ML19208C048
| ML19208C048 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 07/17/1979 |
| From: | Ball S OAK RIDGE NATIONAL LABORATORY |
| To: | |
| Shared Package | |
| ML19208C045 | List: |
| References | |
| NUDOCS 7909240455 | |
| Download: ML19208C048 (7) | |
Text
..
?
t r
k h
Three Nile Island - 2 Technical Support Boiling Detection t
t S. J. Ball I.
Problem:
Perform noise measurements on the core exit thermocouples to determine:
(a) if boiling is occurring in the core; (b) if there were substantial chan3c.s in the noise signals since the initial measurements were made Jf (Apr. 6-7, during forced circulation mode), and (c) if other useful I
information could be obtained from the seasurements.
_ Description of Measurements:
L.
r 1.
During the period April 27-30, noise measurements were made on each r
of the 51 core outlet thermocouples (C/A grounded junction) using Princeton Applied Research (PAR) amplifiers (Gain '30,000), and an HP S420A nois e
analy er.
The sign.
' were analyzed two-at-a-time and recorded simultaneously on a Brush recorder.
Power spectral densities (PSD), cross PSD, and coherence were stored on tape for later plotting and analysis.
i Integrated j
power from 0.0122 to 1.53 H:
f was recorded for each PSD signal.
j 2.
Except for one case (see item 3 below), the signals appeared to be I
uncorrelated.
The amplitudes of the fluctuations ranged from <0.1*F peak-to-peak (p-p) for most cases, to ~1*F p-p (5 cases), to ~6*F p p (1 case).
The signals tere not all stationary, as occasionally on repeat analyse s
the magnitude and character of some PSD's would be quite differ ent.
3.
In one case, two ther ::ouple signals analyzed'(9H and 12F 9
, see core map,, Figs.1-2) had oscillatory (rather than random noise) characteristi cs, a 'SO sec period, were strongly correlated (coherence ~0.9 at 0 02 Hz)
P00R OHK1 K90924 0ff6 m6 zu
2 I
and were clearly related to similar oscillations in the loop A and B
[
p r n.s.u..
p pressure signals recorded elsewhere.
These,, oscillations werc ~2-3 psi p-p, t
s', n < L 1.:.
(
lightly damped, of a slightly different frequency, fe e'"pe*'<", and r
s appeared to be provol.ed by the actions of the primary water makeup system.
Brush recorder traces of 911 (the hottest of all the signals, ~315*F)
Ij showed what appeared to be a canec11ation effect during periods when
(
the two " opposing" loop pressure signals had a certain phase relation-r i
t ship.
(Th.= phas c
- t.. nct readily uutainable sin,. the - s ynels w e recorded on 3 separate recorders.)
i s
4 k
Tests were made of the effects of various thermocouple Icakage resistance paths on their readings. Abrupt " shifts" in several thermocouple l
E readings had been cbserved, so the tests were made to see if the behavior could be attributed to intermittent leakage.
A thermoccuple that normally read 180*F read ~40*F with the low side shorted directly to s
l l
ground, and ~140*F with the high side shorted to ground.
Shorting the I
high side through a 10 C resistor reduced the reading ~4*F, and through a 2 KG resistor ~16*F.
Observations and Tentative Conclusions Related to the ORNL Measureme 1.
The thermocouple noise measurements did not suggest, nor did they 3
l rule out, the possibility that boiling is taking place somewhere in the core.
In many cases, thermocouples with the larger PSD's had relatively i
low temperature readings. The magnitudes of the noise signals were surprisingly low, especially considering the large thermal gradients present in the upper plenum.
Saturation temperature for the operating pressure during our tests (~900 psi) is ~530*F.
It would be possible for steam bubbles to recondense before reaching the thermocouples, and for boiling to occur in non.onitored areas.
Probably the best way to infer Di M E 1006 247
~..
3
?t boiling would be to closely monitor a group of the most suspect (noisies t) c
[l thermocouples, then lower the system pressure and watch for significant k
increases in the noise power.
7 k
2.
The case of the oscillatory and correlated thermocouple signals (9H and 12F) cannot be reasonably e.splained by a pressure-affecting-temperature boiling argument, since:
(a) the start of the thermocouple oscillations 1
coincided eith the start of the pressure oscillations; any changes in si;.nl due to a w 211-n a.::.:r ed. ' ciling region '.. auld prv: ably have taken e
o sev'r:1 minutes to beccme established; (b) the signals were quite " clean" r
and uncharacteristic of a boiling noise signal; and (c) the 180*
phase relation betwcon the two si;;nals would not be readily explainabic by boiling. One theory 1hich can account for this behavior is as follows:
The upper plenum durin;; the natural convection mode is relatively i
stagnant and has numerous " plumes" of widely-varying temperatures emanating from the core regions.
A hydraulically-induced pressure transient sets up wave-like oscillations in the two loops, causing back-and-forth sloshing in the upper plenum, and thus lateral " waving" of the plumes in the vicinity of the 9H and 12F probes. This could account for the "immediate" i
response to the pressure change, the cleanliness of the signals and the 180*F phase.
This model also requires more than one free surface in the system in order to set up the wave motion, so it infers that there may I
i be noncondensible gas pockets in the reactor and/or steam generator upper plenums.
Followup measurements and studies based on this model may also be useful in deriving more detailed information about the condition of the system.
For example, the periods of oscillation of the two loops might be related to the size and location of voids.
If so, one may infer 200R_0R M
-(
I' 4
t E'
changes in void si:e due to a system pressure reduction.
g F.
In any case, 2
h it would be useful to record the two pressure signals and thermocoupl e
(
9H noise on the same instrument.
y.
4; 3.
The " waving plume" model may also be used to rationalize the behavior t
b of several thermocouple signals in the three hours folle. ing the transition i
to the natural circulation mode.
The t xtbook behavior of core outlet k
temperature would be an abrupt rise (corresponding to t?* reduction i t
(
n flow) followed by a slow decline as the full natural circulation flow is established (Fig. 3, ilS).
E Ilowever, several signals insicad showed t
abrupt drops following the pump trip (f;,. 4, H5).
Considering the plume modal and noting that the upper plcra:= i'on distribution patterns would a
be quite different for single-pump forced circulation ar.d natural
?
circulation, it appears likely that the thermocouples su:h as H5 are simply monitoring a different combinati:r of plumes.
4.
The abrupt shifts observed in scr.e thermocouple readings (notably 3F) can more readily be attributed to a costulated shifting of the l eakage resistance of a damaged probe than to pcstulated scenes in which hot debris settles temporarily near the sensor.
F I
{
Work Requested By:
t NRC-TMI (Stello, Ackerman).
I.
Results Reported To:
NRC-TMI, ad hoc Industry Advisory Group (IAG), General Public Utiliti es, and Babcock and Wilcox, April 30, 1979.
Worked Performed By:
S. J. Ball, R. C. Kryter, W. H. Sides, Jr., and G. L. Zigler (SAI)
P00RORGLL
rl E
C 7
g 9
Ic I.
i 8
I I
I g
i
+
gg r
n l
p.
-p t2 f
g I
13 g-189 211 2
8-I4 i
f I
l
['
I C
c 226 193 18 1 (5
=,
I l
D 216 177 176 D
I b.
c-188 231 219 C:
l172 184 202 171 179 142 t
l
! g 232 253 228 236 198 234 h 1F.
205 295 315 164 i.
l t
K.
192 216 181 y
1.
l203 190 172 201 l
I' i
l q
210 254 206 202 190
/fg l
,M 218 198 176 f((
'~ ~
1 I
p 189 180 10.7 175 pg l
t E
p 177 pt (3
I k
197 160 l[( i I2.
4 I
8 I
i 1
i ll b
b i
I'7 fig. 1.
ThfI-2 Cor:4 E::it Thernaccuple Ste: 6y-stcte Rer.dir3r (*r), 4/28/79, 2400 hrs. ::2tural Circulcticn Cooling i
~ 4
T.. '
[
C C
7 8
9 Ic i
i 3
I i
g gg hl l
A-
-A '
I (3
g 1
8- [
i S-10.8
.05
'2.
g 3
i I
I f{.
l C
.06
.05
.08
.44 C
g i
l I
p
.13
.32 3.0 D
f l'
E C
Errat.:
.16
.05
.15 J _..
.24/
l 50 p
F-61
.04
.04 2.8 1
i.
k t
.07
.05
.05
.13
.23 578 E
l 1
l 0 ". l.14/
l g
l f r-
.2/i
.07 13
.06 i.
l.
l N_
I J
K
.06 Ho test hannel
.06
.14 IT i
L
.05
.25
.16
.32 I
. _.I 'l M
.26
.06
.06
.04 10
'O i
f
.07
.49
. 4 '4
'l l
If c
2 i
f
.33
.19
.12
.17 I
i 1
IT l
2 P
.46
.63/
g(
-s 5
l 13 3
i g
3,4
,o3 g
4 I
I (2.
S' i
i I
Il G
7 9
T 10 Fig. 2.
TMI-Z Core Eri;
- Ler:.:::cn; ~ c-1'en Fquare Foise (mV ).
Bandwidti 2
0.0122-1.53 H:.
7.c a ' n : n, -
4 /.'0 / >':
1630 hrs. to 4/29/79, 2230 :
s.
Nature. C h..al aci n C:aling.
OR]HER1 1006 251
.i.
5 h
l k;
Estiraated Cost:
(
20 Mandays plus transportation; $9,600.
t
- t 1i L
6 I
t I
b i
(
t i
I l'
I f
l
\\
l 1006 252