IR 05000251/1987016
| ML17347A541 | |
| Person / Time | |
|---|---|
| Site: | Turkey Point  |
| Issue date: | 05/15/1987 |
| From: | Herdt A NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II) |
| To: | |
| Shared Package | |
| ML17347A539 | List: |
| References | |
| 50-251-87-16, NUDOCS 8706160097 | |
| Download: ML17347A541 (100) | |
Text
7 1P,S RED(
<C+ 1
co O
o
'P**d ~
UNITEDSTATES NUCLEAR REGULATORY COMMISSION
REGION II
101 MARIETTASTREET, N.W., SUITE 2900 ATLANTA,GEORGIA 30323 Report No.:
50-251/87-16 Licensee:
Florida Power and Light Company 9250 West Flager St.
Miami, FL 33102 Docket No.:
50-251 Facility Name:
Turkey Point Unit 4 Inspection Conducted:
March 19 - May 5, 1987 Team Members:
J.
Blake D. Brewer P. Burnett K. Clark B. Elliott W. Kleinsorge J.
Macdonald J.
Menning W.
Ross K. VanDyne B. Wilson Approved by: I ly QOH A. R. Herdt, Team Leader Division of Reactor Safety License No.
DPR-41 Da e Signed Scope:
This special, announced augmented inspection was conducted to monitor the licensee's response and to review the circumstances associated with a
problem identified by the licensee with corrosion caused by deposits of crys-talline boric acid on the reactor vessel head and surrounding areas.
The areas inspected included the sequence of events, effects of failure, metallurgical aspects, chemistry aspects, corrective actions planned, and safety considera-tions for station restart.
Results:
Three violations were identified.
Failure to properly evaluate the leak in terms of the boric acid corrosion of ferritic steel components (para-graph 7b); Failure to properly adhere to the installation and drawing require-ments of the conoseal (paragraph 12);
Leak rate procedure inaccurate in that correction factors were incorrect (paragraph 13)
87Osfeao97 87os>5 FDR ADOCK 05000251
r t P
REPORT DETAILS Persons Contacted
- C. 0.
Woody, Group Vice President
- T. W. Dickey, Vice PResident
- Nuclear Operations
- C. Wethy, Vice President - Turkey Point
- C. J. Baker, Plant Manager - Nuclear - Turkey Point
- J. Hays, Director of, Nuclear Licensing
- H. Paduano, Manager Nuclear. Services
- E. Preast, Project Manager
- D. Chancy, Site Engineering Manager (SEM)
- P. Pace, PTN, Licensing Supervisor, Corporate S. Goliard, Section Supervisor, Codes and Programs
- J. Donis, PTN, Site Engineering Supervisor T. C. Grozan, Nuclear Licensing, Sr. Specialist
- J. Arias, Jr. Regulatory and Compliance Supervisor D.
W. Haase, SEG Chairman
- M. g. Crisler, gC Supervisor
- F. H. Southworth, Maintenance Superintendent
- Nuclear
- R.
H. Hart, Licensing Engineer
- D. Grandage, Operations Superintendent P. Salkeld, Plant Supervisor - Nuclear J.
A. Labarraque, Technical Department Supervisor R. L. Stone, gA Engineer M. O'Meara, Engineering Department R. Longtemps, Mechanical Maintenance Supervisor
- J. Kappes, PEP Coordinator D. Ingram, Mechanical Maintenance Department R. Earl, gC Inspector S. Flynn, Maintenance Foreman, Special Crew Other licensee employ'ees contacted included craftsmen, engineers, techni-cians, operators, mechanics and electricians.
Other Organizations BSW H.
W. Behnke, Advisory Engineer R.
D. Shipley, Supervisor Engineer Westin house R.
E. Tome, Engineer, Reactor Vessel R. Stapleton, Engineer, Reactor Internals D. E. Boyle, Manager, Primary Systems
C
J. DeBlasic, Safety Engineer E. J. Rusnic, Manager Mechanical Eg K. Voytell, Safety Engineer D.
C. Marcharger, Safety Engineer D. Richards, Project Manager Stone and Webster J. Fleming, Level II Inspector US Nuclear Re ulator Comoission
- J. N. Grace, Regional Administrator, Region II
- M. L. Ernst, Deputy Regional Administrator, Region II
- J. H. Sniezek, Deputy Director, NRR
- G. C. Lainas, Assistant Director, Region II Facilities, NRR
- D. G. McDonald, Project Manager, NRR
- Attended exit interview on May 5, 1987 Exit Interview The inspection scope and findings were summarized on May 5, 1987, with those persons identified in paragraph 1 above.
The inspectors described the areas inspected and discussed in'detail the inspection findings listed below.
No dissenting comments were received from the licensee.
a.
Violation 251/87-16-01, Failure to properly evaluate the leak in terms of the boric acid corrosion of ferritic steel components (paragraph 7b)
b.
Violation 251/87-16-02, Failure to properly adhere to the installa-tion and drawing accuracy/control of the conoseal (paragraph 12)
c.
Violation 251/87-16-03, Leak rate procedure inaccurate in that correction factors were incorrect (paragraph 13)
The licensee did identify some material as proprietary during this inspec-tion, but this material is not included in this inspection report.
Unresolved Items There were no unresolved items identified during this inspection.
Augmented Inspection Team (AIT) Activities An Augmented Inspection Team (AIT) was formed and dispatched to the Turkey Point site on the morning of March 19, 1987, to review the circumstances associated with an instrument seal leak identified by the licensee with possible corrosion caused by deposits of crystalline boric acid on the reactor vessel head, associated components and surrounding areas.
The
team arrived on site about 3:00 p.m.
on March 19, 1987.
The team met with plant management and staff to assess the operational status of the unit and NRC inspection assignments were outlined.
Florida Power and Light (FPL) agreed to provide any assistance required by the team.
In addition, the ground rules to be applied by the inspection team regarding quarantine of equipment and components were discussed.
Florida Power and Light agreed to seek NRC concurrence before any work was accomplished for restoration of systems.
The team conducted inspections through March 24, 1987, to ascertain the causes and affect of the instrument seal leak.
The AIT did not conclude its inspection at that time due to the ongoing activities by the licensee to develop a root cause analysis, engineering evaluations, and any necessary repairs, which required subsequent inspec-tion activities.
Onsite AIT activities continued until April 23, 1987.
Media Interest in the Event Media interest in events at the plant was already high because of two previous unrelated problems when technical members of the NRC's AIT arrived at the site.
The unit had been placed on hot standby a
week earlier while a minor problem with a valve on the inner door of the personnel air lock was fixed and was held down while an isolation valve in the containment purge system was calibrated.
Both events resulted in extensive media scrutiny and set the stage for intense coverage of the boric acid problem.
An NRC Public Affairs Officer was dispatched to the site from Region II on Friday, March 20, to assist the technical team with the growing number of media information requests.
Numerous television, newspaper, wire service and radio talk show interviews were held over the weekend.
The NRC Public Affairs Officer kept the licensee s information office informed of materi-al released to the media by the NRC and arranged for a joint news confer-ence in the corporate office in Miami on Monday, March 23.
A model of the leaking instrumentation tube and conoseal was provided for the news conference, along with a one foot sample of stud and locking nut which hold the vessel head in place and a one minute video tape of corro-sion damage to studs actually in place.
Media coverage was generally factual and reflected mostly accurate accounts of the problem and status of the NRC inspection.
The Public Affairs Officer retur ned to the Regional office on Tuesday, March 24.
Media interest is expected to continue until the unit success-fully restarts.
Overview of the Event On August 30, 1986, with Unit 4 in Mode 3 (hot standby),
FPL maintenance personnel noted water and steam leakage from one of the four reactor vessel level instrument ports (conoseal 4MCS-02, northeast corner of the vessel head).
The leak was located at a mechanical clamp (conoseal)
on the instrument port assembly.
See Figure I for a description of conoseal
I
assembly.
FPL engineering prepared a safety evaluation which considered the leakage minor and within Technical Specification limits.
FPL engineer-ing recommended that the conoseal clamp be reinspected in six months.
On October 24, 1986, the conoseal leak was inspected during an unscheduled condenser maintenance outage.
The inspection indicated that the leakage was within Technical Specification acceptance limits.
The boric acid buildup on the leaking conoseal clamp was removed to inspect the clamp for corrosion.
No significant corrosion was found and the safety evaluation performed on August 30 was deemed to still be valid.
FPL engineering, at this time, recommended that the conoseal clamp be disassembled and re-paired during the next available shutdown of sufficient length.
On February 26, 1987, an engineering evaluation of the conoseal inspection results indicated that the conoseal clamp inspection could be deferred to April 24, 1987 (six months from October 24 date).
At this time, the licensee was also evaluating the possibility of extending the required conoseal inspection date to the next refueling outage.
On March 13, 1987, site engineering was informed by Westinghouse (W) that the corrosion 'rates from the conoseal leak on Unit 4 as evaluated in Safety Evaluation JPE-M-86-077, were inaccurate and may actually be double those previously assumed.
Unit 4 was taken from Mode 3 to Node 5 (Cold Shutdown) to assess the leak.
Unit 4 was already in hot standby (Mode 3)
for problems involving a valve on the inner door of the personnel air lock.
An inspection of the reactor head area revealed a large buildup (approximately 500 pounds) of boric acid crystals in the conoseal/reactor head area.
Twenty-eight of the 58 reactor vessel head studs were affected by the boric acid leak; eight of the studs were encrusted with three showing some thread damage above the castellated nut.
The licensee started cleaning the area and a detailed inspection was initiated by the licensee, Westinghouse, and BSW representatives.
After some visual and nondestructive examinations, the licensee decided to remove the reactor head to facilitate the cleanup, evaluate the effects of the boric acid crystals on the reactor vessel components, and provide meaningful data together with observations and evaluations to assure safe restart and operation of the plant.
Sequences of Events In early August 1986, Unit 4 began a restart following a lengthy refueling and maintenance outage.
The outage began in January 1986, and was extended from April to August due to diesel generator load capability concerns.
The reactor was finally taken to Mode 3 (RCS temperature greater or equal to 350'F on August 10 and following the completion of heatup, a visual leak rate inspection was performed on August 12.
The procedure used was OP 1004.1, Reactor Coolant System
- System Leak Test Following RCS Opening, which complied with the requirements of IWA-5240 of ASME Section XI code.
The inspector reviewed a
copy of the completed procedure and interviewed the
individual who performed the inspection in the vessel head area.
The inspection was performed by three contractor personnel who were all Level II certified inspectors.
At the time of the inspection, on August 12, the reflective insulation over the vessel studs was not installed although all other components were assembled.
Also, the RCS was at elevated pressure, approximately 2335 psig, and at normal operating temperature.
The inspector performing the procedure stated that he was very sensitive to leakage concerns because of a previous vessel 0-ring leak on the Unit 3 vessel.
He also stated that there was no indication of leakage from any of the conoseals.
On August
and 14, the plant was cooled down to perform work on CROMs at locations L-7 and H-12.
This work was over the vessel head but not in the immediate area of the affected conoseal.
On August 16, the unit again began a heatup and went critical on August 18 but shut down and cooled down again three days later due to Technical Specification requirements on Auxiliary Feedwater (AFW).
The following day a heatup was again started and terminated due to AFW problems.
The plant finally reached Hot Standby conditions (547'F, 2235 psig) at 0730 on August 23.
The following day the turbine was rolled but severe vibrations were experienced.
The reactor was maintained in Mode 3 from August 25 until September I while the generator exciter was replaced.
The conoseal leak was apparently first noticed by a Maintenance Foreman who was performing a containment closeout inspection early on the morning of August 30.
He stated during an interview that it has been routine to perform several closeout inspections prior to taking the reactor critical.
The closeout inspection is now proceduralized by O-SMM-051.3, Containment Closeout Inspection; however this proce-dure has undergone at least two revisions since August 1986.
He stated that he specifically looked at the four conoseals from the
foot (refueling floor) elevation using a flashlight.
He noticed a
"fine mist and spitting" from the affected conoseal.
The information about the leak was then passed on to the Maintenance Superintendent, who, along with a gC inspector, went into containment to inspect the conoseal at about 0800 on August 30.
This Superintendent stated that it was very difficult to see the leak from the 58 foot elevation (could not be observed without a flash-light) and that one had to know where the leak was in order to locate it visually.
The Maintenance Superintendent then notified Site
'ngineering personnel and requested an analysis of the leak.
An engineer was called to the site (August 30 being a Saturday)
to evaluate the leak and write a Safety Evaluation (SE).
This, SE (JPE-M-86-077, Rev.O)
was completed and issued the same day.
An analysis of this SE is addressed in Section 7b. of this report.
On August 30, September
and 3,
Westinghouse site representative
r t
informed Westinghouse managers in Pittsburgh about the conoseal leak and the licensee's technical decision to operate with the leak.
On September I, the plant went online and operated until October 24, when it was shutdown for unscheduled condenser maintenance.
An inspection of the conoseal leak was conducted by the engineer who wrote the SE and the Mechanical Maintenance Supervisor on October 24; however, maintenance people had been sent into containment earlier to cleanup the area around the conoseal in preparation for the inspec-tion.
The engineer and the supervisor who inspected the conoseal were interviewed by the inspectors and stated that, aside from the leak, no other abnormal indications were observed.
The plant was at reduced temperature and pressure and the vessel head area was un-observable due to installed insulation and the Control Rod Drive Mechanism (CRDM) ventilation shroud.
As a result of this inspection, a letter, JPES-PTP-86-1550, was wr itten which stated
"no significant corrosion was found."
It also concluded that the original SE was still valid and that an inspection should be carried out by February 28, 1987.
A maintenance worker who cleaned the conoseal area on October 24, was interviewed by the inspectors.
He stated that the conoseal area itself was relatively clean, there was no major accumulation of boric acid and that the residue was
"dampish" and could be removed with rags.
There was, however, a significant buildup of boric acid on the surfaces of the reflective shielding and the cavity floor.
This individual estimated the layer of boric acid was less than 1/2" at the maximum point and extended over a 6-8-foot wide area.
He cleaned up this residue with a cfust pan and rags.
He stated that he saw no areas of corrosive damage.
No other documentation of the "as-found" condition of the conoseal was made at the time, although a post-inspection report was written on March 21, 1987, while the AIT was on site.
As the February 1987 inspection date approached, the Plant Manager requested a reevaluation of the need for the inspection since, it would require shutting down Unit 4.
The original reinspection date of six months from August 30, 1986, was based on:
(I)
an engineer-ing judgment of the effects of the assumed corrosion rates, and (2)
the projected shutdown of both units for the Unit 3 refueling outage and the integrated safeguards test.
Both the Unit 3 refueling outage and the safeguards test were delayed.
Therefore, it was deemed reasonable to try to maintain Unit 4 operating until April, if possible.
Site engineering wrote a
memo dated February 26, 1987, (JPES-PTP-87-368)
which concluded that since the October 24, 1986, inspection of the conoseal leak was satisfactory, the six-month inspection requirement may be revised to six months from that date rather than from August 30, 1986.
The licensee also contacted
Westinghouse representatives to request more information on boric acid corrosion rates.
On March ll, 1987, Unit 4 was manually shut down as required by Technical Specifications due to a leak in the containment personnel hatch inner door.
Unit 4 was maintained in Mode 3, Hot Standby, until March 13, while repairs to the personnel hatch door were in progress.
On March 13, 1987, site engineering was informed by Westinghouse (W)
that potential corrosion rates from the known conoseal leak on Unit 4 reactor vessel head penetration, as evaluated in Safety Evaluation JPE-M-86-077 were inaccurate and the actual corrosion rates may be double those previously assumed.
Unit 4 was immediately taken to Node 5,
Cold Shutdown, to assess the leak and the extent of boric acid contamination and subsequent surrounding corrosion areas.
The engineering personnel interviewed stated that this (March 13) was the first time consideration was given to cor rosion of components other than the conoseal clamp.
In addition, no visual inspection of the conoseal assembly was made in March 1987 before the plant was taken to a
cold shutdown condition.
This precluded any direct observation of the leak by knowledgeable personnel;.i.e.,
had the leak rate gotten progressively worse over this time period?
b.
Review of Safety Evaluation (DPE-M-86-077)
The SE was written to justify continued operation of Unit 4 with the known conoseal leak.
The engineer who wrote the SE was interviewed by the inspectors and it was determined that the SE was requested verbally and not in response to a deficiency documented by a Noncon-formance Report (NCR) or any other formal mechanism.
The engineer visually inspected the conoseal from the 58 foot level of containment (refueling level) where he saw minor water and steam vapor leakage.
The primary concern was the location of the leakage and the possibi 1-ity of the leakage getting worse.
Also, the intention was to insure that the leakage rate was within the limits of applicable Technical Specifications.
Based on a review of the SE and an interview with the responsible engineers, the inspectors had the following.
observations:
1.
The walkdown conducted did not appear to be sufficient in that the leakage was observed from a considerable distance.
The plant was hot and pressurized which, from a personnel safety standpoint, precluded close observation.
However, the effects of the borated water dripping on the vessel head were not observed nor evaluated.
2.
The "conservatively assumed" corrosion rate of 30-50 mils/year was inadequate.
Sufficient information existed in the industry with regard to the effects of boric acid corrosion on ferritic steel to show that the assumed corrosion rates were far from conservative.
A post-event memo from Westinghouse stated,
"In
one test series performed by Westinghouse, aerated 25K boric acid solutions were shown to corrode steel at about 400 mils/.
month at 200 F."
3.
Only corrosion of the carbon steel clamp on the conoseal was considered.
The visual leak test procedure, OP1004.1, states
"...particular attention shall be given to the insulated areas of 'components constructed of ferritic steels to detect evidence of boric acid residues resulting from reactor coolant system leakage."
In addition, several IE Notices (80-27 and 82-06)
and IE Bulletin 82-02 have been issued which discussed the effects of primary system components exposed to boric acid from small leaks.
4.
It is evident that the SE was performed and reviewed in a
hurried and superficial manner.
As evidence of the urgency to perform the evaluation, the Plant Supervisor-Nuclear log for August 30, 1986, had the following entries:
1055 (a.m.):
Received report from maintenance superi nten-dent...
that minor leak on conoseal is to be evaluated by engineering prior to our going on the line.
2345:
JPE provided engineering evaluation of conoseal leakage
-
ok to go critical - evaluation PNSC'd via telecon.
Engineering procedures required that the SE be reviewed by all other disciplines (e.g. electrical, ISC, civil, and technical licensing).
These reviews were obtained via telecon on August 30.
It was also determined that the PNSC approval was obtained the same day via individual phone calls to the commit-tee members.
Conference phone calls with PNSC members is employed on an as-needed basics at the plant, but individual approval such as occurred in this instance, precludes interac-tion between the members.
Minutes of a PNSC meeting held the following day, August 31, show that the SE was again reviewed by a quorum of committee members.
The inspection of the conoseal was documented in the October 24, 1986 letter, JPES-PTP-86-1550.
Although not required, this inspection was prudent in order to monitor the leak.
The inspectors believe that more thorough documentation of the "as-found" condition of the seal should have been performed.
The responsible individuals did not visually observe the conoseal until after the area was cleaned although an interview revealed the maintenance supervisor who in-spected the conoseal was briefed by the cleanup crew.
No mention was made in the October 24 letter of the extent of the spread of the-boric acid residue and apparently no one considered the leakage down the ICCS penetration to the vessel hea On February 26, 1987, JPES-PTP-87-368 was written to justify extend-ing the inspection recommendation from February until April.
In light of the incorrectly assumed corrosion rates in the original SE, this extension was not unreasonable and apparently Site Engineering took this opportunity to investigate more fully the effects of boric acid corrosion.
Interviews revealed that the extension request precipitated further discussions with Westinghouse.
These discus-sions led to the March 13, 1987 letter, JPES-PTP-87-518, which recommended
'immediate inspection of the conoseal clamp.
Based on this recommendation, the plant was then placed in cold shutdown.
Technical Specification 4.0.3 requires that inservice inspection of ASME Code Class 1,
2 and 3 components shall be performed in accor-dance with Section XI of the Boiler and Pressure Vessel Code and applicable Addenda as required by 10 CFR 50, Section 50.55a(g).
Paragraph IWA-5250(b) of the Code requires that the detection of boric acid residues on ferritic steel components 'shall require the location of the leakage source and the areas of a general corrosion, if any.
Operating Procedure 1004.1, Reactor Coolant System - System Leak Test Following RCS Opening, contains the following instruction:
"During this examination, particular attention shall be given to the insulated area of components constructed of ferritic steels to detect =evidence of boric acid residues resulting from reactor coolant leakage."
It should be noted that a
SE is not required by Technical Specifica-tion if the leak rate is less than 1.0 gpm (TS 3.1.3).
However, the SE that was performed was deficient for the reasons stated previous-ly.
The licensee's failure to properly evaluate the leak in terms of the boric acid corrosion of ferritic steel components is a violation (251/87-16-01).
8.
Observable Wastage and Corrosion NRC personnel entered Unit 4 containment to observe the conoseal, the reactor vessel, and other areas affected by the conoseal leak.
Preliminary observations were made during this general inspection.
The affected components and a brief discussion of each follows:
a
~
Conoseal (I)
Clamp - Apparent steam cutting on the inside axial surfaces of all three clamp segments at the male/female flange interface.
(2)
-
No apparent cutting or cleaning of gasket seating surfaces.
Both male and female flanges were discolored.
Three
axial marks were seen on the outer diameter of the female flange immediately below the location of the clamp.
(3)
Weld ring - Apparent steam cutting at one azimuthal location.
(4)
Gaskets - Discoloration of lower gasket only with one indication of steam jet pitting.
No indications on the upper flange.
(5)
Clamp shim - Severe corrosion, totally destroyed at some azimut-hal locations.
'.
Instrument Tube/Head Penetration The lower hillside of the penetration appear ed not to have any damage or corrosion.
The upper hillside needed to be cleaned for a thorough inspection but there appeared to be a shallow gouge.
c.
Reactor Head The upper part of the head was not fully accessible; however, some slight pitting and loose rust were observed.
d.
Top of Head Flange Negligible damage/corrosion was observed.
The transition area looked good.
e.
Side of Head and Vessel Flanges Negligible damage/corrision.
f.
Faces of Both Flanges and Studs Between Flanges Boric acid was observed on studs Nos.
5 through 37.
The boric acid appeared moist on stud threads.
Boric acid appeared to be behind studs in the counterbore area of the vessel flange face.
g.
Studs and Nuts Above Head Flange Three studs (Nos.
24,
and 26) were significantly corroded as were their associated nuts.
In addition, a fourth nut (No. 27)
had some loss of castellation.
h.
CRDH Forest Three CRDN penetrations were observed in boric acid and the RPI stacks appeared to have some boric acid spray on the Head Vent Shroud One of three support lugs was badly corroded.
Approximately a
degree arc of the vent shroud bottom was corroded away.
CRDM Cooler Ventilation Duct Boric acid was observed in significant quantity at the flange of one cooler.
k.
Head Lifting Lug Some boric acid was obser ved...
after cleaning no damage was apparent.
Vessel Head Insulation This insulation consisted of two 1.5-inch layers of B8W Kaowool (3 pounds per cubic feet)
covered by 36-inch wide Fiberfrax Cloth (Type L-144T).
The Fiberfrax cloth had been coated with fiberwax cement (gF-150)
and a
water proof cement coating (silicon emulsion).
Preliminary visual observations indicated that only the Fiberfrax cloth (ceramic and inconel fibers)
had been resistant to the boric acid, while the Kaowool blankets had been partially decomposed and
'mixed with the crystalline boric acid.
Kaowool is the trade name of a mineral insulating material that consists principally of aluminum silicate (45% Al,0, and 53K Si0,).
Kaowool is manufactured by Babcock and Wilcox; however, this company did not fabricate the remainder of the insulation used on the reactor head.
Kaowool or aluminum silicate is not soluble in hot or cold water or in any acid other than hydrofluoric.
m.
Vessel Below Floor n.
Oo Annulus area was observed to have boric acid encrustation in approxi-mately 120'rc.
Other Conoseals Boric acid appeared to have settled on one other conosea'l.
Other Components/Areas Containment Coolers/Emergency coolers had white'eposits which may be boric acid contamination.
Instrument line off Hain Steam Line appears to have had boric acid contamination.
Containment coolers coils had the appearance of a boric acid coatin Swipe about purge isolation valve to determine whether boric acid contamination is present.
p.
Shroud support and ring - Severe corrosion of the support and espe-cially the support ring, wher e boric acid solution had apparently flowed through the Kaowool insulation to the base of the shroud.
q.
Lower portion of the reactor head - The lower portion of the reactor head (below the CROM shroud)
had been covered, in a fan-shaped area, with boric acid crystals (probably mixed with Kaowool).
In the lowest regions, in the flange area, boric acid crystals had been deposited to a depth of approximately one foot.
When the CRDM shroud was lifted from the upper part of the reactor head the area below the leaking conoseal was observed to be coated with a layer of boric acid on the insulating material.
After the boric acid was chipped off, the insulation appeared to be intact.
Subsequently, when the insula-tion had been removed the reactor head appeared to be slightly discolored but not corroded.
9.
Metallurgical Aspects The reactor vessel pressure boundary (RYPB) components and materials that could be affected by the boric acid solution are listed in Table l.
TABLE 1 RVPB COMPONENTS AND MATERIALS THAT COULD BE. AFFECTED BY BORATED AQUEOUS SOLUTION COMPONENTS MATERIALS Closure Flange Region Studs, Nuts and Washers
"0" Ring Shell Flange Head Flange Conoseal Penetration CRDM Penetration Reactor Yessel Nozzles Reactor Yessel Shells, A-320L43 (Type 4340 low alloy steel)
A-213 (Type 304 aust. stainless steel)
.
A-508 (MnMoNi ferritic steel)
A-508 (MnMoNi ferritic steel)
A-182 (Type 304 aust. stainless steel)
B-167 (Type Inconel)
A-182 (Type 304 aust. stainless steel)
B-167 (Type Inconel)
A-508 (MnMoNi ferritic steel)
A-508 (MnMoNi ferritic steel)
Upper and Lower Head Discs A-302 (MnMo ferritic steel)
C, t
'eactor Vessel Closure Flange Pressure Boundary Materials The reactor vessel closure flange pressure boundary consists of low alloy ferritic steel fasteners, manganese-molybdenum-nickel ferritic steel head and shell flanges, and an austenitic stainless steel
"0" ring gasket (Figure 2).
Visual examination of the flange region indicates that the leak from the conoseal affected the fasteners and the head and shell flanges.
If the reaqtor vessel operated with reactor coolant below 350'F (hot shutdown),
the boric acid solution could have penetrated to the "0" ring gasket.
Our evaluation will consider the effect that the boric acid solution could have on each of these materials.
Austeni tie stainless steel was chosen as a gasket material because it is resistant to corrosive attack from borated reactor coolant solu-tion.
Austenitic stainless steel is susceptible to stress corrosion cracking when it is sensitized from heat treatment or when attacked by borated water solutions containing considerable amounts of chlo-rides, fluoride or sulfides compounds.
The "0" ring gasket was not heat treated to a sensitized condition.
Based on the chemical analyses performed on the borated crystals (Attachment I and Ref. 4),
it appears that significant amount of compounds were not present in the boric acid solution leaking onto the flange region.
Therefore, the "0" ring gasket should not be affected by the leaking boric acid solution.
The "0" ring gaskets were replaced following removal of the head from the reactor vessel.
The ferritic steel materials used in the head and shell flanges are susceptible to general corrosive attack (wastage)
from boric acid solution but are not generally susceptible to stress corrosion cracking.
Stress corrosion cracking was observed in this material in the Indian Point 3 (IP-3)
steam generator (Reference I), but the harsh environmental conditions in the IP-3 steam generator are not present in this case.
Therefore, the stress corrosion cracking is not considered likely from the leaking boric acid solution.
General corrosive attack has been observed in the laboratory at approximately 400 mils/month when exposed to 155 boric acid at 210'F (Ref. 2).
The amount of wastage caused by the leaking borated aqueous solution was determined during examination of the shell and head flange following removal of the head from the reactor vessel.
The reactor vessel closure studs at Turkey Point 4 have eight threads per inch and the pre-stressed installation results in a stress across the minimum shank diameter of 42.6 ksi.
The low alloy ferritic steel materials used for fasteners with eight threads per inch and pre-stressed to 42.6 ksi are susceptible to stress corrosion cracking in numerous aqueous solutions when quench and tempered to strength
I e
t
levels of 150 ksi and above (Ref. 3).
The material certification from the Turkey Point 4 closure flange fasteners indicates that they were, heat treated to yield strengths greater than 150 ksi. Tests performed by Westinghouse indicate that the quenched and tempered low alloy (4340) stud material is not susceptible to stress corrosion in boric acid solution.
Westinghouse indicates that for quenched and tempered martensitic steels such as the 4340 studs, SCC susceptibi li-ty is controlled by the yield strength and exhibits very little environmental specificity.
When quenched and tempered to strength levels of 150 ksi and above, these steels can fail by SCC in numerous aqueous environments.
Dedicated tests in boric acid solutions are relatively rare, but Westinghouse has conducted several.
A pertinent observation from the Westinghouse proprietary data base is that bolt-loaded, pre-cracked wide open load (WOL) specimens of 4340, heat treated to 148 ksi yield and stressed at 120 ksi/square root inch showed no crack growth in 29% boric acid solutions at 210'F.
When quenched and tempered to a relatively high yield strength of 174.4 ksi, precracked WOL specimens loaded to 80 ksi/square root inch cracked in distilled water but again did not crack in 25% boric acid solution when loaded to 100 ksi/square root inch.
The boric acid solutions, however, severely corroded the stressed WOL specimens and blunted the pre-crack tips.
These observ 'tions indicate that stress corrosion cracking of the stud material is not a primary issue in the boric acid leakage at Turkey Point.
The general corrosion observed in these tests of 4340, however, reinforces the general conclusion that wastage remains the principal concern.
However, the effect of boric acid solution leaching into the stud lubricant has not been evaluated.
The stud lubricant used at Turkey Point Unit 4 is
"Never-Seez,"
Nuclear Grade which contains nickel. in an oil carrier.
The total sulfur and halogens are required to be below 1500 ppm and 200 ppm, respectively.
The low alloy ferritic steel materials used in the fasteners are susceptible to general corrosive attack when contacted by boric acid solution.
Visual examination indicates that studs 5 thru 37 (Fig. 3)
had been contacted by the leaking boric acid solution.
Three studs-
. (Nos.
24,
and 26)
have been severely damaged in the threaded region above their nuts.
The damage was from general corrosive attack.
The severest amount of wastage caused the studs to be reduced in the radial direction by approximately 0.300-inch.
FPL has ultrasonically examined seven studs (Nos.
22 thru 28) that were contacted by boric acid solution and three studs (Nos. 2; 46 and 54)
that were not contacted by boric acid solution.
The ultrasonic examination revealed indications in studs Nos. 22, 23, 24, 26, 27 and 28.
These indications could result from either wastage or stress corrosion cracks at crevices between the studs and the nuts, washers, and head flange or ultrasonic examination anomalies.
Studs Nos. 2, 25, 46 and 54 did not appear to have any indications.
(See paragraph 10 for details of the nondestructive examinations.)
I'
~
I
b.
FPL has examined studs Nos.
5 thru 37 and their nuts, washers, and
,
adjacent flange threads following removal of the head from the reactor vessel to determine whether these components have had either wastage or stress corrosion cracking resulting from the leaking leaching of boric acid solution through the stud lubricant.
Other Pressure Boundary Components The materials used for conoseal and CRDN seal penetrations are Type 304 austenitic stainless steel and 'inconel.
The inconel is seal welded to the reactor vessel head and the austenitic stainless steel is welded to the inconel.
The inconel to stainless steel bimetallic weld may result in sensitization of the heat affected zones (HAZ) in the austenitic stainless steel.
Sensitized stainless steel could be susceptible stress corrosion cracking when contacted by boric acid solution.
The licensee has examined the bimetallic weld and HAZ to determine whether the leak casised amu stress corrosion cracking.
Inconel's corrosion resistance in boric acid solution is equivalent or better than that of austenitic stainless steel (Ref. 1).
There-fore, the conclusions in Section a.
above for austenitic stainless steeI are relevant for the inconel components.
The materials used for the reactor vessel -nozzles, shells, upper head and lower head discs are 'ferritic manganese-molybdenum and manganese-molybdenum-nickel steels.
The effect of boric acid solution on these materials is equivalent.
As discussed in Section a., these materials are susceptible to wastage when contacted by borated aqueous solution.
These components were examined for wastage following removal of the head from the reactor vessel.
c ~
Conclusion Based on the licensee's commitments to inspect the RVPB components (to the extent discussed above)
and to either repair or replace any component damaged as a result corrosive attack from the aqueous borated solution, we conclude that the reactor vessel has been made acceptable for service.
10.
Nondestructive Examination of Studs Procedures and Past NDE History a ~
Background Turkey Point Unit 4 is currently in the first unscheduled outage of the first 40 month period of the second ten year Inser vice Inspection ( ISI) interval.
Unit 4 commenced commercial operations on
J r
September 7,
1973, and entered into the second ten year interval on-April 15, 1984.
The applicable codes for Preservice Inspection (PSI)
and ISI are as follows:
PSI and ISI through the end of the second 40 month period of the first ten year interval are American Society of Mechanical Engineers Boiler and Pressure Vessel (ASME B8PV) Code,Section XI, 1970 Edition through the Winter 1970 Addenda (70W70);
ISI for the third 40 month period of the first ten year interval is ASME BEPV Code,Section XI, 74S75; and ISI for the second ten year interval is ASME BSPV Code,Section XI, 80W81.
The PSI and ISI examination requirements for the Reactor Pressure Vessel (RPV) closure head studs, nuts and washers are specified in 70W70,Section XI Table IS-260, Category G-l, Items Nos.
1.8 and 1. 10; 74S75,Section XI, Table IWB-2600, Category B-G-l, Item Nos. B1.8 and B1.10; and 80W81,Section XI, Table IWB-2500-1, Category B-G-1, Item Nos. B6.10, B6.30 and B6.50.
The licensee performed the PSI examination of the RPV studs, nuts and washers reporting results on November 27, 1972, consisting of visual examination (VT), or magnetic particle examination (MT) and ultrason-ic examination (UT) of the RPV studs and nuts, and UT and VT of the washers.
ISI examination of RPY stud Nos.
1-35 and 38 with the associated nuts and washers was performed during the first 40 month period of the first ten year interval in accordance with code Section XI, 70W70.
These examinations consisted of VT, or MT and UT of the RPV studs and nuts and VT of the washers.
ISI examination of RPV stud Nos.
36, 37 and 39-58 with the associated nuts and washers was performed during the second 40 month'eriod of the first ten year interval in accordance with Section XI, 70W70.
These examinations consisted of VT, or MT, and UT of RPV studs and nuts and VT of the washers.
ISI examination of RPV stud Nos.
1-21 with the associated nuts and washers was performed during the third 40 month period of the first ten year interval in accordance with Section XI, 74S75.
These examinations consisted of MT and VT of.the RPV studs and nuts and VT of the washers.
ISI examination of RPV stud Nos.
1-20 with the associated nuts and washer was performed during the first 40 month period of the second ten year interval in accordance with Section XI 80W81.
These examinations consisted of MT and VT of the RPV studs and nuts and VT of the washers..
Inspection The inspectors reviewed the inspection plans, procedures and records for PSI and ISI of RPY studs, nuts and washers.
A detailed examina-tion was performed on the procedures (indicated below)
and records associated with stud Nos.
20-3 l g
I yt
PROCEDURES Procedure No.
Tit1e SwRI-NDT-900-1, Rev.
SwRI-NDT-900-1, Rev.
SwRI-NDT-300-2, Rev.
Visual Examination of Nuclear Reactor Components by Direct or Remote Viewing Visual Examination of Nuclear Reactor Components by Direct or Remote Viewing Visual Examination of Nuclear Reactors SwRI-NDT-'300-2, Rev.
Fluorescent Magnetic Particle Examination SwRI-NDT-300-2, Rev.
Fluorescent Magnetic Particle Examination SwRI-NDT-300-2, Rev.35 Fluorescent Magneti c Par ticl e Examination SwRI-NDT-600-18, Rev.
Manual Ultrasonic Examination of Pressure-Retaining Studs and Bolts 2" or Greater in Diameter Containing Access Koles SwRI-NDT-600-18, Rev.
Manual Ultrasonic Examination of Pressure Retaining Studs on Bolts 2" or Greater in Diameter Containing Access Holes FPL-NDE 5.7, Rev.
Ultrasonic Examination of Pressure Vessel Studs and Reactor Coolant Pump Studs FPL-NDE 5.7, Rev.
Ultrasonic Examination of Reactor Pressure Vessel Studs and Reactor Coolant Pump Studs
(I)- Procedure Review (a)
Visual Inspection (VT)
The inspectors reviewed the below listed procedures to ascertain whether they had been reviewed and approved in accordance with the licensee's established gA procedures.
The procedures were reviewed for techni-cal adequacy and compliance with licensee commitments/
requirements in the following areas:
type of visual examination used, direct or remote; lighting levels; cleanliness of surface to be examined; results are compared to acceptance criteria and required correc-tive measures taken.
Procedure No.
SwRI-NDT-900-1, Rev.
SwRI-NDT-900-1, Rev.
SwRI-NDT-900-7, Rev.
(b)
Magnetic Particle (MT)
The inspectors reviewed the below listed. procedures to ascertain whether they had been reviewed and approved in accordance with the licensee's established gA procedures.
The procedures were reviewed for techni-cal adequacy and conformance with ASME Section Y,
Article 7, and other licensee commitments/requirements in the below listed areas:
examination method; contrast of dry powder particle color with background and surface temperature; suspension medium and surface temperature for wet particles; viewing conditions; examination overlap and directions; pole or prod spacing; current or lifting power'yoke);
and accep-tance criteria.
Procedure No.
SwRI-NDT-300-2, Rev.
SwRI-NDT-300-2, Rev.
SwRI-NDT-300-2, Rev.
(c)
Other Procedures The inspectors reviewed the below listed procedures to ascertain whether they had been reviewed and approved in accordance with the licensee's established gA procedures.
The procedures were reviewed for techni-cal adequacy and conformance with ASME Section V,
Article 5, and other 'licensee commitments/requirements in the below listed areas:
type of apparatus-used; extent of coverage calibration requirements; search units; beam angles; DAC curves; reference level for monitoring discontinuities; method of demonstration of penetration; limits for evaluating and recording indications; recording significant indications; and acceptance limits.
~
Procedure No.
SwRI-NDT-600-18, Rev.
SwRI-NDT-600-18, Rev.
FPL-NDE-5.7, Rev.
FPL-NDE-.5.7, Rev.
(2)
Record Review The inspectors reviewed records of the VT, MT and UT examination performed on RPV studs Nos.
20-30 during both the first and second ten year ISI intervals.
The records were compared with the code and the applicable procedures in the areas indicated above for each specific method (i.e., VT, MT, UT).
(3)
Observation of Work On March 19, 1987, after some cleaning, the RPV studs were YT examined by the direct visual technique in the partially cleaned condition in accordance with FPKL Procedure NDE 4. 1, Revision 1, FCIA, "Visual Examination VT-1 Welds/
Bolting/Bushing/Washers."
This examination revealed some boric acid contamination on RPY stud Nos. 6-36 and multiple areas of wastage, 1/16 to 5/16 inch deep, in the threaded area, above the nuts approximately 360'n RPV stud Nos.
24,
5 26.
On March 20, 1987, the inspectors observed the UT examina-tion of RPY stud Nos.
23-27 in place prior to detensioning.
This examination was performed in accordance with FPRL Procedure NDE-5.7, Revision 1, "Ultrasonic Examination of
Reactor Pressure Vessel Studs and Reactor Coolant Pump Studs."
(Calculated unit tension load of 45.6 ksi.) The temperature of the RPV was 122'F as measured by control room instrumentation.
This was the first UT examination performed on the suspected studs following the discovery of the deterioration, and the first time that this utility had inspected installed tensioned studs.
The examination was a
manual, contact, pulse-echo shear wave angle beam technique incorporating the use of a special search unit that was manipulated inside the drilled access hole in each stud.
The examination.was performed using:
a Sonic MK1 instru-ment; with an Aerotech 1/4 inch diameter round 2.25 mhz search unit; examination angle of 60'ith demineralized water employed as the couplant.
The calibration block was a 13-inch long, threaded section of a RPV stud with 1/l6 inch wide grooves cut parallel to the threads one thread depth below the thread root in four locations.
With the instrument at primary refer ence level, and with the search unit in the calibration block (inside the center drilled hole)
the signal from the calibration reflector (I/16 inch notch)
was 50K of full screen height (FSH).
The examination was conducted over the entire threaded area of the studs inside the nuts and washers, and the head flange, the threaded area in the gap between the flanges of the head and the vessel, and approximately two to three inches into the vessel flange.
This examination was expanded on March 21, 1987, to include the studs indicated above and studs Nos. 2, 22, 28, 46, and 54.
The following results were noted:
Stud 2 - No recordable indications Stud 22 - One "crack like" indication in the back of the stud at the bottom of the nut.
Runs approximately 45'ircumferentially Stud 23 - Crack like indications located at the bottom of the nut running 360'nd at the junction between the washers and the flange for 180'tud 24 - Similar to No.
22, but indication runs circumferentially for 90'tud 25 - Similar to No.
22, but indication runs circumferentially for 90'tud 26 -
No recordable indications
Stud 27 - Similar to No.
22, but indication runs circumferentially for 90'tud 28 - Similar to No.
Stud 46 - No recordable indications Stud 54 - No recordable indications During the period March 26 - 27, 1987, RPV stud and nut and washer Nos.
24,
and 26 were removed from the RPV, and replaced with Unit 3 studs, nuts and washers one for one, one at a time.
On April 2, 1987, at 12:00 a.m.
stud Nos.
22, 23, 27 and 28 were UT examined after removal from the RPV with no tension load and an ambient temperature of 65'F, all other conditions were as described above.
This examination was repeated later the same day and wit-nessed by the inspectors.
No reportable indications were identified by either examination performed on April 2, 1987.
Starting April 4, 1987, the RPV head was removed in accordance with FPL Procedure No. 1407.7, of October 30, 1986,
"Reactor Vessel Stud Tensioner Operation, Closure Nut/Stud Removal, Guide and Stud Hole Plug Installation."
On April 4, 1987, the removed RPV stud Nos. 24, 25 and 26 were UT examined.
The conditions for the UT examinations were the same as those described for stud Nos. 22, 23, 27 and 28 performed on April 2, 1987.
On the same day stud Nos.
24, 25 and 26 were VT examined by the direct visual technique in the "as removed condition" (prior to cleaning) in accordance with FPL Procedure NDE 4. 1.
The results of these examinations were as follows:
UT examination revealed no recordable indications; VT examination, revealed with the exception of minor rust deposits, no further damage to the studs since the visual examination performed prior to detensioning (March 19, 1987).
By April 9, 1987, all the RPV studs, nuts and washers had been removed from the RPV and RPV head.
During the period April 10 - 15, 1987, RPV stud and nut Nos.
5-23 and 27-37, were VT examined by the direct visual technique in the
"as removed condition" (prior to cleaning)
in accordance with FPL Procedure NDE 4.1.
The RPV studs and nuts were then cleaned by Hydro Nuclear Services, Inc.
(HNS) in accordance HNS Procedure 3360-01, Rev.
1.
This cleaning process employs HNS Module 200-C Freon Tool Cleaner, which washes the parts with liquid Freon TF (Trichlorotrifluoroethane) at a pressure of 200 to 500 ps ~~
l
Following the cleaning operation, the RPY studs and nuts were MT examined in accordance with FPL Procedure NDE 2.2, Revision 1,
"Magnetic Particle Examination."
This examination was performed using a Parker Probe magnetic yoke, in the AC mode, with the continu-ous wet fluorescent method.
Following this examination, the RPV studs and nuts were UT examined.
The conditions for the UT examination for the RPV studs were the same as those described for stud Nos.
22, 23, 27, and 28 performed on April 2, 1987.
The nuts were UT examined in accordance with FPL Procedure NDE 5.10, Revision 1, "Ultrasonic Examination Nuts Two Inches in Diameter or Greater."
The examination was a manual, contact, pulse-echo shear wave angle beam technique.
Scanning was done clockwise and counter clockwise circumferentially around the OD of the nuts using a Kraut Kramer USK-7 instrument; with an Aerotech 0.5 inch diameter round 2.25 mhz search unit; examination angle of 45; with the ultragel II employed as the couplant.
The calibration block was a
7k inch long half section of a RPV nut, with three circumferential square bottomed grooves 1/16 inch wide, one thread depth below the root of the threads, and one axial square bottomed grooved 1/16 inch wide transverse to the threads one thread deep.
Following the UT examination, the RPY studs and nuts were again cleaned in accordance with HNS Procedure 3360-01 and coated with Neolube.
During the same time period, the RPY washers of the same numbers were VT examined by the direct visual technique in accordance with NDE 4.1 in the
"as removed condition," cleaned in accordance with 3360-01, and after cleaning and coated with Neolube.
The results of the above examinations were as follows:
The VT examination revealed no recordable indications with the following exceptions - light oxidation and boric acid residue was noted on most of the studs, nuts and washers, nut Nos.
23, 27 and 28 had damage to the castellations and the small No.
27 washer, exhibited pitting; UT and MT examinations of the studs and nuts revealed no recordable indications.
On April 16, 1987, seven RPV replacement studs, nuts and washer sets were received by the licensee.
The replacement RPV studs were released from receipt inspection on April 17, 1987.
Between April 18 and 20, 1987, the replacement RPV studs, nuts and washers were examined in accordance with ASME B8PV Code Section XI, 80W80 Table IWB-2500-1 Category B-G-1, Item Nos.
B6. 10, B6.30 and B6.50.
The seven sets of RPV studs, nuts and washers were marked 22R-28R.
The above examinations form the PSI for the seven sets and consisted
r f
'
of the following:
UT of RPY studs in accordance with FPL Procedure NDE-5.7; UT of RPV nuts in accordance with FPL Procedure NDE 5. 10; MT of RPY studs and nuts in accordance with FPL Procedure NDE-2.2; and VT RPV studs, nuts and washers in accordance with FPSL Procedure 4. 1.
For all of the above PSI baseline examinations, no recordable indica-tions were identified.
Results The licensee UT examined, prior to removal, while in tension, the three studs with wastage (24, 25 and 26),
two on each side (22, 23,
and 28),
and three others, unaffected by the boric acid (2,
and 54).
This examination identified recordable indications in six RPY studs (22, 23, 24, 25,
and 28).
After removal, the licensee
~ cleaned, UT and MT examined RPV stud and nut Nos. 5-21 and 29-37 and VT examined washer set Nos 5-21 and 29-37.
No recordable indications were identified.
The licensee procured replacement RPY studs, nuts=
and washers marked Nos.
R22-R28 of which the RPV studs and nuts. were UT and MT examined and the washers were VT examined.
No recordable indications were identified.
RPV stud No.
24 was sent to the Brookhaven National Laboratory (BNL)
for metallurgical evaluation.
The results of this metallurgical evaluation will be documented in BNL Technical Report A 33054-87.
RPV stud No.
was sent to Westinghouse for evaluation under the sponsorship of the licensee.
After removal and cleaning, RPV stud Nos.
22, 23,
and 28 were UT and MT examined.
No recordable indications were identified.
The recordable UT examination indications noted on stud Nos. 22, 23, 24, 25,
and 28 could not be repeated after detensioning and removal from the RPY.
Conclusion The licensee has performed a complete PSI/ISI baseline examination of all the RPV studs, nuts and washers that exhibited any boric acid contamination plus one on each side (5-37).
Any RPV stud sets (stud nut and washer set)
that had any recordable indications, including the UT indications that were identified in the tensioned state, which could not be repeated in the untensioned and cleaned state, were replaced with newly procured sets (22-28).
The newly procured sets were also subjected to a complete PSI/ISI baseline examination.
All of the RPY stud sets that were contaminated with boric acid are, or were replaced with RPY stud sets that are acceptable to the PSI/ISI examination acceptance criteria specified in the ASME BSPV Code Section XI, 80W8 r l
11.
Chemistry Aspects a ~
Background During the current fuel cycle, the reactivity control of the reactor had'been maintained by use of control rods as well as by adjusting the concentration of boron in the reactor coolant.
The desired concentrations of boron were achieved by adding to removing calculat-ed amounts of boric acid (H~BO~) from the reactor coolant.
The concentration of boron had been greatest (approximately 1800 parts per million (ppm)) at the beginning of the fuel cycle (September 1,
1986)
and had been decreased to approximately 1350 ppm by October 15, 1986, and to
- approximately 1000 ppm by March 1, 1987.
[One ppm is equivalent to one milligram (0.001 gram) 5of boron or 5.7 mg of boric acid per liter of water or roughly lx10 pounds of boron or 6x10 pound of boric acid per gallon.]
Chemical control of the reactor coolant system (RCS) also had been maintained to prevent corrosion of the reactor coolant pressure boundary piping inconel (stainless steel and 600)
and to minimize out-of-core radiation fields caused by activated trace elements such as cobalt-60.
Gaseous hydrogen had been added to the reactor coolant system to react with and eliminate dissolved oxygen that, otherwise, might cause general corrosion of the reactor coolant pressure bounda-ry piping.
Small amounts, no greater than 2 ppm, of lithium hydrox-ide had been added to keep the acidity of the boric acid solution at approximately pH of 6.5 to 7.0 (Westinghouse had provided guidance to the licensee for maintaining the ratio of boron and lithium within definite limits as shown in Figure 4).
The presence of the relative-ly small amount of lithium hydroxide also plays an important role in controlling the solubility of radioactive
"crud" (oxides of iron and nickel) formed in the RCS.
However, lithium concentrations were so low it was not considered to be significant in the analysis of this boric acid leak.
b.
Results of Conoseal Leak A reconstruction of the licensee's activities associated with the failure of the conoseal revealed the following:
(1)
At the time of startup for the current fuel cycle on September 1, 1986, a 'small'mount of steam was observed in the vicinity of the conoseal clamp.
Consequently, it is probable that small loss of boron inventory was already occurring through transport of boric acid that remained dissolved in the wet steam that formed when the reactor coolant escaped and flashed.
(2)
When the conoseal was inspected in October it was observed that white crystalline material (boric acid)
had been deposited on the top of the reflective insulation (in an area approximately six feet in width)
and on the floor of the reactor cavit Also, a film of white solid covered the conoseal clamp and the adjacent region of the conoseal penetration.
No indication of corrosion had been observed during this inspection; however, the inspector had not been able to observe any portion of the reactor head that was covered by the reactor shroud.
(An annulus of approximately 1/4 inch around the conoseal penetra-tion provided a
pathway for ingress of liquid that escaped either directly through the conoseal gasket or formed as the result of condensation of steam in the region of the conoseal.)
Consequently, it is not known if boric acid attack of the Kaowool insulation, on the upper portion of the reactor head (and hidden by the shroud),
had begun at this time.
When the reflective insulation was removed.from the reactor head below the CROM shroud in March 1987, there was visual evidence of large amounts of white solid material and accompanying loss of insulation and metal portions of the shroud support and other metallic protrusion on the reactor head.
White solid material was also observed in the ductwork of the CRDM shroud cooling system.
The crystalline material in the duct was analyzed to be H3B03 (approximately 90Ã) and iron/iron oxide.
c.
Analysis Westinghouse provided the licensee (by telex dated March 20, 1987)
information relative to the corrosive effect of various concentra-tions of boric acid on steel.
This information, however, was ob-tained under conditions where boric acid was 150 to 250 times as concentrated as in the reactor coolant.
The AIT obtained additional corrosion data more representative of the Turkey Point conditions, from BNL.
These data indicate that wastage of steel in the range of ten to 400 mils per month has been observed depending on the experi-mental conditions.
The only protection that had been provided against corrosive liquids was that the shroud support ring had been painted before the shroud was originally installed.
The type of paint used could not be determined at this time.
The licensee has replaced this ring because it suffered severe corrosion.
The new ring was painted with SP.10 (a surface pretreatment solution)
and then coated with Carboline CZ-1156, a zinc-containing corrosion-resistant coating material.
The upper portion of the reactor vessel head (under the CRDM shroud)
had been covered with Kaowool insulation to reduce thermal effects on the C,RDM.
d.
Cl eanup Observations of the areas of the reactor head that were affected by the conoseal leak (achieved visually from the containment refueling floor and from photographs taken from the reactor cavity both before and after cleanup)
showed that a pie-shape area of the head and
/
~
P I
pa
e.
flange regions below the faulty conoseal had been covered with a white/gray solid.
The licensee subsequently removed as much of this solid as possible to facilitate examination of the reactor head, CRDM reactor shroud, and the vessel flange regions.
Cleaning had been achieved by scraping techniques and then the exposed metal had been hydrolazed with steam.
Most of the solids had been removed by these procedures;,
however, further examinations were to be made to deter-mine if any solid material remained in crevices or hidden parts of individual components and sections of the upper portions of the reactor head.
The crysta'1line material that collected in the shroud cooling ducts was removed in the same manner.
The laminated reflec-tive insulation that covered the flange and lower portion of the reactor head had been removed and stored on the containment refueling floor.
Four of the nine panels had been *partially filled with white solid material.
The stainless steel outer skin (22 gauge 302 SS) of one panel had been discolored and the internals of four panels also appeared rusty colored.
Leak Rate Calculation A preliminary estimate of the amount of boric acid that might have been lost through the conoseal leak has been prepared.
Assuming a
boric acid concentration of 1150 ppm and a leak rate of 0.5 gpm, the loss of boric acid would have been approximately 1300 lbs.
per 30 days.
12.
Conoseal Failure
'a
~
Background The reactor incore thermocouple system utilizes 52 thermocouples to measure fuel assembly coolant outlet temperature at preselected core locations.
The thermocouple conduits enter the reactor vessel closure head through port columns which protrude through four vessel head penetrations.
(The location of these penetrations is shown on Figure 3 at locations 4-CS-Ol, 02, 03, and 04.)
The thermocouple conduits are welded to a seal plug/port column seal adapter assembly to provide a
path for the thermocouple leads through the reactor closure head.
This path is through the conoseal assembly which is a mechanical connection with two gasketed seals to prevent reactor coolant leakage.
The conoseal failed through the lower of the two seals.
The instrument port mechanical connection is shown in Figure 1.
This connection contains two stainless steel gasket seals to prevent reactor coolant leakage.
The lower gasket seal is established by applying a compressive force on the assembly to seat and compress the gasket.
The compressive force is applied using a special fixture and a hydraulic ram port-a-power.
While the assembly is compressed (with a load of 5720 to 6310 psi)
a three piece clamp is bolted around the beveled flanges of the assembly to maintain the compressive loa I
After the loading device has been removed, the upper seal is com-pressed by installing the jack screw plate and split ring at the top of the assembly, as shown of Figure 1.
The jacking screws on the upper seal are required to be torqued to 100 in.-lb.
and the bolts for the clamp on the lower seal are required to be torqued to 125 ft.-lb.
Per discussions with Westinghouse representatives on site, the materials for the conoseal assembly are as follows:
Female Flange Male Flange Lower Gasket Upper Gasket Clamp Spacer b.
Design Considerations Type 304 SS Type 304 SS Type 321 SS Type 321 SS SA 508 Class 2 or SA-541 Class
Unknown As shown in Figure I, the gasket seals are established without the application of any sealing compound.
The upper seal is loaded by a jack screw assembly which applies the force in the axial direction.
The lower seal is a more complex arrangement in that the compressive load is applied in a circumferential direction by the clamp.
This circumferential force is transferred (through beveled surfaces on the clamp and flanges) into an axial force which maintains the compres-sive load on the seal.
The contact surfaces between flanges and clamp are coated with neolube to enhance the transfer of the load from circumferential to axial.
The conoseal assembly shown in Figure 1 contains a spacer or shim between the clamp and the upper flange surface.
This spacer is not shown on any of the drawings available at the plant (and is in fact on only three of the four conoseals,on the Unit 4 reactor head).
In response to the event, the licensee reviewed historical data in order to determine the requirements for the spacer.
This review identified a memorandum dated August 4, 1972, which provided a method of measuring the clamp to ensure that it would provide the proper clamping action to the seal joint. It also stated that with the use of a stainless steel shim an unacceptable clamp could be used tempo-rarilyy while a new clamp was being mad P l
The items of particular interest in this memo were the statement that
"-use of shims is only temporary measure until new clamp can be made" and "the shim is to be used one time only, the same as conoseal."
I These two statements apparently did not get transferred into site procedures because procedure No.
4-GMM-043.2, "Installation of Reactor Vessel Instrument Ports," for three of the four conoseals on Unit 4, still has the original clamp put together with the temporary shim; and the procedure does not list a replacement shim as material required.
The joints have been reassembled for approximately
years using what is apparently the original shim material.
Failure Mechanism The initiation of the fai lure appears to have been an improperly assembled lower conoseal ring.
The exact cause of the improper assembly is still a matter of conjecture, but the fact that neither the seal ring nor the flange seal surfaces showed any evidence of damage leads to the conclusion that for some undetermined reason the clamping force on the conoseal flange was not adequate to prevent leakage past the seal.
Once the seal started leaking, the borated water/steam began to attack the clamp thereby decreasing the clamping force even further.
This is supported by the fact that the clamping surfaces and the shim between, the clamp and flange show evidence of corrosion.
After a demonstration of the conoseal assembly procedure, by mainte-nance personnel using a full scale training mock-up, the inspectors asked for information concerning the design, the materials and a copy of the written installation procedure.
After the licensee provided procedure 4-GMM-043.2, "Installation of Reactor Vessel Instrument Ports" dated November 22, 1985, and a
copy of the August 4, 1972, memorandum discussed in paragraph 12.b.,
above the inspector request-ed copies of the original maintenance procedures for this application and all major revisions to the procedure since original issue.
The licensee provided Maintenance Procedure 1407. 15, Rev. 0, dated April 3, 1972, which was applicable to both Units 3 and 4.
The following revisions to procedure 1407.15 were also reviewed March 14, 1975; October 1, 1976; January 1, 1980; February 2, 1982; May 30, 1984; September 26, 1984; and March 15, 1985.
The March 15, 1985, revision cancelled procedure 1407.15 by referencing new plant specific procedures 3-GMM-043.2 and 4-GMM-043.2.
Items of particular note are as follows:
(I)
The August 4, 1972, memorandum provided a
means by which a
nonconforming conoseal clamp could be used temporarily while a
new clamp was manufacture ~h
s
(2)
The original installation procedure 1407.15 dated April 3, 1972, predated the August 4, 1972 memorandum.
(3)
The March 14, 1975 through September 26, 1984, revisions to procedure 1407. 15 did not acknowledge that any extra materials, such as the shims were required to assemble the instrument ports.
(4)
Procedure 4-GMM-043.2 included a step involving replacement of the shim on three of the four instrument ports on Unit 4, but did not require that the shim material be replaced as suggested by the August 4, 1972 memorandum.
(5)
A revision to procedure 4-GMM-043.2 dated November 22, 1985, changed the sequence of operations involving the torquing of the conoseal clamp.
The change in sequence went from torquing the conoseal clamp bolts to the full 120-128 ft.lb torque with the flange loaded by hydraulic ram to 6000 psi to installing the conoseal clamp bolts hand tight with the flange loaded to 6000 psi and then torquing to 120-128 ft.-lb after the hydraulic ram had been depressurized and removed.
After the historical review of the procedures for the installation of the instrument ports, the inspector reviewed the justification for the change in operational sequence which resulted in removal of the hydraulic load prior to torquing of the clamp bolts.
The listed reason for the change to the procedure was as follows:
"To upgrade procedure and allow maintenance and gC department to complete procedure correctly and with a minimum of radiation exposure.
To maintain continuity of job."
\\
The reason listed above is misleading in the fact that it is called an upgrade to allow the procedure to be completed correctly.
The change does reduce radiation exposure in that the gC department can witness the torquing of the flange bolts and jacking screws during one trip onto the reactor head instead of two.
In retrospect, this would not qualify as an ALARA decision because the decrease in safety margin for the torquing of the conoseal flange would make the decision. unreasonable.
d.
Disassembly of Unit 4 Conoseals On March 24, 1987, the disassembly of the three remaining conoseals was observed.
The as-found conditions from this disassembly inspec-tion are as follows:
(1)
NORTHWEST CONOSEAL (a)
(b)
(c)
(d)
Jacking screw breaking torques 3 at 75 in/lbs I at 95 in/lbs 2 at 135 in/lbs Clamp bolt breakaway torque 85 ft/lbs The spacer was not properly installed. It was found located between the male and female flange on top of the large conoseal gasket.
The licensee subsequently determined on April 23, 1987, that the clamp spacer may have actually been installed properly.
The responsible gC inspector stated that due to the amount of activity in progress (the simultaneous disassembly of the remaining three conoseals),
his location relative to the northwest conoseal, and the difficulty in communicating while wearing respirators, it is conceivable that the conoseal was properly installed.
The small and large conoseal gaskets were properly oriented.
(2)
SOUTHWEST CONOSEAL (a)
Jacking screw breakaway torques 4 at 90 in/lbs 2 at 95 in/lbs (b)
(c)
Clamp bolt breakaway torque 155 ft/lbs No spacer was observed.
This was as expected as a spacer was known not to have been installed after previous maintenance.
(d)
Small and large conoseal gaskets were properly oriented (3)
SOUTHEAST CONOSEAL (a)
Jacking screw breakaway torques 1 at 70 in/lbs I at 80 in/lbs 4 at 140 in/lbs
(b)=
Clamp bolt breakaway torque 155 ft/lbs (c)
Spacer was observed, as designed, on top of the male flange.
(d)
Small and large conoseal gaskets were properly oriented.
(4)
NORTHEAST CONOSEAL The northeast conoseal was the failed seal, which allowed the boric acid to leak from the RCS and accumulate on the reactor vessel head.
It was removed by maintenance personnel on March 18, 1987.
The northeast conoseal was not inspected for the as-found condition or for disassembly data collection.
Based on interviews of maintenance personnel, the following information was obtained:
(a)
Jack screws were tight and impacted by boric acid.
(b)
Deterioration of the clamp nuts was evident and the clamp ring showed signs of erosion over full circumference.
(c)
The spacer was highly deteriorated and had corroded into two pieces, each approximately 180'f the circumference.
After disassembly of these conoseals, the shims were checked to determine whether they were stainless steel as originally described.
The shims on the NE conoseal and one other conoseal were found to be magnetic indicating carbon steel rather than the originally specified stainless steel.
The licensee confirmed by chemical means that these shims were in fact carbon steel.
An investigation revealed that during the 1984 Unit 4 refueling outage, a
new shim was fabricated by maintenance
.personnel from apparently carbon steel mater ial.
During this inspection, no Plant Work Order (PWO) or other documentation had been located author-izing the fabrication of the'ew/replacement shim.
The carbon steel shim was used by the Westinghouse personnel who reassembled the conoseal during the 1986 refueling outage (March 1986), which was the last assembly before the conoseal leak in question.
Conclusions The failure of the conoseal appears to have been the result of a
series of problems, dating back to 1972 which were unchallenged. until the cumulative affect resulted in the problem discovered in March 1987.
The series of problems can be summarized as follows:
alt
~
V J
(2)
(3)
(4)
1972 - Original conoseal clamps found to be in nonconformance
-'SSS authorized installation with stainless steel shims until new clamps could be received.
1972 to 1985, - Nonconforming conoseal clamps apparently continue to be installed without controls on shims (in fact, installation procedure does not mention shim).
1984 -
Carbon steel shim fabricated by Maintenance Personnel without instructions and re-used in March 1986, which was the last assembly before the conoseal leak in question.
March 1985 - Nonconforming conoseal clamps still in use, but installation procedure now revised to include step for installa-tion of shim.
(5)
(6)
November 1985 - Unit 4 procedure revised to change installation sequence so that conoseal clamp is torqued after release of the 6000 psi preload used to seat lower seal.
August 1986 - Safety Evaluation for conoseal did not account for the fact that a
shim of unknown material was a part of the clamping arrangement and that corrosion (wastage)
of this shim could further relax the flanged joint and increase the leak rate.
The problems described above were allowed to occur because of a
flawed program that allowed weaknesses in the preparation of and adherence or procedures.
This conclusion is supported by the follow-ing facts:
The conoseal clamps were installed from 1972 to 1985 without any indication that anyone thought it abnormal to have an extra part and an extra step in the assembly that were not described in the procedure.
(2)
(3)
In March 1985, the procedure was revised to include a step for the installation of the shim on the top of the male flange prior to installation of the conoseal clamp, but there is no indica-tion that anyone in the entire review cycle asked why there was a part that did not appear on the parts list, did not appear on any drawing, and did not appear in earlier revisions of the procedure.
In November 1985, the procedure sequence was changed to allow torquing of the conoseal clamp after relaxation of the installa-tion preload in order to make it easier for mechanics to reach the clamp bolts with a torque wrench and to reduce radiation exposure by eliminating one trip onto the head area for gC inspectors.
There does not appear to have been any technical
4v r
~
J
review of the reduction in safety margin that the change would have on the installation.
(4)
Carbon steel shim fabricated in 1984 by Maintenance Personnel without instructions and re-used in March 1986, was the last assembly before the conoseal leak in question.
The activities described above appear to be a violation of the Turkey Point Technical Specification paragraph 6.8 requirements for prepara-tion of and adherence to procedures.
(Violation 251/87-16-02)
Measurement of Reactor Coolant System Leakage The basis of the measurement of reactor coolant system (RCS) leakage is an accounting of the changes in water mass in the RCS proper, the volume control tank (VCT), and pressurizer (PZR) over a period of time.
Although the RCS, which in this case is the entire high pressure system except the pressurizer, is always full, changes in the average system temperature do affect the inventory.
At full power operating conditions, a change in average temperature of one'F will change the mass by about 590 pounds in the 8043 cubic foot system of Turkey Point 3 or 4.
This is equivalent to
.71 gallons of water at standard conditions.
Since an increase in tempera-ture corresponds to a decrease in inventory; an increase of less than one degree over the period of a one-hour leakage measurement would totally obscure a leak in excess of the one gallon per minute allowed by Technical Specification 3.1.3.
That specification does not define the parameters of the gallon; that is the temperature and pressure conditions to be considered.
Throughout this inspection, it has been assumed that the gallon of water is at room temperature and atmospheric pressure, and weighs 8.33 pounds.
Usually pressurizer level varies over the period of the measurement.
The level instrument,'a differential pressure cell, is usually calibrated for operating conditions, 2250 psia and 653'F.
Thus, the indicated change in volume is for water at these conditions, that water has a density of about 5 pounds per gallon.
The VCT level is measured by a masonelian gauge, a displacement device that is sensitive to temperature changes.
Therefore, the indicated level must be corrected for density effects to determine the actual mass or standard gallons represented.
Fortunately, density is not very tempera-ture dependent in the operating region of the VCT, and a one time correc-tion for typical operating conditions is usually sufficient.
The density of YCT water is 8.25 pounds per gallon.
A test dur ation of two to four hours is most desirable.
The test results are derived from the beginning and ending readings of the parameters.
The longer the test the less sensitive the results to instrument inaccuracies.
A practical upper limit is placed on test duration by the need to make up to the VCT.
The water and boric acid flow integrators are not accurate
,j v s
ft'
nor calibrated in most cases; hence makeup should not be performed during the test.
The measurements described above address the measurement of total leakage from the RCS.
That leakage that is collected in the pressur izer relief tank (PRT)
and the reactor coolant drain tank (RCDT) as well as that from measured leaks, such as from valve packing, is defined as identified leakage.
The allowable leakage of one gallon per minute is the difference between the measured total leakage and the identified leakage.
To use the microcomputer program RCSLK9 to calculate the total leakage measured during routine surveillance tests at Turkey Point 4, using data obtained during those tests, the table of plant specific parameters given in attachment 2 was prepared by the inspector.
The sources of the data included the updated FSAR, the plant curve book, and System Description 13.
The program is described in NUREG-1107, RCSLK9:
Reactor Coolant System Leakage Rate Determination for PWRs.
Typical output from the program is shown in Attachment 3.
A summary of program results using data procedure 4-0SP-041.l is given in Attachment 4.
The fitted slope in the plot of leakrate versus time is more apparent than real.
At the 95K confidence level the slope is 0.00073
+/-0.00087 gpm/day.
The correlation coefficient for the fit is 0.2.
In fact, there was no significant trend of increasing leakage rate.
The data plotted are for gross leakage, rather than unidentified.
In general, pressurizer relief tank (PRT)
and reactor coolant drain tank (RCDT). levels were not recorded with sufficient precision to subtract out identified leakage.
Had the level data been available, a plot of unidentified leakage 'might have shown less scatter, and, of course, a lower magnitude.
It appears that the licensee was always in compliance with the Technical Specification of less than one gpm unidentified leakage, and that if there was an increase in leakage rate, it was too small to detect.
The evaluation and plotting of the RCSLK9 data was performed using a
least-square spread sheet with the SUPERCALC3 microcomputer program.
Some differences from licensee calculations of gross leakage were ob-served.
The most significant is the correction factor of 40 used by the licensee to convert a
change in average RCS temperature to standard gallons.
For full power operating conditions the inspector calculated a
correction factor of 71.
At no load conditions the factor should be 60.
There is no universal factor.
The licensee used a multiplier of 60 to equate changes in pressurizer level to standard gallons.
The inspector calculated a value of 51.
Discussions with licensee personnel also revealed that they believe xenon transients affect RCS leakage tests results even when constant power and average temperature are being maintained.
To test that obser vation, the inspector considered a core in which the linear heat rate was constant and coolant temperature increased uniformly as it rose through the core.
This
a
was compared with a core in which 52.5X of the power was generated uni-formly in the bottom of the core and 47.5% uniformly in the top.
The axial offset would be 5X in the second case compared with 0 in the first.
For the same inlet and outlet temperatures 546.2 and 602. 1 respectively, the core average temperatures were 574. 15 and 575.55 for case one and case two respectively.
With an estimated 536 cubic feet of water in the core,
,the change in inventory was nearly 50 standard gallons.
Since the temper-ature rise is not uniform, because the power distribution is not uniform, this is not an accurate calculation of the correction "for a 5X change in axial offset.
However, it does demonstrate that changes in power distri-bution, at constant total power, during a
RCS leakage measurement do affect the measurement results.
With a requirement to measure leakage daily, measurements during xenon transients are as inevitable as their results are dubious.
At the end of the inspection, the licensee had not been able to justify the correction factors in 4-OSP-041. 1, or to refute the values calculated by the inspector.
Consequently, the procedure has been found to be inadequate, and, hence, in violation of the intent of Technical Specifi-cation 6.8.1.
(VIO 251/87-16-03)
Had the licensee's procedure contained proper correction factors, it probably would have been adequate to perform the surveillance to identify a leak-before-break RCS pipe failure, and that is the intended purpose of the surveillance.
However, neither the licensee's method nor RCSLK9 contain the detailed analysis of system variables and variable measurement to resolve changes in leak rate to a few tenths of a gallon per minute.
14.
Licensee's Corrective Action a 0 Stud Removal The licensee elected to remove and replace stud Nos. 24, 25 and
with the remaining studs in place and tensioned prior to removing the vessel head.
Three replacement studs were available from Unit 3.
The intent was to replace the three studs with damaged threads and then proceed with head removal using FPL's normal procedures for'tud detensioning and removal.
The licensee developed a temporary opera-tion procedure, TOP 322, 3/25/87, Reactor Vessel Stud Removal and Replacement, for removal of these damaged studs.
Review of this procedure by the inspectors revealed that the studs were to be detensioned via electrical heating.
Following verification of proper elongation, nuts were to be rotated.
In the event that the nuts could not be rotated by mechanical means, the procedure provided several contingencies for nut removal (i.e., torch cutting or mechan-ical splitting).
On March 26, 1987, the insp'ectors witnessed detensioning of stud No.
25 in accordance with TOP 322.
This stud was the first of the damaged studs to be removed.
The inspectors verified that applicable provisions of TOP 322 were met, with particular emphasis on
requirements for thermal protection of neighboring components, elongation, and temperature.
Stud No. 25 was successfully detensioned with electric heat and the associated nut was rotated by mechanical means.
b.
Nonconformance Reports The licensee initiated 24 nonconformance reports (NCRs), indicated below, for the recovery effort.
During the period April 14 - 17, 1987, the NCRs marked * were reviewed by the inspectors to determine whether the records adequately documented the status of the NCRs; NCRs were legible, complete and promptly reviewed; records associated with the NCRs were properly stored and retrievable; and NCRs included the status of corrective action or resolution..
On April 22, 1987, the inspectors performed a walkdown inspection of selected hardware, the subject of the NCRs marked '.
These observa-tions, combined with the observations made by the inspectors during a
walkdown inspection of the same hardware on March 20, 1987, were used by the inspectors to evaluate the licensee's satisfactory resolution, the accuracy and completeness of the discrepant conditions, signifi-cance, recormended disposition, and compliance with disposition.
NCR NO.
87-0071
'7-0072
'7-0074 87-0075
- '7-0076
- '7-0077 87-0078
- '7-0079 87-0080 87-0084 *
87-0088 *
Title/Sub ect
"A" CROM Coolers Surface Prep for MT (ISI)
Conduit/Mireways 14'levation Normal Containment Coolers
"C" S/G Spring Can RPI Cables Conoseals (C-0316-87)
RX Vessel Reflective Insulation RX Vessel Permanent Insulation (C-0310-87)
RX Vessel Annulus Nozzle Support Plates
l A
\\
J
87-0089 87-0090
- '7-0099 87-0100
- '7-0106 87-0107
- '7-0115
'-0246-87 C-0295-87 C-0330-87 C-0364-87 C-0365-87 C-0378-87
Excore Detectors RX Stud 18 Dust Cover RX Head Shroud (C-0309-87)
Studs/Nuts/Washers (ISI)
Cavity Seal Ring Clamp RX Head Shroud Support (C-0318-87)
HVAC System Support CRDM "A" Fan Shroud Broken Cable Anchor Bolts Vs.
Drwg Cables Tray Reactor Shroud RPI Coils Lead Shielding on RX Head Seal Ring Fan Motor/Sliced Jacket The licensee examined the reactor vessel head in the area covering studs 22-28 as well as penetration No.
53 counterbore where the NE conoseal was installed.
The results of the ultrasonic examination showed no appreciable wastage on the reactor vessel head dome.
This wastage was evaluated in accordance with the ASME Code and the design basis was satisfied.
Moreover, the ultrasonic examination indicated wastage on Penetration No. 53 up to k".
The detail of the wastage is shown in Figure 5.
The enlargement due to corrosion was evaluated in accordance with the ASME Code.
The design criteria for head minimum thickness, and penetration reinforcement requirements were satisfied.
Some wastage on the reactor vessel head flange was observed between studs 23-24 and 24-25.
The evaluation performed on these areas was acceptable because the ligaments between studs are not a highly stressed area.
15.
Review of Process Radiation Monitoring System The inspector performed a review of associated documentation relating to the process radiation monitoring system for the period. of August 1986 through March 1987.
This documentation included LERs, isotopic analyses from grab particulate samples, control room computer printout of R-ll and
II
R-12 readouts, and nuclear plant work orders.
Through this review and discussions with licensee personnel, the. following was determined From November 14, 1986 to February 12, 1987, Unit 4 Process Radiation Monitoring System (PRMS) R-ll (the particulate monitor) repeatedly failed high, resulting in a containment and control room isolation actuation.
The cause of these events appears to be numerous equip-ment component failures, including broken soldered joints, burned out resistors, a failed ratemeter, faulty cable connection, and a
bad circuit board.,
The PRMS is original Westinghouse equipment and in need of replacement.
The licensee stated that this system is sched-uled to be replaced.
During the time that R-11 was out of service, the licensee took grab particulate samples per the interim Technical Specification require-ments.
Inspector review of the data indicated that the increasing R-ll levels were not substantiated by the grab particulate sample analyses.
In geryral, rQioisotopic analyses showed activity levels consistent at 10 to 10 uCi/cc.
While R-11 continued to fail during this period, R-12, the containment gas monitor, showed increasing radiation levels.
Inspector review of the weekly containment air grab samples pulled by chemistry and analyzed on the GELI system supported the R-12 readings.
The licensee stated that the containment air samples corresponded to reactor coolant activity which was also in'creasing during the same time period.
The plant radiochemist issued an inter-office memo to the Operations Superintendent to inform him of the increasing RCS activity in Unit 4 on December 2, 1986.
The memo stated that the increased reactor coolant activity appeared to be due to a loss of fuel integrity.
The memo also stated that the RCS activity would continue to be monitored.
Based upon the above findings, it appears that the licensee probably could not have identified the boric acid leak on the top of the reactor vessel by reviewing and noting the indications and failures of PRMS R-11 and R-12 containment monitors.
16.
Response to Previous Industry Experience The inspectors considered the aspect of Operating Experience Feedback and whether or not sufficient information was routed to proper plant personnel concerning boric acid corrosion.
This topic was discussed in IE Informa-tion Notices 80-27, 82-06, and 86-108, IE Bulletin 82-02 and Institute of Nuclear Power Operation (INPO) Significant Operating Experience Report (SOER) 84-5.
FPL responded to IEB 82-02 by letters dated August 2, 1982, July 15, 1983 and March 9, 1984.
The Bulletin however, did not address programmatic
!
I
/
i
'
requirements for the identification and evaluation of primary system leakage.
It required 1 icensees to:
1.
Develop and implement maintenance procedures for threaded fastener practices.
2.
Identify bolted closur es that have experienced leakage.
3.
Identify closures and connections where fastener lubricants and injection sealant materials have been or are being used.
4.
Inspect those fasteners identified above in accordance with Section XI requirements prior to reuse.
The licensee's actions in response to this bulletin were inspected and closed in Inspection Report 50-251/84-12.
On September 20, 1984, INPO issued SOER 84-5, Bolt Degradation or Failure in Nuclear Power Plants.
Included in this SOER were bolt failures caused by boric acid corrosion due to leaks and by stress corrosion cracking due to local environments and stress.
A number of cases were cited where boric acid corrosion caused degradation of RCS boundary bolts.
One section of the report makes the following statement:
"Failure of reactor coolant system pressure boundary closures in PWRs due to rapid boric acid corrosion of closure bolts is a concern.
This suggests that even the smallest flange or gasket leak in borated systems should be quickly repaired.
Recent research has demonstrated a worst-case boric acid corrosion rate of 1.65 inches of metal per year."
The report also made the following recommendations that specifically related to boric acid corrosion problems:
l.
Ensure that operating and maintenance practices require prompt repair of leaking pressure boundary joints in systems containing borated water.
2.
Training programs for maintenance, plant engineering, and quality control personnel should include...
industry experience concerning bolt failures, including the effects of borated water leakage of closure bolts.
On December 29, 1986, the NRC issued IEN 86-108, Degradation of Reactor Coolant, System Pressure Boundary Resulting from Boric Acid Corrosion.
It should be noted that this IEN was issued several months after the identi-fication of the conoseal leak and issuance of the SE.
FPL responded to
J
p
/
IEN 86-108 by issuing an Operating Experience Feedback Report, FOP-87-001.
This report forwarded IEN 86-108 and stated:
"St.
Lucie, in response to recommendation 1 to SOER 84-5, reported that PWO's written for leaks with boric acid are given higher priority than other repairs.
Turkey Point is still reviewing SOER 84-5."
The report, FOP-87-001, was subsequently cancelled since its substance was reportedly tracked by two other plant action items.
FPL, personnel were requested by the inspectors to provide any additional information concern-ing the two recommendations previously mentioned.
With respect to recom-mendation 1,
FPL stated that
"An inspection for boric acid was included with the Unit 3 isometric walkdown verification procedure dated 3/28/85."
This procedure is document number CIS-PTP-85-103, Procedure for the Isometric Walkdown Verification of Ten-Year Isometric Drawings, and its purpose is described as providing guidance for those individuals conduct-ing the walkdown of the ten-year inservice inspection isometric drawings.
It does not appear to meet the need for prompt identification and repair of leaking pressure boundary joints.
It should be noted however, that the leak test visual procedure, OP 1004. 1, did contain an instruction relating to boric acid residues (see paragraph 7.a of this report).
With respect to the second recommendation, the only training that was scheduled concerned
"full thread engagement, Hilti certification and threaded fasteners."
Training Brief No. 51, Acceptabl'e Thread Engagement, was issued but no mention was made concerning the effects of borated water leakage on closure bolts.
The limited training that was given was direct-ed to journeymen, mechanics, and electricians.
The training aspects of the SOER were closed out by memo dated February 6, 1986.
The issue of borated water leakage on closure bolts was entirely missed and it is apparent that the FPL Operating Experience Feedback program was unsuc-cessful in this specific instance.
Safety Considerations for Station Restart An AIT was dispatched to Turkey Point Unit 4 to review the circumstances surrounding the discovery of large deposits of boric acid crystals on the reactor vessel head and in the CRDM ventilation system.
The boric acid deposits were the result of a leak in a conoseal flange connection which had been leaking from August 30, 1986 until March 13, 1987.
By letter dated April 27, 1987, the licensee submitted a report entitled
"Report on Instrumentation Port Column Assembly Leakage."
This report provides detailed information on the conoseal leak and a recovery plan and corrective actions for NRC review and concurrence prior to station restart.
After discovery of the boric acid deposits on the reactor head area on March 13, 1987, the licensee performed thorough and extensive inspections and evaluations to identify the extent of the components/equipment which
I i~
were subject to boric acid corrosion.
These included inspections of items in the area of the reactor vessel head as well as in the containment that might have been affected by the conoseal leakage (see paragraph 8).
The licensee's corrective actions are detailed in paragraph 14 of this report.
The team reviewed the licensee's analysis of the event and the recovery plan for restart of the unit.
At the same time, the team performed independent reviews in the following areas:
(I)
design of the conoseal connection; (2) procedures for assembly of the conoseal connections; (3)
reactor vessel materials and potential for damage by exposure to concen-trated boric acid; (4)
nondestructive examination procedures and pro-grams; (5)
reactor system leak rate calculations; (6)
containment radiation monitoring systems; (7)
the licensee's engineering analysis of this event; (8)
response to previous industry experience with boric acid corrosion, and (9)
the licensee's recovery program.
Three areas where the licensee was in noncompliance with Technical Speci-fication requirements and one area of programmatic weakness were identi-fied.
Two of the violations (Failure to properly evaluate the leak in terms of boric acid corrosion of ferritic steel components and failure to adhere to the installation and drawing requirements of the conoseal)
were the primary contributors to the event.
The third violation (leak rate procedure inaccurate in that correction factors were incorrect) did not affect the event because the leak rate program is not sensitive to small leak rates like the conoseal leak.
The area of programmatic weakness involved the licensee's response to industry experience.
This inspection determined that the licensee's program for recovery from the event was very comprehensive.
The AIT has reviewed the licensee's report and has conducted an independent inspection effort which is dis-cussed in this report and provides the bases for our concurrence with the licensee's restart plan.
The licensee's report is comprehensive in the area from once the event was discovered (March 13, 1987)
through the cleanup and evaluation phase.
The report does not specifically identify the cause of the event nor specify definitive corrective actions as indicated by Section 6.2, Improvements in Leak Detection, Evaluation and Repair, which states
"these events were then analyzed to determine the actions which mi ht 'e taken to improve leak detection, evaluations and repair at ur ey Point."
Discussions during the exit meeting on May 5, 1987, clarified when Turkey Point will take several additional steps to ensure leaks in borated water systems are detected as soon as (I)
Emphasis Added
C.
J V
possible.
These steps will be instituted for the Turkey Point plant (Units
and 4) commencing with this outage (before startup):
(1)
The accessible reactor head area will.be visually inspected for boron deposits and other evidence of primary system leakage whenever the plant is taken from Mode 2 to Mode 3 (prior to return to Mode 2) if an inspection has not been performed in the last 30 days.
Procedures wi 11 specify type of inspection and acceptance standards.
(2)
Current requirements for leak inspections wi 11 be revised to include appropriate leak inspections with acceptance criteria during RCS filling and venting operations of components inside containment that have undergone disassembly or maintenance.
(3)
Leak inspection procedures will be revised to add more specific instructions for inspecting components that could leak on the reactor head.
Moreover, the licensee amplified that leaks that occur in the reactor head area which can be retained by insulation or other structures so that the boric acid can accumulate and create corrosion problems will be promptly repaired once detected even if the unit needs to be shut down to complete the repair.
t In conclusion, the AIT concurs with the actions the licensee has taken or plans to take prior to station restart.
ATTACHMENTS 2.
3.
4, Chemical Analysis of Boric Acid Crystals dtd March 22, 1987 Table of Plant Parameters Reactor Coolant Leakage Typical Output Reactor Coolant Leakage Summary of Program Results
C.
C I
g4
~
FIGURES 1.
Thermocouple Penetration and Conduit Seal Assembly 2.
Reactor Vessel Closure Flange Pressure Boundary 3.
Reactor Head Closure Stud Location 4.
Ratio of B ppm/Li ppm 5.
Detail of Vessel Head Dome Wastage and Penetration No.
53 Counterbore Corrosion REFERENCES C. J. Czajkowski, "Investigation of Shell Cracking on the Steam Generators at Indian Point No. 3," NUREG/CR-3281, 'June 1983 WCAP-7693,
"Corrosion/Carbon Steel in Aqueous Boric Acid Solution,"
G.
D.
Toth and D.
D. Whyte, Westinghouse Nuclear Energy System.
3.
C. J. Czajkowski, "Bolting Applications,"
NUREG/CR-3604, May 1984 4.
Turkey Point Unit 4 Environmental qualification Evaluation of the Conoseal Leak, April 16, 1987
(.
r
ATTACHi'IENT 1 MY'K'RX)F FICE CORMSPONDENCE TO:
C,.
I KA~l eR(w: Z S'.SEA6 E.R
- uumDXW CnS,iaL raOrn u-~
CRhN LOCATION:
~PJ OATEt lAAlt!CN Z'7 l987 COPIES Ta:
F.R. La PIE.IUK Pal-hl t'= 6- 0C OM fAAKC,M Z l l957 A SA lYlPLC 0 F CRY STAL RF NOYES F ROPl THE.
M-4 CRDN VC..UTILATIOPJ SYSTEM-N
@PE S B,ROQG MT 70 THE.
CWEPI lSTP'f QC.PAR:TlAE.MT FOR RJUAVJS)S.
THE.
C.RVS TAL hlAD AM ORANGE. TINT ANb BURIN MLtE VEQ
RE.
8'OklC AC)D.
BRANS Of TWE.
CRYSTAL UUEkG. blSSOLVM lM COO enh
.. -::"".:i.,-'.sv'.-bc~uozemiza'n
-:1NN dz.'t t nff'faCKatQ'xiii"'"i~'~Xi~
LASOMTOt.Y AtUALYS I5-BOSOM 15 20 OS C f-c0.02~a
~,~a%+
> ~ '
~
+
f g ~
) )
~ jy,g g,,ty
~ ~
4 i, ('
4'
g. ~ ~
~,,Q) l
~4
~ I t"
'%
Ittehl, at trc l s csea t.. j,.goldian......;
TF YON W AVC.
ANY QUESTIONS CDNCE.RAIN & THE. RC.S,MLl5.
PLEASE.
CDIVTAC.7 R6.
AT K-Oft).
PEOPLE... SERVING PEOPLE Form 1008B {Stocked) Rw. 10/8S
Ap
ATTACHf1fNT 2 PARAMETER LIST Unit Identification:
Pl ant Name Unit Number Docket Number Nuclear Steam SYstem Supplier Vessel and Piping:
Volume Pressurizer:
Level Units Temperature Compensated Ca)ibration Curve Slope Upper Level Limit Lower level Limit Rel ief Volume Control Tank:
Level Units Calibration Curve Slope Upper Level Limit Lower level limit Geometric Method Available TURKEY POINT 50-251 Nestinghouse 8043 cubic feet
/
No 423.5 pounds per /
91 /
5 /
Rel ief Tank 125. 14 pounds per 100
%
0 /
No Drain Tank:
Level Units Calibration Curve Slope Upper Level Limit Lower level limit Geometric Method Available 33.32 pounds per
%
No Rel ief Tank:
Level Units Calibration Curve Slope Upper Level Limit Lower level limit Geometric Method Available 833 pounds per 76. 65 20. 4 No
p
ATTACHMENT 3 NRC INDEPENDENT NEASURENENTS PROGRAN REACTOR COOLING SYSTEN LEAK RATES STATIONs TURKEY POINT UNIT s
DOCKET s 50-251 TEST DATE s 9-30-86 START TINEs 2345 DURATION 2. 25 hours2.893519e-4 days <br />0.00694 hours <br />4.133598e-5 weeks <br />9.5125e-6 months <br /> TEST DATA System Parameters Pressure, psia T Ave, degrees F-Water Levels Initial 2250 574.03 2250 574.42 Pressurizer,
/
Relief Tank, /
Volume Control Tank, Drain Tank, /
52.72
.
71. 6 34. 31
52. 84
30. 38
Water Charged
= 0 gal
.Water Drained
~ 0 gal TEST RESULTS Change in Water Inventory in poundss Vessel h Piping Pressurizer Volume Control Tank Less:
Water Charged Plus:
Water Drained
'-229
(1 )
-492
0 Rel ief Tank (1)
Drain Tank (1)
Collected Leakage 333
Cooling System-670 Leak Rates in gpm (3):
Gross O. 60 Identi fied 0. 30 Unident ifi ed O. 30 (1)
(2)
(3)
Determined from tank calibration curve.
Determi ned from tank dimensi ons.
The density used for converting inventory change to leak rate was 62.31 pounds/cubic foot based on standard condition o t <
C-
.2
0 300 1"
1":i(3 I
"
"
250 Days Six.L"e St.artrq>
ATTACHMENT 4
FIGURE I THERlVlOCOUPLE PENETRATION AND CONDUIT SEAL ASSEMBLY JACKING SCREWS SPLIT AND JACKING RINGS PORT COLUMN
'EAL ADAPTER SEAL PLUG UPPER CONOSEAL UPPER PORT EXTENSION THERMOCOUPLE THIMBLES LOWER CONOSEAL
. PORT COLUMN r ~
LOWER PORT EXTENSION HEAD PENETRATION REACTOR VESSEL HEAD FfGVRE
l U
FIGURE
(g,g) coNvhoL hoD scccuAHILM HQUIIILIO!4QggCt(3)~
(3) LIPTIAIC Luau CauA<<v SPIC.CD C ICOiAPAIL'T(8Pcr)
1!I,D!A NOLC ('TYP)
(Ce) 4 CLOSI/RC STuDS.
COI/ALLYSPACCD Decl ITlg b.C.
l6V()l(3l o'lur
64
'biecth!CAL WAilcth
I I
g pccr. reer ac/s
,IshroON ~ CC &XIWO II
h t
hhe4 I +4/.
A JJe~clesc4,redeeeeeI At) MAC(l'S gO SPCOrO Ceer IJ'd'Per d('I~CCCg lee YA>Cr)
g 'p
- (4l5UPPDIIT LUGE KQVALLY'SPA(:CD ICOA APART ~!I(8)'>4 ICIPIC Boats ( Nvrs I4 cO I
(
a
'I.
N*. e I
I age (
I I
r C~
Sc oteAp P
-'0 C
ee Ie gx RI C'l
O t Cl mzmI7 IzO mI/I of
-.
I CLOOVILC HD. ASM
~al(CT CA/Tth aAWCCT OCTAILiD OWS. I IllIt MONITOhINOI
'TAP ge Q Ho(
I/'r
/Ar PIL
, ~
CO Crh S +Crt Iel e
aa(C r
~ 0 I4$
DIA.
5((
1(ICec ~dAI l'.AV~
d(C di1 rdflrn /II manning
~ g ~ ~
rr I OCTALiK r/WCI. II74TCt
~ Cr reek re err
~e r
0 It
Pl
FIGURE 3
)tTO ENPANOEO 84NPLE END OF EXAMINATION AREA CONTROL ROD DRIVE tlECHAHIStlS C ~ I )
C'V5) CONTROL RODS C IS) SHUTDOUN CONTROL AOOS C'V) COHO SthLS-TDTAL HEAD PENEThh'TIOHS CSS)
V-CS-OS V5 VT 0-VI VO V~ ~
'l-SL OS-VS SS V-LL-~5-y
/
SO Sl r
EXPANDED ST 5t r
shttrLE QEE rj I-
,1 t. J LIPTINC LUC DETAIL
)
c)> LICTINc I.Iles Eauhlt <
srhcED Al Ipo DEcREE y -0.- q I
oo0, QE
- - E
~
IIEII
/
O
~ II IIE START OF EXAMINATION AREA V LL-Ot 100
~
V-CS-Ot EIIEIIEE EII SANPLE r V-SL ~ I
'
CS) SurPORTINC LUGS EOUALLY 5rhCEO AT I I/IC 1[
)
lt0 DECREES VISUALLV Dhtth CEO THREADS INITIAL Ul.ThhS CHIC
'E 0htlrLE ENPANDEO SANPLE EL.,
fP-U surroAT LUc DETAIL J L.
,-
VENT LINE ISO o-j~-V-SL-St Q
tx r
V CS- ~ I ~
OS tt
- Qo o
I oooooooo" V-LL-Ol TURKEY POINT UNIT CLOSURE HERO CONOSEAL REPAIR PTN"'I-83 FIGURE
$~ C
~
g
~
W4
~
~
~ I I
~
~
~
~
~
I
~
I I
'I I
I '
I
~
~
~
I
~
~ ~
~
~
~
~
'
~
L
'
~ ~
~ i p,
~ gJ
>Ilk
t The sketch below is a plan view ooking down on the penetration.
The numbers circled are dfrnen s, the others are depth measuremen The "Counterbore" gap area had approx. 5/8" gap at the bottom, flaring to t/4", the edges being slightly rounded 4w KI ig 0 o The Hasted area Is the Conoseaf Housing. The other is the
'CounterbtaaP Gap between the head and the Coreseal housing.
+. %e Q A SQ Ctl QQ C4 O Q H 8.5 9/4" gap Distance ftem houshg approxhnateP 3/S" to fI2 Sight Erosion of Gap Edge rounding the FIGURE 5
pc V%
t
~ S
II