ML20236V153
| ML20236V153 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 07/23/1998 |
| From: | Langenbach J GENERAL PUBLIC UTILITIES CORP. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| 1920-98-20394, GL-97-04, GL-97-4, NUDOCS 9807310333 | |
| Download: ML20236V153 (12) | |
Text
_ _ _ _ _ _ _ _ _. _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _
l GPU Nuclear, Inc.
h Route 441 South NUCLEAR Post 0%ce Box 480 7
Middletown, PA 17057 0480 Tel 717-944 7621 July 23,1998 1920-98-20394 U. S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, DC 20555 Gentlemen:
Subject:
Three Mile Island Nuclear Generating Station (TM1-1)
Docket No. 50-289 Facility Operating License No. DPR-50 Generic Letter 97 Response to Request for Additional Information
References:
(1) GPU Nuclear Letter 6710-97-2533," Generic Letter 97-04 Response", dated January 6,1998.
(2) USNRC Letter, " Request for Additional Information, Three Mile Island Nuclear Station Unit No.1 (TM1-1) Response to Generic Letter 97-04, " Assurance of Sufficient Net Positive Suction Head for Emergency Core Cooling and Containment Heat Removal Pumps", (TAC No. MA0054)", dated June 2,1998.
On June 2,1998, the Nuclear Regulatory Conunission requested within 60 days (Reference 2) additional information related to GPU Nuclear's submittal (Reference 1) for Generic Letter 97-04. On the basis of this request and a subsequent clarification conference call held with the cognizant NRC Staff reviewers on l
June 15,1998, GPU Nuclear has revised the Reference I submittal. Attaclunent I to this letter prosides a i
i complete rev.' sed and enhanced version that incorporates all requested clarifications of the infornmtion for TMI-1.
If you have any questions or conunents on this matter, please contact Ron Zak, Corporate Regulatory Affairs at (973) 316-7035.
Very truly yours, Ulka 7
(
Y James W. Langenbach o
Vice President and Director, TMI l
l 9807310333 900723 l
PDR ADOCK 05000289 P
1920-98-20394 Page 2 I, James W. Langenbach being duly sworn, state that I am a Vice President and Director of GPU Nuclear, Inc. and that I am duly authorized to execute and file this response on behalf of GPU Nuclear. To the best of my knowledge and belief, the statements contained in this document are true and correct. To the extent that these statements are not based on my personal knowledge, they are based upon information by other GPi' Nuclear employees and/or consultants. Such information has been reviewed in accordance with company practices and I believe it to be reliable.
4MJ Y
/
James W. Langenbach Vice President and Director, TMI Signed and suorn before me this M
day of //[ /,1998.
/ /
'/bNtalWho' Notary Public '
Notarial Seal ocbra S Klick, NotaryPiMa torowlerry Twp., Dauphlp Courity My Comntssion Exrxres Jufy 4.MKG Memeer, Pennsylvania Assoc!ation at Notanos c: Administrator, NRC Region 1 Senior Resident inspector, TMI-l TMI-l NRC Project Manager
1920-98-20394 Page 1 of 10 ATTACIIMENT 1 TMI-1 Response to Generic I.etter 97-04 1.
Specify the general methodology used to calculate the head loss associated with the ECCS suction strainers.
The available net positive suction head (NPSli ) Of the pumps is CalculatCd using the following A
relationship:
I NPSI1.,.w, = h. - h,.p, + h.e - hr.
where:
h, is the atmospheric pressure above the Reactor Building (RB) Sump liquid. When evaluating NPSil the value of this term is taken as the vapor pressure of the sump liquid.
h,.po,is the vapor pressure of the RB Sump liquid. When evaluating NPSH the value of this term is taken at the maximum liquid temperature following the initiation of RB sump recirculation.
That temperature occurs at the time of recirculation.
- h. is the static fluid pressure associated with the pumps. It represents the pressure associated with the weight of water above the centerline of the pump inlet. It is based on a conservative calculation of the RB water level.
hr.is the head loss associated with the flow through the system. The head loss calculation is an evaluation of the piping length, geometry ar.d fittings that include the sump screens. The friction factors, geometry and fitting loss coefficients are established using methods described below.
Prior analyses of the suction piping were based on IJD values, which express the head loss in a valve or fitting in terms of an " equivalent length" of straight pipe. This approach was used in Crane Co. Technical Paper No. 410 prior to 1976. The primary shortcoming of this approach is that the head loss for pipe varies significantly with Reynolds Number, whereas the head loss for i
valves and fittings does not. The llD values were selected to provide accurate results for fully turbulent flow, aa (, significantly overstated head losses at low Reynolds Numbers.
l The velocity head approach expresses the head loss as a dimensionless "K-factor" equal to the head loss expressed in velocity heads. This method eliminates the problem discussed above, and, has been used in Crane since 1976.
w-___-_-_--.-___-___
J
1920-98-20394 Page 2 of 10 In addition to Crane, there are other references that provide more detailed K-factors, such as K-factors for combining and diverging flow in Tees. The current calculation evaluates these references and selects an appropriate reference for each type of fitting. Information from the following references was evaluated in the development of K-factors for the LPI and BS pump suction lines. The Short Names are used in the follow-on discussion for ease of reference.
References Evaluated for System Resistance Calculations Short Name i
Fluid Flow Data Book, General Electric Corporate Research and GE L
Development, Schenectady, NY 2
Handbook of Hydraulic Resistance by 1. E. Idel'Chik,1960.
Idel'Chik Published by The National Science Foundation as AEC-TR-6630.
Translated from Russian.
3 Flow of Fluids Through Valves, Fittings, and Pipe, Technical Paper Crane No. 410, Crane Co.,800 Third Ave., King of Prussia, PA,1988.
i 4
The Two-K Method Predicts Head Losses in Pipe Fittings, William Two-K j
B. Hooper, Chemical Engineering, August 24,1981, pp. 96 - 100.
5 Calculated Head Loss Caused by Change in Pipe Size, William B.
Hooper Hooper, Chemical Engineering, November 7,1988, pp. 89 - 92.
l 6
Modeling Pipe Networks Dominated by Junctions, D. J. Wood, L.
Wood l
S. Reddy and J. E. Funk.
l l
7 Handbook of Valves, Philip A. Schweitzer, Robert E. Krieger Schweitzer Publishing Company, Malabar, F1,1982.
t 8
Application Guide for Check Valves in Nuclear Power Plants, EPRI EPRI NP-5479S, Rev.1,1993 (EPRI Licensed Material) 9 Internal Flow Systems,2"d Edition, Donald S. Miller, Gulf Miller Publishing, Houston,1990.
10 ASHRAE Handbook,1981, Fundamentals, American Society of ASHRAE Heating, Refrigeration and Air Conditioning Engineers, Atlanta.
1 l
i 0
A
1920-98-20394 Page 3 of 10 Most published K-factors for fittings are based on a total enerev approach and include only irrecoverable friction losses. Ilowever, other references analyze the variation of pressure from point-to-point in a stream in terms of static pressure. This is called the static pressure approach.
Static Pressure = Total Pressure - Velocity Head Thus, K-factors from references using the static pressure approach include both frictional pressure drop and reversible changes in kinetic energy. These K-factors were adiusted so that they could be used to determine friction factors.
Special consideration has to be given to tee and wyc fittings. As discussed in Idel'Chik, the losses in a diverging wyc or tee usually consist of:
shock loss accompanying a sudden expansion at the point of Cow branching, and a.
b.
losses due to stream curving along the branch and the accompanying shock in the straight passage.
The resistance coefficient (K-factor) of the straight passage can have a negative value at certain values of the discharge ratio (Qw/Qw), which can occur in this passage. This is caused by the beanch receiving a larger share of the slowly moving boundary layer than the high-velocity core at a stream division. Hence, the energy of a unit volume of the fluid in a straight passage will be higher than that unit volume in the branch. The energy increase in the straight passage is accomp;tnied by an energy decrease in the branch.
By convent:an, the K-factor for a fitting or valve is based on the velocity head at the inlet.
However, thee are exceptions:
The K-facter for a pipe inlet from a large space is based on the velocity head in the pipe.
a.
b.
The head loss in e wye or tee with combining flow is based on the velocity head in the leg carrying the combinei flow.
When performing head loss calculadons involving a wye or tee with combining or diverging flow, the K-factor is adjusted for use with the velocity head in the appropriate stream. For example, in a tee with diverging flow, the K-factor for tw loss from the in!ct run to the branch is based on the inlet velocity. However, the loss must be asswiated with the branch outlet flow stream at its velocity head. The K-factor is adjusted by the sqare of the ratio of either the velocities or flows.
Reynolds Numbers were calculated uased in the internal pipe diameter for the expected range of LPI and BS flows and process fluid temperature.
i l
l
i 1920-98-20394 Page 4 of 10 The following table represents the evaluation performed for each fitting type in the suction lines from the RB sump to the inlet of the LPI and BS pumps.
i Component Type References Reference Bases Evaluated Selected l
Pipe Crane Crane Chart representing Friction Factor as a
{
function of Reynolds Number representing the Colbrook equation is the industry standard.
Elbow Crane Idel'Chik
- 1. Crane does not provide a method to 90 LR Elaow Two-K calculate K-factor for a welded 45 LR 45 LR Elbow Idel'Chik
- Elbow, 15' Elbow / Bend GE 2.
Crane, Two-K and GE do nct provide method to calculate K-factor for a 15 Elbow / Bend.
3.
Idel'Chik method generates consistent K-factors and is valid for all angles.
Standard Pipe Crane Hooper 1.
Crane does not provide a method to Reducer Idel'Chik calculate K-factor for a " Standard" Hooper pipe reducer.
2.
Crane gradual contraction method would be based on an estimated angle.
3.
Idel'Chik method uses the friction loss through a reducer.
4.
Hooper method is specifically for Standard Pipe Reducers.
Reducing Elbow Idel'Chik Approximation
- 1. Idel'Chik statements are contrary to j
Approximation using 90 LR data presented.
l using 90 LR Elbow i
Elbow (Idel'Chik) + Std (Idel'Chik) +
Pipe Reducer i
Std Pipe (Hooper)
Reducer (Hooper)
Fee (Round-Edge)
Crane Miller 1.
Crane method is for a threaded fitting, Non-Diverging Flow Idel'Chik not fcr a welded fitting.
l I
Run-to-Run GE 2.
Idel'Chik method approximates the loss Run-to-Branch Wood for the Run-to-Run configuration.
Diverging Flow Miller
- 3. GE method only applies to sharp-edge Run-to-Run fittings.
Run-to-Branch
- 4. Wood method only applies to sharp-edge fittings at high flow to branch.
- 5. Results from the other references cluster around Miller at low flow to the branch.
I E
1920-98-20394 Page 5 of 10 Component Type References Reference Bases Evaluated Selected Gate Valve Crane Two-K 1.
Idel'Chik only provides one K-factor Two-K value for all valve sizes.
Idel'Chik 2.
Crane, Two-K and Schweitzer methods Schweitzer are consistent and adjust for valve size.
Manufacturer
- 3. Two-K method adjusts for valve size Data and low Revnolds Number.
Swing Check Valve Crane EPRI
- 1. EPRI method is more representative of (Internals Installed)
Two-K valve maximum open angle, which is Idel'Chik slightly greater than % the typical valve Schweitzer maximum open angle of 70 EPRI Swing Check Valve Approximation Average of
- 1. Crane and Hooper methods gave similar (Internals Removed) using Sudden Crane and values for both the sudden expansion Expansion +
Hooper and gradual contraction.
Gradual Contraction:
i Crane Hooper Inward Projecting Crane ASHRAE 1.
Crane provides only one value for Pipe Entrance ASHRAE inward projecting pipe entrance.
- 2. ASHP AE method allows adjustment for geometry.
The TMI-l ECCS suction strainer consists of a large box in the reactor building sump, covered by a mesh screen with 0.125 inch openings that prevent large particulate material from entering the pump suctions. The opening size is consistent with the pump manufacturer requirements for particulate material entrained in the fluid.
Head loss through the mesh screens is calculated assuming the maximum allowed flow from two trains of ECCS with 50% surface area clogging of the screens. The calculation used the methodology established by Diagram 8-6 in Section VIII in " Handbook of Hydraulic Resistance, Coefficients of Local Resistance and of Friction", I. E. Idel'Chik,1960. Because of the large surface area, even with 50% cloggmg, the head loss would be less than 0.1 ft I
l 1920-98-20394 j
Page 6 of 10 l
2, Identify the required NPSil and the available NPSil.
l The required net positive suction head (NPSlia) is based upon the pump manufacturer's NPSH curves (References 8 and 9) provided for the pump type installed at the site:. The available NPSlix is calculated for each pump based on the configuration for the train of equipment in which the pump operates. The calculation of available NPSH accounts for one train of Low Pressure A
injection (LPI) and Building Spray (BS) taking suction from the RB Sump. If either a LPI or BS pump is not operating the we available NPSila for the other pump would increase because of the common suction line for each train of equipment (i.e., lower head loss for lower common suction line flow).
Suction from the Reactor Building Sump' I
"A" "B"
I Pumps System Pump NPSlia NPSH NPSHA A
Flow' Flow'#
Low Pressure 24'70 gpm 3020 gpm i1.078 ft 14.489 ft 15.047 fl Injection (LPI)
Reactor Building 1448 gpm 1448 gpm 13.659 ft 14.743 ft 15.107 ft Spray (BS)
Suction from a Piggy-Back Source from the Reactor Building Sump l
Pumps System Flow Pump Flow NPSila NPSliA Suction (3)
(3)
Source i
liigh Pressure 954 gpm 954 gpm 30 ft 312 fl LPI Injection (HPI)
Pumps N_ ole 1 1. Accounts for instrument error.
- 2. Accounts for pump recirculation flow.
i
- 3. Based on three 11P1 pumps in operation.
l
- 4. Minor changes to the NPSH calculation since previous submittal.
l
l 1920-98-20394 l
Page 7 of 10 3.
Specify whether the current design-basis NPSil analysis differs from the most recent analysis reviewed and approved by the NRC for which a safety evaluation was issued.
l l
The most recent calculation for the LPI and BS pumps (Reference 1) uses the same methodology approved in the SER (Reference 7) allowing credit for overpressure equal to the vapor pressure of the sump liquid (i c., h, = hy). The LPI and BS pump flow values assumed in the NPSil calculations were varied until the available NPSH exceeded the required NPSHa, accounting for A
instrument error and pump recirculation flow. As a result of these calculations the internals were removed from one check valve in the suction piping of each BS pump to satisfy the NPSH requirements at the procedure flow limit plus instrument error as well as maintain the minimum BS l
flow for various analyses. The configuration change was evaluated by a 10CFR50.59 evaluation and found to be acceptable. A 10 CFR 50.59 safety evaluation (Reference 4) of the adjusted flow l
values was performed to document the acceptability of the change. The plant procedures (References 8 and 9) were modified to account for the new flow limits to satisfy the NPSH requirements. The BS pumps flow limits did not change. The plant procedures throttle BS & LPI to meet the flow limits prior to establishing flow from the RB sump.
j With full flow from all available HPI, LPI and BS pumps the BWST inventory would be depleted voithin approximately 30 minutes. With only one train of equipment in operation it would take in excess of I hour to deplete the BWST. The RB sump switchover process would be completed at
)
that time. The procedure guidance for the RB sump switchover process had been revised to address timing issues. The current guidance (References 8 and 9) is as follows:
If the HPl pumps are still running, transfer HPI pump suction to the LPI pump discharge a.
(piggy-back operation) prior to reaching the BWST level setpoint for RB Sump switchover, b.
When BWST level reaches the low-level setpoint (9'6"), open the sump suction valves and throttle LPI flow.
c.
When BWST level reaches the low-low 3evel setpoint (6'4"), close the BWST and NaOH Tank suction valves.
From tha analysis that generated the reactor building EQ Temperature and Pressure Profiles (Referer.cc 3), RB sump switchover occurred at approximately 1.3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />. The RB sump liquid temperature at that time in the analysis was approximately 225 F. which is the maximum temperature for the remainder of the transient. The vapor pressure corresponding to the liquid temperature would be approximately 18.9 psia. A RB pressure of approximately 27 psia was predicted at the time of recirculation, increasing to a maximum of 29 psia before decreasing. Pa -
Pvapor at the time of recirculation would be approximately 8.1 psi (18.7 ft). Sensitivity cases were performed in Reference 2 to dete.mine the minimum value of Pa - Pvapor. As expected, the combination of equipment used in the RB EQ analysis does not generate the minimum value of Pa
- Pvapor. Neither did cases that maximized RB cooling (i.e., minimum ultimate heat sink temperature and maximum flow). Those cases generated the minimum R.B pressure.
l
(
1920-98-20394 i
Page 8 of 10 However, they generated very low RB sump liquid temperatures. The minimum values of Pa -
Pvapor 3 vere generated with high ultimate heat sink temperatures with 2 trains of BS, I train of LPI, and i RB fan cooler. The following table provides the 8 smallest values of P - Py,y from i
that sensitivity study. It is included for information:
Recire Minimum Maximum Case Time Pressure Liq Temp Pv Pa-Pv Pa-Pv (sec)
(psia)
(F)
(psi)
(psi)
(ft) 415 3041 20.08 208.0 13.58134 6.50 15.01580 515 3041 23.10 218.1 16.58048 6.52 15.06400 416 3041 20.32 208.6 13.74369 6.58 15.19523 516 3041 23.38 218.8 16.80364 6.58 15.19'U4 417 3255 19.31 204.6 12.67011 6.64 15 '34213 517 3255 22.33 215.3 15.68784 6.64 15.34737 513 2871 23.15 217.6 16.42108 6.73 15.54784 413 2871 20.12 207.2 13.36488 6.76 15.60839 The results of these calculations are not used in the NPSH calculations because a P. - Puoy value of 0.0 is used in the NPSH calculations as indicated earlier. The 10 CFR 50.59 cvaluation found the revised LPI and BS pump-throttling criteria to be acceptable based an the following considerations:
The NPSH requirements for the LPI and BS pumps are met while taking suction from the a.
b.
The EQ qualified components were evaluated against the revised Containment EQ Temperature and Pressure profiles, generated using the reduced flows accounting for instrument error, and found to be acceptable.
c.
The short and long-term core cooling requirements are met.
d.
All FSAR Chapter 14 accidents were reviewed against the changes in flow and were found to be acceptable.
The Technical Specifications were reviewed and no changes were required.
c.
f.
Post-LOCA boron precipitation is not affected by the change in flows.
g.
The Containment Peak Pressure is not affected by the reduced flows.
h.
The plant procedure changes were verified to be acceptable using the plant replica
{
simulator. The procedure changes did not alter the operator burden.
1920-98-20394 Page 9 of 10 4.
Specify whether containment overpressure (i.e., containment pressure above the vapor pressure of the sump or suppression pool fluid) was credited in the calculation of available NPSil. Specify the amount of overpressure needed and the minimum overpressure available.
No credit is taken for containment overpressure in excess of the vapor pressure of the reactor building sump liquid.
5.
When containment overpressure is credited in the calculation of available NPSil, confirm that an appropriate containment pressure analysis was done to estat'ish the minimum containment pressure.
No credit is taken for containment overpressure in excess of the vapor pressure of the reactor building sump liquid. Therefore, containment pressure analyses were not required.
4 l
l I
,o 1920-98-20394 Page 10 of 10 References 1.
GPU Nuclear Calculation C-Il01-210-E610-Oll,"LPI and BS Pump NPSH Available from the RB Sump Following a LBLOCA", Rev.1, dated 4/24/98.
2.
GPU Nuclear Calculation C-1101-212-E610-053, "TMl-1 RB Conditions for NPS11 Calculations Using GOTHIC", Rev.1, dated 6/26/97.
3.
GPU Nuclear Calculation C-1101-823-5450-001,"TMI-l LBLOCA EQ Temperature Profile Using the GOTHIC Computer Code", Rev. 5, dated 3/31/98.
4.
GPU Nuclear Safety Evaluation SE-000212-032, " Decay Heat and RB Spray pump Flow Throttling Criteria for RB Sump Recirculation", Rev.1, dated 7/15/97.
5.
Worthington Drawing E-196410A for Pump Model 8HN-194, Serial Number 1621419.
6.
Worthington Drawing E-196665A/B for Pump Model 6HN-134, Serial Number 1621420.
7.
" Safety Evaluation by the Directorate of Licensing U.S. Atomic Energy Commission in the Matter of Metropolitan Edison Company, Jersey Central Power & Light Company, Pennsylvania Electric Company Three Mile Island Nuclear Station Unit 1, Dauphin County, Pennsylvania, Docket No.
50-289", Section 6.2.3.2, Net Positive Suction Head, received 7/11/73.
8.
TMI-l Abnormal Transient Procedure 1210-6,"Small Break LOCA Cooldown" 9.
TMI-l Abnormal Transient Procedure 1210-7, "Large Break LOCA Cooldown".
j 3
i 1
-