ML13330A151
| ML13330A151 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 11/25/1980 |
| From: | Baskin K Southern California Edison Co |
| To: | Crutchfield D Office of Nuclear Reactor Regulation |
| References | |
| REF-GTECI-A-01, REF-GTECI-PI, TASK-A-01, TASK-A-1, TASK-OR NUDOCS 8012020521 | |
| Download: ML13330A151 (15) | |
Text
Southern California Edison Company "0
P.O. BOX 800 2244 WALNUT GROVE AVENUE ROSEMEAD, CALIFORNIA 91770 K. P. BASKIN TELEPHONE MANAGER OF NUCLEAR ENGINEERING, (213) 572-1401 SAFETY, AND LICENSING November 25, 1980 Director, Office of Nuclear Reactor Regulation Attention: D. M. Crutchfield, Chief Operating Reactors Branch No. 5 Division of Licensing U. S. Nuclear Regulatory Commission Washington, D.C. 20555 Gentlemen:
Subject:
Docket No. 50-206 Feedwater Water Hammer Evaluation San Onofre Nuclear Generating Station Unit 1
References:
- 1) Letter dated July 14, 1975, K. P. Baskin (SCE) to R. A. Purple (NRC)
- 2) Letter dated December 27, 1977, K. P. Baskin (SCE) to A. Schwencer (NRC)
- 3) Letter dated February 14, 1980, K. P. Baskin (SCE) to D. L. Ziemann (NRC)
- 4) Letter dated April 15, 1980, J. G. Haynes (SCE) to D. L. Ziemann (NRC)
- 5) Letter dated April 22, 1980, D. L. Ziemann (NRC) to Robert Dietch (SCE)
- 6) Westinghouse Report, SONGS 1 Water Hammer Analysis, October, 1980 (enclosed)
Our letter of February 14, 1980 (Reference 3), committed Southern California Edison (SCE) to evaluate the effects of "classic" type water hammers on the feedwater piping system at San Onofre Nuclear Generating Station Unit 1. This evaluation was originally scheduled for completion by September 30, 1980, and with agreement with the NRC staff, was extended to December 1, 1980. The purpose of this letter is to provide the results of that evaluation.
81012020
Mr. D. M. Crutchfield, Chief NoVember 25, 1980 As requested by the NRC staff, the analysis included the development of forcing functions for classical water hammer incidents, such as valve closure or pump start.
As reflected in our submittal of April 15, 1980 (Reference 4), and mutually agreed to in prior discussions with the NRC, the steam generator water hammer as postulated by Creare, Inc. in the 1976 Final Technical Report on PWR Steam Generator Water Hammer, was not analyzed by this most recent evaluation.
During the early stages of the analysis, a thorough review of previously submitted data on San Onofre Unit 1 water hammer experiences indicated that classical water hammer situations resulting from pump start need not be evaluated. Historical data indicate that no damage to main or auxiliary feedwater piping components or supports has ever occurred at San Onofre Unit I due to pump start.
Water hammer loads associated with the closure of the feedwater control valve (FCV) or the motor operated isolation valve (MOV) in each feedwater train were analyzed by Westinghouse under this program. A copy of this report (Reference 6) is enclosed. It is concluded that the transient loads resulting from the closure of either the FCV or the MOV has no significant effect on the piping stress and support loads inside containment.
This reinforces the conclusions of the April 22, 1980 NRC Safety Evaluation Report (Reference 5) that acequate evaluations and precautions have been taken to minimize the potential for such an occurrence.
Existing administrative controls on the manual initiation of auxiliary feedwater to the steam generators during startup, shutdown and low power operation substantially reduce the potential for water hammer incidents in this system. This consideration was previously reviewed by the NRC with our submittal of February 14, 1980. A system for the automatic initiation of auxiliary feedwater is currently in design and procurement. An evaluation of the potential for water hammer with the modified system will be completed prior to returning to power from this outage.
Previous submittals dated July 14, 1975, December 27, 1977 and February 14, 1980 (References 1, 2 and 3 respectively), have stated that there is low probability of damaging the steam generators as a result of feedwater system water hammer. This conclusion is based on the following considerations:
- 1. The feedwater piping configuration inside containment minimizes the consequences of.a steam generator bubble collapse water hammer.
- 2. The main feedwater supply is not isolated following a unit trip. Rather, the FCVs are throttled to a 5% open configuration, thus maintaining flow and preventing the piping and feed ring from draining.
- 3. Administrative controls established in 1975 and 1978 further reduced the probability of steam generator water hammer.
Mr. D. M. Crutchfield, Chief November 25, 1980 Unless additional questions are raised upon completion of the evaluation of the modified auxiliary feedwater system, we conclude that steam generator and feedline water hammer have a low probability of occurrence at San Onofre Unit 1. Further corrective actions, including modification of the steam generators or feedwater piping, are not considered warranted at this time.
If you have any questions, please let me know.
Very truly yours, Enclosures
SONGS 1 WATERHAMMER ANALYSIS October 1980 INTRODUCTION The feedwater piping/support systems inside the containment for both Loops A and B of. San Onofre Unit I were evaluated for the waterhammer transients loads. Loop C is sufficiently similar to Loop A such that evaluation of Loop C was not necessary.
These transient loads result from the valve closure of either motor operated isolation valve (MOV) or feedwater control valve (FCV). The hydro-dynamic loads were calculated and applied to the piping/support system for the structural analysis. The results of the structural analysis are piping stresses and support loads. The piping stresses are compared with the allowables given by ASME Section III Code and the support loads are verified with their Maximum Expected Loads provided by Southern California Edison.
In the structural analysis of the piping/support system only a worst case was selected for the analysis thereby avoiding unnecessary and repetitive calculations.
First, it was determined that the transient load from the valve closure of the motor operated isolation valve has the larger loads of the two transients. (As shown in Figures 3 and 4.)
Also, the feedwater line in Loop B has longer straight runs which leads to larger momentum forces due to valve closure. Therefore, the Loop B feedwater piping subjected to the transient from motor operated valve closure was used in the evaluation.
HYDRO-DYNAMIC LOAD INPUT The propagation of pressure pulses generated by closure of the MOV and FCV has been determined. Feedline fluid velocity gradients (0)* were calculated utilizing the method of implicit characteristics for one-dimensional fluid flow. These velocity gradients were provided as forcing functions for use in the Piping Structural Analysis.
The analysis of MOV and FCV closure was conducted in several steps as follows:
A. Generate San Onofre Unit 1 feedwater system model as shown in Figure 1.
B. Simulate steady-state operating conditions utilizing feedline model with steam generator modeled as constant pressure boundary, and feedwater pumps as constant velocity boundary.
v is the mean fluid time rate of change of mass flow.
C. Derive valve flow discharge coefficient, CD, for valve performance characteristic from valve velocity coefficient, CV, tables for MOV and FCV provided by SCE. Derive pump discharge pressure correlation in terms of discharge velocity and discharge coefficient.
D. Determine discharge velocity correlation for three (3) individual MOVs from simulation of MOV closure on feedline model Figure 1, assuming MOV performance as determined in Step C. Determine correlation by least squares polynomial curve fit.
E. Repeat Step 0 above for FCV closure for three (3) individual FCVs.
F. Generate San Onofre Unit 1 feedline loop models for Loops A, B and C as shown in Figure 2. Loop models include all piping and FCV from MOV discharge to steam generator feedline inlet nozzle.
G. Determine velocity gradients for the nodal locations (see Figure 2) for feedline Loops A, 8, and C for FCV closure. Calculations were based on constant pressure boundary at steam generator and velocity boundary at the FCV as defined in Step 0 for loop models of Step F.
H. Repeat Step G for MOV closure in Loops A and B.
Applicability of the calculated velocity gradients was substantied by analysis of the calculated pressure pulse propogation due to valve closure.
The pressure pulse results were verified by hand calculation based on the "Formula of Michaud."
The hand calculation approximated a linear fluid velocity. The results were verified to less than 6 percent by hand calculation of the peak pressure for MOV closure. Approximation by hand calculation verified the period of the pressure pulse to within 18 percent. Additional verification of results is presented in the two-to-one ratio of the period of the pressure pulse in Loop B (0.22 sec) to the period of the pulse in Loop A (0.11 sec).
Loop 8 (202 ft) is twice as long as Loop A and C (100 ft).
Hence the periods are proportional to loop length.
The peak pressure pulse generated by FCV closure calculated by the analysis had an amplitude of 0.5 psi in Loop B. This pressure amplitude is approximately an order of magnitude larger than the pressure peak amplified of 0.06 psi in Loops A and C. The peak pressure amplitude occursimmediately following valve closure.
The pressure fluctuations damp out in four (4) seconds.
These results indicate that feedline FCV closure has negligible impact on piping loads.
MOV closure generates a peak pressure amplitude of 19 psi in the long length Loop B. This peak pressure amplitude in Loop B is approximately twice the peak pressure amplitude in Loop A of 10 psi.
These peak pressure applitude results are utilized in the piping structural analysis described below.
PIPING STRUCTURAL ANALYSIS MODEL AND METHODS The feedwater piping/support system of Loop 8 was modeled by WESTDYN code (1) and analyzed using time history modal superposition techniques. The model contains the steam generator B and their supports, the stiffness effects from the rest of the reactor coolant loops and Loop B steam line, and the Loop B feedwater piping/
support system inside the containment with anchor at the containment penetration.
The model is a 3-D piping model that includes the non-linear gap effects at piping and steam generator supports. The piping system definition and support stiffnesses were obtained from drawings provided by Southern California Edison.(2) The steam generator/support submodels were derived from a report in reference (3).
The time history forcing functions were calculated using the FORFUN (1) computer code by integrating the time rate of change of the mass flow rate (v) along the pipe centerline. The hydraulic parameter (a,) is described in the previous section.
The time-history forcing function, then, was input into piping structural analysis computer codes WESTDYN and FIXFM3 (4) for the piping analysis by applying an unbalanced force along each straight pipe segment.
ANALYSIS RESULTS AND CONCLUSIONS The results from the piping analysis, bending and torsional moments in the pipe, was substituted into the ANS B31.1 Code equations for the evaluation of adequacy of the piping. High stresses at a few points in the piping are tabulated with their
'Allowables in Table 1. Calculated support loads and their maximum expected loads are shown in Table 2. From Table 1, the piping stresses are all below their allowables from ANS 831.1 Code. It is important to note that the differences between deadweight stress and deadweight plus transient stress are very small and transient stresses are about only 20% of-the deadweight stresses. From Table 2, the support loads resulting from transient loads are not only below the maximum expected loads but also small compared with the deadweight loads. Also, as shown in Table 3, the maximum displacement for the valve closure transient is 0.01 inches which is smaller than the maximum deadweight displacement about 0.02 inches. Hence, the transient loads resulting.from either of the two valve closures have no significant effects on the piping stresses and support loads in the feedwater piping/support system inside the containment.
REFERENCES:
- 1. Bizzak, R. J., and Batt, T. J., "Manual of Piping Analysis Computer Codes",
WCAP-9256, Rev. 1, Dec. 1978.
- 2. SCE Drawings:
(a) Drawing No. 335342-0 (b) Drawing No. 334538-2 (c) Drawing No. 335343-0 (d) Drawing No. 334539-1
- 3. LaPay, W.S., Lee, Y.S., "Seismic Reanalysis and Design of San Onofre Unit 1 Modified Reactor Coolant System", Vol. I, WCAP-8827, January 1977.
- 4. Vashi, K.M. "Documentation of Selected Westinghouse Structural Analysis Computer Codes", WCAP-8285, April 1974.
WESTINGHOUSE NUCLEAR TECHNOLOGY.DIVISION TTLE PAGE Q.ETAUTHOR DATE CIHiiD. ay DATE CHK*D. By DATE
.. CALC. NO.
FILE NO.
GROUP 0.1 4
L/.A,,
ESOWA 7j
~}
M~l Zo~vw/
Moo. XFb p
-3, P4
-3 NO.
O.AT WU~lIONOUS FOR 5521
66 NGHUSENUCLEAR TECHN&GY D!VIS'ION
.0.
-CALC.
NO IL O.jPiU SQAM CAJ 4 E FCd'/ 1i.Je Lac p A
0 147 10 All VA/IZ-5 ARE
~~LNT R E V. E.
U b J DAT/E' UATcAs
'IPCY DAIL
g~we a..
N.
Figure 3. Flow Control Valve Transient
lips
~
s Figure Moo prae Vav
-ransient
TABLE 1. MAXIMUM PIPING STRESS Allowable Calculated Stress (psi)
Total Stress (psi)
Stress (psi)
NODES (1) Deadweight (2) MOV Transient (1) + (2) 1.2 Sh at KBG-392-1 5431 1232 6663 18000 KB-392-1 5407 1253 6660 18000 KBA-392-1 5394 1253 6647 18000 KBG-392-2 5297 1236 6533 18000 SH-392-1 5149 1255 6404 18000 KBG-392-3 5135 1222 6357 18000
- Where Sh = 15000 psi for load condition under occasional loads.
4-
TABLE 2. SUPPORT LOADS Total Loads Maximum Expect SUPPORTS Calculated Loads (1bs)
(1bs)
Loads*
AT (1) Deadweight (2) MOV Transient (1) + (2)
(1bs)
SW-392-1 400 (EW) 400 (EW) 3200 SH-392-1 1820 (VT) 1820 (VT) 2200 KB-392-1 1970 (VT) 0 1970 (VT) 3750 SW-392-3 and 110 (E-W) 110 (EW) 3000 SW-392-4 KB-392-2 1700 (VT) 300 (VT) 2000 (VT) 4000 KBG-392-1 1970 (VT) 20 (EW) 1970 (VT) 3750 20 EW) 4000 KBG-392-2 1840 9VT) 10 (EW) 1840 (VT) 3750 10 (EW) 2500 KBA-392-1 1930 (VT) 620 (NS) 1930 (VT) 3750 620 (NS) 10500 KB-392-3 1750 (VT) 170 (VT) 1920 (VT) 4000 KBG-392-3 1680 (VT) 210 (EW) 1680 (VT) 3000 210 (EN) 6000 KB-392-4 1140 (VT) 30 (VT) 1170 (VT) 3750 SH-392-2 1200 (VT) 1200 (VT) 1260
- Maximum Expected Load are provided by SCE in a support design package prepared by GRINNELL & BECHTEL.
TABLE 3. MAXIMUM/MINIMUM PIPING DISPLACEMENT FROM-MOV TRANSIENT MAX/MIN DISPLACEMENTS (in)
NODE BETWEEN SUPPORTS EW NS -
VT SH-392-1 and 0.002/
0.009/
0.009/
KB-392-1
-0.004
-0.005
-0.005 SW-392-4 and 0.002/
0.004/
0.010/
K8-392-2
-0.005
-0.002
-0.000 KBG-392-2 and 0.000/
0.001/
0.006/
KBA-392-1 0.000
-0.000
-0.002 KB-392-3 and 0.000/
0.000/
0.0001L/
KBG-392-3
-0.000
-0.000
-0.000