ML20009E303

Transcript of GL Sozzi on Doherty Contention 32 Re ECCS Vaporization Rate

ML20009E303
Person / Time
Site:	Allens Creek File:Houston Lighting and Power Company icon.png
Issue date:	07/20/1981
From:	Sozzi G GENERAL ELECTRIC CO.
To:
References
NUDOCS 8107280021
Download: ML20009E303 (6)

Text

1 July 20, 1981 dr

,i

,T<

L. i y v

)

1 i

1 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION g

2 T ' $.; ] w.

m '

p" gg3DEFORE THE ATOMIC SAFETY AND LICENSING BOARD 3, _S 7

- -4 4

In the Matter of S

g S

su 5

HOUSTON LIGHTING & POWER COMPANY S

Docket No. 50-466 6

(Allens Creek Nuclear Generating

,- (\\

Station, Unit 1)

S

\\

N 7

S S.

1 s

g g ggBl" i DIRECT TESTIMONY OF G.

L.

SOZZI g

ON DOHERTY CONTENTION 32 C

j\\)L s 8*

RE ECCS VAPORIZATIO:: RATE

\\~

M p,

g N

Q.

Please state your name and job positicn. k 10 11 A.

My name is G.

L.

Sozzi.

I am presently Manager of 12 LOCA System Technology within the Nuclear Fuel and Services 13 Engineering Department of the General Electric Company.

14 Q.

Would you explain your job responsibilities and 15 your professional qualifications?

16 A.

I am responsible for experimental and analytical 17 support of safety related characteristics of the Boiling 18 Water Reactor (BNR) system.

The BWR Blowdown / Emergence Core C

ling Program is one such investigation under my direction.

19 l

I received a B.S.

degree in mechanical engineering 20 in 1966 from San Jose State University, an M.S. degree in mechanical engineering from the University cf California, and g

(

I also attended Stanford University where I performed addi-23 tional graduate work following my M.S.

degree.

I have spent 9503

1. 8107280021 810720 l

PDR ADOCK 05000466

/

T PDR

1 1

most of my professional career in thermal hydraulic experi-2 mental and analytical research and development relaned to 3

nuclear reactor systems safety technology.

A major portion 4

of my work has dealt with the two-phase flor phenomenon relat-ed to the reactor loss-of-coolant accident (LOCA).

This work 5

has included critical two-phase blowdown flow, two-phase 6

flashing and level swell, and heat transfer during a vessel 7

blowdown.

g I am a member f the American Society of Mechanical 9

Engineers and a registered professional engineer in the State 10 of California.

Q.

What is the purpose of your testimony?

A.

My testimony addresses Doherty Contention No.

32 which asserts that the General Electric Emergency Core 14 Co ng Systems (ECCS) evaluation model underpredicts the 15

_ation of steam during ECCS injection flow after a loss-9 c..

16 of-coolant accident (LOCA).

The source of Mr. Doherty's 17 postulation was the report of some anomalous test results at I

18 l

GE's Two Loop Test Apparatus (TLTA) facility as part of the 19 Blowdown / Emergency Core Cooling Program.

l 20 Q.

Would you describe the TLTA tests?

2' The Blowdown / Emergency Core Cooling program is a i

A.

cooperative experimental research program jointly funded by 22 23 the Electric Power Research Institute, General Electric and 24 t

i l

,_.. - _ _. - - - _ -, = - -

m

1 l

1 the Nuclear Regulatory Commission.

Tests are conducted by 2

General Electric under this program in the Two Loop Test 3

Apparatus in San Jose, California.

The purposes of the 4

present program are:

5 To simulate the LOCA from the start of the accident through the early interaction with ECCS in a configuration which has performance characteristics 6

similar to a BWR with 8x8 fuel bundles.

7 To obtain test information to support analytical predictions used in the evaluation of BWR's during 8

postulated LOCA's.

9 The Two Loop Test Apparatus is an experimental 1 0 scaled mockup of a BWR.

A single, electrically-heated, full-11 size (8x8) fuel bundle simulates the core and is contained 12 within a pressure vessel simulating a reactor vessel.

Both 13 normal and emergency cooling systems are simulated.

Two 14 loops circulate water to jet pumps within the pressure vessel.

15 Other major BWR components are also mocked up including a 16 steam separator.

Emergency core cooling systems include scaled 17 high and low pressure core spray, low pressure coolant injec-18 tion, and auton.atic depressurization.

Blowdown of the reactor vessel for a simulated LOCA 19 is initiated by operating quick-opening valves in one of the 20 re ir ulation lines.

Blowdown fluid from these lines is 21 dumped into a tank of water which condenses the steam and absorbs its energy.

There are approximately 180 data channels dedicated l

1 to the Two Loop Tests Apparatus which are recorded on magnetic 2

tape.

Detailed data analysis and reduction is performed on 3

the GE computer system, 4

0 What is the source of concern expressed in Mr.

Doherty's Contention?

5 A.

During the Fall of 1978 a TLTA test was conducted 6

with an average power bundle (5.05MW) and with average 7

Emergency Core Cooling (ECC) injection flow.

Results of g

this test were then compared with those from a test with the g

same initial conditions, but no ECC injection.

The comparison showed that the system depressurized more slowly with ECC injection than without ECC injection.

The slower depressuri-12 zation with ECC injection was not anticipated by the NRC 13 Staff and as such, the Staff requested that GE review the 14 results and account for the apparent differences observed.

15 The Staff's concern was based on the preliminary 16 conclusion that the slow depressurization in the TLTA test 17 with ECC injection may have been due to greater steam genera-18 tion in the core than expected, which could result in an

~9 1

unanticipated delay of reflood.

This possibility led to a 20 concern that the vaporization model was in error and non-21 conservative in its effect on the calculation of peak clad 22 temperature (PCT).

However, upon completion of its review 23 of the data, General Electric resolved the Staff's concern 24 about increased vaporization.

The slower depressurization - - -

1 for the test with ECC injection was actually due primarily to the fact that the fluid exiting the break was of lower 2

quality than for the test without ECC injection.

For the test 3

4 with ECC injection the emergency core cooling water refilled the system and collected in the vicinity of the break and 5

was carried ott of the break with the steam.

For the 6

test without ECC injection, the fluid discharged from the 7

break was primarily steam.

Some additional steam was generated g

by the ECC fluid cooling the metal masses which are part of the test assembly for the test with ECC injection.

This effect would not be as significant in an actual reactor because the 11 ratio of metal mass to ECC fluid mass is higher in the test 12 assembly due to the scaling parameters used in TLTA testing.

13 Q.

Was GE's account of the slow depressdrization 14 anomaly otherwise corroborated?

15 A.

Yes, data collected on the density of fluid flow cut 16 the simulated break, and calculated mass and energy balances

~7 1

on the system confirmed that the observed differences in 18 depressurization rate were due to differences in fluid quality 19 the break and not due to core vaporization rate.

In out of fact, estimates of the steam flow out of the core region, 20 21 from measurements across the steam separator, demonstrated 22 that there was less net vapor generated from the core region 23 for the test with ECC injection.

24 Subsequent to these earlier tests, GE repeated 1

these tests in the TLTA with improved instrumentation.

These 2

latter results provided further substantiation to the earlier 3

results, i.e.,

the slower depressurization rate for the tests 4

with ECC injection was due to a lower volumetric discharge from the break and not due to core vaporization.

5 Finally, a comparison was made between the peak 6

cladding temperature calculated with the GE ECCS model and 7

that measured in the average power TLTA tests, with and 8

without ECC injection.

The calculated peak cladding g

temperatures excee1 the measured values by approximately 1000 F in both cases.

Thus, the calculations indicate that 11 the Jicensing models maintain a large and consistent margin 12 of safety in the prediction of peak cladding temperature for 13 the two test cases.

14 Q.

Would you please summarize your testimony.

15 A.

GE's ECCS model has conservatively accounted for 16 the phenomenon observed in the tests referenced in the 17 contention.

It is clearly established that no revision in 18 the ECCS model is required.

1 19 20 21 l

22 23 l

24 I

.. -.. ~

.,,