ML20023B833

Forwards Response to 830405,06 & 07 Ltrs & 830419 Telcon Requesting Addl Info on Spent Fuel Pool Expansion

ML20023B833
Person / Time
Site:	Rancho Seco
Issue date:	05/02/1983
From:	Latham W SACRAMENTO MUNICIPAL UTILITY DISTRICT
To:	Stolz J Office of Nuclear Reactor Regulation
References
TAC-49241, NUDOCS 8305060429
Download: ML20023B833 (16)

Text

I O

e'~ SACRAMENTO MUNICIPAL UTILITY DISTRICT O 62 suu.

95813; (916) 452-3211 May 2, 1983 DIRECTOR OF NUCLEAR REACTOR REGULATION ATTENTION JOHN F STOLZ CHIEF OPERATING REACTORS BRANCH 4 US NUCLEAR REGULATORY COMMISSION WASHINGTON DC 20555 DOCKET 50-312 RANCHO SECO NUCLEAR GENERATING STATION UNIT N0 1 SPENT FUEL P00L EXPANSION, REQUEST FOR ADDITIONAL INFORMATION Your letters dated April 5, April 6, and April 7, 1983, requested additional information regarding our spent fuel pool expansion. Also, in a telephone conversation on April 19, 1983, Roger Pederson of the NRC requested additional clarification to our March 29, 1983 response to the NRC's request for additional information, dated March 9, 1983, on our spent fuel pool expansion. Attached is our response to your information requests.

s LLL K.

/

W. K. Latham Acting General Manager Enclosures 00l 8305060429 830502 DR ADOCK 05000 1

AN ELECTRIC SYSTEM S E R VIN G MORE THAN 600.000 IN THE HEART OF CALIF 0RNiA

e RESPONSE TO NRC LETTER DATED APRIL 5, 1983 As stated in our letter of March 29, 1983, the old racks will be decontaminated onsite to a level acceptable for shipment under the contractor's license (contractor yet undetermined). The racks will be shipped to a licensed recipient for further decontamination and ultimate disposition.

RESPONSE TO NRC LETTER DATED APRIL 6, 1983 1.0.

SUMMARY

.,e

• In the Rancho Seco plant, new-fuel storage racks are. located on the operating floor and consist of two parallel modules containing 10 spaces each. The storage cells in each row of ten are on a 21-1/8 inch center-to-center spacing, and the two parallel rows are 53 0.5 inches apart.

Criticality cdlculations (123-group AMPX-KENO) confirm that the effective multiplication factor ' f the storage racks, when fully loaded with standard o

fuel assemblies of 4 wt% 0-235 enrichment, will be less than 0.92 (including uncertainties) under all moderating conditions. The moderating conditions evaluated ranged from fully flooded (unborated water) to optimum moderation with a hypothetical low density moderator (e.g., mist, fog, or foam), in conformance with criticality requirements of the NRC Standard Review Plan, SRP 9.1.1.

Details of the confiming criticality analysis are given in the following paragraphs.

2.0 CRITICALITY EVALUATION

(

2.1 GENERAL Previous calculations (See Figure 4.3 of the Rancho Seco Spent Fuel Storage Rack licensing submittal) have shown that, for an infinite array of fuel assemblies on a lattice spacing of 21 inches, the k is 0.904 0.011 eff (95% probability at 95% confidence level) with fuel of 4 wt% enrichment. This represents the upper bound in reactivity for the new-fuel storage rack when fully flooded with clean, unborated water.

Handley(1,2) has calculated a peak in reactivity at very low water den-sities, such as might be postulated to exist if the moderator were a mist, spray, fog, or foam.

In a series of parametric calculations, Cano, et. al.(3) confimed the existence (calculated) of a peak in reactivity for a very low density moderator in the absence of absorber (poison) material between ful assemblies.

In the latter calculations, the maximum reactivity at optimum moderation could, under certain circumstances, exceed that for the fully-flooded condition in an infinite array of fuel elements.

For the finite Rancho Seco new-fuel storage rack (20 storage cells), the maximum reactivity at optimum

moderation does not exceed the limiting k value of 0.98 specified in eff

(

SRP 9.1.1, as confirmed in a series of conservative calculations described below.

2.2 GE0 METRIC MODEL The geometric model used for the criticality analysis of the new-fuel storage rack is illustrated in Figure 1.

Symmetry conditions effectively result in infinitely long rows of storage cells (y direction), each of in-finite height (z direction)..For conservatism, a reflector of full-density unborated water was assumed (x direction) for all cases, and neutron absorp-tion in rack structural material was ignored.

In practice, both neutron leakage effects in the y and z directions, and absorption in structural material, would reduce the reactivity of the actual storage rack to a sub-stantially lower value than that calculated here, particularly in the case of low moderator density.

2. 3 CALCULATIONAL MODEL t

Survey calculations to identify the effects of low moderator densities 4

were perfonned with the NULIF cell-homogenization code to generate cross-sections in four neutron groups; these cross-sections were then used in the 5

PDQ7 two-dimensional diffusion theory code to obtain the reactivity. The calculations identified a peak in reactivity (k

= 0.635) occurring at a eff 3

moderator density of - 0.05 gm/cm.

Because the cell constants and diffusion theory calculations may be suspect at very low moderator densities, the re-activities were re-calculated by AMPX-KEN 0,7 in the region of low moderator 6

densities corresponding to the peak in reactivity. These AMPX-KEN 0 calcula-l tions used the 123-group GAM-THERM 0S library (Nordheim resonance integral treatment in NITAWL for U-238), with each rod in the fuel assembly explicitly described in KEN 0.

l 2.4 ANALYTICAL RESULTS Results of the calculations, illustrated in Figure 2, indicate that the highest reactivity corresponding to optimum low density moderation is 0.792, l

t

4 3

which occurs at a hypothetical water-moderator density of 0.05 gm/cm. As

(

expected, the AMPX-KEN 0 reactivity value (0.792) is somewhat higher than the value obtained by diffusion theory calculations (0.635).

In the actual rack, substantially lower values of k w uld exist, since neutron leakage eff in the y and z directions would be significant.

The data in Figure 2 confirm that the limiting k f r the new-fuel eff storage rack occurs for the fully-flooded case (k

= 0.904 0.011),where eff the maximum l eff of 0.915, including uncertainties, is substantially lower than the limiting k of 0.98 specified in SRP 9.1.1.

This value of k eff eff (0.904 0.011) is essentially that of a single isolated fuel assembly immersed in water, since the lattice spacing in the new-fuel storage rack is sufficient to prevent any appreciable neutron coupling between fuel assemblies. Thus, the new-fuel storage rack design is adequate to assure the racks will not be critical for all moderator conditions (keff<0.92) when loaded with fuel of the highest anticipated reactivity.

_ = _

r t

e 12. 0 * :-

52.5* O 26.25' t

N'\\g Moderator g

Zero Flux 21.125*

= 10.5625' Boundary 2

Condition C

l Fuel A s s embly l

I A

'R e fle c tin g 1

4%

i Boundary y

Enrichment I Conditions L

_.___._J 3

>X i

FIGURE 1.

GE0 METRIC MODEL OF NEW-FUEL STORAGE RACK FOR CRITICALITY ANALYSIS.

m I

I I

~

~

I

..II

~ :

l

: ~.1:: -:

_ ~.::

.. 1:

~~

~

~

_ :_::: ::: :~

~~

~

~~

_ _i:~

1.0 1( E lj_._0 _;C__l

o I___ 0 ri si

, ;-; ___ 7 7 ;

MFl 7

a s,

_M d_1F -l) D 7

I c il l' n t i a n

- I~

~ :

~-

--~~

~

~

O

~

O.9

[

~

l~--

[

[ [

_] [_ [

~

~_

~

?

p T

,,,M,_

~

~

p

~

~.

~

p" 0.8 --p-7 d M P K

- P-E PI'3 f.

7

~

2 p

~~~

ik

..)

::1.:

~_ :-

~

~~~

~

, s x

0.7

~

~

~

y.~

..g i: ::':

T

- -: T- _-

~ !: : y

~

p f,0 M

. 1: -:::

~

~._~

~

~..--_::_:.:._

~

s

--f-

0. 6 -

s w

/

• s

~

~

M

-h 4U L I F--F l} Of

---;;;p 0.6

~

{

__~

. _____t__

i 0.0 O.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Water Moderator D e n si t y.

gm/cc FIGURE 2.

REACTIVITY VALUES OF 4% ENRICHED FUEL IN NEW-FUEL STORAGE RACKS FOR VARIOUS MODERATOR DENSITIES.

k.

REFERENCES 1)

C. R. Handley, Some Effects of Water Sprinklers on Array Criticality Safety Analysis _, Nuclear Technology,14, 71 (April 1972).

2)

C. R. Handley, Criticality Safety of Some Water Sprinkled Non-cubic Arrsys, Trans. Am. Nucl. Soc. 15, 806 (1972),

3)

J. M. Cano, et. al...Supercriticality Through Optimum Moderation in Nuclear Fuel Storage 6 Nucl. Technol. 48, 251 (1980).

4)

W. A. Wittkopf, NULIF-Neutron Spectrum Generator, Few-Group Ccnstant Generator and Fuel Depletion Code, BAW-426, The Babcock & Wilcox Co.,

August 1976.

5)

W. R. Cadwell, PDQ-7 Reference Manual, WAPD-TM-678, Bettis Atomic Power Laboratory, January 1967.

6) Green, Lucious, Petrie, Ford, White, Wright, PSR-63/AMPX-1 (code package),

AMPX Modular Code System for Generating Coupled Multigroup Neutron-Gamma Libraries from ENDF/B, ORNL-TM-3706, Oak Ridge National Laboratory, March 1976.

7)

L. M. Petrie and N. F. Cross, KENO-IV, An Improved Monte Carlo Criticality Program, ORNL-4938, Oak Ridge National Laboratory, November 1975.

v RESPONSE TO NRC LETTER DATED APRIL 7, 1983 1.

Local Buckling of Fuel Cell Walls e

The allowable local buckling stresses in the fuel cell walls are obtained by using classical plate buckling analysis.

The following formula for the. critical stress has been used (1].

2 2

6n E t (1) o

=

2 12 b (y_y2) 6 where E = 27x10 psi., t' =

.065",

b= 10.43".

The factor 8 is suggested in Ref. 1.to be 4.0 for a long panel loaded as shown below:

im g

i A

.1

-p4 r

e b

~

r b

C h

For the given data a

=4 x 948 = 3792 psi cr It should be noted that this calculation is based on the applied stress being uniform -along the entire length' of the cell wall.

In the actual fuel rack, the compressive stress comes from consideration of overall bending of,bhe rack

/

[1]

Strength of Materials, S.P. Timoshanko, 3rd edition, 1956, Part II, pp. 194-197

\\

4

\\

structure during a seismic event' and as such is negligible at the rack top and maximum at the rack bottom. It is conservative to apply eg. (1) to the rack cell wall if we compare a with the maximum compressive stress anywhere cr in the cell wall.

The output of the dynamic analysis programs DYNAHIS and EGELAST provides the time history of critical stresses at various levels in the rack and in the rack support feet.

In particular, an output is provided for the maximum direct plus total bending stress in the outermost cell wall at the bottom of the rack.

This translates into a maximum compressive stress in the cell wall at some critical time in the seismic event.

The output is given in terms of the stress factor R5 or R6 discussed in the Licensing Report. As defined in that document, the stress factor is the ratio of the actual stress to the allowable value (allowable = 15000 psi for 304 S/S based on yield and/or overall column buckling).

Therefore, the maximum

• compressive stress in the cell wall is given as o=R6 x 15000 psi Table 1 below shows critical values of R6 for different cases and the corresponding local compressive stress.

i k

(

t u

I

~

+

i f

Table it Compressive Stress in Rack Wall and j

Comparison with Local Buckling Stress for Typical Loading Cases l

Load Case R6 o (psi) ocr/0 l.

Full rack, type A 200 3000 1.26 2.

Full rack A, x

116 1740 2.18 j

direction quake 3.

Half loaded Rack, 114 1710 2.21 A type 4.

Full rack, type C 140 2100 1.80 i

a The-values of R6 appearing in Table 6.5 (Licensing Report)_are i

the maximum values for the entire structure including the support legs.

Due to the honeycomb construction of the racks, the seismic induced stresses in the support fee t also exceed those in the rack body.

Therefore, the stress factors reported in Table 6.5 are the largest values obtained and occur in the support feet.

The values given in Table 1 are obtained from the detailed outputs at the root of the honeycomb.

t Note:

(Critical loading condition of all cases examined) 9 y

er

-g w

g---

y

--_mm.m.

.m y

w.

m-

In the above, note that all results are obtained for SSE quake conditions.

The lowest safety margin acr/o is 1.26 which is deemed satisfactory since the actual compressive stress reaches 3000 psi only at the very bottom of the rack.

In addition, the calculation is based on a plate simply supported on four edges.

In the actual structure, the boundary condition is probably more nearly clamped on at lea t two opposite edges.

This would increase the factor 8 to a value near 7.0 in accordance with Ref. [1].

2.

Verification of Post Processor "EGELAST" The transient analysis program DYNAHIS computes nodal displacements and stress resultants at critical locations during the simulation of a seismic event.

Since the input acceleration time history was given in.01 second increments, we save the displacement and stress resultant information, for subsequent post processing, also at.01 second increments.

Therefore, during a 20 second seismic event, we would pass 2000 complete sets of results to EGELAST for evaluation. The purpose of EGELAST is simply to evaluate and printout maximum values of nodal displacements and stress factors Ri, based on an input set of data passed to it by a DYNAHIS run.

Verification of EGELAST was performed by a hand computation.

A sample SSE event was run using a typical data set to generate a DYNAHIS output file.

The output was limited to only 4 complete sets of data so that a hand calculation could be done directly from the results of DYNAHIS.

The output from DYNAHIS was visually scanned for displacement maximums and hand calculations were made for the critical stresses at the locations of interest.

The results of the hand calculations were in agreement with the output from EGELAST.

EGELAST, in conjunction with DYNAHIS, has been ilized in seismic qualification for Fermi II (Docket No.

341), Quad Cities I (Dockets (50-254) and Docket (50-265) Le the past.

3.

Analysis of Welded Joints Welded joints are examined under the loading conditions arising from thermal effects due to an isolated hot shell, and due to seismic loadings.

Under both sets of load conditions, the weld stresses are found to be below the

-+

y.

.s

allowable value of 24000 psi in shear that is given in Table NP329.1-1 of ASME Section III, Division 1, Subsection NF, 1980.

A.

A thermal gradient between cells will develop when an I

isolated storage location contains a fuel assembly emitting maximum postulated heat, while the surrounding locations are empty.

The thermal hydraulic calculations show that the maximum water temperature rise is 18*F.

Therefore, we can obtain a very conservative estimate of weld stresses along the length of an isolated hot cell by considering a beam strip uniformly heated by 20*F, and res tra ined from growth along one long edge.

N_b__

HEATbD 4 ELL. WALL.

-- X N

f rxwascarsam ----,

Q

--=

L Wno uME N

Using shear beam theory, and subjecting the strip to a uniform temperature rise AT = 20*F, we can calculate an estimate of the maximum value of the average shear stress in the strip.

The strip is subjected to the following boundary conditions.

a.

Displacement Ux(x,y) = 0 at x=0, all y and at y

= w/2, all x.

b.

Average force N acting on the cross section Hxt, x,

0 at x = L, all y.

=

The final result for wall shear stress, maximum at x=L, is found to be given as Ea AT MAX "

.931 6

9.5 x 10-6 in/in

• F and AT l

where E - 28 x 10

psi, a=

j 20*F.

=

Therefore, we obtain an estimate of maximum weld shear stress in an isolated hot cell, due to thermal gradient, as

' MAX 5775 psi

=

l I

l

B.

The critical weld locations, when the loading is seismic, are at the bottom of the rack (at the connection to the baseplate) and in the welds on the support legs. The results from the dynamic analyses using DYNAHIS and EGELAST are surveyed and the maximum loading used to qualify the welds in these locations.

The ASME code allowable value of 24000 psi is used on allowable weld stress.

All welds are qualified using SSE seismic results.

The worst loading case is the case 1 analysis of Rack A.

(see licensing document).

The welds at the rack base are 1/8" fillet welds.

The shear stress t in the weld throat induced by a normal stress o in the rack is given as ot/1.414h (h = weld size)

(2) t =

The above result assumes a continuous weld at the critical location and a uniform o along the length of the weld.

For the weld between the rack base and the cell walls, we have r

.065" +.049"

.114" t

=

=

h=

.125" Examination of the structural acceptance factors from the transient analysis outputs of all of the simulations yields the maximum value for R6 at the rack base as

.2 (case 1

- Rack A)

R6

=

Since R6 o/15000, we have, at the weld in question:

=

o=

.2 xc 15000 = 3000 psi Therefore, using eq. (2), and doubling the stress to account for skip welding, we obtain 6000

.114 7

=

x

= 3870 psi rack bas 1.414

.125 The welds at the bottom of the support legs (between the legs and the pad) are full penetration welds.

Therefore, no analysis is required in these areas.

At the top of the support legs, we find that the maximum

(

i

value for the structural acceptance factor (or stress is 1.17 (case 1 - Rack A).

Therefore, using factor) R6 this conservative value,we obtain o = 1.17 x 15000 17550 psi

=

Therefore, the weld shear stress predicted from eq. 2 above with t = 1.25", h = (.25" +.875" +.25")/2 =

.6875" (dimensions for weld calculations obtained from the drawings) is calculated as T

= 22567 psi SUPPORTS y

I l

1 i

RESPONSE TO APRIL 19, 1983 TELEPHONE CONVERSATION 1.

The estimated 20 man-rem was obtained by polling utilities and contractors from similar efforts. This is a fairly accurate estimate. Of the 20 man-rem, approximately 70% will be to the divers for removal and instal-lation, 20% to the Decon personnel, and 10% to all support personnel.

2.

The radiation surveys from which the 6-8 mr/hr was obtained shows lower readings at the pool sides and higher in the middle. The major contributor to this higher reading is the fuel handling equipment.

This equipment is scheduled to be decontaminated during the rerack effort.

i b