ML19220C272
| ML19220C272 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 04/11/1979 |
| From: | Satterfield R Office of Nuclear Reactor Regulation |
| To: | Tedesco R Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 7904300123 | |
| Download: ML19220C272 (42) | |
Text
b t
[pnarc%
o UNITED STATES 3V 4'4 NUCLEAR REGULATORY COMMISSION 5 7 Y h
WASHINGTON, D. C. 20555 Yl%, % }f APR 1 1 }g79 MEMORANDUM FOR:
R. Tecesco, Assistant Director for Reactor Systems, a..
THRU:
S. H. Hanauer, Assistant Director for Plant Systems, DSS FROM:
R. M. Satterfield, Chief, Instrumentation and Control Systems Branch, DSS SUEJECT-A'i ASSESSMENT OF PRESSURIZED LE"El INSTRUMENTATIM It'SMLLED AT TMI-2 1.
INTRODUCTION AND SUM:'.ARY An analysis of the recent events at TMI-2 indicate that, for a number of reasons, the information provided by the pressurizer level instru-mentation may have been in error.
This memo describes the instrumenta-tion and provides an assessment made by DOR / DSS staff of the potential for false instrument indications during the event.
The general layout of a typical pressurizer level instrumentation system is given in Fig. 1.
Three systems are installed.
For each, two impulse lines connect to the pressurizer; one near the top and one near the bottom.
The lines are routed to a differential-pressure transmitter, located near the bottom of containment in the annular region between the shield wall and the containment wall.
There are a number of factors that could affect the accuracy of the level instrumentation.
If the liquid density changes due to a tempera-ture change, the calibration in height (level) units will vary; in TMI-2, this is corrected automatically and continuously by a temperature instrument applying a correction in the level readout instrument.
There are several other factors which could affect instrument accuracy in a depressurization event, and which are not normally accounted for.
(1) A rapid reduction in pressurizer pressure could cause liquid to flash in the reference leg (the line connecting the transmitter to the pressurizer near the top of the vessel - See Figure 1).
Such flashing, should it be significant, could cause the instrument to indicate a falsely high pressurizer water level.
(2) Degassing of liquid in the reference leg could also cause an error.
Dissolved gases could rapidly be driven out of the reference leg by this mechanism, and the level instrument would again indicate a falsely high level.
0,)
Uub n
n.
7904300/25
R. Tedesco
?
(2) Should the pressurizer decressurization occur raoidiv. a venturi effect could in principle ce crea:ec a: :ne coint enere tne rererence leg joins the pressurizer vessel, If this occurred, liquic couic be drawn out of the reference leg causing the same inaccuracies in level indication noted above.
The importance of each of these effects has oeen assessed assuming conditions which existed at TMI-2 prior to and during the event.
Calculations were perfarred te estimate the =##e:::
' r ~ + ~ =ad d m ssing.
'9ile ?bc calculations indicated tna; scoe fiasning ca s e cccuc, tne ree c i c..
water level in the reference leg due to flashing is estimated to be less than one foot.
Because the distance between the taps is about 33 feet, tne effect of this reduction would ce smali.
Caiculations also iaci-cate that the effect of degassing of liquid in :ne reference leg is necligiele.
With regard to the venturi effect, it is estimated that gas velocities at the upper level sensing nozzle are too low to produce any significant effect.
We conclude from these assessments that the errors in level instrument indications during the event at TMI-2 were not large.
In particular, potential effects that could cause falsely high indicated level were assessed not to be significant.
Therefore, the increasing level indicated by the instrumentation beginning about one minute into the event at TMI-2 is believed to have resulted from an increasing level of water in the pressurizer.
No inferences are drawn in this memo regarding the level of water in the reactor vessel, as distinguished from the pressurizer.
2.
DESCRIPTION OF PRESSURIZER LEVEL SENSING INSTRUMENTATION AT TMI-2 The pressurizer is a right circular cyclinder with a hemispherical head and base. Operating pressure is 2155 psig and operating temperature is 648 F.
The normal water volume is 800 ft.3 and steam volume is 700 ft.3.
Overall height is approximately 45 ft.
The inside diameter is 84" and wall thickness is 6".
The pressurizer is surrounded by 6"20f reflective insulation whose heat transfer coefficient is 65 BTU /hr/ft. The bottom of the pressurizer is at approximately elevation 312' (See Figure 1).
Three sets of nozzles are provided for level instrumentation, one set for each of three redundant instruments.
Each set consists of a nozzle in the steam space at the top of the pressurizer (at about elevation 349' 6") and a nozzle in the water space near the pressurizar base (at about elevation 316').
The three instrumentation systems are similarly configured.
Attached to each upper (steam space) nozzle is a 1" schedule 160 pipe which extends outward about 3" from the nozzle to a 90 elbow.
A vertical section
(~ 3" long) of 1" schedule 160 pipe is connected to this elbow as shown in Figure 1.
A 1" globe valve is attached to the vertical 1" pipe.
The 87 09
R. Tedesco botto-of this valve is at elevation 3f?' 9" P tached to the valve is a 1" to 1/2" recucer which connects in turr, to 1/2" tuoinc.
This tubin; is presumed
- to run vertically, about 6" from the pressurizer insulation.
from the reducer down to a point near the tubing coming from the lower nozzle and valve (elevation 316').
These two tubes tnen are routed togerner to the shield wall and through a penetration in the shield wall at elevation 292' 8" After exiting the shield wall the tubing is routed to the differ-ential pressure transmitter which is mounted in a rack (centerline of instre rts lccate d at elevaticr. 2?c' The transmitters ara Bailey Meter Co. type BY3B40X-A (level range 0 to 400" of water).
3.
Ai!ALYSIS OF l EVEL SEilSING INSTRUME iT ACCURACY DURIN3 THE TMI ? EVE:.T Three phenomena which could affect the water level in the reference leg of the pressurizer have been postulated.
Each could affect the accuracy of the pressurizer level instrumertation during depressurization and could give a falsely high indication.
These phenon.ena have been analyzed assumina the conditions which existed before and during the TMI-2 event to assess their impact on the accuracy of pressurizer level instrumentation.
3.1 Assessment of !.Se Effects of Flashing in the Reference Leg A calculation of the axial temperature distribution along the length of the reference leg wu mde to cetermine the amount of vaporization that could occur in this line when system decressurization occurs.
The configuration is as described in Section 2 and Fig.1.
The 1/2" tubing below the globe valve was conserv'tively assumed to be equivalent to schedule 160 pipe.
The valve body wa represented by 6" of 1"-schedule 160 pipe. This is a conservative assumption when calculating heat con-ditions and heat losses to containment atmosphere.
Assuming a containment atmospheric temperature of 100 F, a heat balance (See Enclosure 1 for details) was performed to estimate heat conduction from the region of the elbow down the reference leg.
These calculations show that. at a point I foot below the globe valve, the temperature in the reference leg has reached equilibrium with the coniainment tempera-ture.
Elbow temperature was assumed to be 650 F, so that the gradient to 100 F would take place within 2 ft. of the elbow, giving a gradient of -23 F/ inch.
- The drawings suggest this routing, but definitive information was not available. This routing gives a higher propensity for flashing than other routing consistent with the drawings, and is thus conservative for the present purpose.
n nin l
UUG
R. Tedesco Frc-this result, very little (less than 1 ft. ) boil-of# of the uater in tne reference line aoulc eccur cue to system cecressurizaa on ro:.
2155 to 1400 psig.
3.2 Effects of Decassino in the Reference Lee In assessing the possible effects of degassing, it was determined that 0.9E by volume of gas would excape out of tne reference leg.
In the a s s "~ f " % in ec"41 * -
calcul2 tier- "vd cge-r 22P m ic with dissolved nyorogen in tne water in :ne reference ieg at sn=
start of the event.
The depressurization calculated was 2250 to 1500 psig, at 100 F.
This small liberation of gas would not cause any significant displacement of water in :ne reference leg at tne pressure decay rate exnibited during the TMI-2 transient.
Babcock and Wilcox Company has indicated
- that hydrogen in the primary coolant is present in quantities that produce hydrogen partial pressures of less than 40 psig, which is far less than that assumed in the above calculations.
While there are other gases present in the primary coolant (e.g., nitrogen, krypton, xenon and oxygen), their solubility limits are lower than that of hydrogen and therefore have not been calculated separately.
3.3 Venturi Effects The venturi effect on the accuracy of the level measure. ment has also been determined to be insignificant. The basis was a calculation assuming a steam / water mixture at a velocity of 100 ft/sec flowing past a 1/2" pipe secured at one end.
It was determined that 0.6" of water could be displaced at these conditions.. With the reference leg used, the calculated displacement is insignificant.
4.
BABC0CK AND WILCOX TESTS Discussions with Babcock and Wilcox* have indicated that tests were conducted in about 1975 on reference leg performance during postulated depressurization.
- Telepnone communications between NRC Staff and J. Taylor, B&W Licensing, April 10, 1979.
n/D h) bU
?
277:q R. Tedesco
-S-The written results have been mislaid, but recollections are that significant loss of liquid did not occur.
4 ORIGINAL SIG;1ED BY RODNEY M. SATTEnc1Et0 R. M. Satterfield Chief, Instrumentation and Control Systems Branch, DSS cc:
H. Denton R. Mattson S. Hanauer V. Stello G. Lainas D. Tondi E. Wenzinger P. Shemanski A. Szukiewicz C. F. Miller J. Watt T. Novak R. Tedesco S, Istael Z. Rosztoczy R. Audette P. Norian DISTRIBUTION:
v a r, n,... m n i:E NRR READING FILE PS READING FILE ICSB READING FILE DOCKET FILE (TMI) n rc n
'j
'u, V a
RAudette[h PS:I A '
d RSXABX RSXAB RjPSYB RS:AB C
ASzukitwicz:pob CFMiller RAMMMIIM EWenzinger 4/// /79 4/[/79 XW#X@23 4/ /79 4/// /79 f
PS:ICSB[ld AD:PS o Pic= >
RMSatte Wield SHHanauer
- " " " " * *
- 4lljl73 si l73 DATE F NRC FORM 318 (9-76) NRCM 0240 W u, s. novs =N u ANT PMf MT1hG OFFIC E: 1970 = S2M14
ENCLbSURE1 Calculation of Reference Leg Temperature Profile
- PPC0 $$ h.
.J T'~#~*
"A 4-
.TP=t.P/S~g
/d o D.P.= tsig
/
/$c~ = Vf[ft.stS)
C. to 5$t_',
A~ : 59 C.1is)} (f/S$] = o. t3G 4 C:. o.sai c o a 3.c e w z
AA = 77(f.1!.Q3)
- Ip.31 G, ' 56.o gd1,t'
~bs.r :
2 7 reft. //'or/ff l
/= s.27
~[
W H / B /zi~
- hn : 9.27 8'Tif/6,7C * 'f f
I kh + k~ Ac~
~7~;, - T
=
A;f.0~ih~
z e
NrkM
/ 2(i tilno'// h, p p 7 2 % TW/A. *f
~
I-ci. ? S,6 f.
p.27(3.&22w;f'~,,bV71
,,3 9, gyq/g, sp L Acu.
o.2547
~
G
- b. ? '? [4 M - /r')
o M/f.:. ? ?(y,72 ?ls24/)
b /l4 e
=, lo 99 71Y//h 'E
. J l 7/[74 -N :
/0 f f 7'; 6 K) 2
~S : cts Y = /M 217/(6 S*b - X) =._ ir 9 7 tro t T; - pro IN. d - 92 7/E =, o 99 f y Co +- 7))
/ t d. 4 - J '/. 7 =. 3 y 2 X ggo y')J.f^
- I ~ ISI ? :- 9' 7 3,y
},,g_ 5
. J Y>
89 0;i
I he fY hs b, /2 [s* h!vif, i.s 9y s
- o. r
.3 kg ke.v fSTf.I.SZZ d): M)) 9 6 -s i'
L
- r. f
.iyty lAh:.??{3D-/"fo15).27(3.9'6).0s5
^
w o.057
, i 3 9y (</7it 'iT).:. o n f'f* - /**)
t d. ov t
. i 39y% =. o 2 95(a9) +. 095N i
. it e1 % = l &. o 7 - t. o t I
$ c _b i O J Y3 'f
~
g I
. /4 (9 ffq.
ld0 Y) o.L? = 0. 240 k IR - O. y 44 4
'Tp[694)*-6v4c)'] : o. s s 3c C. 2. c c </ n,- # [
de, e T4 G4 64)'
- o. i yo & P : i.ity wi' Aew :
st.667f s s.zz f '
A4 ' P'l h)fA
.=
his A=T z-J a [2 44V* o,MT =, o s a s,
~
d def., 0.3 9
- 9 n i / 1
. e-o o 4 6 I f
.o st 9 7 ~
~
L
. ? ?(:ro -mhs.zzf* ~=.2][3.ic [zQ b As t
=.it99
,o3343(345-l.)c. t ?'79.343 + % - i r,
?
- 11. I 33
, o snesTo =.o 4 stf(\\ q 3 +T;)
.12 b? t Ti = > I.12 3 - 13. u 2 L U sy,& \\
/d
- - 2. 3 W "h l l
,tplDk Ta S C) 0..,l
~M.
l P
VENT N0ZZLE REllEF YALVE N0ZZLE (TYP OF 3) hl 4.75" WIN SPRAY LINE N0ZZLE gg 3"h 1" ElfE STRAY N0ZZLE e-L E V E SENSING N0ZZLE E'EY f '/
l
'M.- 9 b y-ELCV. 34%.6 y-D p -
r I
i
' l/1 Y vq g N';.
, g,,p(,',y g ( ;,,p,g f
I
\\3 STEAM SPACE 7
4 t
A IIII
/
- 6.138" MIN
~ 84" 1 0 -+-
N tp
\\
/
l NJ
//
N,0 R L' A L HATER LEVEL N/
~ S *-
it
.c
/
G INSUL ATioN
/
s VESSEL SUPPORTS
/
l d
1,
/
\\
/
/
f
/
s q
y THERMONELL
/
S
\\
_.9
,HE A TE R BUNOLE
//
v2
\\
}J R0!ATED giggg
\\,.
. ET _
f E
(%
,/
9 R
sii <//-
//</s, i /,,
N
&>/
LEVEL h
SENSING N0ZZLE 4
s
/ / / N,,.
sis n t.
C,
///p w'r w,
/
ELEV. '3 )(,
\\
f 1R
/
ELEV.
310 NA
~~
SC TX SURGE llNE N0ZZLE ELEV.
g, 200' n1 D'"
C O N rE @,A Tiv E Rout \\ N G 07 as a PRE 55URIZER T11REE MILE LSLAND NUCLEAR STATION UNIT 2
.