Similarity study of ROSA-III and fist large break cuonterpart tests to BWR large break LOCA

Nuclear Engineering and Design 103 (1987) 223-238 North-Holland, Amsterdam

223

SIMILARITY STUDY OF ROSA-Ill AND FIST LARGE BREAK COUNTERPART T E S T S TO BWR LARGE BREAK LOCA Hiroshige K U M A M A R U , Mitsuhiro S U Z U K I , Taisuke Y O N O M O T O a n d K a n j i T A S A K A Department of Reactor Safety Research, Japan Atomic Energy Research Institute, Tokai-Mura, Naka-Gun, Japan and J.A. F I N D L A Y a n d W.A. S U T H E R L A N D Nuclear Fuel Engineering Department, GeneralElectric Company, San Jose, CA 95125, USA Received 16 June 1986

A large break test in a recirculation pump suction line with the assumption of LPCl-diesel generator failure was conducted at the ROSA-Ill test facility of Japan Atomic Energy Research Institute. A counterpart test was also performed at the FIST test facility of General Electric Company. The objective of the tests was to develop common understanding and interpretation of the controlling thermal-hydraulic phenomena during a large break LOCA of a BWR. The fundamental thermal-hydraulic phenomena in the ROSA-III and FIST tests such as the system pressure, mixture level and fuel rod surface temperatures agreed well. The FIST test had more bundle uncovery than that in ROSA-Ill since lower plenum steam in the FIST test flowed out of the jet pumps when they uncovered allowing more liquid to drain from the bundle. The ROSA-Ill and FIST tests and a BWR counterpart were analyzed with the RELAP5/MOD1 (cycle 018) code. The similarity of the ROSA-Ill and FIST large break tests to a BWR large break LOCA has been confirmed through comparison of calculated results though they are slightly different in details. It is perhaps desirable to reexamine the DNB and interphase drag correlations and the jet pump models used in the code.

1. Introduction A loss-of-coolant accident (LOCA) is an issue of major concern in nuclear reactor safety. Experimental programs with test facilities designed to simulate a BWR system have evaluated and quantified the controlling phenomena during a BWR LOCA. The Japan Atomic Energy Research Institute (JAERI) has performed a wide variety of experiments in the Rig of Safety Assessment (ROSA)-III facility [1]. The ROSAIII facility has four half length electrically heated bundles and is scaled to a 848 bundle BWR/6-251 [2]. The General Electric Company (GE) has also performed various safety related transient experiments in the Full Integral Simulation Test (FIST) facility [3,4]. The FIST facility has one full length electrically heated bundle and is scaled to a 624 bundle BWR/6-218 [4]. The first objective of this study was to develop common understanding and interpretation of the controlling phenomena observed in large break LOCA tests

performed in the two test facilities. They are recirculation line break uncovery (RLU), lower plenum flashing (LPF), core uncovery and refiooding (fuel rod surface dryout and quenching), counter current flow limiting (CCFL) and CCFL break down at the core inlet and outlet, and so on. Large break tests in a recirculation pump suction line (ROSA-III Test 983 [5] and FIST Test 6DBA1B [6]) were chosen for the counterpart test study [7]. Evaluation of the test comparisons consists of three parts: first, the similarity of the test conditions, i.e. the initial and boundary conditions; secondly, the similarity of system responses to the test conditions such as timing of key events, system pressure, liquid level, and fuel rod surface temperature transients; thirdly, the effect of facility difference and different scaling compromises in the two facilities. The second objective of this study was to examine the similarity of the thermal-hydraulic phenomena between ROSA-III and FIST large break tests and a BWR large break LOCA. For this similarity study, at first, the

0 0 2 9 - 5 4 9 3 / 8 7 / $ 0 3 . 5 0 © Elsevier Science Publishers B.V. ( N o r t h - H o l l a n d Physics P u b l i s h i n g Division)

224

H. Kumamaru et al. / Similarity study of R O S A - I l l and F I S T

transfer rates are preserved. The vessel regions, shown in fig. 2, simulate the lower plenum, guide tubes, bundles and bypass, upper plenum, steam separator, downcomer, and steam dome. The ROSA-III facility is volumetrically scaled (2/848) to a BWR/6-251 with 848 fuel bundles. The four half length bundle concept was adopted in order to study thermal-hydraulic interaction among the bundles. The FIST facility is volumetrically scaled (1/624) to a BWR/6-218 with 624 fuel bundles. The full height vessel and single electrically heated bundle provides full scale fluid conditions, velocities, and static heads. The bundles in each facility consist of 62 heater rods and two water rods in an 8 x 8 square array. The rod diameter and spacing are the same as in the reference BWR. The rods are electrically heated with a chopped cosine axial power distribution yielding an axial peaking factor of 1.4. The ROSA-III bundles have a local

ROSA-III Test 983 and FIST Test 6DBAIB were simulated with the RELAP5/MOD1 (Cycle 018) code [8] to examine the capability of the code to calculate large break LOCAs. Then, a BWR counterpart LOCA was simulated using the same analysis methodology as was used in the ROSA-III and FIST analyses. Finally, the similarity between the ROSA-III and FIST large break tests and the BWR large break LOCA was analyzed by comparing calculated results.

2. Experimental facilities and test conditions 2.1. Test facilities

The flow diagrams of the ROSA-III and FIST test facilities are presented in figs. l a and lb. The important BWR system features that govern the mass and energy

ECCS Lines

ADS Line

To Atmosphere

Steam Dome

~

Separater _

C team Line

Feedwater Line

Upper Plenum POOL

Break

I Pressure Vessel

MRP 2

MRPI

Broken Loop

I n t a c t Loop

QOBV

Quick Opening Blowdown Valve

-~:~-

: Orifice

QSV

Quick Shul-off Valve

~

: Valve

MRP

Main Recirculation Pump

Fig. l(a). Flow diagram of ROSA-Ill test facility.

225

H. Kumamaru et aL / Similarity study of ROSA-IH and FIST

~

1

HE~TKED FEEDWATER

FLASHDRUM AND MUFFLER PCV

/

T T T T T

ST",",ML,NE

DRIVE BLOWOOWN LINE

I ~STORAGE

,~p.~L.JFEEOWATER IN LET

S U C T I O N BLOWOOWN LINE

HPCS LPCS

f

J

NO. 2JP ORIVE

LPCI B

No. 2 JP SUCTION

RECIRC. PUMP NO. 2

RECIRC. PUMP No. 1

~LPCI

LPCS

Fig. l(b). Flow diagram of FIST test facility.

peaking factor of 1.1 and the FIST bundle has 1.042 as shown in figs. 3a and 3b, respectively. The same power is supplied to each bundle in ROSA-Ill Test 983. In both the ROSA-Ill and the FIST facilities, the volume of each vessel region is scaled to the corresponding volume in the reference BWR/6. This approach preserves the relative water and steam mass distribution within the vessel which is directly related to the system depressurization, internal flows, and core cooling. The regional volumes of the ROSA-Ill and FIST facilities are compared in table 1. This comparison is made on a per one full size bundle basis since ROSA-Ill is scaled to two full size bundles and FIST is scaled to one. Height scaling is also an important simulation consideration. The FIST facility, scaled to full B W R / 6 elevations, preserves the regional flow areas, void fraction, and elevation heads. The ROSA-Ill facility, simulating two full length bundles with four half length bundles, halves the core and bypass heights, doubles

their relative flow areas, and decreases the void fraction in these regions. To preserve the core region counter current flow limiting (CCFL) performance characteristics, the flow areas of the four inlet orifices and upper tie plates were reduced from that of a prototype bundle. Half length bundles also require shorter jet pumps to preserve the relative static head. The jet pump suction is at 66% core elevation in ROSA-Ill which is close to the 67% core elevation in a BWR and in FIST. The ROSAIII other vessel internal regions are also shorter than the BWR/6-251 and FIST as shown in table 1. The breaks are placed in the same relative locations as in the reference BWR so that the mass and energy fluxes are simulated. The break areas, given in table 1, are sized to give mass and energy flows in proportion to the volume scaling of the facility. Because the two facilities have different reference systems, the FIST recirculation suction line break is 4% larger than ROSA-III, and the drive line break is 22% larger. Com-

226

H. Kurnamaru et al. / Similarity study of R O S A - I l l and F I S T

STEAMLINE STEAM

DOME

FEEDWATER

i i i,, LO w ER I PLENUM

~y/

I

I FIST

c,l

ROSA-Ill Fig. 2. Comparison of vessel regions in both facilities.

bined, the two break areas are 7% larger in the FIST facility. Another important facility characteristic is the heat stored in the metal structure, which when released to the water, will decrease the depressurization rate. The stored heat in the regions containing water has the greatest impact on the depressurization rate. In these regions, which include the lower plenum, guide tube, bypass, bundles, and upper plenum, the ROSA-Ill facility has 32% more metal per full size bundle than the

FIST. The metal surface area in these regions of ROSAIII is only 3% larger than that of FIST. 2.2. Test conditions and test procedure Comparison of initial conditions is made in table 2 for the ROSA-III 983 and FIST 6DBA1B tests. The LPCI-diesel generator (DG) is assumed to fail in both the tests resulting in the failure of two out of three LPCI systems. On a per full size bundle basis, the ROSA-III test initial power is 64% lower than in the

227

H. Kumamaru et aL / Similarity study of ROSA -Ili and F I S T

FIST due to power supply limitation. Therefore, the recirculation line flow rate is controlled to give the desired lower plenum subcooling which is very nearly the same in both the tests. Steam and feedwater flows in the ROSA-Ill test are lower corresponding to the lower initial core power. The downcomer levels, slightly different in the two tests because of different height scaling, 0.1802 m3 in ROSA-Ill and 0.1698 m3 in FIST, are set to give the correct scaled water volume in the downcomer. The lower initial power in the ROSA-Ill

Bundle

test reduces the bundle and upper plenum void fractions and therefore increases the initial mass in these regions. The boundary conditions are also compared in table 2. The ROSA-Ill test bundle power is held constant until the per bundle power is very nearly equal to the reference bundle power, i.e. 9 s after break. There is no activation of the ADS in the FIST test, whereas in ROSA-III the ADS is opened at 115 s. This does not have a significant effect because the system pressure is

Bundle

A

D

270"

-

Bundle B

/

Region Radial peaking factor Local peaking factor

180"

A 1.0 1.1

B C 1.0 1.0 1.0 0.875

~Bundle

90"

C

W 0.0 0.0

Fig. 3(a), Arrangement of heater rods and radial power distribution in ROSA-111test facility.

228

H. Kumamaru et aL / Similarity study of ROSA-II1 and FIST

~ 2 i 3 ' 4 1 9

5

6 j 7 8" -Bj ~ L _ _ B " 101 tl 112, 13 14 15 ~ 16

Rod Number Rod Type

17] 18 19 201 21 2C2 Za3 24

C

BICiW

B

Rod Type Local Peaking Factor

W O.0

Fig. 3(b). Arrangement of heater rods and radial power distribution in FIST test facility.

low by this time. The recirculation pumps are tripped off at time zero in both the tests. The feedwater valve is closed at time zero in the FIST test and at 2 to 4 s in ROSA-III. In the FIST test, the pressure control system is actuated to simulate the BWR pressure control system which maintains the turbine inlet pressure at 6.7 MPa by regulating the turbine valve. In ROSA-III, the pressure control system is operated to maintain the steam dome pressure at 6.7 MPa. In the ROSA-III test, main steam isolation valve (MSIV) closure is initiated at 13 s. In FIST, MSIV closure is activated by the Level 1 trip with a 2 s delay. The HPCS is activated at 27 s in both tests. Injections of LPCS and LPCI are specified to activate at system pressures below 1.78 MPa and 1.55 MPa, respectively, in ROSA-III test, and below 2.00 MPa and 1.74 MPa, respectively, in FIST. In the FIST test the broken recirculation loop is isolated at time zero and the intact loop at 13 s to prevent flashing from these overscaled volumes.

3. Comparison of test results 3.1. System pressure and sequence of events

Timing of the key events that follow is compared in table 3, and the system pressures are compared in fig. 4. In the ROSA-Ill test, the system pressure was maintained at 6.7 MPa from 6 s to 13 s due to actuation of the pressure control system. The pressure started to increase after MSIV closure from 13 s to 15 s after

break. The mixture level in the downcomer decreased rapidly after break, and the recirculation line outlet in the downcomer was uncovered (RLU) at 17 s after break. Then the steam in the vessel was discharged directly through the recirculation suction break, and the system pressure began to decrease rapidly. At 18 s after break, the system pressure fell below 6.4 MPa, and the lower plenum fluid reached the saturation condition. The depressurization rate slowed down after initiation of the lower plenum flashing (LPF) because of continuous steam generation in the lower plenum. The difference in the system pressure transient immediately after break initiation is due to differences in the pressure control system operation logic, core power curves during the first 9 s, and recirculation suction break uncovery timing. The ROSA-Ill MSIV closed approximately 3 s before the recirculation suction break was uncovered causing a slight increase in the pressure just before the rapid depressurization began. The FIST MSIV closure occurred as the break was uncovered, and the impact was not noticeable. The rapid depressurization due to recirculation suction break uncovery began at 8 s in FIST and at 17 s in ROSA-Ill. Consequently the FIST system pressure remained slightly lower than ROSA-Ill throughout the transient. On the whole, however, the system pressure transients in both the ROSAIli Test 983 and the FIST Test 6DBA1B were very similar. 3.2. Mixture level

The mixture level transients in the downcomer measured with conductivity probes are shown in fig. 5 for the ROSA-Ill and FIST tests. In the ROSA-Ill test, the mixture level in the downcomer fell rapidly after the break. The jet pump suctions were uncovered at 12 s after break and the exit nozzles of the recirculation line at 17 s. The mixture level in the downcomer was recovered considerably later than that in the core since ECCS water was injected inside the core shroud. The downcomer mixture level in the FIST test fell earlier than that in the ROSA-III test. This is due to a smaller initial mass of water in the downcomer and a greater subcooled break flow in the FIST test. The FIST test was controlled at a slightly higher pressure than the ROSA-Ill due to the differences in pressure control system operation logic. The result was greater subcooling upstream of the break and therefore higher flow rate during subcooled phase. The mixture level transients inside the core shroud are presented in fig. 6 for the ROSA-Ill and FIST tests. The mixture level in the ROSA-Ill test was determined

H. Kumamaru et al. / Similarity study of ROSA-Ill and FIST

229

Table 1 Comparison of ROSA-III and FIST facilities Item

ROSA-III Total

Per full bundle

8× 8 4 o) 1.88 1.10 1.40 62 2 12.27 15.01 387.1 119.1 60.8 4.46 (3)

_ 193.5 t2) 59.5 ~2) 30.4 (2) 2.23

1.340 0.438 0.234 0.172 0.062 0.124 0.082 0.052 0.112 0.066

0.671 0.219 0.117 0.086 0.031 0.062 0.041 0.026 0.056 0.033

6.00 1.00 4.51 2.69 1.12 0.69 2.21 1.28

-

FIST Total

ROSA/FIST Per full bundle

Bundles Array No. of bundles Heated length (m) Local peaking factor Axial peaking factor No. of heated r o d s / b u n d l e No. of water r o d s / b u n d l e O.D. of heated rods (mm) O.D. of water rods (ram) Flow area (cm 2) U T P flow area (cm 2) SEO flow area (cm z) M a x i m u m power (MW)

8× 8 1 3.81 1.04 1.40 62 2 12.27 15.01 96.8 79.1 29.9 7.00

0.5

2.00 0.75 1.02

Volumes (m s) Total system Steam dome Downcomer Jet pumps; Recirc. loops Steam separator Upper plenum Bundles Bypass Lower plenum Guide tubes

0.705 0.246 0.141 0.024 0.040 0.044 0.043 0.037 0.088 0.042

0.95 0.89 0.82 3.58 0.76 1.41 0.95 0.70 0.64 0.77

Heights (m) Total vessel Steam dome Downcomer Jet p u m p s Steam separator Upper plenum Bypass Lower plenum

19.42 6.09 10.80 4.51 2.10 1.83 4.39 4.11

0.31 0.16 0.42 0.60 0.53 0.38 0.50 0.31

Break flow areas (ram 2) ADS Recirc. line break Drive line break

349.6 539.6 81.0

174.8 269.5 40.5

186.7 279.6 51.8

0.94 0.96 0.78

(t) Four, half-length bundles were used in ROSA-III. (2) ROSA-III bundle, UTP, and SEO flow areas are for two half bundles. o) The ROSA-III power decay is delayed by 9 s to compensate for the low initial power.

f r o m m e a s u r e d c o n d u c t i v i t y p r o b e signals, while t h e m i x t u r e level in t h e F I S T test w a s e s t i m a t e d f r o m n o d a l densities which were calculated from measured nodal d i f f e r e n t i a l p r e s s u r e s . T h e m i x t u r e level in t h e core for

ROSA-Ill

test p r e s e n t e d in fig. 6 is for t h e B c h a n n e l

core.

I n t h e R O S A - I I I test, t h e m i x t u r e level w a s f o r m e d in t h e l o w e r p l e n u m at 27 s f o l l o w i n g i n i t i a t i o n o f t h e

230

H. Kumamaru et al. / Similarity stud), of R O S A - I l l and F I S T

Table 2 Comparison of initial conditions and boundary conditions ROSA 983

FIST

Total

Per full bundle

6DBA1B

ROSA/FIST per full bundle

Initial conditions Pressure Power Core inlet subcooling Core inlet flow Feedwater flow Steam flow Water level Total liquid mass

(MPa) (MW) (K) (kg/s) (kg/s) (kg/s) (m) (kg)

7.19 3.62 11.6 35.90 1.26 1.30 4.84 (l~ 600 ~21

1.81 17.95 0.63 0.65 300

7.19 5.05 9.85 18.55 2.45 2.63 14.~ 277

Bounda~ conditions Power trip Break initiation Recirc. suction break Drive line break Steam line break ADS flow area ADS trip Pump trip Feedwater line trip MSIV trip HPCS trip LPCS trip LPCI trip HPCS temperature LPCS temperature LPCI temperature HPCS flow LPCS flow LPCI flow Total ECCS flow Broken loop isolation Intact loop isolation

(s) (s) (ram2) (ramz ) (ram 2) (mm2) (s) (s) (s) (s) (s) (s) (s) (K) (K) (K) (kg/s) (kg/s) (kg/s) (kg/s) (s) (s)

9 0 539.0 80

269.6 40.1

0 0 279.6 51.8

0.96 0.78

349.6 115 0 2-4 13 27 50 & 1.77 MPa 50 & 1.47 MPa 322 322 322 0.80 1.38 1.38 3.54 N/A N/A

174.8

186.7 None 0 0 LI + 2 27 35 & 1.97 MPa 35 & 1.68 MPa 322 322 322 0.59 0.72 0.52 1.83 0 13

-

0.40 0.69 0.69 1.77

0.36 0.97 0.26 0.25 1.08

-

0.68 0.95 1.32 0.97

~ Some uncertainty due to flow effects on the downcomer differential pressure measurements. ~2~ Total liquid mass estimated assuming downcomer liquid level at 5 m, mixture level inside the shroud to the top of the separator, and the slip ratio was 1.0. ~3~ Estimated mass. lower plenum flashing, although almost none of the core was uncovered at this time. This indicates the occurrence of C C F L at the core inlet orifice. As the flashing in the lower p l e n u m subsided, the mixture level was formed in the channel B core at 35 s after break, and the level decreased from Position 1 to Position 3 between 35 s and 72 s. The mixture level in the channel B core began to recover from approximately 90 s, and the whole channel B core was reflooded at 115 s following actuations of LPCS and LPCI from 85 s and 95 s, respectively. In the ROSA-III test, while uncovery of

the channel B and D cores was observed, the channel A and C cores were being covered with two-phase mixture throughout the test. Measured core inlet f o w rates showed that the bundles A and C were upflow while the bundles B and D in downflow or countercurrent flow when the mixture level was being formed in the lower plenum. Judging from conductivity probe signals above the upper tie plate, the upper plenum was probably covered with a two-phase mixture throughout the test. In both the tests, soon after lower plenum flashing started, the mixture level was formed in the lower

11. Kumamaru et al. / Similarity study of ROSA-1H and F I S T

Table 3 Comparison of timing of key events Key events

-

ROSA-m,Mixture Level (estin'mted from CP Signals) FIST,Mixture Level (estimated from CP Signals)

-

....

Time (s)

RLUI ~HPCSLPCS1 ;LPCZ~ ROSA-nT RLUI Ih'PCS " ILPCI ~ FIST

ROSA 984 FIST 6DBA1B Break valves open Pump coast-down begins Feedwater line closes Jet pump suction uncovers Recirc. suction break uncovers Power decay initiated MSIV closes Lower plenum flashing begins Guide tube flashing begins Lower plenum level forms Jet pump exit uncovers Jet pump loop isolated HPCS flow begins Bundle heat-up begins PCT occurs PCF temperature PCT level above BHL (normalized by heated length) LPCS flow begins LPCI flow begins ADS actuation Bypass/GT refill begins Bypass CCFL breakdown Bypass refilled Bundle reflooding begins Bundle reflooded UTP CCFL breakdown Bundle quenched Jet pump exit recovers

0.0 0.0 0.0 0.0 2.0-4.0 0.0 12.0 5.0 17.0 8.0 8.0 0.1 13.0-15.0 8.0 18.0 ll.5 20.0 12.0 27.1 18.0 39.0 13.0 27.0 27.0 39.6 41.0 81.2 110.0 575K 656K 0.81 85.0 95.0 115.0 92.3 114.6 120.8 -

~ !.PCS , ;

100 8

6 ~

0

~

50

I

I

t00

8

~ 7°

0

Time

' ; P S~C

~.....~

'. . . . . . . . . . . . . . .

(FIST)

,-'"

~ Recir. Line Suction

i 100

1'~0

200

250

Time (s) Fig. 5. Comparison of mixture levels in downcomer in ROSAIII and FIST tests.

plenum. The significance of this event was that until the lower p l e n u m refilled, the leakage from the b u n d l e was controlled by C C F L at the core inlet orifice. This led to increased b u n d l e drainage a n d core uncovery in FIST, especially after the lower p l e n u m level reached the b o t t o m of the jet p u m p s at 39 s a n d the lower p l e n u m

115.0

115.0 125.0 125.0 125.0 125.0 125.0 145.0

ROSA-?JT, Mixture Level (estimated from CP Signals) Core (B Channel) , Lower Plenum

[

F.IST , Mixture Level (estimated from Nodal Densities) Upper Plenum, Core, Lower Plenum LPF, .Pcs LPcs LPcI " i * i < ' : - ROSA-~ F ,HPCS , LPCI LPF, LPCS ~ ~ ~ FIST

o'3 I 0 0 ~ ROSA-In" ROSA-.~ T/C Pos Position 5 0 L T/C

50 o~ LLJ

250

Is)

Fig. 4. Comparison of system pressures in ROSA-Ill and FIST tests.

.

.

.

.

.

.

. .

I ,i,~

L ~-I to E~ ~=

FIST T/C Position

. . . . . ~----

HOS. Z -" POS. 2 I ~POS. / I P o s 33 ~~ I ~ ~ ""J Pos4~ Pos. 5 ~ / POs.6~ L........ I~r~J'~ ~ . . . .

~ 50 7o 0

I

2()0

\ I~sA)

/--/

SO

0.50 68.0 75.0

I

~ - : -JP s ~

.......................

i/ o

Ll'Leve~

-/- - -(FIST)

---~-1- . . . . . . . . . . . . . . . . . . . . . . . .

~FIST

I

I IR°sA' ~ i

',l

~-~ ROSA-IT

150

~LI Level

5o : - - f I- . . . . . . . . . . . . . . . . . . . . . . . . .

---rF

.

I

~L3 Level

I.

.--

ROSA-IT

L

- - -

i

x::}.~

121MSI V . . . . . - - I0~- ; ;HPCS LPCS;;LPCI o ]MSIV LPCS CL 8~; ;HPCS ;iLPCI

'~

---~ . . . . . . . . . . . . . . . . . . . . . . . . . . .

---

01.

231

* *

&~/

I II ~

Tie Plale

P(~2 Per, 3

* -Poe 5 * P~ 6 " Pos 7

.-I"-~// IuI I,

I ,I

-

Plenum

°i

,

,; ~;

~--]

Lower T i e Plate "Core Inlet Orifice -Inlet from ,Jet Pump

b

io

tbo Time

1.~,o

2ha

zso

(s)

Fig. 6. Comparison of mixture levels inside core shroud in ROSA-Ill and FIST tests.

232

H. Kumamaru et aL / Similarity study of R O S A - I l l and F I S T

steam was vented up the jet pumps. Rapid decrease in the core mixture level from 32 s in the FIST test was probably due to this increased bundle drainage arising from CCFL break-down at the core inlet orifice. On the other hand, sudden recovery of the core mixture level at 125 s in FIST test was due to water falling from the upper plenum caused by CCFL break-down at the upper tie plate. The increased core uncovery in FIST is also attributable to smaller initial fluid mass in the bundle and upper plenum in comparison with ROSA-Ill. The trends of the mixture level transients in the downcomer and inside the core shroud in the ROSA-Ill Test 983 and the FIST Test 6DBA1B agreed reasonably well.

3.3. Surface temperatures of fuel rods The surface temperatures of fuel rods in the ROSA111 and FIST tests are compared in fig. 7. While the surface temperatures presented in the figure for ROSA1II test are those for the fuel rod D22, the surface

I

7ooI 6oo I

~HPCS LF~SS,,LPCI. . . .

i

~i

____

,~ ROSA-Z - -

I

ROSA-~

- 5oo I

17oo o

400~ .....

--~~600

ioo .

.

.

I::;

.

500[

4°0[

,Pcs 0

Po~.io~ I-

[Rod-

LPc "TF's , 50

~00

150

Time

Is}

~o~., P-&Tz P ~ - ~

temperatures for FIST test are those obtained from several rods, as shown in the figure. In the ROSA-Ill test, after the lower plenum flashing became less effective, the fuel rod surface temperatures started to rise successively from Position 1 to Position 3 between 39 s and 48 s. The fuel rod surface at Position 3 repeatedly dried out and rewetted during the period 48 s to 93 s probably due to oscillation of the mixture level about this position, (The mixture level in the channel B core is presented in fig. 6, since the level in the channel D core was not measured in the test.) After initiation of LPCS and LPCI at 85 s and 95 s, respectively, the core was quenched upward from Position 3 to Position 1 between 93 s and 95 s. The peak cladding temperature (PCT) in the ROSA-Ill test was 575 K, and it was observed at Position 2. Core heat-up began at very nearly the same time in both the tests, i.e. at 39 s in ROSA-Ill and 41 s in FIST. In the FIST test, there was drainage through the upper tie plate that cooled the very top of the bundle (positions 1 through 3). The ROSA-Ill bundle D, which showed downflow at the core inlet orifice, heated up at the very top of the bundle (Positions 1 through 3) indicating liquid loss to the lower plenum and little drainage from the upper plenum. The ROSA-Ill bundle B also heated up slightly later in the transient. However, the ROSA-Ill bundles A and C did not heat up at all. These results are consistent with the mixture level transients the core inlet flow directions in the four channels. The FIST ECC systems came on at 27, 64, and 75 s, and the increased upper tie plate drainage cooled the top of the fuel rods thus limiting the PCT to 656 K at 110 s. The hot bundle in both the tests was in counter-current flow with some cooling from the top. The PCT was 81 K higher in FIST since it occurred at a higher heat flux location. On the whole, however, the trends of the fuel rod surface temperatures in the hot bundle in ROSA-Ill Test 983 and FIST Test 6DBA1B agreed reasonably well.

, 200

250

4. Analysis

5~U~P~

7]

D22--D22~D22 [)2-2--DZ2-D22 [D221

~Rod 18 13 13 13 ' 34 4 7 12l ~FIST IEievot,on,,094 078- 0 6510.52-6 32; 022:012] i

,~(Lenglh from Bollom of HeotingLength) /(Heohng Lenglh) Fig. 7. Comparison of fuel rod surface temperatures in ROSA111 and FIST tests.

Post-test analyses of ROSA-III Test 983 and FIST Test 6DBA1B were performed with the RELAP5/ MOD1 (cycle 018) code. The objective of these analyses was to examine the similarity between the thermal-hydraulic phenomena in ROSA-III and FIST large break tests and a BWR large break LOCA. The primary interest of these analyses was to compare the trends in the transients of system pressure and fuel rod surface

H. Kumamaru et al. / Similarity stud),' of ROSA -III and F I S T

The core was divided into two regions radially: one modeled one out of the four channels and the other modeled the other three channels. Each region had seven heat structures which modeled fuel rods and corresponded to the seven steps of the axial power distribution. Since no jet pump model was incorporated in the code, small pumps with the same coastdown and homologous curves as the recirculation pumps were added to the suction lines of the jet pumps to achieve the steady-state flow condition in the system [9]. Measured flow rates were used as a function of system pressure in the analysis for the main steam flow rate, A D S flow rate, feedwater flow rate, and ECCS flow rates. Closure of the MSIV, opening of A D S valve, closure of feedwater line valve, and actuations of ECCS

temperatures. For the study of similarity, R O S A - I l l Test 983 and F I S T Test 6DBA1B were first analyzed with the code to examine the capability of the code to calculate the transients. Then, a B W R counterpart was analyzed by using the same analysis methodology as was used in the R O S A - I l l and F I S T analyses. Finally, the similarity between R O S A - I l l and F I S T large break tests and a BWR large break L O C A was examined by comparing the calculated results for the R O S A - I l l and F I S T tests and the BWR LOCA. 4.1. R O S A -111 analysis

The R O S A - I l l system was modeled with 66 volumes, 74 junctions, and 26 heat structures as shown in fig. 8.

P Pipe S V Single Votume S J Single Junction

TV Time Dependent

• VTO[uri?e 1 e TE Dependent B

Junction " Branch

~U~

°

L~J

S R V Safety Retief Valve MS • Main Steam FW : Feedwater 5~o

SRV r ~

~gvl°~~ MS ~-rl~_3or'T~ ~-JtI [ ] T ' I ~ I ~

~s~l,w~~' L'~°~31 8o, sv TV

TJ J

v ~

MSLine ~DeS° kLine Break vr-~"rv'

I ~

I

V : Va[ve P P Pump SP ' Separate "'

233

Heat Structur,

Fig. 8. Nodalization diagram for ROSA-Ill analysis.

H. Kumamaru et aL / Similarity study of ROSA-1II and F I S T

234

were done at the same specified times or pressures as in the experiment. The pressure control system was modeled to keep the system pressure above 6.7 MPa before MSIV closure as was done in the experiment. The discharge coefficient was set to 0.6 in the calculation of break flow. The calculated system pressure agrees well with the measured pressure during rapid decrease after break, hold by the pressure control system, recovery after MSIV closure and rapid drop after uncovery of the recirculation line outlet in the downcomer, as shown in fig. 9 which compares the calculated system pressure with the measured result. The time of uncovery of the recirculation line outlet in the downcomer is a little earlier in the analysis than in the experiment. The depressurization rate slows down somewhat after the initiation of lower plenum flashing in both calculated and measured results. The overall agreement in the behavior of the calculated and measured system pressure transients is satisfactory. Temporary dryout immediately after the break is calculated at Positions 1 and 2 in the analysis while it was not observed in the experiment, as shown in fig- 10 which compares the calculated surface temperatures of fuel rods with the measured results for the fuel rod D22, the same rod as those presented in fig. 7. This is due to the occurrence of departure from nucleate boiling (DNB) arising from decrease in the core inlet flow rate in the analysis. The core inlet flow rate immediately after the break is closely related with the characteristics of the jet pumps and the coast down of the main recirculation pumps. Hence, this discrepancy in the surface temperature immediately after the break is attributable to one out of the three causes: the imperfection of the DNB correlation in the code or the insufficiency of the jet

~ 21MSI ~ . ~/ HPCS

"~

IMSIV 81 i !HPCS LR:S: :LPCI ~

6"

4I~_

. . . . LPCS ILP¢I

. ~ = ROSA-'I~

Anolysm ~= ROSA-lIT Doto

' ~

~\x~. \xx~ "

.

----

ROSA-11T Analysis ROSA-11T Data

\\

RLU LPF ~ x

2 ~ ~ RL~ LPF

~--~ "~ - '~.~.,

50

100 Time

150

200

250

(s)

Fig. 9. Comparison of calculated and measured system pressures in ROSA-Ill.

i

IHPCS LPCS~LPCI ~:=: ROSA-~I

700 LPF

60C

-

--

I 1 ]"

ROSA-~ Analysis

-- - ROSA-lIT Data

~

50( 40{ .~ 70( 600

~. 500

7oo

400

~oo

u~ 700

500 o~

~g 600

400

5oo

~oo

400

6o0

700

500

I

6

0

0

~

<

.....

j 0

50

1O0

t 50

200

O5

~

I

%00

! 250

Ttme (s)

Fig. 10. Comparison of calculated and measured fuel rod surface temperatures in ROSA-Ill.

pump model or the input for the main recirculation pumps used in the analysis. The fuel rod surfaces at Positions 1 and 2 are rewetted immediately after the lower plenum flashing in the analysis. The time of dryout after mitigation of the lower plenum flashing and the increasing rate of surface temperature in the upper part of the core in the analysis agree well with those in the experiment. However, the dryout after mitigation of the lower plenum flashing is calculated excessively in the middle and lower part of the core in the analysis while almost no dryout was observed in the experiment. This is because the mixture level in the core decreases to the lower tie plate in the analysis due to insufficient calculation of CCFL at the core inlet orifice. Hence, it is desirable to reexamine the interphase drag correlations in the code so that they may take larger values at flow contractions in order to calculate smaller CCFL at the core inlet orifice. On the whole, the surface temperatures of the fuel rods are calculated well in the upper part of the core and is conservatively estimated in the middle and lower part of the core in the analysis.

H. Kumamaru et aL / Similarity study of ROSA-HI and FIST

235

4.2. F I S T analysis

The FIST test was also analyzed with the RELAP5/MOD1 (cycle 018) code using the same methodology as was used in the ROSA-III analysis. The FIST system was modeled with 58 volumes, 65 junctions, and 27 heat structures. The core consisted of one region which modeled the FIST channel. The region had five heat structures which modeled the fuel rods. Measured jet pump discharge flow rates were injected into the lower plenum as a function of time in the analysis since enough information on the characteristics of jet pumps and coast down of the main recirculation pumps was not available. The pressure control system was modeled to keep the system pressure above 7.2 MPa before MSIV closure. The calculated system pressure reproduces well the rapid decrease after break, the hold by the pressure control system, and the rapid drop after uncovery of the recirculation line outlet in the downcomer, as shown in fig. 11 which compares the calculated system pressure with the measured result. The recirculation line outlet in the downcomer is uncovered a little earlier in the analysis than in the experiment as was the case with ROSAIII. The depressurization rate slows down somewhat after the initiation of lower plenum flashing in both the calculated and measured results as was also the case in the ROSA-III analysis. The overall agreement in the behavior of the calculated and measured system pressure transients is satisfactory. As was the case with ROSA-III, temporary dryout immediately after the break is calculated at Positions 3 and 4 in the analysis while it did not occur in the experiment, as shown in fig. 12 which compares the calculated surface temperatures of fuel rods with the

12 )MSIV'

'LPCS }LPCl

lO [M~SIVIHPCS ~

8~; "~

LPCS ;HPCS ,: iLPCl

6~

0t

o

--

J

I

50

~

~ FIST Analysis

~=DoIaFIST FIST Analysis

I --T--"R'5"='=="~..... r .... loo 15o 2oo 250 Time (s)

Fig. 11. Comparison of calculated and measured system pressures in FIST.

700

LPF

LPCS

-

"---.~

|

FISTAnalysis . . . . FISTData

600 _.-.".~

-

-

_L

500 4o0

700~-

~Z#

~ 5o0 g

" .....

. . . . . . . . . . . -. . . . . .-~ 400 -~

6OO E 5OO

. . . . .

~"

700

4OO

h

70 O;

........... .........

600

4oo

5OO 400 I

700 ~

Pos.i................

600

..........................

]500 ~400

500 400

LPCS

_PF

~,e'r <==~,'o,"

i ~HPCS ~4LPCI r

I J

J

50

I I =

I

100 Time

..........

voTa

J

J

150

~

J

200

J

250

(s)

Fig. 12. Comparison of calculated and measured fuel rod surface temperatures in FIST.

measured results, the same temperatures as presented in fig. 7. This is also due to occurrence of DNB in the analysis. Hence, it is desirable to reconsider the DNB correlation in the code so that it may predict DNB lower at high flow rate in order to calculate moderately the DNB immediately after the break. The calculated behavior of the dryout and quenching of fuel rods after mitigation of the lower plenum flashing in the upper and middle part of the core agrees well with the behavior in the experiment. The trend of top-down quenching in the upper and middle part of the core is calculated particularly well in the analysis. However, the dryout after mitigation of the lower plenum flashing is overestimated in the lower part of the core in the analysis while almost no dryout occurred in the experiment, as shown in fig. 12 which compares the calculated surface temperatures of fuel rods with the measured results, the same temperatures as presented in fig. 7. The reason is the same as that described for the ROSAIII case. On the whole, the surface temperatures of fuel

236

H. Kumamaru et aL / Similari(v stud)' of ROSA-111 and FIST

rods are calculated well in the upper and middle part of the core, while they are estimated conservatively in the lower part of the core. 4.3. B W R analysis and similarity between R O S A . I I I and F I S T tests and B W R L O C A

14 }2

L~CS '

,HPCS , '~LPCI

]

,

/

~=

; ;HPCS LPCSi,LPCI

iRLU

C£.

~E 8 cr~

The BWR counterpart was analyzed with the RELAP5/MOD1 (cycle 018) code using the same methodology as was used in the ROSA-Ill and FIST analyses. The BWR system was modeled with 67 volumes, 75 junctions, and 31 heat structures. The core was divided into two regions: one region modeled the central core with 748 bundles and the other modeled the peripheral core with 100 bundles. The central core included seven heat structures divided axially with radial and local peaking factors of 1.4 and 1.13, respectively, and seven heat structures with radial and local peaking factors of 1.0 and 1.00, respectively. The peripheral core had seven heat structures with radial and local peaking factors of 0.7 and 1.00, respectively. The axial peaking factors of these three heat structural groups were 1.4 as was the case in the ROSA-Ill and FIST analyses. The power curve used in this analysis was that evaluated by the General Electric Company [10]. The main steam flow rate and the feedwater flow rate were kept at steady-state values until these flows were tripped. The feedwater was stopped at 4 s, and the MSIV was closed by the L2 level signal with a time delay of 3 s [2]. The ADS and HPCS were actuated by the L1 level signal with a time delay of 120 s and the L2 signal with a delay of 27 s, respectively. The LPCS and LPCI were actuated at system pressures of 2.2 MPa and 1.6 MPa, respectively [2]. The pressure control system was modeled so as to keep the system pressure above 6.7 MPa before MSIV closure as was the case with ROSA-Ill analysis.. The system pressure transient until the uncovery of the recirculation line outlet in the BWR analysis agrees well with that in the ROSA-III analysis, as shown in fig. 13 which compares the calculated system pressure transient for a BWR with the calculated results for the ROSA-Ill and FIST. This is because the pressure control system in the BWR analysis was modeled in the same way as used in the ROSA-III analysis. The time of the uncovery of the recirculation line outlet in the BWR analysis agrees with that in the ROSA-Ill analysis better than with that in the FIST analysis. This means that the liquid level transient in the downcomer in a BWR is closer to that in the ROSA-Ill than that in the FIST. The decrease in the depressurization rate due to the lower plenum flashing occurred at 6.4 MPa is seen in all three analyses. The temporary hold of the system pres-

tRRLU '

LU

LPCS

L; ,HPCS , ,LPCI

6 -~,

~',~. ~.,

--

~--

BWR

Analysis

ROSA-I

Analysis

FIST ~= Analysis BWR Analysis ROSA-IAnalysis FIST Analysis

el_

2 0

0

50

%00 Time

150

200

250

(s)

Fig. 13. Comparison of calculated system pressures in ROSAIII, FIST and BWR.

sure after LPCS actuation is not seen in the BWR analysis, while it is calculated in both the ROSA-Ill and FIST analyses. The depressurization after the uncovery of the recirculation line outlet in the BWR analysis is somewhat faster than that in the ROSA-III and FIST analyses. The reason may be that the stored heat in the vessel walls and vessel internals slows down the depressurization rate in the ROSA-III and FIST. The temporary dryout immediately after the break due to occurrence of DNB is seen at Positions 1 through 4 in the BWR analysis, as shown in fig. 14 which compares the calculated surface temperatures of fuel rods with radial and local peaking factors of 1.0 and 1.00, respectively, in the central core in the BWR analysis with corresponding surface temperatures in the ROSA-III and FIST analyses. This agrees well with the results in the FIST analysis. In the ROSA-III analysis, the temporary dryout immediately after the break was calculated only at Positions 1 and 2. The main reason may be that the ROSA-III core power was held constant, at 40% of scaled steady state power, during first 9 s. However, no dryout was observed immediately after the break in both the ROSA-III and FIST tests. The times to dryout after mitigation of the lower plenum flashing in the BWR analysis agree with those in the FIST analysis better than with those in the ROSA-III analysis, especially in the middle part of the core. On the other hand, the quenching behavior in the BWR analysis agrees with that in the ROSA-II[ analysis better than with that in the FIST analysis, especially in the middle and lower part of the core. The fuel rod surface temperatures in the upper part of the core repeat rewet and dryout after LPCS actuation in the BWR analysis. This repeating of rewet and dryout is intermediate

H. Kumamaru et aL / Similarity study of ROSA -III and FIST ,

i

6 0 0 L LPF ~ . . ',

'

LPCS ~-,

r

i

1

i

-....

BWR Analysis ROSA-11T Analysis

Pas'~"~ ~"-"~'--

700"~

i--

,

4500 ~'

....

500-

/ ~L

1 oo

7700

500 ~

T700~

~ . ;

500

LPF~

4oo

"'-4~=.~= __.~

LPF ,

,,

Analysis

LPCl

~ ~ FIST .

LPCS ~

,

HPCS

0

50

~

~

,

100

Ana ysls

150

i 200

r J

250

Time (s) Fig. 14. Comparison of calculated fuel rod surface temperatures in ROSA-III, FIST and BWR.

behavior between bottom-up quenching in the ROSA-Ill analysis and top-down quenching in the FIST analysis. However, the overall trends of the system pressure and fuel rod surface temperature transients in the BWR analysis agree well with those in the ROSA-Ill and FIST analyses.

5. Conclusions A large break test in a recirculation pump suction line was conducted at the ROSA-Ill test facility of JAERI and at the FIST test facility of the GE Company under as nearly the same test conditions as possible in order to develop common understanding and interpretation of the controlling thermal-hydraulic phenomena during a large break LOCA of a BWR. The LPCI-DG was assumed to fail in both tests. The following conclusions have been drawn from the comparison of results of these tests:

237

(1) Key experimental observations. (a) The agreement between the ROSA-III and FIST system responses is very good. In particular, the system pressure responses agree quite well. (b) The mixture level and fuel rod surface temperature transients in the ROSA-Ill hot bundles are similar to those of the FIST bundle. (c) The half height scaling in the ROSA-Ill facility does not have a significant influence on the system response. (2) Comparison of experimental responses. (a) One slight difference in the system responses is a slightly faster system depressurization in the FIST facility caused by early recirculation suction break uncovering. This also produces earlier actuation of LPCS and LPCI by about 20 s. (b) The FIST test yields more bundle uncovery than the ROSA-Ill test. Steam in the lower plenum in the FIST test flows out through jet pumps when they uncover allowing more liquid to drain from the bundle. The ROSA-Ill and FIST tests and a corresponding BWR LOCA were analyzed with the RELAP5/MOD1 (cycle 018) code to study the similarity of the ROSA-Ill and FIST large break tests and a BWR large break LOCA. The following conclusions have been drawn from the comparison of calculated results: (1) The RELAP5/MOD1 (cycle 018) code reproduced both test results relatively well. This indicates that the code has the capability to calculate relatively well the thermal-hydraulic behavior during a large break LOCA of a BWR. However, it is perhaps desirable to reexamine the DNB and interphase drag correlations and the jet pump models used in the code. (2) The fundamental thermal-hydraulic behavior during the ROSA-Ill and FIST large break test is similar to that during a BWR large break LOCA with only small differences in details, for example, the difference in depressurization rate due to the effect of stored heat in the ROSA-Ill and FIST tests. Therefore, the ROSA-Ill and FIST large break test data are useful to study the fundamental phenomena of a BWR large break LOCA and to assess a computer code for a BWR large break LOCA.

Acknowledgments The authors are grateful to Dr. Y. Koizumi, Dr. Y. Anoda, and Mr. H. Nakamura for discussions about the experimental and analytical results during the course of

238

H. Kumamaru et al. / Similarity study of ROSA-HI and FIST

the study. The authors are also grateful to Miss M. Kikuchi of the Nihon Computer Bureau for typing the manuscript.

References [1] Y. Anoda, K. Tasaka, H. Kumamaru and M. Shiba, ROSA-III system description for fuel assembly No. 4, JAERI-M9363, Japan Atomic Energy Research Institute (1981). [2] General Electric standard safety analysis report, BWR/6, DOCKET-STN-50477, General Electric Company (1975). [3] J.E. Thompson, BWR full integral simulation test (FIST) program test plan, NUREG/CR-2575, EPRI NP-2313, GEAP 22053 (1982). [4] A.G. Stephens, FIST facility description report, NUREG/CR-2576, EPRI NP-2314, GEAP 22054 (1982). [5] M. Suzuki et al., Recirculation pump suction line 200%

[6]

[7]

[8] [9]

[10]

break integral test at ROSA-III with two LPCI failures, RUN 983, JAERI-M 84-135 (1984). W.S. Hwang, Md. Alamgir and W.A. Sutherland, BWR full integral simulation test (FIST), Phase I test results, NUREG/CR-3711, EPRI-NP-3602, GEAP-30496, General Electric Company (1983). K. Tasaka et al., Comparisons of ROSA-Ill and FIST BWR loss-of-coolant accident simulation tests, JAERI-M 85-158, Japan Atomic Energy Research Institute (1985). V.H. Ransom et al., RELAP5/MOD1 code manual, Vol. I, 11, NUREG/CR-1826, EGG-2070 (1980). N. Abe and K. Tasaka, Analysis of ROSA-Ill test RUN 704 by RELAP5/MOD0 code, JAERI-M 9476, Japan Atomic Energy Research Institute (1981). LS. Lee et aL, BWR blowdown/emergency core cooling program 64-rod bundle core spray interaction (BD/ECC 1A), Final report, Vol. I, Large break simulation tests, NUREG/CR-2009-I, EPRI-NP-1782-I, GEAP-24912-1 (1981).