ML19350C580
| ML19350C580 | |
| Person / Time | |
|---|---|
| Site: | Prairie Island  |
| Issue date: | 03/31/1981 |
| From: | Mayer L NORTHERN STATES POWER CO. |
| To: | Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8104060164 | |
| Download: ML19350C580 (15) | |
Text
e O
NSP NORTHERN 5TATES POWER COMPANY M I N N E A PO L.1 S. M I N N E S OTA 55401 03 J
53 t 7 VM g
S @3 N
P O
March 31, 1981 T
QSg 11 4
9h o.S-0-
/g Director of Nuclear Reactor Regulation
[hN U S Nuclear Regulatory Commission Washington, D C 20555 PRAIRIE 'SLAND NUCLEAR GENERATING PLANT Docket No. 50-282 Licensing No. DPR-42 50-306 DPR-60 Supplemental Information - License Amendment Request Dated January 31, 1980 Attached. are additional responses to questions sent to L 0 Mayer from R A Clark on November 5,1980, concerning the Prairie Island NGP Fuel Storage Facility modification.
Additional responses to question ASB-2,5 and SEB 2,3,4,7,9,10 are enclosed.
L 0 Mayer, P Mana;-* of Nuclear Supprt Service s LOM/TMP/bd cc: J G Keppler NRC Resident Inspector C Charnoff MPCA Attn: J W Ferman 8104060($
b
Attachment to March 31, 1981 Letter AUXILIARY SYSTEMS BRANCH (ASB)
ASB-2 ADDITIONAL RESPONSE As a result of tha *
. generation calculation method change identified in reference 1, the p.rameters listed below have been recalculated.
Originally (reference 2), the heat generation used for pool I was that from all the fuel in ' both pools. This was conservative but this situation could never exist.
The following " time to, boiling" for pool 1 assumes that a full core discharge of 121 assemblies and 266 assemblies generate the heat in pool 1.
The "t ime to boiling" for pools 1 & 2 assumes total mixing between the two pools.
Neither of the " times, to boiling" account for the heat removal effect of evaporation prior to boiling.
Previous Values **
Recalculated Values Time to Boiling 2.9 hr 2.9 hr***
Pool 1 Maximum Evaporation 44.7 gpm 51.4 gpm Rate (Pool 1 & 2)
Time to Boiling 9.9 hr 8 hr Pool 1 and 2 The following administrative controls will be adhered to prior to placing a full; core discharge into -pool 1:
- 1. All gates will be removed
- 2. All spent fuel' pool cooling equipment will be operable.
4 The removal of the gates shen making a full core. discharge -into Pool I will
. provide.approximately 100 sq. ft. of flow area between the two pools and 100 sq. f t. between each pool and the canal allowing natural circulation to mix the:two pools.
This will increase the time to boiling to some value between 3.5 hrs and 8.0 hrs (8.0 hrs assumes total mixing). The requirement to have the spent fuel pool cooling equipment' (pumps, heat exchangers, and valves) operable ~ prior to making a full core discharge -into Pool I will make the total
' loss of spenc-fuel pool. cooling extremely unlikely.
EIf=these requirements canns-be met, the full cort disch'arge will be delayed until the requirements.can ce met or.the fuel from the full core discharge
.will beisplit-between tbs two pools or totally placed in Pool 2.
- *Tne 'four removabic racks in the shipping cask area will contain only the full (core discharge. The other 'five racks will be filled with recently discharged fuel.
~
' **See Reference 2 Exhibit C, page 3-63.
' ***The. two ef fect s of f set one -another.
1 M
Y--
h-
_a_
ASB-5 Additional Response The tools commonly used in the spent fuel pools are listed in the table below.
FUEL HANDLING TOOLS Bail Allowed Factor of Safet Max. Drop Height to Ultimate Tool Weight Length Height Above Fuel (Loaded)
BPRA Handling Tool 760 lbs 31.5 ft 15.3 ft 4 ft 4.8 Spent Fuel Handling Tool 343 lbs 31.75 ft 15.05 ft 8 ft 5.4 Thimble Plug Handling Tool 2
1ha 31.0 ft 15.8 ft 8 ft 8.9 These tools will be administratively restricted when handled over the new racks such that 't'eir potential energy will not be greater than 3250 ft-lbs.
Identification mat -. will be located on the shaf ts of the tools to indicate how high the tools may be raised above fuel when unloaded.
In the very unlikely event that a unloaded tool is raised above its maximum height the following items make the dropping of the tool extremely unlikely:
- 1. The Spent Fuel Pool Bridge Crane hoist hook has a keeper
- 2. The Spent Fuel Pool Bridge Crane' hoist has a 6000 lb capacity
- 3. Tne bail f actor of safety (see above table)
NOTE: The fact that the bail will deform prior to failing was not considered i-the calculations of the factors of afety-making the factors of safety i; ;ed con se rva t ive.
- 4. Trained, experienced NSP plant personnel will handle the tools.
The tools very infrequently (less than once a year) need to be handled by the Auxiliary Bldg Crane. The tools will be handled by the 25 ton hoist whien has a keeper.
- In recent refuelings the BPRA has not been needed to be used at all.
Stractural Engineering Branch (SEB) l SEB-2 Additional Response l
The design for the spent fuel storage modules utilized the AISC Code for i
beam and frame structure methodology.
The basic material allowables were
.taken from ASME Section III and Standard Review Plan 3.8.4 for the applicable load combinations. The fabrication and installation of the modules is in accordance with ASME Section III Subsection NF, Articles 2000, 4000, and 5000 except as follows:
- 1. Material traceability for individual pieces to their associated CMTRs is not maintained when the material is incorporated into the fabricated assembly. Traceability is preserved for all hea. s of material comprising each module. All materials used in the construction of the racks are drawn from a controlled lot of traceable, certified material.
2.'The ASME Code does not address neutron absorber materials such as Boraflex.
Howeve r, it has been purchased in accordance with the following:
a) ANSI N45.2, Quality Assurance Program Requirements for Nuclear Power Plants.
- 5) ANSI N45.2.2, Packaging, Shipping, Receiving, Storage, and Handling of Items for Nuclear Power Plants.
c) ANSI N45.2.6, Qualifications of Inspection, Examination, and Testing Personnel for Nuclear Power Plants, d) ANSI N45.2.9, Requirements for Collection, Storage, and Maintenance of Quality Assurance Records, e) 10 CFR 50, Appendix B - Quality Assurance Criteria for Nuclear Power Plant s.
- 3. The welds used to install the Boraflex into the fuel racks are made by the tig spot weld process. These welds are not structural welds.
Their only purpose. is to secure the neutron ab sorber. Each storage tube has approximately 100 of these tig spot welds.
Random f ailure of 50% er more of these welds would not result in dislocation of the Boraflex.
The' ASME Code 'Section III Subsection NF does not address tig spot welds. In order to qualify this process and the welding operator for use on the fuel racks, ASME Code Case N-262, " Electric Resistsace Spot Welding for Structural use in Component Supports," was usesd.
Although
-this code. case addreses the use of electric resistance spot welding, the qualification and test procedures included in the code case can be directly applied to tig spot welding. The test results have shown that this process consistently-produces welds with strength well in excess of that required by the code case.
In addition, two test samples are prepared and destructively examined prior to welding on each storage
-tube.
Because of _ the use of Code Case N-262 and cue extensive testing during fabrication, this welding process meets the intent of ASME Section III Subsection NF.
- 3;-
. =-
- SEB-3 Additional Response The structural evaluation of the proposed racks conforms to the applicable portions of the OT position referenced in the question.
The spent fuel racks are analyzed in accordance with Regulatory Guide 1.92. The fuel pool structure has been analyzed for the additional loads that would result from the proposed modification by utilizing the original l
Fluor-Pioneer calcualtions.
.In this analysis, the additional loads on the canal. wall and the floor slab were computed using the conventional building
- analysis method. Since the proposed racks are not laterally supported from the. pool walls, only the pool floor and the canal vall acting as a deep beam would ' carry additional vertical loads including additional vertical seismic loads. The vertical seismic loads were computed using the vertical seismic response spectrum from the original building seismic analysis report by John A. Blume & Associates.
The increase in the horizontal seismic. loads.on the pool floor and the canal wall as a result of the
- proposed modification is negligible, and so the application of R.G.
1.92 is not applicable (R.G. 1.92 was issued af ter the design and construction of the
. Spent. Fuel Pool).
t
)
S E,
r-
SEB-4 Additional Response t
a) When the rack drops on the built up wide-flange beam at a location s
directly above the point where the beam rests on the concrete wall, a limited amount of concrete crushing is anticipated. The bearing area under the beam is about 7-inches by 12-inches. Hence, crushing will be limited to this area.
Concrete crushing depth reported in Reference 3 was computed using an energy balance equation. Accordingly, the kinetic energy of the falling rack was equated co the plastic strain energy absorbed in the process of crushing the 7-inch by 12-inch concrete bearing area.
In that evaluation, it was conservatively. assumed that all the kinetic energy of drop would be absorbed by crushing the concrete.
b) Major chipping of concrete was not considered in Reference 3 analysis because
. of the following reasons:
i) The presence of the wide-flange beas and the pad between the falling rack and the concrete wall would "so f ten" the Unpact.
ii) The presence of the pool liner vnich is velded to an angle me=ber at the top inside corner of the pool wall would prevent the concrete from chipping and falling in the pool.
~
iii) The presence of reinforcing bars in both vertical and horizontal
~
direct.ons (No,11 bars @ 6" c/c), would prevent major chipping and spalling of concrete.
However, in response staf f concerns, the possibility of chipping the corner of the pool has been evaluated. The method of this evaluation and results are presented in the following paragraphs.
.The impact load resulting from the drop was calculated using an energy balance method; considering the energy absorption by the wide-flange s
beam, the compressible pad and the concrete. The load-defor=ation properties of the wide-flange beam, the pad and the concrete underneath were represented as three springs in series having elasto plastic characteristics. Using the energy-balance equation, the impact load was.
calculated by iterative procedure. 'For_ che 6" drop of the 24.8 kip xeck, an impact-load of 275 kips was obtained. This predicted impact load is very conservative because of the following reasons:
- i) Energy absorption by the cross-stringers'and the. cover plate of the
~
pool' cover were not included.
ii) Energy absorbed by the local crushing.of the concrete was not
-included.
iii) Energy ' absorption'by the reinforcing steel was not considered, even though it was assemed that the impact-load would cause a diagonal crack in the -inside corner of _the wall.
S-
e r
Using this impact 1 sad, and assuming a 45 diagonal crack at the top insile corner of the wall starting from the outer edg of the pad, the late.m? and vertical deformation of the cracked cornet of the wall was c ompu t r.d. The bending and dowel action of the reinfor( ing bars were evaluated by idealizing the bars as beams on elastic fuundation. The
-resulting lateral ani vertical deformation of the cracked co ner was computed to be less than one-tenth of an inch.
Such a smal' deformation will neither compromise the liner integrity nor will it er e chipping or spalling of concrete in o the pool, c) Because of the lack of clearance, the thickness of compressible pad can-not be increased. Nor is it necessary, since chipping or spalling is not predicted and since the =aximum postulated crushing depth is very small f. actual crus 3in.g depth will be smaller than 0.53" predicted in Reference 3 when energy absorption by the existing pad and the wide-flange beam is taken into account).
SEB-7 Additional Response The stress values reported in the original licensing report (reference 2) were based on the assumption that the rack and fuel assemblies move in phase.
Rattling analysis presented in refernce 3 assumed that the rack and the fuel assemblies move out-of phase. This resulted in additional amplification factors which were computed and presented in reference 3.
In this analysis, to obtain upper-bound amplification factors, all the fuel assemblies were assumed to move in phase with themselves but out-of phase with the rack.
The original seismic stress values (based on in phase motien) were then multiplied by these amplification factors to account for the out-of phase
~
i
. motion between the rack and the fuel assemblies and to obtain the f nal design stresses.
Assumption
'b' of the SEB-7 response refers to the fact that a) the basic seismic stresses in the orisinal document were computed assuming that the racks are not free slide (in addition to assuming that the ' fuel assemblies move ' in phase with the rack).
This is a very conservative assumption and predicts higher stresses since sliding
-releases. seismic stresses; and b) the amplification f actors presented in reference 3.were also computed assuming..that;no energy loss took place due to sliding, and all the energy would be available to the fuel assemblies to impact on the rack.1
. Thus,. the ; stresse's reported. in reference 3 represent very conservative and upper-bound values.
.,,.---..-.--..--.---__----.---_.--_-_aa-_w
t SEB-9 Additional R2sponso - (Rick Imptet & Thermal Stresses) 1.0 ADDITIONAL EVALUATION OF RACK-TO-RACK Ih 'ACT PHENOMENON Since the proposed racks are free-standing, they may slide towards each other and impact during a postulated seismic event.
The evaluation of the scenario where two adjacent racks impact on each other has been presented in the original licensing report as well as in responsa to SEB Question 9 (reference 3).
This additional evaluation is to determine the effect of (a) four adjacent racks impacting on the centra? rack (Figure la) and (b) six racks in one row impacting on the seventh or last rack at the end of the row (Figure Ib).
Evaluations of these two scenarios are performed using the results of the original impact analysis as described in the following two paragraphs.
A. Effect of Four Adjacent Racks impacting on the Central Rack Simultaneous impact of four adjacent tacks on the central rack, even though theoretically possible, would be an extrecely rare event, since the to-and-fro sliding motions of the five racks are likely to be quite random in nature. The resultit.; potential damage to the racks. from this postulated impact would be well within acceptable limits for the following reason:
Evaluation results for a single ra:k impacting on the adjacent rack showed that, even if the computed imp et velocity is increased by a factor of 2.7 (i.e., 12 over 4.43), the
-maximum crushing length of rack storage tubes due to impact is only 1.2 inches.. This was computed by an energy balance method in which the kinetic energy of impact was equated to the strain energy of deformation of the impacting outer tube walls.
This calculation was performed (as described in the response to SEB 9) by arbitrarily increasing sliding velocity trom 4.43 inches /sec. to 12 inches /sec. and by ignoring the energy expended in tilting the racks.
Since, in the case of a four-way impact, the outer tube walls'of any particular side of the central rack would be impacted by only one sliding rack, the postulated damage to the impacted tube walls would, remain the same.
B. Ef fect of Six Racks in a Row Impacting on the Seventh Rack The scenario is also very improbable because the to-and-fro sliding motions of the racks are randca in nature and for the six racks to be moving in the same direction is very. unlikely. Even more unlikely is the scenario in which all the six racks are moving at the peak
. velocity in the same direction.
However, based on earlier single rack-to-rack impact analysis,- this scenario is evaluated in a simple bad conservative way to show' the adequacy of the rack.
It is assumed that the combined kinetic energy of the six racks will be available as the energy of impact on the seventh or last rack.
The total impact energy-will then'be six times 3759 in-lb (based on the computed peak sliding velocity of 4.43 inch /sec) or 22,554 in-lb.
.7_
n
1 V
l X
d l
l l
FIGURE la FOUR RACKS IltPACTING CENTRN. RACK l
l l
l l
t I
[
e e
e e
e e
l l
l l
l.
r FIGURE lb
(
SIX RACKS IMPACTING END~ RACK r
8 I
As described in reference 3, the single rack-to-rack impact analysis was performed using an arbitrary sliding velocity of 12 in/see instead of the computed value of 4.43 in/sec. This was done so as to show
' the existence of high margin of safety. The impact energy of a single rack at the 12 in/sec velocity is 26,733 in-lb.
Thus, the rack was already evaluated for an impact kinetic energy of 26,733 in-lb, which is more than the 22,554 in-lb determined for six racks at the computed 4.43 in/sec velocity. Thus, even for the case when six racks in a-row are postulated to impact on the seventh rack simcita-neously, the damage to the rack storage tubes would not exceed 1.2 inches from'the. top.
This was previously shown to be well within acceptable limits.
2.0 ASSUMPTIONS AND METHODS USED FOR IMPACT ANALYSIS The assumptions and method of analysis used to determine the depth of
- the plastic region in the event of a rack-to-rack impact are describe ~
below.
A. Assumptions
- 1. The racks impact on each other when their velocities are at the peak.
'2.
All the: kinetic energy of impact is absorbed only by the
- plastic deformation of the tube walls parallel to the direction of motion. This assumption is conservative because of the following reasons
- a. The energy spent in tilting the rack to a position where the two racks' may impact is not considered.
- b. The. rebound energy, which is quite significant, is not considered.
- c. Elastic strain energy of deformation -is ignored.
~
- d. Elastic and plastic energy of deformation of the tube walls perpendicular. co. the direction of motion is not considered.
- 3. The stainless steel is assumed to have~elasto plastic stress-strain-properties.
'4. The tmpact energy is computed using the combined real and hydrodynamic mass. ' This is conservative, since only the real mass will contribute to the impact energy.
- 5. The two faces of the two adjacent' racks impact only at the ltopjedges deforming each'other. This assumption is 'more s
conservative than assuming. a flat ' surf ace -impact. - Considering-the very small gap between the racks, the possibility of, one
. point impact -is judged to be' small. Also, initial point impact, if'any,.would cause the racks to rotate and:the' result would be a
'line impact. u s
t
$ '- ? n-. '
B. Method of Analysis Presented below is a step-by-step outline of the method of analysis:
- 1. The kinetic energy of impact is computed using the peak sliding velocity and total combined mass.
- 2. Impact energy on a single tube wall is computed by dividing the total kimetic energy by the number of tubes in the impacting side of the rack.
Far each tube, the two tube walls parallel to the direction of motion are assumed to absorb the energy, since the stiffness of the
. walls perpendicular to the direction of motion is small compared to the walls parallel to the direction of motion.
-3. Evaluation is made to determine whether the tube wall geometry permits buckling before compression yielding.
It was found that the compression yield load is smaller than the buckling load, and hence the energy absorption would result from crushing rather than buckling.
4.The top edge of the tube is assumed to get crushed. The extent of the crushing is defined as "a" in Figure 2.
The value of "a", depends on the impact energy. Using an iterative method, the value of "a" is determined by equating the elasto plastic strain energy of the rushed corner to the impact energy.
5 The vertical length of damaged rack "D" is computed from "a".
E
' _ to _
4 1
V 99 A
IMPACT FORCE e
n V
a l
I l
l l
1 FIGURE 2 FUEL RACK CRUSHING LOCATI0tl i
11 o
3.0 OVERALL EFFECT OF IMPACT ON THE RACK STRUCTURE The rack base assemblies are very stiff in the horizontal direction.
Hence, when two racks impact at the base, the racks will rebound. The impact, however, will result in a high impact force at the base having an extremely short duration. This force will cause only local deforma-tion. Due to the short duration of the force, the total energy transmit-ted to the rack will be small. Using an energy balance method and equating the impact energy (22,554 in-lb) to the elasto plastic strain energy resulting from the crushing of the projected edge of the base assembly, this local deformation was computed to be much less than the proj ect ion.
4.0 THERMAL STRESSES Three -types of thermal stresses may occur in a spent fuel rack.
The se are:
Q = Thermal stresses resulting from placing a hot fuel bundle in an empty rack.
T = Thermal stresses resulting f rom the increase in the fuel pool water temperature during normal operation.
T = Thermal stresses resulting from the-increase in fuel pool water temperature during accident condit ion.
Thermal stresses resulting from Q loads have been considered in the design of the proposed racks.
In order to allow room for thermal expansior. of the racks without exerting forces on adjacent racks, the arrangement of racks in the pools will'be modified to provide a gap betreen the rack bases.
. The thermal expansion of the rack base (in the B cell direction) for a temperature increase from 70 F to 212*F is 0.11 inch. A gap of 0.25 inch will be provided between a djacent racks.
Increase in pool water temperature during normal op<eration (T ) and during the design basis accident (T ) would cause enermal loa 8s if it is assumed that the. friction between t$e rack leg and the pool floor liner
~
'does not pe rmit thermal growth' of the rack.
This fri: tion would also resist' the horizontal. movement of the rack during a seismic event.
Howeve r,
the maximum resistance to both seismic loads and the thermal growth will be -
limited : by the maximum frictional force that can develop between the rack and the pool floor..Thus, if the stre'sses/in the. rack components are
, computed for this maximum frictional force applied at the bottom of-the rack, the resulting ' stresses would constistute the upper bound stresses that 'can t result ' from any load combinations.
Maximum' f rict ional force to ' be appl'ied. a't the bottom of the rack legs was-L computed by multiplying the ~ weight 'of the rack withi the ' maximum coef ficient of: friction of- 0.8. cThis frictional force 'was applied in all.four legs in -
two'orthogonal horizontal directions. The stresses in the critical leg.
member. interfacing the pool floor and the Lother critical components were.
idetermined using-conventional elastic structural analysis methods.
- 12
The results of this upper-bound stress analysis are presented in Table 3C for.the three critically stressed components. These rest Lts show that the stresses resulting from the upper-bound analysis described here are well within the allowable stress values.
In performing this evaluation, the allowable stresses were selected very conservatively.
In fact, allowable stresses applicable for the D + OBE load combination (per Standard Review Plan 3.8.4) were used even though per OT position the allowable stresses would be much higher.
It is concluded that stresses resulting from T and T would not cause any overstress condition in the rack when consi8ered to be primary stresses.
RESULTS OF UPPER-BOUND STRESS ANALYSIS PERFORMED TO EVALUATE ADEQUACY OF THE RACKS UNDER THERMAL LOADS Stress index values Rack Upper-Bound Components Analysis Allowable Floor-Leg Connection 0.682 1.0 Vertical Corner L-Beam 0.813 1.0 Base Assembly Girders 0.458 1.0 a
! S
SEB-10 Additional Response A concern was raised by the staf f concerning the spent fuel pool liner.
The liner load will not increase with the installation of the new racks.
The new racks weigh less than present racks. The largest new rack contains fewer asse=blies per rack than the largest existing rack.
The new rack feet have a larger area of contact with :he ficor, than the present rack feet.
For these reasons, the liner has a unchanged or lower probability of being penetrated with the new racks as co= pared to the present racks.
However, if the liner is penetrated, a leak detection system will identify and limit the amount of the leakage.
The leak detec' tion system consists of a series of channe.: in :he concrete floor and walls of the transfer canal and spent fuel pools jest beneath the stainless steel liner.
These channels are directed to twelve inter ediate collection points located along the sides of the transfer canal and spent fuel pools. These twelve collection points are piped to a central collection area which is located in the decontamination area on the ground floor.
Thus, any leakage through the stainless steel liner would drain to the central collection area.
This collection area then drains to the radio-active waste disposal system.
If the liner were penetrated at the boundary between two areas served by different collection points, the maxi =u= leakage though any two collection lines would be less than 120 gal / min. (resulting in a level loss rate of less than 0.12 in/ min).
This slow level loss rate gives the operator z=ple ti=e to discover the leak and close the isolation valves prior to any significant loss of level.
In addition, the maximum loss rate of 120 gym is well within the capability of the make up systems (see reference 4).
REFERENCES 1.
Letter from L 0 Mayer to the Director of Nuclear Reactor Regulation dated March 10, 1981 - Supplemental Infor=ation - License A=end:en:
Request dated January 31, 1980.
2.
Letter fres L 0 Mayer to the Director of Nuclear Reactor Regulation dated January 31, 1980 - License A=end en: Reques:
3.
Letter from L 0 Mayer to the Director of Nuclear Reactor Regulation dated February 3,1981 - Supple = ental Infor=ation - License Amend =ent Request dated January 31, 1980.
4 Letter frem L 0 Mayer to the Director of Nuclear Resctor Regulation dated June 10, 1980 - Supplemental Infor=ation - License A=end=en:
Request dated January 31, 1980.