ML20079L233
| ML20079L233 | |
| Person / Time | |
|---|---|
| Site: | Perry  |
| Issue date: | 10/30/1991 |
| From: | Lyster M CENTERIOR ENERGY |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| Shared Package | |
| ML20079L239 | List: |
| References | |
| PY-CEI-NRR-1390, NUDOCS 9111060181 | |
| Download: ML20079L233 (27) | |
Text
_ _ _ _ _ - _ _ _ _ _ - _ _ . _ - _ .
l CENTERCOR ENERGY PERRY NUCLEAR POWER PLANT Ad ess Michael D. Lyster
"'""*"*" V'cr entsiorni- NuctrAn Ern5$$4Sosi (216) 259 3731 October 30, 1991 L PY-CEI/NRR-1390 L U.S. Nuclear Regulatory Commission Document Cnntrol Desk Vashington, D. C. 20555 Perry thelear Power Plant Docket No. 50-440 Technical Specif!:ation Change Request, Reactor Vater Cleantip System Isolation Actuation Instrumentation Gentlemen:
Enclosed is a request for amendment of the Pacility Operating License HPF-58 for the Perry Nuclear Power Plant, Unit 1. In accordance with the requirements of 10 CFR 50.91(b)(1), a copy of this request for amendment has been sent to the State of Ohio as indicated below.
This amendment requests revision of Technical Specification Table 3.3.2-2 by adding .iev isolation signal setpoints and allovable values for temperature and delta-temperature instruments in the Reactor Vater Cleanup (RVCU) Containment Rooms (i.e. the Demineralizer Rooms, Demineralizer Valve Room, and Receiving Tank Room-nev items 4.c.3 and 4.d.3), and changing the delta-flov timer setpoint and allovable value (item 4.b).
Attachment 1 provides the Summary, Safety Analysis, Significant Hazards and Environmental Impact Considerations. Attachment 2 is a copy of the marked up Technical Specification pages.
If you have any questions, please feel free to call.
S i nc e r e,ly .
(l ,
W Michael D. Lyster MDL:BSP:njc Attachment cc NRC Project Manager NRC Resident Inspector Office NRC Region III State of Ohio
^0]
I a v~-
cae ,
4
- niu#oisi ottoao P ti e^ 't o..riv P DF< ADOCK O*?'000440 k\ l
\ 'i U ' L p POR
,a i Attachment 1
, PY-CE1/NRR-1390 L Page 1 of 26 >
Summary The Reactor Vater Cleanup (RVCU) system provides continuous purification of reactor vater to saintain reactor Vater chemistry limits by removing impurities that vould otherwise remain in the vessel vater since they are not carried over with the steam. It is also used to reduce the fouling of heat transfer surfaces, to minimize secondary sources of radiation, and to maintain vater clarity during refueling. Since the system penetrates the containment boundary, General Design Criterion (GDC) 54 requires that the system be provided with leak detection, having redundancy, reliability, and performance capabilities which reflect the importance to safety of isolating the piping system. At Perry, the leak detection features for the RVCU system currently .
include an RPV vater level 2 isolation signal, backed up by temperature and I differential-(delta) temperature sensors to indicate a potential leak in the portions of the RVCU system containing hot vater, and differential (delta) flov instrumentation to indicate a potential leak in the cold portion (which it was assumed would escape detection by the temperature sensors).
Numerous spurious RVCU isolations from the delta-flov signal have been a significant source of undesirable challenges to safety equipme..t and of losses of RVCU availability. The purpose of this proposed Technical Specification change is to reactivate the isolation signals from the temperature and delta-temperature instruments in the containment rooms containing the co'.t portion of the-RVCU piping and to extend the delta-flov timer length (while leaving the delta-flov setpoint unchanged). Implementation of this proposed change vill result in a number of significant benefits to plant operation and safety. These include a reduced number of challenges to the isolation valves and the pumps, and an increased availability'of the cleanup capabilities of the system both during operations and during plant startups/ shutdowns when the reactor coolant pressure boundary is exposed to the conditions most conducive to stress corrosion cracking and to the release of activated corrosion products into the piping systems, which increases dose rates within the plant.
It vill also reduce the number of times that control room operators need to respond to a spurious, unnecessary isolation of the system, again aspecta'.ly during plant startups and shutdowns when there are many other important safety-related activities ongoing.
Due to the large number of isolations of the RVCU system experienced because of '.he delta-flow instrumentation, a Technical Specification change has been under consideration for some time. In addition to presenting the basis for the specific change which is being proposed, this letter also presents a history of the isolations,.the results of investigations into the isolations and the corrective actions, and the lessons learned from both the continuing isolations and the different options that vere considered as Technical Specification changes. The lessons learned bear directly on the option that has been chosen for change per this license amendmeat request.
L __ __
Attachment 1
, PY-CEI/NRR-1390 L Page 2 of 26 Leak Detection Logic Discussion A simplified diagram of the RVCU system is provided in Figure 1. Because the system ptocesses high energy and high activity coolant, the system components are located within separated compartments (rooms) which can be individually monitored for leakage. Piping that runs between the Ileat Exchanger Room and the Containment Valva Room (which are not immediately adjacent) is embedded in concrete to preclude leakage from the pipe.
The RVCU leakage detection features include multiple diverse sensors to detect leakage from the system. This diversity includes reactor vater level instruments, various area temperature and delta-temperature sensors to -
indicate a leak of hot water or steam, and delta-flow instrumentation to provide detection of any leakage cold enough to escapo detection by tha temperature-based methods.
The RPV Vater Level 2 isolation signal is the primary plant protection feature for a loss of reactor coolant. This signal provides closure of selveted containment isolation valves, including the RVCU valves, when reactor water level drops due to any significant loss of vater. Vith the exception of the containment pressurization case discussed in Section 6.2.1.1.4.2 of the Perry USAR, the RPV vater level 2 isolation is the only RVCU system isolation for which credit is taken in the Accident Analyses of Section 6 and 15 of the USAR. The pressurization analysis discussed in Section 6.2.1.1.4.2 of the USAR takes credit only for isolating hot vater or steam leaks (the function of the temperature-based detection methods discussed belov). USAR analyres do not take credit for delta-flow isolation of the RVCU system. Thus the purpose of the RVCU delta-flov signal is not to protect the reactor (tl.e fuel) against a loss of coolant accident either large or small, or to protect against
-containment pressurization.
The temperature-based leak detection methods serve to isolate the RVCU system in the event of hot water / steam leakage which ci.'d pose a risk of room pressurization or early offsite dose consequence (due to release of radioactive steam). The delta-flov' instrumentation serves to provide detection of cold vater leakage from the portion of RVCU that contains cold reactor coolant that might escape detection by the temperature sensors. The purpose of detecting cold leakage is to detect a vater loss that might eventually lead to offsite dose consequences.
This cold vater portion (see Figure 1 and 2 for system details) consists of the portion dovnstream (i.e. away from the Reactor) of the Non-negenerative Heat _ Exchanger, through the filter /demineralizer portion of the RVCU system and back to the return side of the Regenerative Heat Exchanger where the
- vater is reheated for return to the reactor system. This cold portion also includes the blevdown line to the main condenser and radvaste system, which is used by the operator to adjust reactor system vater inventory. The normal vater temperature in this cold portion is approximately 100-120'F. The purpose of using delta-flow instrumentation for detection of~ cold water ,
b
Attachment I ;
, PY-CEI/NRR-1390 h '
Page 3 of 26 t
leakage is to provide for an isolation before cold vater leakage reaches the l point which may result fn an offsite dose concern. Since the putpose of the '
delta-flov instruments are to detect this cold water leakage, it vould require !
a tremendous amount of cold leakage to occur in order to significantly affect !
offsite doses since the vater is not flashing to steam and becoming airborne.
This is recognized by dose calculation methodologies as discussed in the Standard Review Plan (NUREG 0800) Chapter 15.6.5 Appendix B,Section III.
The RVCU delta-flow logic block diagram is shown on Figure 3. A flov summer compares inlet (suction) RVCU flow against the sum of RVCU return flov to the reactor and RVCU system blovdown flov (going either to the main condenser or to Radvaste). If the output of the flov summer exceeds the trip setpoint, indicating there is more flov going into the RVCU system than accounted for in the return and blevdown lines, a timer (currently set at 45 seconds) and an alarm actuate. If the trip r.tpoint is exceeded for longer than the 45 second ;
timer, an isolation signal is sent to the RVCU isolation valves. !
History of RVCU Delta-Flow Isolations and Corrective Actions Since the Perry Nuclear Power Plans i.eceived its Operating License in March of 1986 there have been numerous spta u,us isolations of the RVCU system caused by the delta-flov measurement. These 1sulations have been documented in Licensee ,
Event Reports (LERs), with four LERs in 1986 (LERs86-039, 86-056,86-068, and 86-085) that discussed 17 different isolations, four LERs in 1987 (87-001,87-013, 87-026,87-074) that addressed 12 isolations, three LERe in 1988 (80-002,88-013. S8-039), which documented 3 isolations, two LERs in 1989 (89-025,89-031) that disev/s 4 isolations, and one in 1990 and in 1991 (LERs90-022 and 91-011) that co'ter 2 1solations in 1990 and 1 in 1991. Despite the ;
tremendous reduction in the number of isolations, complete elimination of these spurious events has not been achieved.
Even prior.to PNPP's receipt of an operating license, the never BVR plants ,
that had'the delta-flow isolation signal incorporated into their leak l detection logic due to a General Electric (GE) design specification requirement (older BVRs do not utilire this isolation logic), had been experiencing numerous spurious delta-flov isolations. As experience was gained with PNPP's RVCU system, the causes of the isolations at PNPP in 1986 and 1987 vere identified, and corrective actions vere implemented, either through design changes or procedural controls. Review of s report issued by 4 i the NRC's Office for Analysis and Evaluation of Operational Data (AEOD) in l March 1987, and subsequent interface with fifteen other utilities, General l
Electric and INPO between April and September of 1987 through an industry group, shovea that the concerns identified at-PNPP vere consistent with those being experienced elsewhere. The reports and interface also shoved that PNPP ,
had developed and implemented for had plans to implement) corrective actions that vere comprehensive and would eliminate the majority of operational difficulties that had been experienced in trying to run the system without flow perturbations.
l
.s Attachment 1 ,
pY-CEI/NRR-1390 L !
page 4 of 26 In the third quarter of 1986, a dynamic compensator was added to the -l electrical summing network of the delta-flov instrumentation in order to .
electronically filter noise that was aggravating flow perturbation signals. '
In addition, the meter for the delta-flov timer on the "A" channel was moved !
to the front of the panel rs that the operator could view the elapsed time !
since the delta-flov timer had initiated. In the fourth quarter of 1986, one i of the main contributors to system flov oscillations was elimir.ated when the -
I'lov orifice that measures the flov in the blevdevn leath to the condenser was ralocated from a location downstream of the pressute control valve in that line to a nev Iccation upstream of the valve. This was important because the i previous location had subjected the ilov element to turbulent flow conditions r that caused the flov element to become uncovered due to the lack of back pressure in the dovnstream portion of the line, resulting in en indication of i large flow oueillations and leading to delta-11ov isolations of RVCU. A third ;
design change, implemented in the first ouartet of 1987, involved the addition i of electroute compensation for the differences in flow that the i instruaentation reads between full-power og rating pressure and temperature conditions and redured preasure and temperature conditions. In all the 1986 and 1987 isolation events, the plant had been in Operational Condition 2, 3 or 4 at less than rated temperature and pressure. Since the system had only been calibrated for the utnnity and temperature of vater at rated conditions, an otror was introduced at off-rated conditions that essentially reduced the delta-flov setpoint. This " reduced setpoint" effect was minimised by the addition of a reactor pressure signal into the logic, which provides a means of compensating ice the density and temperature dif ferences at of f-rated '
conditions. In the second quatter.of 1987, this compensation factor vas revised based on additional testing performed on the system.
Anotner major contributor to isolations in 1986 through 1988 vas the poor throttling characteristics of the flow control valves in the systems stroking o of thr valves induced large erratic flow oscillations. A design change had "
been prepared in 1966 to replace the flow control valves with different valves that have better throttling charatterlatics. However, this major modification to the piping system could not be implemented until the first refueling '
outage, which didn't occur until 1989.
Finally, in the firs *, quarter of 1988, the cause of several 1987 isolations vas eliminated when replmcement square root extractors for the delta flov
- instruments were obtained which had a lover " clip" point. A number of isolations had occurted due to lov system flovs, since the previous square root extractors gave a signal of zero (0) gpm if flow dropped to less than 15%
of normal flov (below the " clip"). Despite procedural cautions, various lov flow system configurationr could not be avoided. The new square root '
extractors had a clip at approximately 2-3% of system flow. In concert with procedure changes, this replacement has proven effective.
c Despite all the investigations into system operation and the extensive design changes descilbed above, by early 1988 it was becoming ovident that although the modifications had significantly improved performance of the delta-flov instrumentation both relative to htty experience and compared to other BVRs operating with similar instrumentation, they had not been successful in
Attachment 1 pY-CEI/NRR-1390 L page 5 of 26 solving the problem of flov oscillations that led to spurious isolations, nor was complete elimination of the conditions that led to spurious isolations likely to be achievable. The possibility for a Technical Specification change began to be considered, since at that point in time it was becoring apparent e that there was a fundamental discrepancy between the operation of the RVCU system and the basis for either the delta-flov setpoint or the timer. The most obvious choices involved increasing the delta-flov setpoint at which the timer is started, increasing the timer length, a combination of both, or elimination of the delta-flov isolation of the system. CEI therefore examined the delta-flov setpoint and the timer to determine whether one or both could be optimized to provide a better balance between leak detection and spurious trip avoidance concerns. Initial investigation focused on increasing the setpoint while leaving the timer unchanged. Subsequent research, calculations and consideration of all the factors involved led to the conclusion in March 1990 that the best solution was a timer length increase while leaving the setpoint unchanged, along with the addition of isolations from the temperature and delta-temperature instruments in the cold piping sections of the system.
The analysis that was performed on the setpoints and timer lengths is discussed in detail in the section of this letter entitled " History of the Technical Specification Change Considerations", below.
While vork van ongoing to develop the basis for a Technical Specification change, m ional isolations of the system continued in 1988. The first isolation was due to the poor throttling characteristics of the flow control valves that vere due to be replaced in the 1989 refueling outage. The second isolation (LER 88-013) occurred during an RVCU system startup following an earlier system shutdovn as a result of a scram. No root cause was directly -
attributable to this event at the time. Based on investigations into subsequent events this isolation was likely due to the same cause as those described in detail for LERs90-022 and 91-011 belov. The third isolation in 10S8 (LER 88-039) occurred due to a situation involving natural circulation of the RVCU system (no forced pucp flov), when an increase in feedvater system pressure caused a deadheading effect on the RVCU outlet flow path while inlet flov remained constant, and is most likely another example of the phenomenon described belov.
In 1989, two of the three events discussed in LER 89-025 occurred while the plant was being shutdown, with the RVCU system operating in the Reduced Feedvater Temperature Mode. The third isolation occurred during an attempt to restart the system following one of the isolations. Another isolation during an RVCU system startup vas documented in LER 89-031. All four of these events ,
were extensively investigated but no definitive root cause could be established at the time. Certain modem of RVCU operation such as system startup and Reduced Feedvater Temperature Mode seemed to naturally induce flow oscillations that could not always be dampened by operator actions before the 45 second timer expired and isolated the system. Again, the knowledge gained from subsequent investigations shovs that these four events vere all likely due to the same cause as described belov for LERs90-022 and 91-011, especially the third and fourth events, for which computer traces of the pertinent carameters vere obtained which are comparable to the later events.
Attachment 1 PY-CEI/NRR-1390 L Page 6 of 26 Since no root cause other than unavoidable system fluctuations was apparent at the time, the development of a Technical Specification change continued to be the proposed corrective action for these events, as discussed in the " History of the Technical Specification Change Considerations" section of this letter, below.
In 1990, LER 90-022 investigated two isolations that closely resembled previous isolations in 1988 and 1989. One occurred during a plant shutdown with RVCU operati.ig in the Reduced Feedvater Temperature mode, and the second occurred during the subsequent restart attempt of the system. Computer traces of both events vere obtained, and it was as a result of revieve on the traces from these events that the voiding phenomenon discussed in that LER and in the following discussion was first suspected. The LER committed to further engineering investigation into the delta-flow indications that vere being experienced. In 1991, LER 90-022 Rev. I and LER 91-011 documented the conclusions that voiding in the regenerative heat exchangers was the root cause of the isolations discussed in LERs88-013, 88-039,89-025 and 89-031.
This voiding phenomanon is discussed in detail in LER 91-011 and Revision 1 of LER 90-022. Several of the isolations have occurred during operation in Reduced Feedvater Temperature mode. Reduced Feedvater Temperature mode is utilized during reactor shutdovns and cooldovns, when feedvater temperature returning to the reactor vessel is colder than during normal operation (because the feedvater heaters are not in service when the main turbine is off-line). Since the return flov from RVCU to the reactor flows into the feedvater line (see Figure 1) the Reduced Feedvater Temperature mode of RVCU is used which bypasses flow around the shell side of the regenerative heat exchangers. This lineup returns flow to the feedvater system with minimal reheating in' order to minimize thermal stratification of the feedvater piping.
This mode of operation was developed as a result of our operating experience review of NRC Information Notice No. 84-87 " Piping Thermal Deflection Induced by-Stratified flow" and INPO SER 05-85 " Thermal Fatigue Cracks Due to Temperature Fluctuations In Vater Mixing Points".
The necessity of the Reduced Feedvater Temperature mode of operation introduced the potential for an unfo eseen RVCU system event. Vhile operating RWCU in the Reduced Feedvater Temperature 2 ode with the system flov bypassing the regenerative heat exchangers, water on the shell side of the regenerative heat axchangers is stagnant at reactor temo..ature and very close to saturation conditions. As plant cooldovn progresces, slight pressure transients can occur enabling this water in the heat exchangers to flash to steam as pressure is reduced bel,v saturation conditions. If such flashing occurs, vater is forced out of t;ie RVCU system, but Licensed Operators have no indication that the RVCU system he partially voided until the void collapses
- - sometime later. The collapse of this void creates room for additional vater l vithin the heat exchangers, which therefore results in a significant decrease in system outlet (return) flov since the void location is at a lover pressure than the feedvater system. During this time, system inlet flow remains constant or increases slightly. The differential flov signal that results vill start the delta-flov timer, and if the void is not filled in time, vill 3
result in a system isolation.
I
Attachment 1 pY-CE!/NRR-1390 L Page 7 of 26 For the isolations that have occutred during system startups, a similar voiding mechanism within the heat exchangets can occur. Since the three vertical RVCU heat exchangets cteate a high point in the system piping and are in an area which has very high radiation dose rates (which precludes a typical high point vent procedure), it is very difficult to alvays ensure that voiding in the heat exchangers has been eliminated prior to starting up the RVCU pump (s). Vhen the teactor vessel is pressurited, merely opening the pump suction valves can also cause an intush of va:er sufficient to start the timer and isolate the system. For isolations that occur during both system startups and vhen in the Reduced Feedvater Temperature mode, the computer traces are similar flov is maintained into the system but does not exit the system.
As noted in the two most recent LER submittals dated Hay 16, 1991, an engineering evaluation was undervey to considet several items. Some of these included issues such as review of the design basis for the Leak Detection differential flow timer, feasibility of design changes to ptevent flashing, alternatives and operational ramifications 11 a solution to preclude flashing is not feasible, and a safety evaluation for a Trchnical Specification change request to permit longer flow ttansients, if the condition is tolerable and if the evaluation determined such a Technical Specification change is necessary.
This engineering evaluation has been completed, and although efforts are still ongoing to develop design or procedural alternatives to minimize the occurrences of these system disturbances, Engineering and Operations have concluded that complete elimination of these types of system perturbations is not likely to be feasible. The Engineering evaluation also shoved that the piping system (including the shell of the heat exchangers) is capable of withstanding the occasional (several per year) challenges to the system from the pertubations that may continue to occur, and that leaks or breaks of the piping system due to thermal stresses or water hammer are not of concern.
Finally, it shoved that the proposed change to extend the timer length vill have no impact on the occurrence of ' t initial flashing / voiding portion of the perturbation (because this does i i even create a signal to activate the delta-flow circuit, since more vater s ;aaving the system by way of Pe system), and that the extension of acceptableflowpathsthanisentering'hactupontherefillportionofthe the timer length vill have a positive in perturbation since it vill avoid the unnecessary isolation valve closure and shutdown of the system. It vill also provide other benefits as discussed more fully belov in the "lienefits That Vill Result From This Change" section of this letter.
l i
Attachment 1 pY-Cp.1/F<R-1390 h page 8 26 history _of the Technical Specitication_ Change Considerations A summaty of the various optlons consideted ior a Technical Spectfication change over the past sevetal years is ptovided because some of the investigations conducted are ditectly applicable to the change being proposed by this letter. Also, it is important to undetstand that othwt options have been considered, and that the ptoposed changes ate felt to be the approptiate Technical Specification revision.
By early 1988, the design modifications to cortect the known contributors to delta-flov isolations had been implemented with the exception of the flow control valves that vete to be chanFed out at the first tefueling outage. The operators vete still experiencing significant flov oscillations in the RVCU system when changing operational mode lineups, and especially during system startups. These led to continuing occasional it olations when opetator actions to shutdown the system or othervise dampen the perturbations vete unsuccessful within the 45-second timer period. Consideration began for a possible Technical Specification change. Initial investigations vete made into the basis for the setpoint and the timer length.
One option that was first examined was a substantial increase in both the setpoint and the timer (ro additions of '.emperatute or delta-temperature isolations vere considered). This option was rejected, ptimarily bec.use although the amount of reactor coolant that vould escape from a gulMotine break of the RVCU system vas significantly less than that fret a reetreulation line break accident or a main steam line break accident, the leakage that vould escape before the timer isolated the system vas still much greater than with the current nrrangement. It was also expected that calculations vould ptove that va';,r escaping itom a guillotine brenk in the " cold" portion of the RVCU piping would not in fact temain cold, since flov tates through the heat exchangers would be great enough that the heat exchangers vould be ineffective in cooling the water. This vould tesult in the water finshing to steam with two direct consequences. The first postriated consequence vas containment pressurization, perhaps in excess of the containment design pressure, depending on the length of the timer increase. The second was the possibility for early offsite dose consequences, even though these vete expected to be significantly less than for other design-basis loss-of-coolant accident scenarios.
'lherefore, attention was turned to a setpoint increase while maintaining the current timer length. The existing Nominal Trip Setpoint for the RVCU delta flow isolation is 68 gpm vith a Technical Specification Allovable Value of 77.1 gpm. These values included allowances for instrument loop accuracy and drift as appropriate, and correspond to an Analytical Limit of 83 gpm. The Analytical Limit was established by General Electric. This Analytical Limit simply corresponds to 20% of nominal RVCU system flov and was conservatively established by GE on the basis of engineeting judgment, with the intent of assuring that any offsite dose effects associated with this amount of cold leakage vould be acceptable. At the time this Analytical Limit was established, detailed plant-specific offsite dose analyses vere not perfotmed to determine a plant-specific Analytical Limit for this isolation setpoint.
i
l Attachment 1 pY-CE1/NRR-1390 L page 9 of 26 Since the existing Analytical Limit of 83 gpm is not based on a plant-specific evaluation it was determined that petformante of such a dose analysis vould provide a firm basis for a new setpoint, which could be set at a value higher than 83 gpm. Prior to performing the dose analyses. CEI pettormed analyses to determine the etfeet of leal: age flov f rom a bteak on the ability of the heat exchangers to continue to cool the votet, so that the plant-specific flov rate that could lead to steam release could be identified. The purpose of this analysis was to ensure that leakage flow rates from the " cold" portion of the system that vere less than the new setpoint vould temain cold, so that they vould not result in any significant early dose effects. Tovard this end, analyses vere performed to determine what the steady-state leakage temperature vould be for various leakage rates from a crack. The result of this analysis is graphically displayed in Figure (4). As can be seen, for leak rates as high as 200 gallons per minute (gpm) the expected leak temperatures vould remain 3 165'r (extension of this Leak Rate / Leakage Temperatute curve up to 212'r was not considered necessary, since the resulting permissible leakage rates vould be substantially higher than the value being contemplated as necessary to avoid spurious trips of the RVCU syst6m). CE1 con. rvatively determined that the leakage temperature limit vould be set at lbO'P (this vould provide a large 62'F margin to the 212'F boiling point). From the analysis this corresponded to a leakage value of approximately 152 gpm. This value was then used as the Analytical Limit. This number was then adjusted dovnvard for instrument end loop inaccuraclea and drift in order to determine an Allovable Value and Trip Setpoint.
Two safety implications vere considered for these nev proposed values. First, the effects of these new values on previous analyses for a complete guillotine break in any portion of the RVCU line vere studied. It was determined that the previous analyses still enveloped this accident. Since the timer setting was not being adjusted, the guillotine break in the cold section vould be isolated after 45 seconds whether the Analytical Limit was 83 gpm or 152 gpm as proposed. Since there are two independent trip logic circuits, each with its own time delay relay, the failure of one relay would not prevent the system from isolating. Also, there is a RpV level 2 isolation for the RVCU system which is the isolation used in accident analysis for pipe breaks. A break in the hot portior of the pipe vould still have been isolated by the temperature / delta-temperature sensors, which would not have been affected by the change request.
The other safety implication reviewed was the eifeet of permitting a RVCU leak just under the trip setpoint to exist indefinitely. It was noted that a leak of 152 gpm is insignificant in comparison to the ability of the plant to maintain Reactor Pressure Vessel (RpV) level in the normal operating band.
Normal Feedvater system flov at 100% reactor pover is in excess of 32,000 gpm, so this leakage rate is less than one half of one percent (<0.5%) of total feedvater flows this leakage remains well within the capacity of the feedvater system to makeup. Leakage from the cold vater portions of the RVCU system would be directed to floor sumps which would give the Operators indication of an increase in leakage.
I l
I Attachment 1 pY-CEl/NRR-1390 L Page 10 of 26 It was also noted that leckage in the cold portions t.t the RVCU system is not Reactor Coolant Pressurt Boundary leakage. By the Technical Specification definition. Pressure Boundary Leakage is leakage through a non-isolable fault in a reactot coolant system component body, pipe vall or vessel vail. Since the cold portion of the RVCU system is all isolable from the reactor vessel, a leak in the cold portion of the RVCU system does not meet the definition of Reactor Coolant Pressure Boundary beakage. Since none of the cold portion is inside the dryvell, this leakage is also not part of dryvell unidentified leakage.
As part of the evaluation of the effect of permitting a RVCU leak just under the trip setpoint to exist indefinitely, this leakage was evaluated from a radiological standpoint. As discussed above, leakage from the system belov the propose:' setpoint was detett.ined to temain cold, and since small leakages such as being evaluated here vould be fully compensated for by the feedvater system, there vould be no fuel damage, or even an iodine spike because there vould be no need to shut down the plant in order to isolate the leak. Since all of the cold portion of the RVCU system monitored by the delta-flow instrumentation is inside containment, this prel'.eintry analysis shoved that the radiological aspects of this size leak, even if t existed for substantial time periods, resulted in lov offsite dose consequences.
However, by August of 1989, it became clear that the setpoint change being considered would not be sufficient to preclude the delta-flow circuitry from isolating the RVCU system during the flov distutbances being exp*11enced.
Computer traces of pertinent parameters were obtained following sne events documerted in LER 89-025 and following several near-isolations during attempted system startups. The flow rates experienced urs higher than the proposed setpoint, and vere lasting loager than the current 45-second timer length.
One idea that was considered at this time but rejected as a long-term solution was to utilize the Leak Detection System bypass switch during periods when RVCU is susceptible to the flov oscillations that vere being experienced. The Technical Specifications for these Isolation Actuation signals contain an Action Statement that provides a one (1) hour period during which the instruments may be considered inoperable without the need to place the instruments in the tripped condition (Technical Specification 3.3.2, Actions b and c). Although use of this one hour period is alloved, it was deemed a non-conservativt approach to solving the problem of isolations, since it bypassed all the temperature and delta-temperature isolations at the same time as @ delta-flow insttuments. Also, it would be difficult to avoid the isolations that occur during Reduced Feedvater Temperature mode, since they can occur at any time during these conditions, vith no prior varning to the operators (other than the alarm when the timer starts). This same censideration made the concept of a design change to provide a separate bypass svi ch dedicated just to the delta-flow isolation impractical as a solution.
1
4 .
Attachment 1
, PY-CEl/NRR-1390 L page 11 of 26 ,
Reconsideration vas therefore given to the concept of both a setpoint increase and a timer length increase, but this time they would be implemented in combination with the addition of nev isolations from redundant ambient temperature sensors and another diverse isolation from redundant delta-temperature sensors in the " cold" portions of the piping. The addition of these redundant and diverse temperature / delta-temperature isolations vould address the previously expressed concern of a larger break in the system that vould release hot water and steam, and they vould preclude adverse effects such as excessive containment pressurization or early offsite dose effects.
The proposed increase for the delta-flow instrumentation actuation was revised dovnvard slightly from an Analytical Limit of 152 gpm to an Analytical Limit of 150 gpm. Since the temperature-based isolations vould address the possible hot leaks from the system, the delta-flov isolation could be returned to its original design intent (as expressed in the GE design specification) of i detecting cold leaks in the system. The 'imer length could threfore be extended significantly with only minor offsite dose consequences, even using conservative dose analysis assumptions, because the leaks to be detected by thedeltaflovinstrumegtsvereagainonlycoldleaks(lessthan150gpm, therefore less than 149 F vater as determined from the Figure 4 curve).
Since the source of the flow perturbations was still considered to be
- expec;teda rystem fluctuations, and since very long timer lengths could be secommodated bssed on the lack of radiological consequences, it was decided to solect a timer length that was certain to allow the flow perturbations enough time to.rettle out, yet was less than half (1/2) of the time that was permitted by the Technical Specification for the instruments to be inoperable (in bypass). A-timer setpoint of 29 minutes was there1 ore selected, with an allovable value of 29.85 minutes, and an analytical limit of 30 minutes. The setpoints for the new ambient temperature isolation vero chosen to be 137.9'F vith an allovable valus cf 1 8 7'P and an analytical limit of 149'F. The volaes for the delta-temperature setpoint vould be 82.6'!' (delta) with an allovable value of 85.1'F (delta), and an analytical limit of 89'F (delta).
In Februory 1990, it was decided that there was no real need to increase the delta-flov setpoint to 150 gpm, since the timer length increase vould allow flow perturbations even as small as the current setpoint of 68 gpm to settle out. It was determined to be more appropriate to leave the setpoint as-is, so that the operator would still get the alarm and the timer start if flow perturbations reached 68 gpm. This vould enable him to check alternate indications such as containment sumps and area radiation monitors in the same way as is currently performed in response to this signal, in order to verify
=there is no real leak from the system that should be manually isolated prior e to expiration of the timer and the tesultant automatic isolation.
During-the ongoing deliberations over the format of a Technical specification change, isolations of the system continued and computer traces of the events began-to shov a repeated similarity. In the response to LER 90-022 in the fourth quarter of 1990, it was determined that engineering evaluations should continue in order to determine the root cause of the isolations and to evaluate possible changes to the system design or operating procedures. Vork on the Technical Specification change vas temporarily set-aside pending the outcome of these investigations. Vith the occurrence of another identical ;
Attachment 1 PY-CEI/NRR-1390 L Page 12 of 26 isolation during Reduced Feedvater Temperature mode of operation in April, 1991 (despite a procedural change that had been implemented in an attempt to avoid the problem), eftorts to support a Technical Specification change vere renewed.
Now that the root cause had been identified as voiding of the regenerative i heat exchangers (which permitted flov into the system to remain high while flov nut of the system was significantly reduced), all that remained to i evaluate was whether the proposed increase in the timer length could result in ,
- an adverse effect on the system integrity such that a break could be induced that vould not have occurred if the current timer length was maintained. The engineering analysis of the system (described more fully in the " Safety Analysis for Proposed Changes" section of this letter, belov) determined that 3 this was not a concern. This engineering analysis also resulted in an even greater understanding of the phenomenon being experienced. It-vas identified -
that even if the shell side of all three of the tegenerative heat exchangers l vas voided, the lovest system flow permitted by the system operatAng instruction'(70 gpm) would refill the voided volueo vithin approximate'ly eight minutes. After this time period, system flows should stabilize fairly quickly. It was therefore determined that the proposed timer le9gth could ba decreased from a setpoint of 29 minutes to a setpoint of 10 minutes. The delta-flov alarm setpoint vould remain at 68 gpm. The proposed temperature and delta-temperature isolations would still be added and-vould utilize the same setpoints as previously discussed in order to detect and isolate eny leaks that might release hot vater / steam. These are the changes proposed in 4
Attachment 2 of the letter.
In summary, the proposed changes to the RVCU isolation logic have been extensively considered. and vere developed as e ensult of multiple system isolations.
Proposed ChenEe5 The propored changes includes i
1 Addition of new RVCU system isolations from redundant and diverse temperature and delta-temperature sensors located in the Containment Valve Room, Demineralizer Rooms and Receiving Tank Room voere the " cold" !
portion of the RVCU piping is routed. The ambient temperature isolation Setpoint and Allovable Value in all these rooms vi? be 137,9'F and 143.7'f respectively. The delta-temperature isolation Setpoint and Allovable value-in all these rooms will be o2.6'F and 85.1'F, respectively. Alarm setpoints are also provided at lover valt.es. This information-is summarized in Table 1 of this letter. These Aliovable Values and setpoints.are based-on an Analytical Limit of 150 gpm (corresponding to a vater temperature of 149'F from the postulated i break). The Allovable Values and the Setpoints vere determined using the
! General Electric Instrument Setpoint Het.:odology described in NEDC-31336.
--c- ww-%,, $.-.... --y , , ~
7- .ws.w v. -.w,6w,.-.w y,egpey,.,g,, p yw,9.,,99, ,yi.i.,.-,.. igg ,y,%ww%,y.y+,-5 g, mywp,wg,%,.w,.-,,_.,c.w-%,,,em,j,,,,y., ,,,y,,9,.,,aw,,,,.,emmew e =.'mygi<-ww gp ire--4
Attachment 1 PY-CEI/NRR-1390 L Page 13 of 26
- 2. Extension of the RVCU delta-flov timer length. The timer setpoint and allovable value vill be 10.0 minutes ano 10.85 minutes respectively. The Allovable Value is based on an Analytical Limit of 11 minutes, with an adjustment for the .nanuf acturer's stated accuracy of the timer device.
The _etpoint of 10 minutes was chosen for human factors reasons, and balances a desire to eliminate spurious isolations of RVCU against a desire to minimire releases from an actual break. A significant margin is provided between the setpoint and the Analytical Limit, and therefore the Allovable Value also, to minimize the possibility of Technical Specification Violations. The choice of the Setpoint value was the controlling factor, since the Analytical Limit could be significantly higher with minimal safety significan a.
Safety Analysis Por Proposed Changes The assient temperature and delta-temperature instiuments in combination with the RPV Vater Level 2 isolation, protect against breaks of the RVCU system that release hot vater / steam, and the delta-flow instruments provide for detection and isolation of cold water leaks. This design intent vill still be met with the implementation of the proposed changen. The following discussion expands upon the safety significance cf the proposed changes.
The General Electric " Leak Detection System" Design Speelfication states thnt "The cleanup sys:em shall have a means of flow con,parison between the system inlet and the outlets. The alarm and isolation setpoints shall be established at a differential between inlet flow and outlet flov vhich equals 20% of system rated flow, A bypass timer shall be provided to override the isolation during system pump and valve surge conditions". The Leak Detection System Design Spec Data Sheet (DSDS) states that "The RVCU system does have a portion of its piping that containo cold reactor coolant and therefore temperature monitoring for leakage vould not be respansive. For this reason, a flow comparison between inlet and outlet flov'is monitored for this piping." The DSDS also identified that the timer setpoint should be 45 seconds. -Similar to the process for choosing the Analytical Limit for the delta-flov setpoint of 68 gpm (as discussed above .tn the Technical Specification change history),
this 45 second timer length was very conservatively chosen by CE baeed on engineering judgment, with the intent of ensuring, without t_he need for any plant-sp,ecific dose calculations, that any offsite dose effects associated vith this leakage vould be acceptable. It was not expected at the time of its establishment that the. conservatively short timer vould cauce the tremendous number of spurious. (non-leakage) isolatinns that have been experienced at the never plante at t'hleh this featore was installed.
The proposed change vill extend the length of the timer. General 1:lectric is in concurrence with this ptopored change, be. sed on the delta-flov safety function, and also concurn with the retention of the delta-flov alarm and isulation features and the addition o.f the ambient tesperatere and delta-temperature isolations into tne cold portions of the piping. The addition of the temperature and delta-temperature $solations to detect hot leaks effectisely returns the delta-flov timer to its original design intent of cold leak detection. The portion of the RVCU system that to monitored for
4 g
Attachment 1 PY-CEI/NRR-1390 L Page 14 of 26 leakege by the-delta-flov instrumt ,tation is a high energy system (due tu pressure rather than temperature in the cold portions), therefore guillotine-breake are required to be assumed, this monitored portion of the system is outboard of the second containment isolation valve and therefore can be isolated from the_ reactor coolant pressure boundary in the event of such a break. Credit is taken for dstection and isolation of guillotine breaks by the RPV Vater Level 2 signal in order to protect the reactor core, and irolations of the system for breaks large enough to leak hot water / steam and pressurize the containment are provided by the ambient tempercture and delto-temperature instrumants. For r. mall breaks in the cold portion of the system, the delti-flow instruments vill etill alarm and start the timer at the same setpoint as always, and vill-isolata the cold water leakage in a timely fashion vell prict to any a4.gnificant offsite dose consequences, even assuming no operator action is taken to isolate the system due to the leak prior to the expiration of the timer.
l In order to estimate the significance of a non-guillotine, cold water break that could potentially exist for the length of the proposed timer, an offsite dose evaluation was performed, and the resulta vere favorably compared to the Regulaciors and the Standard Review Plan criteria. The analysis assumed that the break occurred while the reactor was operating at full power, and that the break occ ared upstream of the filter-demineralizers, since downstream breaks
! vould reduce the dose by a factor of 100 due to iodine removal. The analysis I
assumed the maximum flow rate from a break which could escape detection by the temperature detectors by assuming a 150 gpm leak of water. Since this water vould be at approximate)y 149'F, the leakage is subcooled and the analysis j
therefora assumed tha- 3e iodine fraction that becomes airborne is 10% per ,
i the Standard Review P1 >
.ction 15.6.5 Appendix B Section III (July 1981).
The analysis also con:< 4tively assumed that the airborne fraction that contributes to the do'o is immediately transported outside of the containment l vith no attempts by t e operators to prevent its escape or to filter the leakage, even though we conta2nment purge system (in the unlikely event that
-it shotld happen.to be running at the time of the leak) has an-isolation from L
a radiation sensor and contains charcoal filters, and the Annulus Exhaust Gas l' Treatment System which processes containment leakage on a continuous basis during PNPP operation contains Engineered Safety Feature (ESP) filtration systems. Thecongervativeshort-term (accident)dilutionfactor(X/0)value
-of 4.3xlO~ sec/m from USAR Table 2.3-24 was utilized for the Exclusion Area Boundary-(EAB) calculations, and the Low Population Zone (LPZ) dose l- calculationsalsoutilizedtheconservatgveX/0valuesfromTable2.3-24.
3 If
( an ass. rage annual X/0 value of 2.7 x 10- sec/.a for the EAB case was used, this would reduce the dose rate by a factor of 159. It was assumed that the reac'.or coolant for a small break of this type vould not contain activity due to a reactor shutdown or due to any fuel failures because feedvater makeup
? lvould maintain normal water levela and no shutdown or scram is necessary in I order to isolate the leak. Therefore, the reactor coolant was essumed to L contain an iodine activity equal to that provided in USAR Table 13.1-3 (which t serves as the starting point from which the iodine activities for all the PNPP j USAR Chapter 15 accidents were determined), These values are based on
- realistic levels of .setivity found in BVR coolant during normal operation.
Noble gas releases from a cold reactor water break of this type vould be negligible, as recognized in NUREG-0016 (Rev. 1) " Calculation of Releases of Radioactive Materials In Gaseous and Liquid Effluents From Boiling Vater Reactors (BVR-GALE Code)," Table 2-4. The break vas assumed to be isolated after 11 minutes (the analytical limit for the timer).
- . 3 ' --
Attachment 1-
. PY-CEI/NRR-1390 L. ,
Page 15 of 26
-The-radiological'calcalations-showed that for cold leakage from the RVCU piping,=_vhich is presumed lov enough to not trip the temperature-based isolations, the resultant EAB-inhalation doses vould be 0.004% of 10CFR100 limits. Since iodine activity is the source of concern, A le body doses vould-be; negligible and were not calculated. -A lirting or various. inhalation doses'for EAB, LPZ and' Control Room and their corresponding ace.eptance criteria are presented in' Table 2. As can be seen, the resultant doses an a
-very small fraction of the 10CFR100 acceptance criteria.- The resultsnt radiological effects-are bounded by both the Dt: sign-basis and the Realistic ;
-analyses for the Main Steam Line Break-Outside Containment event (see USAR Tables 15.6-8 and 15.6-11). Inhalation doses for-the RVCU case-that assumes no filter-demineralization occurring prior to the break location produces >
doses more than a factor:of 100 belov the Realistic analysis for the Steam Line_ Break event.- l Larger breaks than the postulated 150 gpm leak would release hotter water or steam, and would therefore be isolated by the ambientitemperature and/or
-delta-temperature = instruments (or in the vors' case by-an RPV Level 2 signal). !
. As discussed above, the allovable value and setpoint _ for. the-ambient temperature elements vew based on= detecting breaks greater than 150 gpm (149'F Analytical t Limit) _and-the delta-temperature analytical limit of 89't
-vas also rbased on 149?F, and on the normal supply air temperature of 60'F.
The_setpoints associated with these Analytical Limits vill clearly serve to isolate hot water breaks (above 212*F).similar to the comparable existing temperature:isolations in the RVCU Rooms. The doses from hot water / steam-breaks of the RVCU system have:always.been considered-to be bounded by the radiological calculations for the Main-Steam Line Break Outside Containment.
-Therefore,_-no additional dose analyses need be performed for the addition of
_the new temperature-based isolations. The containment pressurizatiots analysis i in USAR Section 6.2.1.1.4.2 vould_also remain valid-and bounding for large hot Lbreaks,Ldue'to the isolations~from the ambient-tenperature and delta-temperature = instruments. ,
Another-consideration-vith respect to a postulated leak-or break in this
. piping is flooding. These areas are designed to accept-flooding, since they contain_high-energy piping which must be assumed to break ?he flooding analysis for= breaks inside the_ Reactor Building is discus: o. i_n USAR section -
3.6.2.3.5.1. No larger amounts of flooding are 'ntroducod by this change, since the mass of reactor coolant released froin a guillotine break.before completion of-its isolation is greater than the amount of-coolant released-
.from a:small. break that might be:able_to continus for up to 11 minutes without an: earlier-isolation from the temperature-based instruments. It should be noted here that the containment floor' drain sump indications provide a confirmatory method for-plant operators to determine exictence of actual l leakage, and that existing plant' instructions provide guidance to operators on appropriate actionszin response to a leak or break. In the event of an actual leak. that activates the timer, it is therefore extremely unlikely that- the leak woul.d continue, unisolated, for the duration of the timer.
.,,t ,
Attachment 1
. PY-CEl/NRR-1390 L Page 16 of 26 Equipment qualification in the RVCU Containment Rooms is unaffected by this change, because these rooms are already categorized as Harsh Environment areat.
for the high energy guillotine line break, which envelopes the conditions introduced from a small leak of cold water for 10 minutes.
The above discussions dealt with breaks that must be postulated to occc.r in order to meet regulatory guidelines, with no regard to the actual probability
-of their occurrence. In considering the acceptability of requesting the proposed Technical Specification changes, the possibility that the proposed changes could induce a break that might not othervise have occurred was investigated.- This was done to answer the 10CFR50.92 question as to whether the proposed amendment vould involve a significant increase in the probability of occurrence of an accident (such as a leak or break in the RVCU system).
This particular evaluation centered on the heat exchanger voiding phenomenon which has been the cause of the delta-flov isolations since mid-1988. The investigation examined whether extending the timer length could cause a voiding event _to occur, and whether extending the tirer length could aggravate voiding events that have already been initiated, it also looked at the
. voiding events from a structural integrity viewpoint to determine the capability uf 'he system to withstand thermal transients and to consider the potential for water hammers in the piping. The results of these evaluations strongly supported proceeding with the proposed Technical Specification change to extend the timer length.
With respect to the first question as to whether increasing the timer length could cause a voiding event to occur, the delta-flov leak detection logic should be briefly reviewed. As stated in the " Leak Detection Logic
-Discussion" section of this letter, the delta-flow logic compares flov into the RVCU system against flov out of-the system, and annunciates an alarm and starts the timer when outlet flov is less than inlet flov. This could indicate a leak in the cold piping (a leak small enough that it cannot be
( detected'by the ambient temperature or delta-temperature instruments). Since the timer is only started when outlet-flow is less than inlet flov, it does not activate when a voiding event begins in the shell side of the regenerative l' heat exchanger, since the voiding forces more flov out of the downstream portion of the system. In fact, the computer traces show that a voided condition can continue for fairly long periods-(20 to 35 minutes, possibly l- longer) depending on the pressure fluctuations in the regenerative heat l
exchangers (which are at-or near saturation conditions). It is only when the steam filled space cools and condenses that the RL'CU system flov into the system begins to exceed the outflov, thus starting the timer. Therefore, the length of the timer, vhether 45 seconds or 10 minutes, makes no difference as
! to the creation of the void itself.
Second, the evaluction considered whether extending the timer length could aggravate voiding events that have already been initiated. It is obvious from the isolations that have occurred that the period of time during refill of the
__ voids is typically longer than 45 seconds, which currently results in isolation of the system. The timer extension vould avoid this isolation and allow the refill to be completed. This has one immediate advantage in that i+
L
- c .
Attachment 1
, PY-CEI/NRR-1390 L Page 17 of 26 reduces the number of challenges to the isolation logic and the isolation valves. This also reduces the challenges to the pumps (since they trip on the lov system flows that occur) and to the system itself (since the valve isolations place a load on the piping). Once the voided volume begins to rsfill, the timer vill start. The evaluation concluded that the thermal transient on the shell of the heat exchanger as a result of the refill has been primarily experienced during the first 45 seconds. An increase in the ;
timer length would not change this portion of the thermal transient. The effect of these thermal transients was therefore examined to ensure that their occurrence, whether limited to 45 seconds or not, was acceptable from a cyclic service standpoint. It was concluded that for the number of events that might be postulated to occur over the life of the heat exchanger, that the temperature step that could be accepted (vithout even the need to perform an analysis for cyclic service) was greater than the maximum possible temperature <
difference between the shell temperature and the RVCU refill water temperature. An analysis vould only be required if the limit on the temperature difference given by NB-3222.4(d)(4) of ASKE Code,Section III had not been met. These thermal transients on the shell and the interconnected inlet s7d outlet piping are therefore not a source of concern with respect to induc.ng a' leak or break of the RVCU boundary. A thermal transient analysis vas also performed on the junction point where RUCU joins with feedvater, to address the brief temperature transient that occurs when_the hot vater within the shell side of the regenerative heat exchangers is flushed downstream by the expansion of stetm within the shell. This also proved acceptable, as compared to the acceptance limit of NB-3600 of the ASME Code.Section III.
Finally, the possibility of a vater nammer occurrence due to the voiding phenomenon and the associated system flow diuturbances was considered.
Although this issue is not connected to the length of the timer, the results are provided here for completeness. There are several points in time during a voiding event or system startup when system flows change, however these do not present a vater hammer concern because the bypass valve eround the regenerative heat exchanger is open in all situations to mitigate any pressure rise in the piping system. Vater hammer events generally result from an abrupt blockage of flow in the pipeline, which is not the case in these instances. This is supported by the computer traces, which do not show any sharp pressure rises. It is also supported by operating experience on the RVCU system, since no vater hammer incidents have actually been experienced (vater hammer events vould leave evidence of their occurrence through damaged-supports, grouting etc.). Industry-wide, vater hammer events in the RVCU system have also not been a problem, aa documented in AEOD Report #AE0D/
E91-01 "A Review of Vater Hammer Events After 1985" and NUREG 0927 Rev. 1
" Evaluation of Vater Hammer Occurrence in Nuclear Power Plants, Technical Findings Relevant to_ Unresolved Safety Issue A-1."
- The engineering analysis concluded that the increase in the time delay is acceptable, and would delay and most likely prevent spurious (non-leakage) isolations of the RVCU system, with attendant benefits such as those detailed below.
i r
.. N I ,
l Attachment 1 PY-CEI/NRR-1390 L Page 18 of 26 Benefits That Vill Result From This Change First, there vill be a significant reduction in the number of challenges to the Reactor Vater Cleanup System isolation valves and the pumps. Since plant li. censing, there have been 39 automatic isolations of the system due_to the delta-flov instruments, none of which vere due to actual leaks in the system.
In addition to the automatic isolations, there have been many additional challenges due to tripping the pumps or closing the valves prior to the expiration of the 45 second timer. This leads to unnecessary vear and degradation of the pumps and valves.
Second, the change vill result in increased avr.ilability of the cleanup capabilities of the system. This system is important to reactor operations because it is the sole method for removing the impurities within the vessel.
An extended loss of the system during power operations could lead to a plant shutdown due to vater chemistry exceeding acceptance limits. More of a concern are the losses of the system during plant startup and shutdown. The t are the critical periods for cleanup system operations, for severcl reasons.
One is that conditions are most favorable for stress corrosion cracking due to chlorides during these periods since temperatures are as high as during normal operation and oxygen levels are higher than normal operation. Operation of RVCU during;these periods is critical to remove impurities from the water to minimize corrosion effects. Conditions are also experienced during plant shutdowns that result in a release of Cobalt-60. activated corrosion products into the reactor coolant. -This Cobalt-60 eventually plates out on piping throughout the plant, leading to higher dose rates to plant workers during subsequent outage work in nearby areas. A BVR-6-group which examined methods to reduce the radiation source term from this Cobalt-60 release determined that the ability of the RVCU system to quickly reduce the peak Cobalt-60 activity levels should make the availability and effectiveness of the RVCU system the top priority item during shutdown. Operation of RVCU also serves to reduce the amount of Cobalt-60 released from fuel surfaces by providing a heat removal source for the reactor other than boiling.
Isolations of the RVCU system have other negative operational impacts as well.
Each isolation forces plant operators to respond even when there is no actual ,
leakage from the system. This distracts their attention away from other critical tasks that they perform. This is especially true during plant startups and shutdowns when the system has proven itself to be most troublesome. It also results in increased personnel exposure due to the performance of system valkde'ns folloviag the delta-flov isolation. During reactor heatups and when maintaining hot standby, the RVCU system is the only means of controlling reactor water level if reactor temperature is below the temperature at which steam can be produced. This can cause difficulties during temporary losses of the system such as those due to spurious isolations.
c Attachment 1
, PY-CEI/NRR-1390 L Page 19 of 26 SIGNIFICANT HAZARDS CONSIDERATION The standards used to arrive at a determination that a request for amendment involves no significant hazards considerations are included in the Commission's Regulations, 10CFR50.92, which state that the operation of the facility in accordance with the proposed amendment vould not (1) involve a significant increase in the probability or consequences of an accident previously evaluated, (2) create the possibility of a new or different kind of accident from any previously evaluated, or (3) involve a significant reduction in a margin of safety.
The proposed amendment has been reviewed with respect to these three factors and it has been determined that the proposed changes do not involve a significant hazard becausts
- 1. The proposed change does not involve a significant increase in the probability or consequences of an accident previously evaluated.
The proposed changes deal only with leak-detection systems that respond to a leak or break that has already been postulated to occur. The addition of new temperature-based isolation signals or the extension of the delta-flow timer length vill have no impact on the probability of occurrence of the postulated break, since they do not affect the integcity of the piping boundary.
The RVCU delta-flow isolation is not used in any of the USAR Chapter 6 or Chapter 15 Accident Analyses. The most severe RVCU line break would be a double-ended shear of a portion of the line. For the vorst-case Containment Pressurization analysis, RVCU is isolated by the redundant ambient temperature or the redundant delta-temperature instruments, which vill be added to all the RVCU rooms in the Containment as a result of this change request.- For the vorst-case RVCU Loss of Coolant analysis, in the event that normal feedvater makeup to the vessel is lost or cannot sustain reactor water level such that RPV Level 2 is reached, fuel damage vill be precluded by the resultant isolation of RVCU. Changing the delta-flow timer length and adding new temperature-based isolations has not changed this bounding accident. In addition,_ flooding effects and equipment qualification requirements remain unchanged and bounded by the postulated guillotine break analysis. Dose consequences also remain bounded, by the Main Steam Line Break Outside Of Containment analysis. Therefore the consequences of previousls analyzed accidents has also not changed.
- 2. The proposed change-does not create the possibility of a new or different kind of accident.
The leak detection system provides a monitoring fut.ction only. As discussed above, the simple addition of RVCU leak detection devices or the extension of time delays on leak detection devices does not affect the integrity of the piping boundary. It also does not interf ace vu.a sys tems other than RUCU.
.a Attachment 1
. PY-CEI/NRR-1390 L Page 20 of 26 Loss of Coolant-Accidents (LOCA) both small and large, have been previously analyzed in the USAR. These analyses have included leak sizes up to and including the Recirculation line break. They have also examined breaks in the RVCU system, up to and including guillotine breaks et the
-system. These RVCU breaks have already been postulated inside the
. containment, where these RVCU rooms are located. Any amount of leakage
~
resulting from a leak or break in the RVCU system is still bounded by those exiating analyses, therefore no new or different kind of accident has been created by the proposed change.
- 3. The proposed change does not involve a significant reduction in the margin of safety.
As discussed above, the RVCU Delta Flov isolation trip signal is not used in any of the accident analyses since the isolation is assumed to be provided by the RPV Level 2 isolation or the temperature-based isolations.
This proposed change does_not result in the RVCU Delta flow trip becoming the bounding _ isolation-signe.l. Therefore the margin of safety has not been reduced by the proposed change.
Environmental Consideration The proposed Technical Specification change request has been reviewed against the criteria of 10 CFR 51.22 for environmental considerations. As shown above, the proposed _ change does not involve a significant hazards consideration, nor increase the types and amounts of effluents that may be releasedToffsite, r.or significantly increase individual or cumulative occupational _ radiation exposures. Based on the foregoing, it has been concluded-that the proposed Technical Specification change meets the criteria given in 10 CFR 51.22(c)(9) for a categorical exclusion from the requirement for an Environmental Impact Statement.
= .
~
_ _ _ _ -_ _ ~ , . . _. . . _ . _ _ _ _
9
,, , c ' .
Attachment 1
, FY-CE1/NRR-1390 L Page 21 of 26 Table 1
-Temperature and Delta-Temperature lustruments Isolation Temperature Alarm Allovable Existing: Element Seinoint Value
- 1. RVCU Heat Exchanger Room a) DT F31NO36A/B-in 3 57'F f 78.4'F Cot.tainment E31NO35A/B-out Elevation Above 652'2" b) T E31N034A/B $ 117'F .f138.9'F
- 4. RVCU Pump Room #1 a) DT E?lN038A/B-in f 23'F $ 30.4'F Auxiliary Building E31NO39A/B-out Elevation 599' b) T E31NO37A/B f 130'F $ 137.9'F
- 3. RVCU Punp Room #2 a) DT E31N041A/B-in f 23'F f 30.4*F Auxiliary Building E31N042A/B-out Elevation 599' b) T E31N040A/B f 130'F f 137.9'F
- 4. RVCU Valve Nest Area a) DT E31N044A/B-in f 23'F $ 30.4'F Auxiliary Building E31N045A/B-out Elevation 599' b) T E31N043A/B f 130'F 3 137.9'F
, News
- 1. -Demin. Room #1 - a) DT E31N048A/B-in ~
< 57'T ~< 85.1'F Containment E31N047A/B-out Blevation 664'7"' b) T E31N046A/B f 117'F f 143.7'F
- 2. Demin Room-# 2 a) DT E31N051A/B-in -< 57'F -< 85.1'F Contcinment E31N050A/B-out Elevation 664'7" b) T E31N049A/B f 117'F f 143.7'F
- 3. Demin. Valve Room. a) DT E31N054A/B-in f-57'F f 85.1*F Containment -E31N053A/B-out Elevation above 652'2" b) T E31N052A/B f 117'F f 143.7'F
- 4. Demin. Receiving a) DT E31N057A/B-in f 57'F f 85.1*F Tank Room E31N056A/B-out Concainment b) T E31N0554/B f 117'F f 143.7'F Elevation 642' ,
.. o .
Attachment 1 PY-CEI/NRR-1390 L Page 22 of 26 Table 2 Reactor Vater Cleanup System Loss of Coolant Radiological Evaluation Results Inhalation Dose (Rem)
Vith Filter- Vithout Filter- Acceptance Demineraliza_ tion Demineralization Criteria (Rem)
Exclusion Area 1.34 E-4 1.34 E-2 3.0 E+2 (10CFR100)
(863 meters)
Lov Population Zone 1.5 E-5 1.5 E-3 3.0 E+2 (10CFR100)
(4002 meters)
Control Roota 1.1 E-3 1.1 E-1 3.0 E+1 (SRP 6.4.11.6)
LCS/ CODED /4952 t
_ _ . . . _ _ _ _ - - _ _ ~ . . _. _ _ _ _ _ _ _ .
3 PY-CEI/NRR-1390 L j . i jj Attachment I
=' f n *R %q - Page 23 of 26
~~l 4 4-g
-g ; .*-o l21m: % focl=, W Figure 1
- r f ,.1-
-v a-1. - 'r. - - -- ; /: . ., a 3,! l al: 213 1 i v2 /r-s lc3l gigs i s y y lii!!i i! ~~
h'
- W,
- : C .
- h:Q i._ _.s
%q A) dh
.. .. . . r ..
n ..
A- :
u Il i-I. -!$
8! Ili-3
- i, x-
{fIdr,_. ,
p 3 (;
a -
g HI i ._E n!! :!!E 6 j
a
.D.
j
-:6 R, h
== . g.g. ( n c a m ej$1 .
. 4 g
2 i
M "
m a.
.o m
m-
< s. .a. - - - ,
s z
t s
=a
'Y S
--h.c IIi l h
h 'i.s:(bi[ _ 'l, d 5
..i . _ . i Fi ,fie.'
e t .
- - < i , .- : , m
.:3 W
, _J ~ g $ I
- H._g I
5 p25 a*<
-s1l :.. t ,Ti-IX.,s,.
I n l!
W W 9 '
t ,
3 (
s .A.
gl.2 m
.. m
- g . t .p 3 o
I f -1 ,,
l t ,
sg .. W
~
M
- f s Sg 3
u!Ty Jg A .
n g =
a ,
'- . 6, a a s k' . ,
p.- - . - - ..
q
---~~--q
, i
@E i iin . . !.
l.
-1 i
,; 9 '
1 a
8 1 -r e..~l-l 9- ? il j I I l
-r 1
,r l Ii !
$l e Vi
,i a i =
^g l 1!-Z IL!.-E s!l -
i -
i ' 5I
, l I!-5, 3 4y I.I -
L.........----]
y y , --
j a-.=-. ---w, ,,y,---..,.--,,---y- 4 pe ,y,,,,-*wi,,- .,,,.,-n-. ,7.-,g g -w, rp , - e e
- . A, . -.-- ..
~PY-CEI/NRR-1390.L b - g,_ ~ _ ..;- Attachment l
- . Page 24 of 26 Y s-i *- Figure 2._-
+-
?<
t: e l,.'
e o, .
74 , en r ese .d coasta,
' *' ~'
REACTOR A
is = 1so r
,1 i .,
' metod weste6 wwr.
Teseecature Resenerettwo Reeenerative and Pevssue, 72 a 271 r T3 e 120 r fa a $40 F- doet de ot P14 1040 pela -
Emchensee- te m e e1 a 408 CPM 4% &
PS a 100 pe&a 45 a 533 CPM ,
Cle ed coetane
. nset oe In -
73 a 120 F
. Filtee A +
N- .
Deenkrierellace Leah V
es foote t . f eapeceturgs shown arv Anatial 'steessw state values.
- p e, . . -, .. c i . _ . .e t e.,
I
-+ ., ,..-~,--% ,. o u m. s , m.~,., .,+.3. -. ,- , ., - , - - , , . -=v ,,ew. - . -u % a rP-
?Y-CEI/NRR-1390 1.
Attachment l l 4** ,
Page 25 of 26
(
Figure 3 4- .
(3e 5' ,
RWCU 3rma , EWCU 3rma RWCU 3rsum l tr:4s now anc== now me-dowe Te Reactor now Canen
- C13 TT N040 C33 TT, N604 C33 TE-Nn33
---~~'-~" ' ' ' ' ' " ' '
F- Y Y ~~'-' ~ ~--'~'
Y l
- I -
uco
(
' ~ I woo C7M - e g
,-t-s
- [ _
j_ lr CFM
- I ,
g ca nser i ; c o nso:
I I I I I
I I
I I l Reactor I . Pres sure l -
n** Samme ! B21-N678
[
( I Z h
t I o-im - -
1 Z*
I cm I I , I g G33Ht4144 1 1 4s s 4 EWCU I**' L***
Umse E
l g f -
De new al sama I
1 swcu m l De new m Mamma :
g i
I -
. t 1 .
I g I coman w t. l 1 cam m . i 1 w% g
. I. I I Is1.
I L _ _ ___ _ _ _ _ ___ _ _tk o.uenas srwar____.___.-__!
RWCU Flow Instrumen:stion .
i Ed NDM l
PY-CE1/NRR-1390 L Attachment I 4.; + Page 26 of 26
' Figure 4 I ' ' t I q , 4 I I 4 6 3 i 4 m s@
4 M
- .3 .
..g e
M P
r m
M u '
) a y
' ~ .
- =
g )
be
- Gfl
~_
C *** 4 m.
m y
- d 4 w
~ ** 3 M.:
o v
o f5 L
E ** +
$. 5' } b u
D ,~ u
~ *
- . H u a-
.' < 9 O ru .at
- ** .m3 ..I m a
k b
~
- m. -
- N
- d + .. 4i m" g >.
"O m
e, W
~
amie M
-= ,, m f
s 4
O N O -Q 7,
<o L
-o a mi < .N 'n y 3
- * .= ca .= . .
=
'.:J p oe
- 4 u.%: ^=
f '.:2 o NAv 3 ')
'd 8
1