ML20038A665
| ML20038A665 | |
| Person / Time | |
|---|---|
| Site: | 05000447, 05000550 |
| Issue date: | 11/09/1981 |
| From: | Pfefferlen H GENERAL ELECTRIC CO. |
| To: | Rubenstein L Office of Nuclear Reactor Regulation |
| References | |
| HCP-027-81, HCP-27-81, MFN-199-81, NUDOCS 8111160135 | |
| Download: ML20038A665 (186) | |
Text
{{#Wiki_filter:, GENER AL @ ELECTRIC NUCLEAR POWER SYSTEMS DIVISION HCP-027-81 GENERAL ELECTRIC COMPANY,175 CURTNER AVE., SAN JOSE, CALIFORNIA 95125 MFN-199-81 November 9, 1981 0h [P U. S. Nuclear Regulatory Commission Division of Systems Integration /~ Ngp2 (h Office of Nuclear Reactor Regulation (- 8j q @\\,- u., 7Q.. pol
- 1e Washington, D. C.
20555 '% n., Attention: L. S. Rubenstein, Assistant Director , /7Mllp,,s 6 N For Core and Containment Systems Gentlemen:
SUBJECT:
RESPONSE TO STAFF POSITIONS ON APPENDIX 38 0F GESSAR-11
Reference:
Letter, L. S. Rubenstein to H. C. Pfefferlen, September 19, 1981. Attached are final responses to the Staff positions in the referenced letter. As per a Staff request, copies have also been sent to Ain Sonin and George Maise. l The schedule for submittal of these responses differs from that in the referenced letter but was agreed to in discussions with Mel Fields. These responses will serve as the basis for a meeting with the Staff currently planned on November 20, 1981. Formal submittal on GESSAR is currently planned for November 30, 1981. If you have any questions on this submittal or the planned meeting please give me a call on (408) 925-3392. Very truly yours, //f f CoS 5 H. C. P e erlen, Manager BWR Licensing Programs [ N'iclear Safety and Licensing Operatons /// 6mi[6/ D.5f t cc: M. B. Fields (NRC) J. A. Kudrick (NRC) L. S. Cifford 8111160135 811109 PDR ADOCK 05000447 A PDR s
r 10.0,"*..O.3k Que!$/0! hvo6M 50 N
- q. r.e.. c : 9,,.L, '2 y
We have reviewed a'.! ef the availab:e tes: data and have exa-ined a:1 ef i i the ana;ytica. jus:ificatien previded by Geners! Electric (CE) cencerning 1 the pe:1 sve;; vele: - specificati:n cf A0 ft/se: Our cenclusien is that 4; s: se: pre; veleci:s is in a d e q ua t e '.y supper:ed b:. the available inferra-t:en. Our reasening is as fellevs: t
- ..: :: ;t
- es. cc:2 art n: rirrest : :. : :vt cf ':a r '. ::: crnc :. nc
~- t....:.., . a. t
- <u,.
.c. s.- w.. , L., c. ....c. . c a. . =... sa. ......e.e ver:s rather ths-three: GE's earl:er cer.parisen of tve-vent versus sin-g: t. - v e - : b; rdr.ns s ;:est: ! !.c : thir res: 1:ke: reduces th< perl s'.c:: u. 2..e ,e6 ( w. $...s-s. t d.... ...c.. . a a
- c... c. t.
- 5., e. 2
- c..
c....;. yaic ...t. e \\. t....... .c...... g2.t rist : a n : J." * - i r '. 5 *. ". t C r ;.".' t F r e S '.u r t h i ' : C r '. In the la (r stage? rf
- h p-s k t '... wh er( the 7 t-a'e ve10:itie5 arO Teached. The 3a
.. e t t >,.
- a. s t s. : e,. s. r, e.
n U, s., .t.., .u........t. .s 4 o c.. ".c tes: res..:5 fr - :..c . '3-s:s:e" an: ". i-s: ale" sys:e-: :ndi c t tha: : hc per; su ll ve;::::
- n a :u.;-scale syster will be greater : :. c.r 40 f:/sec.
This fcilevs frcr the tes getretries and the test results. b::5 the "1/9-scale" and "1/3-s:41e" systems had distcr:ed gecretries, wi:5 full-scale fkv-wise dinensi ns tu: with cross-set:icns scaled devr i l by the indica:ed scaling ra:ic. One expe::s that the r.:re the crcss-sections are scaled down, the thicker will be the slug above the buttle, and the slewer the poc1 swell velocity. This is indeed indicated b-tne 1 data. At a Mark III representative drywell pressure, the "l/4-scale' syster indi:ated a velocity of ab:ut !.0 f:/sec. There is reas:n :: r I I I i 360.3 a 32-/ \\ /
r 1 expee: a fur:her increase as one goes free the "1/3-scale" te full-i scale syster. H:vever, the ateunt Of ir. crease cannet be detere.ined in a satisfat: cry fashien. t t 1 Based en the abert discussier. :N staff and :s censu!: ants feel tha:
- is h
ver;. like:y :ha: the p::1 sve'1 velocity is abeve !.0 ft/see end, based en l I cur jud; ent, a;n:s: c e r t a : r '. be :V E f:/sec. i l Therefere, :: s c.: ;::-:: r. that 6 0 f:/se; be used as a desipr pre! swell r l ve eci:s fer Mari !!! ::ntainment, unless fur:her jus:ificatien can be pre-vitee f er :esier va:;es :f re-: sve : ve:e:ities. 3, l i i 2 i I i I. l t i-P l 1 P-l i. I N 3(3o.3.2.32-2. f ~rn-,,,---,ww---,v,,,--.
. _ _ _ -. - = _ _ - _ - _ _ _ -. _ = _ - - -. - - ~ + I RESPONSE 3B.32
1.0 INTRODUCTION
i l As stated in Response 3B.3 and in the August 4, 1981 GE-NRC meetings (Ref 3B.32-1), 40 ft/sec is an adequate peak pool swell velocity for Mark III design conditions. The adequacy of this velocity was confirmed by four independent approaches. First, a 100% DBA break area full scale \\ l PSTF air blowdown test showed the maximum pool swell velocity is 38 f t/sec. Section 1.1 discusses this full scale test result and addresses the concerns raised by the NRC in their question. There is also a large 1/3 area scale 1 PSTF data base which shows the maximum pool swell velocity in one-third scale, with a 100% DBA break area, is 133 ft/sec. Section 1.2 discusses this data. Further analyses have also been performed which provide addi-l tional support for the design pool swell velocity. Section 1.3 shows that a conservative Modified Froude scaling interpretation of PSTF 1/3 area scale data produces a design velocity of 44 f t/sec. Section 1.4 will show that a GE 2D analytical Marker and Cell (MAC) pool swell model predicts the velocity of the 238 Standard Plant to be s39 ft/sec. In summary, four completely separate approaches all support a maximum design I pool swell velocity of approximately 40 f t/sec. } I 1 i i 3E0.3.2.32-3 1
__ - - _ _ ~ ~ 1.1 FULL SCALE PSTF ESTIMATE OF MARK III POOL SWELL VELOCITY. Response 3B.3 shows that the maximum extrapolated velocity in a full scale PSTF 100% DBA break area air blowdown test (Test 5705-4) was 38 ft/sec. Although the full scale PSTF conditions are not identical to Mark III condi-tions, the results may be considered to be representative of a Mark III. The full scale PSTF used two vents rather than three. The effect of the l third vent would have been to cause the top vent to clear earlier and decrease the peak drywell pressure. Since breakthrough would be expected to occur before the third vent opened, the bubble would not be charged through the third vent. Thus, the effect of not having 3 vents is to cause a higher drywell pressure history. Subscale data show that pool swell velocity increases with drywell pressure. This was seen by looking at the variation in velocity with break size in 1/3 scale. The one and two vent full scale data show the 1 vent tests take longer to clear l the top vent and thus have larger drywell pressure histories. These results indicate that.the effect of the single vent is to blow the bubbic further l across the pool, lif ting a larger slug of water and causing a lower final I velocity. Lack of instrumentation on the far side of the pool where the maximum single vent velocity occured prevents a quantification of this velocity l d ec rea s e. l Y j Consideration of the balancing ef f ects of a higher d*1ving pressure with a l larger slug being lifted, and realization that breakthrough will occur before the third vent would clear, lead to the conclusion that having only two vents causes only a second order ef f ect on velocity. l 3B0.3.2.32 4
l ~ The full scale PSTF also had an undersized drywell. The undersized drywell caused the drywell pressure at top vent clearing to be greater than the corresponding 238 Standard Plant value, however, it falls off faster than prototypically. This effect is accounted for by the fact that the blowdown medium was air and not steam. The response to question 3B.3 showed that if a steam blowdown had been performed, the drywell pressure history would have bounded the 238 Standard Plant drywell pressure history while the resultant peak pool surface velocity would have been the same. Thus, the 38 ft/sec obtained f rom this full scale test is representative of a Mark III design velocity. 1.2 VELOCITY CALCULATION USING 1/3 AREA SCALING. l l The smaller scale PSTF tests show lower pool swell velocities than is observed l in the larger scale tests. At a representative Mark III drywell pressure (100% DBA break area) the 1/3 area scale data indicate a velocity of about 33 ft/sec. On this basis a design velocity of 40 f t/sec seems to be a reasonable estimate. Figure 3B.32-3 shows that in order for the 1/3 area scale steam blowdown data to exceed 40 f t/sec, drrwell pressures exceeding design pressure are needed. The response to question 3B.3 shows that the design pressure is very conserva-tive by itself. Thus, the subscale data support a pool swell velocity in the neighborhood of 40 ft/sec. 1.3 MODIFIED FROUDE SCALING ESTIMATE OF MARK III POOL SWELL VELOCITY. l l All of the 1/3 area scale PSTF data may be conservatively scaled to full scale using the Modified Froude scaling relations. Attachment A develops the i Modified Froude scaling relations and Table 3B.32-1 summarizes the results. t 1 3B0.3.2.32-5 l
~ Attachment B provides an evaluation of Modified Froude scaling. This attachment shows that the basic scaling relations may be used to show the Modified Scaling approach is approximately 15% conservative relative to exact Froude scaling. Attachment C shows that Modified Froude scaling has about 7.5% conservatism when using the Marker and Cell Pool Swell Code to extrapolate to full scale conditions. Additionally, this attachment shows the out of scale vent length and spacing have a conservative ef f ect on pool swell velocity and were accounted for by the way in which the data are presented in Figure 3B.32-1 To obtain a measure of the expected Mark III velocity using Modified Froude scaling, the scaling relations summarized in Table 3B.32-1 will be used along with a correction factor on velocity, K, which accounts for the conservatism of the scaling. For conservatism, the lowest correction factor (7.5%) will be used for this analysis. All of the 5 and 6 foot submergence, 1/3 area scale, 3 vent steam data will be used for this analysis. This is the most applicable data because the scaled Mark llI submergence is 4.33 ft. The scaling equations used are: V / K! V V* = test ref V = measured test velocity = scale factor = 1/,5I K = f actor of conservatism in Modified Froude scaling = 1.075 s V* = Ion-dimensional velocity V = ref rence vclocity = 10 f t/sec cf ~ fr test P = maximum measured drywell pressure (psig) st P = M dified Froude scaled maximum drywell pressure (psig) fr 3 6k]. 3. 2.3a-6
~ To make the dif ference submergence curves directly comparable, the maximum dif f erence between the drywell and wetwell pressure was non-dimensionalized with the static head at the top vent centerline (P - is the controlling dw ww parameter for pcol swell). That is: Y* " I I fr pS h Ec where P* = non-dimensional pressure P = Modified Froude scaled maximum drywell pressure g p = density g = gravity h submergence a g = gravitational censtant These data were numerically fit with a least squares second order curve with V* = 0 at P* = 1, and the resulting curve is presented in Figure 3B.32-1. P* = 1 was used since this is the pressure needed to clear the top vents. To be sure that this second order least squares curve is the most appropriate way to fit the data, a correlation of V* versus P* was obtained using the full scale 5702 data. This data base was used tecause it had a large variation in ?*. Figure 3B.32-2 shows that the secorad order curve through P* = 1 provides a correlation coef ficient of 0.96 and is indeed a valid method to represent the data. To obtain a measure of the Mark III velocity f rom Figure 3B.32-1, the maxiium PdP during the pool swell f rom the GESSAR pressure history ~ vs 3 6o. 3. L 32-7 e c r
(19.8 psid) was used to obtain P* = 10.6. The velocity obtained in this ~ manner (44 ft/sec) is conservatively high due to the many overly conservative inputs used in the analytical model. The best estimate peak drywell pressure would be about 15% lower than the design value which would correspond to an even lower velocity (Ref 3B.32-2). Thus, Modified Froude scaling shows the peak poo? swell velocity with the concervative design drywell pressure history is 44 f t/sec. and this result includes several conservatise inputs. m J 1 = 1 3 6 o.3.7. 32-S --v-? m -4 .e- ---e- ,9
~ 1.4,USE OF 2D MARKER AND CELL (MAC) ANALYTICAL POOL SWELL CODE TO DETERMINE EXPECTED MARK III POOL SWELL VELOCITY. The Mark III pool swell velocity was aise estimated using the 2D Marker and Cell (11AC) pool swell analytical model described in Attachment D. Using the GESSAR pressure history and 238 Standard Plant beometry and conditions, the maximum pool swell velocity 13 ft above the initial pool surface was calculated to be 47.1 ft/sec. This velocitj was obtained by assuming no condensation occurred. When condensation is considered, the surface velocity decreases to m39.1 ft/sec. As was discussed in the August 4, 1981 GE-NRC meetings that due to the large pool swell height in a Mark III, the effect of condensatior. is more pronounced in Fbrk III than in the !brk I or Mark II, thus, it is highly conservative to neglect the effect of..ondensa tion. Figare 3B.32-3 shows that at near proto-typical Mark III pressures (16 to 20 psig), the ratio of the pool swell velo-cities due to air and steam blowdowns ic between 0.72 and 0.74*. To be conserva-tive, it will be assumed that Vpool-s: cam /Vpool-air = 0.75 in 1/3 scale. To determine the ef f ect of condensation in full scale, an analysis was performed which assumed the relative ef f ect of condensation may be found by comparing the steam mass condensed with the vent mass flow rate.
- Note that in Figure 3B.32-3, the steam point at x = 19.8 psig, y = 39 ft/sec was changed from its previously reported velocity of 44 ft/sec (Test 5801-12).
This' was due to an overly conservative interpretation of the data. In this test the level probe string 5 ft from the drywell wall had 9 probes which yielded an extrapolated maximum velocity of 39 ft/sec. The probe string 3 ft from the drywell wall had only 3 probes which wet at the identical times that the corresponding 5 f t probes wet. This showed that the pool surf ace was flat with the same velocity at all locations. In the data interpretation, howeve r, a velocity of 44 f t/ set was inferred from the 3 ft string of probes. The correct interpretation of the level probe data is 39 ft/sec throughout the pool. _________ ___ 3 Bo _ 3 _2 3 2 - 9
_. _. _. _ = _ _ _ _.. _ _ 1 M hA t/h a b E Condensation Effect = -Er- = M Ap V v v v l where h = heat transfer coefficient latent heat of vaporization 4 h = g A bubble surface trea b temperature difference between bubble and surrounding fluid At = density o = A = vent area v V = vent velocity = condensation rate c M = vent mass flow rate v full scale fs = 1/3 1/3 area scale = 1 j Therefore, assuming h, at h and p are the same for 1/3 and full scale: g M c ( "v fs Abfs/(^vfs vfs) 0.69 = Scale ef f ect on condensation a = A / (A V ) b c_ l/3 "1/3 "1/3 i M 1/3 l l l This shows that 69% of the 1/3 scale condensation effect is expected in full scale. Using this iniormation along with the knowledge that Vpool-steam /Vpool-t air = 0.75 in the 1/3 scale f acility shows that Vpool-steam /Vpool-air = 0.83 in l full scale. Since the full scale model velocity without condensation was 47.1 f t/sec, this implies a full scale velocity with condensation of 39.1 f t/sec. This velocity was obtained' using the GESSAP. drywell pressure history. l l 3 BO. 3. 2. 32.- /o - ~
In summary, the numerical simulation explained in Attachment D also supports a design pool swell velocity of 40 ft/sec. l 3B0.3.2.32- //
36,32 o.. r e. o a grr. r: r..s-e.e, i M:Nazara, E.J. "Fa:: Suc11 Presenta:icn" General Ele :ric- ..c,e,. p e... . C.... ' e..e '. -. ".. e c + '. e, A"- e- ". s ' *', 1 0. '. e..
- v........c, a.<.2 4 _.....c.i u..e c e pre care
.c. __.ess4,_ Ce-..*.a'...c... e*....'.. c.'..-.c.'", I n *. e.-..a *. d - e... '. Spe l'atis*.c "'d v.ectin: en Pressure Su;;re ssien, G er.anv, June, 1951 3 g ~3 2.3.2 ' / A
~ Table 3& 3,2-/ SCALING RELATIONS FOR LOCA BUBBLE GROWTH A'iD POOL SWELL MODEL = PROTOTYTE X SCALE FACTOR Independently Varied Geor etric Para eters ' Scale Factor 1. Vent Diameter, d 1 d Ce:endent Test Para +te-s 1. Top Ver.t Subrergence, H A 2. Drywell Volume, +d Ild) 3. Dryseli Air Pse's, r: (A I ,5 d 4. Drywell Initial Pressure, p ; 1 a. 5. Stea. Mass Flow Rate,I 5 (i ) 3 d 6. Vent Lengtn, t A d 7. Ve.! Vertical Spacings, H, H, etc. A p 3 d 8. Po:1 Wictn and Breadt; A d 9. Pcol Vclu e (Ad) 10. Containment Airspace Volu e, t (Ad) ci 11. Containter.: Initial Pressure, P I ci Response Parareters Bubble Volure, t ( d} b 2, BubbleVoluetricGrowthRate,dg/dt (Ad) 3. ' dater Velocity, V (Ad) 4 Containment and Dry, ell Pressurization a d 5. Pool Swell Elevation A d 6. Tirne Durations, t ( )1/2 3 60 3 A 3* / 3
.NX PR IEi ARY FIGURE 3B.32-1 VARIATION OF SCALED VELOCITY WITH PRESSURE FOR 1// 3 MODIFIED FROUDE INTERPRETED DATA o RO o o o% __________am____________ oo I 4.0 - I.,. !E l'; o l m le + m (,0 N 3.0 19 (p; 7 9 Q ln 9 I! O 5 FT SUBMERGENCE y d Ig (8.7 FT SCALED) x W l F. A 6 FT SUBMERGENCE + Y j30 l7 (10.4 FT SCALED) 5 l' A -l3 L 1 I;8 a t I" 1.0 -- 1 I I I I o r W O 2 4 6 8 th l'2 l'4
P RO PR ie T '. y FIGURE 3B.32-2 VARIATION OF VELOCITY WITH PRESSURE FOR PSTF SERIES 5702
- e.. o O
3.0 t o $ s$ o/o O 5 0 O Nn 2.0 0 4 h *> e o O 1.0 'i' 2 3 s n 7 8 9 io o ip 7;,I } v
- ,c..
v a
PRotEle 'lMk y Ra vec 36.323E L4 d C.cedesa r & f.2m suli P57F 3 an in J . - y O STE AM TT STS v AIE. TYSTS V CT A Se j / I
- t. 'r v
$'t v xAe d V e C v Y o JE 9e t .v O O 1 k 2r t gg' 1 i 1[ i T c f /c /f 20 2C g FEAM DR Y WELL PAf35 UAC (ps10) go,3 2.33-#6
QUEE~!05 Y. 3~2. L: 57 CT A TTALH mG rjT".5 ^ ti v.. u s e *.,. e....1 .v. s a.c.4 e a r ~.'d.e Ss c.' 4..,- R e ' c- *..' a n. T... *.. e. 4.
- e..
.v. J.e.?. g.2 7.,,. a g, ....eq.".. ....c .g. C. _r. a... _..... .U. .. s. e.J. Tv..J. e T.,... s.,.. e., 4, c. 3./.) a,. c.a .r. .c g. sca'e FE!? da:s :: Tredict the 23E Standard Plant Feel 5 e:: velcci:y. ee...-.. w. 2 r
- c.,. t..
n_. .v.u... r. c. c.; Cs. e.., ..t.. ....s g, ~ 39 262-17
Attachment A, Response 3B.32 - Development of Modified Froude Scaling Relations 1.0 Introduction Modified Froude scaling relations for evaluation of the pool swell transient in the GE Pressure Suppression Test Facility (PSTF) are developed in this Attachment. Attachment B evaluates these Modified Freude relations in comparison to exact Froude scaling relations. 2.0 Sc a l in _c Ar.alysis The pool swell transicnt ca:. be divided into 3 sequential events: drywell pres-surization, vent clearing, and LOCA bubble fermation. Each of these events may be analyzed en a first principles basis to develop a set of scale factors for s each event. The principles of conserv; tion of mass, momentum, and energy are used te develop governing equations fo'r each event. These equations are then non-dicensieralized using a pr edetermined set of dimensional factors. Scalt factors are then deterrined fro-the Iting dimensionless equations. In this Attachment, only the governing equations f or LOCA bubble f ormation will be described in detail. 2.1 Dimensional Factors Tc make the governing equations dimensionless, it cust be decided which test variables are " controllable" and therefore suitable for use as representative parameters, and which are " response" variables. Here, the variables under control are assuced to be: (a) PSTF geometry; (b) initial states of the drywell air, containment air, and pool water; and (c) steam flow rate supplied to the drywell. Other variables may also be significant, for example, pool-15 5.1. 3 2 /C - /
2.1 Dimensional Factors - Continued swell velocity, but since they are not directly controlled, they should not be used to specify test conditions or to rationalize results of various test configurations. The controllable variables can f.e combined in various ways to f orm a set of parameters whose fundam(ntal dimensions are suitable for nondimensionalizing the response variab!es. Those el usen here..re: o pressure: ,gH o taass flow: .d.gH o time: cHd'/d si o length: H o velocity 5 ;/cd' o volumetric flow: d. gli o temperature: T si o gas constant: F s specific heats: c ,c o o mass mai o heat flow: cd i s'g H o enthalpy: i. s1 o heat transfer coefficient: k/d (Any other set of independent groups will lead to the same conclusions). Capital letters will be used as the nondimensional form of the corresponding lower-case letters which represent the various variables; the only exceptions to this are the use of Y to represent nondimensional time and 6 t_o _r ep r e s en t nondimensional temperature. Whenever confusion might result, a superscript bar is used to indicate the nondimensional quantity. 383.2.324 -.
2.2 Governing Ecuations A diagram of the PSTF geometry and the mass and energy flows associated with pool swell is shown in Figure LA. This model is used to develop the govern-ing equations. Specific assumptions used in developing the cathematical models are: the air / steam mixture in the drywell behaves as an ideal gas o o the drywell gas has unifort properties internal motions of the drywell gas are negligible e o the increase in volume of the drywell is negligib2e as the weir water level drops o the steam flow in the vents is nearly incompressible (i.e., low Mach number flow) o the flow in weir and vents is one-dimensional o boundary layer ef f ects on the drywell wall, containment wall, and basemat are small. All theta assumptions are reasonabic and in accord with actual PSTF behavior. Using the conservation of cass, comentum, and energy equations and the ideal gas law, governing equations are developed for each phase of the pool swell phenomenon. These governing equations are then non-dimensionalized using the dimensional factors described in section 2.1. Scaling relations are then determined f ro the dimensionless equations. Development of the governing equations during the LOCA bubble formation and growth phase of pool swell is described in the following paragraphs. The drywell pressurization and vent clearing governing equations were developed with a similar analysis. 3 8 3 2. 2 2 A.2
~ 2.2.1 LOCA Bubble Formation, Growth and Pool Swell Governing Equations Following the LOCA, the drywell pressurizes and the top vent eventually clears, followed by the lower vents. A gas flow into the pool is then established. Af ter the vents clear, the drywell begins to depressurize. Eventually, a balance is reached between the steam flow into the drywell and the air / steam flow through the vents. At this time the drywell to the wet-well pressure difference is approximately the hydrostatic head at the top vent. For simplicity, a two-vent geometry is assumed, with the bottom vent not yet being cleared; this simplification is sufficient to show all the required scaling relations. Governing equations during this period are now presented. 2.2.1.1 Drywell Pressurizcticn Conservation of Mass and Energy Conservation of mass in the drywell gives, in dimensionless f o rm : \\ d.y f2 \\ fR T d d cd igH s si p g _ b) d' k ) dd s c ( si d (2-1) fd6 dR h d + d + (P + LP / di d r d. d R,d T c The energy equation for the drywell during this period is: b \\' ^ s- [(M 'si vd d] =
- d eiH sd)
C +M e di a ig-Q) (M +M I s dc d \\m. m .c T s1 al vs si (2-2) 38 3 2 32 A-4
~ Empirical correlations must be used to compute heat losses through the walls, n e r te steam condensation, i qd, dc" State-Dependent Pa rame t ers The variations of drywell gas constant and specific heats due to steam addition are governed by: M + MC C (2-3) = vd Msd + Ma M II M + sd aa R d Msd + M ~ a The steam and air masses present in the drywell during this time are governed by. nf" (2-5) M = a "v.a. 1 sd cd.5\\ IcHd dM I (M -Mdc) (2-6) di k j s (si ) ai Empirical Correlat ions Empirical correlations are needed to represent the complicated rate-dependent interactions of the drywell with its environment. The forms chosen here for the correlations are meant to introduce correct physical relationships rather than exact quantitative predictions: 38 3. 2 32A -S~
Empirical Correlations - Continued / IIi \\ !K# -hd (2-7a) si d si I (e -9) Q = d d e (pd, H / (m. d/ vg i si si E"*
- Y
(- } fn h = E t. a s d where fn[...] implies a functional dependency. Condensation on the walls is correlated by a relation of the following form (Ref erence 1) : I 9 3 - T )3 7 / ~9 3 ~ j r k (T ga /ui { 'l fn [ geometry] (2-7c) M = dc I cd' *9i l \\ 2.2.1.2 Lower Vent Response The momentum of the steam flow is neglected because of the much smaller density of steam in comparison to water. The resulting governing equation for the lower vent has the f orm: !bsi 2 si [a(1-Z+II)+1) 2 2 SP -3P -LP =Z-a V + d f w2 I v2 (cd 9i ) kcd[giif (2-8) I dV dV j + u(1 - Z) dT di { l 38 3 2. S 2 A-to
~ Here, LP '2, the wcter pressure relative to hydrostatic at the vent exit, can be conceived as being due to two distinct effects: the turbulent mixing pressure loss and the pressure required to accelerate the pool water. Turbu-lent losses are proportional to the vent velocity squared and given by an empirical correlation, Equation (2-10d). The acceleration pressure LP is governed by the vector field equ.r.tions of incompressible inviscid liquids;
- i. 9 =
0. ..... in the pool (2-9a) V . at the top vent exit (2-9b) V = x v1.... V t the bottom vent exit (2-9c) V = x v2 n.Y = 0. ..... at solid walls (2-9d) 0. at the free surface (2-9e) .P = [The x-coordinate, in Equatiens (2-9 b) and 2 -9 c) is directed horizontally.] The pressure JP and the water acceleration are coupled by the cocentu: equa-tion and the conditions at the vent exit: -(6, / d.35) df/d: (2-10a) Y(iP) = 6 \\ ,,2 P, - iP + K $1 'v2 (2-10b)
- ~
I. ( cd'*EIi / The empirical f actor f or the turbulent pressure drop used in Equation (2-10b) 1 i is given by a correlation of the form: 1 2 ud geometry (2-10d) L. -_ fn 1 -"si" l l l 38.S.7 32A-7
2.2.1.3 Bubble Charging Rate After the top vent opens, a gas flow develops and f orms a rapidly expanding bubble in the water, The difference in pressure between the drywell and the bubble is small, and the flow velocity is subsonic. Therefore, modified forms of incompressible flow relations are sufficient to define the important phenomena. }1/2 l . + LP - LP )' - P'F __] j [ gH ) 1/2 j 2[(Pd2 d f M =C l b m RT } -d (*d + tb) l ( s si / R j I ( '.'- 11 ) The change in density of the gas as its te perature decrcases upon contacting the water has been taken into account by using the total drywell pressure rather than the drywell pressurization, IP in Equation (2-11). The bubble d, gas temperature, I is s me average of the drywell and water temperatures. b, An empirical correlation is needed to determine the vent discharge coefficient, C in Equation (2-11).* This equation has the same f orm as that of Sonin and Huber (Equation 5, reference 2). The volumetric growth rate of the bubble is written similarly as:** 1 / ~' 1/ '7 1/2 dV- )l, P P -M'bc [R T R9 .+LP,-LP si) [ db ) g C s di o f h m l d' c \\EH ) \\ l+ I b ~ b d t (2-12) where c = P /R T and steam condensation, 6 at the bubble-water interface b db must be given by an empirical correlation.
- The use of C, in Equation (2-11) is analogous to the discharge coefficient commonly use2 f or nozzles and orifices.
An orifice in the vent would give a smaller C, and a smaller flow rate, term is neglect ed to give
- In the right-hand side of Equation (2-12) the Mbc a conservative volume expansion rate.
.38.3 2.324-8
2.2.1.4 Pool Water Response The governing equations for the accelerated motion of the pool water are: V.$ 0 in the water (2-13a) = lcd 3I d b - 4 fn VdA at bubble interface (2-13b) = l I Ii di (m ) si P P ~ (c/ 8H }/R a u e in erface (2-13c) b b 0 at solid walls (2-13d) n.V = V V at bottom vent exit (2-13e) = X Y7 LP at free surface (2-13f) Lp = 2.2.1.5 Containment Airspace Pressurization As the pool water is lifted by the bubble, the air in the containment is com-pressed. Assuming fot simplicity a polytropic process, the compression is modeled as: dV_ dV d: ~ di ~ " and P + LP dV_ ci c g LP = -y dT ~ ci c Equations (2-1) through (2-14)are adequate to determine scaling relations for LOCA bubble growth and pool swell. 3B. 3. 2.32A-9
3.0 Development of Scale Factors The set of governing equations discussed in Section 2.0 may now be used to determine the scale f actors which will allow subscale PSTF data to predict the full scale Mark III perf ormance. The scale factors are developed by equating each governing equation in both scale model and prototype and determining the scale f actors which will ensure the model/ prototype equality. The results of this procedure are summarized in the following paragraphs. i For this analysis, the following assumptions are made: 1. All gecretric factors and initial flows and pressures can be adjusted at will, regardless of whether such adjustcents are practical in the PSTF. 2. Prototype fluids (steam, air, water) are used in the model 3. Water temperature of the model cannot be changed markedly frem the prototype (e.g., the water would freeze). This implies that all temperatures should be protypical. Scaling relations were developed for each of the phases of a blowdown. For convenience, the relations are presented in terms of the geometric scale factor for vent diameter, i E*" s, d' H. d d prot otyre 3-1 = model ) H H prototype 3-2 H = model .l B 3. 2. 3 2 A -/O
Drywell Pressurization and Vent Clearing 3.1 The scaling relations f or the drywell pressurization and vent clearing phases are tabulated in Tables lA and 2A respectively. These relations were developed by equating the governing equations for the model and the prototype. The model parameter values are defined as X X ~
- P#
YP medel where A is any controllable or response parameter and x its scale factor. The scaling relations thus developed and summarized in Tables lA and 2A do not simulate all physical effects. However, those effects which are not simulated have negligible effect on the validity of the scaled results. These nonsimulated effects are discussed further in Attachment B. The scaling relatiens shown in Table 1A through 2A show that all significant length dimensions are scaled in the same proportion and all velocities are scaled in prcportien to the squarc root of the length reductien. Since the Froude number is defined as ?2/Hg F = r l we see that 1/2 12 1[d T/ t A Hg e del Hg _ prototype H F F = r r model prototype (for T = 1d) H l These scaling relations are thus referred to as exact Froude scaling relations. l 38.3. 2. 3 2 4-H
3.2 LOCA Bubble For=ation Crowth, and Pool Swell From equation 2-1, it is seen that the drywell pressurization during this time is scaled exactly as it is before the top vent clears. The scaling relations are 2 1/2 1 = 'si c H (3-4) m. h d al = y 'd d' A 1 4 H H 2 Equation 2-11, shows that the bubble charging rate will be correctly scaled if k A A (3-5) C F' i Pdi 1 = H where A =A =) =) P LP LP P di d f B Since the 1/3-area scale PSTF is geometrically similar to a full scale Mark III, we know that A =1. Thus, in order to properly scale the bubble charging rate, Cm Equation 3-5 requires that L 1 -A"' (3-6) Edi 35.3.2. 5 2 A -12
This requirement, however, conflicts with the pressure scaling relation derived f or the drywell pressurization and vent clearing transients () =Ag). nu s, a scaling compromise =ust be made. The selected compromise, which is consistent with the PSTF tests is to use full scale initial drywell pressure, i.e. A =1 (3-8) P di but A (3-9) =' => = )LP P LP d d f g The conservatism of this compromise in predicting the full scale pool swell transient is discussed in Attachment B. Equation 2-12 indicates that the bubble volute growth rate, dV b follow mass flew scaling: (3-9) 'dV /d: ( d) (;H) = b Of course, this assumes that pool dimensions, pressure, etc., are scaled appropriately. Scaling cond it ions f or these phenomena are now examined. 1 Equations 2-/. band 2-13b show that pool dicensions should all be scaled in proportion. Otherwise, the LOCA bubble will expand in some directions at the expense of other directions, in an unscaled fashion. Therefore, the scale factors for pool width and breadth should be the same as f or depth.
- Further, when this is true, the submergence H must be scaled in proportion to the This can be seen by the following argument.
The vent diameter, > ^ H d. LOCA bubble volume is proportional to the integrated volumetric growth rate: 2 1/2 1/' I(dV /dt)dt =V; and the integral is scaled as [ ( d H H d H' ~ b l 3 The tota] poo! volume is scalod, however, as >H' "C' l be scaled in accordance with the pool volume, it is concluded that >H d' l 8b5E.2,824-/3
Scaling relat f orir fer the cc.ntcinnent airspace pressurization are developed fro = Equations 2-4a and 2-4b. The initial containment pressure is the same 1. Following through with as the initial drywell pressure, so I Pci Pdi the procedures used previously, it can be shown that the containment pressurization will not be simulated in the sace way as the drywell pressurization (i.e., A p [ j) unless (a) the volume change of the containment airspace is stall compared to the initial volume; and (b) the containment pressurization is small compared to the initial pressure. Since both of these restrictions are cet in PSTF tests, Equation 2 ' shows that the containment pressurization will be scaled in accordance with the drywell pressurization when the containment volume is scaled as: o (id)^ (3-10) V. = C1 Equation 310 expresses an important observation: the containment volume must be scaled in agreement with the drywell, for otherwise the pressitre dif f erential restraining the pool swell will not be simulated accurrtely. It should be noted that when ) 1 and the drywell volute is accurat ely = 2 scaled ( y=Ad d e actual hpell ah mass will aduaHy be scaled by A d 3 (not I as requir d by the governing equations). This results directly from the d ideal gas equation. This lack of siculation directly af fects only the drywell energy balance, equation 2-2, and is relatively snall because the difference between A and A s not large in comparison to the mass aMed to W hpell d d by the steam flow. Later in the blowdown, when the drywell air mass has been exhausted, there is no error in the simulat ion. %W
The scaling relations needed to simulate LOCA bubble growth and pool swell are sum =arized in Table 3A. These relations, except for those for drywell initial pressure and air mass, agree with the pressurization and vent clearing relations (Tables lA and 2A) when the requirement that ). H ^ d E relat ions of Table 3A may be used to scale the entire pool swell transient. This set of scaling relations are referred to as the Modified Froode Scaling relations. t 1 t l 3'83.2.324-15
Table 1A FROUDE SCALING RELATIONS FOR DRYWELL PRESSURIZATION MODEL = PROTOTYPE X SCALE FACTOR Independently Varied Geometric Parameters Scale Factor 1. Top Vent Submergence, H ). H 2. Vent Diameter, d >' d Dependent Test Parameters 1. Drywell Volume, Y (A d d 2. Drywell Air Mass, m l' H ( d g5 3. Drywell Initial Pressure, p i A d H 4. Stean Mass Flow Rate,in (d H si Pressurization Response Parameters 1. Drywell Pressurization, ap d H (A 2. Time Durations, at H 38.32.324-/(s
Table 2A FROUDE SCALING RELATIONS FOR VENT CLEARING MODEL = PROTOTYPE X SCALE FACTOR Independently Fixed Geometric Parameters Scale Factor 1. Top Vent Submergence, H A g 2. Vent Diameter, d g Dependent Test Parameters 1. Drywell Volume, t I d) d 2. Drywell Air Mass, m A ai H'Ad} 3. Drywell Initial Pressure, p i H ~ A a 4. Stean Mass Flow Rate, in (Ad) ( 4} si 5. Vent Length, I
- i. H 6.
Vent Vertical Spacings, H, H, etc. A 2 3 q 7. Pool Width, w A d Response Parameters 1. Vent Clearing Times, t (AH) 2. Water Jet Velocity, v (AH) 3. Water Jet Acceleration >' H 4 Water det Penetration A d .38.3.2 32 /).f 7
Table 3A MODIFIED FROUDE SCALING RELATIONS FOR POOL SWELL MODEL = PROTOT/PE X SCALE FACTOR Independently Varied Geometric Parameters Scale Factor 1. Vent Diameter, d A d D_coendent Test Parameters 1. Top Vent Submergence. H A d 2. Drywell Volume, 4 IAd) d 3. Dryweli Air Mass, o (x )2 g d 4. Drywell Initial Pressure, p 1 3 5. Steam Mass Flow Pate, in (* d) si 6. Vent length, t A d 7. Vent Vertice l Spacings, H, H, e c. A 2 3 d 8. Pool Width and Brer A d 9. Pool Volume (Ad) 10. Containment Airspace Volune, t (Ad)2 c5 1.1. Containment Initial ?ressare, P I d Response Paranaters 1. Bubble _ Volume, t (^d} b 2, Bubble Volumetric Growth Rate, d i /dt (Ad) b 3. Water Velocity, V (Ad)1/2 4. Containment and Drywell Pressurization A d 5. Pool Swell Elevation A d [ 6. Time Durations, t )1/2 l l l ? .38 5.2 3 24 - is
m 3 gI " g n H t w w f.G u,' T r E C _t N s A E e F M s_ R, I N U w S e A C T P R= N E A EA C TE OC AR A WA P lc c s P T RI
- =
I E L A y V T c e E E N H ~ L / E = U M T R ~ E n A E NI R w M, T A U R A o E W d T S P S N S w E M S n O A C o C R E I E v. r T M 3 T L l A s B b T s T S us F sP f d T vn b S E P c h
- 1 i
I f 2 o w v .Ns l sO* sI e D "oT A t t d 4 E e l RC w R o l CE M (S A h A d d r T d R .l l e W 2 i 9v i z =
t w a w i , E T 54 l d r R N Y.y 3 V = U A a T T e L = e S d A L E n N I E M u R O r E W U C h i 1 s P WL E M S \\ f O n E A \\ A O V e T G 1 Y W T O I C L e O F r L N u S E S V g A i M W F O M L n A F D 5 E E l r T T R e S E = f e h T a e R C 4 d TE I A NR A W EU R S MT E E 1 y-E NA T T T OR A A A M E C I L S P I D d s I, L N IVM D D E E NE N Y WD ET I NI A h> P YN ,S M T L RO ,,S AE A OC ,O EL H L TN T SI N E E TON hgg Dtahrg J
LIST OF PRINCIPAL SYMBOLS Nondimensional equivaients are shown in parentheses. a(A) area c.c (C,C ) crecific heats at constant pressure and volume p v p v C vent discharge coef ficient 7 d vent diameter g acccleration of gravity h(h) convective heat transfer coefficient 11 top vent submergence H ' }t3'"2'"3' "I"'" 2 i, enthalpy of stear flow i (I ) latent heat of vaporization f f k turbulent viscosity }:1 vent friction loss coefficient E (L) vent length m(M) mass m(M) mass flow n e:< pen en t of Leynolds number in heat transfer correlation p(P) pressurc q(Q) heat flow 393 2.32s).2o
LIST OF PRINCIPAL SYMBOLS (Continued) r chugging bubbic radius b R(R) gas constant T(T) t ire e T(:) te peratur( vlV) v e l o c i t :, 4 (V) volute z(Z) water level in weir with respect to initial level a ratio of vent arca to weir cross-section area ratio cf specific heats scale factor (model to prototypc) of jth parateter J p viscosity c density (or water density without subscripts) c surface tension 33 3.2.524-2/
LIST OF PRINCIPAL SYMBOLS (Continued) SUBSCRIPTS a air b LOCA bubble c con t a inmen t air space, or condensation d drywell e outside environment f frictional g gaseous state i initial valuc, or condensation interface s stea: v upper vent 3 vp lower vent w water, or weir Note: Combinations of subscripts are also used. For example, "sd" would indicate a parameter associated with drywell (d) steam (s). .FB3.2. 32 A-22
REFERENCES 1. F. Kreith, Principles of Heat Transfer, International Textbook Co., 1958 2. A. A. Sonin, P.W. Huber, "On the Scaling Laws for Air Clearing in Water-Type Pressure Suppression Systems," ASME Journal of Heat Transfer, Vol. 100, November 1978. .3832.52#22
ATTACHMENT B, RrSPONSE 3B.32 - EVALUATION OF MODIFIED FROUDE SCALING 1.0 Evaluation of Modified Froude Scaling Relations The Modified Froude Scaling relations require full scale initial pressures (A = 1) instead of initial pressures scaled by A as required by exact p d Froude Scaling relations. The effect of this modification is evaluated in the following paragraphs. In addition, several other phenomena, such as heat loss and viscous forces, are not accurately simulated. Each of these nonsimilarities and their effect on the subscale prediction of pool swell will be discussed. 1.1 Effect of Full Scale Initial Pressures As showr. in Attachment A, exact Froude scale simulation requires the initial pressure to be scaled as A p,=AH, out this conflicts with the pressure d1 scaling relation for LOCA bubble mass flow (A =A H ). A c mpr mise Pdi was suggested to be A = 1; that is, use full scale initial pressures. p As a result of this compromise, the initial drywell air mass scales as d instead of ).d. This compromise affects ill 3 phases of the pool swell A transient. The effects of this compromise are discussed in the following paragraphs. .78.3.2.328-/
l.1.1 Vent Friction Loss Coefficient The Modified Froude scaling laws derived in this response do not require increased friction losses in the sub-scale vents, as required by exact Froude relations. Exact Froude relations require a reduction in the initial drywell and containment air pressure in proportion to the vent diameter scale (Ad). This implies that gas densities are also reduced in proportion to A The ratio of the sub-scale d. frictional pressure W:p in the vet system to the prototype is thus (ap }m * ( l Og d )m r ( l-l ' (apf)p = (K) (O V g d )p Since A =a ' A ,C, the scale factor on the vent loss coefficient q"Ad, nd A = p d d y is A
- l/ A g]
d (1-2) That is, the sub-scale vent must have a larger resistance than the protype in order to correctly scale the vent frictional pressure drops. The scaling relations proposed in this response attempt to maintain gas densities at near full-scale values by maintaining full scale initial drywall pressure and temperature; when )- = 1, the implied scale factor for K) is unity - in other words, the sub-scale vent resistance is the same as the full scale vent. 1.1.2 Bubble Charging Rate Sonin and Huber (reference 3) have shown that with exact Froude scaling relations / (i.e.,gdi, A' k*Ad ) the bubble charging rate scales as A Thus, to = d d m achieve exact simulation of pool swell in a subscale test ( A <l), orifices must d be placed in the vents to decrease the vent flow such that the ratio of model to .283. 2. 32 B-L
full scale bubble charging rate is A /. In the PSTF, however, A d Cm because there were no orifices ir. the vents and ). =' As dei.onstrated below, p this results in an excessive bubble charging rate. For the 1/3 area scaled PSTr tests, A = 1/ = 0.577. Thus, for exact d Froude scaling, the model bubble charging rate is (0.577) /2 = 0.15 of the full scale bubble charging flow. For a design basis LOCA, the initial drywell pressure is 14.7 psia and the peak drywell pressure at the time when the top vent clears is ~32.2 psia.
- Thus, pi = 14.7 psia d
(1-3 )
- 3. p
= 32.2 - 14.7 = 17.5 psia = 1.2 pdi d The bubole pressure, p, at this time is equal to the sum of the containment b airspace pressure, the hydrostatic head, and the pressure needed to accelerate the water to confom to the bubble growth; an estimate of the water acceleration pressure is about 8 psi.* For a submergence of 7.5 feet, p e 25.9 psia = 1.8 p i' b d For a Modified Froude scaled PSTF, Ap = (.577)(17.5)= 10.1 psia = 0.7 pdi. The d submergence is (.577)(7.5)=4.33 ft., so the bubble pressure, p ' iS b 1 p = 14.7 + 4.33 (62.2) + (.577)(8)= 21.2 psia = 1.4 pdi b4) b 14* Equation 1-5, the dimensional fom of the bubble charging equation (Appendix 1 A, l equation 2-11) may now be used to evaluate the ratio of the model to prototype bubble charging rate. 6 =C rd 2 (pdi + OP ) -U b m d b 4 R (Td+T) J (1-5) b
- This pressure is not known with certainty; the value used here was inferred from drywell wall pressure measurements (Reference 5).
The exact value is not i crucial to the argument; practically the same answer is obtained for any value such that pb is somewhat less than pd, which it must be to charge the bubble. 38.52.52s-3
Remembering that the model and prototype C are equal and d m MODEL d proto'** see that 9y bbmodel_(.577)2.((1.7pdi) - (1 4 Pdi) / ,2
- b/proto
~2[(2.2Pg )2 -(1.8 Pg)2 = (.577)2( ),72, ),4 ) b 2 2 2 (2.2 , ),g ) = 0.25 [ Modified Froude Scaling) (1-6) The bubble volumetric expansion rate is characteristic of the pool swell velocity The volumetric expansion rate, as described in Attachment A, is written as f=bP (I-7) b/ b where / " P /R T and steam condensation at the bubble-water interface is ignored. b b d b Since temperatures are full scale, the bubble density is directly proportional to bubble pressure. From the above calculation of m ' b , model _ P F bmodel _ l.4 = ~8 Y l'O fproto bproto l l Thus, for Modified Froude scaling, l P .odel bmodel bproto _ 0.25
- 31
= k roto _ b _P 0.8 p bproto bmodel Sinilarly, for exact Froude scaling, l Dbmodel =A = 0.577 d Pb roto p (
so model 15 = 0.26 = g .577 proto Thus, the Modified Froude scaling relations will result in pool swell velocities which are on the order of 191 ((.31.26)/.26x100) too large. In Reference 6, a first order approximation of the conservatism of the Modified Foude scaling relations is made by comparing the dimensionless tem / % t.s 1 F % d n. (1-8) ? in the prototype and model. This parameter was assumed to be controlling because it models the bubble volumetric expansion rate. For prcper scaling, T) = TT) Reference 6 shows that r, l' Pd (1,9) P model b model 2 7 1*T 2 proto d Ib roto By assuning p ' P in both model and prototype, Reference 6 concludes that b d TTImodel 1.06 = TT2 roto p 38 3. 2. 32 B-S~
Mich leads to the conclusion that the Modified Froude scaling relations are on the order of 10% conservative. However, using the above values for bubble and drywell pressure (p 'N P ) in equation 1-9, i.e. b d Pd Pdi redel p
- I4 P b
di g p
- 2'2 P d
di prou p
- I'b Pdi b
it is seen that Tr I model = 1.13
- 2 proto Thus, the use of theTf ratio, which is a first order approximation that j
neglects other terms in the bubble volumetric expansion rate equation, predicts - 13: conservatism in pool swell velocity. A more detailed analysis described previously, predicted ~195 conservatism. Therefore, it is concluded that the l Modified Froude scaling relations are on the order of 15% conservative for predicting pool swell velocity. i l.2 Heat t.osses The suggested Modified Froude scaling of mass flows does not accurately duplicate prototypical Reynolds numbers. Therefore, since heat transfer coefficients typically depend on Reynolds numbers, heat losses may not be accurately simulated by Froude scaling. This distortion is now examined. .38.3 2.328-6
A typical model of convective heat transfer is: d7~ 5h (1-8) d 86 " Pg] ~ M/# Here, P is the perimeter of the flow cross-section in question. For Modified Froude scaling of length, time and mass flow, the scaled equation in terms of prototype is: dAhbI* rg ~[ C d y (1_g) Therefore, accurate simulation of convective heat transfer requires that the /. Experimental values heat transfer coefficient scalina relation be ).h d
- A of h can usually be correlated as a function of the Reynolds and Prandtl numbers: h = (k/P) C)(Re)"(Pr)" (Reference 4 ).
The Prandtl number is dupli-cated exactly, but !1odified Froude scaled Reynolds numbers for water are ().d times prototype value, and for gases are between (>d) and (>'d mes prototype values. (The variation for gases depends on whether the density is accurately simulated.) Since "n" in the heat transfer coefficient correla-tion is approximately one, tne convective coefficient is concluded to be simulated accurately for water, but may be too small for air or steam. Thus the energy loss may be simulated inaccurately, and the energy available to drive the LOCA bubble may be too large. ftdified Froude scaling is therefore conservative with respect to convection heat transfer. Heat transfer directly affects pool dynanics in only a minor way, so the degree of conservatism is i small. i l 38.3 2.328-7
1.3 Viscous Forces Modified Froude scaled water and air / steam Reynolds numbers are less than prototypical, as shown above. Hence, viscous friction is overemphasized relative to other phenomena. Numerical estimates of this distortion cannot easily be rtade, although it would tend to make test results nonconservative. But for short ducts, such as the weir and the vents, viscous losses are small compared to geometric-induced losses (entrances, exits, branches), and these geometric losses are scaled accurately. In the pool, viscous forces are small throughout. Consequently, the overemphasis of viscous effects by Modified Froude scaling should result in a negligible difference between model and prototype condi tions. 1.4 Steam Condensation On solid walls, the rate of steam condensation is proportional to the parameter 1/4 a [: gk (T - T ) )Ji bl , where a is the wall area and b is a s c fg significant length dimension (Reference 4, Chapter 10). For Modified Froude 1 I
- scaling, condensation on the walls therefore has an implied scaling factor of (Ad) /.
The desired scaling relation is (Ad} the same as other mass flow rates. Since (Ad} (A IU " Ad<1, the d simulation of steam condensation is larger than prototypical, less steam is available to pressurize the drywell and Modified Froude scaled tests could be ( nonconservative. For this reason, the drywell should be pre-heated, as is done l l in PSTF testing, to minimize condensation. In fact, the PSTF drywell pressuriza-tion tends to be conservative because there was negligible wall condensation. i i 3B.22<.32B-8
There is also condensation at the bubble water interface, which is modeled as k. ha (T - T)/i The heat transfer coefficient for this kind of = 3 f. condensatior. is a functior. of the Reynolds ar.d Frandtl number and of air content of the stea, decreasing with increasing air content. Modified Froude scaling, wher.>pg5 = 1, overestinates the relative air content initially and under-estirates tne :.ejncids ru.ter. Tre sa e kind o# argJ ents used to estirate heat losses car. be used here tc sho< that the siEJ1atee bubble water stear c or.de n s a ti c e is therefore nct greater than prototypical and may be less.
- Thus, Modified Froude scaling gives an accurate, or perhaps a conservative, simulatiori c' the e'fect of battle water interface condensation on loads t
i l l 1
- 38. 5. 2 32 B-9
LIST OF PRINCIPAL SYMBOLS Nondimensional equivalents are shown in parentheses. a(A) area (C,C ) specific heats at constant pressure and volume P,c c V P V C vent discharge coefficient d vent diameter g acceleration of gravity h(h) convective heat transfer coefficient H top vent subnergence H 'H ver ca spacing of lower vents 2 3 2' 3 i, enthalpy cf steam flow i,, ( 1, 6, ) latent heat of vaporization .e k turbulent viscosity K vent f riction loss coef ficient y E(L) vent lengt h m(M) mass m(M) mass flow n exponent of Reynolds number in heat transfer correlation p(P) pressure q(Q) heat flow l l .38.3 2.3 2 e - to
LIST OF PRII;CIPAL SYMBOLS (Ccntinued) r hugging bubble radius b R(R) g a s :<. r. s t a n t T( ) time T(r) terpcrature } v(V) eclocity V (V) volume (Z) water level in weir with respect to initial level a ratio of vent area to weir cross-section area y ratio of specific heats scale factor (model to prototype) of jth parameter j L viscosity c density (or water density without subscripts) c surface tension S/532.32-/]
....= - _ _. -.._.--____- _.. d t j ] LIST OF PRINCIPAL SYMBOLS (Continued) ( I I l SUBSCRIPTS I I i a air b LOCA bubble c containment air space, or condensation d dryvell e outside environment 1 f frictional I, g gasecus state i initial value, or condensation interface l c model p prototype i s steam y vent I j w water Note: Combinations cf subscripts are also used. For example, "sd" would indicate a parameter associated with dr>vell (d) steam (s). l p 26 22. 32-/2
REFERENCE 1. W.J. Bilanin, "The General Electric Mark III Pressure Suppression Con-tainment System", General Electric Co., NED0-20533, June,1974. 2. SOLA-V0F: A Solution Algoritm For Transient Fluid Flow With Multiple Free Boundaries --- Los Alamos Lab - LA-8355, B.D. Nichols, C.W. Hirt, R.S. Hotchkins, August 1980. 3. A.A. Sonin and P.W. Huber, "On The Scaling Laws For Air Clearing In Water-Type Pressure Suppression Systems," Trans. ASME, J. Heat Transfer, 100, pp. 601-604, 1978. 4. F. Kreith, Principles Of Heat Transfer, International Textbook Co.,1958. 5. R.J. Ernst, T.R. McIntyre, L.L. Myers and J.E. Torbeck, " Mark III Confirmatory Test Progran. One-Third Scale Three Vent Tests (Test Series 5801-5834)", General Electric Report, (NEDM-13407P),May,1975. 6. A. A. Sonin, " Review of General Electric's Data Base for Predicting Mark 3 Pool Swell", Thermal Reactor Safety Division, Brookhaven National Laboratory, Upton, N.Y., March 21, 1977. 38.3232-/3
Attachment C Modified Froude Interpretation of 1/3 Area Scale PSTF Data To Predict The 238 Standard Plant Pool Swell Velocity. In order to quantify the conservatism of Modified Froude scaling in using the 1/3 area scale data to predict Mark III velocities, an analysis was per-formed using the 2-dimensional Marker and Cell (MAC) code described in Attach-ment D. First, this code was used to find the peak pool swell velocity in the 238 Standard Plant using actual conditions and the CESSAR pressure history. A second run was made using the 1/3 area scale PSTF geometry, but with both the GESSAR drywell gage pressure history and initial submergence scaled down using Modified Froude scaling. In both runs, the maximum pool velocity was taken to be the velocity at 1.73 x initial top vent submergence (13' for 238 Standard Plant). The subscale velocity was scaled up using the Modified Froude scaling relations and compared with the 238 Standard Plant velocity to obtain a measure of the accuracy of Modified Froude scaling. This simulation showed Modified Froude scaling to be 7.5% conservative. The 1/3 area scale PSTF f acility may be used as a 1//3 Modified Froude linear scale test facility. The only geometrical compromises in the facility were that the vents were too long (5 feet as compared to desired length of 2.9 feet) and the spacing between the vents was too large (4.5 feet as compared to the desired spacing of 2.6 feet). These parameters delay top vent clearing time and cause a larger drywell pressure history throughout the transient. The larger drywell pressure history causes a greater pool swell driving force which is balanced by the fact that a larger slug of water could be picked up. An analysis with the t l GE pool swell code showed that the effect of the higher drywell pressure due to the out of scale parameters was to increase the pool velocity N17%. Thus, the increased driving pressure (P*) due to the long vents and spacing has a con-servative effect on the maximum pool swell velocity. The effect of these parameters has been accounted for in the calculation of P* for the individual tests. This explains why so many of the data points in Figure 3B.32-1 have such a large value of P*. The 238 Standard Plant P* was not based on these test data, but rather on the calculated GESSAR pressure history. Experimental l data indicate that the out of scale parameters will not have a major effect on velocity because the 238 Standard Plant design P+ and much of the test data are on the plateau region of Figure 3B.32-1, indicating only minor variations of velocity with pressure. l l 36 0. 3. 2.M c-/
AITACb'ME//[ d GENEML ELECIF.IC'.S 2D M.UJ.EE-AN.>-CE1.L POOL Sk'El.L ANALYTICAL M00EL I. INTE 02'.'CT I O!-- General Electric's peel swell analytical codel evaluates the suppressica pt:1 surface prcfile and vel: cit' in a Mark III centain en: during the pc:1 swe;; transient fc1 cving a ; s:ulated LOCA. The code cceputes the tve-dimensicnal p :ential flew field, representing the suppressien peel and governed by the Laplace e qua t ien, with the Marker-And-Cell (MAC) te:hniques. Taverstle cc paris:ns have been ettained between ecdel predictiens and CE 'ari ::: Press.:re Su;;ressi:r Tes: Ta:ili:- (?!!T) test data. 3,, w a a r, w r.e a;.;. .......C.- u. s. The p:cl swe. : analytical ecdel sir ;ates the two-dimensional ficw field of the s>r. ctr) ;;anc (;1ane AA :n Tigureb) in a Mark 11 centainment unit cell, which is d(fined as the res:ce tcunded by the inner and outer centain-cent va!!s radially, and t', tv: a d.i a c en t s'r... e : rv. -lanes circurferentially. r Thus, a unit cell is a C* se::cr cf the centainecnt. Listed belev are the individual ccepenents of the pool swell ecdel. Scre of these are net part of the pec1 swell code, and therdetails can be found elsewhere. 26o 3.2.32D-l
A $ims \\ rk.*e n = p l a. t U.:4 CcI( f 7 g-%43 p I a..s. t F 'i F CON 7 Af N Wf N"T OL'TE 8 m Ai'. l i i Y I I i i WeL i C ON ? a '% p E Ni I l<.: h [ Q..., / th%ta nA,L
- 4. '. c ).jl
...\\ s.. :... .,t e..... .r e ... =.,.. . ~. s e. 7 l { t '.
- i
.. ;..,: w..... l 1 w.. j A N N.; L UE ws.s WA k = A T1; m L1 Ur.n Cell Flan Vie-M 3 (60.2.7. 3A D - I
II-1. Con; nents cf Pc:1 Svell Mede! The peel swell analytical ecdel is cceposed of the fc11eving ecep:nents: o Dryvell resp:nse rodel o Vent clearing ::dC. c Vent clearing j et =edel o Vent flev : del c e...,., r e.,, <.,.. e., y... >, a..... a ....- s. e,. ....e c s.. t. t.., e. 2....c. . k.. e c... c.~. .a ..e,.,..e ,.e.e... e. e.. c. c o. ..... t _.. e, s.e....,...,,.a a ce.e..c.- r. t -...-. e c '. *. a d.. e.' d ..-.
- h e.-.. *.
.d. e. *. a.. '. v *.
- e.- I
..'. e.' ( F.e ' e c.. e P' ),. -... ' are n:t part cf the p:c: s. ell cCd6 The d-.. ell res ense follew-4..e
- t. e..... C,. e u.....e 4
use.a.. a.... u. 4.- 4., s.. c..= a c. -. af. 5.46 = d 4,.4nb .tCrCe . c: ...r* 6; r--* ..w. 4. .k.e Ve.. [... a. a... L... L. L..* e. C w. s.. 4...a... ...i .e...e a.5. e. ..c.. c. e u..... .e..
- p. r.a. -
t ...e ....,4e ..c.=e 4 e_........;: ....e. 4e. ....e a t. ....e Orde dts: riled in feferentehl t0 C'Olain the j et pene!!atien,Ap. The a.e.c - - * ' - f3 r-. Gt
- '6--*
- >E
- .-.8-.'
c ' - e *.' * %.. e b v L. '..T e '..c C ;.' 4.. d..d e. s '. ' '* d-
- c r w
with a dia eter d and lengt? fy, a s she c in Ti g r e D2. The nese is reunded g.,.,
- u...,
- t. g...<>
p .,.,<u..,e e__-....a _.-...g.::...,.a.......e,.
- 4.. t u.. e..,
y >. ar - a... a., c e..; r.... m. c: del. /\\ l /
- / / p I
r W /// l ,r [r h /! - ,I FigureL2. Initial LOCA Eubtle Prcfile I The following sections describe the analytical =edels in the cede. t t I I i I e e
3 . l..,. \\ e....... v.. 2..e. a .s.e ass.. .,.ir s 4. . h. e s. e........ _.. e, arc. A. Air ficv enly - This re eves any cendcnsatie stea. frer the tubtle and results in ecnservative b thie pressure respense. E. Tann: Tic. - That is, the flew is ent-di.cnsional, co fressible, 3a....s,,..,.....
- 5.. c., e,, e..,
s,-....., 4.. ..3. a .st c 1 p....;t...
- ..t.s.
e C..c.._s.,. s ..,...c t g, s.., +.( pr( s s..r c. .s. .g... c... (. 6 t...
- i..-
.c..... t e,s.
- t., *.
L. '... J. J.... 4. u. .t 4 5.e....:. s.. c.... '.. t .#.'~.. d.(
- ..d.'.c*...t..*."..,
.d*.-*.-.*.e a.e #...'C, .d....d e - d.' s. ' - s
- .e ' a-7.a e t 3.o.4'..
. '.. t e.-.t'.'*..'.e' e",'.s-*.*.. .d.e o...a 7.... 4.. . w... ..s..... , t.. ( c..s... ... c..... }, re. s r. s. .n ..t ....u. - C e.(.. .....a.. n. p. e.1 t L-oa. s eT I l At Free Surfaces: l .,. Z ,gs. ~ l c' - i. e* e- \\. + rJ, P J -. O l <1 s g
- c. '
w w \\ ? = Constant 1 i \\ I 3,@.3 2.D D d l 1
Using finite differences (Figure L3 ), the Lapla:e equation is transforret into a set of sinultaneous algabraic equations. ' liven the initial values of velocity p0:e".tial en the free surfaces, these simulteneous algebraic etuations are then solved by overrelaxatien for the velocity potentials in the interier region, whicn in turn pe --its velccity components and potential rates t: be calculated at the free surfaces. Markers used to describe the pcsiti: a-: sta:e c' ea:~ free surfa:e are then m:ved with their given velocities, pr:. icing t*.e t:u.:arj :.:itices for flow fiel: c p.tation at the nea tire ster. The pro:ess is repeate until finished. b d. S ofa4 t. / / i ?%e .g n n m u o e 4 h MArktr5 WL k - it / r m / '/ A (tl$.5 6 % \\ \\(. r 3 'j $cMt. R lt O 5 0 i 0 6 Rc.gatar Mu k Pat r+s o , __ 3 Legatar M f.s k P t a x FigureL3 Marker-And-Cell Representation of Flow Field gpp.3.2 3 7 b-5
_ _ = 11-4. Bubble Mo:el The assu ption of a perfe:t noncondensable gas in the bubble allows the l deterr.ination of bubble pressure through the relationship big R "I6 ,r- = \\/a 0 i w".e r e F: E.t:1e pressure = w ,z w M = Eabble air cass = tsen. dt o o VC T Drywell te perature (assu e: to re.ain consta.t) = g R Gas constant for air = V = Sut:1e vols e e To define V-, the but:le grextn is civided int tw pnases. In the early pnase one tuttle is relatively small c:rpare t: :ne cell si:e, tn_s the bu:ble growth is pred:-inantly three-di ensional and tne cell walls (i.e., synr.etry surfa:es divicing adjacent bubbles) do nct have a strong influence on the bubble exper.sion. Tne buttle volur.e during this phase is given by the expression (Refer to FigureDe) E{ t
2.a.
/# i -t's L-hi q MN-As f e / X FigureUf. Two-Dimensional Bubble pr: file 360 3.2. 37D-6
b b 6 Kg .A g$_ -- 3 3 ya m in w*.ich the two-dirensional area. A, is given by Green's Tneorer g Ae: f1di-td'3 C andtj is an e ;irical constar.t (to be described later). F:r tne special case where tne ts:-di er.si: al tuttle pr file is an ellipse witn se-i-axes a and :. A = n at g a. V red.,:es :: g V=L ?n at c w*ic* c:*res p n:s t: the vel; e t' a. eili:se rotate: ab ;; the y axis if E iS 5 ' I
- E 1 I U E
The b; ble ; c..: e-ters ine se::r: p;ase one ne lateral ex;ansicr. is cor. strained by tne s1:e wa'is ( t k. a point being when the lateral dir.ension e.,als O a.d is fcr:et : e);and uonare. T.re buttle volu e is then calculated as G V Kv r. =- E 3 ee o For the spe:ial case w".ere the two-dirensional battle ;rofile is ar. ell:pse, the bubble volu e reduces to 4 E V. = 3 V. TT a b w A-e e e c wa.ich is the volure of ar. ellipsoid with se-i-axes a,b, and s/2 if ig is set equal to 1.0. 3/30.3 7.. L t6 7
For the bubble ccnstan: K, a value of 1.25 is rece rended based on data 3 cc parisen and physical argumen:. III. MODEi-0ATA COS:P AE15 CG.: The peel swell analy:ical codel predictions have been ec: pared to PSTF test data. A tctal cf eleven test cases were used in the cedel-data cc parisens. Only air bloud:ur tests were used it. the cerpariscns. The parameters cc pared are the p::1 surfa:e elevsti:r hist:ry, the surf ace vel :it y as a fun:tien cf elevati:n, the p::1 su:fa:e and tuthie pr: files and the water ligarent thick-ness. i FiguresL6 threughlu0 sh:v typical results cf redel-data cerparisens. Thece I cceparis:ns lead :c the f:11 ving c:n:'usiens: i 1. The e: del predi::s 7:: surface eleva:::n accuratel T.e >c'. ...-..c t... -..e. e <. r. e > <.-. <..,-. s.. e.,. '. 3;;,<a:e ve,c:,:. 33 3 f t. c. a. .z ....z.e e.e....i.. 3. The n:dcl pred: cts a fla::er p::'. surftee pr: file and a larger tuttle than is sh:vn b; the da:a as a result cf an ever;redi::1:n cf initial buttle pene:ratien by the :: del. 4 The =cdel everpredicts the pcci ligarent thickness throughout the transient, aEain due to everpredi:tien cf initial bubble penetration. 5. Be:ause of inheren: uncer:ain:1:1es associated with the bubble breakthrough, a cne-to-ene ec parisen be:veen the :: del and the tes: data is difficult. The cede overpredi::s the breakthrcugh heigh: therefere the treakthrcugh elevation cus be user-defined. 1 3 Go.3.2. 3;tD-8
I ~ C" 0 CO ' !E VR / l ',3 -e u u y O s: O 3: - 2* - x-34 - g-.g =~ Ca -E C is - 14 - 14 w 12 to - s-SOC 7 v sn i i e O DATA t O WOOEL e ( 2 8 c-I C g 2 3 TIM E larJ Tigure D6 Feel Surf ace Elevation Versus n e (Run 5506-7) a=- enn e m.. m,.,, 38033 3a0-9 l l l t
I I hl OC $ ' t. W $'} 43 POOL Sva8 ACE VE LOC;T V 5806 Aim TECS o 4: O ss O 2: I O s \\ I b 25 wv( y O m M- ,,L PO SE:6 6 10 , = 5 et O Oa: A b d P.'OOEL t l l l i, - C ^ I I I I t I 16 23 24 28 32 36 40 l E 6E va7.O *. a f o l T18 re $7 ' Poel Surface vela:ity Versus Pse.,.r.3e..-.4 t I ..neh * ~
I F C,0 C f.1 E 'F6 Y i I 32 - 34 i u
== M m= x 1J \\ 78 = s N 25 'S [ \\ 24 = [ \\ l N. I 33 = [ x N E E l 1.0 N 18 E g I 6 16 TW E ba; L i 14 E L ,I 12 ,E E(. IC L 1 570L7 ( l CA*A g 1
=== u :EL i, p.
=
l 1 1 !r 4 i 6 1 C E 2 r k. I i i i i i i i i J '< 0 2 4 4 8 to 12 14 18 18 X IM) L POCL SUM
- ACE PmCFILE I
Tigurel.,; E Peel Surface Eleva:icn Map (Run 5706-7) E I 5 L- ,c l l iC l l 3 6 0 3.9 3 9 b I/ 1
f $ =* C E ' Y 6,Y i n y BUBBLE 8 i8tf ACE Pr.3 FILE u CATA -- w::E; = ,.z. dl 7twE tasci a / g\\ ~/j \\ x \\ m \\ N = l \\ u 1 2e 1.1 l E 1 E{ 18 j 1i t p s i se u I 12 ~ s t it _. r 5: ld / ( / /j lll// / Q // ~ %_/ 3 l I I I I I I I I c 2 4 s a to 12 14 ts is x ml Figured 9 Riibble Surface Elevation Map (Run 5706-7) M ] g.~3,2. 3 2 P R
t ho Chk I$M$f l I .n O[ e so b ~ j k s. c = e. 5 I ac t, 5 5 2 c { j J. e 5 e a O C t 4 z u d T c e E r f l 3 C e ~ 1 ~ w 1 i 0 s ~x I ac i I ~ I c = s f 3 J. / I i $'r n / / / / / o 5 ~ l/ C .x c E I k I 5 //
- i 5
E r s e O A E E i O E = s E E = o c 'I / S' I l I I ec e E, a o 4 o o e o (ONt ky n 1h3 A sa g%3 w3ngats; / (wAosn ce;1gi gb 3. 2 32 D-13
i 1 n i 2:J meritaO s -. L:.....,. n. .z L t.z... u: 1.
- Eilanin,
'a'.J., "The General Ele::ric rk II: Pressure Surpressien Centainnen: Sys ten Analytical M:d el," ND0-20533, June 197.:.. r I
- dy, T.J., "Ans; y:ica: M;de; fer Eiquid Jet Freper:te3 fer predicting 2.
F rc e s er E:;;d Su': c r y; S t ru::ur es," ::E ;-;].7 ;, Septe-ber, 197'. i L 3GO 3 7 3#b-M
360.3.2.33 Gue dm Response 38.33 -o .. r.c.. w......,) , s.* The po:1 impact specification for "small structures" ccres fre PSTF Test No. 5706/4, where i= pact occurred at 21 ft/sec. For this specification te be applicable at the =uch higher i= pact velocities that can occur in a Mark III contain=ent, the citigating effects of poc1 curvature cust be con-sidered. The basic approach cutlined in Respense 3E.4 te First Rcund Ques-c
- tiens,
- .e.,
cce;aring the pr duct ( vnax) (DLF) eb:ained frc: the Mark III spec;::catier v;;r the e n e-frcr the Mark II Acceptan:e Criteria (A:n in u-c- ,. s r ea s c r.a c. e... ty..e impact ve,cc.,:. an; pu,se durations,
- newever, i
... r. w used in cen;ur.c t ien wit h the ".ari II A: ca*.. nc he arr rcr riate. The centerns r e g t. r d ;n;; ev.cci:3 are addressed in Questien I ab:vc. The cencerns ateut aulse r dLratier art a5 fellCWs;
- 4. N r,.
rc.;. 9 . a.. c-..s-lv*6 C .a.C F orf 'fec-t h' - ...a<-. e. :..:
- c. n-r'*-
-a c*+$ + .L. .a. .c.p + - + + r4 c _c e.n. - ,r .c e.. c__-..e ..e .e... z.....a.c.. u.. t ..~.e.e s....:.- <.. ~ c r. .o... e.. 7 he tw ;c g. t) The pulse duratien fer circurferential targets, used in Eespense 3L.4, is taken f rer FE!F zeasurenents where :he side ws11s of the tank re-su :ed in scre circurferentia' Feel curvature that sculd not exist in an actual Mark !!! plant. Consequently, the circurferential pulse duratien, as propcsed in Eespense 3E.a. may be tec large. Therefcre, previde the = edified pulse duratiens that will be used for struc-tures that do not span the entire pecl and/cr are less than 10 feet above the initial pool surface and cuanti:: the pulse duratien for circumferential crientated structures abcve the pec1. 360,3.2. 33-/
REfPONSE 3B.33 The bulk pool impact load specification was derived from a full scale air blow-down test. The response to Round 1 LOCA Licensing Question 3B.4 demonstrated that the margin inherent in this test more than accounted for the fact that the impact velocity was less than the design value. To assure that the bulk pool impact load definition is bounding at increased pool swell velocities under plant conditions, the approach outlined in the response to First Round Question 3B.4 was used. The pulse durations were not modified f rom the basic approach because they are bounding even at an increased velocity. Attachment A proves 6.8 msec is a lower bound on the L= pulse duration expected for a Mark III radial structure. Figures 3B. 33-1 and 3B.33-2 show the conservatism in the GESSAR methodology relative to the modified Mark II metho-dology for dif ferent impact velocities. This result shows that the GESSAR methodology is conservative f or all beams less than 13" in width and all pipes with a diameter less than 18" even if the impact velocity was 60 ft/sec. The same comparison was performed for circumf erential beams and pipes using an i impulse duration of 2 msec. Attachment B explains why this duration measured in the PSTF is applicable to a Mark III circumf erential beam. Since the circum-ferential target was so short (3.5ft) sad the pool was so flat, this 2 msec l l impulse duration is expected to be bounding. Figures 3B.33-3 and 3B.33-4 show the conservatism in the GESSAR impact method f or circumf erential beams and pipes. Additionally, the data points on Figures 3B.33-1 through -4 show that all structures subj ect to pool swell impact in the 238 Standard Plant are sized such that the GESSAR impact methodology produces conservative levels of stress, even if the pool swell velocity is 60 ft/sec. In conclusion, the GESSAR impact methodology is conservative for the 238 Standard Plant radial and circumferential beams and pipes up to an impact velocity of 60 f t/sec. 3 60.3.2.33-2s
38.33-/ FIGURE ^ PLOT SHOWING WHERE GESSAR
g------
IMPACT METHODOLOGY IS CONSERVATIVE
- ~
RELATIVE TO MODIFIED MK II METHODOLOGY _ -. _ _. i FOR RADIAL BEAMS t = L _ _. + t 3 i ..__7 l -g -l- ____t...______.__.__.__._- 3 .J. f / 3 ,l p. _ __a..-__ _. I e i i E_ t 1 l l 6 I RADIAL S E A MS g E i_ I -t. 3 t =.o062 se<. 2 l g ep39#</ fir,sTw*bfe3 SMn y _ _1.--3-_-- I 70 l ( 1 3 i i ? i 3 --g E O Q i ' b 8 O I m N 1% l I;P p-e gg. { _.o -__3 GESSAR NON-CONSERVATIVE O,, g f FOR LARGER BEAMS g yo _.____ __. _ _ _ _ _ _ _ _ _. _ _ _ _ _. { _. _. _ O I ( a._._._.__ _.__m__ J__ k i GESSAR CONSERVATIVE ~-'j~ ~ - - , FOR SMALLER BEAMS g g 1, 4..-._-_.-_.-._ I .g _ j I g l _ __. __ i i -. _.. _m E I i j 2,. .__ J I-i I ,i ,t t i t l i ._l .J _ _ ____ __ 4 1. _ _.._i _ 4 i g i f I I l l-to - ---/c (6~ ao 25 of 6ascu) 360 3 2.EO O O #3 M
38.33-2 -- FIGURE 4 PLOT SHOWING LIERE GESSAR IMPACT METHODOLOGY IS CONSERVATIVE RELATIVE TO MODIFIED MK II METHODOLOGY FOR RADIAL PIPES -'~ t yao. g.__ ___, _ __ _,._. I I g, f I _4_ __. _ _ _ 1__. j RADrA L !Pt PE S-5oe. 3 i 4 I t s.00 GB GEC.,_5 bib.. _ -/; -. .[ 2. ) L _ - _.-_. (9s 2 3pf.aJ,./ Ph.d J 400 C. I . g. i E I 1 1 _ j __ _ _ _. _ _ _ po_ __.4 l 1 5 i t. - l4 4 10 O r i [ p._____.__. __~_. j + l' 3 1 1 ..{. - l. t i I. } GESSAR NON-4 CONSERVATIVE } loe FOR LARGER PIPES y k,o..__...... j s k o ; -- - - - - --~ ~' ~ ~~ ~~~ 7 GESSAR CONSERVATIVE FOR SMALLER PIPES ,____.___I I i go - __ = _. ._-a_.-___ 1 i g. }_ l ji i-l l i l soi t I i q.;._. i l j i g 4 i r 5 I l I t I ! i i i I 1____..__.. 5-so or to sr 30 P lPE 08AMETER 6^>) 3 60,3 1.3 3-V
g. ._-y,33_3 -~ FIGURE 4 PLOT SHOWING WHERE GESSAR IMPACT METHODOLOGY IS CONSERVATIVE '^ ~ ~~ RELATIVE TO MODIFIED MK II METHODOLOGY ~ ' ~ 7,.. 9 __ FOR CIRCUMFERENTIAL BEAMS l -- -- -- - I , __ i _ i_ _ _ _ _. lg l 8 Foo : 1 -f . _. _... f,i f i t m g._ f \\ \\ i \\\\ s N 330 -- -- -- - - - - \\ - \\ 5 \\ t :,002 sec. fI* ni N'** **' ( ~' \\ \\ Z5's54..,dno l ~ \\ g o,. -.. G s N K\\ GESSAR NON-CONSERVATIVE i s N FOR LARGER BFEIS .\\ m O N s a N 's N s k soo N g \\ . N g.. xx k .g ....._.\\ _ _.. \\. 7, s g.. . - __\\. _ - - _ -. _ \\ \\ \\~'-~~~~~~V ~ - - ~ ~ ~ ~ ~ GESSAR CONSERVATIVE ~ \\'~~~~~ FOR SMALLER BEMIS 'I 3 a. -\\. _- -- -: t \\ - -. -. - -- \\ [ i i . 's, ~1 ~ \\ 4.-------- in _-._q-. .p ID + i o g+2 a . -.- 1.-- _ j - - - - y _4 l I l i p_. _ _. ..i l t j I i i ,o l-am wiers (~ca) ~360. 3 2. 3 3-5 _. _ _ ~. _,.. _.. _ _. _ _
- a..
y ____2_.___. FIGURE A PLOT SHOWING WHERE GESSAR IMPACT MET 110DOLOGY IS CONSERVATIVE - ~ ~ ~ ~ ~ RELATIVE TO MODIFIED MK II METHODOLOGY I ~ ^ y______ ______ FOR CIRCUMFERENTIAL PIPES go p g l { l----- ._u_ [l-y ._ _ _ {f C s ac.u reRENT Ai. PIPES "\\_ _i_-._ -.- y_ l T: qOC'A SK g i_ g\\3 i i \\h g o... - - -. _.._ s \\ \\ N ~ \\\\ --- \\ -- - ---i-------~-------~-~'~--~ ~ ~ ^ 200"- GESSAR NON-CONSERVATIVE p \\ FOR LARGER PIPES 7 4
- \\
\\ 4 \\ N N O + a s,. v-k (O _ __ %j g N e .)y .- g ._p Y \\ to \\ \\' \\ i .g \\ i y _. -.. _._- ---.-.._.-..----. g - --\\. i \\ \\ I i \\ gn -g \\. i\\ t 5\\ 5 I .J_ }_. - _..-. -. I GESSAR CONSERVATIVE j j FOR SMALLER PIPES ) \\ l t i l, l __.l
- i....
i i 1 i i i i i i i l i Jo l 8 1 l l I i - L - _.1 - _ _. - l - -_ - I L L._.. -. _.-.._-- ___ _.. _ J _. 1 l t I l 1 l i i i i i I I _ -_ k - _ _. -_ _ f, . l.... 10 so ao Ao Sb to WE NAM cTex (in) 3 /30. 3 2. '53-b
Attachment A, Response 3B.33 Justification of 6.8 msec as the minimum impulse duration expected for a radially oriented Mark III structure. The 6.8 msec used for radial beams was the minimum impulse duration seen in a radial structure in the series 5805 testing. The test producing this duration had a submergence larger than prototypical and produced less than 2 inches of variation in the pool surface height. Figure A-1 demonstrates that the 1/3 area scale pool surface is almost flat while a Ir5 Modified Froude scale inter-pretation, which gives a prototypical surface shape, shows significant surface curvature. Figure A-2 shows that a full scale PSTF blowdown also produces significant cu rva tu re. Thus, Figures A-1 and A-2 show that even short structures or structures below the 10 foot elevation will see more curvature than this mini-mum value. 7'o prove that the short beams located near the pool surface have an impulse duration of greater than 6.8 msec, the slope of the PSTF pool surface was calculated using 5 foot submergence tests. The 5 foot submergence tests were used since they may be interpreted as 5/7.5 linear scale Modified Froude tests with prototypical Mark III surf ace curvature. The lowest elevation that structures exist in the 238 Standard Plant is 6 feet above the initial pool surface. The smallest sttucture in the pool is %4 feet long and attached to the drywell wall. Thus, the slope at the drywell wall at a scaled elevation of 6 feet is desired. Using PSTF test data, the nearest to the drywell wall in which the pool surface slope may be calculated is between 3 and 5 feet from the drywell wall. This corresponds to a scaled average distance of 6 feet from the drywell wall. Figures A-1 and A-2 show that the slope closer to the wall will be somewhat larger, thus this is a conservative analysis. The minimum slope was Cd1Culated 4 feet from the drywell wall (scaled distance of 6 feet) and from 3 to 5 f eet above the pool surf ace (scaled elevation cf 4.5 to 7.5 feet) in each l 3 (3 o.1. 2. M - 7 t
of the 5 feet submergence 5801 steam tests. The average of these slopes was calculated to be 0.16 in/in.* In order to determine what slope is needed in the 238 Standard Plar.t to assure that the impulse duration is greater than 6.8 msec, the velocity vs elevation curve shown in Figure A-3 was utilized. This curve was developed by bounding the velocity elevation profiles of 144% DBA break area 1/3 area scale air and steam tests, large break area Modified Froude scaled air and steam tests, and a full scale air test. Figure A-3 shows that if the peak pool swell velocity is 60 ft/sec, then the maximum velocity 6 feet above the initial pool surf ace is approximately 40 f t/sec. This is a conservative approach since the velocity closer to the wall at a given elevation (where the beam is located) is always less than the peak pool velocity at that elevation. Since the impulse duration will be equal to the surface slope times the structure length divided by the pool surface velocity, it can be shown that the slope must be greater than 0.068 in/in for the duration to be greater than 6.8 msec (slope = 1r V (.0068) (40) 0.068). Since the calculated slope was shown to be --- = = 4 0.16 in/in, 6.8 msec is assured of being a minimum duration for all radial structures in the 238 Standard Plant, even includinr, beams near the pool surface which do not span the entire pool. l
- NOTE - A concern was raised in the September 15, 1981 GE-NRC meeting that the PSTF Test 5801-9 pool surface seemed flatter than the Figure 3B.33-1 pool surface. Actually, both of these teats produced similar pool surf ace curvatures and the minimum slope seen for 5801-9 was 0.24 in/in.
3 @,01 1 D
GENERAL ELECITtIC COMPA.W P1LOPRIETARY 'g 20 0.79 eec 1 A/TSROUDE SCALE INTERPRETED \\ TEST SSC11, M
- 8.7 h
\\
==*= 1/5-ARE A SCALE INTERPRET!D TEST tact.16 M = 7.6 M IS 4 ROOF FOR TEST 6801.16 14 ~ 0 es nec = = = = ? g C 66 esc 3, 52 { OWm 2 [ Ic o ( I 0.53 esc 0 4G sec g b* 5 0 0 40 sec ~ 0.36 esc w d C.2s sec
== == A t
- C.2 7 eac At = 0.19 see w
2 INITI A L LEVE L C ~ I I l l C 2 4 4 8 10 POOL R ADIAL LOCATION (M) ( Figure N ~/ Co=parison of Pool Swells for 1/3-Area Scale Test 5801-16 and 1/ 5-Froude Scale Interpreted Test 5801-1 360.3.I 33d
O s ( t ~ I i a- / i r Ja j 1 - t q / / j't 4 I l 4 1 e / s / 3 -D / / / ,I n Yl e 3 5 y -r i
- g s s
/ j-3 / 3 ~ 0 / d l0 k / i i 5 / / -? I i W / / / E / / i~ ( 0 4m / / , j. i L^ ~ x =b \\[) / l t N l r l 2 ~. t I i i 4 ? i x g*- Q f % O i N A 9 k ft\\ D /N q 4 4; \\ l ~ ~ ~ \\. \\ is l N t k $n \\ \\ y l t o u I [ Cf_ t 1 ,i e o a ~ 4 o 7 d 1 w ~ t V:9
- n
}. &;V c4 4
- Q
( d o cu 'S _jn * ' ; c', s', {t I l 3 6o.32 33-/0
FPOPP IE TM rsy FicuRE A-3 VELOCITY vs ELEVATION DESIGN CURVE 1.0 r A .9 A o o V 3 8 gO li( .7 9y (n k 6 o3 A 5 FT SUBMERGENCE AIR TEST (1/3 SCALE) gh o 5 FT SUBMERGENCE STFMI TEST (1/3 SCALE) .5 0 0 0 7.5 FT SUEMERGENCE AIR TEST (1/3 SCALF) rd U V O 7.5 FT SUBf1LRGENCE STEAM TEST (1/3 SCALE) yy o/ a Y 8 FT SUB?tERGENCE FULL SCALE TEST A W tj 4 o L> s .3 o Y .2 .1 0 0 1 2 3 4 5 6 7 8 9 10 11 12 SCALET) ELEVATION (ft)
w
- NMh1f b hQMV S'N APPLICA51LITY OF CIRCO:FERF2;TIAL BEx:
I,,: n. -,. Dn... .s n 20 n. ..n.:. s2 During PSTT test series SE05, leads en circurferential targets have been repcrted (Reference 1) to be much smaller en the target ends (near the pocl side wa'ls) than in the center. The questien of whether the side walls, nct present in a prc:ctype plant, effect the data is a natural ene. In this res; nse 1 w;12 he sh:w. that ne effect cf the side walls is neglicible, sc that an. circumf e:en:ial curvature cf the ec1 surf ace seen in the PS!? will als; te present in the prc:ctype.
- c evidence cf circumferential cur-va:ure can be found in the test data, however.
In fact, evidence of the p:;1 bein; flat circurferen:ially will be presented. Finally, physical argu-ments en the cause cf Icwered impact pressures en the end of the peel swell .a. =..s .... v e. ,, c. t.. e >.. .3 The F5!? suppress::r p :1 baffels censtrain the suppressi:n pt:1 :: =c:k up a single E' cell cf the p:::::ype. The side walls consecuently exist where n: walls exist in a plant. The PSTF side walls transfer a zere shear or plane cf sy.netry cin: tc a zerc veleci:3 plane. The effect cf this change r 1 can be quantified by takin~ an assessment of the Scundary layer thickness en e these walls. 'a*hile the ficw is certainly turbulent, an order-ef-=agnitude esti ate cf the beundar. layer thickness may be made b"J calculating the boun-j dary layer due to an impulsively started latinar flow. Ref erence 2 gives i the thickness cf the beundary layer f or an 1:eulsively started latinar flow to be (1) e - 1 1 i l l 3 go. 3.2. m/L J
p - Continued m.s m e 6 where bounda _nickness i = kinema xosity = time - = L' sing data f er water 7C*F and a swell time cf I see ...t - u. t i. r t. x 2. (,2 sec)1- .013 ft = 0.15 inches
- = -
= se, Ccnsequently, the thickness cf the boundary layer en the PETF walls is at = cst an incP er tw:. Any flew parancters tere than an in h cr tw: frc the walls will te unaffected b: the walls. Alse any surfa:e curvature of the PSTF po:1 will als: be present in the prc ctype. Wi:t the above results in hand, a search fcr evidence of circunferential pec1 surface curvature in the ?ETF 1/3 area scale peel was perferred. Test data indicates the pocl to be flat, no: curved in the circunf erential direction. Figure 1 shows the pressure time histories frer the a pressure transducers en a 10 inch circunferential I-tear used as the target during test 5E05 run 2E. Figure 2 shcws the ceasurement locaticns. k*hile the impulse seen on the east and west end transducers is orders of tagnitude less than that seen in the center, a small presst.r2 spike is seen on the east and and a rise to l the drag lead is seen on the west end. All of these transducers shew scte l response within less than twc tilliseconds of each other. This duraticn alcng with the =easured peel surface veloci:3 of 29 ft per second, indicates i a circumf erential curvature of less than 3/- cf an inch. I ~ S603.2 33-/3 l
/117adnw/ 8 - Continued As described in Reference 1, the reducticn in inpulse en the ends of icpact targets was seen in both the circunf erential and radial directions. This is due to the flew patterns which exist during poc'. swell and the fact that the surface velocity of the peel is not necessarily the vector velocity of the As the bubble penetrates the pool, water runs off the water slug sc water. that while thc surface cf the pcci is rising, the actual water vcicci:3 is radially and circurf ercntia;1y cutward near the ;crl beundaries. Since the =ccentut vecter cf this water is sideways and not upward, no inpact results. Since a;; evidence is that the 1/3 scale P3!T gives pec; surfaces flatter than prc:ctypical, this effect will be less in the P5!F than during poci swell in a prc ::ype Mark III. ggo. 3 2. 33-N
f,$40'him,nd, b E==.'.=='.'"..=.R i... 1. T.R. mci 6 tyre, et al "Mari III Confirmatery Test Prograr - One-Third Scale Pool Swell I= pact Tests" NEDI-13126?, August 1975 2. Eird, Stewart, 6 ' ightf ect, Trans;::: Phen: ens, p 1;;. Jchn k'iley 6 5:ns, i;ew York, '9:: 380.3.2.33-/5
P Ro PR)t.TARy le (, t 3 ~
- W**Ml '} ~^^^e%wy %vmspWA4', <
y h6 C N l
- a. w a.
I 4 i sb l t t t t *. ' foe - Psi y u b a ge, . ~ - ~,../v w -+ m - m % ~ C tfi i 2 w V 1 b
- r..
M MhM M e**%'k# =/ E o 5 O (WQgf.=,Wt#4 / C2 ~ e 3 R f f f ! ! ' f f 10 0 W G s G& "% ^"MW# N "f O n%=. r- , ::, + :,, v ^ "' ' rY .:b i j O f lcc ogo Sc0 7~,m e. milliseced.s &r s sarea to inc k Grc ~hau d F";goce f-/ X m pae+ on I-bea m in Tect 530 f (2 an 28 360. 3 2 3 3~/h
PPoFFIETAPy Cines III E AST COLUVN ACCELE A ATICN CENTER VCVENT STRAIN GAGES f CENTER ACCELEG ATION -c / WEST COLUvN ACC E L E R A TION e '- / EASTIMPACT / PRESSUAE (NOT USED FCR f 5m.8EAVI / / p /,- CENTER SCUTH IMPACT / / PR E SSU R E / / CENTER NCATH IVPACT PRESSURE -ed s WEST IMP ACT P AESSUA E (NOT USED FOA 5-an. BE Av) I i 1 l Figure [ cd Targe Instru=entation - CirCu=ferential I-Beam Targe:s (10 in. (254t=) Shown, 5 in. (12 7:=.) Similar) 1 l 'h '3.2.33-77
i 1 3BO.3.2.34 Question / Response 3B.34 QUESTION 3B.34 l With regard to froth impact, Response 3B.6 to First Round Questions cites selected runs from Test Series 5801 as confirmatory evidence that the 15 psid froth impact specification is bounding. Other test runs from the same f acility, which also involve froth impact on the PSTF roof, exhibited impact pressures considerably greater than 15 psid (i.e., Tests 5802/1, 2 and 5806/5, 7, 12). 1 Justify the current f roth impact load in light of these other test runs, and the increases in the pool swell velocity (up to 60 ft/sec; refer to question 1), or revise your design value accordingly. 4 I 4 RESPONSE 3B.34 l i Only selected runs from Test Series 5801 were presented in Response 3B.6 to the First Round Questions because these are the only runs that are prototy-pical of a Mark III Containment. Two controlling parameters in froth impact i are the break size and the distance between the slug when it begins to break up and the HCU floor. The 5 foot submergence runs of Series 5801 had proto-typical velocities and slug thickness at breakthrough, prototypical distances from breakthrough to the PSTF roof and near prototypical break sizes. Tests run with a 7.5 foot submergence;such as 5806-12 produced liquid ceiling impacts j due to the lower than prototypical PSTF ceiling, and are thus not appropriate. e In order to provide further evidence that the 15 psid froth impact load specifi-cation is conservative, all air, steam and liquid blowdown,5 and 6 foot sub-mergence PSTF 1/3 area scale tests were investigated. Three of those tests (Runs 5806-5, 7, 10) had blowdown orifice sizes more than twice the scaled Mark III main steamline area (> 200% DBA break area). The large blowdown sizes 3BO.3.2.34-1 1
RESPONSE 3B.34 - Continued i produce very non-prototypical roof impact transients because a larger slug of water is thrown up, causing a larger impact density and more liquid flowing through the roof. Figure 3B.34-1 shows that the 5 foot submergence data con-firm that the impact density increases with break size. Figure 3B.34-2 shows that the larger densities cause greater impact pressures for the larger break area tests, Therefore, those tests with break sizes greater than 200% DBA break area are expected to have impact pressures more than twice the proto-typical impact pressures. Since these tests fall so far out of the Mark III i range, the resulting pressures will not be used in the data correlations to be performed. Additionally, tests 5802-1 and 2 had breakthrough much closer than prototypical to the PSTF roof and are thus also not appropriate. Therefore, all tests have been considered. 4 i In order to obtain a measure of the froth impact load on the HCU floor with the requested 60 ft/sec impact velocity evaluation, the maximum lift pressure was calculated for each test. Test series 5806 had several more sensors than the earlier steam and liquid tests. In order to have a_ consistent set of data, the same pressure transducers were used in both the air and steam tests for this calculation. The measured local pressure (Reference 3B.34-1, 2) from the like sensors (at locations of 1, 2, 3 and 4 feet from the drywell wall) were i time averaged to obtain the net lift pressure time history. Since the 5806 l l data were recorded on an analog system, approximately 80 times as.much data were obtained as were recorded in the previous 5801-5803 series. In order to l analyze the data on a more comparable basis, the 5806 data were averaged in 1 i millisecond blocks. Then, the background pressure readings were subtracted I obtaining the peak net lif t pressure. Since the peak pressure is proportional j to impulse over velocity and Figure 3B.34-3 shows that froth impulse duration 1 { 4 3BO.3.2.34-2 1 3 .~v.v.,,w-... c. r_~.m..2 .m.- c-v,.-v--...e,,,,-,.,.v,w,.~, m,.r._%.,,-,s. --n.%m,.,.,m.. ..-..-,.wm.,_,.-w =, - ce -,w_,--,,
RESPONSE 3B.34 - Continued does not correlate with velocity for the 5 foot submergence steam blowdown data, the peak net lift pressures were linearly increased to 60 ft/sec. Figure 3B.34-3 shows all of the 5 and 6 foot submergence air, steam and liquid blowdown data with break sizes less than 200% DBA break area. This figure also shows that there are 2 six foot submergence tests (5502-1, 2) which had much higher im-pact pressures. These tests are quite non-prototypical because breakthrough occurred very near the roof. Test 5802-2 had breakthrough right at the roof and, due to the similar impact pressures and densities recorded from test 5802-1, it 1 may be inferred that this test also had breakthrough very close to the roof. Attachment A shcas that the breakthrough process in the Mark III is enpected to begin by 13' above the initial pool surface. Therefore, the expected distance between the HCU floor and the breakthrough initiation location is % 7.5 feet for the 238 Standard Plant. In conclusion, all applicable PSTF steam, air and liquid blowdown data were used to show that the maximum net lift pressure, when scaled up to a velocity of 60 ft/sec, is 8.6 psid. This confirms that the 15 psid f roth impact load definition has a high level of conservatism. 4 l 4 I 1 6 i 3BO.3.2.34-3
Ar R A K g A y E 5 R 7
- /,
j B 5 h g 7 H T I W )w n o Y o T rw t o s l yrB E Y D / R M A T A / T C T7 A 5 0 E P n h 0 r t / M / R I P E C O F / R O AE P G A A NR B FR o E D A s M T B A U @e %AE t S n E A t V S 7 00 7 1 4 O .S / 3 C b 3 R OF s r R U G /F / / / i Q 5 0 0 f 2 1 T 8 C A 5 F m 4 H i f b t A (
\\,, AE R A K A E R 2 B / H, 6 T, sW,, 7 E / R y A V O g E R S R 5,o A 6 E, 7 R-g E P / n / J K - g /,O R T A P C O P o E y e 6 A d R / R Mg S P Ig y rS o -g A Mc B oy ',D e D s g T g a A I g j 9 R g A n V avs g 2 43 r /,/ . a 6o / r 3 S E#uRo G ~F /F - o 7 6 i 3 2 O Y / A ) ) o c M D nT es I mo/ r S r e v P f t u ( O ~ a o c L (< s
A t ij k k a s v o l o t N W I w' d i s 1 i 4" oR a t t 2 1 C A Sa e 2 o 8 m e w,a l' v w s st 9 x r= k
- E e
ee 6 <e k ct t 't $ o. t Qg O O i El i =s a L g-s V i l A v O \\ ? 9 1 v A tu n I ,e % a L, + Oe i R 2 e r A A X e a e 3 ~ ~ ~ e O .b d o e N -Jb% { s er y 3S
/'" PROPRIETARY f/G UA C 38 34-4 PEA K P3 **f A 00 F L /F 7" PtfSSURf LINEARLY 5CALED TO GO Ft/ set i w ere : TH t. S E 44C NoN-PROTOT m CAL Duf To te ra c rua.a s s N rA n moor Q* M I l D l i A C kTY k O rios Trsrs rrr svax.,str4x y g D S102 rests G fr soon., 3regp y A Sto3 2 Trst fr7 Suen, Liq wo O fiot 7f373 fy 4 ff sudM., A/R 9 3 ( 3 M 4 Y 6 t 1" =_0 's ' w or l 5 k a i /0 W i i( i Q N l ( r l 4 v-c a p .a i O I 4 e O $~ /0 !I i 4 (HCO -87) DISTA A/CE ter )
A-Hacke,A A: Determination Of The Location Above The Pool In A Mark III Plant When The Breakthrough Process Begins. The breakthrough process in a Mark III begins to occur at 13 ft above the initial pool surface (1.73 x initial top vent submergence). The GESSAR design envelope shown in Figure 1A (18 foot breakthrough elevation) was proposed by the NRC and accepted as a way to assure that structures whi:h see 3 high density froth in the range of 12-18 feet, are designed to a liquid impact s pec if ic a t ion. It was never meant to imply that the impact pressure would not decrease until IS' above the initial pool surface. In fact, the loading envelope which GE originally proposed to the NRC (Ref. 1), which is shown in Figure 2A shows the impact load starts decreasing at 12 feet. This decrease in load occurs before the bubble has completely broken through the pool sur-face. Figure 3A from Reference 2 shows that for a 10" beam, this decrease in loading occurs when the slug is about 2 ft thick. One third area scale data (Ref. 3 4) may be uced to decide where the slug first starts to break up. 3 These data shows that breakthrough elevation is independent of velocity. Fig-ure 4A shows that in the 1/3 area scale PSTF, water slugs are not seen above, and breakthrough occurs at 1.8 x top vent submergence (13.5 ft for 7.5 sub-mergence). When the slug reaches a certain thickness, Taylor instabilit ies dominate and the surf ace begins to be broken up, decreasing the density and impact loading. This process will occur at an even lower elevation in a full scale plant because:
- 1) Figure SA shows that the pool surface will be curved and much more peaked in an actual plant. The dominant mechanism for thinning the s.'ug is water draining of f the top, thus, when there is more curva-ture there will be a greater gravity head, which will decrease the slug thickness faster.
- 2) Figure 4A shows hat the 5 foot submergence tests, which may be inter-preted as Modified Froude scaled " full scale" tests with near prototypical submergences, have breakthrough before the deeper submergence runs.
3 80. 3 1.3Y ' 9
- 3) As presented in the response to the Round 1 NRC LOCA Licensin:;
questions, the full scale PSTF showed breakthrough lower than the 1/3 area scale PSTF (less titan 12 f t), on v
- 4) The MIT tests also show the eff ects of scale breakthrough as small scale pool swell tests showed later breakthrough than in the larger scale tests.
Since there is no evidence of a water slug above 13.5-15 fect (1.8-2.0 x initial submergence) for a 7.5' submergence 1/3 area scale test, and it is known that breakthrough will occur at a lower elevation in a full scale test, 13 feet above th' initial pool surf ace (1.73 x top vent submergence) is a conservative location for the initiation of breakthrough in a Mark III. 3 60. 3. 2. 3 Y-7
REFERENCES
- 1. General Electric Company, Bk'R/6-238 Standard Safety Analysis Report (CESSAR), Docket No. STN 50-447, November 7, 1975.
- 2. T.R. McIntyre, et al., " Mark III Confirmatory Test Program; One-Third Scale Pool Swell Impact Tests, Test Series 5805",
General Electric Company, August 1975 (NEDE-13426P)
- 3. R.J. Ernst, et al., " Mark III Confirmatories Test Program; One-Third Scale Three-Vent Tests (Test Series 5801 through 5804)", General Electric Company, May 1975 (NEDM-13407P)
- 4. T.R. McIntyre, et al., " Mark III Confirmatory Test Program; One-Third Scale Three-Vent Air Tests, Test Series 5806",
General Electric Company, October 1975 (NEDE-13435P) 3 60 3 ;t 3 9~'O
FIGURE LA POO L S LU E LL E MPACT LOA b l N G SPEclFIC ATION FOR SMALL S rR uc ro R ES G E SSA R DE sigil Puu l\\', i i I I I l u i I e I o PEAK I ACTU A L l M Lo A DING l \\ e Co (DI) l l i s l l\\ I 1 13 18 19 H.E/G H 7~ fgan poot S u e ; A c c-(FT) I
jl I 'I ll Jl II l' S 8, D A 5 t P E Nu ~ t a J 3 I 1 n Mt 8 5 i 4 Dt A6 = 5 f Mt w l AS V0 i A e f. G OI 1 5 G 6 NI t t a I L f 0i 1 N P Sf F l 9 8 u f DO c. 3 N N OPI OM .OA f N O 5 T T tt 6 f 5O $f A t 6 S 7 n P8 0 I f f l O J N P A o t N i S I S 8 1 O A L t f l f 8 J f Nl O l e 6 S P t P A 2 P i ( P t N6 i Ap A& 2 A O OAl O0 a MC NNt U. l f ) t 4 l OAIf t f 1 I. E f A L t 6 0 f i A1 C l f O i CS A l. n l Af 7 fa f 0 3 N O A u H U I $ A t N U O M G Vs W *. I T U Nt O D0 I t 8 / f t 4 D AJ He 9 NI i t uI GN,w H u. 4 9 ft fA 5 5fli Dl i 8 e i f u I v V 8 O9 o t t l 8 8 t 4 4 5 c t i MS 9 S ( 1 V5 0 fl I 2 1 4. i 9 9 5 n a .~ 51 n ,I 1 I l I N c, u a n,. a s r o, i ss r ,o jl l I l I fe oo 1 r * ( C . l l a, A t f . C i o t tf 5 . o r r. L . t ) ., e N a m T s ro P M ,l o O m t p f e i r l s e l S i e. G I antr ( i s t o st a tr r> d r osu r r.s e g 0i . m A l. .N t o n o 4 , i s( i A. 4l. ] e _S s i e \\ e 0 9 s {' 8 g w O 5 o2 ~ 8[8- [w*
- 4
>V v b @Q.&
- WSLP
I COMPANY PRIVATE Class III i EL 4 E 1 i I O i n i. w N,stewciaccustat%riat I d ,,,, 4 / CL 3 l / / = i / / .x l 12 l z i / 5 O \\\\ J a e / oi l 4 'N i i y c ^ 1 e p o i 2 3 4 s a i e,.esT sLuc Twiewtss n' e t Oab Impa:t Slug Thickness j Figure I j (P rop rie t ary) L 380. 3. ;t. 34 -13 1
4 s V2. CQKFA.NT F10PR1rDAT Class III I i I l ise' aI I / i ! N'! j ~ !]j I l I i \\ l i c i I 2 y a ro I l c 5 > 1 1, j }t I f y ! tI
- 5
= a i 3- : Ihn i U i / 1 //
== = s .s / a s { al } / / I / 1 / 4 = / 3 / 4 / / } / s 2 l / l / \\ I I h I I l t 0 tChiettu peta Av g3h30s MMDHa&>O mm 9 6 I 160 3.2. W'N l i
'%g IN " gg 1 A/T3 m 3UCE SC.A LE INTE RM ETE D \\ TE s1 58C1 1. M = L' #4 1/3.A.R E.A SCA LE INTE R M E1 E D - *== is Tust bact it M. 7 s n < Roosson Test tact.ts o se = 2 nse = 12 a J( g tu = s ~ f nu = 4> a *e = 2 8 U 2 5 8 -~ o ao = o_2s = ~ tan - A v. cL27==c a t. o 19 est 1. 3 ~ IMITIAL MvtL f f I I I e 2 4 e e to POOL RADt AL LOCATION (M Figure M Coi::parison of Fool Svells for 1/3-Area Scale Test 5801-16 and 1/ T3-Froude Scale Interpreted Test 5801-1 3O'?1.2.34-/T i t I
a---_-.-- d. \\- 1 l l 3B0.2.2.3. ?dI Quest ion / Response 3B.35 l 1 QUESTION 33.35 t l } The st ead; state f roth drag load (11 psid) for the HCU floor needs further i jus. stion. The analytical model that is used to predict the Mark III wetwell pressurization (NEDO-20533-1) is phys'cally unrealistic on several counts, e.g., one-dimensional flow, quasi-steady, froth density determined { by mixing all wetwell air with water above top vent at t= 0. The codel was j " verified" by co=parison with ceasurements; however, the test conditions were 1 L l quite nonprototypical in ter=s of pool velocity, vent submergence and pool-to- [ t 4 I i ceiling distance. t i i Show that the current drag load specification is adequate under prototypical I Mark III conditions, and in particular the increased pool swell velocity that l F i the staff feels is appropriate (60 ft/sec). RESPONSE 3B.35 f t The analytical codel that is used to predict the Mark III wetwell pressuriza-tion (Ref 3B.35-1) is indeed a simplified approach to a complex probicm. To account for any non-con:,ervatists which may arise f ro this simplist ic I approach, a very conservative density and loss f actor is used. This model predicts a 238 Standard Plant HCU floor LP of 7.35 psid. This assures the i 11 psid design specification for two phase pressure drop through the HCU floor e is more than adequate for design. f l 3 60.3.? S S- / l e -w
- ~ =. -_, ___ - - To illustrate the conservatism in the HCU floor AP predictions using the l Rcference 3B.35-1 methodolog, design basis predictions of all PSTF 1/3 area scale 5 and 6 foot submergence 1e sts were performed. These predictions utilized the design loss coefficient (K = 5.0), the actual PSTF geometry, and a density calculated according to the Reference 3B.35-1 methodology.* Thus, thesc predictions use the same methodology that is used for Mark III design i calculatJons. The results, shown in Fig 3B.35-1, indicate that the model has an i average c,nservatism of 3.25 psid. This conservatism is due to the f act that the i =casured PSTF densities and loss f actors were substantially less than the design values. If an actual plant had a smaller HCU floor loss factor or froth density, its level of conservatism would be even larger. The 5 and 6 foot submergence PCTF data are c::pected to provide prototypical Mark III HCU floor LP values. This is bec ause the nonconservatism of t he vent submergence is balanced by the conservatism in the pool to ceil.ng distance. The distance between the PSTF roof and the breakthrougF elevation (1 to 11 feet) bound the expected distance between the Mark III.lCU floor and the expected breakthrough location. Additionally, the peak.est velocities j ranged up to 39 ft/sec, providing near prototypical values. The 7.5 foot submer-l gence PSTF data are nonprototypical since the clearance from the initial pool sur-I f ace to the roof is smaller than prototypical causi.ig liquid roof impacts. An absolute interpretation of the PSTF data nay also be performed. Figure i l
- 33. 3 5-1 shows that PSTF data varies from 1.0 to 6.8 psid (Reference 3B.35-2).
1 j These tests covered a wide range of break sizes (72% to 144: DBA break area) and a range of roof open area fractions of 19 to 37%. The test which produced a 6.8 psid (Test 5802-2) had a breakthrough at the roof which is non-prototypical. Neglecting this non-prototypical test, the data range from 1.0 to 5.3 psid. i {
- NOTE:
The density is calculated by uniformly distributing all of the suppres-sion pool water above the bottom of the top vent in the volun e between the top vent and the llCU floor. Test data show this is a very conser-vative approach. 3 g y 7, g_g
d l L The load definition was originally developed by using an even more conservative l application of the analytical model. This was done by artificially modifying i several of the inputs to the model to obtain a highly conservative vent flow j rate and wetwell pressurization when compared to the test data. When the model is applied using this method for the 238 Standard Plant, the r esulting pressure drop is 10.77 psid. Figure 3B.35-2, shows that the present application of the model does an adequate job in predicting the vent flowrate for a typical test and it always bounds the =casured pressure drop. Therefore, it is felt that the previous application of the analytical codel is excessively conservative and the 3.25 psid level of conservatis: associated with the present application of the codel is adequate for design. l 1 In sunnary, the model for froth drag at the HCU floor predicts an HCU floor pressure drop of 7.35 psid for the 23S Standard Plant. This model was shown to have an average 3.25 psid Icvel of conservatism relative to all the applicabic data. Additionally, the appropriate PSTF data range from 1.0 to 5.3 psid. This l information confires that the 11 psid HCU floor design pressure drop has a high level of conservatism. 1 l l f l 1 i l I 3BO.3.2.35-3 i 4 -.----,-,-,.-.m..,---.,-,,-. . - -. ~. - ~ -. ~.... -..... -. ~. -... -.
{ StLP nSc 5 b. ?$ ke breWK 1. 'a'J Eilanin, "The General Electric Mark III Containment Analy:ical Model" NEDO-20533 and NEDC-20533 Supple:en: 1 2. RJ Ernst, etal, " Mark III Cenfir atory Test Pregrar - One Third Scale Three Ven: Tests" NEDI-13!.07F June, 1975 (Class III). 'k ( n vv3.,-.. ,->.4
PRO PRIET8&'/ l' FIGURE 56.35-/ COMPARISON OF MEASURED AND PREDICTED TWD PHASE HCU FLOOR PRESSURi JROP J O 5 FOOT SUBMERGENCE TESTS A 6 FOOT SUBMERGENCE TESTS 10 / 9 A / / ]s 8 A 7 / i O / a 7 / /[Q E o /,,@' G 6 o O /f E /<cO' 5 oo o O / C C0 / c S / f 5 3 / / 2 / 1 / / / 0 0 1 2 3 4 5 6 7 8 9 10 MEASURED PRESSURE (Psid ) 320 3. 2 35-S~ l I
PROPRIE 71Vy FIGURE 3B.35-2 TYPICAL COP. PARIS 0!1 0F P.0 DEL AND TEST VENT FLOWRATES (TEST 5801-5) 1 j I'STF Data O Ana l yt. f cal Model ; Ai r + Steam Mass Flowrate l \\ c8h'$ J ., w~ w, c pi M + . O T,A =- t 6s too - / a3 d o I o l.o 2.o 3.o TIME (sec) }N C
r I l QUESTION 3B.36 i There still re=ains considerable uncertainty with respect to the frequency scaling to " full" scale for condensation oscillation (CO) loads for the following reasons: l
- 1) The "1/9 area" scale data shows a great deal of scatter and is likely i
to be influenced by one or more effects presu= ably not present in either the " full"- or "l/3-scale".ests. f 11) There are only two " full"-scale runs (3 data points) that can be used 1 i to confir the scaling prediction, and these results all lie above the predicted values based on 1/D scaling. j In view of these uncertainties, it see: likely that the actual " full"-scale 1 frequencies might be as much as 50% h ;her than the scaling predictions. Two alternate resolutions to this problem are possible: (a) The present load might be acceptable if it can be shown that the present load specification does not differ significantly from a prediction which accounts for 50% uncertainty in the frequencies. (b) The load specification could be changed so that it l incorporates the possibility of up to 50% error in the fundamental frequency. l l In order to justif y (a), it will be necessary to compare the present load l specification to an appropriate alternative which accounts for the uncertainty in f requency scaling, and demonstrate that the dif f erences are of no consequence. l l In light of the above discussion, provide the appropriate justification f or the t current f requency specification for condensation oscillation loads. J i i
RESPOSSE 3B.36 Figure 3B.36-1 shows the scatter in the C0 frequency data reported in round I one question 3B.10. The frequency of the four harmonic C0 f arcing function was increased by 50 per-cent to demonstrate that the CO load definition bounds the CO frequencies ob-served in the full scale 5707 tests. Figure 2 presents an amplified response spectra, ARS, comparison of the CO load definition and the 50% increased fre-quency C0 forcing function. Examination of Figure 3B.36-2 reveals that the 15% peak broadened ARS of the load definition bounds the increased frequency CD forcing function. Therefore, the CO load definition is conservative for design, i sp~----o-e-eW-Tyw yrgwwy7m N r we'w'r---p*4.- iper-w e+e-e m, e -wW -ww
l GESSAR II 22A7000 l 238 NUCLEAR ISLAND Rev. 2 GE COMPANY PROPRIETARY 061581 Class III 2 J e J6% y
- 4 C
0 D T Uu S ~ y ,j A ? @g g o E D 0 to e .OO E c: 5 S e c Q de ~ c g g u T E d d o C u w b E b C d< Q N 5 5 6 5 m a e b d ?
- E l $l 4
e 1 3 U 3 4 ,e dO @ { L 3 n O -d h. 1 I I I o o a e n o ADN3nO38d 03 AW3S90 \\
1 Figure 311.3 6 -2 GESSAR Il C 0 FREQUENCY PARAMETRIC STUDY RUGUST 27. 1981 COMPARISON OF GESSAR b CO LOO DEFINITION 3 AND 70 5(: ALE FREGOENCY + 5 off UNTRTAINTT. 1.0 PERCENT DAMPINCr. 200 160 1 sessne r ~ ~ CAD DEFINITION L J )6 5CALE PEDICTION + S'O % &w 120 2 2w a l. [ -pm 80 au j i ( 9 L a-1 s l_ q f yo m / l jl \\ l I ~ 0 10 o 10 i 10 2 FREQUENCY ( HZ ) l
a 4, QUESTION 3B.37 To determine the CD design load, the "C0 methodology" is used to predict G, l C, and T, each as a function of time, based on an assumed set of initial conditions and an assumed break size; the functions G(t), C (t) and T(t) are then used to determine PPA (t) and F(t) from the correlations obtained from i 1/3-scale testing, scaled up to full-scale. To demonst: the conservatism of this approach, the C0 design load should be comparec (in terms of PPA, RMS and ARS plots) to other 238 standard plant predictions, still using the CO methodology, but based instead on a complete i range of possible initial plant conditions and break sizes. All initial parameters which significantly affect C0 should be identified and the ef fect of each should be considered. As one example, since pool temperature exerts a ;trong influence on CG, the combination of high initial pool temperature and large break size should certainly be considered. I l t RESPONSE 3B.37 I i Vent mass flux, pool temperatare, and drywell air content parameters were i l determined for a spect:am of break sizes for expected suppression pool tem-i i peratures of 70 to 100 F and expected drywell humidities of 0.2 to 0.8. The l l break spectrum considered ranged from a design basis accident (DBA), an instantaneous guillotine rupture of a main steam line, to an intcrmediate (IBA) 2 size break of 0.5 ft Condensation oscillations do not occur for breaks smaller than 0.5 ft because the vent mass flux is less than the CO - chugging mass flux threshold of 5.0 lb=/see f t (Ref 1). The assumptions and methodology i l discussed in GESSAR II Section 6.2.1.1.3.3.2 were used to determine the vent mass flux, pool temperature and air content parameters in this sensitivity study. l t i
I i I i t 6 l l. RESPONSE - Continued I I l The CO amplitude and f requency and the resulting forcing function were calculated I from the mass flux, pool temperature, and air content parameters using the i methodology delineated in GESSAR Section 3B. lC. t i i Figure 3B.37-1 shows the ef fect of initial pool temperature on CO 100% DBA peak-to-peak amplitude as function of time. Examination of Figure 1 reveals that the i i peak-to-peak amplitude decreases with decreasing initial pocl temperature. 1 ) Figure 3B.37-1 demonstrates that the CO load definition is conservative because l l it is based on the maximum initial pool te=perature of 100 F and is above the tech spec limit of 95 F. Figure 3B.27-2 shows the effect of initial pool temper-ature on CD RMS amplitude. As expected, the RMS amplitudes follow the same i trends as the peak to peak amplitudes. i i i Figure 3B.37-3 presents the ARS comparisons of the CO load definition and CO 100% DEA forcing function resulting f rom initial pool temperatures of 70, 90, and i l 100 F. This figure demonstrates that the CO load definition peak broadened by i 15 percent bounds CO 100% DBA forcing functions with initial pool temperatures i ranging between 70 and 90 F. Therefore, e CO load definition is conservative and e bounds the expected variation in initial pool temperature. I Figure 3B.37-4 shows the effect of air content on C0 peak to-peak amplitude ar a function of time. Examination of Figure 3B.37-4 reveals the air content has an in-significant effect on CO peak-to-peak amplitude.
RESPONSE - Continued Figures 3B.37-5 and 6 show the effect of break sizes on the C0 peak-to-peak amplitude and the CO RMS amplitude as functions of time. Examination of these figures reveal that the peak-to-peak and RMS amplitudes decrease with de-creasing break sizes in the last half of the C0 transient. The pressure loading during the last half of the transient was found to be controlling in i the response spectra (ARS) analyses. Figure 3B.37-7 presents the ARS comparison of the forcing functions resulting i from the break sizes of 100%, 75%, 50% and 16% DBA. This figure shows that the DBA ARS either bounds or is within the estimated analysis accuracy (5%) of ;he ARS's of the smaller break sizes. At low frequencies (less than 2 Hz), where the small breaks exceed the design curve, the low f requency forcing i functions (1-3 Hz) are insignificant to both containment and equipment response. The response to question 3B.38 has further infor=ation, i The sensitivity results of the CO forcing function parameters of break size, vent mass flux, pool temperature and air content demonstrate the CO load def l.tition is still appropriate and is conservative, i i i I I ] a 1 l l i L t v v w s-m - o e,-. w - ---e . - e,vwm w= m - w-, w,-w-m-m.w--mwere,------,ww-,,,ww-,---. ,,-ww-e-w-we-, w-,---ree--w--w-w-,e.-=v--++--w-
i i 1, f i l 1 l i i l l REFERENCE I i i l 1 i 1. A..M. Varzaly, et al., " Mark III Ccnfirmatory Test Program - 1/3 scale I i Condensation and Stratification Phenomena - Test Series 3807." General l Electric Co pany, March 1977 (NEDE-21596-P), page 5-28 l i I ) l I' i-i i t s t e i i t i I i 4 f I I i h. I k 1 4 1 t f I i 1 } d i 1 ~
n.e., i v.,,..., , s, a. n ~' ~ FIGURE 38.37-1 CO PEAK TO PEAK AMPLITUDE COMPARIS0N FOR DIFFERENT F INITIAL SUPPRESSION P00L TEMPERATURES. ~ ~ P0OL TEMP. = 100 T (DESIGN) d i f + l l ..... [.. l... li.d- _.,.,._.l..l.l.,.._.._.' l l
- i 1
i ,JJJi_ .l. ..I...._.l.. t i 1 l i l' - 1 1 g.. ..i l . ~. [ l. I I I { l. .l. )3g. A" POOL TEMP. = 90*F h -}- g .. L. -J . r -- a t. 4 + t- - y : .... q . L. j .f p. .m
- i. g.
t j I i i i p .1.. y .u. = ll/,g 1./
- { !
J. l r l .l a ..s t
- ,.m
.-. [_... ., p. .i 't l . i. .i .s. i a. m l.).. 4 i .i l Jg,g- . '.7.. ..{.. j ..a._...._.._ m / .. POOL TEMP. = 70*F .. p. ] ggl. t 8 i ..s' .__.i ... p.. .4 3_. ..a J.p . J._ ......i_. ,... p _l... [ ..q I i ..l h,. n. .i.. ..l. 9 I . e .. l.
- u. -. '
u f. 4 .s g, .,= .,. l,. t . )__... _,,.,i.. a. l l in i. i S b l.i.I I
- g. f,p _
.l. .. {.. .g. .. _...i. _.o.. 1 I l l w j. l a. l O .L O 1 BREAK SIZE = 100% DBA 3,'g. u__ l____ ~ 7 7 ,_. i 1 INITIAL HUMIDITY = 0.4 l .f. .l _.a.-...-{.. Q s .l l 7 4..m . f.. .._.._..= _ .; m. L I . I 1.l-4'-;. p L .i. .,,. j.. .m. 1.ju b. i. ag _ c c ;._ 7 _.p._..J.I r = 1 v ag_ m.
- 3..
q.
- r.
l j - H'.. ..I ,~ u n,. 3.0 ..(. -..p.. I .;, 0._ .. j. j j. 4 r -- -l- ~ ~ ~ _ q. t, _ _3...q .r- + a' }. l ! i a. i i I t t gg Jol0i, 15.0 p t.p. L.2$ 0l.4 3 4 op Sig..j..W,0;.MS.On_.50;0,j i.
- 0 O!
. !5.0 } i TIME (SEC.) i
...g.---- 4 I ! __ l _ l { t + i 1 -- - } j i z i l d i _ ; _.. e _ _1..a _L. a e i .I 3 4 f I M 2 b i O i i i i u 8 ~ D l z_ v 8 H .j f _'. F t, or 5 - l i i 1 w 4 cs& 42-oH I i o - i 1 i m -I - !. 1 - i. i o. ! o N -j o en.. .i. i-o j T 1 1 - i : 1- - { n-i j
- i. u.. I' ~ E Nx
. l, 7 - W j
- )
._j. n (. --- ;,. _. 1 y-- j : s... -2
- - t 1
~,,,,) ]. ; 4 -. r t 9, C-i_ +- 2- .f._< l E;n 4 J = -..i.. i. m 3 - i 4 g g W g Wl%- H M i. -l M- .[ j j.' _k H r I Z
- . l.. "1l _h. f
- '.[
_L... i w + -2._ ; - + <H w ' a w- -i ;o Ez 1 w [ 1' 1. E a. w . 1 rm. o. .-o co - 1 .T 3 o 1_ o l O -., a - 1 CL w 0-t._.: _ : -. a Q I ) l ? I f' ~ ' i 1 ~ 3 4 o I I J 1 I t -j I [- .. _ 7 ~~ ~ ~ 'f ' ~k ~ ~ w I l 1 t 1- _f f 1 t .' 4 t l _A i n2 ' ~
- .j -
.[ t-. j . 1 o 1 1 T. 7 D g I a -W J ,1' I Zw 1 -- (, .I 'f 2 -" _..1 ._ _~. [ <E l. .._~~ 3 -. f--- ? d,. } .r.4. m- { t ._j. m L'.. 1. CL 2 7 {- I. ] t
- } -
w I - ' ' ~ 'k j g 4- - -
- _-'
.y .!_. - g... - :,g o g E l W CL I T-OE 1._ I ' ') ---- D-- 1 Ow -3. j. t , ~ 4 .i. ..I.., 1 : t a gg .q. 4 i .3-e-= 2 a- . 2_ _. f4
- --- -- _ : : _._ _j_ -
8 i ^ 2 gg ...1 1.- 1.. 4 .i g3
- v. _ g + 4 4
xo _. < Cm - !~ i _j i j f, r_ __ j -. r y- ._.i.. /. }....- j 7 k ^- om r. 3 .j j. ..p- 'I i..'-*1- -. a . F.. -MU i Eo {- l l i .7 .-,w j o em a. - _ a.. _ h. _. t_ v.r. m 6---- mw - ~. k 2' s-- 4 = c_ t 'l - .-. W .._'I. a <C 4 L.e.J O 6 e -'*~J .2 1. .l e . E !~~-*' EW - ~ '~. __.j _.. g_ Q s I I H L_ . 1 pg \\. }. j - 3 o< 1 a t o- \\..,_. ; l ~ %+ _,2 3 3 oZ . 1. _ f 6' O 8-8 . 4 4 f I 1 1 t Q 3 i k i i i.: :. j-. ki' ,_L kO .: =,. _ - _a_ 1 . _. -. m -L. j i. { 6 g i. 1, i _ .L.._. I.. _. j - '. f - I. O 3 f -j - '..,. f f ' h,$ b [ f E I - 3 a ._~k. .L_ I. .f ./. b. -* _l.. f ' - " D,, h i _o .. i.i a. i t -t ) ..): n-! - 5 ._ _ }' _.. 2 -= -. $ *. J j .1 ~- ~.1._. I I.&- f.' 6 1 -- -I . 5'h. f [,. -_ ^. L.. k. i.' ..-_.U, [. 'J . _ -_ ~ k _..J_ _ '_ _k.._d --_. ) *. *.. -._. b-[._.- bd_.i.. .~ 3 '~_T g* -+ ~.. ~ _ " ' ^~ .._ E. i F,.. t. ~: ? }.M.. F { - .I"- .].1 2 9.. F_.2._9 :
- .-- i
. I ~ i.._. -}..; 9.. g -- t. c '. ~ d. ~. i.. -----s. n y-W ~ ~ l h1 .h-l t .? .2 { 1. ...J, 1 j ..T c_- t .. 1. } } { I: l 'I - I f Q j ~: v f I 4 I I D ? ' _.4 I d i ..-M5 I:d i
- F i
4 ' J_i ~ f i tN (ISd) 30n117dW 3 WADS W3W 100B
_=_. i t 2tl0 FIGURE 38.37-3 C0 ARS COMPARIS0N FOR DIFFERENT INITIAL SUPPRESSION P0OL TEMPEPATURES. j 200 h ^
- >I 160 Ill 100 *F (DESIGN)
- 90 *F l} - - - -- - - 7 0 *F 120 w lI l )\\' E m. a 4 80 E ll P /L ' s) r d I,,' h / _) I i u s I l A \\ l.l \\j -1'0-4 I \\T ~ 40 \\ I,' 'A ll 'fy I,' \\' j'A,ll \\\\ ll \\ ,j hl I \\', \\x =,/, (w s ' b_-- _c gj ~ ~ 0, I ~ ~ ~ ~ ~ ~ ~ ~ 10 o 10 i 10 2 FREQUENCY ( HZ )
I i f
- f. I I.
l -1 I f-t I j f f I { i 1 a ~ ~ ~ I. t i { i 5 i - -- A -. - ? r } O Q' ~ i i ~'[ ~ ' ' .* n n n D l l <C l, p i I... l. . l _I ._i I 3 my ppp f _. t -I.'..} ['.- - T. :: ( . f'.. &-e & & l I O t. e-a - HA OOO a -,,_, f I ] '. ' {( l ~ ~ ~ ' -. L i._ s r A.- I OE "r"" *r"" *r-* .. ; a. - t '--*'3-- ow MH DDD ~_] { D -[' d... -l f. xxx t K k uJ D t' _ i.. I. j- .c l ,.. J.,_ ; O -.J J J
- c J
t t LaJ O < < <C t- _J. 4 i 6 .{ ' - [- ^- I .-}L N CL. t {:: _1-i t. .t i - - I i. ( -l. r e s. _ _ u i l i f. 6 mJ r et .z z z. Q*- t M ew e-i l <C & d C~s ^' $E. l m ] l I_ _a co - i -r. -y p g : I u. .' t a. La. i r I i O i 8 4 i i i C e r f ( g .i l-S e E - 1 *. % . i-c ~. !- i- -j. t ! g j;_ i, 1 . h. 4 - f r.C (n-iQ.' j j. _ i ~~i ! l F -- t L ~N 2 .g M,. + z .\\ .t -r r :l_ i .. ; -.j_.- o e i e r i W 3 -2 r.. t-i l t t j ._g l. i I l i i ~! i i i j-l i .g g l m 8 f 3 f M i f t.J L6.8 l 8 t { t Q* ^I f~ g f q D[ i {~ I 1 l - ~-i b g- -- [ M j y 1 -O 4 r U J- 'l/T i l 6 1- .t - i / O M' 3 C1. r L r i I 1 KD i*. I r e t ! _:. t 2. %x f l [ l' L.ti I b . M 5$ h! !:J l h Y- \\. CL + g i t t _o ou j i~ i M &Q YA i k ~ i i w
- -e ch. H
, _ __p l_____ A l oz O 2 _m 3 3 j l D _.. __. i + 1 i i i. 1 3 i N i g f -*a . I t I
- l.
- M
.,l s s i- .D,_ t I CO l-j _.,D j -
- i
.l l j k..
- l...
- i...
i. m -t l t l - j5 h ~ j-It f.- !_ - [ - l ' i t-W t I i 5 g' ~ f -t i i' D _, g h' -!~'- h L. f. ., * -e. o !I t I' '8 q- ]{. -- Cy- .i J .- f..... r T! T. l ~ E _' ~_ _. _. j _. ~l_. i y t-h - .-i =- - i i'l.~I+~ 1: I b 3 d [' i _; i - ! ~ l I-'4
- -! f
.u - 4 ~ ". j i i ! i-M'; f f. __- 4 :. ; g i .k. - f-f. 4. L r.; f i D i i l i-i i A-. 6 6..A-.5 V 6 A. L.A-._A Lh _-..L _$. %e.t_sti s.3 s. tn.. s as et + o e e. m sn t-3 f (ISd) 3001Ild'W W3d 01 W3d
n c m ,..l . i.. , j,o. i,: g u g. _ i-m l-1l- ,n ..s l '. r,t si g ra _ +- -@,:I gi
- i. i. :. j
- j
" ~ ~
- ,. q. i t
!.i p - :I - I - I'? +1- -l-- /,,.,,. " j,$, I l ;. ;. i ljg;g-g
- I l
.o I l., ,.f 1;
- .. \\,,s.! j::1 ll
.i p,- g v. .., r. v _ ';, L :,.1.,, f -M -}.,,-l r i;'n l l....ui:;: /. y.... '.:' k;}3.-0. /..:.;;;
- t. ;j:.
- i.. -q r
+ ,i ' l ...p.. y. s - l..!; ii M l /..
- ' '
- i!
!i! I il1! fi Il. li ii!; N: !!!,' h.. '\\: 3 g. .,ii i!; p ...e ,.., hf- .I. r--- - i 7.s-l .-. i t :. .,n .t.:. A: i!- ! H,
- !l 3
I... E Tj* :! ~. l;ll 1.: .- %,,l lil:.- l.;!
- f:
.i-1: i;:: ::1. ..i l . I. ! 'l
- l. h
- .!: k: ;;
'gl .w l t '::
- l, j ;. g :
i.
- l w
ll
- j :.
1.: '.i g ..l .;l.. . ell! ll,. [. D 7 ., I' i
- I:: !:
- l!!
':I I!- .r- ,g.,f,- .l,, ~ i It:: i I; a 11 p ': l g !!. ! 6 : .T. s.. i ..,i ..,.,i j n-
- ll;
-f 7. .Il' ll .l l,. .j:t i li. u ..i. n i..,.j..! i .L g .g;; et .:- !!n !iii 'i i-y .:e in j ::: Y
- i ! !'.
- i l
1: m-11 O- ::.: i. o + INITIAL POOL TEMP. = 100 F ,j. ,l
- i. !. l ll
!. i. :..,;!l j.;'..I l.
- 1.. ;
INiilAL HUMIDITY = 0.4 x 3 .. 3y 1 g g; p g 1
- s. u
- li ll glj 1001 DBA(DESIGN) a.
g a...
- g I 3.' '8-
- ,.l ', r:. :p.
- --- 75% DBA 50% DBA I q:I, .l: l'i ,i.',. i 'l ;, l... ll l: e. ! l -- 15% DBA ..g T,T-
- i. I
.i. .n I I i i !!l I .I,. l; .e. I i; ,.1 i.
- f'.
- ,
1:1
- g' g.
~ p.. t .;..l.l
- i,
.l.. 1.: !. I.. 1;
- I.1 I'
- 1
! 'l :. :. ..e ' :l .i g i. i ith
- T~;
U O.7i!' ;: 0. I..l . I P.1 l j '!D.e ' j ip.4 " 40. 5 !O."6 !!lio.7;i! '1 O.8 !- ' O. 4F
- 1. 0:
i p + 1n p-i. t ii. i dine NORMALIZED TIME (total C0 duration) FIGURE 38.37-5 CO PEAK TO PEAK AMPLITUDE COMPARISON FOR DIFFERENT BREAK SIZES.
s l r... i. 1 i e i l i i l l __.. l e i _l g,p. '.5 - i FIGURE 3B.37-6 C0 ROOT MEAN SQUARE AMPLITUDE COMPARISON FOR DIFFERENT 7 BREAK sizes. l l '7,0 - u .i l l '6.5 - l l i i i 6.0-i l l l l m S 'S.5 - ,'\\ I. i i a ,o .0-.;., i i i-i p ' l l 1 n\\ l l l f.5- \\ s.N i /:p1....w - [.*,...... * " o s t .s i i I a N 1 / i, .. t e...- %.1.%x '.N /
- ~,. +, ;:
y .+. o - / g.4 ...*.s.,.. / i D g .*.' \\. i R S.5- <% "..N...***'~". i _w.6
- w.
l s g N l s-3.0 - i. w i x h.5-i g e INITIAL P0OL TEMP. = 100 F l h.0-l INITIAL HUMIDITY = 0.4 l 100% DBA(DESIGN) 15- - --- 75% DBA i i ,1 50% DBA lll.0 - i 15% DBA I i i l 0.5 - 1 } I i i iiii .. 0. 0 ao ' o.ii
- o. 2 a.9 0.4 i o.'s
- 0. 6 0.7 i 0. 8 0.4 l.0 - nL 3
e l
- i l
l I I ~ " ' tiene ~' NOINLIZED TIME (total C0 duration)
O 2t10 FIGURE 3B.37-7 C0 ARS COMPARISON FOR DIFFERENT BREAK SIZES. I I I 200 DBA (DESIGN) I k 75". DBA r-q -- 50% DBA l.k 16% DBA 160-ylIj ;. m 120 4 l 8 q \\ l a: a. 99 e is i b / / W ) i i {%;% 4 i l f,%, / l l by (. ij [j '40 c i~ i A / j t10 [ j l, /g l l '[4 , q- // i \\ i s^i / Q ~ 0-10 2 10 i 10 o FREQUENCY ( HZ ) 1
l QUESTION 3B.38 i In the ARS comparisons prepared in response to question 3B.13(c), it can be s seen that the C0 load specification f alls below the experimental curves at both extremes of high and low frequency. This is particularly of concern at the high f requency end in view of the f act that the lowest structural reso-nance lies around 20 Hz. 1 It will be necessary either to prove that these exceedances are of no con-sequence (as, for example, if they are bounded by other loads) or to modify the load specification so that it has a broader frequency content. RESPONSE 3B.38 In the ARS comparisons prepared in response to question 3B.13(c), the CO load i specification falls below the 1/3-area scale data for frequencies below % 2 Hz I j and above the full-scale data for frequencies greater than % 15 Er (figures l 3B.13(c)-61 and 3B.13 (c)-64). The low frequency forcing functions (1-3 Hz) are in-significant to both the containment and equipment. The natural frequencies of the containment are much higher than this low range and equipment response due to this a I very low frequency (1-3 Hz) is bounded by loading at the higher frequencies t (above 3 Hz). Howet2r, the pool swell load specification and chugging global specification bound these data exceedances. This is shown in F:gures 3B.38-1 and 3B.38-2 Thus, the pool swell load definition bounds the CO data on the lower and higher frequency extreme and the chugging glom 1 definition bounds the CO data on the upper frequency extreme. i
= i I I L_ f 0 0 0 0 0 0 O 0 o o E 2 1 s 2 2 2 2 2 2 2 W N N 'N N N M N N N N I D P P P P P P P P P P SN A 1 2 o 9 3 4 6 5 s 0 DI 1 t 2 2 2 2 2 : 3 L0 L s e N N N N N N N N N t f s 7. 7. 7. 7 7 7. F. F 7. F. MI uU U U U U U U U U T e p G H H H H. H. H H H. H H s E 0 0 0 0 0 0 0 0 0 > N D 8 a a a a 4 a a a m Rf i F 6 S b 6 e 5 6 6 b 6 0 E N 5 5 5 5 5 5 5 5 5 5 R 0 D O I t f l H H H M H H H I P Hl f ( t f t t e e e H 1 e l i e e l A I E SD TL 1
- E t
t t t E 4 t e ?. S 5 S S S S 5 5 S 5 S A NCA I i T T T T T T T T T T T ED k S S S 5 S S S S S 5 I S GL F E E E f f. E t E f e e G T T i i T i T i v v EA /m DE R DA A A I O3 L/ w ?. 1 i L L, EL a WL ,M l ON SA W L G OT N a PI W u T H PE A lC Y M a A EN D 0 N HI l .t TA T N M' O DN E Q 5'% i C E NO H R AC E i F P AE 0 TH Di 1 I AT D N 0O C N ~ FO Qa OI g M O 'V N SE I L
- I 1\\
RE n A Pt j f P. wI Mf OC2 ). a f = 1 h_ 9 8 'i g 3 ilgI I' n 3 l ~ e E R U G I F /' / o 0 m a o S 4 ,I$; a w 5g< 1l
= { 60 % J Pt RCt NT DAldPING 40 GE SSAR 18 00 (M EIGN LOAD TEST SE RIES 6 707. RUN IS Pte 2 0 T E ST SE RIE S 6 707. RUN IS P9013.0 "- TE ST SE RIE S 6 707. RUN 20 PN 2.0 ... - TE ST EE RIE 5 5 70 7. RUN 20 PN il.0 g so E r G J M Sa 1= i 7 [ 30 J CliUGGING ELL) SAL i LDAD DEFINITION W / i se _ p p a j,, A
- p. W o
10 800 S l rRtove Mcy iM FIGURE 38.38-2 COMPARISON OF C0 DATA AND CHUGGING GLOBAL LOAD DEFINITION 2 AND 11 FT ELEVATION ON THE CONTAINMENT WALL, FULL SCALE 1 l 4
. _ _ -.. - _ _ ~ - - - _ _ -. ~ i 3 QUESTION 3B.39 The exceedance value relative to the design value for the " local" chugging load i on the we r wall exhibited by.eir chug pressure trace No. 22 (Figure 3B.18-3) in the 30-40 Hz frequency range continues to be a concern. From the point of l l view of the applied loads, we see no justification for this nonconservatism. This nonconservatism may persist even of the " global" load level. This non-conservatism can possibly be justified on grounds related to structural response, i.e., the frequency range in question (30-40 Hz) may not be significant in terms of the weir wall response. If this approach is not feasible, then the I local chugging design load should be modified to include a post-chug oscil-lation of appropriate frequency to the weir design trace, and the global i design chug loading on the weir wall will require either an appropriate modi-fication to include weir chug pressure trace No. 22, or further analysis to justify neglecting this component for gloLal loading on the weir wall. (The use of Monte Carlo simulations similar to those done for basemat loads (Figure 3B.16-5) is recommended.) l l l RESPONSE 3B.39 i Round I question 3B.18 requested amplified r.tsponse spectra ARS of the pressure-time traces exhibiting the maximum chugging amplitude spikes. Chug 22 run 1 had the maximum weir wall spike. Figure 3B.39-1 shows that the ARS of the chug 22 pressure-time trace exceeds the load definition ARS in the 30 to 40 Hz range. Out-side the 30 to 40 Hz range, the load definition bounds the chug 22 ARS by a factor of two. A structural evaluation of the weir wall reveals that this frequency ex-cursion dota not affect the weir wall structural integrity. Thermal stress is the bounding weir wall load condition. Thus, the 30 to 40 Hz frequency excursion is not a significant factor in the weir wall structural integrity. 1 l
I I RESPONSE 3B.39 - Continued l Figure 3B.3-2 shows an ARS comparison of the load definition and a measured chug number 32 of run 1 which represents the upper two sigma limit of the weir chugging spike amplitude. Examination of Figure 3B.39-2 shows that the load definition is conservative by a f actor of 1.5 or greater. I In conclusion, the weir wall load definition is adequate because:
- 1) It bounds measured PSTF data except in the 30 to 40 Hz range.
- 2) This 30 to 40 Hz f requency exc arsion is not a structural integrity problem because the thermal stress is the bounding load condition.
- 3) The load definition ARS bounds the ARS of the measured chug pressure-time trace which represents the upper two sigma limit or the weir wall chugging spike amplitu.e.
l
FIGURE 3B.39-1 AMPLIFIED RESPONSE SPECTRUM MRRCH 21. 1901. 120- -8 COMPARISON OF GESSAR 11 LOAD DEFINITION AND PSTF TEST 5707 DATA ON WEIR ANNULUS 2.0 PERCENT DAMPING -
- litic 22 Rtiri i 100 1
8 l I f i 80 ~- -< h .s g GESSAR II LOCAL g / g LOAD DCFINITION y 60 f ,\\ / c f+ Si** / E L10 -r [V r CHUG 22 RUN 1 MAX. TEST DATA f ~~ n 20 / / ~ -~' 0 I 10 o 10 i 10 2 10 8 FREQUENCY ( HZ ) PL 1
n I i ) =* O % 1 l I I i .o l l
- N 1
l J l f a I P
- D N 5/
L
- =
8\\ g g2 y5-a c. lh h! i \\ O i N y, ^ ~ l j I k 7 E l lJ I I NI C E I) b = i y l U C tE z E l Z L7 } ~- 'J c E 4 I D L; ) Ci E O L7 'd W Z C c/ O 5 5 h: L L x = (9, L7 $J g u = a e en C g i x O h Q \\ a y '5 ~ f f \\ \\ O E ~ l 4 L-G 3 d l-I i i - ~ x g I ( l l J I I, b 'I $ 0 C c-1 () \\, l ^ I e O { l i O O O O O O o-N O D D 2 m J'* 3SN0dS3B G3I3I'IdKV
- _ ~ - _ - _.. --_ - - ~. l gpestion 3B.40 4 The ;.ticular choice of spatial distribution employed on the wetted boundaries during the chugging phase (Figure 4.8d of the ICLR) is not completely supported by the information thus far made available. Many examples of the measured pres-l sure amplitudes exceeding the load specification can be cited for pressure sensors located on the basemat and containment wall. We recognize, however, that r t the idealized pressure signatures which are used f or design may nonetheless be adequate to bound the measured loads on a power-vs-frequency (ARS) basis. Therefore, provide a comparison between desi;n and measured ARSs for.ny case where the peak spike amplitude exceeds the local design values. These comparisons should be provided for all pressure sensors located in the wetwell during Test runs l 5707-11 and 5707-12, which exhibit such exceedance (including drywell wall, if I any). The comparisons may be restricted to those cases showing only the highest amplitudes (i.e., only one comparison per sensor is required). 1 Response 3B.40 Figures 3B.40-1 through 6 presents a comparison of the load definition pool boundary spatial distributien and the measured pressure data f rom test series 5707 runs 11 and 12. Figures 3B.40-7 through 10 show the cire.umerential distribution reasured on the drywell and containment walls of 5707 runs 11 and 12. Examination of these figures indicates that the following chugs exceeded the load definition spatial distrib ; tion by minor amounts (3% to 15%). l CHUG ( RUN NUMBER LOCATION i 11 52 Containment wall from basemat 2 ft 11 52 Basemat from drywell 12 ft 11 27 Drywell Wall 8.75 ft l 11 27 Containment wall 14.5 ft i l i l l.-
i i Response 3B.40 - Continued j lt is important to note that the load definition spatial distribution bounded 111 out of the 113 chugs used for the load definition. Examination of Figures 3, 4, and 6 show that the spatial laad definition bounds the integrated pressure field of chugs 27 and 52. J Figures 3B.40-11 through 14 present ARS comparisons of the load definition and the pressure-time traces of the 2 chugs ouc of 113 which randomly exceeded the i load definition spatial distribution. Examination of Figures 3B.40-11 and 12 l reveals that the load definition ARS's bound the measured pressure-time traces of chug 52 and 11 on the basemat 12 feet from the drywell wall and chug 27 run 12 on the drywell at 8.75 foot elevation, respectively. Examination of Figure s 3B.40-13 reveals that the load definition ARS bounds the measured pressure-time trace of chug 52 run 11 at containment wall 2 foot elevation except for a minor excursion between 20 to 35 Hz. This minor f requency excursion is not i considered to be a structural proolen due to the slight exceedance which is within the analysis accuracy. Examination of Figure 3B.40-14 reveals that the load definition ARS bounds the pressure-time trace ARS of chug 27 run 23 on the containment wall at 14.5 foot elevation up to 150 Hz. This frequency excursion j j above 150 Hz is not considered a structural problem because the structural l and e tipment responses to frequencies above 100 Hz are very small. Based on 3 multidegree of f reedom and finite element analyses of structural, component, l and piping systems, most of the strain energy due to the forcing function s developed below is 100 Hz. In addition, strain displacements above 100 Hz are small and insignificant. Therefore, very hign frequency input is not damaging l l to structures and most equipment even with high accelerations because these high l frequency pulses do not possess significant displacement or energy content. 3 ( In conclusion, the load definition spatial distribution is adequate for design. I l i
FIGl'RE 3B.40-1 1 4 l DRYWELL WALL SPATIAL DISTRIBUTION RUN 22 ~ ~ ~ l ' _~ ~. _ [' - ~ } l I j - g, 4 i -.._ __ j - - _. 9 _ l _ i ]_ pT*='TW. G~ fg i 1 (- j j f LOAD DEFINITION 1 =-- - -$~ j 4 n j N a a. t I l ]
- s w e=~ w ee oe-
~ j 7 A 4 4, c.;_ 1 J U i
- 2. - AGMNhwen--~GG) @ GG iO
'O O d V i i 4 i s i j b'_ J. [ q-g 4 { A 4 s = j i r 1 0-h ~ t ? I r .i 4 l-3 l 4 E t i t I I EEGDinmSTA@ 4' J + z i e _p F i, ( q. I a 5-i 1 4 i f. J 4 1.__ 4 = j i I f - f i L_ _ _. J - - 9gg 2 4 + i j 4 j _ 3 . 3, { 9 __. } _ ____1 j _. j.__ j 8- + 4 i. _ y -i i J j j 9 1 -o I-j j 4 4 4 4 4 ___._q. ,._ 4 J J_ l. __ _. J _,_ _ y ___-._.3.._ 3 9 . _ __ _ q _,g g. 7 4.___a____j .{ c j c._ j q _ j s - --- J ~_- ---~; _*~d. _ __,, _ _. 4__ g _ g gg g. _ _ _ _C-4. 4 2 p Z ::_ _.jNO LilE PRESSUG. ( l = 100 PSI) '. ~C-
- _-_ _. ____. _ _4
~ --- - i - : 4 t _. 4.. _ _. _ _ 3 4_ ( 2----h -- --F _~ T 5 T-Si h R hi . - -. _ - + _ _im_. _ _. .ig. 1 e au--- - -. Ww6 ew---4
- -ehe
.O_
==._-@6+4 m O -e e a m,e 4.. _m.+4--' e-- 6 h .e e6 N-=4-h4 --m e--e 1 egw+.gie__ese -g M e.e. ._-g -.e g .4-. h e--_. .e --ey.h a 3%= m 4 e.,, aw 4.- anei 6mhii _w6-. 4hW. Oh gap-4 +4M M w =g _-1
- 1
--*4 h '*Wh-4=-.
- MMe9-e.e4 eme M m-
2..-
- " :.V:
- i. : ^ ~...:. :
^
- 1
- t
--i - 1::: 1T. __ __ -.. _..-_._._. _ _ p. - NCnHALIIED FRESSJRE (1=100 PSI) __. ) ...y ~ ' _^ 7_ '.. ~ 9 o 9 .o o P 9 P ._. O. o o o o o o o o ~oe .o 7 y th b g y g CD -b.. -. 5. _ _ J . _.. _O 1 i. 4 _._.. j c.e eg e n. {.. __.... 2._ m,e.u.m.3e qqp . ___ _4 p; f ( 8 l } . i 7, l j .C 4- {._j 1 { 6 7 m 4 { p. l 1, g i 1 i i 9 4 1 _- - } a 4 l l l 1 g ) y e i l
- h. _
l n a s i s + 6 + 1 j j j 4 t i t I t + t 8 i ? I r 4 5 f i 1 i s 6 C-J j 3> Q-t l l O N i 3 N V -t M M V,_ 5 I n co - 5 i en = = 2, r-4,c \\ e J ~ i V l X l I Z C' 1 I A 6 N l* C C p. e e n-,-mm.ims e ee e e q = t k .g = = i i O s n un - t + %,s C l y os O C i V, 2 l n I 4 j l
==. i + i i 3 l r s a au > t-1 .___3 ._,i i . ~ _.._} g- + } j . ~ _ _ _l l ~ i l g g i f' g l' / 4 l i ,r i + s
FTCURE 1R.40-1 I I t i CONTAINY.ENT WALL SPATIAL DISTRIBUTION RUN al i 1 l I ( j-6- - 1 I j. i .t .i 4. _ .j____ { _] ._y, FI; '.'t.W.? (g;. 4 4 3 4 I i j -.J.__ l gg i 1 ( i 3 } I 5-i i i i i 1 J 4 I . ___ y.y_';;.::s.y.(.. e _-.j. p 1 f~ 3 i i i OiM 52. i i 3. uw m d_0AD DEFINITION I O -P WF 52-o- G?-6 & i I' ~ l. a 0 j 9_ i g 2-l b _ _. I y 3 o i CriUG 52 k ~ ~~ -~ } l g m emo m:{c 9: I i l cf - i 4 ) { + i j t r f i, 1 I 4 j 4 4 ,_}. 4 1 I l ) i I t i 1 j I j l ,4 i, . _ -EOIGG8C6GGG-EG C9-O, 0 4t40G-54 ...-4 (l h) p...., g o. _ , CHUG 52 i 4 j I, ~ ~ ~.,1 - ~ _ ~ _ _ ' L 4. z.qw.y g55i 3 y j ~ll i 3 i ~ - ' - - " --0 0.0 t o 02 co3 0.04 o.oS ~ -- ____, _3.. 4 _4 . _ _ ___,._; _ _ 1 1 1 NDRMLIZED PMSSURE C1=100f'51)--r.: _<__ 1. ~~~ - Z _a _ _ ~ _. * - 4 4_. j _w4_ ._su.sp q .-*.w eem in e _
- emgg,
.ee.e >.=_p-.._.____{____.4___.--,w ___4 4____.. ._-e 4 M 4 _i .. _ _. =.. __.__"8- .. _._ _ 4 _ _ _a ._.___{____j____.__3_____._, _. 4 _._..- 4._ . _ ___q___.4_.. _. j _.__q _ -__. j
ElGURE 3B.40-4 t } DRYMELL WALL SPATIAL DISTRIBUTION RUN 12 ~ j ,i 1 j l i I g, i I i i 9 9 - 9 .g. r]' 1 4_._. l {... 4 ~ .mg 1 I'-~ LOAD DEFINITION 4 j 1 { i - 5_. a 1 4 i I { i 1 k' ~ f 1 l I (-!, ]. E @ d? 7'- - 6 I C 3~ caus 27 u i -4 ,.eu.<9 G,a ~ (v 3 - - a -.s. 2-W.. f i cnoc-27 / i - _--g H., j u,, i i ._ __ c. _0 - l i. .i e i i 3 - -m =tE=.' + G e cuus z7 (4 5 % )- ---- l l _3, -s E ___ g.g. _ _ _ _ _ i 4 tu I i - 4 ~6- - cr,x 27 t
- a. :.s i
i i s i 4 2 1 , _ g. i 1 i .i...-- s i i ~ i - 4 4 .._1 -4 4 f,. t _3g-j 4._ 4 1 4 i i. j. - - 4 .4 _(o. 3 c o .1.__ { j j .._ _4.. r___ _. p ~ ~. _ _ _g i c ~~~ h n, p _._cA, 6,5 nl6 n.1 nA a9 ! c_ 1 4 2 4___ 4 _ _l _. e___ _ _ =, - _ ,4 4NOPJ1ALIZED FRESSURE Ci=100 psi) i-_- r. ___ _. 6. _ - _ 3. 3 3 _ _ _ _. _ 1 . j 9. 3 .._ _ 3 __.. ; k l_.. 4. _.._. ._ ___ __ __J. _ _ _ -. j. _ j._ _. _ _ _. _. 1 _. _ _ q _.. _. 4_._j-_.- l .li_ ~_ i _~_.'.~ i IlZZ.. i_2.: : :Z.._. Ti--b~l_.- ~ ~ ~iJ. l1~. i _ _ i f_:.~_ _- ~i_ l---_ ~Z. _: _ _~ i ' j _a___ ... _ _j _ _ _ - j
g e .6 Wh q@W.m_h m..w4.ee.- Mweg _M_i'-+M_m
- -e a
'_*+--e_ h-. No %-4 i ee%_-. .m g e e + ..--e +e-.. -*-m_ .-M.m-a e_ e. 4 . _-._-= ..==e.-_ .__-.i -._--e, ,m w4igW B=es.'-6-->hm &--i &N O 1 h w = ew- .g-ui e-i em.g p,4 e-- e-. m j N o __ m 8 "T1 U g { w i e 4 g. I i ^ 4 i l f i I 4 _. { l _... _ _ ___. 4 l l. _ .g. 9 i ( u-h ) } { 7 i i g g i } f { 4 r } i l __ i i e h
FIGURE 3B.40-6 i i CC6'TAINNENT WALL SPATIAL DISTRIBUTION RUN 12 } 1 i i l l 1 l { } [ -- 6 i ' ~.:. ~. -'I ' GC6 0 l
- ~~:
i ( -- -g, l g g' I { . ZT. I 1 4 J _. _ 5- - ) 1 1 s l i l 4- - t 6 i -- -m e <!Nud -27 E i 3 (3 'A ) i i h U 2- ~- ti 6 lr 4 _ 20AD DEFINITION _j _. 4 0- _ _ y _ __ _ i { o<us :'l ) s c. cp -l-I F t 2 - o i }- _,2 EEGeomsvE l d - e,?e, w i -Cnu&2't { i, t 4- - 1 i l I i 4 i i I f I } j - I, i _7 - 4 j l t 1 j i i i -g.. i 4 04ae 27 I _ J' _ 1 g ] i 2
- --EEEEEs e-e-e,ec e.
1 4 9 i j ege. e. -e - i, V i 3 s --Il_. m,
- _. 3 _ _ __.
-o o.oI o.o2 co3 o.o+ 2 ~ ~~ 4 4 J 4 4 - J e _ - 2 NWALINED PE5SURE Ci = Ice P50_ _- _- - _ - ~ - r. .; _. _ j ~ "ZZ ~~ T i i M.. T_.:i i l_ _~__ _ _i i_. - _ l_. ~l .. _ ' _ ^.-.. - ~_ _ ! :.~ i 12_".Z Z1 '^^^~~:~~ ~ ^~:::.: ~~ l z & L.. :~:: ~.. : :. L. -~
- :::~
. :X :~~~~ XX t -___-4 ._ma 4 -. = = =am- -e
- N='usi--e
-e w -e m#e.4 e e me --ee w
- 4
- W+*
- -+6
^9* N
- 6e6 1
e -u' 4 e -e -e-g
- W6**
- 6a=w%.
m i ha em a m en e 6 -e 4 wm****4 - = = - - -6 - *- 4 8 6 m e e 4 g-4 -9
- re 4
-4 '6 hewi um we-4 e6w= we*- 4 = -e ew* 9
- -4 6
4M-e 6 =m 1
- -O
-w -e W g e e d 44e4 O
i ,.i O 1 1 N U l tIi R N N O O I iI T l ij I!1fI .,5 I T I U B N I I R F T E i, l il,i .,4 S D l ~ I D D l ~ A E ~ 3g i,3] L O ~ L A 4_ ~ I ~ A ,\\i }1 L T @ 6h.. ' .I * - a,2 N P E S Y. L ,1 N A + (( -II I O T R N F E 7 R l
- lj li i!
I {, i, ,0 C E 0 E 4 ~ F M A ~ M l f U T 3 C t { .I. !i ,1 S E ~ R I R D I U L G C l I F n @g . O ye E ,2 A f F L Wk ~ l L r ~ ~ ~ A N l a W l E ~ e v R L yk i ,3 N l L yE E E r M . D 'I W I U Y R fC D E C. ~. ~ 5 g 9, 8, q g 5 y 3 z g o o g, g, g, g g g, o y i: l .,4_g{n9-}@a. c e ~-
- awwm4 o
= g'
- .;_<i.
- i. i.-.;,
i ,l .i a. j.' jil: i. i lii e 1 L j;I .i,. jf. 4ii; i. 4I .:i4 4 4 eiI I:i. i ' l,I 4ii '1 4ii ji}. t, i .l 1 ,l l 6i6i i ,l. l l4' v l
,11 .i i,, ttt ,It t i ,) i 'i', !!!i :: ll11!! itt illi ii iii' it i i i i ilt lip- ,,ji i.;, .,i, 4,, 1,, .. t, i ',+,' ,i, . '. '. '. ,.' l ' ,',',i t,t ' I,, i t + t t t t t rt- 'tttt- + 4 ,,,r i t i 1, ,1 liu+ 3 r r+,- lt.,.,, .1 1,, rn .;(1.:Q T.. lN s f. I.I.(E..G,DN,VK.,lO..W,,Jh,,p33,E,,htD'R, f,D, l,., h,,i. I, ii 1 -l -F +i+i ~ H H. 4,,4 , 4,,, i, t
- [ -
'. f... [*i ];- -h6 1 4 sj,,t ,j,, ,+4 i .- - 1 7 t ' t t fl t [.,,*[',,, ,h E +f-t P'. t ,,i 1 t e, r. - b,1 1 1,, 1 ,r 1,,. i 4 4 1 r,. 1 + g n,. . 1 t r. ,1+, 1, 1, 1 ,,, 1, ..,i 4,4,. 4 t,,
- 1.,.,,,
!.1, 3.-4 c. 1 r .Q c' ,4t t t i ' -a- ? ,4 t J,,t t t 1 t,'.i, i i,,'i,' ,14_-.+-pIq . p., ..j, 4,3.., ,1,, i ,, i,, 3 .,it, 1 r, c, }. 1,.i rt1. i,,, 1 i,. [, i -o;.y-i i n7 i ,1, 10 0 + +,,.,. 14,,. j .li, ..3 g m t, 4 .j,,,,,l,i .r, 1, 11, ,.y J,. 4n. ,s .t o ,- -T. ,t , 1., ,,,i, 1,,. , 4 ,,,t , 3 I l l, ,t ,f,. l!,h ,i >,,' ,fI .- t f j 4--b - ,,,t {,, j{ f'l ,,f + t ,(1.' i,, j ,. i., . i .L 4 .p i, i ,,,l .1 1 [,4 ,6, i,, .o;, l g.t i ,l, i u 1 4;.. m s.2, i 70'0 C '- 1. ti i i, y i,tl I. v i I.,,},- 4-rf-- 1
- t r
-H +H i 1 i, ,t. O,,I.
- ilit, 1,,
i., ,i.,0 t,,t i ,,i 1,,, ,,6 t i ite! .y i {}i, t ht.
- i t' i tt i
'i ii '. e ..Ol i { i 0). n. I IIil Ii
- jti, ll l!'.f I
! i } }-f h'. .,i! t i., i,i, _ i
- 1,1
, f, , i,f ' t.,, c> i 't' ll t -T i~rr-t i,.., ,;,. W, i ; ; ji,; .A ...a [,i;. j ii 8 p fi; 4 j j !j'i: tj0111Nid30 UV07.i' l., l' 'i $&AlH J lll .g - f tl - t 1 i .t , i ny,M;4N l,; t ti f l 'I t,, t t,',. ,I{i ,,It j '{
- i, ittt
. f-r t, lt,, 't t t t,{,li ,t t i,!
- I
,t ,,i t' 1 l t' ,I ' i t i t, i I fl lil !1!! '!t i f f' t lt!! i ! j j l}tt } J- ' II tj08 fl01108181S10 lV11VdS lVilij]83dWQ3813 IN3WNIVifJ03 8-O'/
- U E 3MMI !
t ,l j i; i i . i, i, lt i . t.. i ,i. t. i i ,I,. 4 l r-t 1 ?1 4 k f ,t,, 6,tl'l !,.t,!,'
- ,,t!*
- 1.....I, Ii* i,. ,i 1 4 t-c 'I''L , 'i,W, 't, (. ! T ihi j lj i:
- !:S 9
'E T 'T: 1-4, ,!,t i r.t --- 6' i it ,r, .i ii,jf, i,., .,i, .ti. a 4,.1 1,, 4 i t 1; fri t r t -f jtlf' p I,*,, 1ll +1 i{,,i f, t.{ t,t i, .t,. i8 6 4 i, ', f,;_ ),n t ..ti' i 'i .'t t ti, tt,t, 7 -t,i.l j-i ii. e,.iij i 4 i i .I,.; -,!i jt i; ,ii. .i. i,J t i t E .,,t i,, I, }l,. j j i 1.: 4 .... !f. l l ,} t-~ rt-i ei t l 4 .l,i. i,t t, '. " t e .t..
- ,i i.
. i er j.t, 7 1_ -1 1 1 11 4, ' [t~1-+ d ,i; e.,t{, ,, t, ii. .ii; n; 4 j 'iti rt 1 IH-i- . i .. 4 i: i; -n i i4;, i,!. t i iI t , f, t t4 tit, eie4t. i I t17# h-ied i?, i l+ ,t; j }{,, . t! + t,t,{t,
- {t-t t, I, I
- .f t
t i 1-' + f f. ' ',N' lt,'!)
- t. ', t, t.
a.- +- !. l. iw; - j t**, t tt, ni 1 l H-H, f-t1 .i, t 14 p'i l F. 1 ,tt 4,i. 1 1 Iiii i i. ,ji !,i h, I}I t i i i ;k .i .ti+ titi } it, .jii - t,N,.t 14 j,te t g* 4, t, 1 1 t; i ,,.4
- t 3't -
4., 1 , 4 4 ,1, l t ' }t t tt 4 t tl ttil j eg i r'- ! 'ri! t' t t ..!t.i t t ' t, !t t t t 'i {!' i t,t ifl t, i I t-ttt g i t i to,' ,t it l fM .iii@,,!I 9f' Y,1 M ~ iI I ,i,,'. 'f. lIi,i i4'I ',i IIs II'
- d. r t
"o, t i 1 i.? fi t 3,, .t ' 'i ' i 4 4 rnt t -t-- - tft, ' ', l{ !tl3 ,. I , * '.O' f*t-t ,t it t 3i te,. -f -.f--f f gt -]-1 I i t t,tt t,t,l ,.' f. t 4 t ,'t. 1, t r. 1 t g t, 1t 4,. + t 2 4 4+ i I l i? .,II i- ]4 tjjt +, l i,.t'i,.ill i ii . t,i. .i,. iii.i liii ,:f,I -i f.h i i, t. t 1, ty] t-t ' l-gi t,t,t. ,!lt j t o t, jjit, t,I.t t, i,l,, .l '. ',t t..'. iiiI tt 1 *' t -i 4 f i 4- -I
- t Wi 4-
.-{ H I i f-I, ii, i,: t, t,ii, iii! i:j t, i i, li i i i,,, ti, t i vi j t t-- .}} ,J,,, ,t ,? t , t,,t f.i6 ,,f ,,Il
- 6 I
1,.,,' ..i q g ti 3'. 't;1r i., .{;. ,t t' 1 i it a , i .t ii.r i.
- .J i
{', ti Aj-
- 4j
-i4 1.
- ll, il iiii i
ij,t .ji. {i;i t e .r } 4 ,i t.,. I ,i 'IlI t,I i !, t,1 + t, t t t '
- 1 t,
- t. -
t t + t, 4... ,t il3 l' i, . i 1
- tt1, 0
i t t-t I4.. .i. ,, N0111NId30 QV07 - i' t t 8,i09 ( .t .ti p+- .j, i j f
- .t;;
t 44 i.4 .i ,i.
- 4li,
,,.i ,1 t r,. i jit ,ti, li. NJGMU37 l jt.. S' { T f ..t it i 11 i iN! ;,j1 i ET l':' 'tii 1 WW,NMd k,% M -H-i,i., ii ,,i ,i;i i i ili i{i i+it ,i %yft-t-i
- iii, i
it, ,l, i,,,!, m.'.tj
- i.,,.
i 1 1 1 V, ,t 'i, It,1 f ?, t t, i .,,1# i r.., 3,, t f }i -h -h f,f* .,,i t,, t,fif ki 14k. +- Il ', ,8 ',l,, , i t t ti i, t.,i
- t. t.,.,,
t,,t,l ,, t,.',I. t I, 1, , t,, t, f, j., t, t f t-t ,l, !, { -t, t, t ; ,,+, t, i 1 t-4,.,
- U U N08 N01108181SIQ lV11VdS 7VIIN383dWA3813 TlVM T13MABQ,
_4p } 0 tut -t t - 9 [-f 6-O'/ 11'/ 31111Dici i f F LAN SI 83el f.ft MA ti i I I McWS .a a =aisa ie es n s.. <Ases+4a ia .2 wi.aa Nhmmrd B d
F IGl'RE 315.40-10 CONTAINMENT CIRCUMFERENTIAL SPATIAL DISTRIBUTION RUN 12 t 4. Q_ gg g g g__ - - - + - - - - - -.. - E , LOAD Df.FINITION i
- ' - -- - - * - - r ~ 1 - -- ~~ ~ I O 04
.+ ll i I o. (,) .(). i (.) 0.02 - .,i g, i 9 g 3 0.* ??.. O t r g ji 4 o 3 .. s e i f 7.1;;, .si <y , s .i i;g. .) ,4 y Q
- ?e s
2 f+... 3 .. + -O
- . { j 6, { { ! 5-{!;4! yy L ;
9 1 2 p 4 ! 'i + t' ElRrtINFERENTlAL DISTANCE F00ti VeHT 4 CFT) i.'-
- 1..t :
i i 44 : 1., 4 ,t +.
- i..
9 1.4 } }, ,e i .+, .i 4 j{6 { j . e 9 i {l+. 6 4, 4, 4 + 9 l. p+s. {. .4 4 ? p p i
~ m I { 1 e ~ l FIGUltE 3B.40-11 OMPLIF11 D Bl:SPfliPT SPITIMUt1 i St.PILMlH Il 01. I'sfIl l9_ = COMPARIS0tl 0F GESSAR 11 LOAD DEFINITI0fl I i Af1D PSTF FULL SCALE TEST DATA Off MSEfMT l ,THO PERCEf1T DAMPING, I2 FT F5 001 DRTWELL 15-l}_ E L j y g_ GESSAR II LOCAL - ] ag LOAD DEFINITION ) b / i )( / h / 6- / f l / \\ \\ / s 3 -- / 7-7 l Y %/ / 999 52, pyy g; j TEST DATA I 0-10 o 10 i 10 2 10 0 1 FREQUENCY ( HZ ) fr-T
c s ~( a 's I I ~c l f l l I l C t i.. - d_,< '1 L 7 Z< 3Q E k-sM JZ c'= W h i O W - {- C ) J~ 3 2 Iv =L i U l 2O . !,] MO cu i f ~d Wo I 8 \\ l C _g-i ca \\ i i y bgs 3 i i \\i i N ~ .di I i N, i ) N,- z-6 e b5 l I l\\ l \\ a u =- f\\ f (s l \\ 7 CO l N I 2 U \\ / Z s -m as
- 3
,[ C L' =W 7 r, A C N j[ .= c_ m<_ (. m emt l u- .c; i Uv = LJO O I O! r C I =- y zt= Il ~ % O t; 8 i i ) i M La. U ^ i O l i i \\i \\ _re ~ L-x T Y I s \\s \\ ~ b_ i rm, l l c O=2 I I l l l\\ l t u<- I I I I IN t l \\ l \\ \\ l e o' I 4 6 i
- Q o
o o n a w w = m a. o O_ k 3SN0dS3B 03I3IM;V l ] ~_,,.,,._.w-,--
~. n n j l'I GU RE 311.40-13 HiIPL IF ((ll nar nibC SPEC 1f.1 filt.II:.I 19. 1941 l2 COMPARISON OF GESSAR !! LOAD DEFINITION l AND PSTF FULL SCALE TEST DATA UN CONTAINMrNT i. TWO PERCENT DAMPlNG j 10-- .I. i i I i 1 8_.- 1 W I E i e ( i y 6 q-j e / 3 i E l { N / 11-GESSAR II LOCAL -- / \\ ' LOAD DEFINITION p I j / l f CNua s2, RUNIL / ~ / / / %[ MAX. TEST DATA / - H~, -/b l 2 AT 2 FT ELEVATION e / 1 / / 0 10 o 10 i 10 2 10.' FREQUENCY I HZ l r_i _ p [ ic j
~ i i i E 6 i O L i i i ~- I I i ( Ei l i I j j { l l i I l = I l f = l I i I C: i I N< i z< DC x .\\ p g s C*) l sw N* j C H 2 c L l I _; A iO =2y C-1 I I i T t i ~* 5 I w09 I i \\ l J %DT l l I {
- {
N -v-t E. = I I I \\ l = =- ec l I s-1 c-c- q = U l J~ g 3-i 7 -6 7 y f 3 ~$ I ~ =~ l { i O 'I n 2 - ~ ~ L! eO { s = ~ Qw: e< / CC o r e:= l j '9 =J u_ L-6 e<- 5 id LU1 l i en -- CeE l p ( -- p
- a _
u.
- = ;
5,. O C =3 l ,_l a- [Z 5 ' Z. T N i N O i N i i 3 eau __ --= =vu t Xi i \\ G fI i K i \\ m E== i i l \\ ) = v<- i
==x I l l l l \\! s kl I I i 1 i j l\\ l l l l l l l, l l l o , O Ie C C C =- N O *-* 3S.';0dS3E G3IlI M Y e P r -m-- -m-- me,---,ewr-a,-,n.,e-,.,ne,- - - -m,-mm,,,n,~ e,...,,,,...--_,-..,,,vn.,- s, ,-n-w s,-,,-
Question 3B.41 1 No comparisons c f measurement vs design have been presented for the spatial variation of peak pulse pressure on the weir annulus during the chugging phase. This comparison should be supplied using maximum measured values at 11 available j pressure sensor locations. (See sketch of page 5-23 of NEDE-21853-P). Any 1 exceedance should be justified in a manner similar to that suggested in Question 3B.40 above. Response 3B.41 Figures 3B.41-1 through 4 presents a comparison of the load definition weir wall 1 spatial distribution and the measured pressure data from test series 5707 runs 1 and 2 which are the basis for the weir wall load definition. Examinat ion of these figures reveals that only one chug, number 22 of run 1, exceeded the weir 4 wall spatial load definition at one location by one percent. The ARS of chug number 22 is presented in Figure 3B.41-5 and shows that the measured chug ARS exceeds the load definition ARS in the 30 to 40 Hz range. As discussed in the response to Question 3B.39, the fact that measured chug exceeds the load defini-tion in the 30 to 40 Hz range does not have any impact on the weir wall structural integrity. This frequency excursion is not significant because ther-mal stresses instead of pressure stresses are the bounding load condition for i the weir wall. Examination of Figure 3B.41-1 reveals that the spatial load definition bounds integrated pressure field of chug 22. Therefore, the weir wall load definition including spatial distribution is conservative for design. l l l -N. prw ye y-=-w "M "www-w--v--ww y - g-e+-.-ww--.=v v 71-gy ep-y.y~u pe-4 --e-mw--e +-e ..m -%we
FIGURE 3B.41-1 l } I i BEIR WALL SPATIAL DISTRIBUTION RUNJ i I' i j i + i l . 1_-3.. I t. _- LOAD DEFINITION 2 fcHUG 2: i i-EGUITG$D O O O i n t .-- M -Q [ C[: $ C [: O C @I ^ @C +v 4 CH i M y -__ _l - IE> O__GID O O_D.O_ _O.O._O O.G_D. O._ .__ -_uG 2 2 _ _C H u G Z Z. { HUG:2 -2 h.I_ - E. - _. EEi> - -O 5 -3 i .w. -4 [. <:~ >. ~ _ _.W' D - - _ _. _ - - -....-- -.-_ -. _ - - _. _ _ -. _. _ - - ~ aD .]< 8 1 - i i -{< I i 1 i 1 ^L a i n 4 t e 4 e i i i ) i i t 1 i i i i i i - } i' i -Il 1 i i O 03 CZ 03 64, c.S 04
- c. 7 c.8, 4.9 l0 J-4 j
1 A j I u. 2 I 3 4 j j . _ J, __.. _ _ _ _.__! t 4 j._ j j _ I i. _._4_ 4 ,i 4 t____ t l .-.---t r j 1 { 4 { I t { 1-1 4 _ q__ y 1 ._____p__.__. _. _. _ -4 _. _, __. g __}. .m .e e g W
- _w
. =. * * . _.. _ - __.*h. _. ^ .___.3____.- _ = _ _ _ .__ j__ .___j
s 1 a 4 f l l l 1 f i t J l l f 1 4 I i J + I i A l } . g 2 I i i O + m i = u. +> i m 4-a .. _ _ y ~ h l M O< cz: j = o C_ o 1 Q l x g 8 l E.A l 9 C { 9 a OO O GD @ GD1tD @M N$ i z g g e e e m ee can n n if, s J { g g o, m e-- m =e .3_ _@ c m. _O _G a.n_._ g* g m __ _ o 20.o a aw u a oe c e s e somm -N & 1 = .r u V j -J$ d a u a - -t x w m 3 I l } l 4 0 el p I 9 6 e,. t
- 3.. -
iu l i j 5.' {' 1.-, 4__._._-_.. _ _ { j g g s. l l ---k-h N %) 9h fD N Q
=
~ *[~k 6 d d d 6 d D f j ..I_~ l l T __.._...t. i j i .. _- _t (19d E+== 1 ) W0SW- 097I'WJb, N. _4 1 i .___p .. _ }. _. _ .. _ l ..._ _...;.._4 .- 2 ._i ..j_.. .4 ._= .= a L
i i I. l l l 3 t I 7 1 ( } i t I i 4 .i l t l t .j i. p___. 1, ,f _. l ( t"-- t i i j 4 i .e i z i t 8 .r O i I a I = - .g i w ,1 _v = .m W q w 8 : 2 l = __ut o o c o a o ..+ g = = l I 3 a y = = a O C@ O OO OO2 GM Nd y f._. ~Z \\ -p ~ O O O O O O COnWTT n l l s 7G O ( i Q __ i C@ O' O O GO CO QD @
- z E "
d! l 0 90N N 2 j fC ~ i i d xG 1 4 ~ 6 i s { 3 i i 6 ( Z { i f t i w iJ 3 a i I t i i i f I i l } f { i + j g L t i 4 q s .t t f f t t t f 9 [. ~ L i l j p 9 j -_. - -..-_ _ [* _ _ J_ { l'_ 4__ N - _ _n -n o - n_
- .--.o.
o-e e o a m A. -.:.. g - - -
- -- '__-'.._C_
. 6 _ d ~. <5 ~ d _. 6..' ( . ti y. I-- [ ___ } } }_ _ _ (Iql E+ = I) 9&SSMd ~09H.1WS. N --._. _. _- r-( g_._ __ {_ _ _ - [ _ _ _ J __ _. _ _ _._.p.._ l .___.L_._.. i.&, wh6 ..% w W 8""'*** .ee.A#.pe_._. i-.a%e ++ .O 4+w M 7_ .. _.. = =. -.s_4...% w
ric"RE _3LAl-4 a l WEIR WALL SPATIAL DISTRIBUTION RUN 2 l 4 I r n 1 l i i 4 1 i 4 4 t j j_ 7 ~ _ _ _ a 4 j_ _ _Q. t i e i j l 4 .4_._ i i j 4 . -. _- j _ - _.3 t i i 1 j 4 i LCAD DEFINITION l 2 i F l< ilD @ O CID GD O 00 ,CO i i [-c -- ::0 CEO OO-OC = 3 v 4 5 0 -l< sD_CE. _O C_ C_O__O __@ O_O__ O g 5 ~ iecascoana e ee 3 I_ 1 h "I ~ 2 -- S - H ~ ~E c> u -7< ,i W I t i J _g. i 4 l 1 1 t t 1 i t i -b l l l l i [ i o i 1 i i -;e-l { 4 4 f i t i i j i i t j [ -- i I J i 1 ' l -fi 4 O W1 C 03 0:!,4 c.S CA d7 - c.8' o.9 Lo y --. 'T 3 ]' l + m.- -. ~ ~ i ~~~ .. _..' _ ~ ~ ~ ~ o 1, _. ___. 1 j 1' l I l _ __ __ _ a ____4 1 9 .j. q ___j. __q. j ._ _.J._.__ __4 .. _ _. = _, _.. { 3 3 3 ___3 _ _. _..____3_ j ._.__J____- 1 _.4 _ _ _.. j. _ - J. _ _ _. j _ _ _ -,... - - _. . _ -___.._ ~_ _. _ _ _ _ _ - _. -,.__ 4 _ _ __ _, r_._ __a__ f - - _ _ - _ _.__j-.._ __.-4
_n 170 inn NOff : 2 0% o AWPl84G => /\\ h no / \\ tocat taan g j or v.n s.on m U Y a } E v/ c..... e ....,~,,,, o m n I so 4 ,s' 1 I I I I l I I I I t 10 trv) toor) FmFOtMNCY fHel I .' Figure 3B.41-5 Amp 11 fled Renponne Spectrum, Comparinon of Load Definition and PSTP Tent Series 5707 Data on Weir Annulus 8 9
QUESTION 3B.42 The M:nte Carlo sirslations prepared in response to our first round question nu-ter 15 r.vice aceauate justification for the " global" chugging load de" nition on the wette perimeter for a symmetric loading case. However, they do n t con:letely justify the absence of any asymretric chugging load. specificati0n. There are three reasons for this conclusion: a' Ine : :aris: e-tornin; n: ents witn :ne as)Tretri; p::I snell re;;: 3e is at:r;;<iate ce.1, if botn 1: ads are treated tne sa e in l l 1:50 Cc5bi"atic"s. We kn:w that Chsgging loads are Combined with seis-i; ar.d single SEf. ICads. We have not been able to ascertain fr:- tre ICL; i' :ne as - etric 00:1 swell load is siri7arly treated, t: The targins betasen tre tw: ica:s are t:: s a'.I to decide ahether the I as_c stric p::1 selli 1:ad is b andin; on tne basis on only 22 trials. l Pa. t:re M:*te Carl: tria's w:f d te neeced to ha e su"icie-; a:s.'a :e ina: t s s .e case. l l ~ne se siti.it;
- ins Monte Carlo Sitalati r.s 10 ch0 ice Cf time windon has net been exa-ine: sufficiently. A large-rarge (say, E0 to 500 ms) rather than the 25: anc 300 es runs sh:ulc be tonsideres.
l If as a result of tre studies suggested above it transpires that the asymmetri; po:1 swell leads are n:t bouncing, it may still be passible to sh:- that the design load specification is adeauate. Inis coulc be accomplished by showing i a comparison of actual containment res0cnse, at sore key locations, to the i desie.n load, on the one hand, and that obtained from several (hic.h non-i I exceedance) Monte Car 10 trials where overtu-ning moments exhibit peak
i i i i i Question 3B.42 - Continued spectral values in critical frequency ranges. The procedure suggested here is similar to the SRVA methodology described in Attacnment M of the ICLR but dif fers in that the selected trials are not used for design evaluation they are utilized to show that the design chugging load can bound them in terms of the excitation imparted to a standard Mark III Containment. 3B.42a Response i Figures 1 and 2 show ARS comparisons of asymmetric pool swell, single safety I relief valve bubble pressure, and asymmetric chugging x and y moments. It can be shown in Figures 1 and 2 that asymmetric pool swell horizontal amplified i response spectra, ARS, bounds the combination asymmetric chugging and the 5 asymmetric single valve SRV ARS's. Examination of this figure reveals that i there is significant margin between the asymmetric pool swell ARS and asymmetric chugging and SRV combination ARS. This margin has an adequate allowance for seismic loads included in load combinations. 1 } i 3B.42b Response i Figures 3B.42-1 and 2 sho is that the load definition asymmetric pool swell x'and y I moment amplified response spectra, ARS, bound the asymactric chugging moments by factor of 2. The asymmetric chugging ARS's were generated by using: 1) the l worst asynchronous time window, 80 ms, from part c of this question, 2) 59 Monte Carlo trials which represent a high confidence level (95-95) and is number of trials recommended in Attachment N to Appendix 3B for SRV Monte Carlo analysis, 3) and a random chugging amplitude for each vent which was varied i between the maximum and mean amplitude. i { In order to verify that the 80 ms asynchronous time window is conservative, five i Monte Carlo trials were performed (see Figures 3 and 4) using a 200 ms asychronous time window and a random chugging amplitude for each vent. Examination of Figures 3B.42-1 through 4 indicates the 80 ms asychronous time window is conservative. The conservatism of the random amplitude chug in each vent assumption was verified f by an ARS (see Figures 3B.42-5 and 6) which placed the higher amplitude chugs on one ~ ~
Question 3B.42 - Continued side of the drywell and lower amplitude chugs on other side of the drywell. Examination of Figures 33.42-1, 2, 5 and 6 indicates that placing random amplitude chugs in each vent is a conservative assumption. In conclusion, this conservative analysis including the high 95-95% confidence level demonstrates that asy=matric chugging should not be used for containment structural evaluation since it is bounded by asymmetric pool swell by a factor of two. 3B.42c Response Figures 3B.42-7 and 8 show the sensitivity of asynchronous time window on asy= metric chugging x and y moment ARS's. Examination of these figures reveals the maximum moments are produced by an 80 ms time window. The asynchronous time window asymmetric chugging Mante Carlo simulations were generated by 22 trials by each time window and by assuming? cach vent chugs at the 95-95% confidence level a=plitude. As discussed in the part b response, the worse asynchronous time window, 80 ms, asymmetric chugging x and y moments are bounded by the load definition asymmetric pool swell overturning moments. Therefore, the load definition asymmetric pool I swell should be used for containment structural evaluations instead of asymmetric l chugging.
FIGURE 3B.42-1 T AMPLIFIE0 RESPONSE SPECTRUM AUGUST 29. 1981 720 Sl lPFHE! S1 ON '00L X l 90*1 -MOMEb T AN ILY':ilS 660 du nuntunnses huh Ut Ms iI ME7mU lH
- 2. 0PEFCENT
- )AMPING OF RAN00F DRY 4 ell DATf 600 m
[ \\ 540 ASYMMETRIC LOADING 0F 480 P0OL SWELL LOAD DEF. 2.0% DAMPING WITH d P0OL SWELL ATTENUATION 8 420 A m 360 .,err v g is 300 4 ^ -240 f ^ l SINGLE SRV 180 FIRST ACTUATION f Y YL .I b uwsdihhhbEhkh / 8 = gegggwn~ j / 0 ~ i 10 o 10 i 10 2 to a FREQUENCY [ HZ )
FIGURE 3B.42-2 AMPLIFIED RESPONSE SPECTRUM RUGUST 29. 1981 720 SUPPlsELSl Oti 300L YI 61 -MGMEt, T AN 'lL YSIS 660 vmu 0rtus wiin ut Ms n srmnuuw
- 2. 0 PEI CENL JAMPING OF RAND 0F DRY 4 ELL Di1TF 600
[ h 540 ASYMMETRIC LOADING 0F ( L180 -~ POOL SWELL LOAD DEF. \\ 2.0% DAMPING WITH P0OL SWELL ATTENUATION \\ a h L120 %\\ b, e A m 360 -,-,ar g 2 300 u SINGLE SRV FIRST Ar.TUATION [ 180 H &Mr s a=ema'euw- ~ - 0 10 o 10 i 10 2 10 3 FREQUEh,CT ( HZ ) 4 e e
r m I O 8 l i ,nW Cw=rW w c_ uO j Y ') i 1 l l-0 i N ~ _/N O {_ i i i I R ) i I I I( M ~~* l I ~ ~I b i h i l} <<(( _Eb I k I k [d) x =u i i i c i i l ! wi\\ - t a a L' O I \\ Ii U z 5 5. I V U L;.) ar C_ OPC 7 )' s LLJ O C w 2_ 3 D r7= 0 u, E L_ @ l (l } O u
- = a to
= 3 \\... I s z O F w l \\. ll_ O, 79w .1 c o x LO De L1J x j$ E N z e C J hO hh S ~ o ~ U
- 2
/ l i j.i n O ~ [ EOE l l I I I IlI ~ 5iO I I I I I \\FC I i \\ll d l 95b l l -i E l 55$ l / l l \\ cFd /l 8 i. m I o o ( L c. I Cw za=z e c.r. ci i <<3-oo waCD zz U J a--* w _a c. r crwrs -3<< wmC E w && >- o o w Ao 3 < a. m m I f i i i i O O O O O O O O O O O O O O~ DJ LO O O C3 DJ LO O r CD OJ LO LO LO LD r 7 (O CO GJ -4 3SS0dS3'd C3IiIMY J
FIGURE 3B.42-4 AMPLIFIED RLSPONSE SPECTRUM h. SEPTEMBER 23. 1981 720 SUPPitEESIGN 300L Y I O* 1 MOMENT R ifiL 1 $1S 660 anyM DDENITUETF7E 7TENFF 50 K 5 ~RMPING OF RANDOF RMPLITUDE Of TF 2.0 PEFCENI ) 200 ms TIME WINDOW S ti O - 480 ASYMMETRIC LOADING OF P0OL SWELL LOAD DEF. EN 2.0% DAMPING WITH POOL k E 420--- SWELL ATTENUATION 360 d^A rme; ec v g e d 300 240-180 120 w kk l 60 sl I gp w _gh 2b#MD ~~ ~ p~ " 0 10 o 10 1 10 2 10 3 FREQUENCY ( HZ ) ..-/l
FIGURE 3B.42-5 AMPLIFIED RESPONSE SPECTRUM SEPTEMBER 11. 1981 720-ONE ini nt UF RANDOMLY F ORCED PRES ;; tin! I 660 rT-7 rOnE li a.u em cEnr1annar --l 600 + ~~ ~- m 540 j 480-- ASYMMETRIC LOADING 0F POOL SWELL LOAD DEF. g 2.0% DAMPING WITH P0OL m g 420- - SWELL ATTENUATION ~ 360 O^^"e ~ W d 300 4 240 1 180 120 s 60 - r -w 0 10 o 10 i 10 2 10 3 FREQUENCY [ HZ l
i 3 1 8 0 9 1 1 1 1 R E B M E T P E S \\/ -r ^ = 2
- ii jl j
i 0 1 N 1 fl \\ ) y Z b H M U { T l n R RI /, C i L l Y S t W n S 1 w E E C H P ' i N E A I D-U E I E Q 6 E M S R E E 2 N O C R 4 O I i l F T B P Y 1 3 L - S U. M E E O R R D 2 U N G A D R l i I I i E F l E f 0 l f 1 L i F O T t F P lRT n / M n E r NOr L F O O 0 FP GE NDHN I TO DDI I AAWT OO A LLGU NN CLI E I LPT REMT, TWAA ESD M L ML%L YO0.W E S0 AP2S o 0 0 0 O 0 0 0 0 0 0 0 0 01 6 0 L 8 2 6 0 I 1L 8 2 6 6 6 S 3 3 2 1 1 t L 1 1 wu $3e 8 mag ) 1 1 l l
FIGURE 3B.42-7 AMPLiFIE0 RESPONSE SPECTRUM I j F40 GUST 27. 1961 422 i .i. t_ i.._ i i i i SUPPRESSION Po0L X(90*) MoMcNT ~
- 396, ENVELOP OF 22 RANDON CriUG START TIME CASES. FOR 40 VENT SY5 TEM WITH 360 95-75 'LlHIT DRYWELL D6TA A5 INPUT FORCING FUNCTION 324 288 80 ons TtHE WINDOW 2o0 ms me Winoow 252 g
4o0 ms TINE WlMDOW 8 -~ ~ ~ ~ ~~ Soo ms TIMC wtuDow 0; 216 a 180 o 4 - = l fZ3 u/ 3 9 /A( y 1y h r 108 g 72 36 i 0 10 8 10 2 10 : FREQUENCY I HZ )
1 FIGURE 3B.42-8 AMPLIFIED RESPONSE SPECTRUM j AUGUST 27. 1981 ti?2 ^ SUPPRESS 104 Pool YCo') MOMENT ENVELOP OF.22 RANDOM c106 START 396,- TlHe CASES FOR 40 VENT Sy5 TEM WITH 95-95 LIMIT DRYWELL DATA 45. INPUT FORCING FUMCTlON 80 ms TIME wiNooW 324 - 200 rns TINE WINDOW 400 nas TIME Wl20W I 288
Soo m3 Tine WINDOW l'
252 e 216-b /\\ /A i L l 1" / } '/% N 9 [ YL l/1}EL,)Wh jQ7-1 $,f4 108 w v vv i d 72 36 l 0 l 10 i 10 2 10 s FREQUENCY I HZ ) f ,S-
QUESTION 3B.43 We require additional justification for the chug source strength used for design. The response to first round question No. 31(d) does not appear adequate. The chug strength selected fails to bound all of the pressure data in the " full"- scale tests. Moreover, the claim that 2.65 psi " bounds 100% of the peak contain-ment wall chugging pressures at the 2-ft elevation" is contradicted by the data shown in Figure 5-36 of NEDE-21853-P. Also, the uncertainity in these measure-cents is stated to be about 12% (of design) inplying a possible increment to bound of about another 0.4 psi. Furthermore, the statistical nature of chugging may require further extrapolation of the data beyond the highest observed chug to arrive at a sufficiently low non-exceedance level. For these reasons, we do not c nsider the selected strength (ip r ) = 2.53 psi-ft to be adequate. An oo appropriate modification of this design load will be required. RESPONSE TO QUESTION 35.43 The source strength selected for use in definition of the chugging load on sub- =crged st ructures is adequate for design. This will be demonstrated by assessing conservative factors in the submerged structure load conditions, not accounted for in the GESSAR methodology. The major conservatism of the GESSAR methodology is that the submerged structure load accounts only for the absolute pressure on the swept area of the structure, not the integrated pre sure around the entire structure. Theorettically, the pressure field on a submerged structure member in a finite rectangular pool due to a single thug source may be expressed as N N N b)
- E e
- w.. AN sic #
2. (1) 2 Po F = s c tst 2 r oA ot %:-N is M is N E bb* u a
RESPONSE TO QUESTION 3B.43 - Continued while the pressure fiele used in GESSA:, is N f) rd 'Y () P = [ Z. W i:.y s;n a..)~ s 4:-w J - ('JE. u t "R Comparison of these two equations shows that the ( C Lt em is not inclu M in the GESSAR methodology. This term is a result of the integration of the pressure field around the submerged structure. It is possible to show that the conservatism in the GESSAR methodology due to this integration of the pressure gradient is such the design source nearly bounds the test data. The GESSAR methodology specifies the chugging acoustic spike to be triangular in shape with a 2 millisecond base width. If the acoustic speed locally is 4000 ft/sec, the wave length of the pulse is 8 f eet. For a structure submerged in this pulse. it is the pressure gradient, acting over the structure diameter which is important rather than the absolute magnitude of the pressure pulse. Figure 3B.43-1 illustrates a 2 ft diameter scructure submerged it, a pressure pulse of 8 ft wave length. As may be seen, the force on the structure is only a fraction of the absolute pressure of the pulse. Integration of the triangular pressure ~ pulse around a circular cross section structure results in the ('. R ) term of C5 equation 1. For the 2 ft diameter structure y p, (fr) (I ft) ( 1/2.9 Ct. t (4000 ft/sec) (2 x 103 = sec) This factor, if it had been in the GESSAR methodology, would reduce the sub-merged structures load by a factor of 2.54. This same factor of conservatism holds so long as the wave length is four times the structure diameter (a 1 f t diameter structure in a 2 msec pulse with an acoustic speed of 2000 f t/sec, for example). Smaller structures are even more conservative.
RESPONSE TO QUESTION 3B.43 - Continued Applying this factor to the source of 2.53 psi-f t g *ves an ef f ective source strength of 6.43 psi-ft. Using the GESSAR methodology, a source strength of 6.43 psi-ft is equivalent to a wall pressure of 3.2 psid. The maximum chug seen in runs 11 and 12 of PSTF Series 5707 was 3.4 psi *, and only this one chug (out of a total of 113) is not bounded by 3.2 psi. Hence, the effective GESSAR design source strength is at 99.9% nonexceedance probability level. Based on the above discussion, the GESSAR methodology for use in definition of the chugging load on submerged structure is very conservative. The current specified source strength, when applied to the GESSAR methodology, is well within the conservatism inherent in the equation, and therefore, a further increase the conservatism by increasing the scurce strength is unnecessary.
- This is not in agreement with Figure 5-36 of NEDE-21853-P. A thorough review of the test data has shown the maximum pressure at the 2 ft elevation to be 3.4 psi.
t f 3 + 2 Jt l e a l I j i l i i g i d ( I I w Strafste y PL l 8 +c = = l 1 D ; s +a nce. 1 FIGURE 3E.43-1 - LONG WAVE LENGTH PRESSUDE PULSE ACTING ON A SUBMERGED STRUCTURE l
.~ .. - -. - _, -~ - s I t I t PSTF roof impact pressure data for: ) l' 5 5801 - 1, 5, 6, 9, 13, 15, 18 5502 - 1, 2, 3 f j 5803 - 2 1 5806 - 1, 3, 4, 5, 7, 8, 10, 11, 12 i l 4 j Plots of roof pressure time histories for all of the tests except 5806-1 4 I are attached. The roof icpact data are unavailable for 5806-1. j j The locations of the pressure transducers in Test Series 5801, 5802 and r i 5S03 are shown in Figure 1. The transducer locations for 5806 are shown 4 J in Figure 2. Each time point is plotted in the 5801, 5802 and 5803 data. i i The reference time is the start of the blowdown. The Series 5806 impact I t i pressure data was recorded via an analog systen, so much more data was obtained. The attached 5806 pressure histories are actually plotted every l 10 timesteps (.l.2 msec). Reference time for this series is the ti=c i l that the level probe which was 5 feet from the dryvell wall and 31 feet f above the pool floor (x5,Z31) first becauc wet. For both test series, t negative pressures were plotted as zero. i These plots are preliminary and unverified. i 4 l l l t I t i
- M*Nym p *= ewes --wm New-*me p wg
~-ee=e'r--
7 = O nu I I Q ?nb l-O 3*o l5 N o C &l T C $ o4r s
- a e'
us %) N "J i C 3 2 N < O h" (Oh w t. o s Iv$ ~ O$D 2r H s) O
- 7'*
O Roof l l..- M e 4 TN 3 3 810 313 ( 1 V W 3 N y r _a
phv Ih l s Ongsur O v i" new i O x a.o neo O v7 5 9" O n o 57" Q / 3.4 y/.2.6 a 2 0 ' 4- <v 3 a Y d 2 t b s 3 8$ F t o E-odc e H ed \\) 60 i U I B13 313 (> l y y g pi g e
e 'I
- A i
N I e-I 2 s j ~ \\ m = 1" s OO% % % 4 s % dq%wD D %/ q k R L. Sc \\' s ~ ~ c~ i m u.- s kN*% \\ \\ l. s,, >- w E 4 4 l i, h-ek ~ ,N \\ ' \\\\ m ? 8 ~ 'i e m / I. t. s is ~ Q s. ~ 4 C L t... / %.~ ) L-.1 y~ '~ ,} f;;I C3 L,; '. [d [\\ % C'*.. .: *) i t :.. L-n ca .t o tad 1~ c) 1
- +
[ L_ - ,t.. o ( I C" L. '. ~. u. h-I i h> 4 y o O s ~ (l$) ) e)VUd)l A* T ~
O i-5 gi.g f - 4 s t / Accr sov7H 2 Acer 2mo 3 acor 3(c 4 C30r 474 l 1 5 5 'ct r f./ : A M I S 9 Lbw-I 1, i e t* I' ,I ,g,- pr- -[.r,.. 6s,~~,
- t,.
i k, ~ I i N I k l e h .\\ o. ? t; j ; 4, 'I 1 k' -7 e t i l i t 3 l p 'e P ~( ,. n - o. y L t ,.i i t t \\ \\ l r 1 ,1 1I ) 'l. V s. ]* e 1 b l i u 1 A a , 'b i Ik [ e' 7 ? a i <f r' l -l {\\! I l i l. j]" / l u \\' 1,d -- \\,f r / l ,J/ d%d& t5 4;T h; l 'N " m ' ~. I 'l :b, ', w
- gf.
D." s n' l I S )i /, l t 3. _ f _n'g /I 5 t /* .( ] g' ,i / / 1 l 3 f bi l11 t 1 \\. Ljp? W@. ~ .r i j t' i , :\\ a e ,A = ( /4
- t. $
l l 1 r: : i l l l
A A A JLL lffbfYI f- , 0;. y - (, / fo0F SouT4 I 2 roar xno of, b Opf 3rp 4 PocF 0 IM 24 5 fear NotiH O ,; n x. U 8 /* ).. i '. ! 5 i i li - G i' 41 1. I '
- l t,',il F:',',1 ! r s i.. i l j i [ g l i gg N OT T U U E U3ill) IN DL.'.+
2 6 if, 3 k, i tz '. \\ t S t 6 l is' /N 3 ! l' ix)Y $Y x.,n m;~ l , A_vx_L- _jf'*y -JTx; ^ ~ ~ p-W- ,s. in in n.. Tks e (5r c )
Z o 7I \\ Is 3 iT h 83 'b$ 7 $ z 4 "n s 0 W 0 0 D $n3 i I M oC eg h M l l; ~ a.,, s, e ) { /'4 \\ ^ ~ \\ h i / \\ l ~ c u S h:3 j; v 3 M ') is \\ 5.f '3 k II h.3 K C-, e.:
- c..
3 ~) C. ~
- .3 C..,
L E';3 \\ g.t, L'i:) [(j: II ,q w w ~ / 3% 1 4 l (( il ] \\ n '\\ m T o s ~ u w q ~ ~ ( > T c/ ) 2 ) **i 5 2) ) da3
M % l N ~, / d / 1 e -J l = 1 L oot
- k
,7' O A A 2,R ex o ~ W L. % % Q n s e u ( QQ k k n \\ ,a' N D N N s' /i
- 1 f
w p g t j 1 \\. 3 l \\: ~ c ^ v .e ,4 y f, h l%
- W G
~ n s Is. ~ c. L
- , - ~ _
1 c fj-
- y L
g; { p ~.. -.= r u. l C - [ ' s' 4 c=..
- - u+
C. .~ l 4 fai= [ l ',,) . u e a j. l (.. L-Y / s lt' M t / h. I N N 8 = T ty g N (:5c! D' ' % ) AD 7
pQp4isiv.hannv n n FRELU. ':',- ri.,'c ugy:R!FIFD ,----+3 W8 3 L.di.'LL'$liQhg'sjg,,-l - r _.j-i j $ 0 $ c~ S:Ck 1 1 KODr NC 3 to: r -KD 4 A00r did [ har /4 A * ~ ,f l l I
- )
.i 6 I eN l I I i i i l I. a \\ 3 4 ~ 4 t1 s ~i- ) I i. i I 1 r. k tJ t j f ' ja g r +. i i /t ' s A,' s 1' ,m.?, 1 ,t j \\ I, 1 s 1 ,k*%g ~i. / [I-p' w 4 f* ,i* ,.+ b r 'y^ - ~ 's! .f.:,. e g il, .y v "' T r 3. A7./, ,/~~
- /y
- 4
. 44,i g/ J { '. l '.
- j.' f Q/q a
\\ q--N- - - - tj -,rs -..J 4 -,.1 g \\ 7 ? I;l L / \\ ,.( I ' u ',,1,[ '\\ l' % Ff f f I.5 g,% s1 /( l / Io n' 2*
flf lll: 1 t. L H J M T f Dp M K e p Nt 7 C i / 0$ 234 N l ffrff o a o o 0 o o o 3 [t PgP P 3 23 4 S I ^ a l 5 s i X ) C i E e S n ( e n m s o nG 2 r y x h D IEN IFG I /, RS EE y VD N M N UI T DD I E tA 3 P U s P Y i RE O AB R NO P I T M T IL O ER N P 2 O / I S c 2 0 C YI Q LbH
- i
.-2a 2, _m -____m em__m ..m.u.-24__r_m__s_._1mm.u-m..w___m m.-2u.m__.r_amm e l I t T O N n 't f., ( 4 i 7 I I { 4 J t-L l 3 'b IW oawW o i ? 4 N *O V D C i w -- 1Qk k o o o C U) o C o o
- A44 kW i
i - g og 4 u 20 i en g, e "~ \\ l CO e 7,.. c:: <L pu >3 1 g,., t.3 a: C L' 20 l O,.-.
- r -
\\ c e i o gC .i 9 gN A h ;; u ~ u O N 4 i y o I 4-s n m Y i b i f b p, N b f i i l4 a m 4 T [i i I. ~ i p 4 '4-N w -1 a p w:. .4 o to~ Q 4, i t 4 = -b 4 o a a ~ ~ (1:s) a s n :7C)) d's q
- TW e
br, <s as \\ s 1 1 ba 4 Q4 D 4 a o2r u ~ a 9 D NM D 4 k k
- t. q *
- o , O D Q O g i N %44% A N _m Y t \\, i : s, t. / P s ry ) r, l l 2 a w \\ \\ F" \\('O + o w f o \\= u,- ie -4 } Ad)x f ~ k e. (') \\ D-5 77 Lw lE Pe ?".3 I C" f.Z '\\, e a: =a L.5 ZJ <0 j
- D l
GJ t} tu Q i fC dO l C-7 % f ad io w I rio i' $Z I-c. dN e e c w m e m o u s bi 4 7 4 e4 N C }} f $;$)f dq IS
4 s 4 w N / b$a4. n s ~. '\\ kkg % s, 4
- 3 4 4 tt Ag N
-s , g i. \\ 5y~ vf r, a 1 / r sl% i /,X /. x o ~ l 2 t, N x. s \\e N s u I u ( 4k, k l s- {i N t a C. t: ac:;
- ' ~
W L-- .ta s 4 g j asm C. [ '- 1l' i C.: W 7 l L <I Mw b, "- \\ J u y. L4 C. d - - - ~ - ~. _. \\ N M 4 Y w I SO, hl " b E) g *g
me s ~ c 7-N , y NJ N x3 7 i \\ 1 x \\ ) t T g 4 = = N i s D h e ? uO 4 0 t y 6 x e. <M4T 9 f) S L 1 /i &qkq % ~D E R *O f 4 \\ m~ f !\\ \\ %. Q t ,w [ f N l ^ \\ l "} l / %n Il t. e s a i.y A f .y \\~$ / j ,Il, L g m b( j ,x
- e
~ I v. \\, L= 5 ey 5 $5' Ez N /s'd c 'N C %~ L:.*. 30 L 2 t c-C l,s N o Q 8 i ('S 0 h 3J n s; ajg g,y L.
a- - _ - 4._- -a_- _.am_A_ a _.o Ld___a -cs-4
,.asm-aaa.huA--M JF _4aw.-w.
Aa_ / / k ? s p ~ lessih 2\\~' s a e n wct 9 / O O ~m,g4 ~ ~ E w. A %A% k, i J M / M b a/> is % y '? _f 1 i / ~, / ~ ~ ~, ~ f k -gu; I n i x ts ( ~~ T, V y I / s I-1 '4 !: 7. 4 [ j 6 f i f,,- %e h. f +s I 1 n e e, ~\\. f C.. C~ :; by;y Ud f._,.~ . - #x t & " O / i, L y-E ^% Eh = c. I e 1 t 4 T N g 4 M a w y g ( ' b d, $$ 0 X ? d'r/] . - ~.
~ d m 9 m E e g $g = E v ~nw%C' w / - ~ 'f / "%A ~ ~~ vpp>y < - ~ ' ,f s s E $^ g v. u N' \\ \\ /,, y / s C. 69 ./ N "k ) s .= / Z )$ yN f .e ( Ch. n 5o 8 y' N C-x \\ \\o N a C (15$i g e s
- 7
OEZ .m A. LO EU5 p we g, go 2e v b; ca ..~~- a-kN / / 1 m- >0 C Cw C-C <C ~ h j o N ( N( Er s t' EZ b. k/ ~ /- . t ~- n 'p : f >y. / Ch m. v s,. :: d r O 4 W o t n g)- C 2 p 4 = > w + aT E N u - &< $ '~ f 7 k' .M ^.w ,s,> r s,A N x k y n k X j> \\ , m >u4 ' e i I ( I /s( '8 mef o ? C 350655$ W" .I
J e pe[> 9* U i 4 \\ \\\\ k Y. ( l( ) kr o / l O a O h // 1 14 w. O 's N N .I a % =d %wtCE ( '3 j! 3 ouo o,> ( I o % n o
- 4. ~ =>
~ ~% X% % h N X \\ I { ~ % m wk 9 & \\, t /4 sj ( g .\\ ) s., ,- {,; r l ds ', o r f i l l l \\ 'y o j u ) x / ~ y g u - / u L w i 7 c s e\\ Q 7... ,'S w \\ b l e-cc -.W G -i Y v_ 4 i m ca. C C;,) c - o C,'^ l' w n E-o o. i C. - [. I i C e D C D C M
- e
.$+ 4 %3 a .\\ W s 4 W2 bo ~ t a C to [ H QO h$ w a >. O CD C L'J Z $U CL. &O { n> e'- ,2 = X W v Y L \\ O u o 10 N ~> L NQ! ~ p g N E saw 3h bW EE ( a n [4 uts t w o ouo9w* \\! ?EEC 5% <4 ,. m o s e e C k I } Y Y Y (isd) 3?n S S??/$ c/WZ
m N h% f geas";?? i. R f { 9 -c / n mvw n o. .w h \\ ~,smos Q .f <x v,s? o .x fjt k 2 I I e g l W:' E-l x / n c= -I.. / ht' Q c: ~, ,s ~ x 'N .A t i c N r \\ [ l 4 I / ( g/, '^ N %x ~XA ~ x: o b ~~2 '*o y y a SI' $$ 5 5) } y,.
s W 6 k k i L Ys I i s ) 1 C 'w n Q q l' % ~ y S g p k O s s ' C P 7 4 E A A 4 C C
- O G
, 7 rs W. m z m 'A A A 'a y ~ ... ~ b L e N !.. q y 9 i /' x3 ?Q / / / C. \\ ~ M,, 'f w.. .- %'_L - ,~ w ~ 1 qp / <' . N s j\\ m b n-C 5 f e g; w L. g [ y
- c. '.
i L2 c:n. p_w M; m} C-E ~ /;$ / / v. s c;
- a
/ QT Q'y / m c_ zo ~/ c: W m._ -W NP N O, W .- a C" ma. -M -mv r a S pp r :: I$$ W* , : :/ ' s /
sdA$ 23m h2*q5&3<" a "a nw%i% 7 m o -s> --~ T$ g o e n' 4s-n w w x, x 2-x j 44 >Mo' f l i ? / ~ lf' / ,/ ~ + k !l / Q i hJ O / n M 20 \\ N \\* Cu 1 e r_ e:. F~ CO \\ \\, : k 2 / T LM 7% 'N \\ c~ w <3 \\ i v. t. t, \\ n x 1 c C,a b (*a c u, j } t-
- ,-~
e, i ((d c ^ ,i l -'O A(7 p' / / %2 4 l E / / \\, 1 j c r% / ) / 8 \\N _' \\ ~ ,/ E m ): \\ k b J-Q O u %A s I SA f s C t N (7 g j >>'15 G h J}}