ML20236P793

Non-proprietary WCAP-14472, Steam Condensation Events at Osu AP600 Test Facility

ML20236P793
Person / Time
Site:	05200003
Issue date:	02/28/1996
From:	Carter M, Dumsday C, Roidt M WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
Shared Package
ML20236P767	List: ML20236P793 NSD-NRC-98-5738, Forwards non-proprietary TR WCAP-14472, Steam Condensation Events at Osu AP600 Test Facility. Encl Is non-proprietary Version of TR WCAP-14471 Which Was Submitted in
References
WCAP-14472, NUDOCS 9807170149
Download: ML20236P793 (46)

Text

.

i.

},

1 L

t f

lL

.f.

\\

.a

~

l r

\\.

I

{

]

Y

\\

y

\\

e b '

h s

s/

f e

a i

T e.

Westingho.useEnergySystems W-

~

%)

M

...te.,,

-...--w.

.a g.

{,'

• l'

\\

l

\\

e k

p T

.s

N.

d.. h,...,......p.

,,,..e

. 'i l.*

. ;., -.,,,i,-

v, :.

...,.,, j,, s ';, m

k. g.

f,

, '...,b

.>,v

,c...'.,-

.s,

' N. ['.' ; ' ' -,

N'V ' {.

(

dy 5,h

b. y.

go h 't y I

• 9,,'h[h..,..

l

' - f, [ '.),

M h

]Ye

. h

• f,

e

. t%

s

~ ~ ~.; 3

,'h.

.h.

...f p. 7. ' w d,. 4., gr,.,7,P.P.,

,'r-

• ../.-,...., p c. p.?,,

,.)

'.q s,.

.y a

. '.*^.

n.. y.j.) f *.}V,j,.f :

',~

8

w. -r Y

. ( % ' s. ',..., ' ' _..,

y 'f,

\\{.,',, (f..N.}]x j Q'AN.e}-}:.p'.,;'.g y.4 }$ g '

' f e

((. '. "'.' - 7..o u ;. ' ' '.; '., k.g,,

. _' [.( _..

C ' y,. %.. ',...';

t hi L,f

!. ',.:,: A* <o b

s

' i,3>

[,g,

+%.

kj j

q;y, J. 4y{y: : 3'.

'._fip,'yf.c.p

L

',*p

., _,,, g\\

~f m.Q

.x

..r ' ( f.. ; ', L.,, :;, y ;."M g y;;:dg t y,p

,y

,p

.,,y,..,,

-Qp i

J

.Q y.

- e,,,, < l.

xb,

,?

i>

','- /

7

_, } };.,., '.,.l 1,[ ' :, j,, 'r,.c.

.'.),

,. - 7,

.fr. u.A. <, L ' <,.,

'}.

,.,.}

%,g, c.d..

3.

./,

[

[

" g,1 n'

?,

, '. ', E,. [ '. '

.h

,f h, N '

[ ' '

[,;

f {f

. '..- ~ '

a

, t

~

e

? ' '.

W I.

w' >'w. n-MMw. R.e. /. %,eg.s f~FWW 9 f:

- t.,,

s.

, :,Y

. ! @$. ',f.% p

• h b'

Y l -, "x ; w,i c,4 n,y.,.p34'

/

, '.,.- l Q ' ;.,. r,. ',,

x. 3.,;

g....-

'.*'r d

,j

,g e.

.~

n h_ qs ; {n,.u' J.f ' < p,M i'

m.

y.

1,-

., l ){

? l.

i * '-

. '.,t

,J )

i, ';

f [

f; %.." ' /t? l,' 'l h, 81

' ?'

m j.Q;...,,;, p W

,7 e

q

^; ;

.-?

_.Q

• ;8{l };"l

,f,,l

....'.,. '. ~

'. J i

• l l...;

l-i

• l.

..y/ ' f ' <... 4.:.._

ff.;t

. *~._,\\

._n.[-

&p, ry.,

.g.

yy k,r..,,.

K.

r W/Ng..-.*

i.

s.

h w c:p 4,. :t. w, sg ' i,y 4,,{A.,,7-

'..*..N' i,

/-

v.,. y

.','.O..',s"h"S,'.Id,d. '. /,

k I'-

p'*tle 'M'M,'[f)N' N,' 'U k

+r d'l

., '. ".., -,, f 'd ['#'h,. r.m y f

y N,'.-'.t-

,y,,y, e

iE?k.'b."jk'[ ?"'f)f l'?... dh } '-

'b

• 3-

'..,,,s.

., }. r

' ' i - "

' M w. [M h; '4 P

. j'

~

'[

N'i. i., ' m N,.,'l.E.6...'k. e s. '7.-M5l.*

M. M..,' A!,

,3...

b

., * ] Q,. ".,s

['

,7 e 4

.i,

,-4;-

I:

c

..k

't f

e.

i '

,.. ; [,

.(....f,q',,.

/,5.

> c,y, -

.q..f;,g[. 98 Il

.((, ;

,, i 'y.g y ;; s.. ' -

., f,3,s},,l,,.v,

..o

'_ :[.;,2. ','Q' f ^ ' ' 'v,, {; _. e f.;;, -).f ; :,. L ; j ' y ' ;g 'Y).h.j[[!;

.3

.l 9

f

, s; ',..

j,j-

' ;. g.

g i ' '_ :. ? [iy %:

- ; y,

[ g[,.'s..'

L:. c

• g ; :';. : ; f

. i T'.

?y

?'

' 7

. ?

1

[x; 3,([1,q i,;p,n.,.

_J.,.

.y.-

q,

.(.

.,.h.,.,.,a.,.,y

.',..t>-

2

i,

,..J,.'.

p

.?

psy '

w j<

y.

..y...

e, 2.

f9q p'; m:sq a...,g

,. 3.., T c.. M.

.,.. 1, - i

~ '..'

l~

'h.,

, ;. ~. } 2 } ' ',' ;., N I

53

'^

4 j,

{*

g

'1 5

1 y,g,,.,

f,,,,

.y

. g s.

1,

.m i -...,'3

, Q.,' '.' r: ' 1),'.j). yn - N. :i7<d

.r

-, ;E V,. z. w-

,f,3 r;i ; 4,,.>s

, y n. ;);f,. z y

-[

t_, _-, _,..

.'w. *.. - -,

.. - - ', ' '. i.

• j * ? ? ? ;,-

s.,x n ),, d

'q.

^

YO. ':]..',. 5 A )f.

,',, ? ',. ;.. ' - '- :..-

.. ;,'. ll :.

• ' L~

\\

-[ } yy ;.

z

'n ;,

[

' ' _ ~ }&;'l:k,

_',,'v_

u-o

',. '^

' ~,. '_

....;...l i' ' '.'i A'..

,(,

.{.

q

.o '

a,

- M

+ t

-g.

..,g

. ggs

.-('l

~

?/.

y_,

? 1,-....,- - -,o, f.,'..y.'^

?y ^.,; ;;.,

l !sy K j y.,y&,

l',' '. h l,.

' l ^,, Q:C ' ' ? !'. $ 1 4, f A

.. , + _

,., -..' ~, -. ;...,,.,.. _;. - -

f,

., ';. ! - \\

t

• v'n,,

l: '.

4 1

l.

r

- 4.. q.

. a +~.

.,.t w.ms';.x m77.. ". y.*. A (,

.. l: 4o

~

..u.,

f l13.,y, 3

...c..

6

- e

....,.,..,,.,.,,;es

,.c

.....o...

.. m, r a

'1

't j

.'_[1.

?.*.

.. f l_ : ;;...'.. ) ~ ~._..

p's u s.,...c

' l zj

}.j

.e, 7

hr; ;,.

m,

~...

I_ y 9.n,

. f ' ;. :., :.., ;;,y,. t.u

'p

~

t: lt j, et

,'r

l.r. * '_.. !';

l.l9'...{ '}. L{

s

', _j[

s.... _,

's.;...

y [,.,. ; u., 'm; w

..m. r.,,,.,.K;. ;;;3

~-

., ~.;..

j.

~.;-

7

~

y

.e.

e q,4 f.', J

\\7Q.,Q.\\ Q

'l y

..x.,

s :

'I e

.q.

~ '

.,: fk '..., '.,

.ei

.'..'v"'n g.

a n,,,.,..,.

- -'. T [..

.',1

- ' /.

v

... >.....w+.'.',

..c.e;.;p l -

.h

. ? f,s ' ' \\ ' ' ;, '

'*ob

?'?

{.

.,3,

L'.' ' .. ',,,'I' ' ~,. 'V',s', Ll ?

.'".f 1'. ",.$

,. =.;

^ ',.

.e e

' *.;.. 'C

' e l'

y., ;/5. ' -.'v' 4

,'. f,p,,.(/. ! 'y '- w < g t

• '.2

./ <

-a

'..i.: lf.e

..a.

1 :,

p,.,,:: W ',.f'.,* ; :

.., ', - +.f \\.:..'. / f ~ " *l ib l

e u,

L.'

s e','

. a.

~

., an..m.....,...., :.. #. t..

7.. A; ;.. yy1.a.

..+

,u rgw,gwan w

.3 s.

WESTINGHOUSE NON-PROPRETACY CLASS 3 1

I

)

. WCAP-14472 STEAM CONDENSATION EVENTS AT THE OSU AP600 TEST FACILITY l

February 19%

M. Carter I

C. Dumsday M. Roidt l

(-

WESTINGHOUSE ELECTRIC CORPORATION j.

Energy Systems Business Unit L

P.O. Box 355 Pittsburgh, Pennsylvania 15230 4355 C1996 Westinghouse Electric Corporation All Rights Reserved I

i i

i l:

l-u:0735w.non:Ib4313% -

_. _j

4 TABLE OF CONTENTS Section Title P. age l

1.0 INTRODUCTION

1-1 1.1

' Derwtion of Condensation Events 1-3 l.2 Measurement of Pressure Spikes during Category III Tests 1-3 1.3 Analysis of Upper Head Noise Source 1-4 2.0 ANALYSIS OF UPPER HEAD NOISE DURING OSU MATRIX TEST SB01 2-1 2.1 Video Record 2-1 2.2 Event Timing 2-1 2.3 DAS Scan Rates 2-2 2.4 Data Analysis 2-2 2.5 Pre-Transient Conditions 2-3 2.6 Analysis of Transient 2-4 3.0 ANALYSIS OF UPPER HEAD NOISE DURING OSU MATRIX TEST SB18 3-1 3.1 Purpose of Test 3-1 3.2 Analysis of SB18 and Comparison to SB01 3-1 4.0 ANALYSIS OF UPPER HEAD NOISE DURING OSU MATRIX TESTS SB03, SB05, AND SB07 4-1 4.1 Similarities with Matrix Test SB01 4-1 4.2 Analysis of Matrix Test SB03 4-1 4.3 Analysis of Matrix Test SB05 4-3 4.4 Analysis of Matrix Test SB07 4-4 4.5 Empirical Scoping Calculations 4-4 5.0 Conclusions 5-1 Appendices A

Transient Data Summary for OSU Test SB01 A-1 B

Transient Data Summary for OSU Test SB18 B-1 C

Summary of Upper Downcomer Temperature Transients for OSU Test SB03 C-1 D

SB05 Supporting Plots D-1 E

SB07 Supporting Plots E-1 F

Some Empirical Scoping Calculations Related to the Oregon State University AP600 Low-Pressure Integral Systems Test Facility F-1 G

Drawings G-1 m:\\2735wamn:Ib.031396 iij

L

1.0 INTRODUCTION

j Westinghouse Electric Corporation and the Nuclear Engineering Department, Oregon State University (OSU) designed and built a one-quarter scaled model of the AP600 plant that included simulations of the reactor coolant system (RCS), steam generators (SGs), passive safety injection systems, and

~ nonsafety injection systems, in the Radiation Center at the university in Cc vallis, Oregon. The purpose of this test facility is to simulate the performance of the AP600 passive safety systems at a reduced size to provide test data for validation of the safety analysis codes. The test facility, fabricated completely from austenitic stainless steel and designed for normal operation at 450*F and 400 psig, is scaled using the hierarchical two-tiered scaling (H2TS) analysis method developed by the U.S. Nuclear Regulatory Commission (NRC). Simulated piping breaks are tested in the hot leg, cold leg, pressure balance line between the cold leg and the core makeup tank (CMT), and the direct vessel injection (DVI) line. Decay heat that scales to 3 percent of the full power (about 2 minutes after shutdown) is supplied by electrically heated rods in the reactor vessel. Simulated accidents are programmed by the control system to proceed automatically.

The facility design models the detail of the AP600 geometry including the primary system and the l.

pipe routings and layout for the passive safety systems. The primary system consists of one hot leg and two cold legs with two active reactor coolant pumps (RCPs) and an active steam generator for each of two loops. Dere are two CMTs, each connected to a cold leg of one primary loop, and one pressurizer connected to the other primary loop, as in the AP600 plant design. Gas-driven accumulators are connected to the DVI lines. De discharge lines from a CMT, an accumulator, and

[

one-of-two in-containment refueling water storage tank (IRWST) and reactor sump lines are connected to each DVI line. The two independent tiers of ADS stages 1 to 3 are combined as a single tier. The two-phase flow from the ADS stages I to 3 is separated in a swirl-vane separator and the liquid and vapor flows are measured to obtain the total flow rate. De separated flow streams are then i

recombined and discharged into the IRWST through a sparger. Thus, the mass flow and energy flow from the ADS into the IRWST is preserved.

. The time period for simulation includes the IRWST injection, as well as draining of the IRWST and the sump injection to simulate the long-term cooling mode of the AP600. De time scale for the OSU test facility is approximately one-half, that is, the sequence of events occur approximately twice as fast in the OSU test facility as in the AP600 plant.

To measure the mass flow and energy of the breaks, the two-phase flow fmm the break is separated in

- a swirl-vane separator and the liquid and vapor portions of the total flow are measured. The liquid fraction of the flow is discharged to the teactor sump, as in the AP600 plant. The vapor is discharged '

to the atmosphere, however, the equivalent liquid flow is capable of being added to the IRWST and sump to simulate the condensate return from the passive containment. The same approach is also used j

for the two ADS stage 4 valves on the hot legs. De two-phase flow is separated in a swirl-vane separator, the two single-phase flows are measured, the liquid phase is discharged into the reactor sump while the vapor phase is discharged to the atmosphere and the liquid equivalent is capable of l

. uA2735w. mon:lt431396 l.)

1 L___________._________

d l

1 being added to the IRWST and sump. The IRWST, reactor sump, and separators are interconnected and can be pressurized to sirriulate the containment pressurization following a postulated LOCA.

A multi-tube passive residual heat removal heat exchanger (PRHR HX) is located in the IRWST. The j

~ heat exchanger uses the'same C-tube design as the AP600 and has two instrumented tubes to obtain wall heat fluxes during the tests. There are primary fluid thermocouple, wall thermocouple, and differential pressure measurements located on the heat exchanger. The IRWST is also instrumented with strings of fluid thermocouple to determine the degree of mixing within the tank and to record the temperature of the IRWST water that is delivered to the reactor vessel.

l The reactor vessel for the OSU tests includes a 3-ft. electrically heated core consisting of forty-eight i

1-inch diameter heater rods. De heater rods have a top skewed axial power shape. Some of the heater rods have wall thermocouple swaged inside to measure the heater rod temperature. There are l

also five thermocouple rods in the heater rod bundle, including fluid thermocouple to measure the axial coolant temperature distribution. De scaled flow area in the core is preserved as well as the flow area in the vessel upper plenum. There are simulated reactor intemals in the upper plenum to preserve the flow area and the scaled coolant volume. The reactor vessel includes an annular downcomer into which the four cold legs and the two direct vessel injection lines are connected. The two hot legs penetrate the reactor annulus and connect with the loops. De AP600 reactor vessel neutron reflector is simulated using a ceramic liner to reduce' the metal heat release to the coolant.

Dere is approximately 600 kW of electrical power available at the OSU test facility, which corresponds to approximately 3 percent AP600 decay heat.

1 De OSU experiments examined the passive safety system response for the small-break and the large--

)

break LOCA transition into long-term cooling. A range of small-break LOCAs wm simulated at different locations on the primary system such as the cold leg, hot leg, CMT colc leg pressure balance line, and the direct vessel injection line. De break orientation, top or bottom of cold leg, was also exammed. Different single failure cases were examined to confirm that the worst situation was used in the AP600 SSAR analysis. Selected tests continued into the long-term cooling, post-accident mode in which the passive safety injection came from the reactor sump as well as the IRWST. A large-break, post-accident, long-term cooling situation was also simulated.

The testing at the OSU facility progressed through several phases. The first phase was the pm operational testing in which characteristics of components and systems were determined, such as volume and flow resistances. Also during this phase, the facility's data acquisition and control systems were checked out. The testing then proceeded to the integrated response of the facility's

. systems during hot conditions. This phase of testing was referied to as hot functional testing, or HFr.

Durmg HFT, data was obtamed at various plateaus of temperature and power levels. Data was taken during HFT at both steady-state conditions and transient conditions. Data taken during steady-state conditions was used to assess both component and system performance. De data taken during transient conditiums was used as a mechanical checkout of the systems and components, as well as a Y

mA2735w. mon:ll>031396 1-2

m---_ _

i l

shakedown of the facility's control and indication systems. The transient test conditions consisted of simulated small breaks in cold leg no. 3 at several power levels, and an inadvenent actuation of the automatic depressurization system (ADS).

t The final phase of testing was the matrix testing that primarily simulated small-break LOCAs at various locations. The tests were classified as category I, category II, and category III according to their importance in the licensing process for AP600. Category I and category 11 tests were completed first before continuing with category III tests.

1.1 Detection of Condensation Events During the transient tests of HFT, loud noises or hangs were heard in the vicinity of the upper head of the reactor. The source and cause of these noises were unknown at the time. Before beginning matrix testing, We upper head was removed from the reactor vessel and an inspection was made of the reactor internals. During the inspection of the upper internals, the upper support plate was found to be lifted and cocked from its correct position. The design of the OSU test facility upper internals was atypical i

of the AP600 plant design. Although the upper internals are correctly scaled, the upper support plate in the OSU test facility was held in place by its own weight versus being permanently affixed to the upper intemals as in the AP600 plant design. The OSU design was intended to allow the maintenance of reactor heaters without removing the upper intemals. Instead, the upper support plate could be lifted and removed by itself allowing access to the lower internals and the reactor heater rods. Before starting matrix testing, the upper intemals were modified. The upper support plate was permanently affixed to the upper internals by a bolting arrangement, thereby prohibiting movement of the upper support plate.

However, the modification of the upper internals and upper support plate had no effect on the noise j

heard in the upper head during matrix testing. During category I and category II testing, plans were made to test for the source and severity of what was assumed to be condensation events.

1.2 Measurement of Pressure Spikes during Category III Tests Durmg category III testing, the reactor coolant system (RCS) was instmmented with four fast-response model 211B3 Pieztron pressure sensors from Kistler Instrument. The purpose of the fast-response transducers was to record pressure pulsations or spikes during the noise-producing events. The Rosemount pressure transmitters permanently installed in the facility were not fast enough to record the high-frequency spikes that accompany either water hammer or depressurization events. The signals from these transducers were amplified and stored on a digital recorder for later evaluation.

i The four pressure sensors were installed on permanent transmitter sense lines connected to the top of f

the reactor on root valve 101 (RV-101), the bottom of the reactor (RV-122), and cold legs no. 2 and 3 (RV-216 and RV-213, respectively). Refer to Figure G-3 and Figure G-5 for the location of the root valves. These locations were selected as sample points because the reactor was the apparent location ui2735w.non:Ib4313%

l-3

of the noises heard during HFT and matrix tests. Cold leg no. 3 was selected to represent the side of the plant containing cold legs no. I and no. 3, hot leg no.1, and steam generator no.1. Cold leg no. 3 also contained the break location for one of the category III tests. Cold leg no. 2 was selected to represent the side of the plant containing cold legs no. 2 and no. 4, hot leg no. 2, and steam generator no. 2. The PRHR HX and pressurizer are located on this side of the plant.

Five category III tests were conducted while instrumented with the fast-response pressure transmitters.

They are:

SB18 Repeat of SB01: 2-inch cold leg no. 3 break simulation with long-term cooling operation SB26 PRA case: Multiple ADS failures without PRHR HX, no break simulation and long-term cooling operation SB28 PRA case: Multiple ADS failures with DVI line double-ended break simulation and long-term cooling SB29 Inadvertent ADS actuation: No break simulation at higher power with long-term cooling operation SB31 Spurious S signal: No break simulation and without ADS actuation and long-term cooling operation No significant pressure spikes were observed during these tests. 'Ihe largest pressure pulsations were detected during category III test SB18. The pressure peak during this test was less than [ ]"" psi.

All other test data indicated a pressure peak of less than [

]"' psi.

1.3 Analysis of Upper Head Noise Source The pressure measurements during the five category III tests listed above indicated that no significant pressure spikes or pulsations occurred during the tests. However, it was still important to understand the source of the noise heard in the upper head during some of the matrix tests. The Westinghouse test crew at OSU had filmed several category I and category II tests as an unofficial record of some events of interest. Although the videos were never intended to be used for analysis, they have proven to be a valuable tool in understanding the source of the noise.

To understand the noise in the upper head, an analysis of test data from five tests was conducted.

They are:

SB01 (category I): 2-inch cold leg no. 3 break on bottom of pipe. Test was videotaped, u:\\2735w.non:1b 0313%

l-4

SB18 (category III): Repeat of SB01; only analyzed tests instrumented with fast-response transmitters. Test was not videotaped.

SB03 (category II): 2-inch cold leg no. 4 break on bottom of pipe. Test was videotaped.

SB05 (category II): 1-inch cold leg no. 3 break on bottom of pipe. Test was videotaped.

SB07 (category II): 2-inch cold leg no. 3 break on bottom of pipe. Test was videotaped.

Matrix tests SB01 and SB18 were selected because they are identical tests and one was videotaped (SBI) and the other was instmmented with fast-response pressure transmitters (SB18). Between the two tests all required information needed to perform an evaluation is available. SB01 was the first matrix test performed, SB18 the last. The loudest bang of any of the te.,ts was heard in SB01 and SB18. SB01 was videotaped. The other three tests, SB03, SB05, and SB07, were selected for analysis because they were also videotaped. Thus, any noise heard on the tape could be reviewed and compared to approximately 750 channels of test data in an attempt to identify the source of the noise.

uA273Sw.non:Ib-0313%

l-5

1 l

I 2.0 ANALYSIS OF UPPER HEAD NOISE DURING OSU MATRIX TEST SB01 2.1 Video Record The video provided an audio record of events that could be referenced to events on the data acquisition system (DAS). h video clock also recorded video time in hours and minutes. Although

. the video clock is not precisely set with the DAS clock, benchmarking times and events was possible.

l The video of SB01 recorded a sharp, metallic-type noise in the upper head region during the early phase of the test. The video time for this event was 2:40 pm. Just prior to the noise, the energization of the solenoid for ADS no. 3 valve (RCS-603) could be heard. Coincidental with the upper head noise was the sound of the energization of solenoids for IRWST injection valves RCS-711 and l

RCS-712. Based on this information from the video, it was detennined that the upper head noise occurred at approximately [ - )*** seconds on DAS rack no.1. Time = 0 seconds is defined as the time when the break valve opened.

I h opening of the IRWST valves coincidental with the upper head noise was an important clue to

.]

understanding the transient. In SB01, the IRWST valves received a signal to open when the reactor i

coolant system (RCS) pressure decreased to [

]'*' psig. The input signal to the valve opening logic came from pressure transmitter PT-107, which measures the pressure in the upper head of the reactor.

hs, the event that created the upper head noise also created a quick depressunzation in the upper head region of the reactor, thereby generating a signal to open the IRWST valves.

l The video record of upper head noise provided a starting point for the analysis of test data. The DAS l

at the test facility at OSU has about 750 channels of instrumentation, providing a large amount of data l

- for analysis to determme the source of the noise.

l l

2.2 Event Timing l

The DAS had three separate systems or racks, each rack with its own intemal clock. When the TEST pushbutton on the control panel was depressed, a signal was sent to each rack to start data acquisition.

[

The start signal was provided to each rack via a communication network, so there was a small delay i'

between actuating the TEST pushbutton and the first time stamp of each rack.

In addition to the three systems that recorded data from transmitter output, another system recorded the signals generated by the programmable logic controller, or PLC. These signals included TEST pushbutton actuation, valve position, and alarms. Wonderware software recorded the time of any input and output signals activated by this microprocessor-based system.

The break valve opened 120 seconds after the TEST pushbutton was pressed. W PLC both monitored the status of the TEST pushbutton and provided the signal to open the valve. The scan rate of the PLC was 100 milliseconds, so it did not introduce any noticeable delay in opening the break uA2735w.noa:n-0313%

2-1

]

4 valve. De break valve position was monitored by the PLC via the valve's limit switches. The time i

when the break valve opened, that is, was not fully closed, was considered the start of the LOCA and, therefore, the start of the test.

During data reduction, a time = 0 seconds was obtained for each rack by adding approximately 120 seconds to the first time stamp. Note that this time did not exactly coincide with the opening of the break valve due to the small time delay in starting the data acquisition via the communication network. Thus, events monitored by the PLC and recorded by the Wonderware program must be compared to determine the timing of events. The upper head noise was heard at the same time the IRWST valves opened. Using the time recorded by Wonderware for valve actuation and the video record, it was determined the noise occurred at approximately [

]**# seconds using the data files of DAS rack no.' l.

2.3 DAS Scan Rates The DAS had two scan modes for data acquisition: continuous mode and burst mode. De continuous mode scan rate could not be varied by the user; it varied for each rack, but generally the rate was on the order of one scan per channel every 10 seconds. The burst mode scan rate was user-defined. The scan rate for SB01 was one scan per channel every 2 seconds. The event creating the upper head noise was relatively fast, its duration less than the period of the continuous mode. It was necessary to use the burst mode for this analysis.

2.4 Data Analysis In order to postulate the source of the noise in the upper head, it was necessary to determine the status of the fluid in the RCS before, during, and after the event. Several parameters were used: level, temperature, pressure, and differential pressure. In addition, the phase of the fluid is measured by a heated phase switch (HPS). A tabulation of these key parameters is found in Appendix A, " Transient Data Summary for OSU Test SB01." De table lists the parameter measured, the instrument measuring the parameter, and the magnitude of the parameter before, during, and after the transient.

The source of data for the table was the burst test data files of SB01 that had acquisition rates of one scan per channel every 2 seconds. Plots of the SB01 data for the selected channels can be found in Appendix A.

De following discussion of SB01 test results is based upon the data plots and data summary of Appendix A. The reader can find relevant information by using the description of the parameter in the table. The table is divided into sections for level, temperature, flow, differential pressure, and fluid phsse.

uM735w. mon:lb431396 2-2

2.5 Pre-Transient Conditions Matrix test SB01 was initiated by a simulated 2-inch break in the bottom of cold leg no. 3. By

[

]'*# seconds, the time of the upper head noise, the core makeup tanks (CMTs) had emptied sufficiently to actuate the automatic depressurization system (ADS). ADS valves no.1, no. 2, and no. 3 had opened. ADS valve no. 3 opened about [

]"# seconds before the event occurred. The accumulators injected into the vessel at the maximum, or near maximum, values they could reach during the test. The accumulator pressure and flow were sufficient to stop CMT injection, thus the only injection into the core during this time was from the accumulators.

The tubes of both steam generators were empty. The plot of steam generator level indicated the tubes were full and had actually refilled after draining. This was an erroneous indication due to the voiding of the steam generator level transmitters' reference legs during the draining of the steam generators.

Steam generator tube temperatures confirm that they were empty. The tube temperatures were approximately [

]'***F; saturated temperature is approximately [

]'***F. This means that the tubes were empty and the fluid in the tubes was superheated steam.

The reactor vessel upper head was also superheated; its temperature of [

]'***F less than that of the steam generator tubes, but still approximately [

]'***F superheated. The collapsed level, or liquid level, of the upper head, upper plenum, and hot legs was difficult to discern due to inherent errors in measuring levels in steam / liquid environments. However, by combining the data from level I

instruments and thermocouple, a reasonable assessment can be made. Note: The upper plenum was defimed as the region above the upper core plate and below the upper support plate.

Both temperature indication and level indication revealed that the upper head above the upper support plate was voided. By level measurement, the upper plenum level was between the middle of the hot legs and the top of the hot legs. The top hot leg thermocouple at the reactor flange are mounted to measure pipe temperature near the centerline of the pipe. The top thermocouple of botn hot legs indicated a subcooled fluid temperature. Thus, both temperature and level measurements indicated that the level was at least to the hot leg centerline, but below the level of the upper support plate.

The level in the downcomer annulus was measured by two independent level transmitters. The data from the two transmitters was in good agreement. At [

]'*# seconds, the level transmitters indicated the level in the annulus was approximately [ ]'*# inches below the cold legs, or [ ]'** inches above the top of the direct vessel injection (DVI) lines. This level measurement was confirmed by temperature measurements in the annulus. Two independent sets of fluid thermocouple were installed 90 degrees apart, but at the same level. The upper thermocouple were mounted 2 inches above the lower thermocouple. The level transmitters showed that the liquid level was between the upper and lower thermocouple. The stratification between the top and bottom thermocouple was apparent, showing steam at the top thermocouple and a subcooled liquid at the bottom thermocouple. Thus, temperature measurement indicated the liquid level was between the upper and lower thermocouple.

This wa ir. agreement with the level indication.

u:\\2735w.non:1b 0313%

2-3

4 Both the level and temperature measurement indicated the cold legs were empty at [

]** seconds.

Cold legs no.1 and no. 3 were connected to steam generator no.1. The cold leg channel head of steam generator no.1, cold leg no.1, and cold leg no. 3 all indicated superheated steam and were therefore empty of liquid. This agrees with the annulus level instrument indications that the liquid level in the annulus was below the cold legs. The indication of cold legs no. 2 and no. 4 is different.

Dese cold legs connect to steam generator no. 2. The effluent of the PRHR HX discharged to the steam generator channel head of cold legs no. 2 and no. 4. At [

]'6" seconds this flow rate was approximately 2 gpm. This water was [

]***F at the discharge of the PRHR HX, and the effect of this cold water could be seen as a stratification in cold legs no. 2 and no. 4. The top of the pipe was superheated while the bottom was subcooled. The cold leg channel head levels at steam generator no. 2 were small but measurable.

A differential pressure (DP) transmitter measured the DP between the upper head and the downcomer annulus.. De transmitter could indicate a positive or negative flow. He negative indication for DP at

[

]*" seconds indicated a flow of superheated steam from the upper head to the downcomer annulus.

In summary, the test data at [

]*' seconds indicates temperature stratification existed in the RCS coolant prior to the analyzed event. The upper head, upper portion of the downcomer annulus, upper

- plenum above the hot legs, cold legs no. I and no. 3, and the steam generators contained superheated steam. Cold legs no. 2 and no. 4 were also superheated, except for a small subcooled flow on the bottom of the pipes as PRHR HX flow retumed to the core via the cold leg channel head of steam generator no. 2 end cold legs no. 2 and no. 4. He downcomer annulus liquid level was below the elevation of the cold legs and above the DVI line.

2.6 Analysis of Transient Evaluation of the test data during the time period between [

]*# seconds to [

]'b' seconds revealed the probable cause of the upper head noise. De analysis of the data was limited by the scan rate of the data acquisition. De entire transient occurred in only about 10 seconds. With a scan rate of I scan every 2 seconds, the evaluation was performed with 6 scans. However, even with these imposed limits, the sequence of events is clear.

At [

]*' seconds, the superheated bubble in the upper portion of the downcomer annulus began to collapse, accelerating the liquid upward in the downcomer annulus. Within 2 seconds, the downcomer annulus liquid level had increased sufficiently to overrange the 2 annulus level transmitters. De upper taps (reference leg taps) for these two transmitters are located 6 inches below the cone barrel flange where the core bypass holes are located. The accelerated liquid impacted the bottom of the core barrel flange, producing the loud noise, or bang, heard in the video.

De thermocouple in the upper portion of the downcomer annulus also sensed the collapse of the steam bubble in the downcomer. De thermocouple was installed at an elevation 2 inches below the a:\\2735w. mon:11431396 2-4

elevation of the upper level taps of the annulus level transmitters. The thermocouple measured a superheated temperature at [

]"" seconds. At [

]** seconds, the thennocouple measured a subcooled liquid as the steam bubble in the downcomer collapsed.

When the steam bubble collapsed in the annulus, it left a void that resulted in the upward acceleration of liquid in the annulus. The low pressure created in the upper annulus also resulted in a rapid increase in the steam flow from the upper head to the annulus. The DP transmitter across the bypass holes measured a step change in flow that peaked within 4 seconds. At the same time, a rapid increase in the DP across the upper support plate was measured as steam flowed from the upper plenum into the upper head. His DP was almost sufficient to lift the upper support plate had it not been attached to the upper intemals. In the initial stages of the test program, the upper support plate was not attached to the upper intemals and was, therefore, lifted by the DP.

He mass required to replace the void left by the collapsed steam bubble in the annulus originated in the upper plenum. The outsurge from the core peaked within 4 seconds, and was measured by two narrow-range and one wide-range transmitters that measure levels within the core barrel. Temperatures below the core increased during the outsurge as warmer water from the core flowed downward. The temperatures then decreased as the flow changed direction, taking the normal direction from the downcomer annulus to the core.

Within 4 secorxis after the initiation of the event, water drained via gravity from the downcomer annulus to the core. The core was at a lower liquid level as a result of the outsurge from the core.

The remamder of the transient was slow compared to the collapse of the steam bubble. The force caused the initial upward acceleration of liquid in the downcomer annulus as a result of the local depressurization when the steam bubble collapsed. The downward acceleration of the fluid in the downcomer annulus late in the transient (4 to 10 seconds after initiation of the event) was a result of the differential head between the downcomer annulus and the level in the upper plenum. Ten seconds after the initiation of the event, the system had essentially returned to its pre-transient conditions.

i l

uA2735w. mon:lb4313%

2-5 t_______________________.__

i-l 3.0 ANALYSIS OF UPPER HEAD NOISE DURING OSU MATRIX TEST SB18 3.1 Purpose of Test l

l Matrix test SB18, the last matrix test performed at OSU, was identical to test SB01, the first test performed at OSU.. Both tests simulated a 2-inch break in the bottom of cold leg no. 3, which is on the opposite side of the RCS from the pressurizer and the PRHR HX. SB18 was designed to confirm the results of SB01, and although not the original intent, it also provided an opportunity to determine if the upper head noise in SB01 could be repeated. Matrix test SB18 was one of five Category III tests instrumented with fast-response pressure sensors intended to measure effects of any water hammer events. Refer to Section 1.2 of this report for a list of Category III tests performed with the fast response pressure transducers installed.

The remainder of this section will compare the test data results of SB01 to SB18. Dere was no video record kept of SB18; however, the upper head noise was heard by the test crew at approximately the same time it was heard in SB01. De test data comparison between SB01 and SB18 will demonstrate the same sequence of events occurred in SB18 as in SB01 at the time of the Upper Head noise.

3.2 Analysis of SB18 and Comparison to SB01 A' video record was not kept of SB18. However, using the data from the fast-response pressure sensors, the event timing could he determmed 'in the same manner as SB01. Instead of using the audio of the videotape to deermne the time of the event, the small-pressure spike (approximately

[- ]d* psig) could be used as the event marker. With this information and the test data files, the transient was found to start at [ - ]"' seconds on DAS rack no.1.

~ De system conditions in SB01 before, during,.and after the upper head noise are remarkably similar to the system conditions in SB18 during the same event. Appendix B, " Transient Data Summary for J

~ OSU Test SB18" is a listing of system conditions before, during, and after the transient. This table is similar to the table for SB01 found in Appendix A. The only noteworthy difference between the l

transients in SB01 and SB18 was the time of the event with respect to the actuation of ADS valve j

no. 3. In SB01, the event occurred approximately [

]"' seconds after the opening of ADS valve L

no. 3; whereas in SB18, the event occurred approximately [

]"' seconds after ADS valve no. 3 I:

opened.

j A k.p.rison of the data tables in Appendix A and Appendix B indicate that the conditions in the two j

tests immediately preceding the transient were essentially identical. In both tests the level in the downcomer annulus was below the cold legs and above the DVI line. De upper head, cold legs,

)

steam generators, and upper portion of the downcomer were drained and superheated. De upper j

plenum level was somewhere between the hot leg centarline ami top of the hot leg. Superheated steam flowed from the upper head to the downcomer via the bypass holes in the core barrel. The accumulators injected cold water into the DVI lines at the maximum injection flow seen in the test.

mA2735wman:Ib.031396 31 L___o_________.

Just as in SB01, the collapse of the steam bubble in the downcomer initiated the transient. As the

' bubble collapsed, a liquid outsurge from the core decreased the level in the upper plenum. At the same time an increased steam flow from the upper head to the downcomer created a differential pressure across the upper support plate that tended to raise the upper support plate. As discussed before, the plate was modified, before the start of matrix testing, to mount it rigidly to the upper intemals. De test data of SB18 captured a transient DP value across the upper support plate that would have lifted the plate had it not been affixed to the upper internals. His data then confirmed that the transient that created the upper head noise could have been the same transient that lifted the upper support plate in hot functional testing. The plate was not lifted by the momentum of liquid as it impacted the upper support plate, but rather the momentum of steam as it flowed into the upper head during the collapse of the steam bubble in tne downcomer annulus.

' Although small differences exist in system parameters during the transient in SB01 and SB18, they are not significant. The test data from the two tests provide two important conclusions. The first is that

. the transient that produced the upper head noise is repeatable, and therefore predictable. The second, and most important conclusion, is that the test results show that even though the noise of this transient is the most severe of any tests performed at OSU, the pressure pulsations that it produces are inconsequential. The momentum of the accelerated liquid in the downcomer annulus is dissipated in a way so that the initial and reflected pressure spikes are so small, they are difficult to measure.

u:\\2735w. mon:1b-0313%

3-2

O 4.0 ANALYSIS OF UPPER HEAD NOISE DURING OSU MATRIX TESTS SB03, SB05, AND SB07 4.1 Similarities with Matrix Test SB01 The source of the upper head noise in SB01 was the impact of liquid in the downcomer annulus impacting the bottom of the core bound flange as it accelerated to fill the void left from the collapse of the steam bubble in the downcomer annulus. The collapse of the steam bubble was detected by a thermocouple located in the upper portion of the downcomer. It initially measured the temperature of the superheated stc an bubble and then detected the temperature of the subaoled liquid as the steam bubble collapsed. This temperature change was very rapid, taking 2 seconds or less.

' The test data of this same thermocouple in matrix tests SB03, SB05, and SB07 reveal several potential incidents of steam bubble collapse. A video record exists for the first portion of each of these tests.

The videos were reviewed to determine if a noise could be heard in the upper head when the j

downcomer thermocouple predicted a collapsit g steam bubble. If a noise did exist, an attempt was made to determine the severity of the noise, as compared to the one heard in SB01.

4.2 Analysis of Matrix Test SB03 Matrix test SB03 was a simulated 2-inch small break at the top of cold leg no. 3, and was recorded on video for approximately the first hour of the test. While test SB03 was taped, there were six -

occurrences when the thermocouple in the upper portion of the downcomer showed a significant te:nperature decrease in a relatively short period of time. From the conclusions of SB01, this temperature decrease could signify a collapsing steam bubble in the upper portion'of the downcomer.

The audio record of the video was reviewed to determme if a noise similar to the one heard in SB01 could be heard during these events. The test data was also reviewed to ascertain whether the decrease in the downcomer temperature indicated the same sequence of events that occurred in SB01 during the upper head noise.

A#ir C, " Summary of Upper Downcomer Temperature Transients for OSU Test SB03," details the results of the review. 'Ihe appendix contains a plot of upper downcomer temperature (TF-168) versus time for the first 4000 seconds of the test (page C-4).' Numbers 1 through 6 are annotated on i

the chart, representing the time when 'IF-168 decreases. Also contained in Appendix C is a tabulation of the data for the six periods (page C-3). 'Ihree separate clock times are included in the table: video clock time; time from DAS rack no.1, which contains the data for the downcomer level; and time j

from DAS rack no. 2, which contains the downcomer thermocouple. The times from DAS rack no. I and DAS rack no. 2 have been couwied to video clock time to enable a comparison of audio events recorded by the video to the physical parameters measured by the DAS.

uA2735w. mon:Ib 0313%

4-1

4 Event no.1 on the plot is a relatively slow decrease in downcomer temperature, taking over

]'***F, indicating a superheated vapor. This event f

.[

]** seconds. The final temperature was [

did not produce a noise and was not expected to, because the steam bubble did not collapse.

Two more events, nos. 2 and 5, did not produce a noise, or did not produce one loud enough to be heard on the video recording. De events were very similar; at the beginning of each transient, the

- fluid in the upper part of the downcomer was superheated. De temperature in event no. 2 was initially [

]*F; it was [

]""*F in event no. 5. Both temperatures decreased to saturation (211*F). However, the annulus level measurement did not indicate a sharp increase during the event as was seen during the transient in SB01 when the loud noise was heard in the upper head. De absence of this level increase suggests that the steam bubble did not collapse, which is supported by the absence of noise. The saturated condition at the top of the downcomer couhl have been saturated liquid, not saturated vapor. Another possibility is that a small localized collapse took place in the upper downcomer without collapsing the entire bubble, without incicasing the level in the downcomer, and producing little noise.

I The data from events no. 3 and 6 correlate to the data from SB01 when the upper head noise occurred.

In both events of SB03, the fluid in the upper portion of the downcomer was superheated at the beginning of the transient. Initial temperatures were [

]""*F for event no. 3, and [

]*Ffor event no. 6. In both instances the fluid in the upper downcomer became subcooled, indicating that the downcomer steam bubble had collapsed. De minimum temperature during the transient was

[

]""*F for event no. 3; and [

]'***F for event no. 6. The downcomer levels spiked to overrange the level instmments, providing further evidence that the steam bubble collapsed. Data obtained for the DP across the bypass holes in the core barrel flange and DP across the upper support plate had similar magnitudes and patterns of behavior to that of SB01.

The data derived from events no. 3 and 6 suggest that a noise should be heard during the transients as

' a result of the steam bubble collapse and, in fact, the video recorded the noise. The time when the downcomer annulus level transmitters overranged coincided exactly with tne noise heard on the video.

Event no. 6 produced one sharp rap on the video recording. Event no. 3 produced two smaller raps within 2 seconds of one another. This could suggest that the steam bubble collapsed, partially reformed, and collapsed within 2 =~-k The frequency of this transient exceeds the DAS scan rate of 2 seconds, so it is difficult to determine the exact behavior of the system between the 2 second scans.

The remaining transient in SB03 was event no. 4. One sharp rap was heard during the transient.

Event no. 4 was different from events no. 3 and 6, however, in that no significant increase in downcomer water level was detected during the transient. In addition, the fluid in the upper portion of the downcomer did not subcool, but only decreased in temperature to saturated conditions. De data from this event can support either of two conclusions. First, the transient of a collapsing steam bubble in the downcomer could have been so quick that even if a noise was heard, the only evidence of the collapse of a bubble was the resultant decrease in fluid temperature, in this case, from superheated to mA2735w.non:Ib.031396 4-2

saturated conditions. The second conclusion, is that localized depressurizations could have occurred in the upper downcomer without collapsing the entire steam bubble, but still creating a noise.

In sununary, the transients of SB03 discussed previously support the conclusions drawn in the analysis of SB01.. When the transient produced data similar to SB01, the noise of the collapsing steam bubble could be heard. His data consisted of subcooling in the downcomer at the level of TF-168 and a sharp increase in downcomer level within seconds of the temperature increase. On the other hand, the data from SB03 suggests that localized depressurizations in the downcomer could produce a noise without collapsing the entire steam bubble.

4.3 Analysis of Matrix Test SB05 As in the review and analysis of SB03, the data of SB05 was examined to determine if there were any instances in which the steam bubble in the downcomer annulus collapsed. De data of TF-168, upper downcomer temperature, suggested two occu Tences, one stanting at approximately [

]'** seconds after the break valve opened, ai;d a second at a time of [

]'*# seconds after the break valve opened.

Appendix D, "SB05 Supporting Plots," contains a plot of TF-168 temperature for the first 4000 seconds (page D-3). Event no. I began at approximately [

]'** seconds. It is similar to event no.1 in SB03 (reference Appendix B) in that although the temperature decreased, it never reached saturation conditions during this early transient. His indicates that the steam bubble in the downcomer annulus never collapsed in the time frame of [

]'** to [

]'*' seconds. In fact, saturated or subcooled conditions were not detected until event number 2 at about [

]'** seconds.

Event no. 2 in the 0- to 4000-second plot on page D-3 of TF-168 in Appendix D is also plotted on a different time scale in the same Appendix D (page D-5). The 3600- to 3680-second time plot indicates a superheated vapor in the downcomer annulus at [

]'** seconds. Within 4 seconds, the fluid temperature was subcooled as the steam bubble in the downcomer collapsed.

l i

Also included in Appendix D is a plot of downcomer annulus level versus time for the 3600- to 3680-second time period. De independent downcomer level channels did not reflect the same conditions.

One transmitter indicated an increasing level during the time when upper downcomer annulus temperature recorded a transition from superheated steam to liquid. These data are typical of the incidents in which the downcomer annulus liquid accelerates upward to fill the void left by the steam l

bubble collapse. On the other hand, one transmitter reveals a decreasing water level during the same i

time. Given how closely the two transmitters tracked the level immediately preceding the event, the transmitters appear to be recording accurate data. Thus, the data implies a localized depressurization occurred, rather than total collapse of the steam bubble in the downcomer.

The audio of SB05 recorded a sharp rap at 2:55 pm, which coincides with the data time of

[

]'*' seconds, nis event provided further evidence that localized depressurizations occur in the t

uA2735w. mon:1b-031396 4-3 LC----____-

down:omer annulus, but still create a noise similar to that heard during the total collapse of the steam bubble. This event is similar to event no. 5 of SB03, which was described previously. In event no. 5, the temperature in the downcomer indicated a bubble collapse without a resultant effect on the level data, but the audio record of SB03 recorded the noise.

Once again, the video recorded the steam bubble collapse in the downcom-r annulus when the

. temperature data in the downcomer predicted that the collapse would occur.

4.4 Analysis of Matrix Test SB07.

The last test analyzed was SB07 a simelated 2-inch break on the bottom of cold leg no. 3. SB07 was identical to SB01 and SB18, two of the tests previously discussed, except that one of two tiers of ADS stages no. I through 3 was isolated. In addition, one of the two ADS no. 4 flow paths was removed from service.

Upper downcomer temperature data is plotted in Appendix E, "SB07 Supporting Plots," for the first hour of the test (page E-3). The results are unremarkable because the temperature in the upper downcomer remamed at a superheated condition during this time. In other words, the temperature data never revealed an instance in which the steam bubble in the downcomer annulus collapsed.

4.5 Empirical Scoping Calculations

'Ihe analysis of the test data to this point has been limited to discussions of observations, whether the observations are from test data collected by the DAS or observations of the audio from video recordings of tests. The andysis reveals most of the noises are in the upper head, and all indications are that the loud noises heard in SB01 and SB18 are from the collapse of the steam bubble in the downcomer annulus. It is also believed that the force that lifted the upper support plate of the reactor internals is directly attributable to the transient when the steam bubble collapsed.

It is beneficial to validate these conclusions using an analytical approach with the empirical test data.

The results of the analytical approach are included in this report as Appendix F, "Some Empirical Scoping Calculations Related to the Oregon State University AP600 Low Pressure Integral Systems Test Facility."

u$2735w. mon:Ib431396 4-4

t.

5.0 CONCLUSION

S The study of steam condensation and the resultant depressurization during testing at the Oregon State University (OSU) AP600 test facility concluded that the effects of the depressurization are not significant and had no adverse impact on the test facility. This conclusion is based on three evaluations:

1. A review of test data from the fast-response pressure sensors installed for several category III matrix tests.

2. A review of test data from the DAS acquired during several matrix tests.

3. An empirical study developed for a depressurization event in matrix test SB01. The reactor vessel was modeled to show the effects of the depressurization event in the upper plenum, upper head, and downcomer annulus.

Pressure pulsations, or spikes, were measured during five category III tests. The data obtained from these tests indicate a maximum pressure spike of [

]'*# psi, a magnitude that is of no adverse consequence to design. The empirical study mentioned in number 3 above, and detailed in

~ Appendix F, demonstrates the relatively low kinetic energy of the accelerated fluid during the event.

y l

A detailed analysis was performed on two corresponding depressurization events in two matrix tests, l

SB01 and SB18. These two tests were selected because SB18, the last matrix test performed, is a repeat of SB01, the first matrix test performed. The Westinghouse test crew observed that the loudest upper head noise occurmd during these tests.

i

' A video recording of SB01 identified the time this loud noise, or bang, occurred. The test data revealed that the collapse of the superheated steam bubble in the upper portion of the downcomer annulus resulted in the downcomer fluid accelerating upward and impacting the bottom of the core barrel flange where the core bypass holes are located. He impact of the downcomer liquid upon the j

solid surface of the core barrel flange produced the bang heard during the test. The mathematical tiratment of the event in Appendix F describes the physical conditions in the reactor vessel and their effect on the accelerated fluid that created the noise.

)

SB18 test data indicated the same occurrence of downcomer steam bubble collapse at approximately the same time after ADS valve no. 3 opened. Deze was no video recording of this event, but the fast-l

. response pressure instrumentation detected a pressure spike of less than [

]'** psi. His very small i

pressure spike corre ponds to the loudest noise experiereced in the upper head durmg testing at OSU,-

l which supports the conclusion that the depressurization event has no adverse effect on the design.

l Test data and video recordings of three additional matrix tests were reviewed, showing several additional instances of downcomer steam bubble collapse. The video of these events recorded the uA2735w. mon:lkO313%

5-1

a resultant noise as the downcomer liquid impacted the core barrel flange. The review also revealed instances in which the energy from the collapse was sufficiently low, so the video did not record a noise. Data from several depressurizations also showed that the event could be localized or was too fast for the DAS scan rate to record the entire event.

l l

l l

1 uA2735w. mon:Ib-0313%

,>-2

4 APPENDIX A TRANSIENT DATA

SUMMARY

FOR OSU TEST SB01 n \\2735%.non:1bO313%

A-1

Due to its proprietary nature, the following text has been deleted from the non-proprietary version of this document.

u:u735w.non:lb4313%

A-3

e a

APPENDIX B TRANSIENT DATA

SUMMARY

FOR OSU TEST SB18 a:\\2735w.noa:lt>C313%

B-1

Due to its proprietary nature, the following text has been deleted from the non-proprietary version of this document.

l

(-

- mM735w. mon:Ib0313%

B.3

{.'

uu____________z__________

~

l

)

i

.4 I

l APPENDIX C

SUMMARY

OF UPPER DOWNCOMER TEMPERATURE TRANSIENTS FOR i

OSU TEST SB03 l

J l

l 4

li amaswaan:1b.03:396 C-1 (t-_-_-----------

1 Due to its proprietary nature me foil 8

has bn deleted from the non-proprietary version of this document.

n i

i i

l i

l:

f u M 5w m ut:1b-031396 C-3

,1, s

= ~ - - _ - _ - _ - _ - - _ _ _ _ _ - _ _ _

APPENDIX D SB05 SUPPORTING PLOTS mA2735w. mon:lb 0313%

D.1

1 Due to its proprietary nature, the following text has been deleted fT '"

• non-Proprietary

. version of this document.

1 i

i mA2735w. mon:lH31396 D-3

}

l 1

l s

1

!~

t APPENDIX E SB07 SUPPORTING PLOTS 4

i ll l:

i I

. mM735w. mon:IM31396 E-1 ti'.

i i

L Due to its proprietary nature, the following text has been deleted from the non-proprietary version of this document.

1 i

um35w. mon: b43:396 -

E-3

I J

l j

i APPENDIX F SOME EMPIRICAL SCOPING CALCULATIONS RELATED TO THE OREGON STATE UNIVERSITY AP600 LOW-PRESSURE INTEGRAL SYSTEMS TEST FACILITY Mike Roldt, Westinghouse STC l

i i

i l

a:u735wmon:tb4313%

F-1 l

1.0 Introduction To improve the understanding of the thermal-hydraulic phenomena associated with the noises observed at the Oregon State University (OSU) facility, it is necessary to develop some mathematical guidelines to interpret the various data that have been collected.

Some care must be exercised in the data selection, however, since most of the data was taken at 2-second intervals and the characteristic times associated with steam void collapse are on the order of 1 to 10 percent of a 2-second interval. Hus, some of the transient measurements may only be indications of the actual physical conditions at points in time rather than a truly continuous picture of the transient.

In short, the basic hypothesized scenario is as follows:

After the system pressure has been sufficiently reduced following the opening of the break, the automatic depressurization system (ADS) is activated, resulting in a more rapid rate of pressure reduction than that which develops through the break.

This pressure reduction results in flashing of the high-temperature (420*F) water in the reactor vessel head which is vented out the break and the ADS. The temperature of the steam remains high due to radiation and convection from the walls of the upper head.

Along with ADS activation, flows from both the core makeup tanks (CMTs) and the accumulators are inserted into the reactor through the direct vessel injection (DVI) line.

The water level in the reactor core barrel and downcomer annulus could drop close to the DVI injection level.

De relatively cold water from the DVI line is exposed to the steam in the upper part of the downcomer annulus. The steam condenses on the cold water interface and, due to the relatively small core bypass holes in the core barrel support flange, is unable to supply an adequate steam flow rate from the upper head region to maintain the annulus pressure. Hence, the pressure above the annulus, steam-water interface drops considerably.

This reduced pressure in the upper annulus causes the upward acceleration of the water in the lower part of the annulus, which impacts the bottom of the core barrel support flange to yield the " bang."

L De purpose of this work is to develop a simple model of this scenario and determine if it can provide a framework as a predictive tool.

uA2735w.non:IM313%

F-3 j

i

2.0 Selection of Relevant Data As was stated previously, care must be taken in the use of transient data due to the long times between data sampling. Figures F-1 and F-2 for the upper head differential pressures (DPs) and upper

' downcomer annulus temperature, TF-168, respectively, appear the most promising.

In Figure F-1, the " dips" at [

]d# seconds can only be read off a steeply increasing or decreasing value of the DP cell signal and are thus open to question as minimum values. However, the

]*# inch DP read across the cose bypass holes between 500 and 700 seconds must be a seasonable

_[

approximation to the actual, since it remains so steady with time.

In Figure F-2, thermocouple TF-168 reads essentially the same temperature as TF-120, the thermocouple located in the upper head of the reactor, until the time of about [

]"# to[

]**

seconds is reached. From the hot leg thermocouple, SC-141, in the top of the hot leg, it is apparent that high-temperature steam from the steam generators is flowing back along the top of the hot leg and mixing with the steam in the top of the reactor vessel and, to a lessor extent, with that in the annulus.

These temperatures vary only slightly over a minute or so and, therefore, can be taken as an accurate measurement of the superheated steam temperature in both the upper reactor region and in the upper annular region of the downcomer.

Figure F-2 also contains the saturation temperature based on the pressure measured by PT-107 in the upper head of the reactor. Since [

]"#*F is considerably less than the measured value of [

]***F in these regions, it is plausible that the water vapor is superheated by about [

]*#*F over the period of interest, i.e., that time when the bangs m d.

Figure F-2 also shows the rapid decrease in temperature of TF-168 at [

]"# seconds. Such a rapid response from a thermocouple in a steam environment can only be possible if the thermocouple has been immersed in water at a much lower temperature than the steam. Furthermore, since the shape of the curve is somewhat " rounded," it couk! represent an actual minimum value.

3.0 Esti==e*= of Mass Transfer

- This nomenclature is used in the following sections:

pc = pressure u = velocity T = temperature k = ratio of specific heats, k = c,/c, c, = specific heat of steam at constant pressure c, _ = specific heat of steam at constant volume R. = gas constant uA2735w. mon:Ib 0313%

F-4

,+

g = gravitational constant

,;, = mass flow rate

= mass in upper reactor head m

A = flow area V = volume of the upper core region Consider a system of three interconnected volumes:

Volume above the model core which generates the stagnation conditions within the system, denoted by the subscript " sat" Volume of superheated vapor in the upper reactor head region, denoted by the subscript "1" Volume of steam in the annulus, denoted by the subscript "2"; a sketch of the system as shown in Figure F-4 The volume of the upper head region is taken to be 4 ft'. The flow area between the core and the upper reactor head through the upper core support plate, A,, is composed of eight 0.838-inch diameter holes for a 4.4123 in.2 total. The area between the upper reactor head and the annulus, A, is 2

composed of eight 0.375-inch diameter holes for a 0.8836 in.2 total. An estimate of the steady-state flow rate through both these areas is needed.

Shapiro' gives as an expression for the velocity of a compressible flow between pressures p and po under isentropic conditions:

r u-I 2gkRT'.

T 1-. P_,

u=

S k-1 3

p, where the upstream gas stagnation temperature is To. Since little heat transfer is expected as the steam moves through the orifices, the isentropic expression should be close enough.

If [

]*# inches of water (from Figure F-1) is a reasonable estimate of the steady-state pressure drop actuss the core bypass holes, calculate the mass flow, rh,, through the ring, since:

th, = x p,

u, A2 where K represents a loss coefficient.

If it is assumed that the pressure developed in the upper head approximately corresponds to the

[

]*#*F saturation temperature, the pressure will be about [

]*' psia. The (steady-state) uA2735w.non:Ib 031396 F-5

e pressure drop of [

]*# inches of water across the core bypass holes implies that the pressure in the annulus will be about:

Additionally, take k = 1.3, R=1545.45 ft. Ih/lb mole *R/18 lb per mole = 85.85 ft1 R. The absolute temperature is approximately [

]*"*F + 460 = [ - ]*"*R and g = 32.17 ft/see.8 or 386 in/sec.2

' He velocity through the core bypass holes may then be estimated as:

To estimate the mass flow throt) the holes, estimate the density in the upper head. Since the vapor is superheated, approximate it as a perfect gas:

From Shapiro,m assume a loss coefficient of 0.6 and calculate a mass flow rate of:

under steady-state conditions.

When similar expressions are written to equate the mass flow through the upper support plate, pi need only be [

]*# P to match flow rates, so the assumption that the pressures in the core and -

upper head regions are close is acceptable. His concludes the steady-state calculations.

The interpretation that because the shape of the curve is "ium.ded" representing an actual minimum value (Section 2.0), is not crucial to the analysis. The only interest is in the minimum pressure in the upper part of the annulus after condensation has commenced. Assume that any temperature within the annulus could suddenly be exposed to the steam with' the corresponding saturation pressure.

~ Figure F-3 indicates the temperature ranges of the water within the annulus, from [

]***F to

[

]"#'F, depending on the elevation. For this estimate, assume that the annulus temperature drops to 210*F so that, under saturated conditions, the pressure will be 14.123 psia.

- s:\\2735w.noa:lb031396 F-6

4 For steam, the critical pressure ratio, PR,,, may be approximated by:

so that tue flow will be sonic and independent of the pressure in the annulus with a flow rate given by:

3 k+I kg( 2 rer, p, 4 th

=

8

$ RT k+1, o

Calculate this value to be:

or about 6 times greater than steady-state flow.

If the measured range of the cool water in the annulus is considered, note that for saturation pressures of [

]'** psia in the annulus (corresponding to a saturation temperature of [

]d**F), the flow will always be sonic. With the highest annulus temperature of [

]'***F, the Mach number, M, can be calculated from:

where [

]*# corresponds to the pressure at a saturation temperature of [

]'***F. The small reduction in Mach number at the highest annulus temperatures will not greatly affect the numerical results.

4.0 Temperature and Pressure Changes in the Upper Head From a study of the data, a reasenably steady and stable operating condition had been set up prior to and following the bang. This implies that a rough equilibrium had been reached between the flows into (from both the core and the steam generators) and out of the head (to the annulus) and the heat addition from the steel reactor head. The time rate of change of both TF-120 and SC-141 is only on the order of degrees per minute and that of TF-168 is practically zero.

uA2735w.non 1b4313%

F-7

With exposure of cold water in the annulus, the outflow suddenly, for a brief period, became 6 times greater than the inflow. Treating the water vapor as a perfect gas:

pV = mRT 1

and dp RT dm RT

= _. (th,-th )

_=

dt -

V dt V

Inserting relevant numbers:

If the density of steel is 0.281 lb/in.8 and the upper support plate thickness is 3 inches, then the pressure required to lift the plate will be 3.0 x 0.281 = 0.843 psi. Using the value of dp/dt calculated previously, a sufficient pressure dmp to lift the plate is developed in only [

]**# milliseconds after

. exposure of the cold water in the annulus.

His pressure _ drop will be only momentary and will last until quasi-steady processes develop in the upper head when the pressure retums to near saturation.

He physical mechanisms that occur within the vapor in the upper head region, under these quasi-

. steady flow conditions, will be approximated in the following manner: Water vapor enters the upper head region from the reactor at T, and a flow rate approximately equal to that of the flow leaving the upper head and moving into the annulus. Additionally, assume that complete mixing takes place

- between the two fluids prior to exiting the head. nis process can be written as:

m cg AT = rh, cg T,g At - rh, cg At where "t" represents the time, his is a differential equation when written as:

dT rh r

N - T) annswmib-osu96 F-8

=

and has the exponential solution:

f 3

t T = T, + (T - T,) exp o

r In this equation, t* = mhh repmsents the characteristic time - the only quantity of interest.

Assume that the mixed vapor obeys the perfect gas law so that:

This number indicates that the temperature in the upper head will reach near saturation conditions within several seconds, as shown in Figure 2. Note that the error introduced by not varying the mass

. flow with decreasing temperature is on the order of:

or approximately 5 percent.

5.0 Reduction in the Annulus Elevation due to Accumulator Flows Several aspects of the situation at the time approaching that of the bang are indicative of a worst-case scenario:

j De temperature in the upper reactor head is approaching a high value.

I ne temperature in the annulus is decreasing, resulting in a large temperature difference.

e l

De liquid level in the reactor and annulus is being lowered substantially due to operation of

=

ADS valve no. 3.

He first two phenomena contribute significantly to the possibility of steam void collapse and the latter, to the forces generated if a water-hammer event is initiated.

A fourth contribution develops as a result of the accumulator flows being at their highest levels (about L

[

]* gpm) just prior to the event. After the DVI line penetrates the annular region of the reactor, it tums downward. He high injection rate, in effect, creates a downward-facing ejector pump, reducing the pressure in the vicinity of the outlet and entraining local fluid in its jet.

u:\\2735w.non:lt4313%.

F-9 w_-____--_____________-_-___-___

The pressure reduction will result in a drop in the liquid level above the nozzle given by:

a 2g The velocity is given by:

where the ID of the DVI line is 1.16 inches so that:

Thus, if the annulus level has dropped to within six inches of the DVI line, it is possible that the annulus steam could become entrained in the cold water, greatly increasing the condensation.

6.0 Time Durations and Forces Associated with the Postulated Steam Void Collapse in the Upper Annulus Assume, for the purposes of the following calculations, that the model reactor core acts as an energy and volume source which can develop the necessary void volume to accommodate any motions of volumes of water within the system and maintain the pressure in the core at satumtion.

For these calculations, assume that the water level in the annulus is at approximately the 65-inch level, just to the bottom of the cold leg. 'Ihe cross-sectional area of the annulus is taken to be 177 in.2 Additionally, assume that the cold legs are empty. The cold legs are assumed to be about 5 ft. long and have 3.5-inch ids. The initial volume in the annulus is then 4225 in.' and that of the cold legs, 2310 in.'

Ignore the frictional losses of the water rising in the annulus, and consider only the momentum change requirements as fluid is accelerated from the lower reactor plenum into the annulus. The momentum equation must be written for the general case as:

d(W v), g,

pg,

dt where W and v are the weight and (upward) velocity of the water in the annulus, respectively.

um35wmon:lM31396 F-10

The forces acting on the fluid are the pressure difference between the pressure at the bottom of the annulus and the vapor pressure at the top of the annulus and gravity. However, some corrections must be made. For example, if it is assumed that the pressure at the bottom of the heater is substantially P,,,, then the pressure, P*, immediately at the entrance to the annulus will be the saturation pressure, minus the head loss required to accelerate the fluid to the velocity within the annulus, that is:

Pressure Force = (P' - P ) ' A = (P,- p v / 2g - P2). A 2

2 2

2 and the gravity force is simply the weight of the fluid within the annulus.

Additionally, since the overall length of the column of water in the annulus is increasing as it moves upward, the left-hand side of the momentum equation must be written:

. d(W v),, y, dW +W b=V-vA+W b 2

dt dt dt dt Taking W = p x A, where x is the height of the water in the annulus, the momentum equation can 2

be written:

dv P,,,- P,

1. ft

_=g.

-1

_. y, + v3 dt px x

2 The reason for retaming the two velocity squared terms in the last set of brackets is that the first, with the coefficient of 0.5 is always required, but the last, with the coefficient of 1, is required only when the column of water is growing in length.

- Some simple calculations show that, in the above momentum equation, p x P - P (by a factor 2

of about 12) for cases of interest, so the unity term in the first set of brackets can be ignored. For this order of magnitude calculations, also ignore the nonlinear terms, which are zero at the start of the motion of the annular slug, and assume that over the first few inches:

where the column length has been assumed constant at 65 inches. With this value of acceleration, the cold legs are filled in about a tenth of a second, during which time the velocity of the fluid moving up the annulus has increased to about [

]* ft. per second.

.A273sw on:tbest396 F-11

s The maximum velocity that can be obtained by the accelerating slug of fluid is asymptotically approached as the right-hand side of the momentum equation approaches zero, or:

The pressure generated by the impact of a water slug moving at this velocity on a rigid surface, such as the core barrel support flange, is given by:

where c is the speed of sound in water, about 4800 ft/sec., and the kinetic energy stored in the rising slug of water is 12,200 ft/lb.

The high value of the calculated impact pressure is a purely theoretical one postulated on the impact of two perfectly flat surfaces, one of which is immovable. Such an event is hardly credible and, should one be concerned, we may note that the energy available to damage structure is very low.

7.0 SemaPnary From an extensive study of the data accumulated in the various experiments at OSU, it appears that in exj-a.ts SB001 and SB018 several physical phenomena have evolved during the course of the runs that combine to form a scenario, becoming 'a worst-case event. There appears, however, little cause for concem related to either father OSU experiments or operation of the prototype.

'Ihe argument and cahulational procedures described previously provide a consistent and reasonable framework for evaluating the potential results of steam void collapse water hammer in the AP600 low-pressure integral systems test facility at OSU. They appear consistent with experimental observations and provide a physical basis for understandmg the mechanisms involven m the rather complex phenomena under consideration.

=vns.

ibesue6 -

F-12

1 EM Shapiro, A. H. The Dynamics and Thermodynamics of Compressible Fluid Flow. Vol. I, he Ronald Press Company,1953, pp.109.

l i

I l

I t

i l;

i l

i eM735w.noa:Ib-031396 p.]3

d I

- Due to its pmpnetary nature, the following text has been deleted from the mn-proprietary version of this document.

mm35 man:lw3Im F-14

APPENDIX G DRAWINGS

- uA2735w. mon:Ib 031396 G-1 u____.:_ __ _ _ _ _ _ _ _ _. _ _. _ _ _ _ _ _ _ _ _ _ _.

1 1

L 4

l l

I I

E

(

l

=

i

=

=

=

5 l

{

Figure 1 eM735w. mon:lt4313%

G-3

.J

s i

t m

X 5

l I

i l

lI i

i i

i g

i g

R I

I d_

\\n 5

$E3%

[

g

[

rilnlg!

E I

i i

1. w g

i

%~

l~

~

j

[

E I

,, \\

,, j\\

i 1

g s

1 11

\\

I I

=

a i1 F

i 2

i I

\\

l l

J\\l

(

S E

ri --(-

-F-ri Py fg D

f 1

- (y s

i t4

'N

"\\

i i

i 3'

s N

g

_ _ _'i

_ t._p 8 i ii

~'

,,'ar' sM y'

/

5 r

D g

h Figure 2 sA2735w. mon:!M313%

G-4