ML19260C465

Forwards Draft Acceptance Criteria for Mark I Containment long-term Program.Requests Comments by 790814.Meeting to Resolve All Comments Scheduled for 790814

ML19260C465
Person / Time
Issue date:	08/02/1979
From:	Charemagne Grimes Office of Nuclear Reactor Regulation
To:	Office of Nuclear Reactor Regulation
References
REF-GTECI-A-07, REF-GTECI-CO, TASK-A-07, TASK-A-7, TASK-OR NUDOCS 8001030566
Download: ML19260C465 (57)

Text

f,b 9

kR Rf C

/

o UNITED STATES

~,,

NUCLEAR REGULATORY COMMISSION o

{

,E WASHINGTON, D. C. 20555

%e

/

s,,e AUG 2197d MEMORANDUM FOR: Those on Attached Distribution FROM:

C. I. Grimes, A-7 Task Manager

SUBJECT:

PROPOSED ACCEPTANCE CRITERIA FOR THE MARK I CONTAINMENT LONG TERM PROGRAM Enclosed is a draft of the acceptance criteria for the Mark I Containment Long Term Program. These proposed positions have been assembled from the recommendations and assessments of the individual staff reviewers and consultants. Those sections of the attachment with a vertical bar in the right margin are evaluation sections on which certain criteria are predicated.

In order to expedite the issuance of these criteria and to begin imple-mentation of this program, your comments should be completed by August 14, 1979, at which time the staff and our consultants will meet to discuss, and hopefully resolve, all of the comments. This meeting is presently scheduled for August 14 at 9:00am in Room P130A. A subsequent meeting with the Mark I Owners Group representatives is tentatively scheduled for August 15, 1979, to discuss these criteria. A copy of the transmittal letter to the Mark I Owners Group is contained in Enclosure 2.

T C. I. Grimes Plant Systems Branch Division of Operating Reactors

Enclosures:

~

As stated

[O G

80010 @

1700 119

v-w 01 INTRODUCTION The purpose of the Mark I containment Long Term Program is to perfom a complete reassessment of the suppression chamber (torus) design to include suppression pool hydrodynamic loads which were neglected in the original design, and to restore the original intended design safety margins of the structure. This reassessment will be accomplished by a Plant-Unique Analysis (PUA) for each Bi(R plant with a Mark I containment, using load specifications and structural acceptance criteria that are appropriate for the life of the plant.

The fo11owin6 acceptance criteria have been developed from the staff's review of the Long Tem Program Load Definition Report (LDR), the Plant Unique Analysis Applications Guide (PUAAG), and the supporting analytical and experimental programs conducted by the Mark I Owners Group. These criteria specifically address the dynamic loading conditions. Unless otherwise specified, all other loading conditions and structural analysis techniques (e.g., dead loads and seismic loads) will be in accordance with the plant's approved Final Safety Analysis Report (FSAR).

CONTAINMENT PRESSURE AND TDiPERATURE RESPONSE The pressure and temperature transients for the drywell and wetwell shall be detemined by the use of the analytical models and assumptions set forth in Section 4.1 of the LDR. These techniques have, in the past, been found to provide conservative estimates of the containment response to a LOCA, by comparison to the staff's CONTEMPT-LT computer code. Plant-specific results, for each break size, shall be presented in the PUA, along with t.he input conditions, in sufficient detail to allow the staff to perfom confinatory analyses to assure proper application of these models.

1700 120

)

02 For the Design Basis Accident (DBA), the mass and energy release rates from the primary system are to be calculated with the Homogeneous Equilibrium Model (HEM), applied in a non-mechanistic reactor system which does not s

take credit for pressure reduction in the piping during the early portion of blowdown and conservatively assumes all liquid flow during most of the remainder of the blowdown. By comparison of the mass and energy release rate predictions of the GE model to those of a conservative RELAP-4 analysis, we conclude that the GE model will provide a conservative prediction of the critical flow rates for a postulated double-ended recirculation line break for BWRs with a Mark I containment system.

The staff has determined that the application of HEM to calculate the mass and energy release rates from the primary system will not necessarily provide conservatively high release rates for the Intermediate Break Accident (IBA) and the Small Break Accident (SBA). However, the purpose of the IBA and SBA is to provide a spectrum of event combinations, where the priman loading conditions are steam condensation and SRV discharge loads. The primary loading condition affected by the HEM model, and the assumed break sizes for the IBA and SBA, is the containment pressure and temperature response. For the IBA and SBA, the containment response is of secondary importaice to the loading condition and the primary loading conditions are calculated independent of the containment response. On this basis, we conclude that application of the HEM model for the IBA and SBA event combinations is acceptable.

The timing and duration of specific loads are based primarily on the plant-specific containment response analysis for the pool swell-related loads, while the condensation periods are non-mechanistically maximized. However, 1700 121

4 03 the duration of SBA condensation loads are assumed to be limited by manual operation of the Automatic Depressurization System (AIS) at 10 minutes into the accident. Therefore, as part of the PUA, each licensee shall identify the procedures (' including the primary system parameters monitored) by which the operator will identify the SBA, to assure manual operation of the AIE within the specified time period.

VENT SYSTEM PRESSURIZATION AND THRUST LOADS The vent system pressurization and thrust loads during pool swell shall be defined in accordance with the procedures set forth in Section 4.2.1 of the LDR. These loads have been derived by assuming a distribution of flow losses which will maximize the vent system pressures, and an examination of the QSTF test data to maximize mass flow effects. The vent clearing time is significant, especially for those plants that propose operation with differential pressure control, because it establishes a transition in the vant load cal culations. Therefore, the vent system pressures and the vent clearing time shall be determined by the GE containment response model, which incorporates a virtual mass (equivalent to an extended downcomer length) in the vent clearing model. These procedures will provide conservative estimates of both the initial vent system pressure transient and the subsequent vent system flow effects and are, therefore, acceptable.

1700 122

04 NET TORUS VERTICAL PRESSURE LOAIS The downwarti and upward net vertical pressure loads on the torus shall be derived from the series of plant-specific @TF tests, in accordance with Section 4.3 1 of the LDR. However, based on our review of the pool swell tests conducted by the Mark I Owners Group and confirmatory tests performed for the NRC by the Lawrence Livermore Laboratory, we will require that the following margins be applied to each loading phase UP

+ 0.215 (UPmean)

UP

=

nean DOWN,,

+ 2 x 10-5 (pag

)2 D0ilN

=

where "mean" refers to the statistical average of the UTF plant-specific test runs. These margins shall be applied to the QSTF "mean" load function prior to scaling the load function up to full-scale equivalent conditions.

The margin on the upward pressure load includes 15% to bound the uncertainties arising from comparisons of all the two-dimensional and three-dimensional upward load test data. The remainder of the upward load margin and the margin for the downward load reflect the randomness observed in the QSTF test results.

The margins specified above may be reduced or omitted where minimum conservatisms in the QSTF tested conditions for a specific plant can be demonstrated by the application of the QSTF sensitivity test series (NEIE-23545-P). The sensitivity tests may not be used to adjust the mean torus vertical pressure loads. If the plant configuration is changed to the extent that the QSTF test series no longer represents a conservative configuration of the plant, then a new series of QSTF tests shall be performed.

1700 123

t 05 For those plants that use drywell/wetwell differential pressure control as a load mitigation feature, an additional structural analysis shall be performed assuming a loss of the differential pressure control to demonstrate the capability of the containment to withstand this extreme condition, as specified in Sections 5 3, 5 4, and 5.6 of the PUAAG. For this analysis, a single plant-specific @TF test run may be used to define the loading function; however, the downward and upward loading phases shall be increased by the margins specified.above for the base analysis.

TORUS SFELL PRESSURE - POOL SVELL The spatial distribution of the torus shell pressures during pool swell shall be defined from the plant-specific @TF test results and the azimuthal and longitudinal distribution factors defined in Section 4.3 2 of the LDR.

However, the UTF results shall be adjusted to incorporate the margins specified for the net torus vertical pressure loading function; i.e.,

the asarage pool pressure shall be increased by the margin specified for the downward load during the downward loadin6 phase and the airspace pressure shall be increased by the margin specified for the upward load during the upward loading phase. Although the distribution factors have been based on avera6ed test results (W tests for the azimuthal and 24 for the longitudinal), we conclude that the margins in the base load function adequately cover the uncertainty in the local pressure definition.

1700 124

06 COMPRESSIBLE FIDW EFFECTt p' SCALED 300L SWELL TESTS The QSTF plant-unique and sensitivity test series are based on a " split-orifice" vent flow scaling relationship. Preliminary calculations performed by EPRI and GE indicate that compressibility effects, which could not be accurately scaled in the testing program, could result in a higher loading condition at full-scale conditions than that derived frcs " scaled-up" test data. The original intent of these analyses was to provide justification forthescaledflowdistributionintheEPRI1/12-scale, tree-dimensional pool swell test progras.

The loading functions predominantly affected by this finding are the tents downward and upward vertical pressure loads and the vent header pool swell impact timing. Based on our review of the preliminary analyses performed by EPRI and GE, which were presented in a meeting with the staff on July 24, 1979, we conclude that there is sufficient margin in the torus vertical' pressure loads (previously specified) and in the header impact timir4 techniques to accomodate this uncertainty. Further, we do not believe that the Mark I implementation should be delayed during the time that it will take to resolve this concern. We will require, however, that the Mark I Owners Group complete the assessment of compressible flow effects and justify the conservatism in these load specifications prior to the issuance of our Safety Evaluation Report, which is currently scheduled for December 1979 In the event that this conservatism cannot be demonstrated, these two loading conditions will have to be reassessed.

ON 1700 125

07 VENT SYSTD4 IMPACT AND DRAG IDADS A. Vent Header Impact and Drag Loads The load definition procedures set forth in Section 4.3 3 of +,he LDR are acceptable, subject to the following clarifications:

1. The experimental data of local vent header pressure in each of the Mark I plants shall be obtained from the QSTF plant-unique tests.

2. The specification, for each Mark I plant, of the pressure inside the vent header relative to that in the torus airspace at the time of water impact on the vent header shall be determined from the QSTF plant-unique tests.

3. The plant-unique header impact timing (i.e., longitudinal and circumferential time delays) shall be documented in each PUA, and are acceptable subject to confirmation from the assessment of compressible flow effects in scaled pool swell tests as previously discussed.

s B. Downcomer Impact and Drag Loads The load definition procedures set forth in Section 4.3 3 of the LDR are acceptable, subject to the following clarifications. A pressure of 8 psid is to be applied uniformly over the bottom 50 of the angled portion of the downcomer, starting from the time at which the rising pool reaches the lower end of the angled section and ending at the time of maximum pool swell hei ht. The pressure is to be applied perpendicular 6

to the local downcomer surface. The structural analysis for the downcomer impact shall either be dynamic, accounting for the approximate virtual mass of water near the sulnerged parts of the downcomer, or a dynamic load factor of two shall be applied.

1700 126

d8 C. Main Vent Impact and Drag Loads For the main vent, the acceptance criteria specified for impact and drag loads on other structures above the pool apply, except that the structure shall be subdivided into smaller sections and the impact and drag loads calculated separately on each subdivision.

6 1700 127

L POOL SWELL IMPACT AND DRAG ON CITER INTERNAL STRUCTUPES The impact and drag loads for int,ernal structures, other than the vent header, downcomers, and vent header deflectors', as specified in Section 4 3 4 of the LDR shall be modified on the basis of a cylindrical (e.g.,

pipes) or an exposed flat surface (e.g., "I" beaas). The load specification for these two geometries is as follows:

A.

Cylindrical Structures For cylindrical structures, the pressure transient which occurs upon water impact and subsequent drag is depicted in Figure 1.

The parameters in Figure 1 shall be defined as follows:

1. The pulse duration ( T ) is specified as a function of the impact velocity:

t = 0.013 D for V { 16 ft/sec t = 0.21 (D/V) for V>

16 ft/see where D is the cross-sectional diameter of the target in feet.

2. The hydrodynamic mass per unit area for impact loading shall be obtained from the correlation (cylindrical target) depicted by Figure 6-8 in NEDE-13426-P.

3. The impulse of impact per unit area shall be determined by:

H V

h I

p T 144 gc where I is the impulse per unit area (psi-sec), M /A is the H

2 hydrodynamic mass per unit area (1bm/ft ) and V is the impact velocity (ft/sec).

1700 128

10

4. The pressure due to drag following impact shall be determined by:

PD" 2

(144 g j where P i8 the aver 88e drag Pressure acting on the projected area D

of the target (psi), C is de hag coefncient as denned by Figm 2, D

3 and [ is the density of water (lbm/ft ).

5. The maximum pressure (P

) shall be calculated from the impulse per unit area, the pulse duration, and the drag pressure as follows:

(I - 0.5 T P }

p D

P,n =

• P D TI a where P is the maximum pressure averaged over the projected area, and I, is a dimensionless empirical parameter defined by Figure 3.

6. The line connecting P to point A in Figure 1 is defined by:

1 - exp[a(1-t/T )]

line A =

(P

-P)+P g

D where a is a peakedness parameter defined by Figure 4, t is time (sec), P is the maximum pressure averaged over the projected area, and P is the average drag pressure acting on the projected D

area.

1700 129

ll B.

Flat - Surfaced Structures For flat-surfaced structures,.the pressure transient which occurs upon water impact and subsequent drag is depicted in Figure 5.

The parameters in Figure 5 shall be defined as follows:

1. The pulse duration ( t ) is specified as a functica of the impact velocity:

T = 0.0016 W for V { 7 ft/sec t = 0. 011 W for V > 7 ft/sec V

where W is the width of the flat structure (feet) and V is the impact velocity (ft/sec).

2. The pressure due to drag following impact shall be determined by:

D [8V D* T, 144 g j where P is the average drag pressure acting on the frontal area D

of the structure (psi), the drag coefficient CD = 2 (flat strips normal to flow, independent of Renolds number), and [ is the 3

density of water (1bm/ft ).

3. The hydrodynamic mass per unit area for impact loading shall be obtained from the correlation (flat targets) in Figure 6-8 in NEDE-13426-P.

4. The impulse of impact per unit area shall be determined by:

bI I

V p

T 144 g l e

where I is the impulse per unit area (psi-sec), g/A is the hydrodynamic mass per unit area (1bm/ft ), and V is the impact velocity (ft/sec).

1700 130

12

5. The maximum pressure (P

) shall be calculated from the impulse per unit area and the drag pressure as follows:

2I

+P

' P,

=

D C.

Loaa Application For both cylindrical and flat structures, a margin of 35% shall be added to the loads derived above to obtain a conservative design assessment load specification. This load specification corresponds to impact on " rigid" structures. When performing the structural dynamic analysis, the " rigid body" impact loads shall be applied; however, the mass of the impacted structure shall be adjusted by adding the hydrodynamic mass of impact. The value of the hydrodynamic mass shall be obtained from the appropriate correlation in Figure 6-8 in NEDE-13426-P.

1700 131

13 P

m=

E c.,

e UW O.a

>.e c00 e

5 point A o

~

< P e-D I

p P

s s

/

s e

i time 7

Figure 1.

Pulse Shape for Water Impact on cylindrical Targets

,i..i-

,i.

i i

i.

i i 4

i.,

g

.,2

.i i

I i

l

...... i i,i... i i-

%.N i.

i,

iiiv i i ' i

'i

,i,

,i,

.i.

i i x

2+

.i7

.. i e i t

,i

\\i i

i,.,

i

-

• i i,

- r' i i '

,,i.

1 i -

i Ni T

. i..

- i - '

,,i, i

. '. i il '

i-i x

ii,,

i.,

i.,

, i >

' 'i' i

.i

.i, ii.,

.i

,i i

A i,,

.i.

i e

N i

i

'i-

'A i i

• ixi C

i i

i D

i,.

', ' '...'o

' i

i' i,

i.

i i

' < i i:.

i i-i,n g

i,,,

tei i

I

\\

. ' i-i-

'i i T ' ', '

o<

6-i i i i

i ii.

,,n iV i..

,,,i o

i

. i.

> i '

,i<

i.

i.

,,.i i,,

i i

t i i...

i e

i.

i, i.

i 1.. '.

5

.i-,

, i.

i i-4.,

1.

^i

,1 i

'i

, i - i '.

i i i i > i i '..

i*

,,i.

- i i

..n-i

... i i

i ii.

I I

' I

,,6, i,

,i,

,.l.

, i

,i i i,,.

.i a

i..

i i

..o n

+

, i,..

i

- i i.,

.... i. - i i i i-i 4

i i i..

ii

,1

'io

- i i

i.-

i.

,'i i

. i

,ii, i

i.

i,

i.... > - >

i.

,i.

.i..

1 i

- i. i i-

+ i t i -

i,

.n...

[

,l6, i

0 0

I

' l '

,, 4 - i l

l i

l 0

2.

4.

6.

8.

10 Re(D)

Figure 2.

Drag Coefficient for Cylinders Following Impact 1700 132

2

14 m =..

.:.:_n4==

=. I=.. u

=.;n a

..}:... =.. : r...m....=. =...u -.

.. t _=

,.:...m_ =..::. m..

t

_4

.4

~

,...... ~.!..., _.._. _... _... _... _...

..a

.. _.. j....!.._.

.I..

..J.

...j

-- m _...+1_.._-.;!..+t._.=..

~

....... -.:.:d: n

n._.~-:..f.';n_.m._

--j+n:.

_~.~.

..._..'L'*"..'s.]..

-+ - :

t.

_._.4...

. m_...

~:':

2-- 7:!..

....!_..-.*.T_

2::1=

......_.::- in..n.- n r.....~. _;- +.. ;..._: ::!-_ _.... i_

___1.__.;. art:~:

_.d :...

1..

..... - - -.___. _4_

...3.$._._

. 33..._. g _-

T*

-- ! C '..a._

._.t,._.._.

g.

._.t._

--1

._,._..._ _1. _

.3..._.

..4 - _ _. _

.4

...)._.

.1.... _.

__-3.__

.. 1.......

j.

1

.. +. _._ _ _ _.. _.. _ _. _.. _

. 1 __..

_w.

- 2. ' a _.

~

i

?"

~:t----

..__=. g7:.

.s.

-_c. u

l. _

_1. :.:..*

.__t._._._

t.

l: -

1-,

. J,.,: -

1:,.,..

..o...

....g.

._g._

..g..

l-

t-

~.1 =: h, n_i--,,,, -. _ _

___I'~.,

r b ---

-'^

..._ J..+.....

.t

..I

.i...

..4...

t.._.

. 1....

.j.

..I

. l.. _

_ j.-

g.4.

g..

_..j..

. l.._

___.}... A,. a_,

_. ; g...

...1,..

.g.

g_.

.,, - 7.-_._.,

. 1_

77

.t.-

t. u r.

.1:

.? ! -

{:

T: _.

_ ;.. -._O.._.__....

.. _1 t.

. t.

:-~.:: =.-_: :__

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

v..

..__..__.__.!_._~_.__.

. l..

_ 1

. _.. 1. 2..,........ 1,6.... _

...4_..

O.

. _. _.... Fd,._.. _._

_.s.u.4.

1 _..

6.. [

.1.._

t-_

.s.

.. 3. _4 4 w gs. ch.,,,....I...

.7.

g..,.,....

.._.n_._. 7__

.. _. ~...,... -

_.3

_.__.p__

7 - -

r.._Ta.gure 3,,

.D.s..mnsionless -D.epu.he Par.ameter =..... _... _.

._: r 4.

_.._._+_. _..._

.1.

...__. _ _. g _

.. _.. _.._.t...

....t... _._.. _. _.

1

.__.__..__I,..__.

. _ _ ~. _,

....__r___. _ _ _. _. _.

...4..

. a.

t Oe _.

.. t...

.L____

._.....1

..__.____.___t_.._____

__p.

1

__..._.4

.....t...

t...

_5 e _. __.

.. /...

. _. _.,_ _. _. f. _..

_4

.__j_.._.._4

. L. _

4

._._1._.._ _. _

. _ _.. _. _...______...I___

.._4...

..L_..._..T-___-

- - - - + - - - -

__.t.___..._-_ _ _-

_-h_--

- - +

_____-b._.__--

f.__.__._._._._.._.__._

.___a.__.

_m

.___ _. _.. _ _ ~. _. _ _

..t __.

_.4____

. _ _ + _

1

.. _ _. _ _ _ + _ _ _.. _

g _

__4._

J o __..

__1

_ _ _ _. _ _ _ __.__.L____._._..

.__4._..

.3 I._._._..__. _

..... ~.

_ _ _ _ _ _. _ _ _ _.._._.,.._.,O..._

... -..... _ _ _ ______j

.___w,_..._.._._-,.._...__

.___L_____._.

._.._.__._.__.m.6,.__.

__*,._.____..__...__1 t/ 2 eC J. _. ____.._____ _ _ _ _.. _ _. _. _ _

_ _ _._. N..V.

.4

. _, _.... _.. __. - ~ -

Figur_e.._A.g_____D_ _im6.ns. f.un]

es. 3....N... ise. 4... hape..~.Pa. ramete, r. :_". _=__2

_._-_._..=__4.

4 2

...__.9.._._

.. _.. _ _.__.__{.

~.1_

s O &<

~~

15 max E

o.

E.

En E

I gPD P /

/

/

/

\\

l t

time Figure 5.

Pulse Shape for Water Impact on Flat Targets 1700 134

16 FROTH IMPINGEMENT AND FALLBACK LOADS Froth is generated by (1) impact of the rising pool surface on the vent header and (2) bubble breakthrough, as described in Section 4.3.5 of the LL9. The following load specfication was derived from the high-speed film records of various pool swell tests and an analysis of pool acceleration following vent header impact. The impingement loads for Region I and Region II and the froth fallback loads, as described in Section 4.3.5.

shall be defined as follows:

P Y -

f p

144 gc where:

P = froth impingement pressure (psi) f

= froth density (lb /ft )

of m

V = froth impingement velocity (ft/sec) g = gravitational constant c

Region I: The froth velocity shall be based on a source velocity equal to 2.5 times the maximum pool surface velocity prior to vent header impact, which is corrected for subsequent deceleration due to gravity starting at the 45 tangent on the bottom of the vent header, as shown in Figure 4.3.5-1 of the LDR. The froth density shall be assumed to be 20% water density for structures or sections of structures with a maximum cross-sectional dimension of less than or equal to one foot, and a proportionately lower density for structures greater than one foot; i.e.,

The load shall p = (0.2/x) pg, where x is the dimension in feet.

be applied in the direction most critical to the structure within the 90 sector bounded by the horizontal opposite the vent header to the vertical upward. The load.shall be assumed to be a rectangular pulse with a duration of 80 milliseconds.

)7

17 Region II: The froth velocity shall be based on. source velocity equal to the maximum pool surface velocity prior to vent header impact, which is corrected for subsequent deceleration from the elevation of the bottom of the vent header. The froth density shall be assumed to be 100% water density for structures or sections of structures with a maximum cross-sectional dimension less than or equal to one foot, 25 water density for structures greater than one foot, and 10% water density for structures located within the projected region directly above the vent header. The load shall be applied in the direction most critical to the structure within the 15 sector of the upward vertical.

4 The load shall be assumed to be a rectangular pulse with a duration of 100 milliseconds.

Fr.11back:

The froth fallback velocity shall be based on the freefall velocity from the upper surface of the torus shell directly above the subject structure. The froth density shall be assumed to be 2% water density, with the exception of the projected region directly above the vent header which is 10% water density. The load shall be applied in the direction most critical to the structure within the 1 5 sector of the 4

vertical downward. The load shall be assu=ed to directly follow the froth impingement load, with a duration of one second.

1700 136

VEC HEAIER IEFIEC70R IDADS The load definition procedures set forth in Section 4.3 9 of the LDR are applicable only to the four deflector types'shown in Figure 4.3 9-2 of the LDR, and are generally acceptable, subject to the following constraints and/or modifications:

A. An individual plant may choose to use deflector load data taken directly from the QSTF plant-unique tests. This technique is subject to the followin6 requirements:

1. If the QSTF deflector load measurement does not have a sufficiently fast response time to resolve the initial impact pressure spike for the deflector types 1 - 3, inclusive, the loading transient shall be adjusted to include the empirical vertical force history of the spike shown in Figure 1.

This impulse need not be applied for the type 4 deflector.

2. The QSTF plant-unique loads shall be adjusted to account for the effects of (a) impact time delays and (b) pool swell velocity and and acceleration differences which result from uneven spacing of the downcomer pairs. The correction technique shall be evaluated for the instant when the undisturbed pool surface would have reached the local elevation of the center (half-height elevation) of the deflector. The proposed three-dimensional load variation and timing is acceptable, subject to confiznation from the assessment of compressible flow effects in scaled pool swell tests, as previously discussed.

3. In applying the load to the deflector, the inertia due to the added mass of water below the deflector shall be accounted for. The adde?

mass per unit len6th of deflector may be estimated by:

1700 137

19 I

"a V w where: I = total impulse per unit len6th associated with the impact transient, Y = impact velocity w = deflector width (as ~shown in Fi ures 2 - 5) 6

3. The deflector load definition which is based on empirical expressions for impact and drag forces together with semi-empirical, plant-specific definition of the pool swell velocity and acceleration transients, as described in Sections 4.3 9 1, 4.3 9 3, and 4.3 9.4 of the LDR, is acceptable, wi'th the following modifications:

1. The impact transient and " steady dra6" contributions to the load shall be computed from the cor: elations shown on Figures 2 - 5.

for deflector types 1 - 4, respectively. For times past the periods shown, the last value shall be extended for the duration of the transient.

2. The proposed three-dimensional load variation and timing is acceptable, subject to confirmation from the assessment of compressible flow effects in scaled pool swell tests, as previously discussed.

3. The gravitational component of the acceleration drag shall be included in F, as defined in NEDO 24612.

A

4. In computing the deflector response to the load, the added mass of the water shall be accounted for, as described in the previous load definition technique.

1700 138

t0 m

DR\\f1 F = vertical upwani force on deflector per unit length d = diameter of cylinder in deflector types 1-3 V = impact velocity P = water density t = time from beginning of impact 6

4 oc e

g-J w w Ju o JB,2..b/d m f

e-1*O 4

V Y

4 A

w 0

6 4

n e,

o.i a

~

w d

Figure 1.

Impact Force Transient for Addition to the Empirical Data for Deflector Types 1 - 3.

1700 139

21 DR\\fi "2D0l'Y$.

2D**2D

2..t

.m s

.r J

W5 l

v

-u u,,,,.-

u V e

.4

n

.... s... x a-

.n c <.

( c k, 4 S s. 2 L..

o i

' t i

O

'O O.5 f.O.

f V&

7 s

Figure 2.

Impact & Steady Drag Force Correlation for Type i Deflector 1700 140

7' q

d g

he I

h.h-I

~

i gA 1

E

-g j

44 b

.1 o

t

.g n

.'g'

~.

i,

'h4 l

041

.. ; i

).

j N

a Y

x N

00 c

V

~

\\

o N

N k

D

[

i' i o

T

. >.i..: '

k.

o

&~

u m

2 fi

,g O

s j

s

• P i.

t p*. ~.+. --s

,o

..u. s t'

ob t

' ' z.

e e

i

g.,. 4 j-J g

s;$, i..A.,L E

't.1

\\

~.

y n)

IN.

,s

'e

-p L.

I

n

~

J f

s, O

r r '., s :-

M.'

hr,

.:.- t, -

1700 141 8

.- N s

f.

sc s

1

j b

~

W-g 1

d

(

}g d o.me s- -. 4.ss

.f UY

} 7,7 // r~7 g' I

/

T-

.3 (.

j s t

V l

bl4 g_

l r

1-j..

t

/\\.

',-h=

_ o..y l. ;.,

p..,/

.. a

,T s

1..

.\\ -

! \\

.l.

l Vf-o.

~ a0 n

i 05 p'

to 4

n -

b k

=

f,'

f l

Fi ure 3 Impact & Steady Drag Force Correlation for Type 2 Deflector 6

1700 142

24 2002 Ogggg

~

f Q', W-W YMO*$b V t

_ g4 i ' ' g s ri r r g'

_-#.9o y

I f

f I

A.

I ii l

4-I f

7 I.

g 1

/

I L

i l

~ o.7

/

L'..

i

, ye a

.I

.t...

l a

g l

9

..,I.c f g

3 s

l.l 4,, l

....,l

. '; -A te Figure 4 Impact & Steady Itag Force Correlation for Type 3 Deflector 1700 143

25 DR\\f1 C22EE Lr

.s.

V

' '{'[

JT r r.up, r

ar ea r-

- g,g y

S P

?-[.*

l 3

g l.

d j

m c.7

.s i.

a

___t RW g

J.4 i.(.

-l f r:

f g,

gr-g-

4 M

d:.

4

.{

Figure 5 Impact & Steady Drag Force Correlation for Type 4 Deflector 1700 144

26 CONDENSATION OSCILLATION LOADS The following criteria have been developed in consideration of the fact that the " condensation oscillation" loa'ds ('i.e., higli vent flow rate with low air content) have been derived from a single FSTF test run (M8). The condensation oscillation regime is_ a hamonic fonction and, therefore, stastical variance or load magnitude uncertainty cannot be established from one test run.

Although we conclude that the M8 tested conditions are conservative and prototypical for the Mark I design, a reasonable measure of the uncertainty in the loading function is necessary to assu:e the margins of safety in the containment structure. However, based on our assessment of the phenomenological studies conducted by the industry and the NRC Office of Nuclear Regulatory Research, we believe that the following load specifications are probably conservative and form a sufficient basis to proceed with implementation of the Mark I Long Tem Program. We will require that the Mark I Owners Group confirm this assessment by performing at least two additional large break, liquid blowdown tests in FSTF. A commitment to perform these tests and the associated schedule will be required before the issuance of the staff's generic Safety Evaluation Report, which is currently scheduled for December 1979.

Ob 1700 145

CONIENSATION OSCILLATION TORUS SHELL_IDADS The load definition and assessment procedures set forth in Section 4.4.1 of the LDR for the condensation oscillation loads on the torus shell are acceptable, subject to the following confirmatio'n

1. Provided that the " rigid wall" load derivation technique described in NEDE-24645-P is demonstrated to be conservative, in response to Question 7 in our request for additional information (D. Eisenhut, NRC, to L. Sobon, GE, dated July 30,1979).

2. Provided that sufficient justification can be provided to exclude a condensation oscillation asymmetric loading condition, in response to Question 2 in our request for additional information.

3. Provided that the uncertainty in the load magnitude is demonstrated to te less than the demonstrated conservatisms in the load specification, by the testing program described above.

For clarification, the load specification set forth in Section 4.4.1 of the LDR shall be used in conjunction with a coupled fluid-structure analytical model. The condensation oscillation lom &E for the IBA is a continuous sinusoidal function with a peak amplitude and frequency range of that specified for the " pre-chug" load. We will require that the conservatism in the IEA condensation oscillation loads be demonstrated as part of the response to Question 7 in our request for additional information for the

, specific flow regimes of interest.

CONDE'ISATION OSCTTIATION DOWNCOMER LATERAL IDADS A. Untied Downcomer Lateral Loads The condensation oscillation downcomer lateral loads for untied downcomers shall be defined as described in Section 4.4.3 of the LDR, based on the methodology in NELS-24537-P. However, in computing the dynamic load factors 1700 146

28 P

=P aax g

i P,y = maximum static equivalent lateral load for plant-where unique downcomer P

= maximum static equivalent lateral load in FSTF g

DLF = plant-unique downcomer dynamic load factor DLFg = FSTF downcomer dynamic load factor the plant-unique loading condition shall be derived as follows.

1. The damping factor assumed in the DLF shall be 2% (for both the plant-unique and FSTF) if the value of % for the plant is larger than that in FSTF, and 5% (for both plant-unique and FSTF) if the 94for the plant is less than that in FSTF.

2. Assume that A is in the range between 4 and 8 H:, and specify the loading condition as that which produces the maximum response in this range.

As used here, A is the driving frequency associated with the condensation oscillation lateral loads and e is the natural frequency of free lateral vibration of the downcomer (the one closest to resonance with A, if the two major frequencies differ).

B. Tjpd Downcomer Lateral Loads For tied downcomers, the structural response shall be assessed by a dynanic analysis of the tied downcomer pair and its tie bar. The applied load on each downcomer pair will be sinusoidal, with a frequency in the range 4 - 8 H: (the worst loading case shall be taken as the design condition).

For the DBA event combination, the amplitude of the load on each downcomer shall be 1360 lb to occur synchronously on each of the two downcomers.

f The direction of the load en each downcomer is to be within + 22 5 of the plane of the downconer pair, as in the case of the untied downcomers 1700 147

29 and the two downcomers shall be assumed to move away from and toward the torus axis synchronously, although their angular direction within the i 22 5 sector is random. The design etrains on the downeomer and

~

the tie bar are to be taken as the worst load combination, given the 4 - 8 Hz frequency range and the i 22 5 angle of application.

For the IBA case, the specification is the same as that for the DBA case, except the amplitude of the load on each downcomer will be 854 lb and the direction of application shall be random within i 45 f

of the plane of the downcomer pair.

CONIENSATION OSCILLATION VENT SYSTD4 PRESSURE LOAD _S The load definition procedures set forth in Section 4.4.4 of the LDR for the oscillatory pressures in the vent system during the condensation oscillation period, are acceptable subject to confirmation by the additional testing as described above.

1700 148 O

30 CHUGGING TORUS SHELL LOAM

'k The load definition and assessment procedure set forth in Section 4.5 1 of the LD3 for the chug 5 ng condensation loads on the torus shell are acceptable, 1

provided the " rigid wall" load derivation technique is demonstrated to be conservative in response to Question 7 in our request for additional information (D.Eisenhut,NEC,toL.Sobon,GE,datedJuly 30,1979). This load specification shall be used in conjunction with a coupled fluid-structure enalytical model.

CHUGGING DOWCTER LATERAL LOAIE A. Ultied Downcener Lateral Loads The chugging lateral loads on untied downcomers shall be defined as described in Section 4.5 3 of the LDR, which is based on the methodology in NEDE-24537-P, with the following exceptions:

1. The load specification shall be based on the appropriately scaled maximum measured resultant static equivalent load (RSEL) in FSTF, rather than on the 95th percentile RSEL.

2. The multiple downcomer loading shall be based on a nonexceedance probability of 10, rather than 10".

B.

' Tied Downcomer Lateral Loads For tied downcomers, the strains in the downconer itself shall be evaluated exactly as in the case of the untied downcomers. The strain in the tie bar shall be evaluated by assuming that one of the two tied downcomers is subjected to a dynamic load of triangular shape, with an amplitude oft RSEL F,,x = tt t fd 1700 149

1 31 where RSEL is the maximum measured RSEL for an untied downcomer during shugging, T is the lowest natural frequency of vibration of an untied f

downcomer for the specific plant, rad the duration of the load, t, shall d

be assumed to be 3 milliseconds. The load direction shall be taken as that (in the horizontal plane) which results in the worst loading condition for ee tie bar.

CHUGGING VENT SYSTEM PRESSURE IDADS The load definition procedure set forth in Section 4 5.4 of the LDR for the oscillatory pressures on the vent system during the chugging period are acceptable.

DRiFI O

1700 150

s RTI 3AFETY-RELIEF VALVE DISCHARGE DEVICE The acceptance criteria set forth* below for the quencher discharge loads and submerged structure drag load source strengths are applicable only to the "T" quencher configuration described in Section 1.1 of NEDE-24542-P.

For plants using other types of quencher discharge devices, the SRV discharge load definition, submerged structure drag load source strength, and pool temperature limits will be evaluated on a plant-specific basis.

SRV AIR-CLEARING OLTENCHER DISCHARGE SHELL PRESStrRE LOADS A.

Methodology for Bubble Pressure Prediction The load definition procedures described in Section 5.2.1.3 of the LDR and the methodology in NEDE-21878-P for predicting the quencher bubble pressure are acceptable, with the following exceptions:

1. The load definition procedures described in Section 5.2 of the LDR are applicable only to SRV line submergences less than 13.5 feet.

In the event that the submergence exceeds 13.5 feet, additional justification will be required for the proposed load definition procedures.

2. The proposed methodology for predicting bubble press. ts due to SRV subsequent actuations is not acceptable. The pressure amplitude predicted for the SRV first actuation shall be used in conjunction with the bubble frequency range for subsequent actuation, as specified below, for structure, equipment, and piping assessment in response to events containing SRV subsequent IJl actuations.

33 B.

Methodology for Torus Shell Pressure Prediction Based on the predicted air bubble pressure-time histories, as discussed above, the torus shell pressures at various locations in the supprescion pool shall be calculated by the load definition procedures described in Section 5.2.2.3 of the LDR in conjunction with the appropriate pressure attenuation model. For quenchers located on the torus center-line, the pressure attenuation model described in Section 2.4 of NEDE-21878-P in conjunction with the bounding factor presented in Section 3.2 of NEDE-21878-P shall be used.

The load adjustment and attenuation factors proposed for the "off-center" T-quencher configuration presented in a meeting with the staff on May 30, 1979, are acceptable. We will require, however, that this load specification and its bases be documented in a supplement to the LDR.

C.

Multiple - Discharge Loads The torus shell loads due to multiple SRV actuations shall be calculated as follows:

1. The peak values of bubble pressure due to a single valve actuation shall be combined by linear superposition (ABSS method) with the appropriate pressure attenuation model, as discussed above.

All bubbles shall be assumed to oscillate in-phase with the frequency ranges specified below for both first and subsequent actuations.

2. In the event that the combined peak torus shell pressure exceeds 1.65 times the local predicted peak bubble pressure due to a single valve actuation, the resultant torus shell peak pressure for the design assessment may be taken at the lowe value 9

D.

Frequency of Pressure Wave Form The pressure wave form predict'ed by the methodology described in Section 5.2 of the LL.t within tne following uncertainty ranges (streched or compressed time scale) that will produce the maximum structural, equipment, or piping system response shall be used for the design assessment:

1. First Actuation - the frequency range shall be 0.75 timer the minimum predicted frequency to 1.25 times the maximum predicted frequency.

2. Subsequent Actuation - the frequency range shall be 0.60 times the minimum predicted frequency to 1.40 times the maximum predicted frequency.

1700 153

35 SRV DISCHARGE LINE CLEARING TRANSIENT The load definition and assessment procedure, described in Section 5.2.1 of the LDR, for the pressure and thrust loads on the SRV discharge line and quencher, which is based on the methodology presented in NEDE-21864-P und NEDE-23749-P, Addendum 1, is acceptable.

SRV DISCHARGE LINE REFLOOD TRANSIENT The transient analysis technique to compute the plant-specific reflood heights in the SRV discharge line following valve' closure, as described in Section 5.2.3 of the LDR. and based on the methodology in NEDE-23898-P and FEDE-21864-P, is acceptable.

SRV AIR AND WATER CLEARING THRUST LOADS The load definition and assessment procedure for the quencher and quencher support thrust loads, described in Section 5.2.6 of the LDR, is acceptable.

SRV DISCHARGE LINE TEMPERATURE TRANSIENT The transient analysis technique to compute the maximum temperature loads on the discharge line and quencher device, as described in Section 5.2. 7 of the LDR, is acceptable.

RAFT 1700 154

+

36 SRV DISCHARGE EVE!TT CASES The kind and number of SRV discharge events shall be based on the plant-specific system configuration and a conservative. assessment of plant ope. rational history. The following load. cases shall be considered for the design assessments

1. A first actuation, single valve discharge shall be considered for all event combinations involvin6 SRV events. Sing 1,e valve subsequent actuations shall be considered for the SRV, SEA, and IBA event combinations.

2. Asymmetric SRV discharge', both first and subsequent actuations, shall be considered for SRV, SBA, and DA event combinations.

The degne of asymmetric discharge for each event combinatien shall be detemined from a plant-specific primary system analysis designed to maximize the asymmetric condition.

3. ADS valves discharging on first acutation shall be considered for the SEA and DA event ccabinations, followed by subsequent actuations determined from a plant-specific primary system analysis.

4. All valves discharging shall be considered for the SRV event combinations, followed by subsequent actuations determined from a plant-specific primary system analysis.

All of the event combinations above include the earthquake events (OBE and SSE) in combination with the SRV discharge events.

1700 155

SuredION POOL TD4FEdATUFS L3MITS As part of the PUA, each licensee i,s re. quired to either demonstrate that previously subitted pool temperature analyses oi provide plant-specific pool temperature response analyses to assure that SRV discharge transients will not exceed the following pool temperature limits.

A.

Local Pool Temperature Lim".t The euppression pool local temperature shall not exceed 200 F throu6 out h

all plant transients involving SRV operations.

B.

Local and Bulk Pool Temperature The local to bulk pool temperature difference shall be based on either the existing Monticello pool temperature data or plant-specific in-plant tests, and shall consider (1) the quencher configuration, (2) the SRV discharEe locations in the pool, and (3) the RER suction and discharge geo=etry.

The " local" temperature is defined as the temperature in the vicinity ofthequencherdevicedt$ringdischarge. For practical purposes, temperature measurements from locations on the reactor side of the torus, downstream of the quencher end-cap holes, and at the same elevation as the discharge device, may be considered local temperatures.

The

" bulk" temperature, on the other hand, is the temperature calculated assuming a uniform distribution of the mass and energy discharged from the SRV.

1700 156

33 C.

Suppression Pool Temperature Monitor System The suppression pool temperature monitoring system is requiced to ensure that the suppression pool is within*the allowable limits set forth in the plant Technical Specifa*. cation. The system shall meet the following general design requirements:

1. Each licensee shall demonstrate that there is a sufficient number and distribution of pool temperature sensors to provide a reasonable measure of the bulk pool tempe; ature.

2. Sensors shall be installed sufficiently below the minimum water level, as specified in the plant Technical Specifications, to assure that the sensor properly monitors pool temperature.

3. Pool temperature shall be monitored on recorders in the control Two sensors from each sensor group shall be recorded.

room.

The difference between the measurement reading and the actual local pool temperature shall be within 1 2 F.

4. Instrument set points for alarm shall be established, such that the plant will operate within the suppression pool temperature limits discussed above.

5.All sensors shall be designed to seismic Category I, Quality Group B, and energized from onsite emergency power supplies.

1700 157

39 LOCA VATER JET IDADS The load definition and assessment.prooedure descirbed in Section 4 3 7 of the LDR, which is based on the 'Hoody Jet Model" (NEIE-21472-P), is acceptablesubjecttothefollowingconstraintsand/ormodifications:

A.

Theplant-specificjetdischargevelocity,V(t),andacceleration, D

a (t) = dV /dt (t), from the QSTF plant-specific test series shall D

p be used as the driving sources for the jet model.

B.

Forces due to the pool acceleration and velocity induced by the advancing jet front shall be computed for structures that are wit h four downcomer diameters below the downcomer exit elevation, even if the structure is not intercepted by the jet. The flow field shall be computed by modelling the moving jet front as a hemispherical cap centered one downcomer diameter (D) behind the 'Hoody" jet front positions, containing the same amount of water as the " Moody" jet, and moving with the velocity of the 'Hoody" jet front. The fomulas for the hemisphere radius (R ) and the trajectory of the hemisphere 3

center (x } "

• c R(t)=f(f+3(x(t)/D)

)1!3 for x (t)> D s

f f

9x (t) )

f # *f(t) { D f

3 R (t) = y (

3 2D x

" *f

'#*f(

c x (t) = 0 for x (t) $ D f

where x (t) is the position of the " Moody" jet front as a function f

of time, as computed in NEIE-21472-P.

The equivalent unifom velocity and acceleration at the location cf the structure (x,y) shall be obtained from the time dependent potential M (x,y,t) induced by the jet front:

1700 158

E\\FT r

s-i (x - x ) dx (x,y,t)=

c kr/

dt (r /

dt

, -x,f + y and y is the transverse distance of the (x

where r =

structure from the jet axis, and (x-x,) is the distance from the structure to the effective jet front center along the jet axis.

The local unifom flow velocity is U(x,y,b)=V@)

as in EDO 21471, while the acceleration is a(x,y,t) =

[,

This calculation need only be performed for r> R, and x) x. If either of these conditions are not satisfied, the methodology in the LDR will bound the load and is, therefore, acceptable.

LOCA BUB 3LE DRAG LOAIE The load definition and assessment procedures descirbed in Section 4 3.8 of the LDR, which are based on the methodology in EDO 21471 and experimental confirmation in EE-23817-P, are acceptable subject to the following constraints and/or modifications:

A.

Flow Field

1. @TF plant-specific test results (EE-21944-P) will be used.

2. Model E in NEE-21983.-P will be used for the method of images simulation of the torus cross-section.

3, After contact between bubbles of adjacent downcomers, the pool swell flow field above the downcomer exit elev* tion will be derived from the UTF plant-specific te" s.

1700 159

9 4i B.

Drag Load Assessment

1. Dra6 forces can be computed for circular cylinders as given in

= 1.2 must NEDO 21471, but a conservative drag coefficient of C3 be assumed, independent of the Reynolds number.

2. Drag forces on structures with sharp corners (e.g., rectangles and "I'" beams) must be computed by considering forces on an equivalent cylinder of diameter D,q = (2 L

), where Lmu is the maximum transverse dimension.

3. Long slender structures must be considered in segments of length (L), which do not exceed the diameter (D or D,q).

4. For structural segments the centers of which are separated from each other by less than three diameters of the larger structure, interference effects shall either be considered in detail or a bounding load shall be established by multiplying both the acceleration and standard drag loads by a factor of four.

QUEICRER WATER JET LOADS The load definition procedure described in Section 5 2.4 of the LDR, which is based on the methodology in NEE-25090-P. is acceptable, subject to the appropriate documentation of the confirmatory tests discussed in NEE-25090-P.

QUCICHER BUBBLE DRAG LOAM The load definition and assessment procedures described in Section 5 2 5 of the LDR, in NEDE-21878, and in NEDO-21471-2, are acceptable subject to thefollowingconstraintsand/ormodifications:

1700 160

42 A.

Flow Field

1. The detemination of the charging, formatien, and rise of the oscillating bubbles is subject to the same conservative factors that are used for the quencher torus shell pressure loads, as described in NEDE-21878-P.

2. Drag loads on the quencher arms and the SRV discharge line shall be computed on the basis of asymmetric bubble dynadics.. Either a full 180 phase shift shall be considered for full strength bubbles on opposite sides of these structures, or a more detailed assessment of the asymmetry of the bubble source strengths and phasing must be obtained from the experimental infomation in NEDE-21878-P.

3. Mod::1 E.in NEDE-21878-P shall be used for the method of images representation of the torus cross-section.

B.

Drag Load Assessment

1. Drag forces for circular cylinders will be computed on the basis accelerationdragaloneundertheconditionthatU,T/Di2.74, where U,is the maximun velocity, T is the period of bubble oscillation, and D is the cylinder diameter.

For U T/D > 2.74, the standard drag shall be included with the drag coefficient D = 3.6 in order to bound the relevant experimental data.

C

2. The constraints specified for the LOCA bubble drag load assessment also apply to the quencher bubble drag loads, with the exception of the drag coefficient.

1700 161

43 LOCA CONDENSATION OSCILLATION DRAG LOADS The load definition and assessment procedures described in Section 4.4.2 of the LDR and the methodology described in NEDO 25070 are acceptable subject to the following constraints and/or modifications:

A.

Flow Field

1. A maximum source strength shall be established by using the deduced " rigid wall" pressure (p,) at the bottom center of the torus (equation B-4 in NEDO 25070, with f(r) evaluated by equation B-7).

An average source strength shall be established by considering equal source strengths at all eight downcomers in equation B-4 in NEDO 25070.

For each structure, the loads shall be computed for both the flow field produced by the average source applied at all downcomers,and the flow field resulting from the maximum source applied to the two nearest downcomers on one side of the structure alone.

2. The fluid-structure interaction effects shall be included for any structural segment for which the local f2uid acceleration is less than twice the torus boundary acceleration.

This may be accomplished by adding the boundary acceleration to the local fluid acceleration.

B.

Drag Load Assessment

1. The constraints and modifications specified for the quencher bubble drag loads apply.

2. These loads may be applied quasi-statically to structures, only if the highest significant Fourier components occur at frequencies less than half the lowest structural frequency.

1700 162

~

DRAFT LOCA CHUGGING DRAG LOADS The load definition and assessment procedures d,escribed in Section 4.5.2 of the LDR ~ and the methodology in NEDO 25070 are. acceptable subject to the following constraints and/or modifications:

A.

Flow Field

1. The maximum source strength history shall be obtained by using the maximum measured pressure (not necessarily at the bottom center) in a Type 1 chug in equation B-4 of NEDO 25070, with f(r) based on the single nearest downcomer. For each structure, the phasing between the two nearest downcomers that maximizes the local acceleration shall be established. The local acceleration shall then be computed on the basis of the two nearest downcomers chugging at maximum source strengths at the above established phase relation.

2. The fluid-structure interaction effects shall be included for any structural segment for which the local fluid acceleration is less i.han twice the torus boundary acceleration. This may be accomplished by adding the boundary acceleration to the local fluid acceleration.

B.

Drag Load Assessment

1. The constraints and modifications specified for the quencher bubble drag lords apply.

2. Unless the lowest structural natural frequency times the duration of the " spike" in the source strength is greater than 3, the loads shall be applied dynamically. Either sufficient Fourier components will be included to bound the " spikes" or the load shall be applied in the time domain using the source time history.

1700 163

SECONDARY IDAM The following loading conditions may be ne6 ected for the PUA:

1

1. seismic slosh pressure loads

2. post-swell wave loads

3. asymmetric pool swell pressure loads

4. sonic and compression wave loads

5. downcomer air clearing loads SONIC AND COMPRESSION VAVES Immediately following the postulated instantaneous rup+-ure of a large primary system pipe, a sonic wave front is created at. the break location and will propagate through the drywell and into the vent system. The subsequent compression of air in the drywell and vent system will cause a compression wave to be generated in the water leg inside the downconer.

This compression wave propagates throu6h the pool and results in a differential pressure loading on sutnerged structures and the torus wall.

These loading conditions were observed in the FSTF data. The desi6n of FSTF was such that a conservative estimate of the loading condition could be established, because the shulated drywell volume did not allow si nificant attenuation of the wave front. The maximum observed loads 5

were approximately 20 psid in the drywell and 10 psid in the torus, with a maximum duration of less than 5 milliseconds.

This loading condition preceeds all other dynamic loads, and is insignificant in comparison to the other dynamic loads. In addition, a more realistic attenuation of the wave front, based on the actual configuration of the drywell would result in even lower loads than those observed in FSTF.

1700 164

46 On this basis, we conclude that neglecting the sonic and compression wave loads is acceptable.

SEISMIC SIDSH Seismic motion induces suppression pool waves which can (1) impart an oscillatory pressure loading on the torus shell, and (2) potentially lead to uncovery of the downcomers,"which would result in steam bypass of the suppression pool and potential overpressurization of the torus, should the seismic event occur in conjunction with a LOCA. To assess these effects, the Mark I Owners Group undertook the development of an analytical model which would provids plant-specific seismic wave amplitudes and torus srall pressures. Thismodelwasbasedon1/30-scale"shaketest" data for a Mark I torcs geometry.

Based on the results of the plant-specific analyses, the Mark I Owners Group concluded that (1) the seismic wave pressure loads on any Mark I torus are insignificant in comparison to the other suppression pool dynamic loads, and (2) the seismic wave amplitudes will not lead to downcomer uncovery for any Mark I plant. This conclusion was based on the maximum calculated pressure loads and the minimum wave trough depth relative to the downcomer exit.

We have reviewed comparisons of the analytical predictions with scaled-up test data, the small-scale test program, and the seismic spectrum envelope used in the plant-specific analyses. Based on this review, we conclude that the seismic slosh analytical predictions will provide reasonably conservative estimates of both the wall pressure loading and the wave amplitude, for the range of Mark I plant conditions.

1700 165

~

47 Since the maximum local wall p:sssures were found to be less than 0.8 psi at n 99C upper confidence limit, the Mark I Owners Group has proposed that the seismic slosh loads may be neglected in the structural analysis. We a6ree that the seismic slosh loads are insi6nificant by comparison to the other suppression pool dynamic loads. On this basis, we conclude that neglecting seismic slosh loads in the pUA is acceptable.

The zesults of the slosh wave amplitude predictions indicate that, within the local area of maximum amplitude and with maximum suppression pool drawdown (resulting from ECCS system flows), the slosh waves will not cause downcomer uncovery. We have reviewed the assumptions used in these analyses and conclude that they are sufficiently conservative to provide assurance that seismic slosh will not result in downcomer uncovery.

POST-POOL SVELL WAVES Following the initial pool swell transient, pool wave action will result from continued flow through the vent system. This wave action will result in pressure loads on the torus walls.

For the period immediately following the downward and upward vertical pressu:e transient, the Mark I Owners Group has concluded that this wave action is inherently included in the QSTF pressures and are negligible.

Although the scaling relationships by which the QSTF was designed are not applicable following bubble breakthrough, we agree that the QSTP results provide a reasonable estimate of the wave loads during this period.

We, therefore, conclude that the post-pool swell wave loads may be neglected in the PUA.

1700 166

48 Durin6 the subsequent condensation period, the pool wave action is inherently included in the condensation oscillation and chugging load specifications.

These loads were derived from wall. pressure measurements from full-scale steam cc.adensation tests, and, therefore, a separate load specification for condensation wave loads is unnecessary.

ASYMMETRIC VE?rr SYSTEM FLOW The effects of asymmetric flow rates in the vent system have been considered with respect to unequal vent flows (e.g., vent blockage) and unequal vent flow composition, to evalus,te the potential for asymmetric pool swell loading conditions.

The three-d hensional pool swell tests conducted by EFRI for the Mark I Owners Group, and the cenfirmatory three-dimensional pool swell tests conducted for the NRC by the Lawrence Livermore Laboratory, included specific tests to assess the effects of partial and full blockage of one main vent. The results of these tests indicate that the distribution of the pool swell pressure loads are relatively insensitive to the main vent blockage, because the vent header tends to equalize in pressure and, therefore, equalize flow through the downcomers. Due to the configuration of the Mark I vent system, the main vent entrance is the principal location where flow blockage could occur, if at all, and, therefore, flow blocka6e assumptions have not been considered for other locations in the vent system.

~

To assess the effects of potential asymmetric vent flow composition, we have considered the extreme case of localized steam flow through the vent system, without benefit of any steam-air mixing. For the DBA, the earliest 1700 167

U time that a steam front could reach the downcor'r exit is shortly before the peak upwarc load. However, an additional time delay would occur befoze the steam could reach the bubble-pool interface and affect the local pressuze loads, due to the circulation of t'he steam in the existing air bubble. In addition, as tre steam condenses reducing the bubble growth, the air conpression, which is the major contributor to the upward loading phace, would tend to equalize. This would result in a reduced potential for an asymmetri.c loading condition and lessen the severity of the pool swell loads.

We have also considered this extreme case for smaller breaks to assess the potential for localized peol heating, leading to overpressurization of the wetwell. Based on the results of FSTF, we conclude that the increased vent flow rates secompanying higher energy deposition in the pool will provide sufficient mixing to prevent overpressurization of the wetwell.

Basid on the results of this assessment, and the extreme nature of the assessment, we conclude that neglecting asyneetric pool swell conditions for the PUA load definition is acceptable.

D0',TNCOMER SUEMERCE';CE A';D THE?f!AL STRATIFICATION One method of suppression poc1 hydrodynanic load mitigation that the Mark I Owners Group has adopted for the LTP is reducing the initial sulnergence of the downcomer in the suppression pool to a minimum of three feet. The pool volume (i.e., thermal capacity) of the original desi n would be 6

maintained. This approach, however, raises concerns regariing the increased potential for downconer uncovery and stean condensatien capability, both of which could lead to wetwell overpressurization.

1700 168

/

The potential for downcomer uncovery was addressed by the preceeding seismic slosh assessment. This assessment was perfom ed at the most extreme conditions that could potentially lead to downcomer uncovery and was predicated on a minimum three foot downcomer subergence.

Condensation capability of the suppression pool, for the spectn:n of IDCAs, is a function of the local pool temperature in the vicinity of the downcomer exit. FSTF test results and foreign test data have shown that themal stratification occurs, and becomes more severe as the downcomer su bergence is reduced. The most severe themal stratification has been observed in low flow tests with a quiescent pool.

In actual plant condi.tiens, the Residual Heat Removal (RHR) system and SRV discharge will provide sufficient long-tem pool mixing to minimize thermal stratification. As previously discussed, for asymmetric vent system flows, we have determined that, for the short-tem, the increased vent system flow rates with higher energy deposition will prevent overpressurization.

This assessment included consideration for vertical thermal stratification as well. In addition, the analytical predictions of the wetwell pressure and bul?.emperature response have been found to be conservative by comparison to FSTF test data for plant-simulated initial conditions.

The local temperature variation in the pool which has been observed in the test data are not significant to the structure, and, therefore, need not,

be considered in the structural analysis.

Based on this assessment, we conclude that a minimum in' 4 %i downcomer sulnergence of three feet is acceptable, and there ha a ficient conservatism in the containment response analysis techniques,. ace.,d ate the effects of themal stratification.

1700 169

5l DOWNCOMER AIR CLEARING LA7 TRAL LOAIS During the initial pahse of a IDCA, the. rapid clearing of air from the vent system causes the downcomer to be subjected" to a lateral load as bubbles are being formed in the pool, in addition to the thrust loads on the vent system previously discussed. Conservative estimates of the The Mod I air clearing lateral loads were obtained from the FSTF data.

Owners Grup has proposed to neglect the air cleardng lateral load because it is bounded by the repetitive steam condensation lateral loads on the We concur with this assessment and, therefore, conclude that downcomer.

neglecting the air clearing lateral load on the downcomers is acceptable.

6 0

1700 170

DIFFERENTIAL PRESSURE CONTROL REQUIREMENTS Those licensees that use differential pressure control (AP) as a pool swell Joad mitigation feature for the LTP, shall demonstrate conformance with the following design criteria as part of the PUA:

1. There shall be no unacceptable change in the radiological consequences of an accident as a result of the inclusion of the AP system.

2. Steam bypass of the suppression pool via the AP system shall be eliminated by appropriate system design, or such bypass shall be demonstrated to be acceptable by calculation.

3. Design and installation of the AP system shall be commensurate with other operational systems in the plant.

4. When the AP system involves the addition of containment isolation valves, the additional valves shall be included in the plant's rechnical Specifications and the valve design and arrangement shall conform to the requirements of General Design Criterion 56 in Appendix A to 10 CFR 50 and the regulatory positions in Standard Review Plan Section 6.2.4.

Subsequent to the PUA, a license amendment shall be submitted to incorporate the following Technical Specification requirements for the AP system:

1. Differential pressure between the drywell and suppression chamber shall be maintained equal to or greater than "X" (where X is the plant-specific differential pre ure and values less than one psid will not be credited for load mitigation), except as specified in 2 and 3 below.

2. The differential pressure shall be established within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after placing the plant in the RUN mode, during plant startup. The differential 1700 171

DRTT pressure may be reduced below "X" psid 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> prior to a scheduled plant shutdown.

3. The differential pressure may be reduced to less than "X" psid for a maximum of four hours during required operability testing of (specify here those safety-related systems for which operability tests either release significant amounts of energy to the suppression pool or cannot be performed with the AP established).

4. In the event that the specification in 1 above cannot be met, and the differential pressure cannot be restored within six hours, an orderly shutdown shall be initiated and the reactor shall be in a cold shutdown condition within the subsequent 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.

5. A minimum of two narrow range instrument channels shall be provided to monitor the differential pressure. Error in the A P measure-ment shall be no greater than 1 0.1 psid. The instrument channels shall be calibrated once every six months. In the event that the measurement is reduced to one indication, operation is permissible for the following seven days. If all indication of the differential pressure is lost, and cannot be restored in six hours, an orderly shutdown shall be initiated and the reactor shall be in a cold shutdown condition within the subsequent 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.

1700 172

5$

I STRUCTURAL ANALYSES AND ACCEPTANCE C

t. <

The staff finds the general analysis techniques and proposed structural acceptance criteria set forth in the " Mark I Containment Program Structural Acceptance Criteria Plant Unique Analysis Applications Guide," (PUAAG),

NEDO 24583, Revision 1, dated July 1979, acceptable. The proposed criteria will provide a sufficient basis for demonstrating the margins of safety required for steel structures and piping in the ASME Boiler and Pressure Vessel Code and for concrete structures in the American Concrete Institute Code.

Revision 1 to the PUAAG was presented to the staff in a meeting on June 29, 1979. We will require that this revision be fomally submitted to complete the documentation required for this program.

ALLOWABLE STRESS LIMITS The structural acceptance criteria set forth in the PUAAG which will be used to evaluate the acceptability of existing Mark I containment systems or to provide the basis for any plant modifications to withstand suppression pool hydrodynamic loading conditions are generally contained in Section III of the ASME Boiler and Pressure Vessel Code through the Summer 1977 Addenda.

The application of these stress limits to the Mark I design will provide 7dequate margins of safety to insure the containment structural integrity for all anticipated loading combinations and to insure that the containment and attached piping systems will perform their intended functions during those loading conditions expected to occur as a result of a LOCA or SRV discharge and are, therefore, acceptable.

1700 173

Rift Additionally, the ratio of the dynamic collapse load to the static collapse load was established for the torus shell during LOCA pool swell pressure loads and for the vent header dur ng pool swell, impact loads. These values, in conjunction with Code Case N-197, were used to establish the allowable stress values for the torus shell and the vent header local strasses. The staff and their consultants have reviewed the analyses used to establish these factors and find them acceptable.

PIPING FUNCTIONALITY Recent studies by Battelle have demonstrated that theoretical collapse and piping wall buckling limits can potentially be exceeded for certain components in ASME Class 2 and 3 piping systems if Level C and D stress limits are used. Since the piping of concern in the Mark I systems is carbon steel and does not fall in the classification of diameter to thickness ratios that are greater than 50, the area of concern with respect to exceeding the theoretical collapse moments is piping elbows. For A-106 Grade 8 piping the Level D stress limit, 2.4 S, is 36 ksi or approximately yield for the h

materi al. Since the study has shown that the theoretical collapse load could be exceeded in elbows when the stress calculated by Code Class 2 and 3 rules reaches approximately 3/4 of the material yield stress, the theoretical collapse load could be exceeded by approximately 1/3. However, this theoretical limit and the supporting test data is based on a deflection limit of a few times the yield point, which would not significantly alter the piping flow area.

In addition, when a segment of a continuous piping system begins to yield a load redistribution will occur to other areas that would 1.imit the deflection.

1700 174

P hr he dynamic loadings under Therefore, the staff concludes consideration in the Mark I reassessment, the linear elastic analysis of piping systems using ASME Code Class 2 and 3 Limits provides reasonable assurance that sufficient margins exist to preclude excessive piping deflections which could impair flow in safety related piping systems and, therefore, consider these limits accpetable.

DAMPING The damping values used in the analysis of dynamic loading events will be those specified in Regulatory Guide 1.61.

Since these values are specified for seismic analysis of structures and components for OBE and SSE conditions, the values used will be consistent with the stresses expected under similar loading conditions. The staff considers the use of the Regulatory Guide 1.61 damping values acceptable for the Mark I dynamic analyses.

OPERABILITY OF ACTIVE COMPONENTS Active components, as defined in Section 2.2.9 of the PUAAG, shall be considered to be operable if Service Limits A or 8 are met, unless the original design criteria establishes more conservative limits.

If the original component design criteria establish more conservative limits, conformance with these more conservative limits shall be demonstrated even if Service Limits A or B are met.

If the original component design criteria are silent with respect to operability limits, satisfaction of Level A or B Service Limits shall be considered as sufficient to demonstrate operability.

1700 175

57 Active components which do not satis ry Service Limits A or B, and therefore either Service Limits C or D are satisfied, require demonstration of operability. If the original component design criteria for operability exist, confomance with those criteria shall be' demonstrated.

If the orignial component design criteria are silent with respect to operability limits, operability limits shall be established and confomance with those criteria shall be demonstrated.

The operability requirements are necessary to assure the active safety-related components will be able to perform their intended functions.

It is the staff's position that loads calculated by elastic analysis which produce stresses in excess df the material yield stress can produce excessive deformation in a component which can cause interference 6f mechanical motion.

We recognize that the designation of Service Limits A and B do not, by themselves, guarantee the operability of active components. However, the scope of the Mark 1 Conatinment Long Term Program is directed toward the effects of the incremental load increase due to the definition of suppression pool hydrodynamic loads, and the restoration of the original intended design safety margins. The criteria for operability specify that the origir.al component design criteria must be met where they are more conservative than the Service Limits A and B.

We believe that these criteria are reasonable and practical and are sufficient to accomplish the objectives of the program.

1700 176

.e COMBINATION OF STRUC?'IRAL RESPONSES The structural responses resulting from two dynamic phenomena will be combined by the absolute sum method. Time phasing of the two responses will be such that the combined state of the stress results in the maximum stress intensity. However, as an alternate, the Cumulative Distribution Function (CDF) method may be used if the absolute sum does not satisfy the structural acceptance criteria. The C0F abcissa value corresponding to an ordinate value of 84% (i.e., the combined stress intensity value corresponding to an 84% probability of nonexceedance) will be used to compute a reduction factor which will be applied to the stress intensity computed by the absolute sum method. An 84% probability of nonexceedance corresponds to a mean plus one standard deviation for two dynamic responses.

The CDF method is more conservative than Criterion 2 of the Newmark-Kennedy Criteria proposed for use in the Mark II Containment Program.

The rationale for the use of this methodology is similar to that contained in NUREG-0484 and is, therefore, acceptable.

1700 177

, ENCLOSURE 2 s

'o UNITED STATES NUCLEAR REGULATORY COMMISSION n

WASHINGTON, D. C. 20555

-E g

k.....,/

August 2, 1979 General Electric Company ATTN: Mr. L. J. Sobon, Manager BWR Containment Licensing, MC 905 175 Curtner Avenue San Jose, California 95125

Dear Mr. Sobon:

Enclosed are fifteen copies of the first draft of the staff's acceptance criteria for implementation of the Mark I Containment Long Term Program.

In our meeting with the Mark I Owners Group on July 24, 1979, we indicated that copies of the draft acceptance criteria would be provided to permit the Mark I Owners Technical Review Committee an early opportunity to assess As program coordinator, please distribute the impact of the staff's positions.

the enclosed copies to the appropriate Mark I Owners Group representatives.

The draft criteria are being provided to expedite implementation of the Long Term Program. These positions have not been reviewed, or concurred Those in, by staff management and are, therefore, subject to change.

sections identified by a vertical bar in the right margin are parts of the staff's evaluation on which certain criteria are based and would not appear in the final acceptance criteria, but rather, would be in the staff's Safety Evaluation Report.

Following our internal review and comment, we will meet with the Mark I Owners Group representatives to discuss these criteria and a schedule for program closure. As discussed with L. Steinert of your staff, this meeting is presently scheduled for August 15, 1979. The objective of this meeting will be to identify those areas, if any, where the Owners Group significantly disagrees with the proposed criteria, and to develop any infortnation necessary for NRC managenent to decide whether a change to the criteria is warranted.

We intend to issue the final criteria shortly after this meeting to assure that the proposed plant modification schedules will not be jeopardized and to avoid the need to consider extending he exemptions that have been granted to the operating Mark I facilities.

J f

hl

~

D.G.Eisenhut,ActingDirdctor Division of Operating Reactors 1700 178 Ag b3,b 79cnocm

~~

.