ML20195H724
| ML20195H724 | |
| Person / Time | |
|---|---|
| Site: | Farley  |
| Issue date: | 01/14/1988 |
| From: | Mcdonald R ALABAMA POWER CO. |
| To: | NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM) |
| References | |
| NUDOCS 8801200327 | |
| Download: ML20195H724 (13) | |
Text
- - - - - - - - _ - - - - - -
2%"Vfsse?"'
Post Offee Box 2641 Birmingham, Alabama 352914400 Telephone 205 250-1835 LLT/.1,n, Alabama Power the scuhern ekctrc spttin 10CFR50.46 January 14, 1988 Docket Nos. 50-348 50-364 U. S. Nucicar Regulatory Commission ATTN: Document Control Desk Washington, DC 20555 Gentlemen:
Joseph M. Farley lluclear Plant - Units 1 and 2 Evaluation of the Ef fect of Reduced Safety Injection Flow Corresponding to Spilling to Containment Backpressure on the Snall Break LOCA Analysis Recently the Joseph M. Farley lluclear Plant Unit 2 experienced a crack in a safety injection line attached to the reactor coolant system (RCS) cold leg. As a result, Alabama Power Company personnel have discussed the event with the Nuclear Regulatory Commission (NRC) Staff. This has resulted in a number of questions concerning the small break loss-of-coolant accident (LOCA) analyses documented in the Joseph H. Farley Final Safety Analysis Report Update (FSAR).
Specifically, the questions concerned the assumptions regarding the safety injection flow employed by Westinghouse in the August 1974 small break LOCA analyses using the WFLASH computer code as documented in the Joseph M. Farley FSAR.
Provided as Attachment 1 is an evaluation of the effect of reduced safety injection flow corresponding to spilling to containnent backpressure on the small break LOCA analysis.
This evaluation concludes that the effect of the lower safety injection flow results (conservatively) in a 46*F increase in the peak cladding termerature.
These new results, however, were found to still maintain considerable rnargin to the 2200*F Ifmit of 10CFR50.46 and to continue to be bounded by the i.arge Break LOCA result of 2013*F.
Furthermore, the small break LOCA analyses using the WFLASH computer code contain a number of conservative assumptions in addition to those specifically required by Appendix K to 10CFR50.46. Attachment 2 identifies some of these additional conservative assumptions.
a801200327 %%4g
\\
k i
U. S. Nuclear Regulatory Commission January 14, 1988 Page 2 An FSAR update will be made at the next applicable revision to reflect the peak cladding temperature effects contained in this eyeluation.
This update will also include the correction of the peak cladding temperature reported for the 6-inch break case as discussed in Attachment 1.
If there are any questions, please advise.
Respectfully submitted, ALABAMAFf iPANY l
j, ) r{ Gut l Q
m R. P. Mcdonald RPM / JAR: dst-D-T.S 7 Attachments cc: fir. L. B. Long Dr. J. N. Grace Mr. E. A. Reeves Mr. 11. H. Bradford l
ATTACHMENT 1 Evaluation of the Effect of Reduced Safety Injection Flow Corresponding to Spilling to Containment Backpressure on the Joseph M. Farley Small Break LOCA Analysis The current Westinghouse small break LOCA analysis methodology employs a number of conservative assumptions.
The current methodology assumes minimum safeguards safety injection (SI) flow in conjunction with the loss of offsite power and the failure cf one train of pumped flow in the emergency core cooling system (ECCS).
The Westinghouse small break LOCA analysis methodology includes analyses of a range.of break sizes and locations.
Westinghouse has found that small break peak cladding temperatures are bounded by the results of large break LOCA analyses.
Westinghouse has also found that small breaks in the cold leg result in higher peak cladding temperatures than small breaks at other locations since SI flow to the faulted loop could be diverted from reaching the core. For conservatism, Westinghouse currently assumes that all safety injection flow to the faulted loop spills and is lost for breaks in the cold leg.
The amount of safety injection flow delivered to the non-faulted loops is a function of the backpressure and the resistances of the injection lines.
For break sizes smaller than the inner diameter of the safety injection line, the current Westinghouse methodology assumes that the SI flow delivered tc the faulted loop cold leg spills to the RCS backpressure.
For break sizes greater than or equal to the SI line inner diareter, Westinghouse currently assumes that the SI flow delivered to the broken loop spills to the containment backpressure.
At the request of Alabama Power Company, Westinghouse reviewed the Joseph M.
Farley emergency core cooling systen piping configuration and the specific small break LOCA analyses which form the basis for the information in the Joseph M.
Farley FSAR. The review of the ECCS piping configuration showed that pumped 51 flow from the high head SI (charging) pumps and the low head SI (RHR) pumps enters the cold leg through a common 6-inch line.
Passive SI flow from an accumulator enters the cold leg through a separate 12-inch line. For this ECCS piping configuration, the current Westinghouse small break LOCA analysis methodology would assure that the broken loop SI flow spills to containmnt backpressure for break sizes greater than the inner diameter of the 6-inch Si line (5.187 in.).
A review of the small break LOCA analyses which form the basis for the information contained in the Joseph M. Farley FSAR indicated that the analyses were performed using the Westinghouse August 1974 snall break LOCA ECCS Evaluation Model which incorporated the WFLASH computer code. The highest peak cladding temperature calculation resulted for a 6-inch equivalent diameter break at the bottom of the cold leg.
The analyses assuned that no 31 flow or accumulator flow was delivered to the broken loop cold leg for breaks in that loca tion.
This UTLASH Evaluation Model is conservative relative to the calculation of peak cladding terperatures.
The conservatisns in the Farley WFLASH analyses are discussed further in Attachment 2.
Evaluation of the Effect of Reduced Safety injection Flow Corresponding to Spilling to Containment Backpressure on the Joseph II. Farley Small Dreak LOCA Analysis Page 2 The detailed review of the conputer analyses results revealed that the broken loop SI flow was assumed to spill to RCS backpressure for all breaks including the 6-inch analysis.
This is inconsistent with the current small break LOCA analysis methodology as discussed above. However, the small break LOCA analysis results documented in the Joseph M. Farley FSAR remain limiting for all small breaks in the RCS, except for a break in the ECCS injection line equivalent to or larger than the double ended severance of that line.
An evaluation was performed to assess the effect on the Joseph M. Farley small break LOCA analyses of a reduction in the SI flow equivalent to the assumption of spilling the broken loop SI flow to containment backpressure.
The evaluation Indicates that the peak cladding temperature for the limiting 6-inch case can be conservatively estimated to increase by approximately 46*F.
The limiting case continues to be the 6-inch break. Details concerning the calculation are contained in the evaluation section.
The detailed review of the analyses also revealed that the peak cladding temperature for the 6-inch break case was incorrectly reported to be 1703*F for the downflow barrel baffle design which exists on Unit 2.
This number was determined from a table containing information from the analysis calculation of cladding temperature versus tine in two second increments.
Calculation sunmary information was also found in the computer analysis output which identifies the peak cladding temperature to be 1712*F at a time that was between two time increments of the cladding tenperature versus time table.
The peak cladding temperature should have been reported to be 1712'F at 292.8 seconds.
The Joseph M. Farley FSAR also s eports the peak cladding terperature for Unit 1, which was converted to an upflow barrel baf fle design, as 1820*F.
The reporting discrepancy also affects the peak cladding temperature estimate for the upflow design.
The peak cladding temperature for the upflow barrel baf fle design in Unit 1 should have been reported as 1829'F.
Based on the above, consicerable review effort was devoted to assuring the accuracy of the remainder of the reported results.
To the best of our knowledge, the remainder of the information contained in the Joseph fl. Farley FSAR correctly reflects the analysis results.
Evaluation An evaluation was performed to conservatively estimate the effect of the spilling assunption on the peak cladding temperature calculation of the small break LOCA analyses in the Joseph M. Farley FSAR.
Of the three analysis results docunented in the Joseph it. Farley FSAR, only the 6-inch break analysis would be affected by the spilling line assumption. Consequently, an evaluation was performed to assess the ef fect of a reduction in the S! flow equivalent to the assumption of spilling the broken loop SI flow to containment backpressure for the 6-inch break analysis docunented in the Joseph M. Farley FSAR.
r Evaluation of the Effect of Reduced Safety Injection Flow Corresponding to Spilling to Containnent Backpressure on the Joseph M. Farley Small Dreak LOCA Analysis Page 3 The RCS pressure remained above the shutoff head of the residual heat removal (RHR) pumps for the 6-inch break analysis, which precluded RHR injection flow.
Therefore, this evaluation need only review the high head injection flow rate differences.
A review of Joseph M. Farley minimum safeguards SI flow data indicates that the delivered SI flow for one high head SI (charging) pump differs betwecq the two spilling backpressure assumptions as a function of RCS pressure. The mass injected to the reactor coolant system (RCS) for each broken loop spillage assunption along with the difference between the two spillage assumption injection rates is sunmarized in Table 1.
The evaluation assumed that the RCS behavior would not be significantly altered by the change in SI flow rate.
Since the small break LOCA analyses assumed the loss of offsite power, the RCS pressure drops to a pressure slightly above the steam generator secondary safety valve setpoint until the break flow is capable of removing all of the decay heat.
This typically occurs when the break flow becomes all steam.
Prior to the transition to all steam break flow, the reduction in SI flow would not affect the RCS pressure in the WFLASH small break LOCA analyses. Af ter the transition to all steam break flow, a reduction in the SI flow rate would only have a secondary effect on the RCS pressure due to a slight reduction in the steam condensation capability.
A reduction in the SI flow would result in a smil increase in the enthalpy of the fluid at the break.
The slight increase in fluid enthalpy would have only a small effect which would tend to reduce the break flow.
If the ef fect en the break flow is conservatively neglected, the reduction in the SI flow would result in less mass in the RCS as a function of time. The reduction in the RCS mass inventory as a function of time provides the primary effect for reduced SI flow.
The 6-inch computer analysis calculation which forms the basis for sone of the information contained in the Joseph M. Farley FSAR was reviewed to determine the amount of additional SI flow which would be supplied when spilling to RCS back-pressure is assuned instead of assuming spilling to containnent backpressure.
Safety injection is calculated to start in the 6-inch WFLASH analysis between 40 and 45 seconds after the pipe breaks, For conservatisn, the evaluation calculations will assune that the SI flow begins 40 seconds af ter the pipe breaks to give the maximun dif ference in the amount of SI flow on an integrated basis.
The RCS pressure which determines the amount of SI flow to be injected was reviewed as a function of time af ter 40 seconds.
This information was used to determine the difference in the amount of integrated S! flow to the RCS for the transient.
This information is contained in Table 2.
Since the primary effect of the reduction in SI flow is a reduced RCS rass inventory, the information of Table 2 was used to conservatively estinate an earlier tine of core uncovery when spilling to containment backpressure is assumed instead of spilling to RCS backpressure.
Evaluation of the Effect of Reduced Safety Injection Flow Corresponding to Spilling to Containnent Backpressure on the Joseph M. Farley Stall Break LOCA Analysis Page 4 l
Calculation of Earlier Core Uncovery Tine The core mixture level during a small break LOCA with the reactor coolant pumps tripped depends upon the n3nometric head balance between the reactor inner vessel (core and upper plenum), reactor vessel downconer, steam generator inverted U-tubes (T-hot side and T-cold side), and crossover leg loop seal (downflow side from steam generator outlet plenum to the bottom of the horizontal section and upflow side from the bottom of the horizontal section to the reactor coolant pump). As RCS nass is depleted by the break during a small break LOCA, distinct mixture icvels are formed in various locations of the RCS as the system begins to drain from the top elevations down.
The steam formed by flashing and decay heat boiling displaces liquid in the inner reactor vessel.
T-cold side of the steam generator inverted U-tubes, and downflow side of the loop seal, while entraining liquid in the T-hot side of the steam generator inverted b-tubes.
There are, of course, dynamic ef fects in the larger small break sizes, but basically the static head of the collapsed liquid level in the inner reactor vessel in combination with the static heads in the steam generator T-hot side and the loop seal upflow side balance with the static heads of the T-cold side of the steam generator inverted U-tubes, the downflow side of the loop seal, and downcomer. Therefore, as the mixture level in the T-cold side of the steam generator inverted U-tubes and the downflow side of the loop seal decreases due to inventory depletion, the mixture level in the inner reactor vessel is depressed.
This leads to uncovering of the reactor core.
A reduction in the SI flow rate results in a decreased RCS system inventory as a function of time. Consequently, the mixture level in the steam generator T-cold side of the steam generator inverted U-tubes and the downflow side of the loop seal will be lower for a case with reduced SI flow.
'his results in uncovering the core slightly earlier.
A review of the core and loop seal mixture levels as calculated in the 6-inch srall break LOCA analysis for the Joseph M. Farley plant was performed.
This information is provided in Table 3.
The core mixture level is given in feet above the bottom of the core. Mixture level elevation information for both the downflow and upflow sides of the loop seal in the broken loop is given in feet above the botton of the horizontal section of the crossover leg.
The maximum elevation of the upflow side of the loop seal is 12.75 feet.
The maximun elevation of the downflow side of the loop seal is 20.52 feet.
The r
level information for the intact loop seal is similar to the broken loop.
A review of this infornation shows that the core mixture level drops rapidly between 100 and 130 seconds.
This rapid core uncovery is caused by the dropping of the level in the downflow side of the loop seal due to RCS mass inventory depletion. Changing the SI flow rates could af fect the loop seal downflow side level depression and tine of core uncovery.
Decreasing the S1 flow rates could i
cause the loop seal depression to occur earlier for this calculation because of the lower RCS mass inventory.
Attachment l' Evaluation of the Effect of Reduced Safety Injection Flow Corresponding to Spilling to Containment Backpressure on the Joseph M. Farley Small Break 1.0CA Analysis page 5 The level of loop seal downflow side falls below the level of the upflow side between 90 and 95 seconds. To determine how much earlier this could occur, the integrated SI difference for 100 seconds was used to be conservative. At 100 seconds the difference in total mass injected into the RCS between the two spillage cases is approximately 480 pounds as seen in Table 2 If this mass were not present in the RCS, the loop seal downflow side mass may have been depleted earlier forcing the core mixture icvel down earlier.
To conservatively estimate how much earlier the loop seal downflow side level could decrease, the following assumption was made:
All of the difference in integrated Si to the RCS is assumed to be contained in the downflow side of the broken loop seal.
The integrated SI mass difference will be removed from that location for this evaluation calculation.
This is conservative because if the mass difference were assumed to be uniformly distributed throughout the RCS instead of just in the broken loop seal, this would result in the calculation of a smaller increase in core uncovery time.
When calculating the mass depletion rate of the loop seal downflow side, information was used from 85 to 110 seconds in order to bracket the significant events of the transient (when the downflow side level decreases below that of the upflow side level and when the core mixture level decreases below the top of the core).
The average rass depletion rate of the loop seal downflow side from 85 to 110 seconds is 181.8 lbm/sec.
This was calculated by cividing the amount of mass depleted over this time range by the time range:
(7632 lbm at 85 sec - 3086 lbn at 110 sec)
(85 sec - 110 sec).
If there were 480 pounds less in the loop seal downflow side, the mixture level decrease could start approximately 2.64 seconds (480 lbm /181.8 lbm per sec) earlier.
Assuming that the core does not recover earlier because of the earlier uncovery time and corresponding earlier decrease in RCS pressure, the earlier core uncovery could result in an additional 2.64 seconds of core uncovery before the accumulators inject to reverse-the cladding tenperature rise.
Calculation of Effect Once Accumulators Inject The potential exists that for the lower 51 case the core will not be recovered as quickly as the case shown in the FSM once the accumulators start injecting
Evaluation of the Effect of Reduce) Safety Injection Flow
= Corresponding to Spilling to Containment Backpressure on the Joseph M. Farley Small Break LOCA Analysis Page 6 because there will be less mass in the RCS as a result of the lower SI. A calcul. tion was conservatively performed to. determine the extent and ef fect of this de'.ay.
In the 6-inch analysis the core recovers at approximately 296 seconds. A review of Tatle 2 shows that at 300 seconds the difference.in integrated SI flow would be atproxima.tely 1800 pounds.
A calculation was performed to estimate how long it woula take the accumulators to inject this amount of water. The core mixture level could stay down this much longer because the accumulators would have to make up this mass before the level would climb.
The average accumulator in.iection rate between 270 and 290 seconds is 1603.6 lbm/sec. Assuming a conset vative injection rate of 1500 lbm/sec, it will take 1.20 seconds (1800 lbm /1500 lbm per sec) to inject the mass before the core mixture level will increase.
The total estimated additional time of core uncovery caused by the lower SI flow rate is the sum of the estimated earlier uncovery and delayed recovery times.
This total is 3.84 seconds.
In order to estimate the maximum possible increase in the peak cladding terperature, the additional core uncovery time of 3.84 seconds will be multiplied by the maximum cladding heatup rate which occurs shortly before the peak cladding temperature is reached.
The following cladding heatup information is taken from the 6-inch LOCTA run.
The peak cladding temperature occurs at 292.8 seconds. The cladding heatup rate between 280 and 290 seconds is:
(1683*F - 1564*F) / 10 sec = 11.9 *F/sec The estimated effect of the increase in core uncovery time on peak clad temperature is:
11.9 *F/sec
- 3.84 sec = 45.7 *F This estimate is rounded up to 46*F.
Conclusions The effect of the lower SI flow curve on the Joseph M. Farley 6-inch case was conservatively estimated to result in a 46*F increase in the peak cladding tempera ture. 'This results in a Unit 2 peak cladding temperature of 1758'F when applied to the correct analysis peak cladding temperatur( (1712*F) for the downflow barrel baffle case.
1
'l
' Attachment l' Evaluation of the-Effect of Reduced Safety Injection Flow Corresponding to: Spilling to Containment Backpressure on the Joseph-M. Farley Small Break LOCA Analysis
~Page 7 The increase in peak cladding temperature that results from the reduction in thel SI flow is applicable to both the downflow and the upflow barrel baffle plant configurations. This results in an estimated Unit 1 peak cladding temperature of:1875'F for. the upflow barrel baffle case, These' new results maintain considerable margin to the 2200*F limit for 10CFR50.46 and continue to be bounded by the Large Break LOCA result of 2013*F for Units 1 and 2.
b 1
t
ATTACHMENT 2
'Conservatisms in.the Joseph M. Farley l
Small Break LOCA Analyses Using the WFLASH Computer Code The small break LOCA analyses documented in the Joseph H. Farley FSAR contain a l
number of conservative assumptions in.Jdition to those specifically required by Appendix X to 10CFf.69.46.' This atta.nment identifies some of the important ones.
The break is conservatively modeled to be at the bottom of the cold leg.
Representing the break at the top of the cold leg, which would be appropriate for a break in the ECCS line and the spilling to containment backpressure situation, would provide benefits in terms of an earlier time to the transition to all steam break flow. Modeling the break at the top of the cold leg would also provide benefits in terms of tha s tatic head of fluid in the downcomer which would tend to result in higher core mixture levels.
Both of these effects would result in reductions in the calculated peak cladding temperature.
Furthermore, the small break LOCA analyses contained in the Joseph M. Farley FSAR conservatively neglect all of the broken loop passive accumulator flow.
This assumption is conservative for the Josep'. *. Farley ECCS design which has separate lines for the accumulator flow and the safety injection flow.
The minimum safeguards assumption accounts for hypothesized degradation of the safety injection pumps.
In reality, the flow which would reach the reactor coolant system could be considerably higher.
It should be noted that the August 1974 small break LOCA ECCS evaluation nodel, which was superseded by the December 1974 small break LOCA ECCS evaluation model, and by the October 1975 small break LOCA ECCS evaluation model, conservatively calculated the peak cladding temperature when compared to later models using WFLASH.
Furthermore, analyses with the May 1985 small break LOCA evaluation model using the NOTRUMP computer code have shown additional margins to peak cladding temperature calculations of over 200*F.
This indicates that the Joseph M. Farley peak clad temperature calculations are very conservative.
l l
l l
l
o Table 1 Joseph li. Farley ECCS High Head Injection Datal Flow into RCS (ibm /sec)
RCS Pressure Spill to RCS Spill to Containrent Difference (psia)
Backpressure Backpressure (lbm/sec) 2015 25.0 6.0 19.0 1815 29.0 13.0 16.0 1615 32.0 19.0 13.0 1415 35.0 25.0 10.0 1215 38.0 30.0 8.0 1015 41.0 34.0 7.0 815 44.0 38.5 5.5 615 47.0 42.0 5.0 415 49.0 46.0 3.0 215 50.0 49.5 0.5 1 Mininum Safeguards, One Charging Pump Operating l
t l
Table 2 Difference in Integrated Safety Injection Flow (1)
(2)
(3)
(4)
(5)
Injection flode Delta SI Delta Time X Sun of Tire Pressure Flow Column (3)
- Column (4)
(sec)
(psia)
(1bm/sec)
(lbm)
(lbm) 40 1151 8.0 45 1141 8.0 40 40 55 1127 8.0 80 120 65 1126 8.0 80 200 75 1131 8.0 80 280 85 1139 8.0 80 360 95 1143 8.0 80 440 100 1143 8.0 40 480 110 1142 8.0 80 560 120 1126 8.0 80 640 130 1093 7.5 80 720 140 1083 7.5 75 795 150 1064 7.5 75 870 170 996 7.0 150 1020 200 896 6.5 210 1230 240 757 5.5 260 1490 260 642 5.5 110 1600 270 514 5.0 55 1655 280 386 3.0 50 1705 290 272 3.0 30 1735 200 247 3.0 30 1765 325 335 3.0 75 1840 350 341 3.0 75 1915 375 340 3.0 75 1990 400 334 3.0 75 2065
- Higher of the two Delta SI flow values used
Table 3 RCS Mixture Level Information Loop Seal Loop Seal Core Mixture Downflow Side Upflow Side Time (sec)
Level (f t)
Level (ft)
Level (f t) 0.0 23.91 20.52 12.75 25.0 17.09 20.52 12.75 50.0 17.10 20.28 12.75 75.0 17.05 19.93 12.75 85.0 17.04 16.05 12.75 90.0 16,85 13.16 12.75 95.0 15.72 11.16 12.75 100.0 14.12 9.65 12.75 110.0 9.67 6.49 12.75 120.0 4.15 2.43 12.70 130.0 0.11 1.29 12.49 150.0 0.09 3.49 11.66 170.0 0.10 3.82 9.98 200.0 0.08 3.33 9.94 250.0 0.00 1.73 11.61 l
290.0 0.00 0.95 11.73 1
300.0 13.10 1.28 11.78 l
325.0 10.72 0.15 6.78 l
350.0 11.60 0.00 3.58 375.0 11.69 0.00 0.01 l
_ - -___ - - - - -___ - _________ - - _ _ _ -