ML20055G034
| ML20055G034 | |
| Person / Time | |
|---|---|
| Site: | Grand Gulf  |
| Issue date: | 07/16/1990 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20055G033 | List: |
| References | |
| IEIN-87-043, IEIN-87-43, NUDOCS 9007200075 | |
| Download: ML20055G034 (7) | |
Text
- *
+
g *,
+I J. UNITED $TATES sW ' ,n NUCLE AR REGULATORY COMMISSION
- g[ . ij wAswiNcf ON, D C. 20b66
%,*..,*/
SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATED TO BORAFLEX GAPS IN SPENT FUEL RACKS ENTERGY OPERATIONS. INC.
GRAND GULF NUCLEAR STATION UNIT 1 DOCKET NO. 50-416
1.0 INTRODUCTION
High density spent fuel racks were installed in the Grand Gulf Nuclear Station, Unit 1 (GGNS 1) during the first fuel cycle (1985). On September 8,1987, the NRC issued Information Notice No. 87-43 alerting all operating licensees that gaps had been found in the Boraflex panels of tie high density spent fuel storage racks at Quad Cities Unit 1. In the tight of this information, the GGNS-1 licensee (System Energy Resources Inc. before June 6, 1990 and Entergy Operations,Inc.onandafterthatdate),initiatedneutronattenuationmeasure-ment (blackness testing) on the Boraflex panels in the GGNS-1 high density spent fuel storage racks. On November 21, 1938, the licensee reported the results of these measurements which confirmed the presence of gaps in about 40% of the irradiated panels investigated with an average gap size of 0.8 inches and a maximum gap size of 1.4 inches. The licensee has, therefore, performed a criticality analysis to demonstrate the safety of the storage racks and submitted it to the NRC by letter dated February 17, 1989.
On February 27 and December 5, 1989 and A]ril 30, 1990, the licensee proposed a program to monitor the performance of tie Botaflex panels. The surveil-lance program consists of using blackness tests to examine the Boraflex panels in at least 50 fuel storage cells where freshly discharged fuel assemblies will be stored for 10 to 14 months. The surveillance will be conducted once per fuel load cycle. The blackness test can detect gaps (separations) of wider than half an inch in the Boraflex panels. This surveillance program supplements the existing inspection program in which nine pairs of Boraflex coupons are installed in the fuel storage pool, periodically removed, and then sent to an independent laboratory for physical testing.
On June 22 1990 the licensee committed to perform additional surveillance andanalysIsfollowinganearthquaketoassurethatgapsintheBoraflexare not significantly changed for a seismic event exceeding the postulated OperatingBasisEarthquake(OBE).
The staff's evaluation of the criticality analysis, the spent fuel rack surveillance program and the additional surveillance for seismic events is provided separately below.
l MR72 p
MR M.M PDLa
e 2
2.0 EVALUATION 2.1 Criticality Analysis The KENO Montt Carlo computer code was used to calculate the reactivity of the GGN$-1 storage racks when fully loaded with the most reactive Cycle 4 fuel.
Neutron cross section data from the 123-group SCA1.E library was generated for input to KENO using the AMPX code. Small incremental reactivity effects due to manufacturing tolerances were calculated with the two dimensional multi-group, transport theory code CASMO. These models have been bench; narked agaInst experimental data by the licensee, as well as others in the nuclear industry, and have been found to adeountely reproduce the critical values.
The spent fuel pool reactivity calculations were conservatively based on a water temperature corresponding to the maximum reactivity within the design '
range, no credit for radial neutron leakage, no credit for the neutron absorption of minor structural components or burnable poisons, unirradiated '
I fuel of the highest reactivity, and Boraflex width, length, and composition yielding the maximum reactivity. Reactivity effects of manufacturing tolerances were determined at the 951 probability 951 confidence (95/95) level, thereby meeting the NRC requirements.
The reactivity effect of gaps in the Boraflex panels was based on the following assumptions. All of the gaps were assumed to occur at the cxial midplane, maximizing the reactivity effect. The maximum gap size modeled was 6.0 inches.
This is conservatively larger thah the predicted 5.75 inches based on a max' mum 41 shrinkage (suggested by Electric power Research Institute) and the 1.4 inch maximum gap measured in the GGHS-1 racks. Gap size measurements of panels with multiple gaps were treated as a panel with one large gap since the reactivity effect of one large gap in a given sanel would be expected to be greater than that of two or more smaller gaps. inally, a measured gap falling within a gap size interval was assumed to be at the upper limit of that interval thereby maximizing its reactivity effect.. The results indicate that a 4 inch gap can be accommodated at the axial midplane of all 4 panels per cell and still main-tainarackeffectivemultipiteationfactor(k-effective)belowtheacceptance criterion of 0.95. In addition, a gap size as large as 6 inches can be acecnrnodated at the axial midplane of 2 Boraflex panels eer storage cell and still meet the 0.95 acceptance criterian.
The blackness test data from GGNS-1 and Quad Cities indicate that the size and 1ccation of Boraflex gaps are nearly randomly distributed throughout the rack.
Since reactivity is sensitive to the gap size and axial location as well as the number of panels with gaps, a worst case distribution was developed by the l licensee by multiplying a gap size distribution by a gap frequency distribution.
I The staff has reviewed these distributions based on the GGNS-1 and Quad Cities measurements and concludes that they are adequately conservative. The results of an analysis assuming these maximum distributions of gap size and gapped panels per cell yielded a 95/95 upper limit k-effective of 0.9496 for the storage rack. Therefore, the licensee has confirmed that the maximum reactivity due to a conservatively high distribution of gap sizes and gapped panels per w- - - - _ *'____I- _ __ '
___a +-____-__m.__m__ e -_ =__ L
. -3 cell will be less than the NRC acceptance limit of 0.95, including i uncertainties, with the racks fully loaded with the most reactive GGNS-1 Cycle 4 fuel (9x9-5 LTA) and flooded with clean unborated water at the temperature within the operating range corresponding to the highest reactivity.
i In sumary, the staff has reviewed the assumptions, analytical methodology, and results of the criticality analysis including the effect of Boraflex gaps in the GGNS-1 spent fuel storage racks. Based on this review, the staff concludes that the storage racks can safely accommodate the maximum reactivity of GGNS-1 Cycle 1 through Cycle 4 fuel with the presence of gaps in the Boraflex panels at a conservative worst case distribution.
2.2 Spent Fuel Rack Surveillance program !
Nuclear reactor plants provide storage facilities or pools for the wet storage I of new and spent fuel assemblies. The fuel storage pools must maintain the stored fuels in a subcritical array during all credible storage conditions.
In the fuel pools where the fuel assemblies are closely arranged (the so-called high-density fuel storage pools), neutron absorbing materials (poisons) are used to maintain subtrit'cality. Sheets of Boraflex, a methylated polysiloxane elastomer uniformly filled with finely divided boron carbide powder, are often used as poisons in the fuel storage pools. $1nce 1980 23 comercial nuclear powerplantsintheUnitedStateshavebeenusingBoraflexastheneutron absorbing material in the fuel storage pools, in pressurized water reactor plants, boron is also added as a neutron absorbing material to the water in the fuel storage pools. In general, in boiling water retctor plants, no boron is added to the fuel storage pool water.
i i in the fuel storage pool of GGNS-1, the fuel storage racks consist of l individual 6-inch by 6-inch square cross-section, rectangular cylinders, l each of which can house a single 8 by 8 nuclear fuel assembly. The rectangular cylinders, or cells, are arranged ir modules of varying number of cells. Each of the four walls in a fuel storage cell consists of a Boraflex panel. Each Boraflex panel consists of a Boraflex sheet sandwiched between two stainless steel plates. The space containing the Boraflex is vented w the fuel storage pool to allow escape of any gas which may be generated from the polymer binders in Boraflex during heating and irradiation, thus preventing possible bulging or
- l. swelling of the Boraflex panels. in the fabrication of the Boraflex panels,
! thin beads of Dow Corning RTV adhesive (a silicone-based cement) is applied down the length of the assembly. The bead is smeared thin and a Boraflex slab is attached to it. The stainless steel plates in the Boraflex sandwiches are then welded together on the edges.
The water in the GGNS-1 fuel storage pool conta is no boron. The pool water temperature is maintained below 140 degrees Fahrenheit under normal operating conditions.
l-__--_____-
4-Boraflex has undergone extensive testing to determine the effects of gama and neutron irradiation in various environments and to verify its structural integrity and suitability as a neutron absorbing material (Bisco Products, I Inc. Technical Report No. NS-1-001, " Irradiation Study of Boraflex Neutron l ShieldingMaterials," Revision 1. August 12,1981). The evaluation tests have !
shown that Boraflex is unaffected by the environment of fuel storage pools and '
will not be degraded by corrosion. Boraflex in a solution of 2000 parts per million of boron was exposed to a cumulative gama radiation dose of 1.03+11 (1.03 times ten raised to the eleventh power) rads with a concurrent neutron ,
flux of 8.3E413 neutrons / centimeter square /second. The results show that Boraflex maintains its neutron attenuation capabilities in the environments of borated water and gama and neutron radiation. However, irradiation caused some loss of flexibility and shrinkage of the Boraflex material.
Long-term soaking of Boraflex in borated water at high temperatures was also performed (Bisco Products, Inc., Technical Report No. NS-1-002, "Boraflex Neutron Shielding Material Product Performance Data," sugust 25,1981). The -
test results show that Boraflex withstands a temperature of 240 degrees Fahren-heit in a solution of 3000 parts per million of boron for 251 days without visible distortion or softening. The Boraflex showed no evidence of swelling or loss of ability to maintain a uniform distribution of boron carbide. The water temperature c'. the fuel storage pool in a nuclear power plant is not expected to exceed 180 degrees Fahrenheit under credible conditions. This temperature is well below the 240 degrees Fahrenheit of the test temperature.
In general, the rate of a chemical reaction decreases exponentially with decreasing temperatures. Therefore, the staff does not enticipate ?ny signifi-cant deterioration of the Boraflex due to high temperat%res and water soacing i
at nuclear power plants over the typical design life e,.' the fuel storage racks.
nor Boraflex The tests (ibid) composition hasalso show thateffect a discernible neither on ir-adiation, the neutronenvironment,f attenuation o the Boraflex l material. The tests show that Boraflex does not possess leachable halogens l that might be released into tne pool environment in the presence of radiation.
Similar conclusions are reached regarding the leaching of element boron from the Boraflex. Boron carbide of the grade normally present in the Boraflex typically contains 0.1 weight percent of soluble boron. The test results have '
l confirmed the encapsulation capability of the silicone polymer matrix to prevent the leaching of soluble species from the boron carbide.
Recently, anomalies and degradations of Boraflex were observed in two nuclear power plants that use Boraflex in high density fuel storage pools. Minor physical changes in color, size, hardness, and brittleness were found in Boraflex surveillance coupons in the Point Beach Nuclear Plant, Unit Nos. I and 2 (letter from C. W. Fay, Wisconsion Electric Power Company, to George Lear, U. S. Nuclear Regulatory Comission, dated February 11,1987). Gaps or physical separations of up to 4 inches were discovered in Boraflex panels of the Quad Cities St6 tion, Units 1 and 2 (Northwest Technology Corp, Report No.
l NET-042-01, " Preliminary Assessment of Boraflex Performance in the Quad Cities Spent fuel Storage Racks," April 10,1987).
L
I y'. l
-5 )
The exact mechanisms that caused the observed physical degradations of Boraflex have not been confirmed. The staff postulates that gamma radiation from the spent fuel assemblies initially induced cross-linking of the polymer in l Boraflex, producing shrinkage of the Boraflex material. When cross-linking scissioning(aprocessinwhichbondsbetweenatomsare '
became broken) ofsaturated,lymer the po predominated as the accumulated radiation dose increa Scissioning produced porosity which allowed the water in the fuel storage pools to permeate the Boraflex matrix. Scissioning and water penneation could embrittle the Boraflex material. In short, gama radiation from spent fuels is the most probable cause of the minor physical degradations of Boraflex, such as the changes in color, size, hardness, and brittleness that were observed in the surveillance coupons of the Point Beach plant. The staff does not have ,
sufficient information to determine conclusively what caused the gap formation in some Boraflex panels of the Quad Cities plant. However, it is conceivable ,
that if the two ends of a full-length Boraflex slab are physically restrained, then shrinkage caused by gama radiation can break up the panel and lead to gap formation. The elements of such a postulation are present in the Boraflex panels of both the Grand Gulf and Quad Cities plants.
Subsequent to the reports on degraded Boraflex by the Point Beach and Quad Cities plants, NRC issued Information Notice No. 87-43, " Gaps in Neutron-i Absorbing Material in High-density Spent fuel Storage Racks," on September 8 L 1987, sotifying nuclear power plant licensees of the potential gap formation of Boraflex. The GGNS-1 surveillance report was a response to the Information l Notice.
The Bisco Products, Inc. published additional test data on Boraflex in Technical Report No. NS-1-050 (Interim), " Irradiation Study of Boraflex Neutron Absorber i I
Interim Test Data," Revision 1, November 25, 1987. The interim test data show that shrinkage of Boraflex of up to 3.4 percent in length and up to 4.5 percent in width appear to have saturated at an accumulated gama radiation dose of
- IE+11 rads. In December 1988, the Electric Power Research Institute published a report, EPRI No-6159, "An Assessment of Boraflex Performance in Spent-Nuclear-Fuel Storage Racks," sumarizing the behavior of Boraflex under various test conditions and in nuclear power plants undar field conditions.
The staff determined that the adhesive used to cement the Boraflex sheets onto the stainless steel plates can create physical restraints in the Boraflex i slabs in the fuel storage racks of GGNS-1. When Boraflex shrinks as a result of radiation exposure, it is likely that these physical restraints will cause
- local stresses leading to separations (gaps) of the Boraflex sheets. However, during the projected life of the fuel storage racks, the proposed surveillance program using neutron attenuation measurements should be able to detect such gaps of wider than half of an inch in the degraded Boraflex panels.
The GGNS-1 Boraflex Surveillance Program uses a fast neutron source and thermal neutron detectors to determine the size and distribution of gaps in the Boraflex panels. A section of at least 50 rack cells will be designated i
as the test area. Freshly discharged spent fuel will be placed in this test area during each refueling and left there for 10 to 14 months, so that this test area will receive the maximum accumulated gamma radiation within the
- - - - - - --,----w~,
i 4 spent fuel puol. Gap sizes and distributions used in the criticality analysis are based on a conservative extrapolation of industry and GGNS-1 test data.
Prior to each refueling, the spent fuel assemblies are removed from the test area and blackness measurements are made in the rack cells. The data will be used to confirm the validity of the gap sizes and distributions used in the criticality analysis. Gaur.a fluence calculations are performed for incore fuel to determine maximum residence times allowed in the racks. The residence tiees are limited to the maximum fluence in the test area associated with the latest blackness test. The accumulated genna dose for each rack cell will be tracked and recorded.
If the gap size and distribution are found to be not bounded by the criticality analysis, the licensee is required to report this unanalyzed condition under 10 CFR 50.73 and take whatever corrective actions are needed to maintain the effective multiplication factor required by GGN$-1 Technical Specifications (k-effective less than 0.95). For example checkerboard loading of fuel assembliesintherackscouldbeuseduntilalong-termcorrectiveactionis taken.
The staff concludes that the proposed surveillance program is acceptable because it provides a conservative means to detect Boraflex degradation due to gaps and to take corrective actions in a timely manner. However, the gap size and frequency distribution used in the criticality analysis is based on a sample of from one to two percent of the rack cells. Should the assumptions in the criticality analysis not be confirmed by the data, it may be necessary to expand the measurements to additional rack cells. The licensee has committed to continue the surveillance program until data shows that the gap size and distribution is stable and until a report providing the data supporting its discontinuance is submitted to and approved by the staff.
2.3 Additional Surveillance for Seismic Events The Boraflex panels of the high density spent fuel racks consist of a sheet of Boraflex sandwiched between two sheets of stainless steel. The basic element used in fabricating the modules are cruciform, eli and tee sections of these panels. The nominal dimensions of the cell cross section, Boraflex sheet thickness and stainless steel sheet thickness indicate that the space within the sandwich panel may be large enough in some panels for the Boraflex sheets to slump downward or for Boraflex sheet segments to slip downward in a seismic event leaving the upper portion of the panel void of Boraflex.
In response to staff's questions regarding this concern, the licensee stated that Boraflex panels are fabricated using closely - dimensioned edge spacers so that the maximum space available in the sandwiched panel is 0.006 irrhes. The Boraflex sheets have a nominal thickness of 0.070 inches, and therefore, snere would be no appreciable slumping and no downward motion of segments of Boraflex except for closing of the gaps developed by shrinkage. The criticality analysis to account for gaps assumes the gaps occur in the central section of the panel.
During a seismic event movement of Boraflex segments downward would result in a less reactive configuration because the gaps would be increased in the top section of the panel and decreased in the central section. Neutron leakage is greater at the top of the rack compared to the central portion of the rack.
s '. -
.y.
The staff concludes that during a seismic event, potential movement of Bornflex within the panels of the high density spent fuel racks would not decrease the shutdown margin below that calculated in the criticality analysis. This is based on the use of a conservative assumption in the criticality analysis that gaps in the Boraflex occur in the most reactive central section of the Boraflex.
The licensee has committed to perform additional blackness surveillance tests and analyses following a seismic event exceeding the postulated Operating Basis Earthquake to determine any changes in Boraflex position and criticality analyses.
3.0 CONCLUSION
S The staff has reviewed the licensee's criticality analyses and surveillance programs which are used to account for gaps in the Botaflex of the high density spent fuel storage racks. The staff concludes that:
(1) The storage racks can safely acconnodate the maximum reactivity of GGNS.1 Cycle 1 through Cycle 4 fuell (2) The GGNS-1 storage rack surveillance program is acceptable for long term storage of spent fuel, and (3) The additioNil surveillance following an earthquake exceeding the postulated Operating Basis Earthquake is an acceptable means of assuring an adequate margin of subcriticality following a significant seismic event.