IR 05000528/1995098
| ML17311B289 | |
| Person / Time | |
|---|---|
| Site: | Palo Verde  |
| Issue date: | 11/22/1995 |
| From: | Vandenburgh C NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION IV) |
| To: | |
| Shared Package | |
| ML17311B288 | List: |
| References | |
| 50-528-95-98, 50-529-95-98, 50-530-95-98, NUDOCS 9512060187 | |
| Download: ML17311B289 (84) | |
Text
ENCLOSURE U.S.
NUCLEAR REGULATORY COMMISSION
REGION IV
Inspection Report:
50-528/95-98 50-529/95-98 50-530/95-98 Licenses:
NPF-41 NPF-51 NPF-74 Licensee:
Arizona Public Service Company P.O.
Box 53999 Phoenix, Arizona Facility Name:
Palo Verde Nuclear Generating Station.
Units 1,
~
and
Inspection At:
Tonopah.
Arizona Inspection Conducted:
August 3 through November
~
1995 Inspector:
C. J.
Paulk, Reactor Inspector.
Maintenance Branch Division of Reactor Safety L. J.
Smith. Reactor Inspector.
Engineering Branch Division of Reactor Safety Approved:
an en urg
.
ie
.
ngineering rane Division of Reactor'afe y
LL-Z2 T5 a e 95i2060i87 95i20i PDR ADQCK 05000528 G
I,(
EXECUTIVE SUMMARY
.
DETAILS
.
INTRODUCTION
.
SYSTEM CONFIGURATIONS TABLE OF CONTENTS lv
EXPERIENCE REVIEW 3.1 Licensee Use of NRC Generic Communications
.
3.2 Use of Licensee Experience
.
.
.
.
.
.
.
.
.
3.3 Vendor Experience 3.3.1 Condensate Removal
.
.
.
.
3.3:2 Governor Valve Stem Corrosion 3.3.3 Use of Vapor-Phase Inhibitors in Oil 3.4 Terry Turbine User's Group
.
GOVERNOR SYSTEMS
.
4. 1 Background
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4.2 Governor System Designs 4.3 Timing Issues 4.4 Periodic Testing
.
4.4. 1 Standby (Cold Quick-Start) Turbine Test Requirements 4.4.2 Governor System.
Governor Valve, and Steam Supply Valve Performance Trending
.
4.5 System Design Verification Testing
.
.
.
.
.
.
.
.
.
.
.
.
.
4.6 Governor Preventive Maintenance Practices 4.7 Governor Modification Control
,5 GOVERNOR VALVES 5. 1 Background
.
5.2
,Stem Materials
.
.
.
.
.
.
.
.
.
.
.
5.3 Validation of New Stem Materials
.
5.4 Governor Valve Maintenance Practices
CONDENSATE CONTROLS 6. 1 Background
.
.
.
.
.
.
.
.
.
.
.
6.2 System Design 6.3 Steam Supply Valve Leakage
.
.
.
6.4 Operating Status Information
.
.
6.5 Safety Classification 6.6 Steam Trap and Drain Maintenance
LUBRICATING AND HYDRAULIC OILS
.
7.1 Background........
7;2 System Designs......
7.3 Aeration
.
7.4 Oil Filtration
.
7.5 Oil Viscosity Requirements
'.6 Maintenance Practices
3
6
7
10
10
11
12
14
15
16
18
19
19
21
22
22
23
24
4'l~
'>I'I l~
,<J
, l!
t If r I I
TABLE 1
- TURBINE SURVEILLANCE REQUIREMENTS...
TABLf 2 - TURBINE DESIGN VERIFICAIIGN TESTS
.
TABLE 3 -
GOVERNOR SYSTfMS TABLf 4
-
CONDENSATE CONTROLS TABLE 5 - LUBRICATING AND HYDRAULIC OILS ATTACHMENTS:
,ATTACHMENT 1 - Persons Contacted and Exit Meeting ATTACHMENT 2 - List of Acronyms
27
29
111
\\
I
EXECUTIVE SUMMARY This inspection was performed using Inspection Procedure 93801.
"Safety System Functional Inspection."
The primary objective of this inspection was to assess the operational performance capability of safety-related steam turbine-drivers.
which were supplied by the Terry Corporation.
This initiative was prompted by recurring. fai lures of safety-related steam turbine-driven standby pumps at several facilities in Region IV.
An in-depth engineering review was performed concurrently at several Region IV facilities to assess the scope of design.
maintenance.
and testing practices related to these safety-related steam turbine-drivers.
Previously identified generic safety-significant
,
findings were pursued at each facility.
The inspection examined several aspects of applicable experience review processes to determine why similar failures continued to occur.
The inspection found a wide variation in system designs, which has reduced the effectiveness of NRC generic communications related to Terry turbines.
Further.
the inspection found that most facilities did not have a programmatic requirement to formally review NUREGs for applicability to thei r facility.
As a result, many licensees had not evaluated NUREG 1275, Volume 10.
"Operating Experience Feedback Report
- Reliability of Safety-Related Steam Turbine-Driven Standby Pumps," to identify failure mechanisms and potential actions, which could be taken to prevent the failures.
In additions many licensee personnel stated that the turbine vendor has not provided a good focus for emerging technical issues.
The inspection also found that licensees were not consistently implementing vendor recommendations.
While the Ter ry Turbine User's Group was attempting to work with the vendor to provide a nuclear focus for technical issues.
these licensees indicated that the user's group cannot be relied upon to solely solve the problems'ecause they do not represent all licensees.
As a results the inspection found that similar steam turbine failures and problems continued to occur.
Most licensees did not rigorously address vulnerabi lities until thei r equipment was directly affected.
For example, the importance of condensate removal for operation of the steam turbine-driven safety-related standby pumps has not generally been understood fully unti 1 after experiencing a mechanical overspeed trip.
Similarly, an industry-accepted root cause for corrosion-induced governor valve stem sticking has not been determined.
even though approximately 18 failures of this type have been observed nationally.
In additions the inspection found that licensees were not consistently monitoring the governor valve stems for sticking or consistently replacing the stems with a material which 'was less susceptible to corrosion.
In general the inspection found that licensees'id not maintain the reliability of safety-related steam turbine-driven standby pumps with the same
1Arkansas Nuclear One. Units 1 and 2 (ANO 1 and -2). Coooer Nuclear Station (CNS).
Comanche Peak Steam Electric
)tat)on, Un)ts 1 and 2,(cP)Es.)
and -2}
Plaojo canyon Po<<er Plant. Units 1 and 2 (ocPP-1 and -2). palo verde Nuclear uenerat)na station Upits
an(w 3
~Pv )M >..-z and -3)
)(:ver'Bend stat)op (R )s). san onofre'Nuclear Geperat)no stat)On, Jn)tSj and
~
t>OAGS..2 an-3 Ourn'eXaS PrO3eet.
Un)tS 1 )lnd '
iP-1 aOO -2),Naterfprd Steam r.leotrig hat)on', unit 3 (Noir ), <<'asnington Nuclea~ ~o'r.z (NNP-F).'and '<<ol. iree<< Nuc)ear 8eneratlng stat>on (fC))GS 1V
J,
~
rigor as other safety-related equipment, such as the emergency diesel generators.
For example:
licensees had not fully tested the existing designs at pressurized water reactors to support extended operation under station blackout conditions:
the system designs had not always. included instrumentation or alarms for alerting the operators to steam-line drain fai lures.
which could prevent the standby pumps from performing their safety
.
function; and routine surveillance testing had not always detected degradation of the safety-related standby steam turbine-drivers.
Detailed Plant-S ecific Results:
Maintenance Licensee personnel at eight units (ANO-2; CPSES-1 and -2; PVNGS-l. -2.
-3: and STP-1 and -2) relied on vendor technical information to perform governor valve maintenance (e.g..
maintenance practices for stem packing instructions.
valve bonnet alignment pins, and valve stroke/linkage adjustments)
and found that the vendor manual did not always provide sufficient detail.
The vendor representative stated that they had prepared the technical information for maintenance with the assumption that an experienced turbine professional would be onsite di recting the maintenance activity.
The vendor indicated that the licensee was responsible for procuring expert technical services if they did not have a turbine professional on staff (Sections 4.4.2 and 5.4).
Licensee personnel at eight units (CNS; CPSES-1 and -2; DCPP-1 and -2; WAT-3: WNP-2:
and WCNGS) had not established a preventive maintenance requi rement to refurbish or replace the standby steam-turbine governors on a periodic basis.
At the exit interview personnel from CPSES stated that they planned to establish a preventive maintenance program for governors (Section 4.6).
The inspectors found that licensee personnel at two units (WAT-3; and WCNGS) had not established a periodic preventive maintenance program for steam traps.
At the exit interview personnel from WAT-3 stated that they planned to establish a preventive maintenance program for steam traps.
In addition.
none of the licensees had established a preventive maintenance program for drains (Section 6.6).
Personnel at four units (CPSES-1 and -2:
and STP-1 and -2) removed their steam traps after experiencing condensate-induced mechanical overspeed trips caused by poor preventive maintenance programs.
Personnel at PVNGS-l, -2, and -3 upgraded their preventive maintenance programs for their steam traps after experiencing condensate-induced mechanical overspeed trips (Section 6.6).
Personnel at three units (CNS.
RBS and WAT-3) did not filter thei r turbine and governor oil through a 5p filter prior to adding to the system (Section 7.4).
0, I
En ineerin As a result of the lack of uniformity in system design and system complexity, licensee engineers at eleven units (CPSES-1 and -2;
~
DCPP-1 and -2:
PVNGS-1.
-2 and -3:
SONGS-2 and -3; WAT-3, and WNP-2) did not always correctly evaluate NRC information notices related to steam turbine-driven standby pumps (Section 3. 1).
While NUREGs (such as NUREG 1275.
Volume 10) were sometimes routed for review. the inspectors did not identify any licensees that routinely documented experience review associated with NUREGs (Secti.on 3. 1).
The inspectors determined that only one facility (ANO) was monitoring success-on-demand (including surveillance test results),
which is an important indicator of turbine reliability. for comparison with the reliability estimates used in probabi listic risk assessments (Section 3.2).
The inspectors found that the drain conAgurations at three units (ANO-2; and DCPP-1 and -2) were not consistent with the vendor recommendation to continuously drain the steam lines.
In addition. the turbine casing steam trap at WAT-3 was designed to allow a small amount of water t'o stand in the turbine casing following a turbine run.
This also conflicted with the vendor recommendation that drain lines remain open when the turbine is idle to prevent corrosion of internal parts (Section 3.3. 1)
The inspectors found that oils with vapor -phase inhibitors were used at nine units (CNS; DCPP-1 and 2; PVNGS-1.
-2.
and -3; RBS:
and SONGS-2 and 3).
Further, the inspectors noted high turbine standby temperatures at SONGS, which made them the most susceptible to problems associated with the out-gassing of the vapor-phase inhibitor (Section 3.3.3)..
'icensee personnel at several facilities stated they relied on the turbine vendor (Dresser-Rand, Terry-Turbodyne, a joint venture) to evaluate emerging technical issues from a nuclear perspective.
However, the vendor provided recommendations from a commercial perspective and.
did not conservatively and promptly identify issues related to nuclear applications to all affected licensees (e.g.,
condensate removal, governor valve stem, and use of vapor-phase inhibitors in oil)
(Section 3.3).
Licensee personnel at three units (CNS; and DCPP-1 and -2) were not members of the Terry Turbine User's Group.
A Terry Turbine User's Group officer estimated that. nationally, 30 percent of the utilities were not members (Section 3.4).
Licensee personnel had not proceduralized the requirement for cold-start testing at four units (ANO-1 and -2; CNS: WAT-3).
However.
personnel at these facilities stated that they do test from the standby condition.
Personnel at ANO record the turbine standby temperature prior to each run (Section 4.4. 1).
Licensee personnel had not performed any type of dynamic governor valve performance trending at three units (CNS
~
~ WAT-3).
Although personnel at two units (RBS and WAT-3) had used manual valve manipulation to detect stem binding.
as discussed in Information Notice 94-66, this testing was not predictive at STP (Section 4.4.2).
~
The licensees had operated the safety-related steam turbine-driven standby pumps at only three units (PVNGS-1. -2, and -3) for an extended period of time in a configuration representative of a total loss-of-alternating current to the safety-related steam turbine-driven standby pump and supporting equipment (Section 4.5).
Based on a review of the safety analysis reports.
the inspectors found that licensee personnel at nine units (ANO-1 and -2:
CPSES-1 and -2; SONGS-2 and -3; STP-1 and -2; and WCNGS) had not demonstrated.
by testings that the safety-related steam turbine-driven pumps were capable of running over the full range of steam inlet pressures (Section 4.5).
O.
Licensee personnel at two units (CNS and WCNGS) were using incorrect assumptions to determine whether, or not. they had the liquid-nitrided governor valve stems (Section 5.2).
At the time of the exit, potentially suspect stems were installed at four units (PVNGS-1,
-2, and -3; and WNP-2).
Personnel a't three units (PVNGS-1,
-2 and 3) planned to replace the stems with a material which was less susceptible to corrosion.
At the exit interview, personnel at WNP-2 stated that they planned to inspect the installed governor valve stem during the next outage (Section 5.2).
Licensee personnel at the WCNGS had no plans to restrict the use of the corrosion susceptible spare governor valve stems'hich were in stores (Section 5.2).
The inspectors identified that periodic inspection and dynamic testing capability are important to demonstrate the long-term acceptability of the new stem materials (Section 5.3).
~
The inspectors determined that the system designs for seven units (ANO-1 and -2:
DCPP-1 and -2; and, PVNGS-1,
-2.
and -3) did not include any instrumentation or alarms to alert the operators to steam-line drain fai lures.
In addition, only personnel from RBS and WNP-2 had installed alarms or high level indicators for the turbine or turbine exhaust side steam traps or condensate pots/drains.
(Section 6.4).
Licensee personnel.
generally, did not recognize that steam-line drains had a safety function to remove condensate until after discussing a
condensate induced overspeed trip with NRC personnel.
Condensate removal is an important safety function because condensate accumulation upstream of the turbine governor valve wi 11 cause an overspeed trip to occur and prevent the standby pumps from performing their safety.
function (Section 6.5).
I f
ilj l
(il I
~
The inspectors identified that none of the licensees had performed a
4-hour run of the safety-related steam turbine-driven standby pumps after changing the oil type.
This was a concern because the vendor had stated that oil aeration was detected on some turbines during the initial 4-hour qualification run and that the affected turbines required modification prior to shipment to the licensees.
The susceptibility to oil aeration varies with oil type (Section 7.3).
l iaaf f
I
DETAILS
INTRODUCTION Historically. there have been several occurrences throughout the industry of safety-related steam turbine-driven standby pumps fai ling to start. failing to continue to run after starting.
and tripping to a "lockout" condition which required manual operator actions at the turbine to return the turbine-driven pump to an operable status.
More recently, there have been three additional
. examples at two Region IV plants (South Texas Project (STP)
and Comanche Peak Steam Electric Station (CPSES))
where turbine-driven pumps have not operated as designed.
These continuing fai lures have raised concern because the safety-related steam turbine-driven auxiliary feedwater pump is normally the only source of core cooling for pressurized water reactors during a station blackout.
This inspection compared the programs that the licensees had implemented to assure the reliability of the safety-related steam turbine-driven standby pumps to the level of attention they have given to other risk-significant safety-related equipment.
such as the emergency diesel generators.
The inspection also reviewed specific industry-recurrent f'ai lure mechanisms for turbines which were initially provided by the Terry Corporation.
These failures included governor system fai lure or loss-of-speed control margin, governor valve stem binding, excessive condensate and/or moisture accumulation, and lubrication and speed control problems associated with oils and hydraulic fluids.
The inspection also included an evaluation of some overspeed trip device malfunctions.
This inspection specifically reviewed safety-related steam turbine-driven standby pump applications in the auxiliary feedwater (AFW) system or the emergency feedwater (EFW) system at 15 Region IV pressurized water reactors.
The inspection also reviewed steam turbine-driven standby pump applications in the reactor core isolation cooling (RCIC) system at 3 boi ling water reactors.
The inspection did not review data associated with the high pressure coolant injection system turbines (also supplied by the Terry Corporation) at boiling water reactors because these steam turbines were significantly larger than the turbines used in the AFW ~ the EFW ~
and the RCIC systems.
Fort Calhoun Station
.
was not included in this review effort because it does not have a
safety-related steam turbine-driven standby pump that was produced by the Terry Corporation.
The Callaway Plant and Grand Gulf Nucl.ear Station were.
also not included in this review effort because they were not in Region IV at the time of the inspection.
Therefore, the information presented in this inspection report involves 18 individual units in Region IV.
The inspection was conducted at eight sites.
Information gained during recent NRC inspections at CPSES and STP was also included in this report.
(Reference NRC Inspection Reports 50-445/95-13; 50-446/95-13 and 50-498/95-10; 50-499/95-10.
respectively.)
Personnel at CPSES, STP, RBS and WNP-2 were contacted by telephone during the inspection.
An in-office review was performed of documentation supplied by personnel from all 18 unit SYSTEM CONFIGURATIONS The Terry Corporation supplied similar commercial-grade steam turbines for use in the AFW.
EFW and the RCIC systems at 18 Region IV units.
The Terry Corporation became the Terry Turbine Division of Ingersol-Rand, and is now Dresser-Rand, Terry-Turbodyne.
a joint venture.
The turbine vendor will be referred to as "Dresser" in this report.
In addition, the Woodward Corporation supplied various commercial'-grade mechanical and electronic governors, which are used to control turbine speed.
The inspectors found that each pressurized water reactor unit had a unique configuration for the layout of the steam supply piping for the AFW.
EFW turbines.
For example, some systems used the trip-and-throttle valve as the steam admission valve. while other systems used a separate steam admission valve.
At some facilities. the steam isolation valve was located close to the turbine.
At others, the steam isolation valve was located a long distance away.
Steam traps and/or condensate drain pots.
upstream of the steam admission valves, were included in some systems.
The designs varied because these systems were designed by different architect engineers.
The system configurations for the boiling water reactors were much more similar because they were designed by a single nuclear steam system supplier.
EXPERIENCE REVIEW Region IV performed this inspection to evaluate the implementation of the
.
licensees'xperience review process and to determine the status of safety-related steam turbine-driven standby pumps with respect to selected industry-recurrent fai lure mechanisms.
The licensees.
industry organizations, vendors, and the NRC have performed several studies in an attempt to identify and correct the causes of turbine failures.
NRC issued NUREG 1275, Volume 10,
"Operating Experience Feedback Report
- Reliability of Safety-Related Steam Turbine-Driven Standby Pumps."
and the following information notices to discuss events related to safety-related steam turbine-driven standby pumps:
"PWR Auxiliary Feedwater Pump Turbine Control Problems,"
dated March 10.
1986; Information Notice 86-14, Supplement l. "Overspeed Trips of AFW, HPCI, and RCIC Turbines." dated Decembei 17, 1986; Information Notice 86-14.
Supplement 2,
"Overspeed Trips of AFW ~
~
and RCIC Turbines," dated August 26, 1991:
~
"Reduced Reliability of Steam-Driven Auxiliary Feedwater Pumps Caused by Instability of Woodward PG-PL Type Governors,"
dated March 18, 1988; Information Notice 88-67,
"PWR Auxiliary Feedwater Pump Turbine Overspeed Trip Failure." dated August 22.
1988;
'
"Overspeed of the Turbine-Driven Auxiliary Feedwater Pumps and Overpressuri zati on of the Associ ated Pi ping Systems."
dated July 6.
1990; Information Notice 90-76.
"Failure of Turbine of Overspeed Trip Mechanism Because of Inadequate Spring Tension," dated December 7.
1990:
~
"Repetitive Overspeed Tripping of Turbine-Driven Auxiliary Feedwater Pumps,"
dated July 9, 1993;
~
"Overspeed of Turbine-Driven Pumps Caused By Governor Valve Stem Binding," dated September 19, 1994;
~
Information Notice 94-66, Supplement 1,
"Overspeed of Turbine-Driven Pumps Caused by Binding in Stems of Governor Valves," dated June
'995:
and
~
~
Information Notice 94-84 'Air Entrainment in Terry Turbine Lubricating Oil System."
dated December 2,
1994.
The inspectors reviewed these publications.
vendor information'nd Terry Turbine User's Group Newsletters to identify the actions licensee personnel could take to prevent the selected industry-recurrent failures.
The
~
inspectors reviewed previous NRC inspection reports.
licensee maintenance documentation, and interviewed plant personnel to determine which actions licensee
'personnel had taken to provide assurance that their safety-related steam turbine-driven standby pump(s) would perform the intended safety function(s).
3. 1 Licensee Use of NRC Generic Communications During the inspections the inspectors requested that each licensee provide copies of the documentation of their review of NRC Information Notices 94-66 and its supplement; 93-51; 86-14 and its supplements; and.
'olume 10.
The inspectors also sampled responses to some of the other Information Notices listed above.
As a result of this review. the inspectors found that licensee personnel do not routinely document experience reviews associated with NUREGs.
The system engineers at some units (STP and Arkansas Nuclear One (ANO)) had participated in the industry review of the document prior to publication.
While these personnel were very familiar with the content of NUREG 1275, Volume 10. they stated that, in some cases.
NUREGs are routed for information at their facilities. but no formalized evaluation is required.
One other system engineer (CPSES)
had received the document without any type of action item associated with completion of the experience review.
The system engineers at Cooper Nuclear Station (CNS) and Waterford Steam Electric Generating Station, Unit 3 (WAT-3). were unaware that the document existed prior to this inspection.
The inspectors fou.id that personnel at several facilities (Diablo Canyon Power Plant (DCPP):
Palo Verde Nuclear Generating Station (PVNGS):
San Onofre
e
j
Nuclear Generating Station (SONGS):
and, WAT-3) incorrectly stated that NRC Information Notice 88-09,
"Reduced Reliability of Steam-Driven Auxiliary Feedwater Pumps Caused by Instability of Woodward PG-PL Type Governors,"
was not applicable to their facility because they did not have a PG-PL-type governor.
The information notice discussed the misapplication of buffer springs internal to the governor and subsequent speed instabilities.
The inspectors noted that this failure could also occur in the EG-R-type actuators and in the PG-A-type governors'hich were installed at these facilities.
This was explained in NUREG 1275, Volume 10. but generally over looked by licensee personnel.
(Reference Table 3,
"Governor Systems." for site-specific information.)
The inspectors concluded that personnel performing experience reviews at these facilities did not understand the internals of the governors sufficiently to draw correct conclusions regarding the applicability of the information notice.
Similarly, the inspectors found that personnel at Washington Nuclear Project-2 (WNP-2) had performed a review of Information Notice 86-14 'upplement 2,
and incorrectly concluded that an inspection of the governor sump was not necessary because the turbine oil was found to be clean.
The information notice pointed out that the problems which had occurred at ANO were not detected by sampling the turbine oil every month and changing the lube oil filter every 6 months.
The design of the EG-R-type actuator sump is such that changing the turbine oil will not change the oil in the actuator sump.
Therefore, the sump is more subject to long-term accumulation of contaminants and should be inspected separately.
The inspectors also noted that the CPSES personnel performed a review of Information Notice 93-51,
"Repetitive Overspeed Tripping of Turbine-Driven Auxiliary Feedwater Pumps'
" that discussed the importance of minimizing steam supply valve leakage based on failures which occurred at STP.
The CPSES reviewer incorrectly determined, as discussed in NRC Inspection Report 50-445.
446/95-13. that the issue was not applicable to CPSES because of system design differences between the facilities.
The licensee stated that the system design differences made it difficult to do an effective experience review for AFW issues.
The inspectors concluded that, as a result of the lack of uniformity and the complexity of the equipment, licensee personnel'were not always correctly evaluating NRC information notices.
Careful analysis of problems associated with'safety-related steam turbine-driven standby pumps is necessary to determine the applicability of experience review.
3.2 Use of Licensee Ex erience In NUREG-1275, Volume 10, the NRC reported that the industry-wide demand probability of fai lure for the turbine-driven auxiliary feedwater pump was 6.5E-2 (excluding maintenance unavailability),
as compared with a value for the Surry Probabi listic Risk Assessment in NUREG-1150 of 1. 1E-2 for auxiliary feedwater.
These fai lures were primarily caused by overspeed trips.
The inspectors interviewed licensee personnel to determine the availability of plant-specific fai lure data at each facility and to determine the consistency of the data with the individual plant examination submittals.
The inspectors
l t
)
l
'1
found a wide variation in the availability of failure data.
In most cases..
licensee personnel had not systematically established methods for tracking these failures.
Most of the other licensees had access to recent failure history in some form. but they had not put the information in the context of total demands to develop a reliability estimate.
Several utilities were tracking total availability. rather than success-on-demand.
Engineers at one facility (ANO) had completed a preliminary engineering analysis of the reliability data for the turbine-driven EFW pumps in preparation for implementation of the maintenance rule.
The inspectors noted that the plant-specific failure data used in the utilities individual plant examination submittal was an order of magnitude less conservative than the results of the recent engineering analysis.
Licensee personnel stated that the base probabi listic risk assessment model had not been updated with this new data.
The licensee did plan to update the model as more reliability data, which was being developed for the maintenance rule, becomes available.
The licensee was monitoring the effects of turbine reliability changes on core damage frequency on a monthly basis using a simpler model to approximate core damage frequencies.
Licensee personnel stated that this information was being used to develop operating and maintenance strategies.
Licensee personnel stated that for, AN0-2, the loss of the turbine-driven EFW pump was more important than the loss of one diesel generator.
The reverse was true.for AN0-1.
Conditional Probability that Turbine Driven EFW Pump Will Be Available, Start and Run for a Valid Demand Signal UNIT ANO-1 ANO-2 PROBABILITY
- 95K 87 - 93K 1989 - 1994 1989
- 1994 EFW System Conditional Probability Analysis 5/26/94 EFW System Conditional Probability Analysis 5/26/94 TIME PERIOD SOURCE OF INFORMATION Individual Plant Examination Data for Turbine-Driven EFW Pum UNIT FAILURE HODE Fail to Start ANO-1 Fail to Run Maintenance Fail to Start ANO-2 Fail to Run Maintenance MEAN 5:76E-3 9.37E-5 3. 11E-4 7.27E-3 6.45E-3 2. 91E-4 ERROR FACTOR 5.21 1.81 1.28 5.21 9.62 1.81 TIME PERIOD Pre
- 1990 Pre
- 1990
I I
f
As stated in NUREG 1275, Volume 10, most of the failures which have occurred for standby steam turbine-drivers are related to the start sequence.
The inspectors concluded that success-on-demand (including surveillance test results)
was an important indicator of turbine reliability,.and that most licensees were not monitoring this data for comparison with the reliability estimates used in probabi listic risk assessments.
3.3 Vendor Ex erience As discussed above.
most safety-related auxiliary feedwater and RCIC system standby steam turbine-drivers were supplied by Dresser as commercial-grade items.
Dresser did not perform the original design of the steam supply, the steam exhausts or the condensate removal systems.
Licensee engineers have typically accepted full-design responsibility for these systems from the original architect engineers or the nuclear steam. system supplier.
Howevers licensee engineers typically lack specialized turbine expertise.
The inspectors evaluated the design interface between the licensee and the turbine vendor for three emerging technical issues:
mechanical overspeed
=tripping due to inadequate condensate removal; mechanical overspeed tripping due to corrosion of governor valve stems:
and.
use of vapor-phase inhibitors in oil.
This inspection was performed to assess the effectiveness of the vendor/1icensee interface with respect to assuring reliable turbine operation.
3.3. 1 Condensate Removal'he inspectors reviewed nine vendor technical manuals for Terry turbines.
The inspectors found that six of the nine technical manuals (ANO-2. CPSES, DCPP, STP, PVNGS.
and WAT-3) contained recommendations for condensate removal.
In Section 10, "Operation." of these technical manuals, under the paragraph titled.
"Emergency or Quick Start-up." the vendor stated, in a note that,
"(ijfemergency quick starts are anticipated provision should be made for steam lines to be continuously drained
.
.
.."
The inspector s also noted that the three remaining technical manuals (AN0-1, SONGS, and Wolf Creek Nuclear Generating Station (WCNGS)) did not include this statement.
The inspectors discussed the inconsistency in the recommendation with the vendor representative.
He stated that the turbine was designed to run with very low quality steam.
However, provisions should be made for steam lines to be continuously drained at any installation which anticipates emergency quick starts.
He stated that the drains should ensure that 'condensate does not accumulate upstream of the governor valve. since this could lead to an overspeed trip during star ting.
He also stated that Dresser would evaluate the need fop updating the remaining technical manuals.
The inspectors evaluated the drain configuration at the applicable units and only found three units which did not comply with the vendor recommendation to continuously drain the steam lines (ANO-2: and DCPP-1 and -2).
At AN0-2, the safety-related steam turbine-driven standby pump turbine casing drain valves were normally closed.
The operators opened the valves once per shift to drain any accumulated water out of the turbine.
The licensee used a
test to demonstrate that this operator monitoring approach was sufficient to
l I
ensure that condensate did not accumulate upstream of the governor valve.
The licensee ran the turbine during a period when the steam isola.ion valves were leaking significantly.
Without opening the turbine casing valves.
licensee personnel secured the turbine. waited 17 hours1.967593e-4 days <br />0.00472 hours <br />2.810847e-5 weeks <br />6.4685e-6 months <br />.
and then successfully restarted the turbine.
Licensee personnel also stated that maintenance was performed to reduce the steam isolation valve leakage.
The inspectors noted that this approach depended on continued operability of the nearby steam traps.
The licensee stated that operators routinely monitored trap performance once per shift.
At DCPP. the turbine casing drain valves on each unit's safety-related standby-steam turbine were also normally closed.
The licensee stated that if the
.
steam admission valves leaked.
then the turbine casing drains would be opened every 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.
Cold quick-starts were performed at DCPP once per quar ter and prewarmed starts were performed twice per quarter.
If the turbine was started for warm start testing.
operators opened the drain valves for approximately 1 minute. while warming the steam lines and turbine. to remove any moisture.
The cold quick-start tests simulated an automatic turbine start (i.e., the drains remained closed).
In Section 3 of most of the technical manuals for the Terry steam turbines, under the paragraph titled, "Auxiliary Piping," the vendor stated that. "drain lines are to be open when the turbine is idle to prevent accumulation of condensate in the turbine. which will result in corrosion and rapid deterioration of internal par ts."
The inspectors noted that the, turbine casing steam trap at WAT-3 was designed to allow a small amount of water to stand in the turbine casing following a turbine run.
The inspectors found that the system configurations at four units were not in accordance with the vendor recommendations.
which were provided to the licensees.
The personnel at both DCPP and ANO-2 were attempting to meet the intent of the vendor's recommendations with respect to condensate accumulation by use of operator monitoring.
The inspectors concluded that operator monitoring at DCPP and ANO-2 was critical.
Otherwise.
these turbines were more vulnerable to excessive condensate accumulation because the turbine casing drains were closed.
At the exit interview. personnel at WAT-3 stated that they planned to evaluate the need for modifying their drain system to ensure water would not stand in the turbine casing following a turbine run.
3.3.2 Governor Valve Stem Corrosion In Region IV. four units (AN0-2, STP-2.
CPSES-1, and River Bend Station (RBS))
have experienced overspeed trips, which were caused by corroded valve stems.
At least 18 such events have occurred nationally.
The inspectors found that stems.
which had been manufactured with a liquid-nitride surface treatment, were present in each of the failures associated with governor valve stem sticking.
The inspectors reviewed vendor recommendations related to this issue to determine if they adequately characterized the risk associated with the use of the valve stems manufactured with a liquid-nitride surface treatmen ! j ij
'I
e At the time of the inspection.
Dresser had not recalled the suspect stem material for nuclear applications because they believed it was suitable.
provided the licensee could control moisture and steam chemistry.
In a March 24.
1993. letter to Surry Nuclear Power Station (with copies to other facilities) Dresser discussed the vulnerability of some 410 stainless steel governor valve stems to corrosion in the presence of moisture and corrosive steam chemistry.
Licensee personnel at two units (STP-2 and CPSES-1)
stated that
~ prior to the overspeed trips at their facility. they had discussed the potential vulnerability with Dresser.
System engineers stated that when they contacted
. Dresser directly about the valve stem corrosion issue, the Dresser representative stated that over 100 of the stems were in service with only a few fai lures.
Thus.
system engineers at STP and CPSES incor rectly concluded that the possibility of fai lure at their unit(s) was remote.
(Reference Table 3.
"Governor Systems,"
attached, for site-specific information.)
Both licensees believed that gross leakage was necessary for the corrosion phenomena to occur.
In the STP-2 design, the steam isolation boundary was close to the turbine:
however.
the steam isolation boundary valve did not leak measurably at the time of the overspeed trip.
The steam isolation boundary at CPSES-1 was further from the turbine and the leakage was approximately 13.75 Lph [3.5 gphj.
The inspectors reviewed steam supply valve leakage controls at all of the facilities and determined that the amount of moisture necessary for valve stem corrosion could be intermittently present at most units.
t At the request of industry personnel, Dresser is currently performing qualification testing to develop a replacement valve stem.
which is more
'orrosion resistant.
3.3.3 Use of Vapor-Phase Inhibitors in Oil On April 21.
1978. the NRC issued IE Circular 78-02,
"Proper Lubricating Oils for Terry Turbines."
This circular reiterated the vendor recommendation to use turbine lubricating oil with vapor-phase corrosion inhibitors, such as Mobil Vaprotec Light, to prevent internal corrosion of turbine parts.
Vapor-phase inhibitors out-gas from the oil onto all surfaces to form a protective barrier.
At temperatures above 48.9'C f120'F], the vapor-phase corrosion inhibitor in Mobil Vaprotec Light oil out-gases and plates out on any surface, forming a sticky, varnish-like substance.
If the oil with vapor-phase inhibitors is also used in the governor system, speed control problems can result.
The formation of a sticky, varnish-like substance can interfere with the proper operation of the overspeed trip tappet.
The tappet can bind, preventing an overspeed trip or making resetting the turbine difficult.
On August 26.
1991. the NRC issued Information Notice 86-14, Supplement 2,
"Overspeed Trips of AFW. HPCI.
and RCIC Turbines." to address overspeed tripping due to fouled control oil.
The information notice did not refer to the use of Mobil Vaprotec as a potential cause of the failure at ANO-2 because there was not enough industry data at that time to support this conclusion.
e I
The Hay 1994 Terry Turbine User 's Group letter, however, stated that the following plants have had oil problems with Mobil Vaprotec Light:
ANO -1 and-2: WAT-3: STP-1 and -2: Clinton Power Station.
Unit 1: LaSalle County Nuclear Power Stations Units 1 and 2; Pilgrim Nuclear Power Station Unit 1; and
~
St. Lucie Plant. Units 1 and 2.
In 1993 'resser reviewed an industry consultant report that indicated about half of the utilities, which used vapor-phase inhibitor oils, were experiencing problems.
In a September 21.
1993, letter to the Electric Power Research Institute/Nuclear Maintenance Applications Center, Dresser acknowledged the problem of solids forming in vapor -phase oil as a result of
. high standby temperatures.
They stated that their turbine was never designed for standby temperatures between approximately 49-93 C [120-200 Fj.
Dresser further stated that..if it was not reasonable for the owners to maintain lower standby temperatures by maintaining the steam supply valves free of leaks, then conversion to a high-grade turbine oil should be considered as an alternative to the solid forming problems associated with high oil temperatures.
The inspectors found that oils with vapor-phase inhibitors were used at nine of the 18 units (CNS [Mobil Vaprotec Lightj; DCPP-1 and -2 [Shell VSI-68j; PVNGS-1,
-2, and -3 [Shell VSI-32j; RBS [Mobil Vaprotec Light3; and, SONGS-2 and -3 [Mobil Vaprotec Lightj).
The inspectors also found elevated standby-temperatures at 2 of these units (SONGS-2 and -3).
The inspectors determined that the SONGS units were currently the most susceptible to problems associated with the out-gassing of the vapor -phase inhibitor.
The inspectors noted that standby temperature was directly related to steam isolation valve leakage.
Most licensees did not routinely measure either standby temperature or steam supply valve leakage.
If steam supply valve leakage increases in the remaining units that use vapor -phase inhibitors, they may also experience the solid forming problems.
The inspectors discussed the issue with the vendor representative.
Me stated that the Dresser recommendations for use of turbine oil were intended to provide flexibility to the licensee.
The recommendations allow the licensee to select the correct turbine oil, depending on plant-specific conditions."
The inspectors asked if the switch to high grade turbine oils had resulted in any new failures.
The vendor representative was not aware of any.
Based on interviews with licensee personnel, the inspectors found that, licensee personnel lacked turbine expertise.
They relied on the turbine vendor to evaluate emerging technical issues from a nuclear perspective.
However, the vendor lacked detailed system installation information and, as a
result.
provided recommendations from a commercial perspective (i.e.,
restating the equipment limitations without making conservative recommendations for nuclear applications).
Licensee personnel also stated that the vendor did not routinely provide updated vendor information to all licensees when a
new issue was identified.
The inspectors also found that licensees did not consistently implement vendor recommendation.4 Terr Turbine User's Grou During the summer of 1993.
licensee personnel from ANO-2. STP.
and other facilities worked with the Nuclear Maintenance Applications Centers which is operated by the Electric Power Research Institute, to establish the Terry Turbine User's Group.
The Terry Turbine User's Group initiated several programs to improve standby turbine reliability.
On an intermittent basis they publish a newsletter.
which addresses ongoing technical concerns.
As discussed above.
a May 1994 Terry Turbine User's Group newsletter thoroughly addressed the use of vapor-phase inhibitors in Terry turbines.
~ The Terry Turbine User's Group sponsored two maintenance workshops to provide hands-on training covering governors, trip-and-throttle valves, overspeed trip devices, and the turbine itself.
Through the Nuclear Maintenance Applications Center, operated by the Electric'ower Research Institute, the Terry Turbine User's Group produced the
"NMAC Terry Turbine Controls Maintenance Guide (NP-6909)."
In many cases, this manual provided specific quantitative guidance (e.g.,
clearances related to governor valve stem linkage assembly)
and measurements (e.g., appropriate spring tension for the emergency trip spring).
A companion troubleshooting and performance monitoring guide is
'eing developed.
The inspectors determined that this group was effectively addressing technical issues and were providing a useful forum for the dissemination of technical information regarding the use of Terry turbines.
The inspectors found that all licensees were not represented in the Terry Turbine User's Group.
Licensee personnel at three units (CNS and DCPP-1 and -2) in Region IV were not members of the Terry Turbine User's Group.
A Terry Turbine User's Group officer estimated.
that nationally, 30 percent of the utilities were not members.
Licensee personnel explained that the current-rate structure for joining the Electric Power Research Institute is based on total megawatts produced.
As a result, utilities with a heavy total investment in fossi 1 and hydro-electric plants were less likely to join the Electric Power Research Institute and were ineligible to become members of the Terry Turbine User's Group.
The inspectors were told that the Electric Power Research Institute was working to change the rate structure for nuclear activities so that all nuclear facilities will be charged a comparable fee, making the information equally accessible.
The inspectors noted that the licensee representatives were usually working level personnel (system engineers, maintenance engineers).
not licensee management, and were not positioned to direct changes at thei r unit(s).
The inspectors determined that it was not appropriate to rely too heavily on this organization to resolve safety issues.
GOVERNOR SYSTEMS 4.1
~Back round Earlier NRC and industry studies had shown that the most significant factors in the failures of safety-related steam turbine-driven standby pumps had been the failures of the turbine drivers and thei r controls.
The governor system of these turbines had played a large role in the failures of the turbines to start or to keep running.
In particular.
a corr'ect governor response is
e li t
0 critical to prevent mechanical overspeed trips of the steam turbines during the start (or restart)
sequence.
The majority of standby turbine failures were the result of malfuncti'ons of the turbine governor during cold quick-starts.
Overall.
system dynamic problems must be fully considered to prevent malfunctions.
The inspectors found that several contributing factors often combine to cause mechanical overspeed trips.
4.2 Governor S stem Desi ns The governor system consists of the governor.
the governor controls.
the governor valve, and the linkage connecting the governor to the governor valve.
. In standby.
steam is isolated from the turbine.
Governor valves usually go full open when the turbine is secured.
and remain full open in the standby condition.
In general, the turbine controls were designed so that a safety signal.
such as an engineered safety feature actuation, opens a steam supply to the turbine.
Turbine rotation was necessary to develop the hydraulic pressure used to move the governor valve.
After turbine speed increases, the governor acts to throttle close the governor valve and take control of turbine speed at a preset minimum speed..
Then the governor will ramp open the governor valve at some predetermined rate until the turbine reaches full speed.
4.3
~Ti i The inspectors found that the timing of the start sequence of steam turbine-driven standby pumps was critical.
The turbine controls must be designed to coordinate the opening of the steam supply valve(s) with the throttling of the governor valve.
The governor valve must throttle closed before the steam supply valve(s) fully open to prevent a mechanical overspeed trip of the turbine.
The design of the timing sequence was also influenced by the design of the steam supply piping.
Condensate formed when the steam isolation valves open and steam passed through the cold-steam supply piping.
More condensate formed at units which have remotely located steam supply valves (as much as
m [250 ft] away from the turbine).
The vendor stated that Terry turbines were designed to run reliably with very low-quality steam, but they were not designed to star t (or restart) with excessive condensate.
The vendor stated that excessive condensate accumulation could increase the likelihood of an overspeed trip for a number of different reasons.
Huch of the condensate that passed through the turbine flashed to steam.
resulting in erratic turbine speed changes.
The design of the governor system was not responsive enough to control the speed changes caused by the water steam mixture; therefore.
the turbines could trip on mechanical overspeed.
The governor valves were also designed for a steam application and not for closing against water.
Licensee personnel from several units reported extensive testing and modifications during the initial licensing phase to address this design vulnerability.
Licensee personnel developed a variety of timing strategies.
For example.
four units'esign (ANO-2; and PVNGS-1,
-2 and -3) incorporated automatic warm-up valves into the start sequence so that condensate would slowly be introduced to the turbine.
Two units'esign (CPSES-1 and -2)
opened the steam supply valves quickly in an attempt to sweep the condensate through the turbine before the governor system developed enough hydraulic pressure to control speed.
With either strategy.
turbine reliability was very sensitive to changes in governor valve/steam supply valve/bypass valve coordination and to changes in the amount of condensate.
4.4 Periodic Testin The inspectors reviewed licensee surveillance tests to determine if the licensee engineers had developed effective, periodic testing programs which would detect:
(1) changes in governor valve/steam supply valve/bypass valve.
~ coordination:
(2) changes in the amount of condensate which initially passes through the turbine; or.
(3) the onset of governor valve linkage/stem binding.
The inspectors found that. in all cases.
the Technical Specifications requi red the licensees to perform periodic pump flow tests:
however.
these tests were not always written to confi rm the readiness of the safety-related standby steam turbine-drivers.
4.4. 1 Standby (Cold Ouick-Start) Turbine Test Requirements The inspectors noted that periodic testing must duplicate actual demand actuation conditions to adequately demonstrate the operational readiness of the turbine-drivers.
The test should be performed in the standby condition.
(i.e.. without preconditioning the system by prewarming or draining the steam lines).
This type of testing had historically been referred to as cold quick-start testing or cold-start testing.
However, the inspectors noted that some licensee personnel maintained their turbine and steam lines in a prewarmed condition to minimize condensate formation during turbine starts.
Licensee personnel from every unit stated that it was their normal practice to perform testing to demonstrate standby readiness.
Some licensee personnel (DCPP-1 and -2) did not perform a standby start.for every test.
Some of the time they prewarmed and drained the turbine to mitigate aging effects.
The inspectors reviewed, the licensees'urveillance test procedures and found that licensee personnel had not proceduralized the requi rement for periodically testing without preconditioning at four units (ANO-1 and -2:
Cooper Nuclear Station (CNS);
and WAT-3).
At DCPP-1 and -2; cold quick-starts were performed once per quarter and prewarmed starts performed twice per quarter.
At STP-1 and -2. cold quick-starts were performed following maintenance.
Personnel at STP determined that 80 to 90.
percent of the cooldown occurs within the first 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> after a run.
They routinely perform all survei llances at least 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> after a previous run.
Personnel at ANO-1 and -2.
had included a requirement to measure turbine inlet temperatures prior to a run
~ but the procedures did not specifically include precautions to prevent preconditioning.
As stated in Section 3.3. 1, the operators at ANO-2 open the valves once per shift to drain any accumulated water out of the turbine.
ANO-1 and -2 does not link the operator action to drain the lines with the monthly surveillance test.
Operators may drain the lines before or after the turbine run.
(Reference Table 1, "Turbine Surveillance Requirements."
attached, for site-specific information.)
ll
)
ll
)j
!
i 1~
,',I f
f
Since the majority of standby turbine failures were the result of malfunctions of the turbine governor during cold quick-starts.
the inspectors concluded that the failure to specifically requi re periodic standby testing (i.e.,
without warming or preconditioning)
was a minor program weakness.
4.4.2 Governor System, Governor Valve.
and Steam Supply Valve Performance Trending The inspectors noted that Dresser recommends that governor valve coordination be verified quarterly.
The inspectors verified that each unit operated the standby turbines at least quarterly.
This testing provided a baseline assurance that the standby turbines were operable, but it did not detect loss-of-reliability margin.
The inspectors found that those licensees that had trended speed trace data were best able to confirm governor valve/steam supply valve/bypass valve coordination.
Licensee personnel were able to use speed trace data to identify potential governor valve binding issues.
timing sequence changes, and changes in the quantity of condensate formed during the start sequence.
For example, the system engineer at ANO-2 used speed trace data to identify a slight governor valve alignment problem, thus.
avoiding a failure of the governor valve.
The alignment problem occurred during a governor valve bonnet replacement.
Licensee personnel at ANO-2 incorrectly assemb'led the valve and did not install requi red valve alignment pins.
The replacement bonnet had not come drilled to accept the required alignment pins and the vendor drawing did not show that the holes needed to be drilled.
The inspectors discussed the, lack of detailed information with the vendor representative.
The vendor representative stated that they prepared the technical information assuming that an experienced turbine professional would be onsite directing the maintenance activity.
He further stated, it was the responsibility of the licensee to procure that expertise if they did not have a turbine expert in-house.
The inspectors determined that the vendor interface was ineffective in this case.
The inspectors noted that speed trace data for the STP-2 turbine indicated speed control anomalies prior to the December 19, 1994, mechanical overspeed trip.
Personnel at ll units (ANO-1. and -2; PVNGS-1,
-2, and -3; SONGS-2, and -3: STP-1
~
and -2:
WNP-2; and WCNGS) were trending speed trace data.
The engineers at 2 other units (CPSES-1 and -2) were monitoring governor valve
.
performance during the start sequence using strain gauge data and valve position data.
The engineers at 2 units (DCPP-1 and -2) used pump discharge pressure for trending governor valve performance.
The inspectors noted that system conditions.have to be duplicated during every test for a known relationship to exist between speed and flow.
The inspectors found that the engineers at three units (CNS, WAT-3, and RBS)
were not performing any type of dynamic monitoring.
The system engineers stated that they did not have the equipment to perform meaningful dynamic monitoring of governor/governor valve performance.
The inspectors were concerned with this situation because a recent Terry Turbine User's Group Newsletter noted that slow degradation of governor systems is difficult to
if/I 4,
t, i
III
diagnose without transient monitoring and recording equipment or a thorough performance monitoring program.
The inspectors determined that licensee personnel at WAT-3 and RBS had relied on successful pump starts to validate the valve coordination and manual governor valve manipulation necessary to prevent binding problems.
However, the inspectors noted that manual valve manipulation was not capable of predicting governor valve stem binding at STP.
Personnel at CNS had stopped the manual valve manipulation in mid-May 1995: since then, they had relied solely on thei r routine flow test.
The inspectors were concerned that these units did not have an optimal method for identifying governor system
- weaknesses prior to failure.
(Reference Table
~ "Governor Systems."
attached'or site-specific information.)
4.5 S stem Desi n Verification Testin As stated in NUREG-1275.
Volume 10.
some governor instabilities are only exhibited during stand-alone operations.
Based on interviews with licensee, personnel.
the inspectors found that personnel at only three units (PVNGS-1,
-2, and -3) had performed an extended stand-alone turbine run in conjunction with a total loss of alternating current.
(Reference Table 2,
"Turbine Design Verification Tests'
" attached, for site-specific data.)
The inspectors noted that NUREG 1154.
"Loss of Hain and Auxiliary Feedwater Event at the Davis-Besse Plant on June 9, 1985," discussed the importance of testing all design bases steam-line configurations.
The inspectors reviewed the safety analysis report f'r each unit and found a wide variation in the design verification testing requirements.
Although some units had tested the steam supply configurations over the full range of steam inlet pressures, not all units had incorporated this verification.
Personnel at some units performed endurance tests followed by a restart test and at some'units personnel demonstrated only the capability of the pump to produce rated flow at normal operating temperatures and pressures.
(Reference Table 2. "Turbine Design Verification Tests," attached, for site-specific data.)
The inspectors found that licensee personnel at nine units (ANO-1 and -2:
CPSES-1 and -2; SONGS-2 and -3: STP-1 and -2; and WCNGS) had not demonstrated, by testing, that the safety-related steam turbine-driven pumps were capable of running over the full range of steam inlet pressures.
Personnel at STP had tested over part of the steam inlet pressure range.
They had tested the turbines as low as 400 psig
~ however, steam inlet pressure was expected to go as low as 100 psig.
The inspectors were also not able to establish that restart capability was fully verified by test.
However, the inspectors did not review the associated preoperational and hot functional test data packages.
In some cases, more testing than was clearly described in their safety analysis 'report may have been performed.
The capability to restart can be important.
As an example, NRC documented in Information Notice 86-14 that three steam turbine-driven standby pumps started following a reactor trip at the Turkey Point plant.
The operators secured the turbines when they were no longer needed.
When the turbines subsequently received another auto-start signal. all three turbines tripped on overspeed.
While licensee personnel believed this event occurred because the governors were inadequately reset.
the event highlights the importance of demonstrating the restart function.
The inspectors noted that auxiliary feedwater standby turbines are frequently secured by the operators after steam generator levels stabilize.
because it is easier for the operators to control steam generator levels with motor-driven pumps.
The inspectors also noted that emergency operating instructions at pressurized water reactors typically do not include precautions to not secure the turbine-driven auxiliary feedwater pumps unti l after decay heat is reduced enough to allow time for the operator to manually restart the turbine.
Also,
~ boiling water reactor designs include automatic reactor water level controls that stop and start the turbine-drivers used in the RCIC system.
Therefore, it is important in both pressurized water reactor designs and boi ling water reactor designs that condensate removal be functional after the turbine is secured so that the water will drain out.
4.6 Governor Preventive Maintenance Practices In NUREG-1275, Volume 10, the NRC reported that several governor problems had been traced to inadequate maintenance.
As a result of this inspection, the inspectors noted that licensee personnel at eight units (CNS:
CPSES-1 and -2; DCPP-1 and -2: WAT-3; WNP-2:
and WCNGS) had not established periodic maintenance requirements for the turbine governors.
At the exit interview.
personnel from CPSES stated that they planned to establish a preventive maintenance program for governors.
Personnel from WAT-3 stated that they had replaced their governor in May 1994 due to vapor-phase inhibitor buildup.
The engineers at the remaining uni,ts planned to ship the governor to the manufacturer for refurbishment or planned to replace the governor on a
specified frequency.
The majority of these licensees specified -a frequency of 5 years or every third refueling outage for this preventive maintenance task.
(Reference Table 3,
"Governor Systems."
attached.
for site-specific information.)
4.7 Governor Modification Control In NUREG-1275 'olume 10, the NRC reported that several governor problems had been traced to inadequate modification control between the various vendors (Woodward and Dresser)
and the utilities.
The inspectors interviewed licensee personnel and determined that Woodward's controls for tracking the governor
.
design configuration at each unit had been adequately implemented.
Licensee personnel stated that the nameplates of recently purchased components were marked with a "9903-" prefix, which indicated that configuration control would be monitored by the vendor.
The nameplates of components that had been modified in the field by Woodward were marked with a
"US" beginning in 1993.
This marking indicated that the field configuration would not match the documentation at Woodward.
When the component was returned to Woodward for refurbishment the documentation was updated and the
"US" was removed from the nameplate.
The inspectors verified that several specific upgrades had been implemented at the units.
The inspectors verified that all units. which used an electronic
I I
j
governor.
had installed Terry Design Improvement 6.
This modification added an oil sump for the EG-R-type actuator of an EG-M-type governor control box to improve governor control for quick-starts.
The inspectors also noted that all remote, speed control bellows for mechanical governors were vented to prevent speed changes with changes in ambient temperature as discussed on page A-1 of NUREG-1275, Volume 10.
Licensee personnel verified that external wiring for electronic governor EG-M-type control boxes and EG-R-type actuators was sized and shielded as specified by Woodward.
The licensees also verified that the direct current power supply for electronic governors was connected to the battery as specified by Woodward.
All governor control systems had a speed ramp provision to minimize overspeed during quick-starts.
In addition, all licensee personnel had upgraded to the latest overspeed tappet design.
GOVERNOR VAL'VES 5.1 Back<around As discussed in Section 3.3.2. failures of the turbine-drivers caused by corrosion-induced governor valve stem binding occurred at four units (ANO-2, CPSES-1.
RBS, and STP-2) in Region IV.
Similar events have occurred at approximately 18 sites throughout the country.
On June 16.
1995. the NRC issued Supplement 1 to NRC Information Notice 94-66,
"Overspeed of Turbine-Driven Pumps Caused by Binding in Stems of Governor Valves," to provide additional information to licensees regarding these fai lures.
Based on metallurgical examinations, licensee personnel at some Region IV units have determined that a
1976 change in valve stem material processing (i.e..
from gaseous to liquid-nitride surface treatment),
in conjunction with conditions conducive to corrosion.
leads to rapid stem failure.
However, an industry-accepted root cause for governor valve stem sticking had not been formally determined at the time of the inspection.
5.2 Stem Materials Licensee personnel at each site, that had experienced the stem failure.
had replaced the corrosion susceptible governor valve stem material (usually a
liquid-nitride surface treatment) with a more corrosion resistant material.
. Entergy Operations, Inc.,
management also studied their other units including WAT-3.
They identified that the gove'rnor valve stem in place at WAT-3 had the liquid-nitride surface treatment.
Personnel at WAT-3 installed a more corrosion resistant stem during the inspection.
The only other units that had the stems upgraded to a material less susceptible to corrosion, without having experienced a failure, were SONGS-1 and -2.
The inspectors found that older valve stems, which most licensee personnel believed to be manufactured using the gas-nitride surface treatment process, tend to corrode slowly via pitting.
As an exceptions based on a metallurgical examination.
personnel at CPSES believed their older stem was manufactured using a liquid-nitride surface treatment process.
When visually inspected, the inspectors observed that the older stems had a uniform black coating with
.
some pitting.
l
~
I t
I J
Region IV licensees determined that the valve stems that had recently failed were manufactured with the liquid-nitride surface treatment process.
Licensee personnel noted th failed stems appeared to rapidly corrode via a general corrosion mechanism or possibly galvanic corrosion.
Some licensee personnel attributed the failures to incrpased sulfur levels in the carbon spacers.
The inspectors observed that the failed liquid-nitride surface treated stems had striped black corrosion marks that paralleled the position of the carbon spacers and stainless steel washers.
The liquid-nitride surface-treated stems also had brown porous corrosion produc'ts in the area of the valve stem leak off.
The inspectors found that while industry personnel had not reached agreement regarding the precise cause of the corrosion, the corrosion occurred on recently replaced valve stems manufactured with the liquid-nitride surface treatment.
Personnel at four units (CNS; DCPP-1 and -2; and WCNGS) believed that the stems they currently had in use had the gas-nitride surface treatment.
Dresser personnel stated that a change in the surface treatment process was first allowed in 1976.
Licensee personnel at DCPP and CNS determined that the valve stems installed in thei r units were manufactured before 1976.
The licensee for WCNGS believed that a gas-nitrided stem was installed because they had not replaced the original stem.
However, the vendor representative stated that this was not sufficient basis because some of the turbines were originally shipped with liquid-nitride surface treated stems installed.
CNS personnel had also evaluated that a valve stem that was stored in, the warehouse was acceptable because it was manufactured in 1980.
After the inspectors discussed the vendor-supplied date with the CNS system engineer, he stated that he would evaluate this information before using the valve stem in stores.
The system engineer for CNS stated that the installed governor valve stem bound when the operators attempted to start the turbine after the previous refueling outage.
CNS personnel believed that the stem binding occurred because the turbine sat idle for an extended period of time.
They did not believe the binding was caused by corrosion.
The CNS engineer s determined that the old carbon spacers most likely had a low sulfur content; therefore.
the valve stem was less susceptible to corrosion.
The licensee also revised the operating procedure to improve moisture control in the valve stem leak-off region.
The licensee had inspected the valve stem via the goverrior valve stem leak off and had not noted any signs of corrosion.
The inspectors were concerned that this method of inspection was not adequate, because it was not possible for the licensee to detect corrosion in the vicinity of the carbon spacers and stainless steel washers.
The clearances between the carbon spacers/stainless steel washers and the valve stem were too tight to allow visual inspection without disassembly.
After discussions with the inspector, CNS personnel agreed that the inspection of the valve stem was inadequate.
As corrective action, they planned to perform a full inspection during the next refueling outage (November 1995)
and to evaluate the need for replacing the valve stem at that time.
At four units (PVNGS-1.
-'2 and -3:
and WNP-2). licensee personnel were not able to positively determine which type of valve stem material was installed.
They stated that Dresser personnel did not consider that the material processing change would affect the form, fit. or function of the stem and.
as a result.
Dresser personnel did not initially track stems manufactured by the two processes separately.
The PVNGS personnel planned to replace the present stems in each unit with a more corrosion-resistant material during the next refueling outage for each unit.
The licensee for WNP-2 was the only licensee that believed that the stem material could have-been susceptible to th'e corrosion process.
but had no plans to evaluate the stem for replacement.
At the exits personnel from WNP-2 stated that they planned to inspect the governor valve stem during the next
. outage and make a replacement determination at that time.
The inspectors also noted that two spare valve stems in stores at WCNGS were believed to be of the susceptible material.
However, WCNGS personnel had no plans to place any type of engineering hold on the use of the suspect stem material.
(Reference Table 3.
"Governor Systems,"
attached.
for site-specific information.)
The inspectors concluded that licensee personnel were not consistently replacing the suspect governor valve stems with a material which was less susceptible to corrosion.
5.3 Validation of New Stem Materials The inspectors found that a variety of different stem materials had been used to improve corrosion resistance.
such as:
Inconel 718; 410 stainless steel.
coated with chromium nickel; 422 stainless steel
~ coated with aluminum nickel and ferralium.
Licensee personnel stated that no overspeed trips had occurred as the 'result of corrosion of valve stems made of the new materials.
However, the hew stem materials have not been in service very long.
Engineers at the units with the new stem materials have a variety of inspection and test programs to determine the long-term acceptability of the new stem materials.
Some licensee engineers were performing routine surveillance tests which were sensitive enough to detect the onset of binding.
Other licensee engineers were inspecting the new stem material on a periodic basis to establish confidence that it is an appropriate material selection.
The inspectors determined that periodic inspection and dynamic testing capability were important to demonstrate the acceptability of the new stem materials.
5.4 Governor Valve Maintenance Practices The level of detail in governor valve maintenance instructions ranged from copyi ng a page out of the vendor manual to step-by-step disassembly/reassembly instructions.
The inspectors found that the vendor information provided to
.
licensee personnel.
related to the installation of governor valve stem packing.
was not very detailed.
The vendor typically supplied a drawing which indicated the general arrangement of the carbon spacers and stainless-steel washers.
The drawing did not specify how many washers and spacers should be installed.
nor did it specify the final acceptable, clearance.
The inspectors
,
noted that missing spacers and,washers can result in cocked spacers and washers and an associated increase in friction forces.
Licensee personnel at eight units (ANO-2: CPSES-1 and.-2:
PVNGS-1,
-2
~ -3; and STP-1.
and -2) relied on vendor technical information to perform governor valve maintenance (e.g.,
maintenance practices for stem packing instructions, valve-bonnet alignment pins. or valve stroke/linkage adjustments).
These licensees found that the'endor technical information did not always provide sufficient detail for maintenance to be successful.
After maintenance errors related to the installation of the governor valve stem packing occurred at these facilities. licensee personnel upgraded thei r governor valve assembly instructions.
Personnel at these units developed more detailed instructions, which included clearance specifications and counting the number of washers and spacers installed in the packing assembly.
Licensee personnel stated that the vendor had provided subjective information for adjusting the linkages.'s stated above.
the Terry Turbine User's Group had developed guidance documents to provide clarifying information, but this guidance had not been implemented at every facility.
The inspectors concluded that the information in the vendor manual did not provide sufficient guidance. for licensees to reliably perform maintenance on the governor valve.
CONDENSATE CONTROLS 6.1
~Back round As discussed in Section 4. Terry turbines were designed to run reliably with very low-quality steam.
The turbines were not designed to start (or restart)
with excessive condensate in the turbine.
As a result, multiple mechanical overspeed trips occurred at several plants (ANO, CPSES, STP, and WAT-3) during the preoper ational-test phase.
Licensee engineers had l edesigned the supply-side condensate removal systems to assure the capability of the turbine-driven pumps to start following the initiation of a safety signal.
The inspectors noted that condensate.
which formed in the steam supply piping during a cold start, was an especially significant problem for units with long runs of piping between the steam admission valves and the turbine.
6.2 ~St
A variety of design approaches were used by licensees to control the condensate formation and removal.
The inspectors found that half of the units (ANO-1: CPSES-1 and -2; DCPP-1 and -2:
PVNGS-1
~ -2.
and -3:
and WAT-3) had long runs of piping between the steam isolation valve and the turbine (greater than 15.24 m [50 ftj).
The steam isolation boundary for the remaining units was located close to the turbine.
For example.
the inspectors noted that:
~ 't AN0-1, the licensee included several steam traps to drain the steam lines during standby conditions.
In addition, the licensee located the steam isolation valve at ANO-2 approximately 6.1 m L20 ftj from the turbine.
After repeated overspeed trips, the licensee's engineers had revised the design to include an automatic warming valve in the steam 1 1ne,
1 J
~
At CPSES-1 and -2. the licensee initially addressed condensate formation and removal by adjusting the rate and timing of the condensate flow through the turbine.
However.
as a result of the recent mechanical overspeed trip. the licensee plans to upgrade their steam line drain system.
~
At DCPP-1 and DCPP-2.
the licensee ensured condensate removal by using three steam traps upstream of the steam line isolation valves and one steam trap between the steam line isolation valves and the trip-and-throttle valve.
At PVNGS-1.
-2.
and -3. the licensee ensured condensate removal by using steam traps and drain lines upstream of the steam admission valves, between the steam admission valves and the trip-and-throttle valve, and between the trip-and throttle valve and the turbine.
Additionally. the licensee reduced the effects of'ondensate formation by adjusting the rate and timing of the condensate flow through the turbine.
At SONGS-2 and -3 and STP-1 and 2, the licensees used the steam admission valve as the trip-and-throttle valve (short distance to turbine).
Drain lines were installed upstream of the trip-and-throttle valve at both facilities.
~.
At WAT-3 ~ the licensee used heat tracing to minimize condensate formation by prewarming the steam supply piping.
~
At WCNGS, the licensee kept the steam line warm by a small bypass line around the steam admission valve.
This minimized the effects of'oisture in the line.
6.3 Steam Su
Valve Leaka e.
The inspectors found that the steam supply valves at most units had leaked at least part of the time.
The inspectors noted that most units had not established any upper bound for steam supply valve leakage.
Only the engineers at PVNGS had established a quantifiable leakage rate (227 kg/hr
[500 ibm/hrj or approximately 3.78 Lpm [1 gpm], total leakage from the four isolation valves),
which specified when the valves should be repai red.
When questioned.
all licensee engineers contacted agreed that steam leakage significant enough to roll the turbine would be repaired.
The inspectors noted that approximately
L [20 gallons3 of condensate will fi11 the turbine and render it inoperable.
Therefore; if the steam supply valve is leaking at 0.75 Lpm [0.2 gpm], then the turbine could fill up in less than 1 day if the steam traps and drains malfunction.
The inspectors also noted that licensees typically had not established an upper limit on steam supply valve leakage as it related to steam trap and drain capacity.
Therefore, the inspectors
~ determined that it was important to have good operating status information about the readiness of the drain system.
,
f
6. 4 0 era tin Status Informati on The inspectors noted that the condensate removal system at several units had supply-side condensate high level alarms which would alert the operators to drain system fai lure.
In fact
~
CPSES had recently reinstalled the supply-side condensate high level alarm. following the June 13.
1995.
mechanic'al overspeed trip of the Unit 2 steam turbine-driven auxiliary feedwater pump.
The inspectors also found that seven units (ANO-1 and -2; DCPP-1 and -2: and, PVNGS-l, -2.
and -3) did not have supply-side condensate hi'gh level alarms.
Nevertheless, operating personnel at both ANO units routinely blew down the
. steam traps once per shift.
In addition, DCPP-1 and -2 personnel checked the supply-side steam traps monthly by physically touching the lines to compare the temperatures.
If the steam admission valves leaked.
the DCPP-1 and -2 operators would blow down the turbine casing drains every 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.
Licensee personnel at PVNGS-1,
-2.
and -3 performed thermography once per week to verify that the steam traps were functioning correctly.
Since minor steam supply valve leakage was allowed at these units, and since minor steam supply valve leakage can accumulate within I day to fill a turbine, the inspectors concluded that the operating practices at DCPP-1 and -2, and PVNGS-1.
-2, and -3 were not optimal for detecting excessive condensate accumulation.
The recent mechanical overspeed trip at CPSES and the STP events highlight the importance of also maintaining exhaust-side traps and drains.
The inspectors determined that only personnel from RBS and WNP-2 had installed alarms or high level indicators for the turbine or turbine exhaust-side steam traps or condensate pots/drains.
Personnel from CPSES and CNS planned to add high level indicators for the turbine exhaust-side drains.
(Reference Table 4.
"Condensate Controls.'" attached, for site-specific information.)
6.5 Safet Classification As stated above, condensate removal systems must function correctly to ensure the capability to automatically start and restart the standby turbines.
However, the inspectors noted that most licensees have not formally recognized the safety function associated with condensate removal.
The inspectors found that the licensees had upgraded the design classification of condensate removal components at the units that had experienced mechanical overspeed trips caused by inadequate condensate removal.
For example, STP personnel established that the supply-side steam drains were safety related following, discussions with the NRC Augmented Inspection Team (reference NRC Inspection Report 50-498/93-07:
50-499/93-07).
Similarly.
CPSES personnel were in the process of evaluating supply and exhaust drain systems to determine if they were safety related for condensate removal (reference NRC Inspection Report 50-445/95-13:
50-446/95-13).
Personnel at other facilities partially recogni zed a safety function associated with the condensate removal system.
For example, the inspectors noted that the steam trap at WCNGS is considered safety related for only the pressure boundary capabi ity: however.
a level control valve in parallel with the steam trap is considered to be safety related for both the pressure boundary and condensate removal capability.
At'he remaining units, licensee
l f
I l
personnel that recognized a safety function for the steam traps and/or drains had only formally recognized the pressure boundary function.
(Reterence Table 4.
"Condensate Controls' " attached, for site-specific information.)
The inspectors concluded that all of the licensees did not recognize that steam-line drains had a safety function to remove condensate until after discussions with NRC personnel.
The inspectors were concerned that condensate removal was an important safety function because condensate accumulation upstream of the turbine governor valve will cause an overspeed trip to occur and prevent the standby pumps from performing thei r safety function.
. 6.6 Steam Tra and Drain Maintenance The inspectors reviewed six auxiliary feedwater and one RCIC pump technical manuals.
The inspectors noted that the licensees.
for most of the units that used steam traps in the condensate removal system, had some type of preventive maintenance program for the steam traps.
However. the inspectors noted that the licensees for two units (WAT-3: and WCNGS) had not established a
preventive maintenance program for steam traps and drain pots.
At the exit interview, personnel from WAT-3 stated that they planned to establish a
preventive maintenance program for their steam traps.
Personnel at WCNGS noted that they conditionally perform maintenance when the level alarm indicates that the steam trap is not functioning.
In addition, none of the licensees had established a preventive maintenance program for the drains.
The inspectors also noted that the licensees at four units had determined that inadequate preventive maintenance programs for thei r steam traps contributed to condensate induced mechanical overspeed trips (CPSES-1 and -2; and STP-1 and -2).
Personnel at these sites redesigned their condensate removal system to eliminate the steam traps in the turbine drain system (Reference Table 4.
"Condensate Controls." attached, for site-specific information.)
LUBRICATING AND HYDRAULIC OILS 7.1
~Back round The inspectors historical review identified that use.of the proper lubricating oil played a large role in the successful operation of safety-related steam turbine-driven standby pumps.
Because oil provides the lubrication for moving parts.
as well as the motive force for the governor valve, a failure in either could render the equipment inoperable.
Various factors affect the performance of the oil. If the oil is too thick (viscous),
the governor valve response could be sluggish.
If the oil is aerated, lubrication of bearings', could be lost and an erratic response of the governor valve could be experienced.
If the oil chemically breaks down due to environmental factors (heat.
humidity, and contaminants),
the loss-of-speed control could occur.
7, 2 S~tO The inspectors reviewed the designs of the lubricating and hydraulic oil systems for 18 of the units surveyed.
All of the units with EG-M/EG-R-type governors had a single oil sump that supplied oil for both lubrication and hydraulic controls.
The units with PG-A-type governors had separate sumps tor
,
I
'h
each function.
In each of the units. the lubricating oil was distributed by way of slinger rings and the hydraulic control oil was distributed by way of'
positive displacement pump.
The inspectors noted that the vendors (Dresser and Woodward)
had provided the users of thei r equipment with guidelines for the type of oil to use for lubrication of moving parts and hydraulic control of the governor valve.
In a letter. dated September 21, 1993.
Dresser provided recommended viscosity requirements for oils at 37.8'C [100'F] and 98.9'C [210'F].
At 37.8'C. the viscosity range was 190 to 510 Saybolt Universal Seconds.
At 98.9'C, the range was 43 to 65 Saybolt Universal Seconds.
This generally corresponds to
. an International Standards Organization ( ISO) Viscosity 32; however, oils with other ISO viscosity numbers fall within this range.
For reliable governor actuator response during cold quick-starts (approximately 37'C [98 F]).
Woodward stated that the maximum oil viscosity should be 300 Saybolt Universal Seconds-.
The inspectors noted that all but two units used ISO Viscosity 32, or equivalent. oil for the turbine and governor applications.
The two units'CPP-1 and -2.
used ISO Viscosity 68.
The inspectors verified that the viscosity of this type oil was within the specifications provided by Dresser and Woodward.
7.3 Aeration As previously mentioned, the inspectors noted that oil aeration had caused problems with both the lubrication and the hydraulic control of the turbine.
Aeration will occur with excessive oil in the turbine bearings or during addition of oil to the governor.
The loss of lubrication and possible bearing damage can result from aeration in the bearings; erratic speed control can result from aeration of the governor actuator.
It was for these reasons that the vendor specified oil levels for the components.
For example, the vendor required the oil level in the bearings to be at least 6.35 mm [0.25 in] above the bottom of the slinger ring, but no more than 1.5875 cm [0.627 in] above the bottom of the slinger ring.
The inspectors rioted that all but two of the units attempted to control the level within these guidelines.
At WCNGS, the minimum and maximum levels were not annotated on the sight glass.
At CNS the inspectors noted that the turbine oil level was approximately 6.35 mm
[0.627 in] above the high level mark.
The system engineer at CNS was not aware that a high oil level could be detrimental.
After discussing the issue with Dresser, the licensee lowered the oil level to within the guidelines.
As stated in NRC Information Notice 94-84, "Air Entrainment in Terry Turbine Lubricating Oil System,"
dated December 2',
1994, the use of a 3.81 cm [1.5 in]
oil return line from the bearings minimizes the effects of aeration in the bearing oil.
However, the inspectors noted that two units (ANO-2 and WNP-2)
had installed a 2.54 cm [1 in] line.
In additions ANO-2 personnel stated that they had observed the aeration phenomena after overfilling the oil system by 3. 175 mm [0. 125 in].
The licensee had tried unsuccessfully to vent the oil to improve drainage.
During subsequent troubleshooting, ANO-2 found one internal orifice missing.
As corrective action.
ANO-2 personnel reset the pressure regulating valve and replaced the missing or>fice.
They now believe that they
,
have resolved the oil aeration issue.
The licensee for WNP-2 reported that they had not experienced any oil aeration problems.
The inspectors also noted that aeration of the oil for the governor actuator could occur during the addition of oil to the system.
For this application.
aeration could result in erratic operation of the governor actuator.
The inspectors verified that procedures to remove the air prior to returning the system to an operable status were in place at all units.
The inspectors did not consider any unit to be currently vulnerable to aeration caused by incorrect oil addition practices.
Several units had changed thei r turbine lubricating oil to address the problem of excessive solid formation due to vapor-phase inhibitors plating out at high standby temperatures.
(Reference Table 5. "Lubricating and Hydraulic Oils,"
attached, for site-specific information.)
The inspectors asked the licensees if they had performed any type of extended run to ensure that the change in lubricating oil did not inadvertently introduce new aeration problems.
The inspectors noted that the original qualification at Oresser was 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> in length.
The vendor stated that during initial turbine testing a 4-hour run was performed and used to detect aeration problems.
If aeration occurred, the turbine oil system was modified prior to shipment.
The inspectors noted that none of the licensees had performed a 4-hour run after an oil change out.
Most turbines were run approximately 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> after the oil change.
The inspectors noted that this testing was not comparable to the original qualification tests.
7.4 Oil Filtration The inspectors noted that the location of the steam admission valves could also have a negative effect on the governor actuator response because the elevated temperature contributes to oil breakdown.
Also, steam admission valve leaks can result in elevated temperatures and moisture.
Contaminants can be controlled by such means as the use of filters when adding new oil, and the sampling of the oil on a periodic basis.
The vendor recommended that oil be filtered through a 5p filter prior to adding to the system.
Licensee personnel at three units (CNS.
RBS and WAT-3) were not filtering'heir oi 1 prior to addition to the system.
The inspectors found that the oil was sampled on a periodic basis at all units and no actual.contaminated samples had been identified.
7.5 Oil Viscosit Re uirements As with any hydraulic control. system, the viscosity of the oil affects the response of the system.
As used in the safety-related steam turbine-driven standby pumps.
the oil is usually more viscous when the component receives a
start signal (because the temperature is cooler) than when operating.
The thicker oil has trouble flowing through the small passages in the governor actuator.
resulting in a sluggish response.
As stated above.
the inspectors found that an oil with.an ISO Viscosity 32 was used at all but two units (DCPP-1 and -2).
At.those units.
an oil with an ISO Viscosity 68 was used.
While the ISO Viscosity 68 would not function at as
l
low a temperature as the ISO Viscosity 32. the environment at DCPP was such that the lower temperature capability was not a factor.
However. the oil with the ISO Viscosity 68 was capable of operating at a higher temperature.
which was an added benefit.
The inspectors concluded that licensee personnel had specified 'appropriate oil viscosity limits.
7.6 Maintenance Practices Maintenance practices related to oil included sampling, replacement.
and filtration prior to addition.
In a September 21.
1993. letter to the licensees, Dresser recommended a maintenance program to sample the oil at intervals of, 30 days and renewal of all additives at 6-month intervals.
-The inspectors found that an oil sampling program to evaluate the condition of the oil in the standby steam turbine-driven pump systems was in place at all of the units.
Licensee personnel sampled the oil periodically (usually annually)
and no oil had been replaced as the result of unacceptable oil test results.
i I
TABLE 1
- TURBINE SURVEILLANCE REQUIREMENTS TURBINE SURVEILLANCE ANO POST.
HOOIFICATION TESTING REQUIREHENTS ANO-1 ANO-2 CNS CPSES-1.-2 OCPP-1. -2 PVNGS-1.-2.-3 RBS SONGS-2.-3 STP-1. -2 WNP-2 WAT-3 WCNGS COLO QUICK START TEST REQUIREO BY PROCEDURE (INFORHATION NOTICES 88-09 and 93-51)
INITIALTURBINE TEHPERATURE IS RECORDED INITIALTURBINE TEHPERATURE IS RECORDED NO QUARTERLY QUARTERLY YES RECOHHENOEO QUARTERLY
'YES YES ANNUALLY YES
.1
)
il l,
t
'
~ ~
~
~
)
~ t
)
~
~
f I
~
g
/
~
~
~
a I
I
~ ~
I '
g
~
$
~
~
~
)
~
~
I I'
g
~
~t
~
'f
~
~
I If
~
TABLE 3
-
GOVERNOR SYSTEMS STEH HATERIAL ( INFORHATION NOTICE 94-66 and SUPPLEHENT I)
PLANNED STEH HATERIAL CHANGE (INFORHATION NOTICE 94-66 and SUPPLEHENT 1)
GOVERNOR VENDOR PREVENTIVE HAINTENANCE ACTIVITIES PGA/PG-PL SHUTDOWN ASSEHBLY INSTALLED ( INFORHATION NOTICE 86-14 and SUPPLEHENT 2)
GOVERNOR/STEM TESTS ( INFORHATION NOTICE 94-66 and SUPPLEHENT I)
ANO-I
)NO-2 CNS 410 STAINLESS STEEL (SS)
WITH NICKEL CHROHE STEH TREATHENT F ERRALIUH 410 SS WITH GAS NITRIOED STEH TREATHENT FERRAL IUH N/A INSPECT INSTALLED STEH IN NOVEHBER LIQUID NITRIOED STEH IN STORES REFURBISH EVERY THIRD OUTAGE REFURBISH EVERY THIRD OUTAGE NONE N/A N/A N/A QUARTERLY SPEED TRACES HONTHLY. SOON TO BE QUARTERLY SPEED TRACES STOPPED WEEKLY HANUAL STROKE TESTS IN HAY 1995 CPSES-1
.
OCPP-1.
~ 2 PVNGS 1.-2.-3 RBS SONGS-2.-3 STP-1. -2 INCONEL 410 SS WITH GAS NITRIDED STEH TREATHENT 410 SS WITH UNKNOMN TYPE OF STEH TREATHENT, PROBABLY LIQUID-NITRIDED 422 SS WITH ALUHINUH NICKEL STEH TREATHENT 410 SS WITH NICKEL ALUHINIZING STEH TREATHENT INCONEL N/A N/A INCONEL 718 N/A N/A N/A
~
REFURBISH EVERY FIVE YEARS NONE REPLACE EVERY FIVE YEARS REPLACE EVERY THIRD OUTAGE REPLACE EVERY 10 YEARS REFURBISH EVERY FIVE. YEARS YES NO. ADHINI STRATI VELY CONTROLLED N/A N/A N/A YES HONTHLY VALVE POSITION INDICATION. STRAIN GAGE.
DISCHARGE PRESSURE.
VISUAL HONTHLY FLOW TRACES QUARTERLY SPEED TRACES
&-HONTHLY HANUAL STROKE BIWEEKLY STROKE
& BREAKAWAY TORQUE SPEED TRACES AT DISCRETION OF SYSTEH ENGINEER HONTHLY SPEED TRACES WNP-2 WAT-3 MCNGS 410 SS WITH UNKNOWN TYPE OF STEH TREATHENT FERRALIUH 410 SS WITH GAS NITRIOED STEH TREATHENT INSPECTION PUSNED DURING NEXT OUTAGE
.
~
N/A 410 SS WITH LIQUID NITRIDED STEH TREATHENT IN STORES NONE REPLACED HAY 1994 NONE N/A N/A N/A QUARTERLY SPEED TRACES
& HANUAL VALVE STROKE WEEKLY 0.125 HANUAL STROKE SEHIANNUAL SPEED TRACES
I
"
f
~
~
TABLE 4
-
CONDENSATE CONTROLS ANO-I ANO-7 CNS CPSES-I..2 OCPP-l.. -2 STEAM SUPPLY INDICATION FOR CONDENSATE REHOVA!
( INFORHATION NOTICES 86-14 and 93-51)
NO YES YES EXHAUS I SIDE INDICATION FOR CONDENSATE REMOVAL ( INFORHATION NOTICE 93-51)
PLANNED PLANNED NO LICENSEE MAINTAINS STEAM SUPPLY VALVE LEAKAGE CRITERIA WHICH IS I ESS THAN TURBINE ROLL (INFORMATION NOTICE 86-14)
FNGINEERING JUDGEHENT fNGINEERING JUDGEMENT NOT DETERMINED NOT DETERMINED NOT SPECIFIC OPERATOR MONITORING ( INFORMATION NOTICES 86-14 and 93-51)
DRAIN ONCE PER SHIFT DRAIN ONCE PER SHIFT HONITOR UPSTREAM LEVEL ALARH MONITOR UPSTREAM LEVEL ALARM HONTHLY TRAP TESTING STEAM TRAP PREVENTIVE HAINTENANCE ( INFORHATION NOTICES 86-14 and 93-51)
YES YES ONCE PER REFUELING CYCLE YES YES, QUARTERLY SONIC TES'fS CONDfNSATE REHOVAl. SAFETY FUNCTION RECOGNIZED PARTIALLY SAFETY-RELAIf0/
PRESSURE BOUNDARY ONLY PARTIALLY SAFETY-RELATED/
PRESSURE BOUNDARY ONLY NONSAFETY-RELATED SAFETY-RELATED LEVEL INSTRUMENTATION SAFETY-RELATED PVNGS ~ 1. -7.-3 NO 500 LBH/HR (-1 GPH)
DRAIN ONCE PER TOTAl. LEAKAGE FROM ALL SHIFT FOUR STEAM ADHISSION QUARTERLY VALVES THERMOGRAPHY YES SAFETY-RELATED DRAIN RBS SONGS-2.-3 STP-1. -2 WNP-2 WAT-3 WCNGS YES YES YES YES YES YES YES NO NO YES NO NOT SPECIFIC NOT SPECIFIC NOT OETERHINfD NOT OETERHINED NOT DETERHINEO NOT SPECIFIC LEVfL ALARHS.
QUARTERLY LEVEL INDICATION VERIFICATION CHECK LEVEL ANO/OR TEMPERATURE ONCE PER SHIFT N/A LEVEL ALARMS DRAIN ONCE PfR SHIFT LEVEL ALARM YES N/A N/A YES YES CONDITIONAL PH LEVEL INDICATION SAFETY RELATED PARTIAL!.Y SAFETY-RELATED /
PRESSURE BOUNDARY ONLY NONSAFETY-RELATED NONSAFETY-RELATED NONSAFETY-RELATfD SAFETY-RELATED LEVEL CONTROL VALVE FOR CONDENSATE REHOVAL
0'
TABLE 5
- LUBRICATING AND HYDRAULIC OILS ANO-1 ANO-2 CNS CPSES-1.-2 DCPP-I..?
PVNGS. 1.-2. -3 TURBINE /
GOVERNOR OIL CHEVRON GST-32 CHEVRON GST-32 VAPROTEC LIGHT HOBIL DTE-797 SHELL VSI-68 /
GST-68 SHELL VSI-32 STANDBY TEHPERATURE
) 120'F YES YES NO NO CHANGED OIL TYPE YES YES YES YES YES
HOUR RUN AFTER OIL CHANGE N/A N/A Sy FILTER ( INFORNTION NOTICE 86-14.
SUPPLEHENT 2)
YES YES YES YES GOVERNOR OIL SmP INSPECTION (INFORHATION NOTICE 86-14.
SUPPLENENT 2)
YES YES YES YES YES GOVERNOR OIL SYSTEH FLUSH AttD CLEAN ( INFORNTION NOTICE 86-14.
SUPPLEHENT 2)
YES YES NO YES YES YES CHANGE TURBINE OIL AND FILTER ( INFORNT ION NOT ICE 86-14.
SUPPLEHENT 2)
YES YES YES YES YES YES OIL LFVEL INDI CAT ION SAT SAT SAT SAT SAT SAI RBS VAPROTEC LIGHT EVALUATING FOR CHANGE DURING NEXT REFUELING OUTAGE N/A NO NO.
BUT NO SAHPLED GOVERNOR OIL YES SAT SONGS-2 -3 STP-1. -2 VAPROTEC LIGHT HOBIL RAD-797 YES YES N/A NO YES YES YES YES YES YES YES YES SAT SAT WNP-2 MAI-3 HOBIL DTE-797 HOBIL DTE-797 NO YES YES YES NO NO YES NO REPLACED EGR YES YES SAT SAT WCNGS HOBIL RAD-797 YES YES YES YES YES YES UNSAT
I
ATTACHMENT 1 PERSONS CONTACTED Arizona Public Service Com an K. Chavet.
Industry Operating Events Coordinator
- B. Eklund
~ Consultants Compliance Group A. Fernandez.
Mechanical Maintenance Engineer N. Hallas, Materials Engineer C. Landstrom.
System Engineer C. Lewis. Senior Engineer.
Instrumentation and Controls 1.2 Dresser-Rand C.
1.3 Slater
~ Service Engineer Enter 0 erati ons Inc.
1.3. 1 Arkansas Nuclear One. Units 1 and
- T. Mitchell.'nit 2 System Engineering Manager
- T. Morse.
System Engineer, Unit
- D. Nilius. System Engineer, Unit 2
- B. Shor't
~ Licensing Engineer
. Smith, Licensing Supervisor
. Wrape. Unit 1 System Engineering Manager 1.3.2 River Bend Station
- D J.
- E
- W.
- J wR
- G Gi lley, System Engineering Supervisor Maher
~ Licensing Engineer Roan, Acting System Engineering Supervisor Stuart.
System Engineer Summers'icensing Specialist'est.
System Engineering Manager Zinke, Quality Assurance Manager 1.3.3 Waterford Steam Electric Generating Station, Unit 3
- R
- p
- T
%J
- R R
- D
- D
- D
- H Burski. Director Nuclear Safety Gropp, Mechanical Design Engineering Supervisor Gaudet.
Operations Licensing Supervisor Hologa, Manager of 'Mechanical and Civil Design Engineering 0;Quinn.
Mechanical System Engineer Quinnold
~ System Engineer Shipman.
Plant and Scheduling Manager Urciuoli. Senior Licensing Engineer Vinci, Licensing Manager Waldschmidt.
System Engineer
1.4 R.
D.
1.5
- M.
%J T.
- p
- R.
- O B.
1.6
- J W.
- M C.
- p
- L 1.7 J.
- B
- J
- G L.
1.8 Houston Li htin and Power Com an Asbury, Auxiliary Feedwater System Engineer Chambers, Acting Power Production System Engineering Supervisor Schulker, Compliance Engineer Kanavos.
Manager Mechanical Fluids Division Nebr aska Publ ic Power Distri ct Bennet.
Nuclear Licensing and Safety Supervisor Gausman, Plant Engineering Manager Gauthier, System Engineer Graham, Senior Engineering Manager Jones'enior Manager Safety Assessment Olson.
Core Cooling Supervisor Victoria Licensing Engineer Pacific Gas and Electric Com an Bard, Senior Engineer, Condensate and Feed Barkhuff. Piping Engineer Coward, System Engineer Harbor.
NRC Interface Milne, System Engineer Parker.
Independent Safety Engineer Southern Cali fornia Edison Com an Blakeslee.
Supervisor.
Heat Removal Diamond, System Design Engineer Hirsch, Manager, Power Generation Kaplan, Compliance Engineer Marr, Cognizant Engineer Plumlee
~ Compliance Supervisor Pressey, Business Administration Supervisor, Maintenance Texas Uti 1 ities Electri c Com an
- R
- T
- B A.
1.9 Flores.
System Engineering Manager Hope, Licensing Supervisor Reppa
~ System Engineer Saunders, Nuclear Steam Supply System Engineering Supervisor Washin ton Public Power Su
S stem
- L
- D
- T
- B Fernandez, Licensing Engineer Gi roux. Previous System Engineer Hancock.
System Engineer Pensee Previous System Engineer
/"v f
t
1.10 Wolf Creek Nuclear 0 eratin Cor oration
. Ferrel.
System Engineer
. Grieves, Supervisor, Auxiliary Systems S. Hatch'ngineering Specialist
- W. Norton, Manager.
System Engineering
- L. Ratzlaff. Engineering Supervisor C. Reekie.
Compliance Specialist III
- J. Yunk. Engineering Specialist
- S. Yunk, System Engineer 1.11 NRC
- W. Smith. Senior Resident Inspector The personnel listed above.
which are marked with an asterisk.
attended the exit interview
'by telephone.
EXIT MEETING An exit meeting was conducted by telephone on November 9, 1995. with personnel from each facility.
During this meeting, the inspectors reviewed the scope and findings of this report.
Licensee personnel expressed positions on the inspection findings which have been documented in this report.
The licensee did not identify as proprietary any information provided to. or reviewed by, the inspector ~C t
f J
ATTACHNENT 2 LIST OF ACRONYHS AFW ANO CFR CNS CPSES DCPP EFW NRC PVNGS RBS RCIC SONGS STP WAT-3 NGS WNP-2 Auxiliary Feedwater System Arkansas Nuclear One Code of Federal Regulations Cooper Nuclear Station Comanche Peak Steam Electric Station Diablo Canyon Power Plant Emergency Feedwater System Nuclear Regulatory Commission Palo Verde Nuclear Generating Station River Bend Station Reactor Core Isolation Cooling System San Onofre Nuclear Generating Station South Texas Project Waterford Steam Electric Generating Stati Wolf Creek Nuclear Generating Station Washington Nuclear Project-2 on, Unit 3
4l