ML20127L555
| ML20127L555 | |
| Person / Time | |
|---|---|
| Site: | Rancho Seco |
| Issue date: | 06/06/1985 |
| From: | Reinaldo Rodriguez SACRAMENTO MUNICIPAL UTILITY DISTRICT |
| To: | Martin J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION V) |
| References | |
| 85-06, 85-6, RJR-85-286, NUDOCS 8506280056 | |
| Download: ML20127L555 (11) | |
Text
--
o
$ suun "TP SACRAMENTO MUNICIPAL UTILITY DISTRICT O G201 S Street, P.o. Box 15830, Sacramento CA 95852-1830,1916) 452-3211 AN ELECTRIC SYSTEM SERVING TH RgF CALIFORNIA 4 JE17 RJR 85-286 gmGli M June 6, 1985 J B MARTIN, REGIONAL ADMINISTRATOR REGION V 0FFICE OF INSPECTION AND ENFORCEMENT ATTN: DOCUMENT CONTROL DESK U S NUCLEAR REGULATORY COMMISSION WASHINGTON DC 20555 DOCKET NO. 50-312 LICENSE NO. DPR-54 SPECIAL REPORT NUMBER 85-06 The Sacramento Municipal Utility District hereby submits Special Report Number 85-06 to the Nuclear Regulatory Commission as an itam of interest.
On April 21, 1985, during routine preventive maintenance on the 3A switch-gear bus, eight bolts connecting the incoming line to the breaker were identified as defective.
These bolts were found with their heads nearly or completely sheared off.
To investigate this incident the following actions either were, or will be, taken:
(1) The Plant Review Committee (PRC) requested an analysis of the incident by the Electrical Maintenance Department before the initial reactor startup following the current refueling outage. The analysis report (attached) was reviewed by the PRC on May 23, 1985.
The PRC concluded that the results indicated the failure was limited to these specific bolts and, with their replacement, there are no safety concerns.
(2) The Assistant General Manager, Nuclear directed that an interdepart-mental task force be formed to thoroughly review this incident for generic implications and root cause. The results of this task force study will be forwarded to the NRC for review when completed. Addi-tional corrective or investigative measures may be implemented after consideration of the interdepartmental task force recommendations.
8506280056 850606 PDR ADOCK 05000312 S
PDR 1
I l
3-D/
T 4
%s J. B. Martin June 6, 1985 4
).
The foregoing actions were reviewed and approved by'the Management Safety Review Committee (MSRC) on June 3, 1985.
~
Ifyouhaveany. questions ~ concern'ingthisShecialReport,pleasecontact Mr. Ron Colombo (916 452-3211, ext. 4236) at the Rancho Seco Nuclear Generating Station.
\\
I
+
a R. J. R RIGUEZ ASSISTANT GENERAL MANAGER NUCLEAR ENC cc: Director, NRR L-
y.
SACRAMENTO MUNICIPAL UTILITY DISTRICT OFFICE MEMORANDUM TO: PRC Members DATE: May 23, 1985
'nkhart/B. Wichert FROM: C.
i
SUBJECT:
SILICON-BRONZE BOLT FAILURES
Reference:
ODR #85-111
SUMMARY
A failure has occurred in the 3A switchgear involving Silicon-Bronze bolts. The failure appears to be caused by intergranular stress corrosion cracking. All of the bolts of the same size, type, and service used in the 3A switchgear have been replaced.
An inspection effort checked 100% of the remaining Silicon-Bronze bolts in the 3A switchgear, all the Silicon-Bronze bolts in the 3B switchgear and a large number of additional Silicon Bronze bolts installed in various areas of the plant.
No additional failures have been identified.
Based on the absence of other failures, we conclude that the problem is isolated to a small area of the 3A switchgear. We recommend that plant startup from this outage not be made contingent on any further work, inspection, or analysis of the bolt failures.
FAILURE Under direction of our electrica1' maintenance group, an NPS electrician was wiping down the exposed busbars of the 3A
-switchgear on Sunday April 21, 1985. When he wiped over some bolted connections from the incoming bus from X43A to breaker 3A05, several bolt heads broke off. The electrician was using lint free rags and no solvents or cleaning agents. See f.igure 1 attached for the configuration of the bolts which failed.
An initial direction was given to the electrical maintenance group to emove all of the bolts on the affected connection to inspect for damage.
In the process of removing the remaining bolts some more were broken in the act of untorquing them.
Five of the initial twelve bolts were left unbroken efter removal.
These five were looked at for visual signs of cracking.
After finding no indications on four of the five, they were reinstalled along with eight new bolts. Two of.the original four replacement bolts broke when retorqued.
At this time, all remaining original bolts were replaced with new bolts.
a,-
PRC Memb2rs Page Two.
May 23, 1985 The broken bolts were identified as 11 inch long, 3/8 inch diameter, 16 threads per inch (National coarse thread) Silicon-bronze manufactured by the Coast Industrial Supply Company.
They failed at the junction of the head to the shank of the bolt.
Four of the failed bolts were sent to a metallurgical laboratory for analysis.
USE OF SILICON-BRONZE BOLTS AT RANCHO SECO The 3/8" bolts which failed in the 3A switchgear were installed during original plant construction.
Silicon-Bronze i ardware is commonly used to make electrical connections and was used extensively in the construction of Rancho Seco. More than 75 were used to make field connections to the 3A bus.
Splices at all 480 volt and 4160 volt motors are made with Silicon-Bronze bolts.
Over 90%
of equipment grounding connections are made with 3/8" Silicon-Bronze bolts. Approximately 80% of the Silicon-Bronze bolts currently installed were manufactu-ed by Coast Industrial Supply Company.
The vast majority of silicon-bronze bolts used at Rancho Seco are 3/8" or 1/2" diameter.
They are almost always installed with no lubricant. Occasionally, on an aluminum to copper connection or on a battery connection, a corrosion inhibitor such as NO-0X ID is used on the connection which provides some lubrication for the bolt threads. Those which failed were installed dry.
Torque values which should be used on these bolts are 15-20 ft-lbs for 3/8" and 25-35 ft-lbs, for 1/2". There have been no records found which indicate to what torque the failed bolts were installed.
Silicon-Bronze bolts are used in all areas of the plant.
Those that failed were installed inside of a metal panel in a switchgear room 1.n the Aux. Bldg. Many Silicon-Bronze bolts are located outside exposed to the weather.
They are used inside containment; near acid, caustic, and ammonia systems; in open air; and in completely sealed applications.
Until now, no fail'res of Silicon-Bronze bolts in the head to shank u
transition area have been documented at Rancho Seco.
The manufacturer has been contacted and reports that their Silicon-Bronze hardware is in use all over the west coast.
They report that no failures of this type have come to their attention.
INSPECTION PROGRAM Soon after the discovery of the failed bolts, an inspection program was defined for the rest of the 3A switchgear.
Every Silicon-Bronze bolt in the 3A bus was untorqued, looked at, and retorqued.
The 3A bus itself is held together with steel bolts, a random sample of which were untorqued and retorqued.
(It was initially believed that the failed bolts might have been torqued beyond their yield point during a past inspection of 3A bus bolt tightness, so some of the steel bolts were included in the inspection).
PRC Members Page Three May 23, 1985 When we received the preliminary report that the bolts failed due to stress corrosion cracking, the inspection effort was expanded.
Samples were taken from the same room as the 3A switchgear location and sent off for laboratory examination. The 3B bus was subjected to the same 100% untorquing and retorquing of all its Silicon-Bronze bolts. Samples of 3B bus bolts were sent for laboratory exantination.
A widespread visual inspection of Silicon-Bronze bolts was initiated.
Bolt tightness was checked by verifying that the split ring lock washer was flat. Where possible, the terminated cable was pulled on to put extra momentary strain on the bolt. Bolts were looked at in the tank farm, the transformer yard, battery chargers, inverters, DC buses, motor control centers, near the aux. boilers, and the CRD system.
No failures were found.
All bolts appeared to be torqued.
Samples were taken from within 15 feet of the ammonia storage tank and sent to the laboratory for examination. These are thought to be from the harshest environment at the plant.
LABORATORY ANALYSIS In addition to the samples taken from service in the plant, a number of Silicon-Bronze bolts from stock were also sent to the laboratory.
A preliminary report from the laboratory has been received and is attached.
CONCLUSIONS The laboratory analysis of the failed bolts concludes that the probable cause is intergranular stress corrosion cracking (IGSCC). The corrodant suspected is ammonia. Also noted is a difference in grain structure between the bolts which failed and the samples sent which had not failed in service. The difference was that the grains in the failed bolts were aligned perpendicular to the surface while the grains in unfailed bolts were aligned parallel to the surface. The difference may have increased the sensitivity of the failed bolts to IGSCC.
It is our conclusion that the failed bolts suffered from a difference in material properties from the normal Silicon-Bronze bolt. The extent to which this material defect affects Rancho Seco cannot be determined exactly without analyzing every bolt installed.
- However, the facts that no other such failures are known of, samples taken from service do not show the same defects, and hundreds of bolts have been looked at or retorqued with no failures found support the conclusion that the 3A switchgear failures were an isolated case.
The investigation will continue, but it is our recommendation that plant startup not be held by further work on the Silicon-Bronze bolt failure.
CL:pr cc:
F. Sheehan J. Sullivan i
B. Daniels G. Coward
e y
e 9
9 4
\\
II Il ll l
1 l
l F-s s.
l I's ss l l l
s l
l
'N N 1l l
I s.
I O d l
Q~ ~) R l l9'Ks i
l ii O
ll l
D g
'y II
's II l
s A_]><_-
s s_s
\\t
/
r r,
v[J II]#]~' l ]
s ll s
s--
\\
l
's Il ss ss i
l !
e '. '3 ')
^
'N \\
- s
'xs.
N
- s. ss s
N. \\
N N s Figure I
9.
l ANALYT4 CAL Anamet La ao ratories, I n c.'
l1**1,CA<
j e4850 See 5778 M84N TENPERATURE 3417 SEVENTH STAEET SE AKELEY. CALI(CANIA 94 710 APPLIED RESEAACM PNYSICAL TESTitee May 22, 1985 i
i i
LABORATORY NUMBER:
F585.54
SUBJECT:
PreliminaryReportonInvestigation!
i of Failed Silicon-Bronze Boltu 1
I i.
DATE SUBMITTED:
May 6, 1985 s.
REPORT TO,:
Sacramento Municipal Utility District G201 S Street P.O. Box 15830 l..
Sacramento, California 95852-1830 Attn:
Mr. R.P. Wichert Senior Mechanical Engineer l
ff /
l II /
t /-
"/'.4
'.y l
' 1 "-
REPORT
/-
- k..
,y j
t
..j.h,hisreportdiscussesthepreliminary_.findingsofafailure l
,r
~
e s
analysis of silicon-bronze bolts used as fasteners for ele'ctri-e cal bussbars.
The results of a determination of th'e fracture i
. mode, some materials testing, and a discussion of preliminary findings are included.
Bolts from service and from stock were submitted for this ex-i amination.
A description of all the samples received is pra-l[
sented in the following tables.
From Service
- i 4 - 3/8 x 1 1/2 UNC Fractured i
1 - 3/8 x 2 UNC 2 - 3/8 x 1 1/4 UNC Prom harshest corrosive environment in plant l
~
i 2 - 1/2 x 1 1/2 UNC One from same room as failures
- All bolts from service were supplied by Coast Industrial Supply (Coast)
From Stock
[
2 - 1/2 x 1 1/2 UNF (Coast)
L93 ORATORY CERTiflCATC l
. Anamet Laboratories, Inc.
Lab. Not FS85.54 2*27 SEvChiet 51 ACCT. SEAKELEV. CAL &lORNLA 94780 1 - 1/2 x 1 1/2 UNC (ddast 3 - 1/2 x 1 1/2 UNC (Supplier 2) 1 1 1/2 x 1 1/2 UNC (Supplier 3) 4 - 3/8 x 2 UNC (Coast) i 4 - 3/8 x 1 1/2 UNC (Coast) 4 - 3/8 x 1 UNC (Supplier 4) i EXAMINATION AND TESTING:
Visual and Fractographio Examination.
The four failed bolts fractured at the underhead fillet.
The fractures were macro-scopically brittle with no vicible evidence of a shear lip or plastic flow at the fracture.
The radius at the underhead i
fillet on the fractured bolts was about 3/64 in, which was i,
not considered to be excessively sharp for a bolt this size.
No evidence of significant corrosion damage was observed,on the l
fractured bolts.
The original surfac'a.was only slightly j
tarnished, and the fracture. surfaces.nere relatively bright.
f l
The fracture surfaces were examined with the scanning electron l'-
microscope.
This examination revealed a predominantly inter-granular fracture mode.
Intergranular fracture in this materi-al is g'enerally found only due to environmentally-assisted l
I i
l fractures, such as stress corrosion cracking.
I f
The other bolts removed from service were examined visually and by dye-penetrar.t inspection.
No evidence of cracks were ob-served in any of these bolts.
Many of the other bolts from service had sharper radii at the underhead fillet than the failed bolts.
I Metallographic Examination.
Longitudinal / radial sections of one of. the fractured bolts and two bolts from service (3/8 x 1 1/4 in, and 1/2 x 1 1/2 in.) were metallodraphically pre-pared.
Materials from all three bolts exhibited similar micro-structures.
The single phase microstructures observed for the bolts was consistent with the silicon-bronze alloy specified.
l',
i e
n
.u---
-.-,w.,-.
f Lk% ORATORY CERTIFICATE s-Anamet Laboratories, Inc.
Lab. No:
es8s.54 men sevepm stacu. scance cmroam suso Deformed grains in the head area and at the threada indicated l
that the heads had been cold formed and the threads had been i
rolled".
In the fractured bolt the grain flow due co. head forming was approximately perpendicular to the longitudinal axis at the surface in the underhead fillet area.
Thus, at the surface, the grain f1'ow was perpendicular to the applied stresses and parallel to the fracture.
In the boles that were not fractured, the grain flow followed closely to che fillet profile, which results in grain flow more parallel to the ap-plied sti' esses.
5 Chemical Analysis.
Chemical analysis was performed on material j-from three bolts removed from service:
A failed bolt (3/8 x 1 1/2), a.3/8 x 2 in. bolt ind a 3/8 x 1 1/81.in. bolt.
The bolt material was expected to be 1% silicon bronze alloy speci-fied by ASTM B 98 Alloy B (UNS Alloy.C6.5100).
Chemical.analy-ses indicated that the material from a.ll three bolts corre-
~
sponded more closely to the Alloy A designation (UNS Alloy C65500), which has a nominal silicon content of 3%.
t DISCUSSION OF PRELIMINARY FINDINGS:
The intergranular fracture mode indicated environmentally as-sisted fracture, probably stress corrosion cracking.
For l
copper alloys, environments containing ammonia are the most l
l common cause of intergranular stress corrosion cracking.
Am-l monia tanks are reported to be in the area near the building l
where the failures occurred, although exposure of these bolts t.
to significant concentrations of ammonia is considered unlike-i ly.
Other environments identified as stresa corrosion agents for copper alloys include:
Amines, chelate compounds, sulfur dioxide, and polyethers.
Moisture, oxygen, and possibly carbon dioxide, have been reported to aid in stress corrosion cracking, but cracking would not occur in the presence of these compounds alo'ne.
The bolt material was expected to be a silicon-bronze alloy,,
I LA". ORATORY CERTiflCATE
~
Anamet Laboratories, Inc.
bab. No:
rsas.s4 3*2 F SEVENTM STAECT. SE AEELEY. calif 0ANIA 94 FIO UllS Alloy-C65100' (nominally 1% silicon).
Chemical analysis indicated that the material composition corresponded more closely to yNS Alloy CSSsGO, a 31, silicon-bronze.- Among the copper alloys, the silicon-bronzes have relatively good re-sistan'ce to ' stress corrosion cracking.
Some field studies have reporte'd that the higher silicon materials are more susceptible to stress corrosion than the 1% nominal silicon alloys.
Metallographic examination revealed an undesirable grain flow in the failed bolts.
The grain flow (frbm cold forming of the headsr was perpendicular to the longitudinal axis at the surface at the underhead fillet'uhere fracture occurred.
This configuration results in applied tensile stresses perpen-dicular td the direction of grain flow, and the material will have a lower strength in this direction than in the direction l
The lower strength and the higher parallel to the grain flow.
i j
concentration of grain boundaries exposed E.t tha' surface may I
facilitate stress corrosion.
Un-cracked bolts exhibited a grain flow thac followed more l
closely to the profile of the fillet, which results in a grain ~
I flow more parallel to the direction of the apIslied stress.
Thus, the grain fl'ow configuration of un-cracked bolts may have,resulted in a reduced susceptibility to stress corrosion as compared to the failed bolts.
The stress corrosion failures may also be due to higher stresses applied to the' failed bolts.
For example, excessive tightening l
of-the nuts on these bolts may increase the effective stress above a ' threshold for stress corrosion cracking, whereas proper torqueing would result in stress below this threshold.
IU TESTING.AHDipROGRESS:
?
Accelerated stress ' corrosion cracking testing is in progress on bolts submitted from SMUD stock.
Stress is applied by tightening the nut against a spacer on the bolt.
An ammoriia atmospbereisgeneratedfromammoniumhydroxide.
Water vapors
_y.
6
a m u u a w m uu m u s
'Enamet Laboratories, Inc.
Lab. nos rs8s.s4 3*.3 7 SEVENTH STAEET. SEAKELEY. CAuFOANIA 94780 are added to the environment as well as occasional additions of carbon dioxide.
This test has been in progress five days with no failures to date.
At this time a more elaborate test program is also under.de-velopment.
This program would allow better control of' the stress corrosion cracking perameters, stress and environment.
In addition. bolts from stock will be examined following stress i
corrosion testing to determine the grain orientation at the underhead fillet.
9 Respectfully submitted, ANAMET LABORATORIES, INC.
I i
By Larry _D. Hanke Metallurgist
..e Approved byGordon E. Lakso Chief Metallurgist 3c mit i
I f
l i
e i
r