ML20031B635
| ML20031B635 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 09/29/1981 |
| From: | Hukill H METROPOLITAN EDISON CO. |
| To: | Stolz J Office of Nuclear Reactor Regulation |
| References | |
| L1L-273, NUDOCS 8110050279 | |
| Download: ML20031B635 (19) | |
Text
. _ _ _ _ _ _ _ _ _ _
Metropolitan Edison Company Post Of fice Box 480 II Middletown, Pennsylvania 17057 Writer's Direct Dial Number September 29, 1981 alt 273 h
Of fice of Nuclear Reactor Regulations
\\
Attn: John F. Stolz, Chief p' OCT g
Operating Reactors Branch No. 4 i-IS8IA U. S. Nuclear Regulatory Commission 1 p y'
""oJO,gron Washington, D.C.
20555 c
Dear Sir:
Three Mile Island Nuclear Station, Unit 1 (TMI-1)
'O Operating License No. DPR-50 Docket No. 50-289 NI Errors Babcock & Wilcox's letter of October 29, 1980 to the Commission (Ref. 1) concluded that for certain events, the transient-induced neutron flux errors can be larger than those assumed in the FSAR analyses.
For 177 Fuel Assembly plants, the transients of concern are small overcooling events including small steam line break, and the ejections of a low-worth rod.
Your letter of January 14,1981 (Ref. 2) requested information on this potential problem regarding the operation of Three Mile Island Unit 1.
Our letters to you of April 23 and June 15, 1981 (L1L 098 and Llt 188) stated that an investigation was underway and that results would be available by August 11, 1981. Due to unforeseen delays this submittal was further deferred by a phone call to you until September 30, 1981.
Since TMI-1 was in a shutdown condition during the entire interim, deferral of these results did not constitute a safety issue. Responses to your questions of January 14th are attached.
Sincerely,
\\
l.D.
lu ill Director, TMI-l llDil:vj f Attachments
[oo/
.5
/
8110050279 8109297 DR ADOCK 05000289 PDR n Edison Company is a Membei ;' the General PutAc Utihties System
ATTACHMENT ITEM 1 Determine if the high flux trip setpoint for your plant is affected by the accident-induced neutron flux errors discussed above.
Provide us with information establishing that your present accident and transient analyses are valid and that the present Technicc1 Specification limits provide as a minimum the original protective margin derived f rom the safety analyses; if not, provide the following information:
a.
Confirm that only the two non 205FA plant concerns discussed in the B&W letter of October 29 affect your plant; namely, small overcooling events including a small steam line break, and a rod ejection accident.
Response 1.a.
Small overcooling events (including small steam line break), and ejection of a low-worth rod are the only transients that have been identified as being affected by the high flux error.
b.
Provide the effects of the error on your plant, supported by appropriate analyses.
Response 1.b.
The results of our investigation, based on evaluations perf ormed by B&W, are given in Enclosure 1. provides additional material on the rod ejection transient.
For overcooling events the concern was that the induced neutron flux measurement error might increase as reactor coolant temperature decreases.
I Thus, the actual core power level could exceed the assumed 112% overpower l
condition before a reactor trip occurs. However, the evaluation has shown that the increase in core power is offset by the beneficial effect of the reduced coolant temperature on core thermal margins. Departure-from-Nucleate Boiling Ratio (DNBR) analyses performed for the most limiting condition (indicated power at high flux trip limit of 105.5%,
FC pressure at low pressure trip limit of 1800 psig) demonstrate that the minimum DNBR will be greater than 1.43 for conditions under which a reactor trip would be initiated at core power levels up to 130% of rated power.
In addition, TMI-l Cycle 5 calculated core power distri-butions at all allowable rod index and Axial Power Shaping Rod (AS?R) positions for normal full power operation were examined and margin existed for both DNB and Center-line Fuel Melt (CFM) assuming an actual core power of 125%.
Therefore, the induced flux measuremer.1 error does not compromise the safe operation of TMI-l during overcooling events initiated from anywhere within the allowable operating range.
The concern on the rod ejection transient was that the high flux trip may not be activated for an ejected rod of small worth (less tnan 0.2%
l A k/k).
As shown in Enclosure 2, an engineering evaluation of the l
l L
Page 2 ITEM 1 (continued)
Response 1.b. (continued) conservatisms in the original analysis has led to the conclusion that a reanalysis using realistic assumptions, will show that the peak fuel enthalpy limits are not exceeded. Therefore, this concern is not considered to compromise the safe operation of TMI-1.
For TMI-1, the maximum steam dump capacity for six fully open turbine bypass valves is 22.5% of steam flow. This flow is lower than that assumed in the B&W evaluation and would cause a smaller reduction in downcomer temperature and, therefore, a smaller transient-induced flux The assumptions of this evaluation are therefore conservative error.
for TMI-1.
Further, a low pressure trip of 1800 psig was assumed in the evaluation.
TMI-l will have a LP trip of 1900 psig following NRC staff approval of the pending Technical Specification change. Therefore, even larger margins will exist at TMI-l than those calculated.
Provide your program and schedule for mitigating the effects of the error.
c.
Response 1.c.
The investigation has shown that even with transient flux errors to 137.,
TMI-l will maintain significant margins to DNB and CFM limits. Therefore, no additional actions are deemed necessary at this time.
ITEM 2 Provide justification for continued full power operation of your plant until your program to mitigate effects of the error is completed.
Response 2.
It is believed the results of the investigation justify full y'ver operation of TMI-l Cycle 5 since it has been shown that no safety limits will be violated due to transient flux errors during small evercooling and low-worth rod ejection events.
References:
1.
Letter, J.H. Taylor (B&W) to V. Stello (NRC), October 29, 1980.
2.
Letter, R.W. Reid (NRC) to All B&W Licensees, January 14, 1981.
3.
Letter, H.D. Hukill (Meted) to J.F. Stolz (NRC), April 23, 1981.
4.
Letter, H.D. Hukill (Meted) to J.F. Stolz (NRC), June 15, 1981.
L ENCLOSURE 1 W
i c
E c
f m
General Public Utilities Service Corporation c
I Task 78 - N. I. Calibration Error Final Report I
i August 1981 Document Number 12-1127447-00 i
s 0
f t
v
J 9
INTRODUCTION
~
A concern raised durig ce* tain overcooling transients and rod ejection transients is that a transient-induced neutron flux error greater thaq that assumed in the FSAR could exist.
This concern was presented to the Owners in a group meeting in Lynchburc en October 23 and 24. 1980.
The pr: css of Master Services Ta,k 78 - NI Calibration Error is to identify plant safety margins available to offset the additional NI errors, thercoy providing
,iustification for full power operation of TMI-1 Cycle 5.
SUMMARY
AND CONCLUSIONS A concern was raised that overcooling events could result in an induced neutron flux measurement error which might increase as reactor coolant temperature decreases.
l This error could result in the actual core power level exceeding the assumed 112% overpower condition before a reactor trip However, an evaluation has shown nat the increase in core power occurs.
is offset by the beneficial effect of the temperatura decrease on core thermal margins.
DNBR analyses perfonned for the most limiting condition (indicated power at high flux trip limit of 105.5%. RC pressure at low pressure trip limit of 1800 psig) demonstrate that the minimum DNBR will be greater than 1.43 for conditions under which a reactor trip would be j
initiated at core power levels up to 130% of rated power.
In addition, cycle 5 calculated core power distributions at all allowable rod index and Axial Power Shaping Rod (ASPR) positions for normal full power operation were examined and margin existed for both DNB and CFM asstsning an actual core power of 125%.
It is therefore concluded that the induced flux measurement error does not compromise the safe operation of TMI-I for Cycle 5 during overcooling events initiated from anywhere within the allowable operating range.
1 W
L 1
The concern on the rod ejection transient is that the high flux trip may not be activated for an ejected rod of small worth, less than 0.2% A k/k.
Under these conditions, current models m uld show unacceptable results (peak fuel enthalpy > 280 cal / gram).
Althouah no reanalysis has been p
performed, an engin ering evaluation of the conservatisms in the original analysis (such as adiabatic heatup) has led to the conclusion that a reanalysis using realistic assumptions, will show that the peak fuel i
l' enthalpy will not exceed 280 cal / gram. Therefore, this concern is not considered to compromise the safe operation of TMI-1 Cycle 5.
OVERC00 LING ACCIDENT ANALYSIS b
{
The current flux error assumptions in the FSAR are:
2.0% heat Balance 2.0% Steady State Neutron Measurement
- 2. M Transient-Induced Neutron Measurement 0.5% Instrumentation
~
6.5% Total L
j The 2.0% transient induced neutron measurement error is sufficient to f
accomodate all transient induced errors excluding those which are the topic of this report.
The additional transient induced error was discovered
{
during a study for the WPPSS - WP 1/4 FSAR, a 205 plant.
provides an evalJation of the induced NI error during certain overcooling l
events.
m Although no transient analysis has been performed specifically for TMI-1 Cycle 5, an engineering evaluation based on the WPPSS analysis has shown that the maximum transient induced error for moderate frecuency overcooling
(
transients will be approximately 13%. Therefore, for moderate frequency overcooling transients only, the instrumentation error that should be
(
considered is: 5s
2.0%
Heat Balance 2.0%
Steady State Neutro!. Measurement 0 to 13% Transient Neutron Measurement Dependent on Coolant Temperatures 0.5%
Instr mentation 17.5%
Total j
To justify full power operation, one must demonstrate that operation up to 123% full power is acceptable during these overcoolino transients.
This power level is based on a high flux trip setpoint of 105.5% full power plus a total error of 17.5%.
It should be remembered that this power level
}-
could only be reached during certain overcooling transients that provided specific core conditions.
~
The analysis of induced flux errors during overcooling transients led to 9
the quantification of the ratios of indicatec power to actual core power as a
a function of downcomer fluid temperature and core average coolant
]
temperature. The p:imary concern is to determine the conditions that would permit the actual core power to exceed 112% without a reactor trip
}
occurring.
The error calculations were used to determine the maximum actual core power as a function of temperature for the case where the
}
indicated power would be 105.5% which is the high flax trip setpoint j
(Figure 1). A series of heat balance calculations vere then performed,
}
using the minimun licensed RCS flowrate (374, 880 GPM), to determine the corresponding core operating conditions.
When the heat balance is superimposed on the curves of Figure 1, the result is a single line, for dily giVen pressure, which defines the actual core power as a function of coolant inlet and average temperature, consistent with the assumed constant l
indicated power level of 105.5%. This line is shown (dashed) on Figure 1 plotted against downcomer temperature.
It can be seen by examination of l
Figure 1 that core operatio.1 under conditions below or to the left of the a
3
L power vs. temperature line would be less restrictive (lower power and temperature). Also, operation above and to the right of this line would be prevented because the indicated power level would be greater than 105.5%.
thus resulting in a reactor trip.
In order to cuantify core thermal margin for the conditions corresponding to operation at an indicated power level of 105.5%. DNBR calculations were t
performed using the CHATA and TEMP codes.
All 2200 psia RCS pressure points allowed by the RPS and correspondino to operation with indicated I
power equal to 105.5%, lie well above the Tech Spec minimum DNBR of 1.30 for th.
.-2 correlation.
For low pressures corresponding to the RPS low I'
~
pressure setpoint (1800 psig), the variable low pressure trip provides a trip if the RCS outlet temperature exceeds 588.5*F.
This trip function r
I" then provides protaction to a ninimum DNBR of 1.43 for core power levels up to ~ 130% FP (Figure 2).
Any operation at the richt of these limits as I"
plotted is prevented by the Reactor Protection System.
In order to 7
quantify DNBR margin along this line, a parameter study was performed to i'
determine the effect of coolant temperature variation on DNBR at 1800 psig and constant power.
{
The calculations described above are based upon reference design peaking I
conditions (1.714 radial x local peak and a 1.5 core mid-plane axial peak).
{
The applicability of the design peaking to core power shap?.s is ensured by the use of Maximum Allowable Peaking (MAP) curves.
The MAP limits define the maximum peak allowed over a range of core elevation: and uxial peaks.
These maximin allowable peak combinations represent radial and axial
[
peaking for which the calculated MONBR is the same as for the design peaking conditions.
The MAP limits are used in the maneuvering analyses as an acceptance criterion for the development of power imbalance limits (nor.nal operating or RPS limits).
Three-dimensional power distribution calculations were performed to assess the core power distribution perturbation at 125% FP due to an overcooling transient, and to determine the margins to centerline fuel melt (CFM) and departure from nucleate boiling (DNB) limits.
Identical calculations were generated from normal steady-state operation at 100% FP and from ooeration at 125% FP with a 16*F inlet temperature reductior. All power <1istribution calculations were initiated from within or near the normal rod index, APSR and axial imbalance limits of operation, such that the core behavior over
}
the entire allowable operating range was examined.
}
CFM and DNB margins were computed for t"e 125% FP cases to deterrrine if core safety limits would be preserved during an overcooling transient.
}
Since all calculations were performed from near steady-state c.onditions, appropriate peaking factors were included in the 1257. FP margins
}
calculations to account for potential peaking increases due to transient xenon and quadrant tilt. Maximum allowable peaking curves for TMI-1
}
Cycle 5, were used to evaluate the DNb margins.
The applicability of these curves at 125% FP was verified by DNBR analyses performed for the limiting
}
cases.
}
RESULTS
}
The increased power resulting from certain overcooling transists can be accomodated with the preser.t cycle 5 Normal Operating Limits.
l'os itive
]
CFM and DNBR margins are shown in Table 1.
The analysis of the most limiting peaking distribution yielded DNB margins in excess of that required to offset any rod bow DNB penalty for this cycle. The high flux trip provides DNBR protection to the minimum DNBR limit.
Since p0sitive margins are still present, the Normal Operating Limits as re90rted in Section 8 of the Cycle 5 Reload Report (BAW-1509, November 1978) wre not changed. 1
TABLE 1 TMI-1. Cycle 5 Minimum CFM. DNB Peaking Margins (Most Limiting Conditions - 4 EFPD) e CFM DNB Rod Index APSR Index Margin
- Margin *
[
(%WD)
(%WD)
(%)
(%)
]
300 32.0 20.2 7.9 300 25.5 17.3 6.62 271 6.1 7.6 N/A 6
m
- Margin is quoted in terms of relative peaking compared to the
~
k limit value.
L t
n L
T F
~
)
_ L-
f ATTACHMENT 1 L
Since the concern of induced neutron flux error during overcooling events as raised, the importar.t ouestion has been to Quantify the magnitude of the induced error and define the transient (s) that result in such an error.
To this end, a task was undertaken to review current available data, l
determining the bounding moderate frecuency overcooling transient and estimating the resulting induced neutron flux error for that case.
1 The most complete data available to evaluate was from WPSS 205 FSAR
}
analysis. Based on the WDSS analysis, the most severe moderate frequency overcooling transient wnich is terminated by a high flux trip sional is a failure of the turbine bypass system (atmospheric on/off valves). This w
failure consists of a single atmospheric on/off valve on each of the four
)
steam lines opening at power.
The induced " break" results in an increased steam flow of ~ 4.6 x 106 lbs./Hr. The real flux (thermal power)
]
increases to s 122% FP with an induced flux error of ~ 13% FP.
Since the indicated power does not exceed 105.5% FP, the high flux trip will not
)
terminate the transient.
This transient causes a reduction in the cold leg or uowncomer temp. by
- 160F, a reduction in core average temperature by 4
~ 100F and very little change in RCS pressure.
a 1
2 This transient has been analyzed on Power Train IV with moderator coefficients for BOC and EOC covering a range of 0.0 x 10 4 Ak/k/ F to -3.37 x 10-4 A k/k/0F and with varying feedwater temperatures.
]
The change in flux at the out-of-core detector locations due to changes in cold leg temperature were determined using the ANISN transport theory program. The induced neutron flux error was obtained by using the Power Train IV output (real power, Thot, Tave) and the results of the ANISN analysis.
L b
A-1
]
ATTACHMENT 1 Since the Atmospheric Valve failure 's considered to be the most probable ana restrictive overcoolina transient of concern, it is realistic to believe the 13% FP induced neutron flux error is valid for the small overcoolinc, moderate frequency event of interest.
Using the WDPSS Atmospneric Valve f ailure for determinaticn of 177 FA plants induced neutron flux error results, involves the following ass umptions-
- 1) The specific transier.t response for 177 FA plants includina the event timing is similar to WPPSS 205 FA plant.
I
- 2) The failure of all bypass valves on 177 FA plants is similar to the failure of the Atmospheric valves event on WPPSS.
- 3) The magnitude of the induced neutron flux error for 177 FA plants should be similar or conservative because:
e The total bypass flow in 177 FA plants is similar or less than the 205 plant. Also a specific single f ailure can be identified to cause all TBV's to open.
~
e The failure of a single steam generator bypass valve can also occur in c 177 FA plant: however, the resulting steam flow increase will
~
be 50 to 70% less than the WPPSS steam flow increase.
p A-2 N
Figure 1 ACTUAL CORE POWER VERSUS INLET TEMPERATURE RCS PRESSURE = 1800 PSIG 2568 MWin 150 -
\\
\\
ACHIEVABLE CONDITIONS 145 -
\\
BASED ON HEAT BALANCE
\\
- fo e
\\
g 140 -
/,
\\
ACHIEVABLE CONDITIONS
- s
\\
BASED ON ASSUMED TEMP.
g f
\\
COMBINATIONS y
\\
(From Instrumentation
/
\\
Error Analysis)
\\
0 130 -
i
\\
\\
S 2
125 E
E
' +
\\
3 120
- 0
\\
e
\\
\\
115
\\
)
E
\\
110
\\
3 105
\\
\\
]
100 i
500 510 520 530 540 550 560 570
]
Inlet Temperature, F
Figure 2 RPS AND ONBR LIMITS 1800 PSIG
}
3
\\\\
- ]
\\\\\\
ISO -
\\ \\
- ]
\\\\\\
q s\\
J
\\ \\
n0 -
\\x
.)
\\
\\ ',
sign roux 1Rie Il
\\ 3 d\\
(3 5
\\
0 OPERATION FORB100EN 5
\\
BY RPS y\\\\
POWER LIMIT BASED ON
/f
\\\\
2 MAXIMUM PREDICTED ERROR
/
I) g i20 OPERATION
\\\\
~
\\\\
PERMITTED
/
BY RPS
[
3
,5
\\
3
\\ p DN8 = 1.S0
()
ii0
\\
L
~
' DN8 =
\\
kl
- i.c
\\
l 3
\\x VARIABLE
\\
200 f
\\\\
f LOW PRESSURE
/
TRIP AT 1800
\\ \\
f
\\
/
PSIG 90 I
/
I I
]
560 580 588.5 600 620 T
F out -
c i,"
ENCLOSURE 2 l
EVALUATION OF THE SMAtt WORTH EJECTED ROD ACCIDENT Ak Since the ejected-rod accident analyzed in the FSAR for.65%
/k worth rcd is terminated by a high flux trip, concern has been expressed that if no high flux trip occurred during ejection of rods with worths less -
Ak than.2%
/k, that fuel pin failure saight occur.
Current models used far FSAR analysis do not consider energy transfer from the local pin to the wactor coolant, therefore, if used for analysis of the small worth rod ejection accident could show unacceptable results (peak fuel enthalpy > 280 cal /ga).
1 For feed and bleed operation of a B3.W core, the highest worth ejected Sk rod at power is extremely low <.1%
/k, this is because the transient bank of rods (Group 7) is nomally operated > 80% wd from the core.
Therefore, for an evaluation of this low worth rod ejection accident, 1
the tuost important parameters for consideration are the subchannel rate of power increase and the total local power generation.
The cochination of these two parameters will detemine the heat transfer and therefore, fuel-rod enthalpy.
Contrary to the ejected rod analysis in the FSAR, where the local peaking factor doubles or triples in less than.3 seconds, the small rod ejection causes a very small local peaking change.
This results because the power is depressed to the bottom of the core in the ejected-rod location before the rod ejection; af ter rod ejection, the
' cxial flux shape actually flattens resulting in minimal percentage changes in both radial and total power peaking.
-w~
a m-e
-r
.x
,/.
/
For the average-channel power response, the small ejected red accident without trip on high flux provides a corparison with the large ejected ved analyzed in the FSAR.
First, the peak power is considerably lower than the FSNt ejected-rud case and secondly, the local power rate of change is relatively small.
fin illustration for the above justification considers the power rate of A
change for two ejected rod cases.2 and.45%
/k and the resulting energy balance as a function of the time the power is above 100% FP.
~
For a.2% ^I/k ejected md case, the peak power is 135% FP and decreases AI to 112% FP in 11 seconds whertas for a.451
/k ejected rod the peak power is 245% FP at.15 seconds and decreases to 100% FP at 1.4 seconds.
Therefore, the total stored energy in the fuel pin is the difference between thermi power generation and heat reraaval rate.
Since the power response after a snall rod-ejection is slow, and the average-channel power la relatively low with an insignificant total peaking change, the heat trcnsferred out of the pin to the reactor coolant within the 11 seconds should be significant and therefore, the peak pin fuel enthalpy will be less than 280 cal /gns.
Using the above observation and esaintaining nomal RCS flowrate, it chould be conservative to compare the amount of power excursion between small worths and FSAR ejected-rod transients.
The existing data indicates the stored energy should be between 200 to 250 cal /gu which is highly conservative-yet still below the limiting value of 280 cal /gn.
G l
l
,,,.--.,,,,.,,.__,.__.--,..,_-,_._--.7
, _, - - - - _, - --.,_ _-,_,__,-... -_-