ML19329A505
| ML19329A505 | |
| Person / Time | |
|---|---|
| Site: | Oconee  |
| Issue date: | 01/27/1978 |
| From: | Mackay S Office of Nuclear Reactor Regulation |
| To: | Buckley B Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8001060022 | |
| Download: ML19329A505 (9) | |
Text
.
7 1978
&a HOTE TO:
B. C. Buckley, Section Leader, Plant Systems Branch, D0R FROM:
S. D. Macray, Plant Systems Branch, DDR
SUBJECT:
OCONEE POWER OSCILLATI,0HS During a visit to the Oconee power station on December 27, 1977, additional infomation was obtained regarding the nature and cause of reactor power oscillations on Unit No. 3 and the action taken by the licensee to reduce the magnitude of these oscillations.
c The licensee had observed reactor power oscillations in Units 1, 2 i
t
~
and 3 at. power invels between 50% and 75% of full power that had a frequency of 0.25 Hertz and a peak-to-peak amplitude of approximately 1.5% of full power.
These oscillations were considered normal.
Housver, following the first refueling outage of Unit No. 3 in the Fall of 1976, these 0.25 Hertz oscillations gradually increased in amplitude from 1.5% in November 1976 to 7% in June 1977 (see Reference 1).
i, In an effort to determine the cause of the oscillations and to eliminate them, the licensee gathered data on all parameters that were oscillating and ehecked all settings in theidntegrated control system.
Parameters that oscillate at 0.25 Hertz include:
i reactor power (out-of-core detectors), reactor coolant inlet temperature, reactor coolant outlet temperature, steam generator pressures, steam generator water levels, feedwater flows, feedwater pump pressures and reheater drain level. The only abnormality noted in checking the process system parameters was that steam generator A provided 18'F less super-heat than stea geparator B and the least amount of superheat in any once-through tem generator.
It has been pointed out that this may be related to be fact that steam generator A has an open lane three tubefwide rather than one tube wide.
In assessing the effects of the J
integrated control system on the oscillations, the power from the steam i
I generators was unbalanced by demanding the reactor coolant returning j
from steam generator B to be 2*F hotter than the reactor coolant i
returning from steam generator A.
This unbalance stopped the oscillations but the settings were restored for normal opeaation.
3 Similarly, when only three of the four reactor coolant pumps are in operation, the oscillations are not present. This and other information l
together with previous knowledge that the steam generator wal.er level had a natural frequency of 0.25 Hertz led to the conclusion that the cscillations might be reduced by increasing the hydraulic resistance of the downconer region of the steam generator.
Thft is an adjustable i
i orifice in the steam generator for this purpose.
drq OFFICE F
._. g F
oats >.
NRC FORM 318 (9 76) NRCM 0240 W u; s. novsawuawr enmrme orreces e,7s -eauad DU 8001060
'I JAN 2 71978 i- '
It was planned that during the October 1977 shutdown, the orifice plate t
position would be checked and if it were 21/2 inches open, the same as Units 1 and 2, it would be completely closed. During the shutdown I
it was found that the orifice plate was in the expected position, l
2 1/2 inches open. However, attempts to close the orifice failed as I
the plate could be moved only 1/4 to 1/2 inches toward the closed j
position. -It was secured in that position, between 2 and 2 1/2 inches l
open..
After refueling, the plant was started up on December 3rd and the amplitude of the power oscillations was less than one percent. However, after operation for three days at 75% power it was apparent that the osc141ations were increasing in amplitude. Figure 1 shows that on December 18 while operating at approximately 75% power, the peak-to-peak amplitude of tie power oscillations was 2.2 percent of full power.
4 My current knowledge of the phenomenon leads to the following ciualitative understanding. The oscillatory tendency originates in y
the steam generator at the natural frequency of oscillation 6f the water levels in the downcomer and tube bundle regions. This effect 4
may or may not be enhanced by the process system or the control system.
In either case, a change in the water level in the tube bundle will cause a change in the effective heat transfer area coupling the i
reactor coolant system to the steam system resulting in a change in the rate of heat removal from the reactor coolant and thereby causing a change in the temperature of the coolant returning to the reactor I
core. The oscillating water level therefore causes the reactor water temperature to oscillate and results in reactivity changes that cause
' the reactor power level to oscillate. These temperature and power level changes also result in the oscillation of the reactor outlet temperature but it is not clear whether this temperature oscillation in turn significantly influences the water level in the tube bundle region.
The power level oscillations were minimal when the reactor was first j.
placed into operation after refueling because at that point in core life the absolute magnitude of the temperature coefficient of reactivity was relatively,small, i.e., about 1/5 of that at the end of the l
! previous cycle. However, as fbe fuel burned up and boron was removed
/,from the moderator, the temperatuee coefficient of reactivity increased and thereby caused the reactor power oscillations to increase in i
- amplitude. Thus, if no corrective action is taken, the amplitude y
!would be expected to increase throughout thee life and become i
approximately the same as that at the end of the previous cycle.
/ It was also noted that the power icvel oscillations are measured on
~
the~out-of-core neutron detectors whereas the in-coes detectors
, indicate no oscillation. There is no in-core 4sdication of power
/
oscillation because the in-core power level signals are generated OFFICs h
~. eunnamah obs>
l NRC PORM 313 (p.76) NRCM 0240 W us e. eovanmusst riiluvine orricas e,7e-eae4a3
^
JAN 2 71978 y
- by the activation and decay of rhodium powered detectors that have an effective response time of several minutes and therefore do not i
respond to a 0.25 liertz oscillation.. The ott-of-core detectors L
respond quickly to neutron flux level changes and therefore follow
[
the 0.25 Hertz oscillations that could arise from core power level
'3 changes or changes in the density of the water in the reactor vessel downcomer between the core and the detectors.
A calculation of the changes in neotron attenuation due to the oscillating temperature of the incoming water compared to an estimate of the reactivity j
effects of that temperature oscillation shows that the reactiv'.ty 1
effects are predominant and therefore the out-of-core detectors indicate an actual power level oscillation.
A good deal of additional data and some testing are necessary to confirm our present understanding of the phenomenon and allow us to quantify '
h some of the effects that have been observed.
10M The licensee is continuing its evaluation of this matter, devising 4
special tools to permit movement of the orifice plates in the steam d
generators and developing further tests of the instrumentation and i
E controls. Additional data will be obtained within the first 28 effective full power days when the reactor is at 75% power for the
[L correlation test of the in-core and out-of-core deutron detectors.
It is expected that the amplitite of the oscillations, will not exceed j
the value previously experienced, and there is apparently no significant safety consideration at that value. However, it is recommended that t
the licensee be encouraged to obtain additional data and a better quantitative evaluation of this phenomenon so that the oscillations j
'may be eliminates or reduced to an insignificant level.
l S. D. MacKay, Plant Systems Branch Division of Operating Reactors References 1.
" Trip' Report - Oconee Power Oscillations" 10/31/77, R. Woodruff, IE.
- 2.. BAW-10002, "Once-Through Steam Generator Research and Development Report," August 1969.
3.
BAW-10002 Supplement 1 June 1970.
Enclosure l
' figure 1 -'Oconee Unit 3 Power I.evel Oscillations opric a p-l summaus h nav s p-NRC 70RM $1s (9 76) NRCM 0240
- v. s. oovanmuswr painvues orrics sere-one.ead
r e.
JAN 2 71978 t
cc w/ enclosure:
'i V. Stello D. Eisenhut-W. Butler F. Jape IE i '
F. Long IE R. Woodruff, IE J. Sniezek, IE i
K. Seifert, IE 1
N. Moseley, IE R. Baer M. Bendonca J. Buzy, OLB P. Collins, OBL A. Schwencer D. Neighbors S. MacKay 9
i i
s 5
i k
i t
f 1
t
~
uA j
DMb orrscz e ou,,,,j g, SMacKay;pab 01/2ld78 l
z D.TE W W ve s. novannu adT 'FiiiETins oFFICEd 1975== 034.e33 f-NRC PORM 318 '(9 76) NRCM 0240
JAn 2 7197a i
6coxes Umr 3 fo W6g lCVEL.
D S C / L L A '7~/ o N S em o
q'
[ e e j\\\\ o Ju. S.
-3 SYSTEM AVAILABLE 7.589E 01 +7[ 87 f FU44 ToWSR -
='
[k11T #03 12-18-77
\\l b
O'
~
m M
1 k
L-
+,Me{3 (8
7.37K 01 10 0 TIME 19 9 f SfC OND S-->
- 4-NO TE
- THis IS NoT AN ANALO4 REcoRb.
DATA JS PL OTYfD AT eME SfCo@ /NTCR&9tS.
4
\\
f/6)UR E 1 u--.
--m w
'=-w9--T-*v-*W9**-'"yM-*
- ~""--P'"'"""*FT"'**
(.
JrU ? 71973 ADD 6NDL/M.- 7D A61*C pgy.79 7~ 2. Rucktfy s p-e Oconee Oscillations 15CZ)'8*t -
1.
Response of Resistance lemperature Detector (RTD) to 0.25 Hertz temperature oscillations.
On January 18, 1977, I learned from a B&W employee that the response time of the RTDs was three to four seconds but might be as high as
~five seconds. Using four seconds as the time constant, we obtain aT=2 m (0.25) 4
= 6.28 radians A calculation of the response of this simple lag system shows an attenuation of 16 dB and a phase lag of 81.
'InisshowsthattheRTDelementwill experien'ce a temperature oscillation that is 16% as great as the oscillation in reactor coolant temperature at the 0.25 Hertz frequency.
Thus the indicated temperature oscillation of 0.18'F (peak to peak) observed on December 18,1977 at 10 am.
reflected an actual temperature oscillation of 0.18/0.16 or 1.13 F.
2.
Reactor Response to Oscillation of Inlet Water Temperature The transfer function of a PWR is fairly flat in the region or 0.25 Hertz and the reactor should respond with nearly full amplitude to reactivity oscillations at this frequency.
Since the fuel temperature oscillations will be very small, reactor power will respond to T inlet directly.
Thus the power level response is estimated to be approximately dn aK dn 4
UK 10 x
=
naK aT.
nT AT percent power per F.
j t
\\
,. ~ -
JAN 2 71373 Thus.a 1*F rise in downcomer water temperature will cause the neutron detectors to indicate a 0.3% power level increase.
This accounts for part of the amplitude of the oscillations.
4.
Phase Relationship Between Power Level and Downcomer Density
?
The effective transport time from the downcomer to the core is approximately one and a half seconds.
The power level will lag the reactivity change slightly and the total lag time between a 'downcomer density increase and a power level rise will approach two seconds.
With a two-second lag, the power level will be 180 out of phase with the downcomer density, i.e., when the power level is maximum, the density will be minimum and there-fore the two effects will be additive to produce the maximum neutron level reading.
~
1 1
1
- w
s
'j (-
Jfd; 2 ?E3
'The moderator temperature coefficient AK for cycle 3 has aT been reported to be -0.53 x 10-4 Beginning-of cycle an'd
-2.55 ' x 10-4 at end of. cycle.
Thus if the temperatere coefficient were-l.2x 10-4.on December 18 when the temperature oscillation was l.13*F, the resultan't power. level oscillation should have been
'about'l.4~ percent power.
The data shows however, that the power level oscillations were approximately 2.2% power. A temperature power transfer coefficient of 1.8% power per *F would be needed to explain this behavior.
Although this calculation is neither sophisticated nor very c
accurate, it does indicate that the power level oscillations observed could result directly from the inlet temperature
}
oscillations.
3.
Neutron Detector Response to Downcomer Water Temperature Oscillations The downcomer annulus contains a layer of water 81/4 inches (21 cm) thick between the reacter core and the neutron detector. For this thickness of water, the effective attenuation length will be approximately 10 cm (3 meV neutrons) and the change in 1
attenuation per 1 F will be f=-fdx= -j h
- where b = - 0.0014.per F at 550 F at 2000 psi o
f.=hx'0.0014=0.0030per F
= 0.3 percent per *F i
~
JAll 3 7 1973 ocoxcc Omr 3 CO L1>
LEq 7%~/yPERA rt/K. C
~
.s
- d..&. a... -
=
5.627E 01,
UNIT se3 12-te-77 t
s E TC B1 A
A i
~
0./8'F
)
~
\\
I-V
)
s.s21s q I to-e- e rtre te-t- e S Sfc. ->
+
om m
3-J e w JB e Ju 1 \\ L i
~
-