AECM-82-368, Forwards Review of NUREG-0737,Item II.F.2,Instrumentation for Detection of Inadequate Core Cooling, Per Supplemental SER 2,Item 22.2

Forwards Review of NUREG-0737,Item II.F.2,Instrumentation for Detection of Inadequate Core Cooling, Per Supplemental SER 2,Item 22.2

ML20063G887
Person / Time
Site:	Grand Gulf
Issue date:	08/30/1982
From:	Dale L MISSISSIPPI POWER & LIGHT CO.
To:	Harold Denton Office of Nuclear Reactor Regulation
References
RTR-NUREG-0737, RTR-NUREG-737, TASK-2.F.2, TASK-TM AECM-82-368, TAC-49459, NUDOCS 8209010199
Download: ML20063G887 (15)

Text

+g .

MISSISSIPPI POWER & LIGHT COMPANY k Helping Build Mississippi P. O. B O X 164 0, J AC K S O N, MIS SIS SIP PI 3 9 2 0 5 NUCLEAR PRODUCTION DEPARTMENT U. S. Nuclear Regulatory Conimission Office of Nuclear Reactor Regulation Washington, D. C. 20555 Attention: Mr. Harold R. Denton, Director

Dear Mr. Denton:

SUBJ ECT: Grand Gulf Nuclear Station Units 1 and 2 Docket Nos. 50-416 and 50-417 License No. NPF-13 File 0260/L-860.0 NUREG-0737 (II.F.2):

Instrumentation for Detection of Inadequate Core Cooling OL Condition 2.C [44(F)], SSER 2, Item 22.2, II.F.2 AECM-82/368 Mississippi Power & Light Company (MP&L) has performed a review of the draf t BWR Owners' Group (BWROG) analysis of inadequate core cooling for Boiling Water Reactors (BWR) prepared for the BWROG by S. Levy, Inc.

MP&L's review of the draft pre-published report by S. Levy, Inc. titled,

" Inadequate Core Cooling Detection in Boiling Vater Reactors" (Report No.

SL1-8218) is provided as Attachment 1 to this letter.

If you have any questions or require further information, please l

contact this office.

Yours truly,

"' ,n i

L. F. Dale Manager of Nuclear Services

! RAB/SHH/JDR:ac l

Attachment:

Instrumentation for Detection of Inadequate Core Cooling cc: See next page l

8209010199 820830 PDR ADOCK 05000416 E PDR Member Middle South Utilities System (hol

AECM-82/368 MI:0:212CIPPI POWER Q LIGHT CCMPANY Page 2 cc: Mr. N. L. Stampley (w/o)

Mr. R. B. McGehee (w/o)

Mr. T. B. Conner (w/o)

Mr. G. B. Taylor (w/o)

Mr. Richard C. DeYoung, Director (w/o)

Office of Inspection & Enforcement U. S. Nuclear Regulatory Commission Washington, D. C. 20555 Mr. J. P. O'Reilly, Regional Administrator (w/a)

Office of Inspection & Enforcement U. S. Nuclear Regulatory Commission, Region II 101 Marietta St. , N.W. , Suite 3100 Atlanta, Georgia 30303 t

i

ATTACHMENT NO.'1' NUREG-0737, ITEM II.F.2 INSTRUMENTATION FOR DETECTION OF INADEQUATE CORE COOLING Concern By August 31, 1982, MP&L shall submit a report addressing the analysis performed by the BWR Owners Group regarding additional instrumentation relative to inadequate core cooling and shall implement the staff's review of this report. These nodifications shall be completed on a schedule acceptable to the staff.

Response

INTRODUCTION The BWR Owners' Group (BWROG) study of inadequate core cooling detection (ICC) in boiling water reactors (BWR) was commissioned by the BWROG (owners) in response to the Nuclear Regulatory Commission's (NRC) plans developed in response to the accident at TMI-2. Those plans included Task II.F.2, " Identification of and recovery from conditions leading to inadequate core cooling." Later via Regulatory Guide 1.97, the NRC indicated that the installation of core thermocouples in BWRs and certain other changes would accomplish Task II.F.2. The General Electric Co. and most, if not all, BWR owners seriously questioned the usefulness and need for core thermcouples in BWRs. GE directly and with the owners reviewed the technical basis for them with the NRC staff without resolution of the issue.

Subsequently the owners commissioned a study of incore thermocouples as applied to BWRs. The results of that study confirmed the owner's reservation over incore thermocouples. Af ter presentation of the report and their reservation to the NRC staff, it was mutually agreed a general study of ICC detection in BWRs would be performed. As agreed with the staff, the study was to consider means for providing reliable information to limit core damage during accidents. This was to include the detection of trends toward ICC, the existence of ICC, and the return of adequate core cooling. Also, the cost / benefit of detecting local ICC was to be considered. In addition, the owners wished to determine if additional instrumentation for this purpose was warranted. A report on the detection of ICC to address the above issues is currently in draft form. This summary report is based on the pre-published version of S. Levy's (BWROG Consultant) report titled, " Inadequate Core Cooling Detection in Boiling Water Reactors."

/

RELATIONSHIP BETWEEN WATER LEVEL AND ICC To address the above issues the Owners' Group study first examined the relationship between reactor water level and the approach of, existence of and return from ICC. This examination justified reactor vessel water level as a viable indicator of ICC; and the conclusive variable for operational control of BWR reactors for the avoidance or mitigation of ICC.

A more exacting definition of ICC for BWR's was arrived at by the Owners' Group study based on the effects of high temperature on fuel cladding. The peak fuel surface temperature denotative of chemical interaction of the fuel clad material with any available water is the threshold for fuel assembly deterioration affecting the fuel cooling process. The Owners' Group study denotes that peak fuel surface temperatures between 1300*F and 1800*F in an average fuel assembly is a plausibile definition of ICC.

The reactor operational conditions of power level, water level, and recirculation flow were reviewed to identify the operational regimes in which ICC is important or even possible. The constant pressure mode of operation and the controls to assure proper heat removal were outlined.

The only significant regime where ICC can be a concern relative to NUREG 0737 Item II.F.2 is a post shutdown decay heat removal regime. This conclusion was drawn after an investigation of various reactor powers, water levels and recirculation flows. At high power ICC is produced by operation beyond the critical heat flux. This condition is precluded by the power-flow trip on the approach to unsafe conditions. 'At low inventory conditions, ICC is produced by stagnant boiloff as the steam-water mixture height drops below the active length of fuel.

i To illustrate the low inventory condition, consider the case of a j scrammed reactor with recirculation pumps tripped, vessel isolated, no

) RCIC or ECCS and no line break, calculations predict fuel channel bypass and downcomer levels are conservative indicators of core water level. The water level measurement system senses level from'within the core downcomer zone, which gives the most conservative information on core coverage with respect to water level and hence to peak cladding temperature. It can be-concluded from Figure 1 that, for the accident postulated, peak cladding temperature is a function of water level and from Figure 2 that peak cladding temperature has low sensitivity to core'uncovery time. Water level is a reliable indicator of peak clad temperature, therefore a reliable indication of the approach to and the existence of ICC. In the event that ICC does take place, the restoration of water level above the top of active fuel will indicate the return to adequate core cooling.

Various industry studies and tests (listed below) indicate that core damage will not propagate once the core is recovered.

1. NED0-20355A, "The Effects of a Large Bundle Flow Area Restriction on the BWR Emergency Core Cooling System Effectiveness", August 1976

! 2. NEDO-10174, " Consequences of a Postulated Flow Blockage Incident in a BWR", October, 1977. .

3. NEDO-10208, " Effects of Fuel Rod Failure on ECCS Performance",

August 1970.

ADDITIONAL ICC DETECTION

~

i While water level measurement is the primary ICC detection device in a BWR, there are several others which indicate the adequacy of core cooling. Chief among these is the core spray flow rate. Each of the two core spray systems is capable of cooling the core by spray action.

Because they cool by direct spray onto the core, adequate core cooling is i provided independent of water level. Thus, if the measured flow of either spray system reaches or exceeds rated flow, adequate core cooling is indicated.

Flows to and from the vessel are indications which confirm or support water. level indication. Level indication trends must be consistent with the vessel inventory changes inferred by these flows. As an example, consider a case of interest from an inadequate core cooling standpoint.

Assume the reactor has entered an isolation status and that a high-pressure injection system is making up inventory. Assume, also, as would be expected, that the indicated level moves from normal level to high level. The automatic control systems or operator can be expected to turn off all injection as a result. Now, postulate that this upward indication movement was the result of level indication failures which were undetected j by the level measurement validation process. These failures will be

revealed by the operator observing that level indications do not fall even though there is no flow to the vessel and the safety / relief valves periodically opening and removing inventory. As has been shown, it would take over 40 minutes for the actual level to reach the top of the core.

Therefore, there is considerable time available to observe the inconsistency between the indicated level trend and the. lack of inventory j makeup.

Should ICC conditions exist, they would be revealed by containment gross gamma, containment hydrogen concentration, and' reactor and suppression pool water sample activity measurements.

j RISK CONTRIBUTION OF CURRENT ICC SYSTEM t

The risk contribution of the water level measurement system to core degradation probability was evaluated based on modifying an existing PRA for a BWR-4 plant with a Mark-II containment. The core power density and plant pressure rates on a BWR-4 are between those of earlier and later plant designs. The Mark 11 containment in the reference plant represents an evolutionary step between the bbrk I and thrk III containments.

The uncertainty in the calculated best estimate frequency of core melt in the subject PRA had been quoted as approximately a factor of 20.

From this is can be seen that the absolute value of core melt frequency may be significantly higher than the best estimate value quoted hege.

Specifically, the core melt frequency could be in the range of 10- / year and still be within the 90% confidence interval.

i

The selected BWR is therefore broadly representative of the total BWR population. However, this is not meant to imply that the results given here may be applied to a specific plant. There are considerable differences between plants that would affect the calculated core melt frequencies, and plant-specifde analysis would be required.to identify core melt frequencies for a particular plant. The results provided are indicative of the relative contribution of the water level system for a

" typical" plant.

In the subject PRA, three sources of water level indication error were not considered in the original PRA. They are: 1) loss of drywell

, cooling; 2) instrument line failure; 3) level instrument failure. Event trees of the subject PRA were modified to include contribution of the above sources of water-level indicator errors. Comparison of original degraded core frequency to the degraded core frequency calculated in this study revealed the following contributions stated in event frequency and as percentage of the original total event frequency (14 events /million years):

Loss of drywell cooling 0.74 events /million years . 3%

Instrument line break 0.58 events /million years 4..%

! Instrument failures 0.46 events /million years 3.2% ,

TOTAL 1.78 events /million years 12.6%

COST EFFECTIVE BACKUP TO WATER LEVEL MEASUREMENT l To evaluate cost effective backups to water level, alternative core cooling measurement devices were evaluated. A broad spectrum of possible devices (Table 1) were subjected to preliminary screening, performance i

evaluation, and finally a cost comparison to determine four viable possibilities.

Heated Junction Thermocouples in LPRM tubes can detect the presence of water (level) by measuring the heat transfer rate capability at the sensor location via a change in temperature difference between the heated and the non-heated thermocouples. Normal thermocouples in LPRM tubes could detect ICC in some cases but would suffer from ambiguous outputs.

4 Thermocouples located in the steam dome are at best as good an indicator as the above T/C's in LPRM tubes. Normal t'ermocouples h placed in the LPRM tubes or steam dome would have appreciable time lag that would prevent detection of the approach to inadequate core cooling. SRM detectors may be able to sense ICC by a sudden drop in thermal neutron flux when water level drops below the SRM detectors, but considerable development work is needed for these devices. Costs of these four alternatives are summarized i in Table 2.

A cost / benefit exercise based on an NRC proposed technique (SECY 81-513 August 1981) for prioritizing safety issues was used. EPRI RP-1585 published June 1982 catagorizes the priority score as follows:

. . , _ _ - . , ~ . .-_ _ - _ _ _ . - ~ --

l l

S = 1 - 1,000 Low 1,000 - 10,000 Medium ,

greater than 10,000 High Safety Benefit = Ns$ [FR] R

• S= Cost C + NI where, S = Priority score N = Number of reactors affected R = Consequences, in curies released R* = Weighting Factor F = Event frequency in events per reactors years C = Forward looking NRC cost in millions of dollars I = Forward looking industry cost in millions of dollars per reactor

/1 = Mathematical operator to indicate the change in the quantity within the brackets.

The NRC priority score falls well within the low priority range (S=91) based on an alternative device increasing the probability of recognizing the threat of ICC by a factor of 5 (training on procedures) and a cost of $3 million for a new cevice. The factor of five means additional instrumentation would inform the operator of ICC so that he could take manual action not previously taken, and that the operator would successfully perform the desired function. A range was calculated of S=8 to 1030 based on the square root of the sum of the squares of the assumed uncertainty in each term. This analysis supports our position that such a fix is also very low priority for Grand Gulf.

CONCLUSION In conclusion, the draft BWR Owners' Group analysis of Inadequate Core Cooling demonstrates that knowledge of measured reactor vessel water level provides a reliable means of determing whether the core is adequately cooled. Grand Gulf's full range of vessel water level measurement, procedures, and training gives the operator indication and instructions for proper actions. Reliabile information to limit core damage is provided at Grand Gulf. Detection and trending of the approach to ICC is measured by the existing Grand Gulf water level system in that the relationship of water level and approach to ICC was established in the Draft Owners' Group Report. It was determined that it is not cost beneficial to detect local ICC. Although many devices for ICC detection were evaluated generically, they are not expected to survive extreme conditions long enough to provide significant information for local ICC detection. In light of the low degree of risk and existing ICC detection, it is determined that no additional instrumentation is needed to detect inadequate core cooling. As shown in Figure 3, the risk remaining after either additional ICC devices or improved water level measurement reliability improvements is diminishingly small, indicating that the application of both additional ICC devices and water level measurement reliability improvements is not warranted.

a

_(_

TABLE 1 POSSIBLE ICC DETECTION DEVICES NAME OF DEVICE NAME OF DEVICE Source Range Monitor Wave Guide Intermediate Range Monitor Vessel Weight Local Power Range Monitor Vessel Vibrations Traveling Incore Probe Floats Gamma-Neutron Reaction Detector Conductivity Probe Gamma Attenuation Capacitance Probe Gamma Void Meter Sonic Reflection Neutron Modulation Void Meter Loose Parts Monitor Core Reactivity Detector Microwave Probe Fuel Plevum Tracer Mass Balance Primac System Activity Meter Differential Expansion Integral Anemometer Incore Thermocouples Delta-P Bubbler Heated Junction Thermocouples Self-Powered lieutron Detector

, Gamma Thermometers Resistance Temperature Detectors l

Control Rod Drive Thermocouples Steam Dome Thermocouples l Sight Glass Liquid Level and Void Fraction Detector l Cerenkov Light Detector l

l l

l l

TABLE 2 COST

SUMMARY

OF ALTERNATIVE ICC DETECTION DEVICES EXPOSURE MAN / REM COST MIN. MAX.

HJTC $2.9 Million 65 450 T/C (LPRM) $2.5 Million 65 450 T/C (st. dome) $ .8 Million 16 80 SRM $1.3 Million 16 100

O

'I I I I I I I I i i 3400 -

1 l CLADDING MELTING POINT PEAK CLADOING TEMPER ATURES 3200 - 6 ENTIRE CORE O AVERAGE POWER BUNDLE

(.

O LOWEST POWER BUNDLE 3000 -

l0 -

2800 -

(7 2600 -

2400 -

• I i *

() E

$ 2200 -

( _

5

• 1 a-L 3 2000 1

O Q E 1800 -

d j w

1600 -

1 (i

1400 -

\

j l

1200 -

\ \ . I -

l .>

g 1000 -

l .

N * '

NN 800 -

\ -

l N  %

N l' 600 -

400 '  !  !  !  !

i 0 1 2 3 4 6 5 7 8 9 to 11 12 l

WATER LEVEL IFEET ABOVE BAF)

Water Level As An Indicator of Core Overheating Figure 1 1i l

)

0 I I I I i l l l l l l 1400 - PEAK CLADOING TEMPERATURES _

O AVERAGE POWER BUNDLE O LOWEST POWER BUNDLE O d HIGHEST POWER BUNDLE

• = WATER LEVEL 1300 -

p 4 FEET

• _

/

/

r -

1200 -

O 1100 -

4 FEET *

{

e o a 5

m 2 1000 -

c E

iO 8 5

l u 900 -

O 4 FEET *

.9 800 -

_g '

()

% > 6 FEET * , ,

l 700 -

~

&_ ~S s l

l .) ,

,_ - - O ==-- .

- 8 FEET *

^

T i 600 -

I I I I I I l  ! l l [

0 10 20 30 40 50 60 70 80

f. 90 100 110 120 TIME AFTER SCRAM (minutes)

Cladding Temperature Sensitivity to Core Uncovery Time ca Figure 2

- ' ' 7, ;; .

^

u C  :.c. w 0 a ^

O O w ,

Sisk of Core Melt - % >

0 10 20 30 40 50 60 70 80 90 100 I I I I I I I I I I I N N Overall Plant Risk 5' of {

m Core Melt 5 g

\ -

\ \ $

g.

Risk Contribution of Current Water Level Measurement System

{

3, er N

/ l Risk Contribution of Current Water N

Level Measurement System with Additional ICC Detection Devices N

/

Risk Contribution of Improved Water Level

/ ~

y Measurement System with Enhanced Failure Detection

.