ML20116M962
| ML20116M962 | |
| Person / Time | |
|---|---|
| Issue date: | 11/30/1989 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| References | |
| NUREG-BR-0125, NUREG-BR-0125-V01-N6, NUREG-BR-125, NUREG-BR-125-V1-N6, NUDOCS 9608210192 | |
| Download: ML20116M962 (11) | |
Text
,
PDR.
OFFluE OF NUCLE AR RE ACTOR REGULATION
~ U.S. NUCLEAR REGULATORY COMMISSION i,
a m
TECHNICAL nuaEe/en.0,2s NEWSLETTER nonkM Improving NRR's Performance on Licensing Actions by Thomas E. Murley in the past couple of years, I have placed a major emphasis on Holahan, and their staffs. We have reduced the inventory level to improving our performance on licensing actions to clear up some the goal established by the Commission, completed nearly all of the long-standing problems and to hold ourselves to the same technical specification license amendments over 2 years old, devel-standards of accountability to which we hold our licensees, oped systems for controlling additions and for accomplishinglow-priority actions, and maintained accurate records of the TMI To be specific, the following licensing action problems faced generic actions. We are well on the way to improving our records NRR:
of unresohed safety issues, generic letters, and bulletins, and we will soon begin to improve our records of generic safety issue (a)
There was an inventory of old, low-priority actions that actions.
lingered without resolution.
I have asked Frank Gillespie to take the lead in setting up a (b)
We did not have a program for controlling additions to comprehensive management information system and a centrally the inventory, controlled source of accurate records oflicensing actions. Under his leadership, we are well on the way to having those systems in (c)
We did not have a system for working on low-priority place.
amendments and actions.
In order to seek more broadly those activities in NRR that should (d)
We did not have accun:e records of how we imposed have improved management control systems, we have begun devel-generic reqtirements, nor of how the licensees imple-oping an NRR Ouality Program. This program is not firmly mented the requirements and how we finally closed the defined as yet, and we will keep the staff informed as we move issues.
carefully to implement it so that ongoing staff activities to accom-plish our heasy workload are not disrupted.
(c)
We did not have an effective management information system in place to track planned and accomplished work I appreciate the staff's work in making these major improvements and to maintain accurate records of the inventory, in NRR's performance on licensing actions, and I ask for your continued support as we finish the task.
Because of these lingering problems, a perception existed that NRR did not have the licensing action inventory under control and, further, that NRR was not giving the matter adequate Mat's in this issue?
management attention.
See Page 2 In the past fewmonths, we have made substantialimprovements, A
n
'f' largely as a result of the excellent work by Steve Varga, Gary 9608210192 891130 l
j
.. A
~
In mr ny cases an SSFI hns exposed mJjor shorico*mingsin configu-ration management programs. (Configuration' managqpe'nt is the IN THIS ISSUE process of ensuring that plant systems and compon*ents are ma, int tained in keeping with their intended design bases.) SSFIs often found that equipment was improperly maintained or modified Improving NRR's Performance on Licensing Actions because design-basis documentation was missing or was being by Thomas E. Murley.
1 used inappropriately.
Region III Oversight of Licensee-Initiated Configuration Management Programs SSFIs quickly received national recognition as a valuable diagnos-by II. Miller, T. M artin, M. Phillips, and tic tool. Given the safety payoff of these inspections and limited I. Ym...
2 NRC inspection resources, NRC encourages its licensees to con-Region I Assessment of PRA Applications by duct their own SSFIs. This encouragement took many forms, Licensees including NRC participation in a special American Society for by William Kane and Eugene Kelly -
3 Ouality Control (ASOC) industry seminar on SSFIs in 1986 to Inservice Inspection of Steel Containments review the process and the experience oflicensees that had been by Chen P. Tan 4
through an NRC-led SSFI. Utilities that had undergone SSFIs Improvements to Technical Specification Surveillance reinforced the message being sent to the industry at large; SSFIs Requirements 1
are likely to identify weaknesses at most plants, and licensees by R. lobel and T. Tjader.
5 should find and correct their own weaknesses before they became Emergency Response for llurricane 11ug significant problems. The industry through the Electric Power f
by William Rankin, Eldan Testa, and Nancy Hughey. 6 Research Institute, developed NSAC-121, " Guidelines for Per-Maintenance Team Inspections formirg Safety System Functional Inspections," which was issued by K. Hart.....
7 in November 1988.
Requalification Program Evaluations by William Dean.
9 With this encouragement and emphasis by Regional management, Initiative to Improve Licensee Events Evaluation by the end of 1987 nearly all of the nuclear plants in Region III had Programs, Region V m
implemented some form of self-initiated SSFI and configuration byJesse L. Crews.
10 management review. A similar effort may be found in other Regions. This effort has involved a significant commitment of NEWSLETTER CONTACf:
engineering-oriented resources for most licensees. In several Valeria Wilson, NRR 492-1208 instances, these programs will take up to 5 years to complete and will encompass all safety-related systems as weil as selected non-safety-related systems.
Re@n W WersMt Although each licensee has a program tailored to the needs of his a
particular plant, ali the programs have included the following of Licensee-Initiated e'e=ats: remastiteioa of desisa-basis iarormation ana docu-mentation; detailed walkdown of electrical and mechanical sys-Configuration Management tems to ensure as-built conditions match design; and detailed evaluation of selected systems to confirm that they remain func-Programs tionai.
by H. Miller, T. Martin, M. Phillips, and I. Yin, DRS, Region 111 In addition to enhancing operator confidence in system design and reliability, these efforts produce a better set of reference materials In 1985, a new NRC inspection approach, called a Safety and tools for the design engineers who will continue to modify the System Functional Inspection (SSFI), was developed by the plants and deal with equipment-aging issues.
Office of Inspection and Enforcement. This inspection was intended to be a comprehensive " vertical-slice" review to While the licensees' initiative may reduce somewhat the need for evduate whether a particular safetysystem had been designed, NRC to conduct SSFIs, the scope and potential safety significance constructed, maintained, tested, and operated in a way that of these efforts dictate some form of NRC oversight. The two would ensure it met its required safety function. These Principal objectives of this oversight are: (1) to understand the resource-intensive inspections, which typically involve 3 weeks depth and effectiveness oflicensee reviews as well as the findings of field time for as many as 10 inspectors, were very effective that are being made and (2) to assess the promptness and effective-in identifying major design, modification, maintenance, and ness of licensecs' corrective actions and the accuracy of licensee operational deliciencies that could impact safe plant opera. reports of significant findings.
tion. SSFIs are still being performed, on a limited basis, by NRC Headquarters and the Regional Offices.
This oversight is important if NRC is to give appropriate credit to 2
_. _ ~ _ _ _ -
l
~
licensecifortheir cfforts and to identify where such initiatives cooling water during a postulated high. energy-line break. These might b'owcak, making NRC inspection prudent. The challenge and other findings have resulted in licensees submitting many facing Region *lll has been to conduct such oversight without voluntary Licensee Event Reports. Ilowever, some of the prob-discouraging licensee initiative. To accomplish this, the Region lems and limitations that the NRC has found in these programs, l
has emphasized direct communication with senior licensee particularly with respect to corrective action, have strengthened officials on these inspections and their findings. While the kind the Region's perception of the need for continued oversight.
of costly and intense scrutiny that is a part of these reviews can hloreover, the continued interest and invoh'ement of the Region be painful, Region III has reminded the licensees of the safety have helped licensees' senior managers see the benefits of these and other benefits that accrue from competent efforts. For often costly, difficult programs.
i example, licensees have been made aware that selection of candidate sites for NRC-led SSFIs would be based partly on whether the facility had a credible self-initiated SSFI and con-figuration management program. Consistent with changes Region i Assessment of I
made to the NRC enforcement policy,licensecs have been told PRA Annlications bv Licensees that enforcement discretion would be granted whenever pos-rr e
sible for licensee identification of violations.
by William Kane and Eugene Kelly, Region l As a result of the level oflicensee activityin this area, Region Ill reduced the number of SSFis actively performed by Region Region l's Technical Support Staff (TSS) in the Division of Reac-personnel and shifted resources instead to monitoring the per. tor Projects is assessing how licensees apply Probabilistic Risk l
formance of licensee programs. This effort includes periodic Assessment (PRA) to their operational and maintenance pro-meetings with licensees, including senior management, and on-grams. The first assessment was conducted at Limerick in Decem-site review of self-initiated SSFI reports. The on-site review has bcr 1988, and the second at Susquehanna in October 1989. These consisted of approximately one to two person-weeks of effort assessments are not as extensive as those done by the PRA teams devoted to reviewing the licensee's report, conducting staff nor are they intended to focus on the same areas as those done by interviews, and evaluating corrective action taken as a result of the mandatory maintenance teams.
the SSFI findings. (In some cases, licensees have been reluctant to allow the results of their self-initiated reviews to be released These inspections are centered on three parameters: transient i
off site, making it necessary for NRC to conduct its review at the initiator frequency, safety system availability, and human error licensce's facility.) These reviews have shown that the inspec. rate. The numerical values of these parameters are critical to the tions often identifysignificant safetyissues. The Region has also quantification of core damage frequency in a PRA. The inspec-l found that while t he licensee's corrective action has been accept. tions evaluate actual plant experience in light of the values used in able in most cases, sometimes it has been less aggressive than the PRA. Also, part of the inspection is focused in the balance of warranted. There appears to be a tendency on the part of the plant, particularly where plant modifications have been made l
licensees to treat the findings of their inspections differently than to correct vulnerabilities identified by the PRA.
they treat the findings of an NRC inspection, even when the findings are just as significant.
In the case of Limerick, the inspection found that total unavailabil-ity of certain risk-significant systems was, in actuality, significantly On a selected basis, the Region expects to assesslicensee efforts higher than assumed in the PRA. As a result the licensee is now by cctually performing SSFIs on the same systems reviewed by tracking the total unavailability of risk-significant systems and l
the licensee. Region Ill has recently completed an SSFI of the using that information to plan preventive maintenance activities.
Iligh Pressure Core Spray (HPCS) system at one utility to That inspection served to underscore the insights from PRAs, and evaluate the utility's own SSFI of the same system. Based on the the importance that PRA can playin day-to-day operational aethi-findings, the Region III team concluded that the utility initiated ties. These inspections have been well received bylicensees in part SSF) had some limitations in that it did not challenge original because of their focus on safety.
dedgn or construction where warranted nor did it adequately l
vorify that all Technical Specifications required conditions were The TSS intends to conduct these inspections at additional l
Iming monitored for conformance.
Region I facilities, particularly thw that have PRAs and are actively using them.
Currently, most Region III licensees have active configuration management improvement programs and many conduct system evaluations patterned directly after the NRC SSFI model. Region III has seen significant safety payoff from the licensecs' pro-gram. For example, one licensee conducting an SSF1 of the instrument air system identified a design deficiency that could render both station diesel generators inoperable. Another licensee identified a potential for complete loss of component 3
was at r below the minimum specified wall thickness.vl'o "ed'uce ins"^rVICO lnSDCCllOn Of further thinning of the torus shell, the licensee planned.(1) to a' ply r
p Steel Containments a Protective waiing or (2) to reanalyze the minimum wall thick,-
by Chen P. Tan, ESGB ness based on calculated stresses and actual material properties.
The licensee may also install an additional saddle support at mid-span of each torus bay. A survey of BWRs in NRC Region i INTRODUCTION sh wed that some ton had expenenced degradatim of the coating and that cleaning and recoating were required.
Ever since nuclear energy has been used for civilian power gen-f,g g %,
crition, steel has been the material of construction for low-internal-pressure and/or small volume containment vessels.
On August 24,1989, the licensee of McGuire, reported the discov-Steel vessel technology is best adapted for these vessels and for cry of base metal corrosion of the Unit 2 steel containment vessel this reason, most boiling water reactors (BWRs) Mark I and (SCV) at plant elevation 725 feet. The two units at McGuire use pressurized-water reactor (PWR) ice-condenser contamments the PWR ice condenser containment. During the pre-Integrated tre built of steel.
Leak Rate Test (ILRT) inspection, coating failures were found on the outside of the Unit 2 containment shell. The shcIlis enclosed Containments are designed to contam radioactisity in the un-in a reinforced concrete shield building, with a 6-foot annulus. The likely event of a serious reactor accident. Unlike the reactor failure of the coating has led to the corrosion of the base metal up vessel, the containment vessel,is not subject to the design to 0.123 inch. The degradation of the shell, which has a nominal pressure and temperature for most ofits design life, thickness of 1 inch at that elevation, is limited to a 37-foot section Although the containment vesselis the last of several passive defense lines in the complex system of engineering safeguards, On August 29 and 30, the licensee inspected the Unit 1 SCV and there has been no inservice inspection (ISI) requirement for found similar pattern of corrosion.
treel containment vessels simdar to that accorded to reactor vessels. The only requirement is a visualinspection before each integrated leak rate test-The condition that led to the SCV corrosion is ponding of water in some areas of the annulus. The water comes from condensation PROBLEMS and leakage from some of the instrumentation lines. It is difficult to drain water from the annulus floor.
BR Dow#
After NRC inspected the Unit 1 containment, the licensee made a number of commitments and proposed actions to reduce or Steel containment shell corrosion was first discovered in the prevent further corrosion. These can be summarized broadly as sand cushion region of the drywell of Oyster Creek. The drywell follows: coat the lower portion of the SCV, recoat the annulus shcIl thickness had been reduced from 1.115 inch to an average floor, and seal the joint between the SCV and concrete floor, thickness of 0.85 inch, with some local spots reduced to about correct the water / acid leakage improve the drainage of water in 0.73 inch. The root cause of the corrosionis very corros,ive water the annulus, rek>cate interfering equipment such as the HVAC in the sand cushion. The water leaked from the refueling pool on duct, perform more frequent surveillance and inspection, develop the top of the drywellinto the air gap between the drywell and the acceptance criteria for weld repair. The commitments are for both concrete shield, then through the gap-forming material, which units, and are of short term, interim and long term nature.
contains chlorides and sulfides to the sand cushion. There is also some corrosion in the upper portion of the drywell where gap-The McGuire licensee also is the licensee of two units at Catawba forming material is present. Water leaked from the refueling whose SCVs are similar to thosc at McGuire. An inspection of the pool through a defective rubber gasket. To ensure the integnty Catawba SCVs found less extensive corrosion than at McGuire.
of the drywell and to reduce he rate of corrosion, the licensee s
The corrosion at Catawba was limited to about s 15-foot section has committed to periodically measure shell thickness through 1 inch above the annulus floor, ultrasonic testing (UT) during outages when a drywell entry is planned or required. The licensee also installed a cathodic CONCLUSION protection system.
Because of the corrosion of BWR and PWR steel containments BU h described above, adequacy of current requirements for contain-ment must be examined. Currently a visual inspection of the The corrosion of the torus at Nm.c Mile Point I was discovered containment vessel surface before each ILRT is required. With during an NRC special announced team inspection. The Nme the exception of a periodic tendon surveillance required for Mile Point I torus was designed and constructed uncoated. The prestressed concrete containments, no other inservice inspedons UT of the torus shell showed several areas where the thickness are required for stect and reinforced concrete containments. Not 4
.S until recintlydid the American Society of Mechanical Engineers that have been caused by surveillance testing:
Section til Division 1 and Division 2 Codes stipulate any require-ments for ins'ervice inspection. Since no inspections were re-Oconec Unit 1, a Babcock and Wilcox-designed pres-o quired, containments have no easy access for inservice inspec-surized water reactor, tripped from 100 percent power tions or even for visualinspection. Toinspect the drywellin the when an instrument and control technician,who was sand cushion region, concrete had to be removed. In the case of peforming a test of the reactor protection system failed to the steel shell, components in the annulus such as the IIVAC follow procedures and bypass a failed channel while testing duct had to be relocated. Both these operations are time another channel.
consuming and expensive. The staff is in the process of recom-mending ISI for steel containments and seeking input from At Surry Unit 1, a Westinghouse-designed three-loop o
industrial groups so that a realistic and practical procedure for reactor, calibration of a nuclear power range detector ISI can be established. The need for developing ISI requirments caused a blown fuse that resulted in a turbine runback.
for containments is further hightened by licensees' intent to Because of the configuration of the main feedwater extend the life of their facilities beyond that of the original system, the steam generator level rose to the feedwater operating licenses.
isolation setpoint, this produced a turbine trip, follow-ed by a reactor trip.
Improvements to At Millstone Unit 3, a Westinghouse-designed pressurized-o Technical Specification water reactor, an emergency diesel generator surveillance iest resuited in a turbine 1,ip ana,cacto,irip when a Surveillance Requirements b'e^' ' P'"*d *^"'i"8 ' ' P * ' ' ^ " ""i'^' b"'-
by R. Lobel and T. Tjader, DOEA Summary offindingr The NRR staff recently completed a comprehensive examina-Technical Specifications require a large number of surveillance tion of all Technical Specifications surveillance requirements tests. According to the licensee for Limerick Unit 1, the following to identify those that should be improved. This effort is an surveillances were performed there in a two-year period: in 1986, element of the Technical Specifications Improve-ment Pro-with no refueling outage,14,888 surveillances and in 1987, with a gram (TSIP) and the results will be presented in a NUREG, refueling outage,17,540 surveillances. Approximately 98 percent
" Improvements to Technical Specification Surveillance Re-of these were required by the Technical Specifications;the other 2 quirements," in the near future. In determining which re-percent were required by other agreements between the licensee quirements should be improved, the staff had lour criteria:
and the NRC. During the year with no refueling outage, an average of more than 40 tests a day were performed.
(1) The surveillance could lead to a plant transient that could challenge a safety system.
Equipment failures and personnel errors during several types of surveillance tests cause reactor trips. In addition, testing results in (2) The surveillance results in unnecessary wear on many spurious isolations of the control room, fuel handling build-equipment.
ing, auxiliary building, and containment ventilation. Inadvertent emergency diesel generator starts are relatively common results of (3) The surveillance results in radiation exposure to surveillance testing, and act uations and isolations of standby safety plant personnel that is not justified by the safety equipment occasionally occur.
significance of the surveillance.
While some testing at power is essential, safety can be improved by (4) The surveillance places an unnecessary burden on reducing the amount of testing at power.
plant personnel because the time required to per-form the surveillance is not justified by its safety Wear on equipment is also a significant concern. Some instrument t
I significance.
parts (such as connector pins and plugs) experience wear from the amount of plugging and unplugging required for testing. Auxiliary The details of the study are discussed below.
feedwater pumps are subjected to wear because of the small recirculation lines used during testing.
Recent Surveillance-Testing-Related Events Emergency diesel generators are subjected to an excessive amount Surveillance testing is important to ensure the operability of of testing, especially in plants with older Technical Specifications.
safety-related equipment. Ilowever,it can also cause plant transients that could challenge a safety system and other Radiation exposure to personnel as a result of surveillance testing operational problems. The following are events that have ranges up to approximately 20 percent of the total dose incurred at 1
occurred in the last year that illustrate the types of problems a site. (The biggest contributor to incurred dose is maintenance, 5
.*r*
not testing; however, some surveillinces do result in significant On September 14,1989, at 6:00 a.m., Region II began monitoring incurred redittion dose.) Tests that require containment entries the National Weather Service advisorics and 6ther real-time wather while the reactor is in operation (e.g., containment purge and information on the hurricane. As Hugo moved closer to the exhaust isolation valve leak testing) cause significant doses. mainland, the Region tracked it, and closely studied landfall pre.
Walkdowns of systems to check valve alignments and snubber dictions to try to determine where Hugo would strike. Region II operability also contribute significantly to radiation dose.
surveyed plant sites near the coast from Turkey Point in Florida to Surry in Virginia to ensure that adequate NRC presence would be Improving preventive maintenance programs is an important maintained at each.
element in reducing testing at power. A review of Licensee Event Reports and other data shows that many of the failures Hurricanes are classified in five categories (one is minimal, and found from testing are the result of such factors as dirt or five is the extreme). Hugo struck Puerto Rico as a category four impurities in fluid systems, bent or broken parts, loose parts, etc. hurricane, but it had lost so;ne of its fury as it set out on its course Surveillance testing generally only identifies that a piece of for the mainland.
equipment is in an inoperable condition so that the time it is inoperable can be limited; preventive maintenance, however, The Headquarters' Duty Officer was officially notified of Region can limit the number of failures that occur, ll's hurricane '. racking on September 18,1989. After a Region II j
management briefing on Hugo September 20, because of the Recommendations hurricane's position, forward speed, and intensity, a Hurricane Watch Team was established. The Regional Incident Response The staff study produced two recommendations.
Center was continuously staffed from early in the day on Septem-ber 21, and sites on the projected path of the hurricane were (1) A study should be done on the feasibility of using monitored to ensure that hurricane preparation procedures had reliability-based Technical Specifications.
been or were being carried out.
(2) Plans for advanced light-water reactors should in.
When landfall predictions indicated a high probability that South clude surveillance testing to ensure that operator bur-Carolina would be the target of the hurricane, Region Il reactor den, radiation exposure, equipment wear, and plant project staff members went to the Brunswick plant to assist the transients are minimized.
Resident Inspector and to monitor the licensee's preparation. In addition, radiological control inspectors who were at the nearby Benefits GE fuel facility and an inspector already on site for an Emergency Plan inspection were redirected to assist the Senior Resident The approved changes to surveillance testing will be integrated Inspector at Brunswick. Region I was apprised of Region II with of the overall Technical Specifications Improvement Pro-activities and the predicted course of the storm.
gram through individual plant conversions to the new Standard Technical Specifications or individual license amendments.
During the several hours before the hurricane reached the coast,its windspeed increased from 105 mph to 135 mph, and it's forward The implementation of the recommended changes to surveil-speed increased to 22 mph. It again became a category four lance testing should reduce reactor transients, radiation dose to hurricane.
personnel from testing, and wear on equipment. The reduction in testing will also increase the performance and availability of The maximum strength of the storm reached land at the Isle of safety-related equipment. A reduction in thc Technical Specili. Palms, a few miles north of Charleston, South Carolina about cations-related workload will mean that utility technicians and midnight on September 21. Once on shore, Hugo turned north-engineers have more time available for other work more impor-west toward Charlotte, North Carolina, and finally turned north.
tant to safety, such as preventive maintenance, which will result in an increase in reactor safety.
Although the focus of the early NRC response had been Brunswick
-- because of its coastal location was near Hugo's maximum intensity zone -- once a landfall location was established, the focus Emergency Response for of the NRC response changed to inland facilities in the hurricane's Hurricane Hugo by WiHiam Rankin, Eldan Testa, As the storm progressed inland, the Hurricane Watch Team and Nancy Hughey, Region 11 maintained close communication with each of the following licen-see sites: Vogtle, Robinson, Summer, McGuire, Catawba, Surry, Many of us in Region II watched intently in mid-September as and North Anna. The team also stayed in close communication Hugo moved steadily toward the United States and into the with General Electric, Westinghouse, Babcock and Wilcox Navy spotlight on local news programs. Hugo had already left its Nuclear Fuel Division, and Nuclear Fuel Services. The storm's destructive mark on Puerto Rico and the Virgin Islands, progress was monitored using the National Weather Sersice, 6
N
- f reports fr'om litensecs, end reports from the high frequency ham Inspections include (1) detriled plant walkdowns to evaluate gen-rtdlo emergency network.
eral housekceping and site material condition and (2) direct obser-vations of maintenance and maintenance-related s'upport activi-
, Communication was a particular concern; however, sophisti-ties. The requirement for Headquarters support provides an cated, portable, self-contained satellite communications system excellent opportunity for NRR personnel to get out to a site and and a satellite communication specialist from the Boise Idaho become familiar with inspection activities.
Interagency Fire Cache ensured continuous reliable communi-cation from the Brunswick site. (When these were released from Pilot Maintenance Team Inspections were conducted in July 1988 Brunswick, they were redeployed to one of the devastated South at Peach Bottom, Oconee, and Diablo Canyon. The feedback and Carolina counties that had lost its entire telephone system!)
lessons learned from these inspections were incorporated into the final version of Temporary Instruction (TI) 2515/97, which pro-The Region II Hurricane Watch Team was deactivated about 4 vides direction for the conduct of the Maintenance Team Inspec-p.m. on September 22, after a long and busy 34 hours3.935185e-4 days <br />0.00944 hours <br />5.621693e-5 weeks <br />1.2937e-5 months <br />. By this tions. The Tl provides detailed guidance on a maintenance pro-time, flugo had been downgraded from hurricane status, and it gram and presents a logic tree for use in evaluating licensees. The I
was moving north out of Virginia. Subsequent monitoring of the tree contains 43 individual elements to be evaluated during the storm was done by Regions I and III.
course of the inspection; it is useful in pulling together inspection results and presenting the information to the licensee.
During the height of Hugo's fury, the Region II Hurricane WatchTeam was supported bya Headquarters Hurricane Watch When a logic tree is completed,its elements are colored according Team. The NRR propct managers responsible for facilities that to the inspection results; green is for good, yellow for satisfactory, could have been affected by the storm and the AEOD incident red for poor and blue for not observed. A green block shows a response staff provided assistance and support in coordinating a program and implementation that need only very minor improve-focused Agency response. Technical specialists from NRR who ment. Yellow shows an average program and implementation, may have been needed during and after the hurricane -- such as with areas that need im provement balanced by better performance electrical experts in loss of offsite power events, meteorologists, in other areas. Red indicates an element is inadequate or missing.
l hydraulic engineers, and structural engineers -- stood by to Blue shows that a particular element was not evaluated.
provide their expertise.
Most of the elements in the inspection tree are divided into halves.
The Agency's response was successful with considerable team-The upper half of the block represents the programmatic aspect of work and coordination between the various Offices. The extra the element, while the lower half represents the implementation of effort of all those involved is sincerely appreciated.
that element.
The maintenance inspection program really contains nothing new, It is a combination of inspection activities that the NRC staff has been performing in the past, pulled together for a comprehensive Maintenance Team inspections look at maintenance and related activities. Even the inspection by K. Hart, DLPO tree is a concept that has been used in previous inspection pro-grams.
Good maintenance is a key factor in achieving and maintaining This inspection program consists of three major sections which are 5 wn c 6teet Phanchesof tktree: Mant Pedormance I
a high level of safety in plant operations throughout the life of a nuclear power plant. Thus, maintenance was chosen as the clated to Maintenaxe, (2) Management Support of Ete-special area of emphasis in the Mandatory Team Inspection Pro- "^"' ' "" ( )
- E #" " "", "*
gram. This emphasis will continue through 1990 because of the further subdivided into eight categories: direct measures related to increasedinterest in maintenance and the proposed rulemaking "lant perf rmance; management commitment and in p
- nagement orgasatmn; techm, cal support; work control; plant being considered by the Commission.
mamtenance organizat,on; mamtenance facilities, equipment, and i
gg g 7, g
gp material control; and personnel control.
Maintenance inspection teams consist of a team leader from the W
Regional Office; two reactor / project engineers and a radiation specialist. also from the Regional Office; and two engineers from As of the end of fiscal year 1989, inspections have been completed Headquarters. The inspection covers six weeks: one week of at 33 sites and results obtained from 31 sites. The overall evalu-preparation, two weeks of on-site inspection, one week of,
ationsindicate that most plantshave atleast a satisfactory program m
office inspection, and two weeks of documentation and report in place,with about half of the programs evaluated as good. The implementation of the programs, however,was considerably weaker.
- *E' This same overall conclusion applies to nearly every portion of the I
9 -
- inspection tree. Appropriite elements are in p!rce but tre not agement in the following ways:
always effectively infplemented.
Engineers have limited time to monitor the system for which they are responsible. (At o Plant Performance Related to Maintenance one plant, six of seven engineers had not 4
walked down their system in a year.)
In the area ofoperating data and general plant condition noted during walkdowns, the inspection results indicate Responsibilities not well defined.
some need for improvement in the area of housekeeping and plant material condition. About 10 percent of the Management depends heavily on contractor plants were rated poor in this area.
support which dilutes personal ownership.
Little training is provided.
e Management Suppoet of Maintenance Personnel are always in a reactive-not pro-In regard to management support of maintenance, about active-mode.
I 35 percent of the programs were rated good,58 percent satisfactory, and 7 percent poor. For implementation, only 16 percent were rated good, with 77 percent satisfac-o MaintenanceImplementation tory and 7 percent poor.
Maintenance implementation is the focus of at least 80 The area of Icchnical support was clearly the weakest, percent of the inspection effort. In this area, programs were with the engineering support being very weak. The pro-rated 52 percent good and 48 percent satisfactory. Implem-grammatic aspects of engineering support show 26 per-entation was 33 percent good,61 percent satisfactory, and cent rated as gcxxi,55 percent satisfactory, and 19 percent 6 percent poor. The significant weakness was in the area of poor. The implementation of engineering support rates maintenance trending, with nearly 25 percent of the sites 10 percent good,61 percent satisfactory, and 29 percent rated as poor. Weaknesses in this area included:
poor.
(1)
Reports typically indicated gross overall trends and Some of the reasons for the poor ratings include:
did not identify repetitive failures over a long per-iod, subtle trends, or component failure trends.
(1)
Repetitive failures of equipment were not identi-
}
fled as a basis for changes h. the scope of the (2)
Documented information on completed work pack-I preventive maintenance program.
ages was not adequate to assist in root cause analysis or aid in analyses of future trends.
(2)
Inadequate root cause analyses were performed when equipment failed.
(3)
Some programs were fragmented so that a single reviewer did not see all the available information.
(3)
Despite vendor recommendations, some preven-tive maintenance activities were not conducted, and no technical evaluations were performed to Significant problems with procedures were also identified. The support these exclusions.
two that recur frequently are (4)
For some preventive maintenance problems,it (1) procedures are not followed and took more than 2 years to reach an engineering resolution.
(2) procedures are not adequate for the task.
(5)
There is inadequate communication between The subcategory of personnel control stands out as the strongest technical support groups and maintenance groups area,with more than 58 percent rated as good in both program and concerning-implementation. Some of the specific strengths are:
Post-maintenance testing (1) a 3 to 1 ratio of experience maintenance personnel Failure analysis to apprentices Equipment qualification Surveillance and test procedures.
(2) well-defined programs for staffing and promotion, with detailed job descriptions and attributes sought (6)
System engineer concept not supported by man-for new hires 8
1
b.
-Sumntdry
- Nine Mile Point 2, Point Beach, Turkey goint, and Zion), the l
N.
spectrum of problems noted has ranged from a wgakness in the l
.Nearly half the maintenance team inspections have been com-facility staff's knowledge of a particular system's integrated plant i
pleted. Several general conclusions can be drawn from the in-response characteristics to inadequate attention by management spection results and activities to date. All sites have established to continuing training programs for operations. Concurrently,the maintenance programs. All but one have been evaluated as at scope and intensity of NRC remediation actions have been case-least satisfactory, wit h nearly half or them rated as good. The im-dependent, with resulting activities ranging from evaluating re-l plementation of the maintenance programs / process is not keep-training and conducting reexaminations that emphadzed the areas l
ing pace with documented programs. The maintenance inspec-of noted weakness, to performing operational evak onsof allthe tion program will continue to concentrate on the performance remaining crews to determine whether a suffi. i at number of and effectiveness of maintenance activities.
adequately trained and knowledgeable operators were available to maintain the plant in a safe status operationally.
l The NRC anticipated that there would be a " learning curve" among the utilities as they became familiar with the format of the Requalif.ication Program examination and more proncient at generating validated test items i
Evaluations thauncube uituia contained in ESM lt was also expected that facihty trammg staff members would have some difficulty m l
by William Dean, LOLB changing modes from being an instructor to acting as an evaluator.
i This has indeed been the case, and the NRC examiners have expended considerable time and effort in reviewing facility test The revised NRC requalification program described in NRR items and providing direction on how to correctly develop these i
Technical Newsletter Volume l, No.1, dated May 1988, was pilot test items, as well as providing feedback on operator evaluation l
tested from December 1987 to June 1988. The final program, as techniques. Improvements in these areas have been observed as a i
described in Examiner Standard ES-601 of NUREG-1021,"Op-result of communication within the industry about the program l
erator Licensing Examiner Standards," was implemented on and it is expected that these problems will become much less l
October 1,1988. This program consists of a two-phased exami-prominent in subsequent program evaluations.
nation process, with the examinations developed in a coopera-l tive effort between NRC's operator licensing examiners and Once a facility's program is evaluated as satisfactory,the NRC m senior operations and training staff from the facility. Each modify the evaluation process by using a methodology intended to operator is administered a two-part written examination during minimize the use of NRC resources. This change entails using one which facility-controlled reference materialis available, and an NRC examiner to observe the performance of two operators at a operational examination consisting of a simulator test that time in the operational portion of the evaluation (the simulator and emphasizes the operational skills needed by a typical shift crew, walkthrough examinations). This modified evaluation has been and a walkthrough examination focusing on the indhidual op-successfully conducted in a pilot program. The simulator test erator's ability to perform tasks important to safety.
re mains crew-orientr d, wit h critical indhid ual tasks pre-identified.
The walkthrough exainination still focuses on the performance of The results of the evaluations conducted by the end of FY 89 of each individual although shere will be fewer opportunities for the facility requalification programs are given below.
examiner to directly obsen ce each operator perform job related tasks. A lot of program effod indicate this procedure will not affect the ability of the examiner to adequately evaluate operators.
Facilities Operators Programs Operators Evaluated Examined SAT /UNSAT* Pass / Fall The experiences gained during the first full year ofimplementation of the revised requalification program indicate that the program 37 574 29/7 474/100 objectives of (1) implementing a process that meets the require-ments of 10 CFR 55.57(b)(2)(iv), which requires an operator to
- One facility did not meet the criterion in ES-601 pass an NRC-administered written and operating requalification that requires an evaluation of a minimum of 12 exam during the term of his 6-year license; (2) providing the staff operators to determine the status of a requalification with a program that can be used to assess the effectiveness of a program.
facility's requalification training program; and (3) minimizing the potential adverse impact m the safe operations of a facility while conducting this procedure have been achieved. The program is These numbers represent approximately35 percent of the facil-considered to be highly successful as it has rectified industry ity licensees and about 10 percent of the licensed operators.
concerns;it is an innovative process that serves well the NRC's
[i regulatory requirements.
For the seven facilities thus far determined to have an unsatisfac-tory requalification program (Hrowns Ferry, Ginna, Millstone 3, 9
i
,, c.,.,,
Initictivo to impravo Lic3ncoa AU p wu reactor lianus in Region V currently hakrevised.
events evaluation programs that require formal root cause;malysts Events Evaluation Programs in of selected events. As the implementation of these programs matures, licensees have been encouraged to expand the population
- Region V of events requiring root cause anaiysis. During the pasi year, for
(
example, procedures have been revised at the Diablo Canyon by Jesse L. Crews, Region V plant, increasing the number of events requiring formal root cause analysis from approximately 150 to as many as 1000 per year.
Included in the revised program are events invohing balance-of-plant systems and components. A new events evaluation program has been implemented at the Palo Verde site thisyear as well. This During the past 3 years, an initiative has been under way to focus Program, which is a graded (Category 1-4, in terms of safety on events evaluation programs of the power reactor licensees in significance) lacident Investigation Program,is estimated to cover Region V, This effort has been spearheaded by the Senior over 500 events per year.
Reactor Engineer in the office of the Regional Administrator.
l Region V's experience has been that licensees traditionally focus The objectives of the initiative have been to encourage licensees events evaluation effort in the area of hardware performance, to (1) develop improved skills within the plant organization for eften neglecting the area of human performance. In this respect, events evaluation, with particular emphasis on formaf root cause Region V licensees have been encouraged to become active analysis, and (2) increase the population of events subjected to participants in the Institute of Nuclear Power Operations (INPO) root cause analysis.
Iluman Performance Evaluation System (IIPES) program, and J
during 1989 all licensees were participating in this program.
Region V's initiative grew out of experiences at the Rancho Seco plant. In response to concerns identified by Region V and the Over the past 2 years, measurable improvement has been observed findings of a 1984 study of Rancho Seco problems by LRS Asso. in the performance of Region V licensees in the area of events ciated, a consultant to the licensee, the licensee established a evaluation. The term " root cause analysis" has become a common dedicated incident analysis group (IAG). This group was ini. buzz word within the licensee organizations. Expertise in root tially staffed by contract personnel and subsequently, in late cause analysis has improved significantly. There has been increas-1985, by six licensee employees selected on the basis of prior ing support and attention by senior licensee management in terms experience in the areas of plant operations, testing, and accident of staff resources, training commitment, and performance expec-or incident investigation. The IAG staff was also trained in tations in this important area of licensee activities. This senior various techniques of root cause analysis, including MORT management support has resulted,in most instances,in establish-
)
(management oversight risk tree) analysis, causal factors analy. ment of a focal point or a lead group within the nuclear organiza-f sis, change analysis, barrier analysis, fault tree analysis, and the tion for events evaluation and root cause programs.
Iluman Performance Evaluation System process.
This lead group, often a dedicated root cause analysis group, The IAG has been considered a significant strength at Rancho typically takes a lead role in the conduct ofinvestigations of more Seco and an example of a successful approach, which if followed significant events, serves as a resource to the line departments in can result in a substantialimprovement in events evaluation, par. their evaluation of events of lesser significance, and monitors (in ticularly root cause analysis,in a relatively short period. Two some instances, formally concurs in) the performance of events other licensees in Region V, Portland General Electric Com. evaluation or root cause analysis by the nuclear organization. In pany at the Trojan plant in 1986 and Washington Public Power the case of the Trojan plant, the dedicated root cause analysis Supply System (WPPSS) at WNP-2 in 1988, have followed the group reports directly to the Vice President-Nuclear and is used example of Rancho Seco in establishing dedicated events evalu. extensively by the Vice President to assess the performance of ation groups.
various groups within his organization.
At the outset of Region V's initiative, most licensees had mini. Although progress in achieving improved performance by some mal formal root cause analysis capability within their site or licensees has been slower than Region V would have preferred, corporate organizations. Over the past 2 years, increasing substantial improvements have been realized. Often, licensee emphasis has been given to the formal training of licensee management has lacked a sensitivity and an appreciation for the personnel, particularly onsite personnel, in the techniques of need of expertise in root cause analysis. In some instances, this root cause analysis. The number of plant staff trained in root appreciation of need came only through the painful experience of cause analysis techniques today typicallyranges from 50 to more failing to understand the cause of repetitive failures sufficiently to than 100, and the numbers are continuing to increase. Much of permit meaningful corrective actions to preclude further failures.
this trainir.g has been provided by contract personnel experi. Region V's experience has been that most instances of substantial enced in accident or incident investigations.
improvement in licensee programs for events evaluation and root M
l s
cause dnalysis,have come only after the attention and the com-mit' ment of senior c'orporate management were directed toward
, the'se programs. Such was the case at the Trojan plant where, after a period of such experience, the then-Vice President-Nuclear concluded that the formation of a dedicated root cause i
analysis group was necessary to achieve timely improvement. A similar decision was made this past year by the Managing Director of WPPSS with regard to WNP-2.
Regions V's continuing effort is now focusing on the need for a more consistent quality of events evaluation or root cause analysis by licensees. This need becomes increasingly important as licensees expand the population ofevents requiring root cause analysis and places additional responsibility on the line organiza-tion in the implementation of these programs.
i k
1 i
I 11