel Failures Fuel rod failures that may be attributable to waterlogging were identified in Point Beach Unit 1 and reported to NRC by the utility, Wisconsin Electric Power Company (J4).
During a refueling outage, two fuel rods in one assembly were found to have a degree of failure more severe than previously found in any other licensed PWR. Sections of the cladding and fuel pellets were missing in some axial locations, and at other locations cladding splits or holes and missing fuel pellets were observed. The failures were attributed by the Wisconsin Electric Power Company to the following scenario.
Bypass flow through a stitch weld
~
in the baffle plate has been observed in foreign plants and probably occurred at Point Beach. This resulted in water impingement on the corner or near-corner fuel rods of the assembly at this position and probably caused vibration and fretting wear of one or more tael roda. The assembly in which the failures occurred was in this core location during cycle 2.
In cycle 3 the fuel assembly was moved to a core position in which it was subjected to much higher powers. The escalation to power at the beginning of cycle 3 was at a much higher rate than presently used (approximately 40% of full power per hour was achieved in one hour) and failures were observed at a power level between 40 and 50%.
It was presumed that these two rods became waterlogged during the shutdown and then ruptured on the relatively rapid return to power. Westinghouso, on the other hand, has reported that the cause of these failures during the startup was probably due to pellet /
cladding interactions. Although there are differing opinions regarding the failure mechanism, the incident has been discussed here because it is the only known event in a large commercial LWR in which waterlogging failures might have occurred. Examination of the fuel rods adjacent to the failed rods did not reveal any adverse effects.
E.
Irradiation With Intentionally Defected Fuel Rods An irradiation test on an intentionally defected fuel rod was per-formed in the L-12 test loop in the ETR reactor as part of the Bettis Laboratory LWBR/LSBR fuel development effort (j5,). The test rod had Zirealoy-4 cladding and contained four fully enriched fuel materials in 1
four pellet geometries.
The fuel materials were uranium dioxide in com-I bination with thorium, zirconium and calcium oxides, and uranium dioxide, which was used for comparative purposes. The defected rod had a 6.2-mil diameter hole located in the plenum region. The fuel rod was irradiated from June 1904 until March 1965 and received a total exposure of 112.4 ef/ective full power days (EE PD) during this time. After this period the rod was removed for further examination because of progressive increases in cladding diameter in the plenum region observed during earlier in-terim examinations (0.022 in. during the previous examination). Follow-ing the last examination it was reinserted in the loop on October 9, 1965.
At 2.' hours after startup of the ETR a large activity burst was noted.
At the time of the activity increase the ETR power level had been in-creased from 60 MW (34.3 percent of power) to 80 MW (45.8 percent of power). Total fuel rod exposure at the time of failure was 121.2 EFPD.
Examination revealed that the plenum region had developed a cladding split 1.5-in. long.
The densities of the fuel pellets were low (approximately 83% of theo-retical density).
It has been postulated that the presence of considerable open porosity in the fuel permitted water to permeate the open porosity and, upon escalation to power, caused the gross fuel fracturing that was observed from gamma scans and neutron radiographs. It was further postulated that some of these fractured particles periodically became trapped in the defect hole causing closure of the hole and, because of increased internal pressure from entrapped water, led to the progressive diametral increase in the plenum region. Examination of the fuel rod produced the follow-ing information:
(a) Neither excessive hydrogen pickup nor the presence of solid hydride was observed.
(b) Diameter increases and cladding thinning were both observed in the plenum region.
(c) Cladding thicknesses of 0.013 to 0.016 in., compared with the 0.0177-in. original wall thickness, were observed.
(d) Maximum diametral changes of up to 0.045 in. had developed.
(e) Gamma scans indicated some fuel had been lost.
(f)
Neutron radiographs showed fuel accumulations at the top of the plenum and confirmed some loss of fuel.
(g) Maximum fuel rod bow of 0.015 in, was observed.
IV.
Vendor Assessment of Waterlorfing Fa_ilures All the reactor vendors have addressed the effects of waterlogging failures in Section 4.4.3 of their safety analysis reports for standard reference designs.
A.
Westinghouse In the RESAR 3S safety analysis report (16) the potential for waterlogging failure resulting from water entering a defect in the cladding is acknowledged.
The CDC tests at SPERT (4,) and the experience in the Western New York Research Reactor (1Q,11), coupled with the fact that normal operational transients in Westinghouse piarits are limited to about 40 cal /gm-min in the peak rod, are cited as the bases for not expecting fuel rod failure propagation.
B.
Combustion Engineering In the CESSAR System 80 safety analysis report (ll) the potential for water-logging is acknowledged if a combination of a small opening in the cladding, time to permit filling of the fuel rod with water, and a rapid power transient occur. However, since the startup rate after shutdown is controlled, C-E believes the potential for rupture is " minimized."
C-E cites the CDC tests at SPERT (2,1) and states that, since these test rods were filled with water and sealed and since the transient rates used in these tests were several orders of magnitude greater than those allowed in normal operation, the consequences in a commercial reactor would be even less severe than those observed in the CDC tests.
C-E concludes that normal power transients in waterlogged rods are not likely to result in cladding rupture, and, even if rupture did occur, it is highly unlikely that fuel would be expelled or that damage to adjacent fuel rods or fuel assembly structural components would occur.
C.
Babcock & Wilcox B&W in the safety analysis report for BSAR-205 (j8) states that pressure pulses from the rupture of a waterlogged fuel rod will cause no significant damage to the rest of the core. B&W cites the tests run in the CDC at SPERT (1,1,29) as a basis for this conclusion.
D.
General Electric In the safety analysis report for GESSAR (22) GE acknowledges that water-logged fuel elements can have lower failure thresholds than r.on-waterlogged fuel during rapid reactivity excursions from the cold condition. CDC tests at SPERT (2,3) are cited.
It is postulated that the resultant energy release and pressure pulse would be much less for a waterlogged rod than for a non-waterlogged rod that exceeds damage threshold because the failure threshold is much lower for a waterlogged rod.
V.
Discussion Most of the waterlogged tests have been performed on fuel rods completely filled with water and sealed.
It would be expected that the fuel rods with defects would rupture at higher energy inputs than rods that were sealed with water inside because the water or steam could escape through the defect and prevent a pressure buildup within the rod.
Based on the experimental results described in Section II, the rupture threshold is not only dependent on the size of the defect, but also on the location of the defect and the operating conditions at startup.
The test results thus illustrate the difficulty in predicting the failure throchold of waterlogged rods because of the many uncertainties involved.
The Lest results on sealed rods completely filled with water should provide a bounding failure threshold for any given rate of energy input and total energy deposition. However, such results may not be overly conservative given the potential for defects to become plugged and the apparent insensitivity of the failure threshold to the quantity of water in the rod.
Waterlogging failures of interest can be separated into two basic categories:
(1) those associated with relatively slow transients repre-sentative of normal and abnormal operational transients, and (2) those associated with fast transients representative of accidents. For both types of events the worst situations are reactivity insertions that take place either f rom cold shutdown or hot standby conditions. With the reactor in a shutdown condition (i.e., with relatively cold fuel pellets), water could be forced into the defected fuel rod due to a pressure differential.
Subsequently, if a rapid reactivity insertion should take place, the water J
may not be able to escape from the fuel rod through the original defect rapidly enough"to prevent cladding rupture.
A.
Transients The waterlogging fuel rod failures associated with normal and abnormal operational transients have been rare and are less severe than those asso-ciated with accidents because of the lower energy input. The failures in the BONUS boiling water reactor and the questionable waterlogging failures in the commercial Point Beach Unit 1 (See Section III) are the only known events that could be placed in this category.
In both of these events no failures in adjacent rods were observed. The potential for water-logging failures during normal operation is further reduced in commercial reactors because of vendor operating recommendations to the utilities.
These operating recommendations were initiated to reduce pellet / cladding interaction failures, and the recommendations restrict the rate of power increase until the fuel is preconditioned. The restrictions are always in effect after startup following a long shutdown for refueling. The potential for fuel rod waterlogging is also greatest during this period, particularly if the defect is small. Thus the restrictions to reduce PCI failures also reduce the potential for waterlogging failures.
B.
Accidents The greatest potential for damage from a waterlogging failure in a commercial reactor is from a rod-drop or a rod-ejection accident from shutdown (cold or hot) conditions with a defected rod in the vicinity of the ejected or dropped rod. Tests in CDC at SPERT and in NSRR at I
JAERI and experience with pulse operation in the CDC core and in the Bestern New York Research Reactor have provided a good indication of the effects due to a rod drop or rod ejection accident in a commercial reactor. The results of these tests and events indicate that fuel rod failures due to waterlogging will not result in significant failure propagation beyond nearest neighbors or structural damage of the fuel assembly that would inhibit coolability or prevent safe shutdown of the reactor. The only event that has resulted in failure of an adjacent rod and was directly attributable to waterlogging was in the CDC reactor (See Section III-C).
Some bowing of fuel rods adjacent to the ruptured rods has also been observed (1, lj), but here again the damage was not severe enough to affect the heat transfer from these rods to a degree sufficient to inhibit safe operation or shutdown.
NRC Regulatory Guide 1.77 for a PWR rod ejection accident (2) has resul-ted in designs that limit the calculated radial average energy density to 280 cal /gm.
Peak enthalpies for a rod drop accident in a BWR are also calculated to be less than 280 cal /gm (11). Therefore, energy depositions larger than 280 cal /gm are not of practical interest for waterlogging failures in commercial reactors. The waterlogged fuel rod ruptures that occurred in the CDC and NSRR tests at total energy inputs less than 280 cal /gm have all been localized and did not result in sufficient flow blockage, due to cladding ballooning or dispersal of fuel, to cause failure of adjacent rods.
The nuclear-to-mechanical energy conversion and associated pressure pulses provide an indication of the potential for failure propagation.
The energy conversion values reported in Table I for waterlogged rods, in which the total energy input was less than 280 cal /gm, were all 0.2% or less.
This is consistent with the results f.'om non-waterlogged SPERT tests.
In both instances very little, if any, fuel was dispersed into the water, and this is a condition that would be needed for a large nuclear-to-mechanical energy conversion to occur.
The measured pressure pulses in the sealed capsules in these experi-ments have not exceeded 1400 psi.
The highest capsule pressures observed were in two tests in which the fuel rods were intentionally defected (1),
and pressures reached 5300 and 5900 psi.
In one of these tests, in which the capsule pressure reached 5300 psi, there was an adjacent fuel rod that recuived no damage.
Pressure pulses in a reactor should be much less for the same energy inputs because of the openness and size of the core compared with the capsules in these experiments. Since the pressure pulses did not lead to unacceptable results in the capsules and the pulses in a reactor would be less severe, significant adverse consequences are not expected.
Fuel rod waterlogging tests in the CDC core (JS) indicated that less total energy was generated when ruptures occurred than in unruptured rods.
This was because the rupture of the fuel rod and the dispersal of the fuel into the water provided an additional negative reactivity mechanism that aided in reducing the nuclear energy release during the excursion. General Electric in GESSAR (2D) also has concluded that any fuel dispersal that might result from waterlogging failures would reduce the severity of such a transient by reducing t!.e rod reactivity.
~ _.
TABLE I - WATERIACCED TESTS ON FUEL RODS WITH ZIRCALDY C1 ADDING (d)
REFERENCE TEST ENERCY TOTAL REACTOR NAX. FUEL MAXIMtM NUCLEAR /NECH-REMARKS NO.
DEPOSITION ENERGY
- PERIOD, ROD INTERNAL CAPSULE ANICAL ENERGY AT TINE OF DEPOSITION, asec FRES ' <.'E.
- PRESSURE, CONVERSION, 1
- FAILURE, cal /gm pet ps!
cal /gm 1
555 160 225 5.5 6500 440 0.03G CLADDING ANNEALED
. SS THAN 558 160 225 5.5 1560 530 0.059 a ot 0F 556 150 225 5.5 11000
$30 0.031 CLADDING COID WORKED FUEL 559 120 225 5.5 1110 406 0.022 EXPELLED 3
496 60 220 3.0 NR 1200 0.20 CLADDING 101 COLD WORKED 6
401-3 101 141 1.5 NR(b) 781 0.04 401-3b 130 137 1.53 NR 937 0.02 4
552 165 430 3.0 1710 140 0.18 CLADDINC ANNEAL.ED ALL FUEL 547 85 430 3.0 1830 170 0.25 CIADDING COLD WORKED EXPELLED (a) 401-4 119 146
%1.5 19300 767
<0.1 FAILED 401-4c 113 136 St.5 13980 818
<0.1 FAILED 401-5 NA 115 St.5 NA NA
<0.1 FAILURE THRESdOLD (APPROX) 6 402-3 NA (c) 141 1.51 NR NA 0
RADIAL CAP IN ROD FILI.ED WITH M 0 2
411-3 112 139 1.53 NR 933 0.09 15 NIL NOLE IN UPPER FART OF FIN 421-3 NA 139 1.53 NR NA 0
15 NIL HOLE IN HIDDLE FART OF FIN (a) 401-4b 110 147 NR 0
9'O NR FAILED - 1001 FILLED WITH WATER
'402-4 96 145 NR 2986 NR NR FAILED - 72% FILLED WITH WATER 403-1 112 141 NR NR 782 NR FAILED - 321 FILLED WITH WATER 403-3 102 134 NR
!!23 1393 NR FAILED - 111 FILLED WITH WATER (a) 401-6 90 216 NR 8617 1279 NR FAILED - 961 FILLED WITH WATER 402 81 219 NR 4520 1137 NR FAILED - 693 FILLED WITH WATER 403-2 97 217 NR 0
241 NR FAllED - 151 FILLED WITH WATER (4 JAERI Unpublished Results
( c) NA-Not Applicable. No Fallure
- ( b) NR - Not Reported (d) All Cladding Zircel y 2 VI.
Conclusions Waterlogging failures can result in extensive damage to individual fuel rods and the loss of fuel pellets, and tests that simulate severe reactivity 1
insertion accidents have produced the worst observed waterlogging failures.
Pressure pulses produced in these tests are low, however, and the conversion from nuclear to mechanical energy does not appear to be significantly different than in accident tests without waterlogging.
In only one known event was significant damage experienced by adjacent fuel rods, and in most events no damage to adjacent rods has been observed.
Prior cladding damage is needed to permit fuel rod waterlogging to occur, and cladding defects in general are sparse (usually much less than 1% of the rods leak during 'their lifetime). Furthermore, operating restrictions that are used to reduce pellet / cladding interactions reduce the potential for waterlogging failures in defected rods during normal and abnormal operational transients.
The potential for waterlogging failures is thus small, and the only suspected case of a waterlogging failure in a commercial plant took place before PCI operating restrictions were in effect (only two rods were damaged). There is, therefore, no apparent threat from waterlogged fuel to overall fuel rod integrity, to coolability of the core or to safe reactor shutdown.
l l
-20 VII. REFERENCES 1.
" Standard Format and Content of Safety Analysis Report for Nuclear Power Plants," NRC Report, NUREG-75/094 (Regulatory Guide 1.70, Revision 2), September 1975.
- 2.. J.R. Buchanan (NSIC) letter to B. Siegel (NRC), " Transmittal of j
Citations Related to Waterlogging Fuel Element Failures,"
November 9, 1976.
3.
L. A. Stephan, "The Response of Waterlogged UO Fuel Rods to Power Bursts," Idaho Nuclear Report, IN-ITR-105, April 1969.
i 4
L. A. Stephan, "The Effects of Cladding Material and Heat Treatment.
i on the Response of Waterlogged UO Fuel Rods to Power Bursts,"
Idaho Nuclear Report, IN-ITR-111, January 1970.
5.
" Quarterly Progress Report on the NSRR E'.periments, October 1975 to March 1976, Reactivity Accident Laboratory and NSRR Operations bection," JAERI Report, M6635, June 28, 1976.
i 6
_M.
Ishikawa, "First Progress Report of the Nuclear Safety Research Reactor (NSRR) Experiments," JAERI.
l 7.
" Assumptions Used for Evaluating A Control Rod Ejection Accident for i=
Pressurized Water Reacters," NRC Regulatory Guide 1.77, May 1974 l
8.
" Quarterly Technical Report, SPERT Project October thru December 1967,"
l Idaho Nuclear Report, ID0-17279, October 1968.
i 9.
" Quarterly Technical Report, SPERT Project, January thru March 1966,"
Idaho Nuclear Report, IDO-17206, September 1966.
10.
R. T. Burnett (Western New York Research Center) letter to D.J. Skovholt (USAEC), " Report of Fuel Pin Failure," February 11, 1971 11.
P. T. Burnett (Western New York Research Center) letter to D.J. Skovholt, (USAEC),'" Report of Pin Failure:
Part II, and Request to Resume Pulse Operation," August 27, 1971.
l 12.
E.M. King and E.L. Long, " BONUS Superheater Failure," Nucl. Safety, 1, l
1965.
13 H.L. Whitener and ' L. A. Stephan, " Fuel Rod-Failure Experience During Transient Testing," Trans. Am. Nucl. Soc. JA, 32 (19f*3 s
i
- REFERENCES 14.
S. Burnstein (Wisconsin Electric Power Company) letter to B. Rusche
.(USNRC), " License Event Report No. 50-266/75-13 Failed Fuel Assembly D-03, Position K-6, Cycle 3, Point Beach Nuclear Plant," December 30, 1975.
15.
B.A. Hersey and H.B. Meieran, " Behavior of an Intentionally Defected l
' Fuel Rod Which Ruptured During Irradiation," Westinghouse Report, WAPD-TM-628, July 1969.
16.
Westinghouse Reference Safety Analysis Report, RESAR 3S, July 1975.
17.
Combustion Engineering, System 80 PSAR, CESSAR.
18.
Babcock and Wilcox, Standard Nuclear Steam System, BSAR-205.
19.
" Experimental Results of Potentialy Destructive Reactivity Additions to an Oxide Core," Idaho Nuclear Report, ID0-17028, December 1964.
20.
General Electric BWR/6 Standard Safety Analysis Report (GESSAR).
21.
C. J. Paone, R. C. Stirn and J. A. Woolley, " Rod Drop Accident Analysis for Large Boiling Water Reactors," General Electric Report, NEDO-10527, March 1972.
l 22.
D.H. Risher, Jr., "An Evaluation of the Rod Ejection Accident in l
Westinghouse Pressurized Water Reactors Using Spatial Kinetics Methods," Westinghouse Report, WCAP-7588, Revision 1, December 1971.
4 l
l b
e
..,...