ML13333B211
| ML13333B211 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 11/20/1984 |
| From: | Cintula T NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| Shared Package | |
| ML13333B210 | List: |
| References | |
| TASK-AE, TASK-T424 AEOD-T424, NUDOCS 8501220490 | |
| Download: ML13333B211 (3) | |
Text
AEOD TECHNICAL REVIEW REPORT*
UNIT:
San Onofre Station TR REPORT.No. AEOD/T424 DOCKET NOS.:
50-206, 361 & 362 DATE:.November 20, 1984 LICENSEE:
Southern California Edison EVALUATOR/CONTACT:
T. C. Cintula NSSS/AE:
Westinghouse/Bechtel Combustion Engineering/Bechtel
SUBJECT:
FIRE WATER MAIN LEAKAGE INTO 4 kV SWITCHGEAR ROOM EVENT DATE:
June 16, 1984 (LER 361/84-033)
SUMMARY
Water entered the 4 kV Switchgear Room at San Onofre-l from an unexpected rupture of an isolated section of the fire main during hydrostatic test ing. The event was considered an isolated case without generic concern because:
(1) Very heavy construction equipment passing over the fire main contributed to the premature pipe failure, and (2) The flow of water followed a path that normally would be sealed to water flow after completion of construction.
No safety-related equipment was affected by this event and the licensee postulated that continued fire main flooding for a prolonged period would not significantly increase the water level in the Unit 1 4 kV.Switchgear Room.
- This document supports ongoing AEOD and NRC activitiesand does not represent the position or requirements of the responsible NRC program office.
8501220490 841120 PDR ADOCK 05000206 S
-2 DISCUSSION At 1142 on June 16, 1984, during hydrostatic testing of a new section of fire main piping for Units 2 and 3 of the San Onofre Nuclear Generating Station, the hydrostatic test boundry valves leaked and pressurized the entire fire main above its operating pressure (125 psig).
The fire main cracked outside of the hydrostatic test boundry. A jockey pump and the three additional fire main pumps automatically started on low system pressure. Localized flooding occurred in the vicinity of the break and water flowed through newly installed telecommunications ducts (that were under construction) into the Unit 1 4 kV Switchgear Room.
No electrical equipment in Unit 1 was affected by the ingress of water. Units 2 and 3 were not affected by the localized flooding.
At 1515, the fire main leak was isolated and system operability was restored.
Investigation of this incident by the licensee indicated that it was unlikely that the fire main piping failed from a simple overpressurization incident.
First, the piping was subjected to a maximum internal pressure of 220 psig for only a few seconds during the hydrostatic test. This was well within the design pressure of the 12-inch diameter, 0.5-inch wall thickness, cast iron, cement-lined yard main. It had been calculated that for a 220 psig internal pressure that only a 0.37-inch wall thickness pipe would have been required. Secondly, it was evident that the piping break occurred in an area where extremely large and heavy construction equipment had been passing over the fire main. Finally, excavation activity on both sides of the fire main may have decreased the compaction of the soil in the vicin ity of the break and may have permitted unexpected vibration of the pipe during the movement of the heavy construction equipment. Because of these considerations, the failure of the piping was attributed to cyclic stress fatigue that was intensified by a moderate increase in pressure from the leaking hydrostatic test boundry valves.
The cause of the leaking isolation valves has not been determined because the licensee is awaiting arrival of replacement parts before disassembling the valves.
The licensee examined the possible consequences to Unit 1 with the conser vative assumption that the fire main flooding had continued for a prolonged period. They concluded that the presumed additional flooding would not have significantly increased the water level in the 4 KV Switchgear Room at Unit 1 because it is surrounded on two sides by a large level area with numerous floor drains at the same elevation. It did not seem possible that water from the cracked fire main could have entered any other rooms at Unit 1 with safety-related equipment, or could have affected Units 2 and 3 which were remote from the construction activities.
-3 The San Onofre Station work procedures require open conduits to be sealed with tape at the conclusion of daily work activities to prevent dust and foreign materials from entering the conduit.
The telecommunications duct was sealed in this fashion at the time of the fire main break, and it was ineffective in preventing the water ingress into the 4 kV Room. In the future, the licensee will require construction crews to seal open conduits with caps at the conclusion of daily work activities to prevent this unantic ipated path for water transport to other areas.
FINDINGS The licensee's explanation that the fire main piping failed at a reduced pressure because of the damaging effect of the passage of heavy construc tion equipment over the pipe is in agreement with accepted formulae for calculating pipe allowable internal working pressures.
Therefore, with continued passage of the heavy construction equipment, it can be presumed the pipe would have eventually failed from normal water pressure if the hydrostatic test boundry valves had not leaked. Since the underground fire main loop typically is a long length of pipe encircling much of the perimeter of the site, it may be important to predetermine locations where heavy equipment will be passing over the fire main piping to prevent unnec cessary weakening of the fire main piping.
The fire water entered a room with safety-related electrical equipment through the newly installed empty telecommunication ducts. Tape at the end of the duct to keep out dust was ineffective in preventing water entry.
The licensee's proposed station-wide future corrective action of capping all open conduits and ducts under construction at the end of the work day specifically.to prevent the passage of water is appropriate. When comp leted, the telecommunications ducts will be sealed with a standard fire barrier cable penetration seal.
While this seal may not be impervious to water penetration, it will reduce water flow so that the rate of ingress of water will be much less than the analyzed probable maximum flood.
CONCLUSIONS The unanticipated rupture of the fire main piping at San Onofre is an isolated case without apparent generic concern. The licensee's proposed corrective action of capping open penetrations under construction to prevent water penetration is appropriate and should be considered as a good practice for the industry.