ML19247A264

Forwards TMI Run 106,analysis of First 2,550-s of Transient. Calculations to Be Continued After Reactor Coolant Pumps Tripped Off & Heatup Phase of Accident Started

ML19247A264
Person / Time
Site:	Crane
Issue date:	07/03/1979
From:	Behling S, Hendrix C EG&G, INC.
To:	Throm E Office of Nuclear Reactor Regulation
References
NUDOCS 7907300464
Download: ML19247A264 (7)

Text

.

p f [3 l

F 1:

3=e i

DATE:

July 3.1979 f

(

TO:

Ed Throm, MAC-DSS r_f

/

j FROM:

S. R. Behling and C. E. Hendrix. EG&G Idaho

SUBJECT:

ANALYSIS OF THE THREE MILE ISLAND ACCIENT 1

J An analysis of the Three Mile Island (THI) accident has been partially co@leted. The calculation of the first 2Z50.0 r econds of the transient is corplete and it is planned to continue the ralculations past the tilne i

when the reactor coolant pumps were tripped of f and the heatup phase of

}

the accident started.

j The study pmsented hem was designated Three Mile Island Ran 106. This

}

. analysis reflects several important improveraents over an analysis presented

, i earlier to the NRC, designated Three Mile Island Run 9.

These incluoe a

}

nere detailed RELAP4 model of the mactor and changes in the initial ton-j ditions and events.

The results from this calculation are precising. The analysis will be continued i

to find out if the core heatup phase of the TNT accident can be predicted.

1 Resul f.s of a calculation to 2.5 hrs of transient should be donc by July 6,1979.

f f

Initial conditions - Three Mile Island Run 106 5

Reactor Power

- 2568.0 Nt j

System Pressure

- 2156.0 psia (hot leg pressure)

}t i

37562.6 ltr./sec j

Core Flow Hot Leg Flows

- 19153.3 lba/sec j

30.0 ft

}

Pressuri;er level Steam generator mass

- 34900 l be

(

Steam oenerator levels e 25.96 ft 1

A set of trip controls was input that simulated the sequence of events which occurmd during the accident. These controls am:

[

'I Event Setpoint 4

S.G. feedwater flow off 0.0 seconds (12. second ram)

Pressurizer valve opens 2270.0 psia in a loop hot leg l

Reacter scram 2370.0' psia in a loop hot leg HPIS start (2 pumps) 124.0 seconds HPIS off 278.6 seconds 1

I Makeup flow start 278.0 seconds (140 gpm) s Let down flow start 300.0 seconds (140 gpm)

{

5.G. auxiliary feed on 480.0 seconds 1

l Trip pum B-loop 4440.0 seconds 576085 Aux feed to A-loop S.G. off 5100.0 seconds f

I j

g 90'ZS00f C,

u-mi a

t If

'i To: Eo Thrum, MRC l

Page 2 l

l I

Event Setooint j

e Aux feed to B-loop S.G. off

' 5220.0 seceNs l

Aux feed to A-loop S.G. on 5650.0 seconds l

1 Trip pum A-Loop 6000.0 seconds f,

Aux feed to A-loop 5.G. off 7a40.0 seconds Snut EWJY 8280.0 seconds i

t

.RELAP_4 Mode 1 ll The RELAP4 model used in the analysis is showp ]in Figure 1.

The model was originally developed to check out RELAP4/K)DSLl.

It was then nodi fied for analysis of the TMI transient. A description of most of the canponents of the model is contained in Reference (1).

The HPI5 is modeled with Junctions 31. 38, 39, 40, 41 and 42. Juncticns 31, 38 and 39 model the nomal HPI5 flow while Junctions 40, 41 and 42 model the 1

flow with one puw only.

There are uncertainties in the steam generator conditions during the early l

part of the transient. The turbine bypass flow (Junctions 26 and 29) which was given as 15t of nomal flow, the steam generator relief valve flows

{

(Junctions 27 and 30) and the steam generator auxiliary feed ficw (Junctions 1

26 and 29) are uncertain. Steam generator conditions are critical to the j

analysis because the energy recoved from the reactor system is dependent

(

on these conditions.

To eliminate some of the problems in calculating steam generator conditions, volums 23 and 24 were added.

These are steam filled l

time-dependent volumes with pressures set to etch the measured steam

'j generator pressures during thf accident. Insertion of these voltanes will j

hold the pressures in steam oenerator voltsnes 20 and 21 at the ceasured i

pres su res. The trean volume flow and volune quality in 29 and 21 will still be functions of the input turbine bypass flow, relief valve flows, aM auxiliary feed flow.

fi The RELAP4 bubble rise model was selected for volumes 1, 2, 3, 4, 6, 14 f

34, 36, 37 and 38. This was done to allow a calculation of core dry out during the heatup phase of the accident. The computer calculation time for i

the analysis was very long when bubble rise was selected for the entire 5

tran51 en t'.

Therefore, the bubble rise calculation was not started until I

after the reactor coolant pumps were tripped off.

j t

The pressurizer was modeled as a vertical stack of six voltsnes with i

vertical slip selected at the junctions connecting the volumes. This j

method has been proven to be the most accurate and efficient cethod of

(

modeling swelling phenomena such as occurs in the pressurizer during this

}

transient.

J3 57GOEG ia b

P00RORIBIE 1

a

k 4

e E

• }'

.. To: Ed Throc, NRC g

2

p. y.. -

Page 3 l

1 x

g-y,

!}

71 7

"l f*

0, J[

=

o 2

g]'

wm

.=.

e7

.a.

r EM E M.I k-w C

P y

h "

an,

.e

-- _h g:

sg s

r-

~,;.

1 JL g-t iI, s

P' gR

.=

y a

g, 4.

g

.y gib 1

t 1

J.

1 g.

t i

~J

\\ y

. ~ i* 2 -

.m' tt 4

l

..n y

'~_ '"

g'. ~9f

?

O a

o

]

x

_3 E

y

~

l f5 I

f

'g "

T

.m

' [L _-

t-

[

t, 4

C C

g

E N

1 4

~

}

p, m

~

l 1

y-

<.a t

I ? m~

f

? N.

t q

M M

m n

4s Ig m

o k.

U..

5 pgy ng hW_

'QUl$ 0hf))lj)]},\\'

e e

T l.

4 To: Ed Throm, NRC 1

Page 4 d

q I

Preliminary Results Figure 2 shows the B-loop hot leg tmperature plotted versus the reasured data. There la a sharp drop in tecperature after reactor scra at 10 seconds. When the steam generators dry out at 60.0 seconds the temperature I8 increases. The RELAP4 calculation matches the trends shown in the data at these times.

Af ter steam generator dry out the te@eratum increases faster in the code calculation than in the seasured data. Apparently. the HPI5 flow o

prevents a sharp tecoerature rise but this trend did not appear in this calculation. When HPI5 flow is tumed off at 278,0 seconds the seasured temoerstures increase at the same rate as the calculation.

The te@eratures 1

peak at 480.0 seconds af ter the steam generator auxiliary flow starts. The n'

rate of te@erature decrease after (80 seconds is smaller in the calculation

}

than the data indicates.

This is believed to be due to heat transfer I}

1 calculations not reflecting the steam generator tube rewetting.

1 Figure 3 shows the RELAP4 code calculation of A-loop hot leg pressure and measured data. The measured data shows that the pressure in the TMI reactor decreased during the first 350.0 seconds of the transient. A pressure j

increase then occurred between 350.0 to 500.0 seconds. The rise was caused

[

by the liquid mixture in the pressurizer rising to the top. This decreased i

the volumetric flow rate through the pressurizer relief valve. At 500.0

(

seconds the system started to depressurize again due to cooling caused by j

the steam generator aux 111ary feed flow.

As seen in Figure 3, the calculation does not match trends shown in the data i

during the first 500.0 seconds of the transient. This problen is believed

{

to be caused by an inaccurate prediction of pressurizer behavior during the

}

early part of the accident. However, after 500.0 seconds the calculation matched the measured depressurization rate.

l Figure 4 shows the RELAp4 calculation of the reactor core power overlayed g

with the power transferred through both steam generators. This graph 111ustrates the importance of correctly calculating stea:n generator L

behavior during a scull break transfent because there is a mismatch between f

power in and power out throughout the accident.

In this analysis, the h

results indicate that the power out exceeds power in af ter 500.0 seconds

{

vhich resulted in the te@erature decrease seen in Figure 2.

I i

11 11 a

k 57G0$$

i

11

~<.

I O

Q O

. m r0

.-t.

C4 C

CD

^

fJ g

5

/

J E

5

=

g ore

=0-m Gwe g

n 3 t-- Q I

g I I a*

C

./

i g

=w e

m g

a mg

~

gE58 e

a

~

w a

S M

C wo mo i a

CD 1E Q

HN N o

4 (D

O g

J d

=

=

CD g

sN N

d e

o o

e o

o en e

o o

O Cr.

CD r-e W

v Ps N

W W

W"I W

ID W

W W

(D 1A I

14=

l[

t j

.i i

s

'f.

,e

.c.

e e

Q

.L rn 4

O

,G

^

m. ~

J' e

t I

D w

o E

N CD V

p N

OZC

' W e

eu

=QH W

c; G*C' M

c )

A =d W C '

MC D-c; o

M JQ

1. J w

3W E

m m

rJ W

yya m

4" p

-p n

JJD

( J cm g

m gEOC J

g o

m W

C WNE C

J H

wo

%O I m

I C

c

=E E

CD M\\N g

.g m

c.

CL M

CL C

o J

ii y

M C

.b D

)

j[

c)

O O

C C

CD,

O C

G w

v oj O

CD ti., O O

O to O

O C3 O

O v

N O

N N

N N

• =*

CW3 ed 3003DRM1L s2cwo

~

g g

t E

O

=

r o

e c

'4 29 (e

e

,iC k'$

-4 G.

O N

W to Jq

(

g O

b no

%H O

E W

rJ.

4 eg

=

m m

O m N

~Zw W

N y

CW g.

g D

W=

c a. 4

^V W

g 3C O

M O LaJ q'

oq W

W,y%

f2 H En ~

s.)

9 0

V 1

=-

s..

W=r m

e 7

Ou r

c2

=

wyO 8

oo M

O h

J Cl I E

( J.

"~2 I

(J.

i O

NW W

.,. r (1-

~

f gh>

0 c t--

=

G W

rJ- -

ya m

W ca--

g 3

O CM yy 1

( > 4 gg ck m

I

} d H

e O

L N

e 4

,[ ^

D em

. 4 c 2 0 4

( d l

f i C f Js s __

(

O m

o CD W

W M

Q W

'D v

N O

-N w

=

w Q

G, C

O O

{

o i,

X C V 4G7 898kd dM.l. %W 576G91 t.

4

_