ML20092H566
| ML20092H566 | |
| Person / Time | |
|---|---|
| Site: | River Bend  |
| Issue date: | 02/13/1992 |
| From: | Odell W GULF STATES UTILITIES CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| RBGK-36502, NUDOCS 9202210200 | |
| Download: ML20092H566 (7) | |
Text
8
[
G ULF ST/A TES UTILITIES COME'JA NY ~
....u.. n o., m
. w c..wi e m, u,w.uwtume von At!1 A ( ( ':l 40.
$ 4 tvJk4 3448141 February 13, 1992 RBG-30f>03 File Nos. G9.5, G9.25.1.3 U.S. Nuclear Regulatory Commission Document Control Desk Washington, D.C.
20555 Gentlement River Bond Station - Unit 1 Docket No. 50-458 Please find enclosed an Informational Roport concerning cracks discovered in solds in the hydraulic control unit piping at River Bond Station - Unit 1.
This report is submitted to inform the NRC of those conditions and document GSU's investigation and corrective actions.
gor,lyg
.- 3 o'll
.H Manager - oversight E/
/d
/
/C A /kym cc U.S. Nuclear Regulatory Commission 611 Ryan Plaza Drive, Suito 400 Arlington, TX 76011 HRC Resident Inspector P.O. Box 1051 St. Francisvillo, LA 70775 INPO Records Contor 1100 Circle Parkway Atlanta, GA 30339-3064 Mr. C.R. Oberg Public Utility Commission of Texas
.7800 Shoal Crook Blvd., Suite 400 North Austin, TX 78757 9202210200 920213
['
PDR ADOCK 05000458 S
PDR 1
j
1NTl[ODUCTION On February 28, 1991 hydraulle control unit No.36-21 was found leaking. After closer examination of the llCU, it was noted that the water was spraying out from a circumferential crack in a short, horizontal segment of charging water piping between the accumulator pipe coupling and the charging water check valve, near the toe of the weld (see Figure 1, weld #1).
This was documented on condition report (CR) 910074. The plant was not operating at the time of discovery.
The hydraulic control units (IICUs) are part of the control rod drive hydraulle system. They are used to regulate the pressure and flow of water to and from the control rod drives (CRDs).
There are 145 llCUs at River Bend, one for each CRD. During normal plant operation, the IICUs direct and control cooling water and drive water to the control rods. During a reactor scram, the llCOs provide the water needed for rapid insertion of the control rods into the reactor core via scram water accumulators mounted on the llCU skids.
INVESTIGATION The leaking pipe was replaced and liquid penetrant test (PT) examinations were performed on the remainder of the llCUs (144 units) at the location of the leak (weld #1, Figure #1). Sixteen additional linear indications were discovered and reworked; none were leaking (See Table #1).
A metallurgical evaluation was performed on the cracked pipe nipple. The study concluded that the failure mechanism was most likely due to low cycle fatigue at the toe of the weld. General 111ectric issued Rapid Information Communication Services Information Letter (RICSIL) No,056 to document the problem.
The following action plan was developed for the determination of the root cause:
t Review and evaluation of existing design bases, procedures and other pertinent physical data for errors, omissions or anomatics that could be responsible for or contribute to the problem.
Assessment of the data available from the plant through testing to support the investigation.
Collection and analysis of the data to establish a potential root cause.
Confirmation of the root cause through Deld testing.
L Design and implementation of corrective actions.
Reviews of the existing piping stress analyses and application of design basis transients required for River llend Station do not demonstrate the existence of a low cycle fatigue problem, llowever, using observed piping displacements as input to the computer model, failure at the joint in question could be predicted. Furthermore, since the subject piping operates only after a reactor scram or Technical Specification (TS) rod insertion time test, it was hypothesized that the forcing function responsible for the pipe cracking developed during reactor scrams'or TS time test events instead of during normal operation of the plant, it was later determined that there was no significant risk introduced from single rod scrams.
(
Review of the RilS operating and maintenance procedures has been completedt no obvious anomalies were found that would account for or contribute to the observed condition.
A reactor scram history has been developed to help establish the relationship between pipe cracking and positions of control rods and to provide an empirical baseline for the types and numbers of scram cycles related to the frilures. Currently, the pattern of crack incidence appears to be completely random, occurring on both the east and west sides of the reactor and at several different charging water header and branch line locations. Correlations along other common lines are presently being evaluated. Also, experiences from other llWR/6 plants are being reviewed for applicability to the RBS problem. One other BWR/6 plant (KKL in Switzerland) was found to have nperienced a similar pioblem. Their experiences are being reviewed for applicability to River Bend Station.
Tests to collect piping and structural vibration data were conducted during the reactor shutdown scram and the mid-cycle 4 (MDCY4) outage that followed. Through analysis of this data it has been determined that the cracking is due to piping vibrations caused by a hydraulic transient phenomenon of unknown source in the system, it has also been determined that the vibrations emanate from the charging water header down into the vicinity of the cracking, rather than up from the scram water accumulator. The investigations are continuing.
Based on the above and on the results of a fracture mechanics analysis performed by General Electric, an operability evaluation was prepared. Details of this evaluation are presented in the next section.
OPERAHILITY EVAL,UATION GSU has determined that the plant is operable based on the following:
i 1.
PT examinations of the welds at location #1 have been and will continue to be performed following every full core plant scram from power for 100% of the llCUs.
All indications documented during these inspections shall be reworked / replaced prior to restart of the plant from each scram.
2.
As described below, PT examinations were conduc;cd on the next three welds upstream of weld #1 prior to restart from the MDCY4 outage (see Figure 1).
Note that these welds were determined to be susceptible to cracking from review of available pipe stress analyses. All (100%) of the weld #2 and #3 locations were inspected and a selected sample of the weld #4 locations were inspected (only when an indication was found in either of the corresponding weld #2 or #3 locations). All indications were reworked accordingly and no indications were found in any of the inspected welds at location #4 Table 2 summarizes the results of these inspections.
3.
The forcirs function responsible for the cracking has been determined to occur during full core scrams from power and not during normal operation of the plant.
Also, there is no significant risk introduced from single rod scrams, which are required for Tech Spec control rod scram time testing (with valve Vil3 closed l
o
and no Dow in the charging water riser to the accumulators) or from mode switch scrams after plant shutdown. This is supported by the vibration test data and by the results of the General Electric fracture mechanics analysis.
4.
The fracture mechanics analysis performed by General Electric demonstrates that at least 4 additional full core scrams from power can be tolerated without the development of a through wall crack even when a crack of 0.5" long x 0.0825" deep is assumed to exist (pipe wall thickness is 0.147"). Since the llCUs are subject to general visual inspections on a daily basis by Operations personnel, it is likely that leaks would be discovered and isolated before the pipe had the opportunity to break, Therefore, pipe break can be climinated as a potential consequence. This conclusion is further strengthened by GSU's commitment to rework / replace all welds with known indications prior to restart of the plant from future scrams.
5.
Based on the leak rate of a crack computed in the General Electric fracture mechanics analysis, leakage from a through wall crack in a charging water pipe does not present an operability concern.
6.
The recommendations of GE RICSIL #056, PT inspection of welds and general inspection of clamps, have been reviewed and implemented, as applicable to River Bend Station.
7.
In accordance with RICSIL #056, this is not a safety issue. During normal plant operation, automatic and manual scram capability is maintained even without accumulator pressure due to reactor pressure. The most severe impact in the case of using reactor pressure would be slower control rod scram insertion time (to 2.5
- 3.0 seconds). _The safety significance of this has been reviewed and is acceptable on the basis of the total rod density inserted from all rods into the reactor core as a function of time.
8.
A 10CFR50.59 evaluation was performed. Based on the conclusion in item 7 above no unreviewed safety question has been identified.
CI)NCI.USIONS Cracking of the llCUs at RBS is determined to be caused from low cycle fatigue due to a Ould transient of unknown origin which occurs during full core reactor scrams from power. Cracking typically occurs in the first two or three welds on the charging water line off the accumulators on the llCU skid. This is not a safety issue because automatic or manual scram can be achieved during nonaal operation without accumulator pressure by using reactor pressure. Moreover, during startup, the ability to scram is assured by the commitment at RBS to rework or replace all known indications in charging water pipe welds prior to plant startups.
The investigation to determine the source of the responsible fluid transient and to conceptualize corrective actions is currently in progress. Due to the complexity of the problem and lack of significant precedence in the industry, a firm date for resolution of the problem has not yet been established.
. -. _ _ _ _ _ _ _ =
FIGURE I TypicalIICU Charging Water Piping Arrangernent Crackto n ettoe e tain med ch e gee w aru To we scism veu v120
" h,
I/~h#l[.,#
CRApo!NG WATER HEADER u. a;..
u D
i'~~
chusee wate Accumuste M
one van vtor vsw
^
CHAnctNo WATER RI5ER
- WELD # 1 Watp ~
W4 S(a p
m sa QETML 'M N
/
P
\\'
V113 V115 W5LD ACC.
- -T 1
PT EXAMINATION SEQUENCE and CRITERIA a.
100% examination of weld #1, b.
100% examination of welds #2 and #3, c.
Where weld #2 or #3 revealed ign indication, then wcld #4 was inspected.
TABLE 1-Weld #1 Indication Summary from February '91 CRD NO.
REPAIR 12 21 New support riser installed 12 25 Buff, grind, no weld 12-29 Buff, grind, weld 12-37 New support riser installed 20-09 Weld found satisfactory, no rework-20 21 Buff, grind, weld 20 37 Buff, grind, weld 24-17
- Buff, grind, weld 28-05 New support riser i_nstalled 28-13 New support riser installed 28-25 Buff, grind,' weld 28-45 New support riser installed 32-21 New support riser installed -
36-21*
New support riser installed 44-17
. New support riser installed i
q 48 -New support riser installed 52. New support riser installed -
- through wall crack
TABLE 2 Indication Summary for Weld #s 2,3,4 Following MDCY4 WELD CRD NO.
- 2
- 3
- 4 REPAIR 04 41 X
File, buff, no base metal removed 12-29 X
-X
. Weld repair #3, file, buff #2 12-37 X
File, buff, no base metal removed 20-29 X
X File, buff, no base metal removcd 24-17 X
X File, buff, no base metal removed 24-41 X
File, buff, no base metal removed 28-45 X
File, buff, no base metal removed 32 53 X
File, buff, no base mets! removed 36-25 X
File, buff, no ba'se metal removed 36 X File, buff, no base metal removed 36-49
-X File, buff, no base metal removed 36.-53 X
File, buff, no base metal removed 40-45 X
File, buff, no base metal removed X denotes linear indication reported NOTES:
1.
In accordance with the disposition of CR 91-0409, only the indications found on HCUs 12-29 and 28-45 were stress related. For the purposes of the current investigations, all
-are considered to have been stress related.
f.
2.
No indications were found on any of the #4 location welds.
l l
-,... -