ML13333A376
| ML13333A376 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 12/31/1977 |
| From: | Southern California Edison Co |
| To: | |
| Shared Package | |
| ML13333A375 | List: |
| References | |
| NUDOCS 7906200351 | |
| Download: ML13333A376 (5) | |
Text
REPORT CONCERNING SUSCEPTABILITY TO STEAM GENERATOR FEEDWATER LINE WATER HAMMER EVENTS SAN ONOFRE NUCLEAR GENERATING STATION UNIT 1
- December, 1977 9 062Oo S\\
REPORT CONuERNING SUSCEPTABILITY TO STEAM bENERATOR FEEDWAJD LINE WATER HAMMER EVENTS INTRODUCTION This report discusses the results of an extensive review of plant operating history related to water hammer in the steam generator feedlines and provides specific responses to items of interest described in the September 2, 1977 letter from Mr. A. Schwencer to Mr. J. Moore, in docket 50-206.
REFERENCES (1)
February 8, 1974 letter from R. N. Coe to R. H. Engelken regarding Regulatory Operations Request 74-1.
(2)
July 14, 1975 letter from K. P. Baskin to Office of Nuclear Reactor Regulation regarding response to May 13, 1975 request for information.
(3)
September 2, 1977 letter from A. Schwencer to J. B. Moore.
(4)
June 13, 1975 letter from J. B. Willis of Westinghouse to J. G. Haynes regarding Water Hammer in Steam Generator Feedwater Lines. (Technical Bulletin NSD-TB-75-7).
DISCUSSION A thorough review of operating history as it relates to steam generator water hammer was conducted. It was noted that two reports of this type of water hammer were previously reported to the NRC. (References 1 and 2.)
In both of these cases it was noted that the damage reported was a result of design errors and/or improper installation rather than an overstress condition. The first event resulted in a cam follower on a feedwater regulating valve positioner slipping off its cam. This design deficiency was resolved by replacing the positioner with a different type which has the cam follower securely attached to the cam. The second event involved discovery of several steam and feedwater pipe support failures during a unit outage. It was initially speculated that the supports may have been damaged by water hammer. However, a review of the original design calcula tions revealed that lateral loadings had been neglected in the final support designs. Also, it was found that the support anchor plates were improperly installed. Thus, in over 10 years of operation, there has never been a steam generator water hammer which would have resulted in damage to the plant if the original design criteria had been applied correctly to the valve positioner and piping supports. Moreover, the repairs and modifica tions made as a result of these two incidents have successfully prevented recurrences.
Informal surveys were conducted of several operators who have been assigned to Unit I since initial startup. They recalled several non-damaging water hammers during the early years of plant operation. None of these events resulted in damage to pipe supports or pipe insulation, pipe displace ment, or failure of pipes or components. Also, several improvements in the plant design and operating procedures have eliminated even these minor events.
40,.
-2 Specifically, the installation of auxiliary feedwater regulators greatly improved the ability to make the necessary fine adjustments in flow required for low load/off line level control.
However, leakage through the main feedwater regulators still presented some problems.
Therefore, motor operated block valves were installed upstream of the main regulators to assure positive shutoff. In 1975, the mechanism for steam generator feedline water hammer was identified and brought to Edison's attention by Westinghouse (Reference 4).
As a result, several changes were made in operating procedures to minimize the probability of creating potential-water hammer conditions. The operators were cautioned to attempt to maintain level above the feedring at all times and to add feedwater slowly in the event the feedring did become uncovered. No water hammer events of any observable magnitude have been experienced since these changes have been incorporated.
Another facet of the review of operating history included an inspection of steam generator level/feedwater flow charts during startups, shutdowns, and after unit trips from 1967 through the present. These charts contain virtually all situations that have occurred which may have produced the conditions for water hammer (i.e., steam generator level below the feedring in combination with low feedwater flow rate of sufficient duration to cause partial draining of the feedring).
Load runbacks and other load changes do not normally result in conditions which could result in water hammer because of the high feedwater flow rates which maintain a full feedring.
It was noted that level has fallen below the feedring on numerous occasions.
These occurrences can be divided into two broad categories:
- 1) Steam generator level drops rapidly after a unit trip, almost always uncovering the feedring. The main feedwater pumps continue to run (except in the event of loss of offsite power) and the main feedwater regulators are positioned automatically according to average reactor coolant temperature (T avg.).
If T avg. is above 545 0F the valves will continue to be controlled to maintain the normal 30% narrow range level set point. When T avg. falls below 545 0 F the valves are automatically positioned to control feed flow at 5% of full load feed flow unless the high level override is actuated (at 90%).
Normally, the above control scheme results in a momentary level drop to near 0% followed by rapid recovery to above 30%. At this point the valves are placed on manual control and level is maintained near 30%. Since the top of the feedring is at 26%, water level drops well below the ring under these conditions.
Operating experience has indicated that water hammer does not occur under these circumstances. This may be due to continued feedwater flow keeping the feedring full during the low level transient. Also, since the reflood is accomplished with relatively hot feedwater, the potential for rapid steam bubble collapse in the event of partial feedring drainage is reduced due to the low temperature difference between the steam and incoming feedwater.
-3
- 2) Plant operating instructions require that feedwater control be placed on manual after unit trips and below 20% load during startups and shutdowns to minimize feedwater flow instabilities.
The second type of uncovering event occurs when level control is on manual and some system perturbation such as steam dump actuation, turbine roll, or a change in the rate of steam generator blowdown causes a level fluctuation below the feedring.
These events may take place when feedwater temperatures are low and feedflow is below that necessary to maintain a full feedring. Under these conditions, the instructions specify that the feed rate during recovery be slow such that the probability of water hammer is reduced.
Even though the feedring uncovering events of this type have been numerous, audible water hammer has rarely occurred and "damaging" water hammer events have never occurred.
The absence of damaging water hammer may be attributed to the short horizontal piping runs leading to the steam generators.
Steam generators A & C have a 14" x 10" reducer, a horizontal 10" x 450 ell and a downward sloping 10" x 450 ell connected to the 14" steam generator nozzles. Steam generator B has only a 14" x 10" reducer and a downward sloping 10" x 450 ell.
- Thus, the total horizontal dimensions from the steam generator nozzles to the centerline of the downward sloping 450 ells are 31 3/4" on steam generators A and C and 19 1/4" on steam generator B. These short horizontal runs minimize the potential size of the steam bubble which can initiate a water hammer event, thereby minimizing its potential magnitude.
It is therefore concluded that major plant modifications such as installation of top discharge spargers and plugging the bottom discharge holes are not warranted at this time. However, feedring uncovering events can be essentially eliminated during normal startups and shutdowns by procedural modifications as described in the following responses to the items of concern as described in the September 2 letter. (Reference 3).
ITEM (1) "Provide proposed plant design and/or procedural modifications, if any, which are necessary to assure that the feedwater lines and spargers remain filled with water during normal as well as transient operating conditions. Supporting evaluations and analyses, and any necessary procedural testing, should be provided to demonstrate the system performance. We recognize that there may be various design and/or procedural options available which will satisfy this objective and that, in some cases, existing plant designs may already incorporate such provisions.
For example, we believe that provisions such as the use of top discharge feedwater spargers, and automatically initiated operation of auxiliary feedwater systems, could maintain the feedwater lines and sparger full of water. In addition, it appears that steam generator feedwater system design characteristics may, in.some cases, inherently maintain a filled status by virtue of the elevation of the sparger, the closure rate of the feedwater control valve, prompt availability of auxiliary feedwater, etc. Proposed modifications will be reviewed and approved on the basis of their effectiveness to reliably maintain the feedwater lines and spargers filled with water on a plant specific basis."
RESPONSE TO ITEM (1):
The following procedural modifications will be made to essentially eliminate conditions which cause the steam generator feedring to be uncovered during startups and shutdowns.
- 1) While the feedwater controls are on manual and load is below 20%, steam generator level will be maintained at a nominal 50% on the narrow range instrumentation rather than the automatic control set point of 30%. This will essentially eliminate feedring uncovering events which occur as a result of normal level fluctuation during periods of low feedwater/steam flow.
- 2) A precautionary note will be added to the appropriate procedures to point out the necessity for close attention to steam generator levels while on manual control when changes are made to the rate of steam generator blowdown. A review of operating records during start ups and shut downs indicates that changes in blowdown rate cause fluctuations in level which quite frequently result in uncovering the feedring. Blowdown rate is controlled manually at a location remote from the control room. Therefore, a control room operator manually controlling steam generator level must be aware of the timing and magnitude of any changes in blowdown rate to effectively adjust feed rate. All other system perturbations that normally cause level fluctuations are performed by controls adjacent to the feedwater control and should be compensated for by maintaining a higher initial steam generator level.
ITEM (2) "Provide an assessment of your proposed design and procedural modifications in terms of the impact or interaction with the safety design bases and supporting safety analyses for the plant."
RESPONSE TO ITEM (2):
The proposed procedural changes will result in a higher average steam generator level during periods of low feedwater/steam flow when the feedwater controls are on manual.
The automatic level control set point will not be changed and none of the overrides, alarms or trips associated with the system will be affected.
The 30% level set point is designed to maintain level above the feedring while minimizing moisture carryover in the main steam at full load. Since moisture carryover is negligible below half load, maintaining a 50% level below 20% load will not have any adverse affects on moisture carryover.
ITEM (3) "Provide a schedule for implementation of the proposed design and procedural modifications."
RESPONSE TO ITEM (3):
The proposed procedural changes will be incorporated into the appropriate station operating instructions by January 31, 1978.