ML20037B551

Summary of 760323 Meeting W/Ge in Bethesda,Md Re Core Spray Distribution Tests & Proposed Changes to GE-ECCS Evaluation Model

ML20037B551
Person / Time
Site:	Dresden
Issue date:	05/17/1976
From:	Frahm R Office of Nuclear Reactor Regulation
To:	Novak T Office of Nuclear Reactor Regulation
Shared Package
ML20037B550	List: ML20037B549 ML20037B551 ML20037B552
References
NUDOCS 8010170714
Download: ML20037B551 (6)

Text

I E.T.T..

Esd :

UNITE 3 STATss

/

NUCLEAR REGULATORY COMMISSION

==:

[

k WASHWGTtW. O. c. 30ss5

. u-d e

S

• MAY 17 W6

' j=i-r==:

==:-

i ir... u-

. Thomas M. Novak, Chief, Reactor Systems Branch, DSS THRU:

W. Minners, Section Laader, RSB, DSS //[

l5l%

I

~

SUMMARY

OF MEETING WITH GENERAL ELECTRIC COMPANY L

1, A meeting was held with General Electric in Bethesda on March 23, 1976, 1

==

to discuss core spray distribution tests and proposed changes to the GE-ECCS Evaluation Model..

f..E:=

i GE personnel presented a progress report on the core spray distribution tests and their. implications for operating reactors. GE also discussed j

potential i=provements applicable to their inventory and core heetup models.

• ':.~.*[.

l The enclosure provides more detail concerning the items discussed.

2ff Y. :.'. 11~

i Ronn1d K. Frahm hx-Reactor Syste=s Branch l

Division of Systems safety c.

Enclosures:

1.

Me-ting Minutes 2.

List of Attendees 3.

Slids Presentations cc S. Hanauer V. Stello D. Ross T. Novak Z. Rosstoczy R. Baer R. Woods 1

W. Minners i

ACRS (21)

NRC PDR R. Frahm

-e W

t

..l er -

E..U 4

t

.; E s=1 l

cmh

" "E l

I.

Core Spray Distribution Tests

==

General Electric (CE) presented the results of tests with single

9 nozzles in a steam environment,' Nozzles of the high. velocity

".Z $

j type with a fine spray showed some e,ffect of cone contractions.

' ~ ~ ' ' '

Nozzles of the low velocity types with coarse spray were little

= = -

affected by steam. The WC type nozzles, although they produce rE=E=

a coarse spray have a deflector and showed abrupt cone contrac-

_.. =9 -

tions under c.ertain pressure / temperature conditions. For BWR/2,

~;ll.... _

j 3 reactors the fine, high velocity type nozzles are.used to cover T4 E=

the peripheral fuel while the coarse low velocity type cover the

?

remainder of the core. Therefore it is concluded that core spray

.. = =

distribution will be little changed by a steam environment in

. =. = =

j BWR/2, 3 type reactor configurations. BWR/4 reactors use WC (M

j nozzles to cover the peripheral and center region assemblies A.@=;

with the fine, high velocity type covering the intermediate region.

i.E j

However, the WC nozzle. cone contraction that affects spray

.7,

distribution maybe compensated for by the superposition and

=Ei interaction of the two spray spargers.

==

==.

]

, GE also presented material to quantify the observed contraction for

.j different reactor types. BWR/2, 3 and'Vermonc Yankee have two i

3 nozzle types; an atomizing nozzle (covers periphery) and an open j

elbow nozzle (covers remainder of the core). Because the atomizer l

nozzles cover the periphery, the sensitivity of the angle of elevation on cone narrowing was investigated. The results showed

~

that even with a severe position change no peripheral fuel bundles

=1. g l

will starve, even though some of the edge channels are below

=L j

the design min'imum flow rate. The sensitivity of the BWR/2, 3

.ar:

open elbow nozzles (coarse spray) to steam showed that patterns "Ek in air and steam are nearly identical at one atmosphere; showed fr FF a concentration in the center of the cone as pressure is raised

=F-i with a maximum occurring at three atmospheres and remaining constant at high pressures. The tests showed that BWR/2, 3 type j

nozzles are less sensitive to a steam environment than the WC 7

nozzle used in the tests, therefore no further spray distribution i.

tests are planned for BWR/2, 3 reactors.

r; 1

= '

{

The WC nozzle used on BWR/4, 5 reactors showed an abrupt cone con-

==

traction in steam. The WC nozzle, without a deflector in air,

, gave the same spray pattern as an open albow nozzle in steam. The

=== -

i WC contracted pattern in steam is twice as broad as an open elbow E====

nozzle in steam and a WC without deflectors in air. The simulated WC cone contraction in air showed a significant change in spray

...M distribution with WC deflectors removed. For the single sparger

"=:"}-

us _

m..

..^"

E.......

j

l T-l l'

(Y.7.!.!..

[d.u.

.....,7

• *A *.;*

i 77i.t:f.,5 j

==

J VNC (w/o deflector) tests, there was full coverage of the core;

;;.;;;g with a minimum channel flow of 2.09 GPM (with updraft). Between

g=

18 to 26% of the channels had less than the 3.25 spa =4n4== design requirement. The tests showed that air updraft also improved j

i flow coverage in the low flow regions. The double sparger tests, t,;;;;;,.;...

;=e.s with no air updraft, resulted in full distribution with no channels below the minimuth design flow of 3.25 gym. Again there was a

~~ W.~5 significant change in spray distribution with the VNC nozzle deflectors removed. The siculation of the VNC cone contraction in air showed that the superposition of flows from many nozzles contributed to uniformity of spray. distribution and BWR core spray i

'l - -

digtribution can tolerate cone angle changes without degradation

""~.=1 of distribution.

1 3

i l

With regard to the ECCS capability of a EWR/4, the two sparger tests 7

.3-..

showed that the 3.25 gym minimum flow requirement is exceeded. For

, ~"

i a design basis LOCA, assuming a single failure of the LPCI injec-Pl5

{

tion valve (2 core spray systems available), the reflood time is

...]

240 seconds and the peak cladding temperature (PCT) is 2200'F with

.."::l;

=

i full core spray heat transfer. For an assumed diesel-generator

~

single failure (2 LPCI + 1 core spray failed) leaving only one j

core spray sparger'available, and with the test results showing

~

=

1 that only 75% of the channels received the minimum 3.25 gpm; I

the calculated PCT vas 1950'F (at a reflood time of 110 see) assuming l

half core spray heat transfer. GE felt that these results show that the LPCI failure is still the worst single failure and that

=

j

=

j, the current MAPLHGR limits still apply.

j Focure experimental programs to be undertaken by GE include:

distribution tests of all GE nozzles in the single nozzle ASEA g.=,

facility; single steam nozzle tests to quantify nozzle performance;.

~

I and full scale tests with modified nozzles to simulate horizontal

. :12 l7 steam test results. GE wants to simulate the cone angle change

=

i more precisely in air tests (because the effect of cone angle may 1

be overpredicted by the air test).

Future analytical work includes: single-droplet model development single-nozzle model development, and developing a multiple-nozzle interaction model. A core spray distribution model will be N

developed to investigate the sensitivity of BWR/4, 5, and to facilitate the design of BWR/6.

_ 3. g tj II.

Evaluation Model Improvecents

1. o 1.

Reflood Model Experimental Programs

= =:

GE plans a study of steam-water interaction effects in a BWR

.. k system during a LOCA. Separate effects studies include single

= "=1

%=~.

g is l-

=

]*

' - =. ~:

-i

, - +

=::.=:

1 pE iff n=

WEE =h

== 1 d

sui = b and parallel bundle studiss with thermal-hydraulics in the-

[ =iiN upper plantan. The system features incorporated in the test

~~ "

apparatus include representation of two bundle flow paths, one

M

l bypass region flow path and ona jet pump flow path. Adiabatic s

tests are being run with steam injected into bundle flow

~~= :";-

j paths' to simulate vaporization, and the two-phase charac-f}

teristics are being evaluated with attention to counter-flow g "y-of liquid and steam at the flow path entrances and exits.

=

a Heated tests in which bundle flow path steam is intsrnally

~

generated are also being run. This test program is being s

run to quantify parallel channel hydraulic stability and g=g :

CCFL breakdown.

j GE is considering a one-sixth scale reactor simulation for

=5 investigating thermal-hydraulic phenomena in the upper plenum jf@

region. Adiabatic air / vater tests and steam / water tests vill

=27" investigate multiple-channel CCFL effects, including effects on core spray distribution and temperature distribution across

.. g the top of the core as the liquid is he.ated by the steam upflow.

5 j

i The effect of system feedback on CCFL performance characteristics is being investigated in the single bundle /ECC System test apparatus.

In this test CCFL performance is evaluated under m=r transient conditions representative of the LOCA,, including feedback from the system hydraulic characteristics.

m:

Modifications to the two-loop test apparatus (TLTA) are also planned. Interaction effects of ECC injection on t'he bypass j

and bundle exit CCFL characteristics in the upper plenum will 9

i be evaluated. Later testing will examine BiiR reflooding

~

characteristics.

1

{

Proposals for a large scale integral ECCS facility will be made t

to EPRI and NRC in late 1976. This facility will use full scale reactor components including 32 full size bundles incorporating actual tie plates, spacers, channel boxes, control rods, guide tubes, and steam separators.

.

1 Early results of these experimental programs show conservatism in the treatment of CCFL in the reflood calculation; further experiments will be used to develop more realistic correlations to predict reflood times.

2.

Core Heatup Model Potential Improvements T

A radiation model coupled to the fuel clad swelling model using W=?= -

individually calculated radii; and grey body factors recalculated each time a perforation occurs, will be incorporated into CHASTE.

~

=

j I

~

e

a

- mg

((

(r.

A modified Bromley-Ellion correlatica will be used in the post

==

nucleate boiling; post lower-plenum flashing; fuel channel heat l

transfer (when covered by bypass inventory); and post re-

=....

flooding regimes.

?E.........

The cenduction model will be improved by using 11 nodes vs 4

~~ ~ET nodes, constantAvolume vs constantA radius, full implicit integration, and better time step and noding sensitivity.

i 3.

SAFE Code Potential Improvements The improved model will use discrete physical regions (instead

=- -

of large thermodynamic nodes), hydraulic modelling will consider void distribution within nodes, inventory redistribution will be considered in all nodes, CCFL will be considered at restric-

"~~

tions. A homogeneous critical flow model with choked conditions at two regions, to be incorporated into LAMB and SAFE, is now in progress.

GE plans to show the overall conservatism in the evaluation model and provide a basis for reducing operating restrictions, e

4 s

4 e

4 iA :.

A ?.*.

~

e.

f

=v _

g n;

u.

=. = -

ATTENDEES - CORE SPRAY MEETING

-=

=... =..

__=

5?

=::

NRC GE I

R. Frahm P. W. Marriott

=:

W. Minners A. J. Levine R. Wood's R. E. Schaffstall (Bethesda GE)

~

W. Hodges J. Guibert

~

e-J. Thomas

...J Y.Y.Hsu(RSR)'

N. Zuber (RSR)

P.Boe.hnert(ACRS)

O

?

=&:{

4 9

O 5

d

=: ::

'" :."' b.:

'C l

_. _. - -.. _,.