ML19291F241
| ML19291F241 | |
| Person / Time | |
|---|---|
| Site: | Pilgrim, Limerick, 05000000 |
| Issue date: | 01/15/1982 |
| From: | Speis T Office of Nuclear Reactor Regulation |
| To: | Mattson R Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML19264C529 | List: |
| References | |
| IEB-79-08, IEB-79-21, IEB-79-8, NUDOCS 8202110034 | |
| Download: ML19291F241 (20) | |
Text
i
~
- -~'
DISTRIBUTION
/'
Central - File :
/~
TSpeis ICSB - Rdg.
.\\
RSB - Rdg.
I JAN 15 1987 Graves - Rdg.
(C, Gjj RSB - Subject MD10RANDLI4 FDR:
Roger J.14attson, Director
.n
~7 Division of Systems Integration
[
FRGi:
Themis P. Spets, Assistant Director for Reactor Safety Division of Systems Integration StBJECT:
ERRORS IN BUR VI.SSEL MATER LEVEL INDICATION 5[--ZR3 Attaciraent A provides a sumary of the results of wo'rk done to date in the RSB and ICSB under Task Interface Agrecaent 81-21 " Pilgrim 1. Water Level Instrumentation Oscillation." It is emphasized that review of this issue is rat complete, even though we have proposed some short and long-term recomendations. By copy of this meno I am requesting that comments or other relevant feedback on the contents of this nemo, and especially the proposed recommendations, be provided to C. Graves by 1/27/82.
t ?, maul Signed By Themis P. Spds Themis P. Speis, Assistant Directcr for Reactor Safety Division of Systems Integration
Enclosure:
As stated cc:
H. Denton W. Hodges G. Lainas J. Rosenthal T. Ippolito C. Graves S. Rubin B. Sharon L. Rubenstein G. Mazetis C. Scrlinger H. Thocason L. Phillips V. h-.s W.14111s D. I1caann T. Dente (CilR Owners Group)
D. Eisenhut F. Rosa T. Ilovak E. Rossi S. Hanauer COWACT:
C. Graves (x294D4)
J. Rosenthal (x29459) 8202110034 820115 CF ADOCK 05000293 CF
- SEE PREVIOUS CONCLRRENCE SHEET Q N.
RS_B,; DSI
.IC.SB: DS..I RSB:DSI ICSB:DSI ICSB:DSI RSB:DSI
'Spih)5',)F ADR
. 5 o mc,,
sunaut).CGCAv.p~: c s*
JRosenthal*
Hodges*
Rossi*
Rosa *
,.,$' heron *,,
1/8/82 T/Yli/52"."..."."..
""l?B7Ef""' '
"17TU782" *
~ 'T71'1782~
~1712782""' T/75782"""-"
'une ronu aw om escu ma OFFICIAL RECORD COPY usa m m m oso
x met 13RANDifi FOR: Roger J. Mattson, Director Division of Systems Integration FROM:
Themis P. Speis, Assistant Director for Reactor Safety, Division of Systems Integration
SUBJECT:
ERRORS IN BWR VESSEL WATER LEVEL INDICATION Attachment A provides a summary of the results of work done to date in the 'tSB and IC5B under Task Interface Agreement 81-21 Pilgrim 1, Water Level Instrumentation Oscillation."
Themis P. Speis Assistant Director for Reactor Safety Division of Syst::::s Integration ec:
F. Rosa E. Rossi W. Hodges J. Rosenthal C. Graves B. Sheron G. Mazetis CONTACT: C.GGraves (x29404)
J. Rosenthal (x29459) h r
-? u' u c4 9
<TJ I J d7
/)g ( ', !M x N5'B:DS,(,,
,IC,SB[5I']
RSBYbSI ICSB ICSB:DSI R9 F
Adit (DSI omet t sunuxuct CGray.e,sig,L..JBognthal. Bo@ s.,,,,
Ross) 1 Rosa A
.. Br.Cn. Spei g
1/ 72
..U.8.LM.....
.1/d.LU....
l/.6.L62...
.1/10./R 1/!U.82 1/!Ra2 9
om>
OFFICIAL RECORD COPY uso %_mm tac ronu ais oa-sa> wncu ana
e ATTACHMENT A BWR WATER LEVEL INDICATION ERRORS I.
INTRODUCTION On Sentenber 26, 1981, during a routine reacte-shutdown and cooling operation at Pilgrim 1, there were several large oscillations of Yarway level detection indication (reference 1). The first oscillation caused high level isolation followed by low level scram. The oscillations were attributed to high con-tainment temperatures, which caused flashing in the heated reference legs of the Yarway instruments. At the time, the reactor coolant temperature was about 220*F while the t-,perature in the upper part of the drywell was 240 F.
In a Task Interface Agreement of October 1981 (reference 2), NRP was assigned the following action plan items:
1.
Review event to establish the generic licensing implicatiors; (DSI/RSB & ICSB) 2.
Review adequacy of Pilgrim Tech Spec on high containment temperature;
( DSI/RSB) 3.
Determine acceptability of oscillations in safety related instruments; (DSI/RSB & ICSB)
This memorandum summarizes the results oa work in RSB and ICSB to date, provides preliminary responses to the Task Interface Agreement action items and lists some possible short and long-term solutions.
It is emphasized that the in-formation in this memorandum is preliminary since the review is not completg. A report dealing with the problem which was prepared for the BWR Owners was ob-tained from General Electric on 1 2/31/81 and has been given only a cursory rt view thus far.
Detailed discussions with General Electric personnel will be held after staff review of the GE report.
II.
BAC". GROUND As the result of t5 2 TMI-2 accident in March 1979, both the staff and industry
- ave reviewed the adequacy of level detection instrumentation under accident conditions.
In April,1979, IE Bulletin 79-08 (reference 3) requested information from each licensee on vessel level indication.
IE Bulletin 79-21, " Temperature Effects on Level Measurenent" (reference 4) was issued in August,1979.
This bulletin addressed errors in steam generator water level resulting from high energy line breaks including LOCA, inside containment and consequential high containment temperature which caused temperature increases and possible flashing of water in the reference leg of the level indicatcr.
The problem was identified in a Westinghouse letter of June 1979. Although the bulletin required actions from PWR operators, it was also sent as information to all BWR operators.
A staff letter (reference 5) addressing this problem was sent to all EWR liceasees in July 1979.
In July,1979, General Electric notified its customers of false level indication caused by high temperatures and possible flashing of water in the reference legs of Yarway le"31 instruments under post-L0s ; conditions (reference 6).
In September,1980, General Electric again notified its customers of the importance of compensating for these false level indications in Yarway instruments and described false level indications in cold reference lea instrunents caused by flashing in the sensing lines (reference 6).
A staff review and evaluation of level instrumentation errors for BWRs, based on a review of GE information provided in Auoust 1979 in NED0-24708 (reference
- 7) is oresented in NUREG-0626 (reference 8).
Additional information c n the safety significance of errors in or total loss of level indication was provided during 1980 in fiEDD-24708A (reference 9) and f4EDD-25224 (reference 10). Some current information is available in the propcsed emergency procedure guidelines for EWRs which are presently under sieif review (see reference 11 and recent revisions) and in the Shoreham docket (reference 12).
M
I I I..
WATER LEVEL INSTRUMENTATION All level measurement systems in BWRs employ differential pressure transmitters, a reference leg connected to a condensing pot and in turn to the reactor vessel steam space, and a variable leg connected to the vessel at a lower elevation.
Several differential pressure cells share common impulse legs. Temperature compensated and uncompensated reference legs are employed.
Those level measure-ment systems which use a temperature compensated reference leg are called Yarways.
Those level neasurement Lystems which use an uncompensated reference leg are called cold leg instruments or, often, GEMAC.
BWR 1, 2, 3 and sone 4's use two redundant Yarways to generate engineered safety feature actuation signals and cold reference leg instruments for indication and control.
The remaining BWR 4's and all 5 and 6's use redundant cold reference leg systems exclusively.
-5 A.
Yarway (Heated Reference Leg) Instrument A schematic of a Yarway level detector is presented in Figure 1.
Steam condensed
'in the condensing chamber maintains the reference leg water level by overflow to the variable leg.
The condensate heats the variable leg which, in turn, heats the reference leg.
A thermal shield is provided to reduce heat loss to containment and to maintain relatively high reference leg temperatures.
For short column Yarways,
netal clamos have also been used to inorove heat transfer between the legs.
Information in reference 9 indicates that the reference leg temperature is roughly equal to local containment temperature plus 40 percent of the difference between reactor steam temperature and local containment temperature.
For example, a local containment temperature of 135 F and steam temperature of 546 (Tsat at 1000 psia) would result in a reference leg temperature of 300 F.
The sensing lines leading from the Yarway to the differential pressure cell outside of the drywell are 1" schedule 80 stainless steel piping.
Flow in these-lines is blocked by the differential pressure cell.
During normal operation, the stagnant water in these lines should be approximately at local containment temperature.
If the lines are installed close to each other in containment, they should have about the same elevation change and local temperature, Hence, the effects of water density variations along the lines should be cancelled and have a minor effect on level measurement.
The Yarway level detector, which measures the collapsed water level in the outer annulus region of the reactor vessel, is subject to a number of uncertainites.
Those resulting from differences between actual and assumed values of averag coolant density in the annulus (affected by system pressure, subcooling and carryunder) were shown to be small in reference 9.
However,in 1979 the General Electric Company identified rather large uncertainties associated with high reference leg temperatures that could occur under some accident conditions (steam line breaks) for which local containment temperatures up to 340 F are predicted.
. The high reference leg temperatures would result in false high water level signals.
In addition, a constant indicated lower water level could be reached even though the actual water level has dropped well below the low level tap at the reactor vessel.
Hence, GE recorrnenOd that its customers review calibration of the Yarway instruments, increase certain trip points and take other corrective actions to com-pensate for this effect.
High containment temperature combined with reactor depressurization can also lead to false water level readings because of flashing or boiling in the reference leg or the sensing lines within containment lea.dng to the differential pressure sensor.
Flashing in the lines might occur during depressurization if the local containment temperature exceeds the saturation temperature corresponding to vessel pressure.
Flashing in the reference leg might be expected ear ier in the transient because of the higher initial temperatures in the reference leg. The GE communication of 1979 was concerned only with the effects of flashing in the reference leg of Yarway instruments.
Apparently, fleshing in cold reference leg instruments was considered to be of minor importance at the time.
In a later communication (September 1980), flashing in the sensirg lines of cold reference leg instruments was -also cc'nsidered.
Flashing in the reference leg or lines could occur during normal system depressuriza-tion in preparing for initiation of R4R cooling or under accident conditions.
During thc cooldown event at Pilgrim on 9/26/81 (see reference 1), flashing of the reference legs in the Yarway instruments was indicated by several oscillations in the level readings.
At the time, the reactor coolant temperature was 220 F and peak local containment temperatures were about 240 F.
Under accident conditions such as a steamline break, local containment temperatures can reach 340 F.
Hence, when vessel pressure drops below about 112 psig (p t 340 F) flashing could occur in the sat lines.
If it is assumed that the reference leg terrr rature rapidly increases to the steady state value for a containment temperature of 340 F and RCS temperature of 546 F, flashing in the reference leg might occur when vessel pressure drops below about 300 psig (Psat at 422"F).
Another scenario involving flashing in the reference leg could occur for larger breaks and times such that the vessel pressure is about equal to containment pres-sure.
In this case, as discussed in reference 13, the rapid reduction in containment pressure following initiation of the containment spray, combined with the delay in reduction of metal temperatures, could cause flashing in the reference leg. Tests were conducted to confirm that large errors in level indication could occur.
The solution to this flashing problem involved installation of a cooling jacket around the reference leg which was supplied with water from the containment spray line.
Even without a break, loss of the non-safety grade containment coolers would cause the containment to heat up and could cause flashing upon depressurization.
With respect to the flashing problem it should be noted that there would be a time delay involved in the heating of the reference leg and lines under accident condi-tions. A delay in heat transfer would be expected because of the relatively large amount of metal in the walls of the reference leg and lines and the relatively low heat transfer coefficients expected for surfaces in contact with the containment atmos-phere.
In reference 9, the thermal time constant for the Yarway detector was estimated to be about 20 minutes.
This value may have been calculated assuming only high temperature air.
For steam-air mixtures, the condensation ou cold surfaces results in appreciably larger heat transfer coefficients than those for air at the same temp-erature.
It should also be noted that water expelled by flashing in the heated reference leg and corresponding line to the differential pressure sensor may.not be replaced quickly. At the high containment temperatures and lower vessel pressure expected under accident conditions, the condensing chamber could cease to function.
Hence, refill would be delayed until sometime after the vessel water level increases to a point above the tap leading to the condensing chamber.
Even under these cir-cumstances, boiling could occur for a while in the reference leg and lines as the re-sult of continued high local containment temperatures.
In the case of degraded core
cooling when water level remains well below the tap to the condensir.g chamber and noncondensible gases and superheated steam could be present, there could be extended time periods with large false indications of vessel water level.
In fact, purging of the lines could be required to remove non-condensibles.
B.
Cold Reference Leg Instruments A schematic of a cold reference leg instrument is presented in Figure 2.
In this case, the reference leg upper level is maintained by overflow of condensate in the condensing chamber back through the tap to the vessel.
Water density effects and flashing in the lines within containment which lead to the differentia' pressure sensor -
could be of concern.
Changes of elevation in the lines inside of containment range from 1 to 40 feet in operating plants.
Hence, flashing in the lines under accident conditions could cause false water level indications and delay in refill problems such as those discussed in Section A.
Flashing in cold reference leg level instru-ment lines was recognized in the guidelines developed by GE (reference 11). This situation (loss of reliable level indication for both heated and cold reference leg detectors) was treated by operator instructions to initiate ADS and ECCS actuation to fill the vessel and overflow to the suppression pool via the S/R valves.
_9_
I V.
RESPONSE TO SPECIFIC ACTION ITEMS:
1.
Review event to establish the generic licensing implications.
All EWR vessel level instrumentation, to some degree, is susceptible to reference leg flashing and consequential loss of level indicatien following rapid vessel depressurization such as observed at Pilgrim.
The generic BWR emergency procedure guidelines
- include caution and action statements related to loss of level indication.
The suscepta-bility of the level indication system to substantive non-conservative errors during event sequences which include depressurization, and the adequacy of emergency procedures is discussed below.
2.
Review adequacy of Pilgrim Technical Specification on high containment tempera ture.
The Pilgrim Technical Specifications do not include drywell temperature as a limiting condition for operation. We believe such a specification would be crudent to prevent undue equipment aging. However, a LCO on the pre-accident drywell temperature will not preclude post accident loss of vessel level indication.
3.
Determine acceptability of oscillation in safety related instruments.
Engineered safety feature actuation signals are generated using the following process variables:
High pressure core spray (HFCS) - vessel level or drywell pressure Low pressure core spray (LPCS) - vessel level or drywell pressure
- These guidelines are presently under review by the staff and are not, to date, employed at Operating Reactors.
Low pressure coolant injection (LPCI) - vessel level or drywell pressure Automatic depressurization system (ADS) - vessel level and drywell pressure Containment Spray (CS) - vessel level and drywell pressure Reactor Core Isolation Cooling (RCIC) - vessel level only.
Delays in initiation of engineered safety features due to reference leg heatup and boiloff have been considered in response to IE Bulletins 79-08 and 79-21. The staff concluded in NUREG-0626 that for all break sizes, the reactor either depressurizes fast enough to allow timely initiation of the low pressure system on high drywell pressure, or the breaks are small enough that (at worst) ECC functions occurred before the potential boiling of the reference leg fluid.
Furthermore, ESFAS systems employ latching circuitry except on the ADS level permissive to ensure that safety actions, once initiated, go to completion (IEEE 279).
Hence, concerns related to initiation accuracy for dutomatic safety systems due to reference leg heatup and/or flashing and concerns related to potential reference leg fluid oscillation have been previously and adequately addressed for design basis events; however, there are event sequences involving multiple equipment failure which will require manual initiation of engineered safety features.
For some accident scenarios involving a break inside containment, adequate indi ation of actual vessel water level could be lost for all pertinent level instruments as the result of flashing and boiling in the reference legs.
The emergency guidelines (reference 11 and revisions) consider the case
where the operator has recognized that ve" 11 level cannot be determined.
For this case, the guidelines involve actions to depressurize the reactor and to refill the system until it overflows to the suppression pool via the S/R valves.
However, if the operator fails to recognize that he has lost level indication and has a false high reading of water level, he might take action to throttle or stop ECCS systems in order to avoid filling steam lines or to reduce load on emer3cncy power systems.
In this case, the flash-ing or boiling in the reference legs could lead to operator actions prejudicial to plant safety.
V.
RECOMMENDATIONS These are preliminary. Once we have received feedback from people on the distribu-tion list and met with the BWR Owners Group, they will be finalized.
A.
Short-Term Recommendations (1) Operators should be warned that all level indication is susceptible to large inaccuracies. We are concerned that operators may have been trained to unduly depend upon cold leg instrumentation should they recognize errors in Yarway reference leg instrumentation.
A cursory examination of plant procedures at Pilgrim 1 and Browns Ferry show that concerns related to cold leg instrumentation inaccuracies have not been incorporated in their procedures.
The operators may have been warned of these concerns by other mechanisms such as training sessions.
We believe that utilities are aware of potential water level inaccuracies in Yarway and cold leg instrumentation based on staff review of GE docu-ments prepared for the staff and documents prepared for GE cwnerr Early documents recommended reliance on cold leg instrumentation. Later docu-ments warned that these instruments, depending on the plant specific in-stallation, might also exhibit substantive indicated level errors.
We do not know whether or not these concerns and corresponding warnings and actions have been communicated to the control room operators.
(2)
Plant specific emergency procedures should be confirmed and/or modified to:
(a) Clearly identify which level indicators in the control room employ Yarway reference legs and which employ cold reference legs, and direct the operstor to the appropriate indicators.
(b)
Include procedures to help the operator decide when level instru-mentation is to be mistrusted.
Relate specific drywell temperature indication, readily and reliably available to the operator in the control room, to reference leg temperature.
(c)
Include procedures to help the operator recognize those plant conditions and observed instrument responses which indicate successfu' refilling of reference legs following flashing.
(3) Operability limits of the temperature sensors used in (2)(b) above should be included in the plant Technical Specifications.
B.
Long-Term Pecommendations We believe that it is prudent to provide the operator with continuous reliable level information.
Event sequences have been identified during which reliable indication will be temporarily lost. This potential is addressed in the emergency procedure guidelines now under review by the staff.
Hardware modi-fications should be sought to address this problem.
We believe that operator recognition of loss of accurate level information as addressed in the emergency procedure guidelines is cumbersome at best. The operator is to relate indicated water level and drywell temperature using a table contained in a caution statement of the emergency procedures.
Indicated water level values beyond the ranges shown in the table are to be mistrusted.
Automation of these actions and decisions seems in order.
Should the operator decide that the water level indicators are to be mistrusted, the operator is to fill the vessel. Supposedly reference legs would ultimately refill.
At some point in the event sequence, the operator should be previded with positive means to confirm that reliable water level indication has been restored.
This problem may not be adequately addressed in the emergency guide-line procedures which are presently under staff review.
~
Several potential plant modifications are being considered by the staff.
It is not our intent to dictate hardware fixes.
Rather, we give the below recommendations as illustrations that reference leg flashing is a tractable pro bl em.
(1 ) Perform plant specific analysis of susceptibility of cold leg level in-strumentation to reference leg flashing and/or local heatup ad corres-ponding water expulsion.
Those plants which are designed with small vertical drops of reference legs inside the drywell should be satisfactory as designed.
(2) Consider rerouting of reference legs to meet condition (1) above.
(3)
Install temperature measurement of the reference leg. Such measurements could be used to confirm operability following a drywell temperature excursion and subsequent cooldown. The measurement would be of little use should high drywell temperatures be sustained.
(4)
Develop means to cool the referent' leg by establishing flow within the leg.
Two techniques have been suggested:
(1) the temporary opening of equalization valves and/or drain valves, and (2) pumping water with a positive displacement pump from outside the drywell, up reference lines and into the vessel.
Equalization and drain valves are local manual valves. They are hypothetically accessible following an accident. The drain lines are routed to the waste treatment system.
Following vessel depressurization, reference leg flashing and subsequent vessel filling in accordance with emergency procedures, temporary opening of the valves could be used to ensure reference leg filling.
No hardware modifications would be required.
Should a sufficiently large LOCA occur, or should an event sequence involving multiple equipment failure occur, such that the
~
vessel cannot be filled above the reference leg taps, this tec.hnique woald be of little use.
Pumping water up reference legs would ob-viously require hardware modifications. The flowrate need only be high enough to overcome the heat load on the reference legs inside the drywell under accident conditions. This technique would permit refer-ence leg filling even if high drywell temperatures existed and the vessel could not be filled to the reference leg tap.
(5) Develop means to cool the reference leg by using a coolant jacket and diverted ESF flow.
'7 M
REFERENCES
~
1.
Licensee Event Report 81-055/0lT-0, "High Drywell Temperatures".
Pilgrim Nuclear Power Station, 10/15/81.
2.
Task Interface Agreerent, Task No. 81-21, " Pilgrim 1, Water Level Instrumen-tation Oscillation", October,1981.
3.
IE Bulletin 79-08, " Events Relevant to Boiling Water Power Reactors Ident -
fied During Three Mile Island Incident", April 14, 1979.
4.
IE Bulletin 79-21, " Temperature Effects on Level Measurements", August 9,1979.
5.
Letter from T. Ippolito, NRC to C. Reet - Commonwealth Edison Company, "Addi-tional Information Required for NRC Stai. Generic Report on Boiling Water Re-actors", July 13, 1979.
6.
Telephone conversation with General Electric Company personnel, December,1981.
7.
NEDO 24708, " Additional Information Required for NRC Staff Generic Report on Boiling Water Reactors", August,1979.
8.
NUREG-0626, " Generic Evaluation of Feedwater Transients and Small Break Loss-of-Coolant Accidents in GE-Designed Operating Plants and Near-Term Operating License Applications", Janua ry,1980.
9.
NEDO 24708A, Revision 1, " Additional Information Required for NRC Staff Generic Peport on Boiling Water Reactors", December,1980.
10.
NEDO 25224, "GESSAR Assessment Report, Review of BWR/6 Protection In-Depth for Transient and Accident Events", June,1980.
11.
NEDO 24934, " Emergency Procedures Guidelines - BWR1-6", January,1981.
12.
Attachrent to letter from B. McCaffery of Shoreham Nuclear Power Station to H. Denton, NRC, August 18, 1981.
13.
Memorandum to Carl Berlinger, CPB, NRR, and Faust Rosa, ICSB/NRR, from N.
Kondic, ICB, DF0, "Two Phase Ruid Water Level in Nuclear Vessels (Reactor SG, PZR),
November 23, 1981.
DRYWELL ACCUgut ATOR H E AC1DR
~
DRywELL t's E 55URI WALL
~
VELLIL CO ND E N LING CHAveER STEAM
\\ Vl,:
ff N '/
-pEMOTE vt LLE L gno CAllON IN CONTROL LEVIL
./
ROOM 0 </ [ SggtL g
/
/
7 LRENCE VAR TAB L E
/
[
\\
/
d i
/,
/
- HE AT
,/, n [ /
, d TH AN5f ER
/-
/
C LA f/fs (! HO R T
//
p CO L U? '.N ON L Yi
/
[T
//
? L
~
CONTROL MOD
((,
4
=
/ LE VE L
/
/
1AP l/)
Figure 1-Yarv;ay (Heated Reference Le9) Leve1 Detection Instrucment (From NEDO-2470BA)
Cy wN Cona/en rny.
Wa //
ye,x)
CAeehee M
i
& fu ercr we tw
.:e g L osace, Top i
r) P Cc//
d iN
/'
n.ni.'s 4 43 Figure 2.
Cold Reference Le9 Level Detection Instrument