ML17202U984
| ML17202U984 | |
| Person / Time | |
|---|---|
| Site: | Dresden  |
| Issue date: | 02/06/1991 |
| From: | Lougheed V, Maura F, Phillips M, Salehi K NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | |
| Shared Package | |
| ML17202U983 | List: |
| References | |
| 50-237-91-06, 50-237-91-6, 50-249-91-06, 50-249-91-6, NUDOCS 9102130017 | |
| Download: ML17202U984 (12) | |
See also: IR 05000237/1991006
Text
U. S ~ .NUCLEAR REGULATORY* COMMISSION
- REGION III
Reports No. 50-237 /91006 (DRS) ; 50...,249/91006. (DRS)
Licenses No. DPR-19; .DPR-:-25
Licensee:
Commonwealth Ediso.n Company
Opus West III
1400 Opus Place
Downers Grove, IL * 6*0515* * *.
Facilit:y Name:.*. Dr'esden Nuclear Power station,. Units 2 and 3
- Inspectio.n At:
Dresden Site, Morris, Illinois
- In~pection Conducted:
January 23-28~ 1991
Inspectorl:j: .
- _~lJ1
...... 1'-+-/.*._. ;J
......._~ _____ * .
K.~J:f
. y;J~ LJ
~ . . p. ~(]heed
F .. A .. Maura
Approved By:~
.
- M:P:Phi1,....i---=-ip.,.... -s-, -c-h-=-i-ef
- operations Br_anch * ..
Inspection summary
Date'
1
- .
3/~lrr/
Date 1
- _)~ /9/
Dat!e.'
Date
Inspection* on Januarv 23-28, 1991 * (Reports No.** 50-237 /91006(DRS);
No. 50-249/91006(DRS))
.
Areas Inspected:
Special annotinc~d safe~~ inspection by regional
based inspectors of .the licensee's coritainment integ~ated leak
rate test (CILRT) *and local leak*. rate test .(LLRT) programs,
including review of *the events surrounding the failure of the
flange on the Torus Ptirge Exhaust Containment Isolation
Valve 2-1601-20A, and *the data acquisition system failure
during the verification portion of the CILRT~
Results:
The inspection resulted in one apparent yiolation
against Technical Specification 3/4.7.A.2 which requires that
containment integrity be maintained at all times when the* reactor
is critical.
Leakage from the*fla.nge, at accident pressure, was
conservatively calculated to be 65 weight percent per day, which
. was 41 .. times the allowable *. This apparent violation is described
in section 6 of ~his* report.
9102130017 910207
ADOCK 05000237
Q
Dresden Unit 2 had been granted a waiver of compliance until the
end of the cycle 12 refueling outage in regard to testing of the
Reactor Building.Closed Cooling Water System (RBCCW).
results during the'* refueling *outage *showed that this penetration
was acceptable during .the previous. ope_rating cycle.
The licensee is conside~ed to have failed the CILRT in the
as-found condition, due to leakage 'from the inboard* flange o*n
valve 2-1601-20A *
I
'
"
DETAILS
.1.
Persons Contacted-
Commonwealth Edison
- E. -Eenigenburg, Station Manager
- J~ Geiger, Technical Staff, CILRT Lead Test Engineer
- J .. Glover, Nuclear "Engineering Divi.s.ion, Corporate
CILRT Lead *
.
.
- *L. ' Gerner, Technical Superintendent
- M. Horbaczewski, T~chnical Staff Group Leader
- G. Kusnik, Quality Control
- D. Lowenstein, Regulatory Assura~~e Analyst
- K. Peterman, Regulatory Assurance Supervisor
- R. Williams, Senior Quality Assurance Inspector
- K. Yates, Onsite Nuclear Safety Administrator
Illinois. Department of safety .
- R. Zuf.fa, .Resident Inspector
U.S .. NRC
- P* Hills, $enior Resident Inspector
- M .. Peck, Resident Inspector
- J. Monninger, Re$ident Inspector,*NRR
- Attended exit.on January 28, 1991.
The inspectors-also interviewed other licensee employees*
during the course of the inspection~ including members of
the maintenance,'. quality. *control, and technical staff.
2.
Licensee Action on Previously Id~ntified Findings
a.
(Closed) Violation 50-249/90005-01 "Failure to Include
Service 'Air and Demineralized Water in Valve *Lineups":
The licensee revised the CILRT procedure to include
the above systems in the valve lineups.
The inspectors
reviewed the revised.proced~re*_and the test results and
- had no further concerns.
This item.is *considered
closed.
b. *:(Closed> Unresolved Item 50-249/90005-02 "Status of
Small Diameter Pneumatic Tubing-Penetrating Drywell":
The licensee has completed their. evaluation of all
- containment -isolation boundaries and is in the process
of performing a final review prior to submittal to.NRR.
The particular lines in question were handled by
cutting and cappi.ng them, under the licensee's minor
design change program.
The work was accomplished on
Unit 2 during*the Wir:it~r.1990 refueling outage, and
.. ,
- Cl
will be performed on Unit *3 during the*spring 1991
refueling outage.
The inspectors. reviewed the packages
and.discussed the post-modific~tion testing
requirements for *unit 3 with the licensee *. * (Post-
modifi9ation testing for Unit 2 *was accomplished
through the CILRT.)
The -inspectors had no problems
with'the licensee's solution for*th~ pneumatic lines.
This item is consideied closed.
3 -~*
Revi*ew of co~taihment Integrat~d *Leak Rate Test
Proc*edure (70307)
The 'inspector~ reviewed .*surveillance instruction DTS 1600-7,
"Unit 2/3 Integrated Primary *Containment Leak Rate Test,"
Revision 11, dated December 13, 1990,.relative to the
.z:equirements of _10 CFR Part 50 Appendix *J,. ANSI N45.5-1972,
and th~ licensee** s Technical -Specif icatioris. . The valve
- 1ineup problems and fa'ilure *to require venting of some
systems as desci;-ibed in .. Inspection 'Reports
No. 50-237/90006(DRS); ~o. 50-249/90005(DRS) had been
resolved in this revision.
~he inspectors dia not identify
any further probl~ms with the procedure.
4.
Test Re~tilts ~vaiuation (70323)
a.
Review of Data Acquisition Problems Encountered During
the CILRT
During the pressurizat.ion' phase the licensee
experier1ced *a failure of* their Data Acquisition System
(DAS) due to the loss of incoming data information.
.
- .:The licensee attributed this to wetting of 1;he amphenol
, connector. to the DAS computer.
.This same failure also
occurred during the verification portion of the CILRT,
causing the l.icensee to terminate the supplemental
test.
The inspectors discussed the cause of the
failure ~ith the licensee~ .. The apparent root cause was
due to a soap-water soluti.ori used to identify leaks
wetting the connector and *causing it to* *short. out.
The
soap-water solution was used to. check the* fittings on
the CILRT instrumentation penetrations.
Th*ese
penetrations, which exit containment through the
.
personnel airlock, were a source of leakage during the'
last Unit 3 CILRT~
In order to- ensure that a similar
leak did not. occur durin*g. the Unit 2 test, the licensee
used a new Conax fitting, which provided a tighter
se~l, and t6en checked 'for leakage using the soap-water
mixture.
The licensee state~ that excess solution ran
down the cable from t.he fitting to the connector.*
Since the connector was not designed to be water-proof,
the accumulated moisture caused the connector to fail *
- Although the licensee carefully.dried the cable and
connector when *the.failure first occurred during
containment pressurization, the licensee speculated
..
b.
. c.
that sufficient ~~isture ag~in 6ollecied to cause the
failure during the supplemental .test.
The inspectors
also discussed with the-licensee the .intended
preventative measures to ensure that similar problems
did not occur in the future .. The inspectors reviewed
the collected ,.data during both failures and verified*
that it was not isolated sensor* failures, but a
- complete failure of al'i sensors during :that time
period.
The licensee stated that they were*considering
tighter.controls on the use of-soap~water solutions on
electrical.connections as well as providing a
water-proof connector. - The inspectors had no further
questions in this area.
CILRT Data Evaluation
A six hour and ten minute CILRT was performed on Unit 2
on December 17. and 18,. 1990_. at a test pressure of 64. 4
psia following satisf~ctory completion of the required
. temperature stabilization period.
During the
-pressurization phase, the licensee experienced
excessive leakage from the inboard flange of the Torus
Purge Exhaust Containment Isolation Valve, 2-1601-20A.
An evaluation of this l.eakage is described in section 6
of this report~
Data were collected every 1.0 minutes.
The inspectors independently evaluated leak rate data
using total time
(BN~ToP-'1) formulas to verify the
- licensee's calculations of the leak rate and instrument
performance.
There was good agreement between *the
inspectors' and licensee's results as indicated by the
following summary (units are -in weight percent per
day).
Measurement
Measured leak rate during
CILRT (Lam)
Lam at 95 percent Upper
.
Conf iqence Level (95% UCL) *
Licensee
Inspector
0.500
0.500
0.743
0.746
.The 'Appendix J acceptance criterion is that Lam, at the
95% upper confidence -level (UCL), be less than 0.75 La
(i.2 wt%/day).
The test.met this crite~ion.
supplemental Test Data Evaluation
After satisfactory completion o.f . the CILRT, a known
),eakage rate of 13 scfm, equivalent to 1.53 wt%/day was
induce*d *. _The inspectors noted ~hat the licensee
corrected* the reakage rate for containment temperature
and volume changes at the time of the test.
These*
corrections were conservatiye in nature and resulted
'
_.
_, ..
.;,
in.a higher flowrate than the methodology used by the
inspectors.
Since both the inspectors' and the
licensee's calcui'ated induced leakage rates were within
the band specified by Appendix.J, the method was
- . considered acceptable.
The Bechtel test method, described in BN-TOP-1,
requires a one hour stabilization perioc:l* b.etween the
time when *the known flowrate is induced and when the
supplementai'test*is started.* It also requires that
the supplemental test be at least half of the main test
in length.
The licensee commenced the supplemental
test following the one hour stabilization period
required by BN-TOP-L
Data were collected and analyzed
by the licensee every 10 minutes.
After approximately
two and a ha:J.f hours, the licensee's instrumentation
malfunction~d.
Th.e cause of this malfunction is
described in sectiori 4_.a above.
In order to obtain the
required* three hour verification period, the. 'licensee
requested from Region III that they be allowed to
include the one hour.stabilization period in the
supplemental test data.
The inspectors calculated 'the
supplemental test leakage rate, both with and without
the stabilization period data.
All data units are in
.weight pe~c~nt_ per day (wt%/day) *
Measurement
Measured leakage rate (Le)
during supplemental test,
stabilization data not
included*
Measured leakage rate'. (Le I°) _
during supple.mental test, .* ... :
stabilization data included
Induced Leakage Rate (Lo)
Results from main test _
(Lam)
.
Le*- (Lo+ Lam)
Le '
..:. (Lo + Lam)
Licensee
Inspector
1. 756
1. 754
1.814*.
1.812
1.556
1.533
0.500
0.500
-0.300
'.""'0.279
-0.242
-0.221
The.Appendix J acceptance criterion is that the value
of [Le -
(Lo+ Lam)) be within a band of+/- 25 % of La.
For Dresden, this results in an acc~ptance ~riterion of
.;..0.400 < [Le -
(Lo+ Lam)) < 0.400.
The supplemental
test results fell within the band, for both the cases
using the stabilization data ,and those without.
.
-,,
d.
.
The inspectors noted:that the last five data points
during the s_upplemental test* showed a decreasing trend.
Normaliy, *.the NRC expects the verification* results to
stabilize within the band before the test is .
terminated.
In this case,* since.the test was
terminated due to *equipment malfun*ction, the inspectors.
agreed, for this supplemental test only, to accept the
results as valid as long as all.of the points remained
.within tbe acceptance band ..
. CILRT Volume Change Corrections
'
At.the completion of the CILRT and the supplemental
test, the licensee was required to make corrections to
the calculated Lam at the 95% UCL due to changes in
volume of *various* water sources inside containment.
The following.corrections to Lam were recorded and
calculated by the licensee, with the calculations being
verified by* the ,inspectors:
-Water Source
Change in Volume
Equipment Drain Sump
.48 cubic feet
Floor Drain sump,
4*a cubic feet
prywell_ -. Floor
146 cubic feet
The' level at the beginning of the CILRT was 2.67 feet
in both sumps, and at the* end of the test both sumps
were full, with an additional inch of water on the
drywell'basement floor.
The* sumps are six-feet by six
'feet by four feet deep.
The drywell basement has a
radius of 23 .. 6 feet. *The licensee had not identified
the source of the water at.~he time of the inspection.
The total volume change due to water inleakage was
242 cubic feet which corresponds to o.*020 wt%/day.
This results in a total leakage, at the 95% UCL, of
o.766 wt%/day, which is under the maximum allowable of
1.2 wt%/day.
e.
CILRT Valve.Lineup Penalties
Due. to vaive:*configurat.j.ons which deviated from the
ide_al penetration v~l ve 1 ineups for the CILRT, the
- I
I
~'
,: f
. '*
f .
res.ul ts of LLRTs for ~uch* penetrations must be added as
a penalty to Lam at the 95% UCL, per Appendix J *. The
licensee h~d the following penetrations in a
configuration which differed from that which would be
experienced post acciden*t:
X-101
- X-107A,B
X-116A,B
X-119
x~l49AiB:
X-15bA
X-109A B
.
I
X-122
X-310A,B
X-lllA,B
X-138
X-JllA,B
X-113
X-:J..45A,B
- Addition of the as-left minimum pathway LLRT result for
the non-vented penetrations added a penalty of 25.9
scfh to the 95%' UCL limit.* .This leakage was equivalent*
to 0.051 wt%/qay, resulting in a final as-left limit of
0.817 wt%/day."
This value was within the acceptance
'criterion (Lam< 1.2 wt%/day).
The licensee added an
add~tional 2 scfm penalty to its calculations~
~his
was-due*to the CILRT procedure listing service air as
-~*non-vented penetration.
Service air was properly:
vented during the CILRT, however the licensee failed
to remove it from th~ table,*~nd decided to take.the
"additional penalty as *<<;i conservative measure *. This
- penalty* added approximately* o. !)04 wt%/day to the total
and was negligible.
As-Found Condition of Containment
The* as-found condition is* the co.ndition of the
containment at the beginning of the outage prior to any
repairs .or: adjustments to the containment boundary!
This is calculated by reviewing the summary of the
LLRTs and calculating:the amount of leakage rate
improvement dq.e to .r~pairs ;or adjustments using the
minimum pathway methodology.
The inspectors reviewed
the licensee's local leak rate re~ults to ensure that
the minimum pathway *repairs and adjustment calculations
were correctly performed. * The inspectors determined
that these corrections.were worth 140 scfh which
corre.sponds' *to O ;*275 *wt%/ day.
Adding this to the final
as-left tot:al would, have resulted in an as-fo.und leak
rate of 1.092 wt%/day, exclusive of the 2-1601-20A
flange leakage.
Howev:er, due* to the excessive leakage
- from the flange, the CILRT ;was considered to have
failed in the as-found condition.
5.
Review of Local Leak Rate Test.Res~lts (61720)
a.
.R~v.:i..ew of* Proposed Dresden Primary.Containment Pathways
Document
- *
The inspectors performed a cursory review of *the
licensee's -proposed DresdEm* primary contai,nment
0
pathways document issued by the corporate (Nuclear
Engineering Division) off.ice on November 19, 1990.
The
licensee stated that the current status of the document
was that it was being reviewed by the on-site technical
staff.
Following this* review, the document would be
reviewed.and* approved by the Onsite Review Group.
It
would then*be submitted to NRC -
NRR for review and
approval concurrently.with the .Quad Cities document.
The inspectors noted. that this document failed to
identify the requirement.to test the inboard flanges
of the purge . valves. _ No other problems were noted.
b.
Review of Local *Leak Rate Test Results
The* inspectors i::-eviewed* the licensee's LLRT procedure,
DTS 1600-1, "Local Leak* Rate Testing of Primary
Containment Isolation Valves," Revision 14 dated
September 24, 1990.
The inspectors noted that the
.licensee had completely revis.ed the d.ocument to resolve
the concerns discussed .. in Inspection Report 50-
237/90006 (DRS); 50-249/90005(DRS).
The inspectors also
reviewed the licensee's- LLRT data. *The licensee
performed the majority. of the LLRTs by using the flow.
makeup method.
This allowed *.for more accurate
representation of the* actual penetration leakage rates.
The inspectors noted that the licensee still tended to
p*ressurize the penetrations in excess of Pa ( 48 psig) .
This was especially true when the licensee had to
account*for a water head on one side of the penetration
and *was'using an increased test pressure to.overcome
the water pressure, which often resulted in a
differential pre~sure of- greater than Pa on the
This was discussed with the licensee, and
the licensee agreed to keep future -tests closer to Pa.
The inspectois particularly revie~ed the results for
penetr~tions x~123 and X-124 for the Reactor Building
Closed Cooling Water '(RBCCW) system.
The licensee had
obtained a wa.iver of Compliance for the previous outage
from including these penetrations in calculating their
0.6 La total. *The inspectors noted that these
penetrations. ha*d acceptable LLRT results ( o. 4 and
14.6 scfh, respectively, in.the as-found condition.)
The inspectors had no further problems in this area.
'.
6.
Review of Ev.ents surrounding the Failure of Valve 2-1601-20A
(93702) .
On December 17, 1990, the* licensee began the pressurization
of containment in order to perform the CILRT.
A review of
the test log and discussions* with test personnel indicated
.. that at approximately 6: 30 on tbe morning of the 17th, with
the containment pressure at approximately 15 psig, a* loud
"pop" was heard, followed by a *~~siren-like sound".
Upon
- '.
further inspection1* the licensee found the. leakage to be
.
occurring past the inboard.flange on the Torus Purge Exhaust*
'Containm~nt Isolation Valve 2-1601-20A.
The licensee
stopped-the air compressors, obtained approximately.ten data-
point~ prior.to tightening the flange bolts thereby *
eliminating the leak, and recommenced the CII,.RT
pressurization~ The licensee s.ubsequently *e:o:rnpleted the
CILRT with satisfactory results as discussed in section
- ,4. o,
above~ :The inspectors had no concern . with these short-
.. term corrective actions.**
On December.18,* 1990, the*licens~e reviewed. the maintenance
history on the 2-1601:-20A v~lve and.determined that the
valve had not been w.orked on *since the last. refueling
outage, at which time the valve was replaced.
Upon making
this determination, the lic~nsee reported the event to the
NRC under. 10 CFR 50. 72 (b-) (2). (i).
This was subsequently
followed by.Licensee Event* Report LER-~0-018 which was
issued on January .i4, 199_1_.
The inspectors review*ed the work history of the valve and
The valve was replaced under nuclear work request 67528 in February 1.989.
Fol.lowing replacement of the valve,
a LLRT was performed by pressurizing between the 2-1601-20A
valve and check valve 2-1601-31A.
This test, which was the
licensee's normal LLRT method, tested the 2-l.601-20A valve
seat, p~cking,*and outboard flange as well as operation of
the check valve.
How~ver, it .failed to test the inboard
flange as required by paragraph IV.A of 10 CFR Part*50,
Appendix J.
This testing omission was no.t recognized by any
of the licensee personnel involved in development of the
work package, maintenance work_ activities, or conduct of the
post-maintenance t~st.
Following satisfactory completion of
the LLRT, as well as other required valve surveillances, the
valve was returned to operation. * On ,February 19, 1989,
Dresden.Unit 2 went critical following completion.of their
refueling out~ge.
The plant remained basical'ly in operation
until they shutdown for refueling in September 1990.
In order to obtain an estimate of *the leakage through the
flange, the inspectors calculated the leakage rate at
15 psig* using the ten data .points supplied by the licensee.
Using point to point test methodology, the*inspectors
calculated that the leakage rate through.the flange was
within the range of 23.8 to*_.24.6 wt%/day at a* test pressure
of 15 psig.
The inspectors then-corrected the leak rate
to the design pressure*(48.psig).
This involved both a
correction to the test mass and.pressure.
In regard to the
mass change, the calculation was base¢! on the ideal ga's law.
This result was then utilized in the pressure correction
calculation.
In.this.case, calculations were performed
using both a formula described in the licensee's local leak
rate procedures, which., yielded a result of 31 wt%/day; and a
formula developed by the Franklin Rese~rch Institu.te for the
.t
NRC, which yielded a result of 65 wt%/day.
~he later
formula assumes a capillary, or non-orifice, type leakage,
such* as would be the ~ase for leakage through a flange, ahd
the former Gase .~s based on flow through an orifice.
Neither calculation assumed that the size of the opening
- would *enlarge due to increasing*pressure, although that
might be physically.realistic,* if the gasket material was
displaced due. to the, .increased air flow at higher pressures.
At the conclusion of the inspection, the licensee was in the
_process. of performing calculations to better quantify the
leakage at.design pressures.
Th~ inspectors noted that all.
of these values were considerably in excess of the design
allowable of 1.6 wt%/day and that either method of
extrapolating the leakage rate to design pressure resulted
in undesirable leakage rates.
At the time. of the inspection, the licensee had not
completed fts evaluation of the ef.fect of these leakage
rates upon 'the dose releases following a design basis loss
of coolarit accident~* Although direct *1inear interpolations
are not necessarily accurate, the inspectors noted that the
flange leakage rates were five to ten times greater than
tho~e utiiized in th~ dose calculations presented by the
licensee during an enforcement conference concerning opening
ot sampling valves held on October 12, 1990 (EA 96-168).
Those results .are contained as attachments to Dresden
Inspection Reports No. S0-237/90025(DRP);
No. 50-249/90024(DRP).
The inspectors*conducted numerous interviews to determine
- the root cause of the failure.
This event :appeared to
originate in the maintenance program involving installation
of flanged valves.
The inspectors noted that there were no
acceptance criteria for tightening of the flange bolts as
they w,ere replaced.
The licensee had considered this
activity to be within the realm of "skill of the craft."
Discussions with the maintenance department indicated that.
the *bolts were tightened through use of a slugging wrench.
This involved tightening the bolts by hammering' the wrench
until the bolts'would not turn any further.
A.maintenance
worker also indicated th.at there was not a good fit between
the pipe and the valv~ at the time of installation that may
have contributed to the problem.
In addition, as discussed
above, the post-maintenance test did not challenge the
flange so that the. failure to completely tighten the flange .
. was not identified subsequent to* .,the valve installation and
prior to its being returned to service.
Although the
licensee installed eight similar valves with sixteen similar
flanges during the prior refueling outage, this was _the only
flange that leaked.
7.
. I
The inspectors noted_ that the iicensee, at the Quad Cities
N:uclea:r Station, had previously been notified that the
inside flanges of the purge* valves were part of the
containment boundary and that Type C testing* between the
containment .isolation :valves did.not challenge these
flanges. (See Inspection~eports No. 50-254/89024(DRS);
No. 50-2G5/89024(DRS) .issued on June 15, 1990).
Since Quad
cities was on an accelerated test program such that it
performed a CILRT every outage, no actiorr was required at
that site.
During the Dresden inspection, the inspectors*
~pecif ically discussed _with the licensee the. Appendix J
requirement to test the flanges was **similar to that for
other containment isolation valves and needed to be
performed every refueling outage rather than just following
maintenance. * The license~ acknowledged this requirement and
stated that the flanges would be properly tested. during *
. 'future* outages.
Te'chnical Specification 3.7.A.2 requires, in part, that
pi;-imary containment integrity be maintained whenever the
reactor is critical. Section 3.7.A.2.a.(3) defines the
maximum allowable leakage rate (La), at a. pressure o.f Pa,
as equal to 1.6 percent by weight of the containment air
per 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> at 4B psig.
Technical Specification section 3.7.A.2.b states, in part, that, when containment
integrity is required, primary leakage rates will-be limited
. to an overall integrated leakage rate of less than or equal
to 75 -percent of La.- Operation of Dresden Unit 2 over a
'complete operating cycle from February 1989 until
September 1990, with the inboard flange of containment
isolation valve 2-1601-20A having a calculated leakage of
approximately 65 wt%/day at 48 psig, is an apparent
.*violation of Technical Specification 3.7.A.2
( 50-237 /91006--01 (DRS)) * .
.
Exit Interviews
The inspectors met.~ith licerisee representa~ives (denoted in
section 1) throughout _the inspection.
An exit meeting was
held prior to leaving the.site on January 28, 1991.
The
inspectors summarized the scope of the inspection and the.
apparent findings.
The* licensee.acknowledged these
findings.
- . The inspectors also discussed the likely
informational content of the insp~ction report with regards
to documents or processes reviewed by the inspectors during
the inspection .. The licensee did not identify any such
documents or processe~ a_s proprietary .