IR 05000397/1992030

Insp Rept 50-397/92-30 on 920817-21 & 27-29.Major Areas Inspected:Augmented Insp Team Review of Power Oscillation Event at Facility on 920815

ML17289A894
Person / Time
Site:	Columbia
Issue date:	09/29/1992
From:	Perkins K NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION V)
To:
Shared Package
ML17289A893	List: IR 05000397/1992030 ML17289A892
References
50-397-92-30, NUDOCS 9210070002
Download: ML17289A894 (66)

Text

~Re ort No.

Docket No.

License No.

Licensee:

U.S.

NUCLEAR REGULATORY COMMISSION REGION V

AUGMENTED INSPECTION TEAM (AIT) REPORT 50-397/92-30 50-397 NPF-21 Washington Public Power Supply System (WPPSS)

Nuclear Plant, Unit 2 Richland, Washington W Il g N

11,

I Benton County, Washington Ins ection Conducted:

August 17-21 and 27-29, 1992 Team Members:

L. F. Hiller Jr., Chief, Reactor Safety Branch, Region V,

Team Leader L.

E. Phillips, Chief, Core Performance Section, Reactor Systems Branch, NRR J.

W. Clifford, Acting Chief, Program Development Section, Operator Licensing Branch, NRR W. Ang, Senior Project Inspector, Reactor Projects Section 1,

Region V

D. L. Proulx, Resident Inspector, WNP-2, Region V

.

T.

Sundsmo, Project Inspector, Reactor Projects Section 2,

Region V

Dr. J.

March-Leuba, Consultant, Oak Ridge National Laboratory (ORNL)

WNAdb !:

.

Per ins, r.,

erector, ivision o eactor a ety and Projects, Region V

Ins ection Summar

Ins ection on Au ust 17-21 and 27-29 1992 Re ort No. 50-397 92-30 Areas Ins ected:

Augmented Inspection Team (AIT) review of a power oscillation event at WNP-2 on August 15, 1992.

During this inspection, Inspection Procedure 93800 was used.

9210070002 920929 PDR ADOCK 05000397 G

PDR

TABLE OF CONTENTS 1.

Introduction Hang ement Summar and Event Overview 1. 1 Purpose and Scope of the AIT Inspection 1.2 Inspection Methodology

.

1.3 Event Summary 1.4 Findings and Conclusions

.

1.4.1 Findings

.

1.4 '

Concl'usions

~

~

~

~

~

~

~

~

~

~

~

~

~

~

2,

2.

Narrative Descri tion of Event of Au u'st

1992

.

3.

Descri tion and Anal sis of Power Oscillations which Occurred 3. 1 Description of the Oscillations 3. 1. 1 Oscillation Amplitude 3. 1.2 Oscillation Frequency and Decay Ratio 3. 1.3 Oscillation's Effect on Fuel Integrity

.

.3.2 Stability Calculations for the August 15 Event

.

3.3 Causes for Instability Which Occurred

.

4.

Review of Res onses to Generic Corres ondence

.

~

~

4. 1 Review of Generic Correspondence 4.2 Implementation of Generic Correspondence 5.

Evaluation of Or anizational Performance 5. 1 Review of Operator and STA Performance

.

~ 1. 1 Assessment of Procedural Adequacy 5. 1.2 Evaluation of Operator Performance

,

5. 1.2. 1 STA/SNE Performance 5. 1.2.2 License'd Operator Performance 5. 1.3 Assessment of Operator and STA Training Effectiveness

.

~

~

14

17 5.2 Assessment of Engineering Performance 5.2. 1 Evaluation of Core Design

19

5.2.1.1 5.2.1.2 5.2.1.3 5.2.1.4 Core Stabi1 ity Core Nuclear and Thermal-Hydraulic Design

.

Exclusion Region Boundaries Assumptions

.

Reduced Core Flow Capability

'1

23 5.2.2 Evaluation of Core Design guality Assurance

.

.

.

.

.

5.2.2. 1 Review of Supply System Core Design Process 5.2.2.2 Evaluation of Core Design gA Effectiveness

.

25 5.3 Evaluation of Equipment Performance

.

.

.

.

.

.

.

.

.

.

.

.

.

5.3.1 5.3. 2 5.3.3 5.3.4 5.3.5 5.3.6 Recirculation Flow Control System Jet Pumps Feedwater Control System

.

.

.

.

.

Turbine Control System

.

.

.

.

"Loose Parts Detection System

.

Stability Honitor

27

27

28 5.4 Evaluation of Licensee Event Investigation

.

.

.

.

.

.

.

.

.

.

5.4. 1 Technical Analysis

.

5.4.2 Root Cause Analysis

31 5,5 Assessment of the Licensee's Emergency Notification Process

~

6.

Descri APPENDIX A tion of Flow Biased ATWS Tri Set oint Errors Notes

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

APPENDIX B Detailed Sequence of Events APPENDIX C

APPENDIX D Stability Calculations for the August 15, 1992 Event

.

Stability Calculations for the August 30, 1992 Startup

.

APPENDIX E Stability Calculations for the August 2, 1992 Startup APPENDIX F Review of Previous Cycle 8 Startup Reactivity Anomalies APPENDIX G Survey of Control Rod Sequence Practices at Other BWRs

.

APPENDIX H AIT Charter APPENDIX I Persons Contacted APPENDIX J Shift Staffing During the Event APPENDIX K Partial List of Procedures and Correspondence Reviewed

.

APPENDIX L Definitions of Core Physics Parameters

.

ENCLOSURE I INSPECTION REPORT 50-397/92-30

REPORT DETAILS l.

Introduction Mana ement Summar and Event Overview 1. 1 Pur ose and Sco e of the AIT Ins ection

.This report presents the findings of an NRC Augmented Inspection Team (AIT)

inspection of the power oscillation event which occurred on August 15, 1992 at the Washington Nuclear Plant, Unit 2 (WNP-2) facility.

The decision to dispatch an AIT was made by NRC management based on the apparent similarity of this event -to others which have occurred at boiling water reactors, most notably at the La Salle Unit 2 facility on March 9, 1988.

Also, control of power oscillation was the subject of NRC Bulletin 88-07,

"Power Oscillations in Boiling Water Reactors."

The NRC staff and the Boiling Water. Reactor Owners'roup (BWROG) have continued to study this phenomena since that event, to develop technical specifications and other guidance to avoid power oscillations.

The WNP-2 event was of particular concern because it was initiated from a region of power and flow which was outside the exclusion areas defined by the technical specifications, but which was also within the cautionary area suggested in a March 18, 1992 BWROG advisory letter,

" Implementation Guidance for Stability Interim Corrective Actions."

The AIT consisted of six NRC inspectors or engineers, and a consultant from Oak Ridge National Laboratory.

The consultant was expert in the field of power oscillations at BWRs.

The AIT Charter (Appendix H) directed that the team verify the circumstances, identify the causes of the event, and evaluate its significance.

The inspection was conducted from August 17-21 and 27-29, 1992.

An entrance meeting was held with the licensee on August 17, 1992 at WNP-2.

A public exit meeting was held at the corporate headquarters on August 29, 1992.

Appendix I provides a list of attendees at the exit meeting.

1.2 Ins ection Methodolo After an initial briefing by licensee personnel at the entrance meeting on August 17, 1992, the AIT interviewed the operating crew and Shift Technical Advisor, engineers who participated in the fuel reload analysis, and licensee managers.

The AIT reviewed the licensee's core physics information regarding the event, and made independent calculations for comparison.

The AIT also reviewed other relevant charts, logs, written statements, procedures, memoranda, and other documentation during the inspection.

Region V management was briefed daily on the progress and preliminary findings of the inspection.

At 0300:38, with the reactor at 34

% power and

% flow, on August 15, 1992, a manual scram of the reactor was initiated due to power oscillations observed by the control room crew.

The oscillations began at 0258, and were observed by the operators at 0259.

The oscillations in cor'e average power were approximately

% power in amplitude peak-to-peak.

Post-event analysis

confirmed that the oscillations were in-phase, across the core.

No fuel damage occurred, but the operating limit critical power ratio was exceeded.

The reactor tripped normally, and all systems performed as expected.-

An unusual event was declared at 0320 due to the power oscillations, and terminated at 0430.

1.4 Findin s

and Conclusions The AIT made numerous observations, findings, and conclusions which are detailed in this report.

The following findings and conclusions are considered to be the most significant ones identified:

1.4.1

~Findin s

~

There was no evidence of fuel failures or violation of fuel safety limits due to the power oscillations which occurred.

~

The primary cause of the oscillations was very skewed radial and axial power distributions in the reactor.

These were a result-of:

the control rod pattern selected for power escalation and recirculation pump shifting, and the core fuel loading configuration.

~

The large radial and axial peaking factors obtained during the event

~

exeeded the values assumed in current BWROG procedures for stability region boundaries.

These peaking factors were, therefore, not conservative with respect to the less limiting, empirical stability regions specified in the Technical Specifications.

The large peaking factors resulted in instability predictable by stability calculation codes (such as LAPUR).

1.4.2 Conclusions

~

The current reactor can be operated with a very low risk of power oscillations by following the procedures and startup plan proposed by the licensee in their letter to the NRC dated August 29, 1992.

This plan includes operation of the stability monitor when in the region greater than 25 / power and less than

/o flow. It also required the specification of stable rod patterns to be used during power operations with less than

% flow.

"

~

Neither the licensee nor the fuel vendor properly assessed.

the vulnerability of the reactor to instability when operated as permitted by the licensee's procedures, when they designed the Cycle 7 and Cycle 8 fuel reloads.

~

The.licensee did not adequately incorporate.into its procedures the March 18, 1992 BWROG advisory letter that, recommended increased instability alertness outside the TS exclusion regions.

~

Operator training was not effective in ensuring operator understanding of the latest BWROG advisory letter.

The training and qualification program for Shift Technical Advisor/Shift Nuclea Engineers (STA/SNEs) did not adequately address the potential impact of power distributions and rod patterns on reactor stability.

~

Procedural controls to specify appropriate control rod patterns or other effective stability criteria between 20 / power and the target full power rod pattern were inadequate.

2.

Narrative Descri tion of Event of Au ust

1992 On August 13, 1992 at 1655 PST, drywell unidentified leakage was identified.

The licensee commenced a shutdown to identify the leak.

Operators reduced power to

N to make an entry into the drywell to attempt to identify and isolate the leak.

At 0746 on August 14, Supply System personnel entered the drywell, found a leaking valve, backseated it, and stopped the leak.

Subsequently, at 1710, the Reactor Operator (RO)

commenced rod withdrawal to resume power operations.

At 2109, with reactor at

N power, the main generator was synchronized -to the grid.

At 2228, the mid-shift crew relieved the watch.

At 2320, the reactor was at 25 / power, and operators continued to increase power with control rods.

The stability monitor (the Advanced Neutron Noise Analysis, or "ANNA" system)

was not in operation.

The procedures used by the crew during the shift on which the event occurred are listed in Appendix K (Procedures 1-4).

A summary of the shift staffing during the event is provided in Appendix J.

During the reactor startup, two operational constraints limited the reactor power and flow 'conditions where flux shaping and recirculation pump shifting from slow speed to fast speed could be performed.

First, reactor feed flow of approximately 4.5 to 5 million pounds per hour and reactor power greater than

N was required by procedure PPH 3. 1.2,

"Reactor Cold Start-up," prior to the pump shift in order to prevent cavitation in the recirculation system.

Second, the length of time that the recirculation pumps were operated in fast speed with less than 50 / flow needed to be minimized, because excessive pump vibration occurred under those conditions.

To minimize this time period, control rod patterns were adjusted prior to the pump shifts to obtain a

desirable flux shape and reduce the number of fully inserted control rods.'hese adjustments minimized the amount of rod movements after the pump shifts, and would have allowed efficient control rod withdrawal / power ascension within the constraints of the fuel preconditioning limits.

Fuel preconditioning limits could have restricted control rod withdrawal if power density near the control rods exceeded specific limits.

Usually, fully inserted control rods (i.e., less than position 12) are the most limiting. If preconditioning limits had been approached, full power operation would have been delayed while several power changes (via recirculation flow control) were performed to produce power and xenon distributions that permitted control rod withdrawal.

CROA'I, under supervision of the STA/SNE, was adjusting control rods from 1830 (August 14, 1992) until about 0245.

This period was used to adjust the reactor's neutron flux profile, adjust the timing of six control rods, and conduct'

surveillance to exercise a control rod.

Reactor power and total

core flow were maintained at about

% and 30 %, respectively, during these adjustments.

The STA then informed the CRS and SH that the control rod

'djustments were complete.

CRO¹1 closed the "A" recirculation loop flow control valve (FCV) in preparation for shifting the "A" recirculation pump to fast speed.

As the FCV was closed, reactor power decreased from 36.4

% to 33.5

%,

and total core flow decreased from 30.5

% to 26.0 Reactor power oscillations started as the FCV was closing, at about 0258: 18.

Average Power Range Heter (APRH) -oscillations were initially observed by the operators at about 0259:49, and were terminated by a manual scram that was activated at 0300:38.

The SH, CRS, CRO¹l, CRO¹3, and the STA were present in the control room.

CRO¹1 initially identified the power oscillations when he observed the APRMs swinging about

% peak to peak power.

CRO¹1 alerted the CRS and other crew members.

The crew then observed multiple Local Power Range Heter (LPRH) downscale indications, and continued power oscillations on the APRMs.

The CRS recommended initiating a manual reactor scram to the SM, who then directed the scram.

The manual scram effectively terminated the reactor power oscillations.

Following the manual scram, reactor water level went below Level 3 (13 inches)

for about 30 seconds, causing the crew to momentarily enter emergency operating procedure (EOP)

PPM 5. 1. 1,,"RPV Control (Non-ATWS)."

The lowest reactor water level reached was -15 inches; it was automatically restored by the main feedwater pumps without requiring operator action.

CRO¹2 was called into the control room (from the control room back panels)

and assisted with the post scram actions.

The crew continued plant shut down, without complications, using procedure PPM 3.3. 1,

"Scram Recovery."

r

'Plant computer data showed that the power oscillations were core wide and were in phase.

The AIT concluded that power oscillations lasted for 144 seconds from initial onset until the reactor scram.

Operators scrammed the reactor

seconds following the first annunciator received (APRH Flow biased rod block).

Licensee calculations indicated a Hinimum Critical Power Ratio (MCPR) of 1.68 during the event.

The safety limit '(minimum allowed value)

HCPR of 1.07 was not exceeded.,

A reactor coolant sample confirmed that no fuel damage had occurred.

All other plant systems responded as expected.

3.

Descri tion and Anal sis of Power Oscillations which Occurred This section describes the power oscillations that occurred on August 15, 1992, in the WNP-2 plant, and the analyses performed by the AIT to identify the root causes for these oscillations.

The subject of boiling water reactor thermal hydraulic stability has been of interest to designers, operators, and regulators since the early days of BWR design.

Much theoretical, experimental, and operational information has accumulated on the subject.

BWR stability is influenced by several power distribution and operating state variables that change during normal operation and from cycle to cycle.

In particular, the radial and axial power distributions and core inlet subcooling have a strong impact on stabilit. 1 Descri tion of the Oscillations The August 15, 1992 WNP-2 power oscillations exhibited the characteristics o'

a density-wave instability of the corewide type (also called in-phase or fundamental mode).

Of the four types of instabilities that have been observed in HWRs (corewide, out-of-phase, single channel, and control-system-induced instabilities),

the corewide type of power oscillation is the type least likely to result in a significant challenge to the fuel because, under most reasonable operating conditions, the high APRM automatic scram will take effect before any thermal limits are violated.

Figure 1 shows a time trace for LPRH 32-17C during the event.

On this trace, the oscillations started approximately at 02:58:45 and grew for about one minute with a decay ratio of approximately 1.06 until the amplitude saturated to a peak-to-peak value that is approximately

/o of the average local power.

The oscillation frequency was approximately 0.5 Hz (2 second period).

The decay ratio (DR) is a measure of the relative stability of the reactor; DR values less than 1.0 indicate stable operation, while DRs greater than 1.0 indicate an instability.

3. 1. 1 Oscillation Am litude A characteristic of corewide type instabilities is that the oscillation amplitude is proportional to the average value of the local power at each core location (i.e.,

Local Power Range Monitor (LPRM) readings).

In other words, for corewide oscillations, the oscillation amplitude should not be the same in all LPRH signals',

but it should be proportional to the average LPRH reading.

On first evaluation of the data, this appeared not to be the case because LPRHs32-17C and 32-09A had an oscillation amplitude several times larger than all other signals recorded.

Later detailed evaluations indicated that all signals (from the process computer)

except LPRHs32-17C and 32-09A had been conditioned by a 0.3 Hz low-pass filter that reduced the apparent oscillation amplitude by a factor of four.

For instance, the filtered Average Power Range Monitor (APRH) data indic'ated an apparent oscillation peak-to-peak amplitude of only 6 / of core rated power, while other unfiltered recorded data (from the Transient Data Acquisition System (TDAS)) showed that the APRg had oscillated between 22.66

% and 48.91

/o of nominal power, indicating that the APRH had a

peak-to-peak oscillation of at least 26 / of core rated power.

Since the core average power during the event was 33.7

/o of core rated power, the APRH peak-to-peak oscillation measured as a percent of the actual average power during the event was 77 / (i.e., 26/.337),

which is consistent with the observed 80 / relative oscillations in all LPRHs once the effect of the 0.3 Hz low-pass filter was corrected.

3. 1.2 Oscillation Fre uenc and Deca Ratio From, the APRH recorder the power oscillations started at 02:58: 18 and grew for approximately one minute until they saturated.

The estimated decay ratio during the oscillation growth period was 1.06, and the

LPRM3217C I

I I

20

O

0 n5.

!i7:5S 15.

30.

05.

15.

30, n5.

15.

30.

l5.

15 ii(psst 1992 Figure'1.

Time trace of LPRM 32-17-C during the 8/15 event.

This LPRM was not filtered by the D.3 Hz low-pass ltltc oscillation frequency was 0.5 Hz.'Note that superscripts refer to Appendix A, "Notes.")

At 0300:00

, the oscillations seemed to have reached a stable limit cycle (i.e.,

decay ratio of 1.0),

and their amplitude was not growing significantly.

A slight trend towards increased amplitude, however, can be observed in Figure 1 even at the moment of scram.

This small increase is attributed to nonstationary reactor conditions, such as nonequilibrium feedwater temperature and nonequilibrium xenon.

3. 1.3 Oscillation's Effect on Fuel Inte rit There was no evidence of fuel failure or violation of fuel safety limits due to the August 15 power oscillations.

However, the licensee's calculations showed that the thermal-margin operating limit (OLHCPR) was exceeded.

The change in critical power ratio (CPR) during the oscillations was evaluated by the licensee using the VIPRE transient code (under review by the NRC),

and by Siemens Nuclear Power (SNP) Corporation using the XCOBRA-T code.'he results of both calculations indicate that the change in CPR was relatively small and not sufficient to cause a safety limit minimum CPR (SLHCPR) violation.

A calculation was performed by SNP using a bounding analyses according to the licensed methodology (COTRANSA 2).

This analysis resulted in a CPR change of 0.27 and a

HCPR of 1.68, well above the SLHCPR of 1.07, but exceeding the Operating Limit minimum CPR of 1.795.

The above results are consistent with previous experience and calculations, which indicated that safety limits were not likely to be violated for relatively small-amplitude power oscillations such as were observed at WNP-2.

3.2 Stabilit Calculations for the Au ust 15 Event Using the best estimates available for power, flow, power distribution, and the operating conditions at the time of the oscillations, the AIT calculated a

corewide DR of 1.05, a hot-channel DR of 0.83, and an out-of-phase DR of 1.0.

These results indicate that:

(2)

The hot channel was probably stable, but not with much margin (DR 0.83).

A single channel thermohydraulic instability was not likely, but could not be ruled out without further analyses because of the extreme radial power peaking existing in this event.

Even though the event data shows that the instability was clearly of the in-phase or fundamental mode, the calculations indicated that the out-of-phase mode was also fairly unstable.

From these results, it was not certai'n which of the two modes would likely dominate, Therefore, other

startups with these skewed power distributions could result in out-of-phase oscillations.

Oak Ridge National Laboratory (ORNL) performed some sensitivity analyses for the AIT to identify the root cause of this instability.

As stated before, the main root cause was the extreme radial and.axial power distr'ibution.

The other contributing parameter was the mixed-core characteristics present in WNP-2 at the time of the event.

Appendices C,

D, and E provide a more complete discussion of the stability calculations performed by ORNL for the AIT.

3.3 Causes for Instabilit which Occurred The AIT concluded that the main cause of this instability event was the very skewed radial and axial power distribution (1.92 radial peaking factor and up to 1.76 axial peaking factor).

This same core had been started on two previous occasions (July 27, 1992, and August 2, 1992) without oscillations, even though the recirculation pump upshift was performed at higher rod lines on the previous startups.

The power distribution during the August 15 startup was caused by the control rod pattern selected, which included all shaper rods withdrawn and four primary power rods located in the core-center region that were withdrawn 28 notches, resulting in a high power area in the core-center region.

When a more conservative control rod pattern is used for the pump upshift, the decay ratio for this core can be as low as 0.3 (see Appendix D),

compared to a decay ratio of 1.05 with the rod pattern selected on August 15.

The AIT also found by analyses that a contributor to the instability of Cycle 8 in WNP-2'was a mixed core with unbalanced flow characteristics between the new SNP 9x9-9X fuel and the old SxS assemblies.

Under these conditions, the low-power and low resistanc'e SxS bundles were starving-the flow from the high-power and high friction 9x9-9X bundles; this effect can be observed in Figure 2, which shows the relative power and flow of all the channels in the core at the time of the event.

LAPUR -calculations (Appendix D) indicated that if the whole core had been loaded with 8xS fuel, the decay ratio would have been 20 I lower and the instability would have been avoided, even with the power distributions which were in use on August 15.

Noticeably, if the whole core had been loaded with-9x9-9X instead of being a mixed core, the decay ratio would be lower by 10

% and the instability may have also been avoided.

AIT LAPUR calculations indicated that the hot channel was thermohydraulically stable during the August 15 event.

AIT LAPUR calculations also indi,cated that the out-of-phase mode of instability did'not have much margin to instability.

If the LAPUR calculations had been performed before the event, they would have shown that the instability could have been either in-phase or out-of-phase with almost equal probabilit UO

-

August 15, 1992 at 3:00 AM, 14 Sec.

BX9(s)

50

Power = 33.7~ Flow = 26.7

BX8 (Gj

rz LIg

~I~+

t~~

tf o

t ~o f

20

0

10

20 Flow (% Rated)

40

50 Figure 2.

Rclativc power and flow of all channels just before the 8/15 WNP-2 event.

The reduced flow area of thc 9x9-9X tuel compared to 8x8 results in a lower flo.

Review of Res onses to Generic Corres ondence This section of the report evaluates the licensee's responses to generic correspondence concerning core power oscillations.

4. 1 Review of Generic Corres ondence Since the core power oscillation event at LaSalle in 1988, the NRC and BWROG have developed guidance intended to reduce the likelihood of a core power oscillation event, and to mitigate an event should core power oscil.lations occur.

The NRC issued NRC Bulletin (NRCB) 88-07 dated June 15, 1988 and NRCB 88-07 Supplement 1 dated= December 30, 1988.

The bulletin established actions r'equired in response to the core power instability event at LaSalle, including operator and STA training on recognition, prevention, and mitigation of uncontrolled power oscillations.

The bulletin also called for a verification of instrument adequacy; The supplement provided additional actions.

It stated that the specific power to flow boundaries for plants using non-GE supplied fuel should be based on existing boundaries that had been previously approved by the NRC.

For new fuel designs, it stated that the stability boundaries should be reevaluated and justified based on calculated changes in core decay ratio using NRC approved methodology, decay ratio measurements, or applicable operating experience.

The BWROG has issued several relevant letters:

BWROG-8847 dated July 8, 1988, BWROG-8879 dated November 3, 1988, a revision to BWROG-8879 dated November 4, 1988, and BWROG-92030 dated March 18, 1992.

The first letter (BWROG-8847)

documented the BWR Owners'roup understanding of a June 24, 1988 meeting with the NRC, including information to be developed by the BWROG to address the types of issues identified in NRC Bulletin 88-07.

No specific actions were identified in this letter.'he second letter (BWROG-8879) provided interim recommendations, developed by GE to address potential instabilities in BWRs, that the BWROG considered prudent interim actions whil,e the results of ongoing BWROG Stability Programs were evaluated.

The BWROG requested an expedited review and decision on the part of all BWR Owners to the recommendations in the letter.

The actions were grouped according to whether BWR 4 plants had filtered (Group 1) or.unfiltered

.

(Group 2)

APRM signals.

BWR 5 and 6 plants, including WNP-2, were grouped with the filtered APRM plants (Group 2).

Group 2 plants were to immediately scram the reactor to exit Region A (greater than 100

/o rod line with less than 40 / rated core flow).

A third letter (BWROG-92030) modified the second letter (BWROG-8879) to make it clear that the plant operators were to scram the reactor if any thermal hydraulic instability occurred while in specified regions of the power to flow map.

The latest BWROG letter dated March 18, 1992 provided additional recommendations that the BWROG felt were necessary to enhance the effectiveness of the interim corrective actions (from BWROG-8879).

The BWROG

provided this information based on additional analyses and additional operating experience, including an oscillation event at a non-U.S. reactor that initiated outside the recognized instability region.

The letter stated that "the guidance should be carefully considered by all owners" for application to their procedures and training programs.

The recommendations also were based on an increased understanding of the conditions that might result in oscillations.

Based on these considerations, the BWROG highlighted the need for caution when operating near the stability exclusion region.

The additional cautions included:

1) emphasizing operator training to scram the reactor even if oscillations were observed but were less than 10 %,

2) enhanced guidance for recognition of oscillations, including increase or periodicity in noise levels when near the instability region (defined as approximately

% power or 5

% flow from the current potential instability region),

3)

a caution that the stability exclusion region boundaries were not exact, especially at lower power and flow where the uncertainties in measuring power and flow can increase, along with an admonition that it was best to minimize the amount of time spent operating near the stability exclusion region, 4) the importance of the effect of reduced feedwater temperature on core stability, and 5) the strong impact of radial and axial power distributions and core inlet subcooling on stability, as well as the concern that fuel and reload designs have resulted in higher fuel bundle power levels, making oscillations more likely.

4.2 Im lementation of Generic Corres ondence The AIT reviewed the licensee's procedures referenced in this report, and correspondence listed in Items 6-9 of Appendix K, and conducted interviews to evaluate the licensee's implementation of generic information related to prevention and mitigation of core power oscillations.

Based on review of these documents, interviews with plant and management personnel, and evaluation of procedures and training,,the AIT found that the licensee had implemented some of the information available on core power instabilities.

The licensee's actions included the following:

PPH 4.12.4.7,

"Power Plant Maneuvering,"

was developed in response to NRCB 88-07 through the utility's formal procedure review process.

This procedure was subsequently modified to incorporate the guidance of the BWROG November 3, 1988 letter (as modified by the November 4, 1988 BWROG letter)

and NRCB 88-07 Supplement 1, using the formal procedure review proces PPH 3. 1.3, "Plant Startup,"

and PPH 4. 12.4.7,

"Unintentional Entry Into Region of Potential Core Power Instabilities," incorporated the guidance contained in the NRCBs, and the BWROG letters through November 4, 1988 regarding recognition of core power instabilities, and mitigative actions if operating in a region of potential core instability.

~

-

Lesson Guide 82-SgT-9202-L2;

"Reactivity Hismanagement Events,"

incorporated the guidance contained in all the related NRCBs and BWROG letters through the Harch 18, 1992 BWROG letter.

In addition, the lesson plan used several operating events to provide examples of what could occur that could lead to reactivity mismanagement, and the appropriate actions to take in each case.

However, the AIT found that the licensee had not fully implemented some other important generic guidance.

The licensee's procedure for review of generic information was PPH 1. 10.4,

"External Operational Experience Review."

This procedure required a documented review by all potentially affected departments to determine whether or not procedural changes or other action was appropriate.

Information from the BWROG was not required to be reviewed under this procedure.

The information was, however, reviewed using a routine distibution of information list.

This distribution list was used to provide information for documents not specifically covered by PPH 1. 10.4.

In the case of the Harch 18, 1992 BWROG letter, the licensee's representative on the Owners'roup sent the letter to personnel on the distribution list for BWROG information related to stability.

This list included:

Hanager, Engineer Services; Hanager, Nuclear Engineering; Principal Engineer, Reactor Engineering; former Supervisor, Shift Engineering; and STA Trainer.

These personnel then decided if additional action was required on the information provided through the distribution system.

There was not any documentation of the decisions which they made to distribute this information.

A licensee representative stated that none of the personnel on distribution considered that any prompt actions on the information were necessary.

The STA Trainer did, however, incorporate the information into a lesson plan on power instability experience which he was updating.

As a result of the failure to distribute this information throughout the organization, several errors occurred.

These deficiencies were:

The licensee's operating procedures did not contain the guidance from the Harch 18, 1992 BWROG letter discussed in Section 4. 1 of this report.

The SNE/STA stated that he conducted rod time testing and as many rod pulls as possible before shifting recirculation pumps to high speed, staying within a few percent of the 80 / rod line during these evolutions, contrary to the BWROG guidance.

He further stated that this was normal practice, and no recent information had changed the pr'actice.

The Operations Procedure Supervisor was not aware of the Harch 18, 1992 BWROG letter, or its associated guidance, before the event.

The Operations Procedure Supervisor is directly responsible to the

Operations Supervisor for evaluation of operating experience for incorporation into station procedures.

~

The Plant Manager was not aware of the March 18, 1992 BWROG letter, or its associated guidance, before the event.

~

Licensee personnel did not incorporate consideration of the filters on the APRHs and stability monitor, as directed in the BWROG letter dated November 3, 1988, and in NRCB 88-07 Supplement 1, into their consideration of APRM response to core power instabilities, or of the impact on the stability monitor.

Licensee personnel also did not consider the effect of the filters in their event analysis until the effect of the filters was identified by the AIT.

Plant Operating and Nuclear Engineering Procedures were not reviewed under the licensee's formal program for review of operating experience to ensure guidance and directions to operators were consistent with the latest generic information contained in the BWROG letter dated March 18, 1992.

The AIT did note that the information from the BWROG letters of November 3 and 4, 1988, had been evaluated and incorporated in plant procedures under the licensee's previous program for formal review of external operating experience.

The licensed operators and SNE/STAs were trained on the.information contained in the BWROG March 18, 1992 letter, but the information was not provided to the Operations Department for evaluation for incorporation into procedures.

The AIT found that the licensee's program for control of generic information had broken down: it permitted the procedures and training to conflict.

Training was conducted on operating strategies that had not been coordinated with the Operations Department.

The strategies were not the method by which the management expected the plant to be operated.'.

Evaluation of Or anizational Performance 5. 1 Review of 0 erator and STA Performance 5. 1. 1 Assessment of Procedural Ade uac PPH 3. 1.3,

"Reactor Startup from Hot Shutdown," provided guidance for conducting a plant startup from a hot shutdown condition.

Once at power, PPH 9.3. 12,

"Power Plant Maneuvering," provided guidance to the SNE/STA and operators while changing power. using control rods and recirculation pumps.

In addition, PPM 2.2. 1, "Reactor Recirculation

,

System,"

provided detailed instructions for conducting a shift of.

recirculation pumps from slow to fast speed, PPH 4. 12.4.7,

"Unintentional Entry Into Region of Potential Core Power Instabilities" provided actions for the operator upon entry into areas of the power to flow map identified as core instability regions, or the onset of actual core power osci-llations.

PPH 9.3. 12,

"Power Plant Maneuvering," contained guidance for operating

in specified regions of potential core instability on the power to flow map, including use of the stability monitor for operating in Region C.

No specific guidance was provided, however, for operating in areas where

. instabilities were not expected.

Based on a review of these procedures, analysis of available data from the plant computers, interviews with personnel, and observation of a, Plant Oversight Committee (POC) meeting, the AIT determined that the procedures did provid several cautions which, if followed, would have reduced the probability that this event would have occurred.

The specific guidance was:

~

PPH 9.3. 12;

"Power Plant Haneuvering,"

also provided general principles and objectives that included:

1) flatten radial peaking, 2) maintaining a strong bottom peak, but having too large a bottom peak would prevent opening flow control valves after pump shift to fast speed, 3) not being overly aggressive when pulling shaper rods, with a peaking factor of 3.4 as an optimum value, considering recirculation pump vibration and fuel preconditioning limits, and 4) minimizing the amount of time spent at slow speed pumps when performing a rod set following a power reduction, due to concerns for xenon burnup after power increase.

As discussed in Section 5. 1.2, the operating crew did not use this guidance.

In addition, the AIT noted deficiencies with th'e following procedures:

PPH 9.3. 12,

"Power Plant Maneuvering," provided guidance that allowed deviation from the Control Rod Withdrawal Sequence once above approximately 20 / power.

This was provided based on margins between operating fuel rod powers and fuel preconditioning considerations.

The procedure stated that deviations were considered necessary to account for power distribution and stability'onstraints, and to account for various core reactivity conditions and xenon inventories.

Interviews with plant personnel demonstrated that deviation from the rod pattern above the low power setpoint (approximately 20 I. power),

was a

common practice at WNP-2,'ith rod pulls based on calculations by the STA/SNE.

The AIT was concerned because this practice resulted in large variations in stability. during several startups, due to.

excessively peaked power distributions caused by the rod patterns selected by the STA/SNE.

PPM 9.3. 12,

"Power Plant Maneuvering," provided guidance to.

complete most rod motions prior to shifting recirculation pumps to fast speed.

Additional guidance was provided'in this procedure to conduct rod withdrawal to increase power at the left side of the power to flow map.

This required operating at close to the 80

%%u rod line for several hours while rod pulls were conducted, which was contrary to the guidance provided by the March 18, 1992 BWROG letter to avoid extended operation close to this are Licensee procedure, PPH 2. 1.8,

"ANNA System Operation," required a

scram if ANNA indicates that APRH oscillations,peak to peak were greater than 10 /,

and if APRH power range recorders indicated oscillations

> 10

/o.

BWROG advisory letter 92030 dated March 18, 1992 stated that "it may be possible to violate the MCPR safety limit during a regional oscillation while the APRH's are oscillating less than

%%d.

A regional oscillation...

may not always result in LPRH high or downscale alarms."

When discussing training, the BWROG letter stated "this training should emphasize that a scram is. required even if the magnitude is below

%%d on the APRHs and LPRH upscale and downscale alarms have not occurred."

The letter also observed that oscillations may be occurring if APRH noise levels are two to three times normal peak to peak.

- Licensee personnel stated that, in the past, peak to peak noise has been about

/o when the plant-was at approximately

N power.

The licensee's procedures did not incorporate this 'guidance

.

PPM 9.3.9,

"Control Rod Withdrawal Sequence Development and Control" required calculated power distributions to be attach'ed to each control rod order sheet.

Paragraph 7.4.2 stated, in part:

Actual operating conditions, exposure histories relative to'he target or preconditioning strategies may dictate using a

different pattern than provided.

With each rod sequence sheet, include as an attachment the target rod pattern, calculated power distributions and the target power distribution for the sequence review.

However, the STA/SNE did not determine the projected power distributions for each rod sequence sheet as required by this procedure.

Licensee representatives stated that the procedure was incorrect, and that they had not intended to perform these power distribution calculations.

Although PPM 9.3. 12 and PPH 9;3.9 provided qualitative guidance for controlling neutron flux peaking, the procedures did not define adjectives used to prescribe guidance.

Examples of these guidance adjectives that did not have quantitative limits or examples included terms such as excessive, too large, optimal, overly aggressive and fierce.

The lack of quantitative, and specific qualitative requirements on reactor power distribution parameters permitted the STA/SNE to implement an unstable control rod withdrawal pattern that generated the very high peaking

.

factors described in Table l.

Although Technical =-Specification thermal limits provided quantitative power limits, these limits were based on.full power operation, and did not effectively limit axial and radial peaking factors sufficiently to ensure stability during startu ~

Control rod sequencing procedures at other BWR sites were reviewed to determine if the practices at WNP-2 were typical.

The. AIT concluded, based on this review, that the practices at WNP-2 were typical.'his may indicate that some other BWR licensees have a

similar vulnerability to instability when, very skewed power distributions are used.

5. 1.2 Evaluation of 0 erator and STA SNE Performance 5.1.2.1 STA SNE Performance During the reactor startup, the STA/SNE monitored the reactor power profile and control rod withdrawal sequence to ensure that fuel limits were not exceeded, and that the flux profile would allow power ascension with minimal rod motion after pump shifting.

He authorized several changes to the control rod withdrawal sequence specified by the approved Control Rod Withdrawal Order Sheets between 23:55:29 and 00:53:44.

These changes preferentially withdrew control rods in the center of the reactor, while not,withdrawing rods further out in the core.

This caused an increase in radial peaking, which is represented in Table 1 by the increase in CHFLCPR between points 1 and 2.

The Core Maximum Peaking Factor (CMPF)

and Core Maximum Fraction of Limiting Critical Power Ratio (CHFLCPR) parameters in Table

provide a measure of axial and radial neutron flux peaking.

(These terms are defined in Appendix L.)

TABLE 1:

Core Physics Parameters vs Time POINT TIME (HOURS)

23.55.29 1137 3.221 POWER (HWth)

CHPF CHFLCPR 0.711 00'53'44 01'05'44 01'37:30 1231 1221 1212 3.274 3.307 3.500 0.924 0.939 0.933 01'53:29 02:33:14 02'57'14 1211 1153 3.560 3.696 1207 3.782 0.933 0.884 0.959 03'00'14 1119 3.859 0.922

• Recirculation loop "A" FCV closed prior to this dat The control rod pattern during this event was essentially held constant from 0105 until the event occurred; however, neutron flux peaking factors continued to change.

Data at points 3, 4, 5, 7, and 8 of Table 1 were all taken from the same control rod pattern.

The increase in CHPF over this time period was caused by xenon burnout resulting from the power increase that occurred during the previous seven hours.

Xenon burnout acted to increase neutron flux peaking because it was burnt out of the core areas having higher power faster than areas having low power.

The crew did not consider the increasing peaking factors to be a concern, and consequently, took no actions to limit them.

CMFLCPR is well correlated with the radial peaking factor; as it increased prior to the event, the radial peaking factor also increased.

Points 3 to 7 of Table

show how CHPF and CHFLCPR increased prior to the event.

The increases in CHPF show that axial peaking increased steadily between 0105 and 0258, the time of the event.

The data in Table 1 was taken from computer calculations (MON runs) that were used by the crew to monitor control rod positions and reactor power: distribution during the power increase prior to the event.

The AIT evaluated the guidance that was provided to the STA/SNE to control the reactor pbwer distribution during startup (see also Section 5. 1;1).

The "Power Distribution Constraints" section of procedure PPM 9.3. 12, stated that:

Establishing too large of a bottom peak will prevent opening

.

flow control valves (FCVs) to 20,000 and 24,000 GPH respectively following the shift to 60 Hz due to preconditioning or linear heat generation rate (LHGR)

limits.

Therefore, do not be overly aggressive when pulling shaper rods, especially on the first of two ramps.

A (Core Maximum Peaking Factor) of approximately 3.4 usually results in the optimum rod pattern to meet both constraints during a

xenon free rod set.

Even though the basis of this guidance was not core stability, controlling CHPF at a value of 3.4 would have reduced the core's neutron flux peaking factors, and improved stability.

During the control rod positioning that was performed from 23:55:29 to 02:57: 14, the crew did not attempt to limit CMPF to 3.4.

The STA/SNE stated that he did not consider the CHPF values (from 3.22 to 3.78) during this time period to be excessive; he considered a

higher CHPF to be better because it enhanced fuel conditioning and xenon buildup in the bottom of the core.

The STA/SNE's strategy was endorsed by PPH 9.3. 12, which stated that a strong bottom peak power distribution was desIrable during startup, However, the guidance in PPH 9.3. 12 to maintain a strong bottom peaked power distribution during startup was tempered by other

guidance within the procedure that warned of problems caused by excessive peaking.

Section 7.1.5'.e,

"Xenon Constraints,"

specifically identified that prolonged operation with slow speed (15 hertz) recirculation pumps would lead to high peaking factors induced by xenon burnout.

This paragraph stated, in part:

The more time spent at low power, the more fierce the xenon burn following power increase.

Normally, a rod set can be performed in approximately 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />..... If more time than this is spent at low power, the resulting xenon burn may make it difficult to meet peaking factor and RRC (Reactor Recirculation)

pump operation constraints.

Emphasis was also provided by section 7. 1.3.c,

"Minimize Excessive Neutron Flux Peaking."

This section called attention to the Technical Specification requirement to adjust APRM trip setpoints when the total peaking factor exceeded rated values.

This par'agraph stated, in part:

If the core total peaking factor increases above its design

'alue, then per Technical Specifications (3/4.2.2),

the APRHs must be adjusted to make the trip setpoint more conservative..... This may result in an inability to attain the rated load line before a rod block is received.

Consequently, care should be taken to prevent the occurrence of excessive peaking.

In summary, the AIT concluded that the STA/SNE gave recommendations to the licensed operators which caused the peaking factors prior to this 'event to be excessive from a stability perspective.

5. 1.2.2 Licensed '0 erator Performance The AIT interviewed the five licensed operators, and the STA/SNE, that were on shift at the time of the event:

These interviews provided information that was used to confirm data captured by the Transient Data Acquisition System (TDAS), and also identified the following points:

The Shift Nuclear Engineers (SNE) were highly respected by the operators, who felt that their qualifications were outstanding, The operators felt that controlling the reactor's power distribution profile was solely the-'SNE's responsibility.

The operators ensured that Technical Specifications were not violated, but generally viewed the SNE's authority in determining flux shape as final.

Norie of the operators could recall the classroom training they had received in June, 1992, on the Boiling Water

Reactor Owners'roup (BWROG) letter,

"Implementation Guidance for Stability Interim Corrective Actions," dated March 18, 1992 (see Section 5. 1.3).

- The operators considered the area of core instability to be well defined.

That is, they felt there was no possibility of the reactor becoming unstable so long as it was operated outside this area (see Section 5. 1.3).

The operators consistently described the post-scram events as an easy, smooth, and uncomplicated reactor shut down.

The inspectors found that there were no administrative controls requiring peer review of SNE changes to prescribed-.control rod order sheets.

Although the CRS had been formally designated as the Reactivity Manager, his duties were broadly defined.

PPH 1.3. 1,

"Conduct of Operations,"

paragraph 5.4.20 describes these responsibilities as:

Responsibilities include ensuring a conservative approach to operations involving core reactivity changes.

The AIT concluded that the rapport between the STA/SNE and the operators, the lack of procedural requirements, and particularly the lack of specific requirements regarding startup flux shaping, explained the absence of operator review of changes by the STA/SNE to the prescribed control rod withdrawal sequence.

The AIT concluded from a review of the sequence of events and interviews of the operators that the actions taken by the control room crew to terminate this event were prompt, effective, and complied with the licensee's procedures.

Prior to the event, the control rod sequence used by the crew caused very high neutron flux peaking, and did not conform to guidance provided by the licensee's procedures.

Procedural guidance and requirements regarding power distribution during reactor startup were non-

-specific, and relied upon the SNE to interpret broad, qualitative guidance.

5. 1.3 0 erator and STA Trainin Effectiveness The AIT evaluated the effectiveness of operator training related to information, procedures, and operations affecting the event.

The AIT reviewed operating procedures, training attendance records, lesson plans, examinations, interviews, and observation of a video taped.

session of a classroom lecture.

The operating procedures reviewed are listed in Section 5. 1. 1, 5. 1.2 and Appendix K of this report.

The lesson plan reviewed was Lesson Code 82-SOT-9202-L2,

"Reactivity'ismanagement Events," dated Hay 27, 1992, in the STA Continuing Training progra Based on this information, the AIT found:

The training department used the information available in the generic correspondence described in section 4. 1 of this report to develop lesson plans and train operators on relevant operating experience.

The lesson plan approved on Hay 27, 1992 incorporated information from the Harch F18, 1992 BWROG letter.

The training session from a June 2,

1992 requalification class which the AIT observed on video tape provided very good coverage.

of the information from the Harch 18, 1992 BWROG letter.

All licensed operators on shift at the time of the event had attended requalification training that included this training on instability.

All operators took and passed written examinations that 'included the instability topic.

However, the AIT found that none of these licensed operators could recall having attended training that addressed the Harch 18, 1992 BWROG letter concerns regarding core instabilities, or that there was new guidance from the BWROG.

None of the operators recalled any of the specific guidance from the BWROG Harch 18, 1992 letter.

The training conducted was not based on changes to procedures.

The guidance was not sent to anyone in the Operations Department, and no, procedure changes were evaluated or initiated based on the Harch 18, 1992 BWROG letter.

The guidance from the Harch 18, 1992 BWROG letter was incorporated into SNE/STA and,licensed operator training because the STA trainer decided to include it in the reactivity mismanagement lecture.

While the initiative demonstrated by the STA trainer was commendable, the AIT found that the lack of a formal review process for this important guidance was a serious weakness.

The AIT was also concerned with the potential for confusion by the shift staff, since training covered operating concerns that were not incorporated into plant operating procedures.

The AIT determined that con'fusion did not occur, since the shift staff did not recall the information provided in the training lectures covering the guidance in the Harch 18, 1992 BWROG letter.

The SNE/STA on shift at the time of the event had attended prior requalification training on instability events that did not include information from the Harch 18, 1992 BWROG letter.

The licensee used a systems approach to training to develop topics and lesson plans.

This approach was procedures-based, meaning that task lists were developed from procedures, and job-task analyses were performed to determine the knowledge, skills, and abilities necessary to complete the tasks.

Lesson plans were then developed to incorporate the job-tasks into 'an appropriate

training setting.

This process was not followed for the May 27, 1992 lesson plan discussed above.

The AIT concluded that the training conducted on potential core instabilities was ineffective:

the SNE/STA on shift did not attend training on the latest information available from the BWROG, and none of the personnel on shift at the time of the event who were trained on the BWROG guidance could recall either having received the training, or the details contained in the lesson, plan.

5.2 Assessment of En ineerin Performance 5.2. 1 Evaluation of Core Desi n

The WNP-2 Cycle 8 core loading consisted of 76 fresh SNP 9x9-9X fuel assemblies, 112 once burned SNP 9x9-9X assemblies, 8 once burned SNP 8x8 assemblies, 4 twice burned ABB Atom (SVEA-96) lead fuel assemblies,

twice burned GE-11 (9x9) lead fuel assemblies, 144 twice burned SNP 8x8

'ssemblies, 4 three times burned SNP 9x9 lead fuel assemblies, 132 three times burned SNP 8x8 assemblies, 152 four times burned SNP 8x8 assemblies, 120 five times burned SNP 8x8 assemblies, and 8 six times burned SNP 8x8 assemblies.

This multiplicity of fuel types and power histories was somewhat unusual.

Licensee personnel stated that it related to the Supply System's preference for annual refueling outages (with small fresh fuel batches)

and a continuing evaluation of different fuel designs and fuel suppliers for optimum fuel economy.

The Cycle 8 core loading and fuel design were selected by the Supply System to maximize end of cycle (EOC) core reactivity.

The projected control rod patterns and licensing evaluations of shutdown margin, hot excess reactivity, and thermal margins were performed by Siemens Nuclear Fuel (SNP) Corporation, and are presented in the Cycle 8 Fuel Cycle Design Report.

Results of the SNP evaluation of system transient events

- are presented in the

"WNP-2 CYCLE 8 PLANT TRANSIENT ANALYSIS REPORT,"

ENF-92-039.

The analyses were performed using analytical methods which have been reviewed and approved for generic applications by the NRC staff.

The Cycle 8 safety analyses were performed under provisions of

CFR 50.59 and were not submitted for review by the NRC staff.

The AIT performed an audit of the Cycle 8 safety evaluation and identified the following issues

'discussed below:

5.2.1.1 ~bill The Cycle 8 core stability evaluation by SNP consisted of three decay ratio calculations using their single channel time domain code COTRAN.

TABLE 2:

DECAY RATIO CALCULATIONS PERFORHED BY SNP

~/F1 i

fl Decay Ratio Foun'd/

(SNP Design Acceptance Criteria)

~CHe

~CHe

65/45 Intercept of 45

% flow exclusion 0.41/(<.75)

0.42 boundary and APRH rod block line 47/27.6 42/23.8 Two pump minimum flow intercept with 100

% rod pattern line-Region A exclusion boundary Natural circulation flow intercept with 100

% rod pattern line-Region A exclusion boundary 0.77/(<.90)

0.86 0.64/(<.90)

0.81 The

.75 DR acceptance criteria was based on a Hay 10, 1984 NRC safety evaluation of the licensing topical report XN-NF-691P,

"Stability Evaluation of Boiling Water Reactor Cores,"

and its Supplement 1.

The.9 DR acceptance criteria was an SNP guideline.

The 1984'eview also concluded, separately, that COTRAN calculated core decay ratio values greater than 0.75 indicated potential core instability because of code calculational uncertainties.

The staff concluded that the methodology was acceptable for licensing of reload fuel with the condition that acceptable technical specifications were required to restrict operation if the calculated decay ratios exceeded 0.75.

The nature of the technical specification limitations on permissible operation was later modified to be consis'tent with Generic Letter (GL) 86-02.

GL 86-02 directed licensees to evaluate each core reload to assure that it was typical of previously evaluated cores which have accept'able stability margin.

Following the 1988 LaSalle instability, NRC Bulletin 88-07 and its Supplement 1 provided further guidance to reduce reliance on decay ratio calculations for avoidance of power oscillations, to improve training and procedures, and to verify the adequacy of instrumentation relied on for procedural response to oscillations.

Supplement 1 directed that for proposed new fuel designs, the stability exclusion region boundaries (based on GE fuel designs)

should be 'reevaluated and justified based on any applicable operating experience.

On March 18, 1992, the BWROG transmitted a letter to all owners with recommendations for improved implementation of Supplement

based on the insights, obtained from the work of the BWROG Stability Committee.

The letter stressed the need for greater caution when operating near the exclusion boundaries.

Reexamination of procedures and training to reflect uncertainties in the definition of power and flow stability exclusion boundaries was suggested.

During the licensing review of Advanced Nuclear Fuel (ANF) 9x9 fuel, both ANF (the predec'essor of SNP)

and the NRC concluded that the design was less stable than existing 'designs of Bx8 fuel in operating reactors.

During its introduction in the Susquehanna BWRs, the staff required initial startup stability monitoring of each core reload until the transition to a full 9x9 core was complete.

On the introduction of the SNP 9x9-9X design in Cycle

of WNP-2, stability decay ratios of 0.42, 0.86, and 0.81 were calculated using COTRAN for the respective power/flow statepoints identified in the preceding tabulation.

These were not provided to the NRC at the time.

After the August 15 event, LAPUR analyses by the staff and hydraulic stability evaluation by.SNP (based on the ratio of two phase versus single phase pressure drop)

indicated that the 9x9-9X fuel was.less stable than the ANF 9x9, and less stable than other fuels in US BWRs.

The reduced (>.75 decay ratio) stability margin which was calculated by SNP for Cycle 7 and 8 resulted in operation with reduced thermal margin during power oscillations (in fact, the CPR operating limit was exceeded during the oscillations of August 15).

The licensing bases assumed that a design basis accident would not be initiated from a condition which exceeded the CPR operating limit.

The licensee and SNP accepted the reduced stability margin for Cycle 7 and 8 without reviewing the possibility of initiati'ng a design basis accident with this reduced stability margin and potentially, reduced CPR.

The AIT concluded that the application of 9x9-9X fuel in WNP-2 should have received more specific licensing attention to the reduced stability margin, and potentially, may have been an unreviewed safety question.

5.2. 1,2 Core Nucle'ar and Thermal-H draulic Desi n

Historically, a major design goal for new fuel types in an existing core has been to match the hydraulic resistance of the existing fuel in order to maintain thermal-hydraulic compatibility of the mixed core during power operation.

Differences in hydrauli.c resistance tend to starve flow from the more resistant fuel and increase the uncertainty in the calculated core inlet

flow distribution.

While there is no regulatory criterion governing the acceptability of the match, the approximately

%

difference in pressure drops (measured from the inlet orifice to the top of the core support plate)

between the WNP-2 original GE P8x8R and the initial reload SNP 8x8-2 is typical.

The pressure drop mismatch between the SNP 8x8 and the SNP 9x9-9X is about

%, which appears to be undesirably large.

The Cycle 8 core design used the relatively unstable fresh SNP 9x9-9X fuel in a mixed core loading pattern which maximized the flow mismatch with neighboring once and twice burned SNP 8x8 fuel in regulating rod control cells.

The core loading with fuel of several different types and exposure histories also resulted in a core unrodded power distribution with radial peaking which is greater during the early core life than it is in other BWRs fueled by SNP.

However, unrodded radial peaking was also compared to LaSalle Unit 2, Cycle 4,

and found to be comparable.

WNP-2 was higher for the first three GWd/HTU and then slightly lower than LaSalle.

This core nuclear design may make the core more vulnerable to high peaking factors during core power maneuvering by injudicious selection of control rod patterns.

As discussed in Section 3.3 of this report, extremely skewed radial and axial peaking factors which occurred during startup maneuvers were the direct cause of the WNP-2 instability.

In addition, the mixed core design, and the SNP 9x9-9X fuel, in particular, may have made the core more vulnerable to both the core-wide and out-of-phase modes of instability (see Section 3.2).

5.2. 1.3 'xclusion Re ion Boundaries Assum tions The AIT considered whether appropriate design attention had been given to maintaining the core power distributions assumed consistent with the technical specification exclusion boundaries for WNP-2.

The BWROG methodology for exclusion boundary calculations is described in BWROG NEDO 31960,

"BWROG Long-Term Stability Solutions Licensing Methodology,",

Supplement 1, dated Harch 1992.

That methodology assumed a bottom peaked axial peaking factor of 2.0 in Node 3 of 24, and a radial peaking factor based on an End of Cycle (EOC) Haling calculated power shape (typical radial peaking factor of 1.5 for BWR Ss).

However, the exclusion boundary used for the interim stability solution required by NRCB 88-07 and its Supplement 1 was based on operating experience and was considered to be conservative.

Therefore, no limits were specified on power distribution.

Nevertheless,,

the AIT concluded that sufficient information was available to

.

'ndicate a relatively unstable core, and neither the licensee or SNP gave appropriate attention to the design and operation of the core to assure its stabilit.2. 1. 4 Reduced Core Flow Ca abi 1 it The licensee reported that maximum recirculation flow capability of WNP-2 had been reduced from 106

% to 96

% of nominal because of heavy crud deposits in the jet pumps.

The pumps were inspected by GE during the recent refueling outage and there was no evidence of cracks or other structural defects.

The source of the crud was believed to be metallic corrosion products from the condenser tubes, though the evaluation was incomplete at the time of the inspection.

The AIT viewed photographs of the jet pumps showing-heavy deposits at the venturi entrance with gradual reduction in crud thickness proceeding down the pump.

The AIT also noted that the natural circulation flow line for WNP-2 is significantly lower than other flow control valve (FCV)

reactors (for example, about 5% less than the LaSalle units).

This may be due to differing hydraulic characteristics, and makes WNP-2 more vulnerable to low flow instability.

The AIT noted that the plant transient analysis was performed at the 104

% power-106

% flow point which SNP confirmed to be conservatively applicable to rated power and flow 'conditions, but did not address the

% degraded flow capability that was known to exist.

However, the licensee claimed to have considered the applicability of the existing analyses to the reduced flow condition.

Though not documented, this review concluded that the existing analysis was bounding for permissible power/flow map operation based on the techniques employed for the operating HCPR limit determination.

Later, SNP analysis confirmed that the

%

flow condition was bounded by the existing analysis.

The AIT concluded that the WNP-2 Cycle 8 core design contributed to the instability of the core, Also, the impact of the SNP 9x9-9X fuel (initially loaded for Cycle 7), on the stability of the WNP-2 core was not adequately reviewed by either the designer or the licensee.

Nevertheless, the AIT and the licensee concluded subsequent to the August 15, 1992 event, that the core could be operated safely by conforming to the compensating operating restrictions and precautions specified in the licensee's August, 29, 1992 letter to the NRC for future low flow startup and shutdown operations.

These precautions included continuous monitorinp with the ANNA stability monitor below 50

% flow with power >25

%.

'.2.2 Evaluation of Core Desi n

ualit Assurance The AIT'reviewed the licensee's

.and the fuel vendor's core design quality assurance, and staffing as related to the August 15, 1992 event.

5.2.2. 1 Review of Su l

S stem Core Desi n Process Approximately six months prior to a refueling outage, the licensee begins planning for the next cycle reload.

A fuel vendor is

determined by the licensee and contracts are prepared.

For the Cycle 8 reload, the licensee selected SNP as its fuel vendor and reload designer.,

A Final Energy Notice was issued by the licensee on August 28, 1991 that established the final energy requirements and the design criteria for the reload.

The design criteria established the licensing, operational and economic requirements for the reload.

The Energy Notice established the thermal limits,

"reactivity limits and the design margins for those limits.

The notice was revised on November 22, 1991 to update the cycle energy requirements based on Cycle 7 fuel utilization.

Licensee personnel stated that during the cycle reload design process, Nuclear Engineering personnel were in frequent communications with the fuel vendor.

Because of the close proximity of the fuel vendor, the licensee often had direct

'articipation.

in the design process despite the design responsibility having been contracted to the fuel vendor.

Upon completion of the design, the fuel vendor provided the licensee with several design analyses.'pon completion of the design and receipt of the various design analysis reports, the licensee performed a sampling design review for each reload.

For cycle 8 reload, a sampling design review was performed on the fuel mechanical design, thermal-hydraulic design, safety and transient analysis and. core monitoring.

A design review plan was prepared and approved; a design review checklist was prepared and approved; the results of the design review were documented and questions and concerns resolved.

The licensee's sampling design review, in essence a technical audit, appeared to be penetrating in the areas reviewed.

However, the Cycle 8 design review,'nd other previous cycle'esign reviews performed by the licensee, did not appear to have included a review for core stability.,

A Core Operating Limits Report (COLR) was prepared by the licensee upon review and evaluation of the fuel vendor Cycle 8 design reports noted above.

The COLR described the Cycle 8 reload, provided the design thermal limits, and ascertained that applicable limits of the plant safety analysis were met.

The, licensee performed a

CFR 50.59 review for the COLR and submitted a change request to the WNP-2 Technical Specifications (TS) to reflect the Cycle 8 reload.

Amendment 109 to the facility operating license provided the NRC evaluation and approval of the proposed TS changes.

The AIT found that the licensee had performed appropriate sampling design reviews, appropriate evaluation of fuel design analysis results and report, and had appropriately utilized the design information for the reload in establishing the cycle thermal limits.

However, the AIT also found that the licensee's reviews had not thoroughly reviewed Cycle 8 core stability:-

5.2.2.2 Ev'aluation of Core Desi n

A Effectiveness I

The licensee selected SNP as both the Cycle 8 reload designer and as the fuel vendor.

SNP was an approved vendor on the licensee's approved suppliers list based on both Supply System audits and industry audits.

SNP had an.NRC approved Quality Assurance (QA)

program.

Revision 25 to its QA Topical Report was reviewed by the'RC and approved by the NRC in a letter to SNP dated February 18, 1992.

The AIT performed interviews, sampling procedure reviews, and sampling analysis reviews to determine the adequacy of the fuel vendor's design process for the WNP-2 Cycle 8 reload as it. relates to the August 15, 1992 event.

Several weaknesses were observed that appeared to have been m'issed opportunities to minimize the occurrence of the core instability that occurred.

The weaknesses observed were as follows:

Discussions with fuel vendor design management personnel indicated that they were not aware of the BWROG'March 18, 1992 implementation guidance and stability interim corrective actions.

~

,

A Cycle 8 stability analysis (Calculation E5072-N06-3 dated

'anuary 31, 1992)

was performed by the fuel vendor as part of analyses for the cycle 8 reload report.

Four points, approximately in the periphery of the stability exclusion zone, were analyzed to demonstrate core stability.

However, the analysis was based on end of cycle conditions and with all control rods withdrawn.

No consideration of rod patterns, and the consequent effect on stability, was required.

No analysis of the effects of various rod patterns was performed.

No recommendations were provided for rod patterns that may be required during operation in the proximity of the stability exclusion zone.

~

An evaluation was performed by the fuel vendor for "final feedwater temperature reduction" with thermal coastdown toward the end of Cycle 8.

However, the evaluation did not include a specific stability analysis for the Cycle 8 reload.

A stability analysis performed for Cycle.3 was used as a basis for the stab'ility evaluation despite the differences in flow characteristics of the two cores.

The AIT's interviews and procedure and analysis sampling review also indicated additional weaknesses that did not relate directly to core stability but which were design process discrepancies:

The WNP-2 cycle 8 reload design group consisted of seven engineers and one technician.

The design group performed independent verification of the various analyses required

for the reload design.

However, the independent verifier and the analysis originator frequently interchanged functions for various analyses due to the limited number of people in the group.

The effectiveness of the independent verification function was further minimized by independent verification of an individual's work on an analysis that utilized inputs from a different analysis that had been performed by the same'ndependent verifier.

For example, this was the case for the cycle 8 reload stability analysis calculation.

~

The fuel vendor had programmatic procedures for its design process that appeared to meet the requirements of 10 CFR 50 Appendix B and ANSI N. 45.2. 11, "guality Assurance Requirements for the Design of Nuclear'ower Plants."

However, the vendor did not have specific detailed procedures for performance of various specific analysis for the core reload design.

In addition, vendor programmatic procedures allowed design calculations to remain with the designers, uncontrolled, approximately six months after the cycle startup.

At the time of the inspection, the Cycle 8 stability calculations were still on the originator's desk.

5.3 Evaluation of E ui ment Performance The AIT reviewed the operation of the equipment listed below to determine any contribution of the equipment to the event.

Equipment that potentially had a

significant effect on core stability and monitoring was chosen for review.

The AIT reviewed the Shift Hanager's log, Control Room Operator's log, TDAS data, alarm print-outs, control room deficiency tags, associated maintenance work requests, associated problem evaluation reports, and the equipment history computer data base.

The AIT interviewed licensed operators, STAs, design and system engineers, and other members of the licensee's engineering and plant staff to determine equipment performance prior to, and during the event, and to determine their potential'ontribution to the event.

The AIT concluded that no equipment discrepancy was a direct contributor to the event.

5.3. 1 Recirculation Flow Control S stem Flow Control Valves (FCV) - 2-RRC-FCV-60A and 2-RRC-FCV-60B and Recirculation Pumps 2-RRC-P-lA and 2-RRC-P-2B The reactor was at approximately 36.4

% power and core flow was at approximately 30.5

% flow, both FCVs were in their full open position and both recirculation pumps were operating at

HZ (slow speed)

prior to the event.

In preparation for shifting the recirculation pumps to 60 HZ (fast speed),

FCV-60A was closed to its minimum open position in accordance with operating procedures.

TDAS data indicated appropriate recirculation flow response to the valve manipulation from approximately 9600 gpm to approximately 1000 gpm.

Core flow decreased accordingly, Noticeable indication of power oscillations shown in the TDAS plots

started when recirculation train A flow decreased to approximately 2000 gpm and FCV-60A was approximately 35 X open.

FCV-60B was left in the full open position through the event and TDAS data indicated a

relatively constant and appropriate train B recirculation flow of approximately 10,000 gpm.

Both recirculation pumps were at their 15 HZ slow speed prior to and throughout the event.

5.3.2

~Jet Pum s

TDAS data indicated that flow through the A Train jet pumps was approximately 17.37 million pounds/hr and through the Train B train jet pumps was approximately 16. 12 million pounds/hr prior to closing FCV-60A.

Flow through the Train A jet pumps decreased to approximately 10.4 million pounds/hr after the onset of power oscillations.

Prior to.the event, the licensee had been aware of jet pump fouling at WNP-2, had instrumented the jet pumps, was monitoring the pump performance, and was considering possible corrective actions.

The licensee considered that jet pump 'fouling was not unique to WNP-2 and was aware of other BWRs with similar conditions.

The net effect of the jet pump fouling. at WNP-2 has been a decrease of core flow such that WNP-2 is currently only capable of approximately 96 / core flow.

In addition to TDAS data, the licensee reviewed daily jet pump surveillance data since July 18, 1992.

The licensee review revealed no significant jet pump performance variance that directly caused the event.

The TDAS jet pump flow data confirmed the licensee's evaluation.

(See also the discussion in Section 5.2. 1.4)

5.3.3 Feedwater Control S stem Licensee personnel stated that the feedwater heaters'ater level control system was in automatic prior to and during the event.

They also stated that no abnormalities of the feedwater heaters, feedwater temperature and feedwater flow were noted prior to or during the event.

A review of TDAS data confirmed the interview information.

Feedwater loop A temperature was approximately 309 F and loop B was approximately 311 F prior to the event.

Feedwater flow was approximately 2.2 million pounds/hr prior to the event.

Turbine bypass valve testing and control rod drive valve timing were performed a few hours prior to the event.

Feedwater flow changes occurred as expected but resulted in no significant flow and temperature variance that contributed to the event.

5.3.4 Turbine Control S stem

'upply System operations and technical staff indicated that no abnormalities with the Digital Electrohydraulic Control (DEH) system that had an effect on the event were noted prior to the event.

The DEH system controls reactor pressure.

TDAS data confirmed that reactor pressure remained relatively normal at approximately 956.6 psig prior to

'he even.3.5 Loose Parts Detection S ste Prior to and during the event the loose parts detection system was inoperable as indicated in the Shift Hanager's Log.

Technical Specification 3.3.7. 10 requires that the loose parts detection system be operable during startup and operation.

The specification also states that TS 3.0.3 and 3.0.4 were not applicable to this situation.

The action statement for TS 3.3.7.10 requires that with one or more loose parts detection channels inoperable for more than 30 days, a special report be submitted to the commission.

The AIT found that, in this case, the licensee was in compliance with its Technical Specifications since the 30 day limit had not been reached.

No data relevant to the event was available from the loose parts detectors since the system was inoperable.

5.3.6 Stabilit Monitor TS 3.2.7 'requires the stability mo'nitoring system to be operable when operating in the region of potential instability defined in the technical specifications for various core 'flow and core thermal power conditions.

Supply System personnel indicated that WNP-2 had never operated in the defined region of instability and consequently had never had to have the stability monitoring system operable.

The monitor had been operated previously, however, for testing and, on one occasion, for informational purposes.

The STA on-shift at-the time of the event stated that WNP-2 was not in the Technical Specifications defined region of instability prior to and during the event.

Because of this condition, the stability monitoring system was not operable, nor was it turned on, prior to and during the event.

The stability monitoring system at WNP-2 was purchased from SNP and installed in WNP-2 as a non-safety related system.

The SNP stability monitor, called the Advanced Neutron Noise Analysis (ANNA) monitor, was evaluated by SNP for suitability for use at WNP-2 based on the LPRH and APRH inputs to ANNA.

Interviews held with SNP personnel indicated that SNP had recognized that the LPRH and APRH inputs to ANNA would have a 0.3 HZ filter.

SNP considered the effects of the filtering of the LPRH and APRH inputs to the stability monitor.

They tested for the filtering'effect using a

single sine wave, and determined that the filtering would have no effect on the decay ratio (DR) determined by ANNA.

However, the amplitude of oscillations determined by ANNA would be affected.

SNP determined that the amplitude warning and alarm setpoints for ANNA should be changed, and made the changes accordingly.

However, SNP personnel indicated that the change was only verbally communicated to Supply System personnel, but not documented in writing.

The 10 'percent peak to peak oscillation amplitude ANNA alarm was reset by SNP to 4.2 percent due to the filtering of the input data.

That is, a

10 percent amplitude oscillation would be shown as a 4.2 percent amplitude oscillation by ANNA.

WNP-2 surveillance procedure 7.4.2.7.3,

Revision 2, dated December 9,

1991,

"Core Stability Monitoring,"

paragraph 8.2 step 4, required a manual SCRAM if both ANNA APRH peak signal oscillations were greater than 10 percent, and one or more APRHs indicated a peak to peak amplitude greater than 10 percent of rated.

Steps 7 and 8 of the 'procedure required various steps to be taken should the DR exceed 0.6.

The procedure did not appear to have accounted for the 0.3 HZ filter nor the 'SNP evaluation and subsequent change to the, ANNA alarm that was necessitated by the 0.3 HZ filter.

Licensee personnel stated that licensee post-installation testing of ANNA was performed by putting digital data in to ANNA, downstream of the

. 0.3 HZ filter.

Consequently, the post installation testing did not identify the variance that the filter caused.

The post installation testing of ANNA did not appear to have been comprehensive, in that it did not appropriately test the total system's interaction with the input data.

Subsequent to the event, the Supply System recovered TDAS data and reanalyzed the data to determ'ine what indications may have been identified by ANNA.

LPRH and APRH data and corresponding licensee generated graphs were reviewed by the AIT.

The AIT noted differences in signals for two of the

LPRHs and a potential discrepancy in the data due to filtering.

Subsequent licensee inspection of the LPRH and APRH inputs to ANNA determined that LPRH 32-09 and 32-17 had a

5 HZ filter installed in lieu of the required 0.3 HZ filter.

Licensee and SNP evaluation of the processing of ANNA data based on independent processing both at WNP-2 and at SNP confirmed that the WNP-2 ANNA, except for the two LPRH feeds that had a

HZ filter, was working appropriately.

Furthermore, event data processed through ANNA after the event indicated a higher than normal DR would have been shown by ANNA a few minutes prior to the event had ANNA been operating.

Also, the SNP evaluation determined that the 0.3 HZ filter made a negligible difference when the DR was 1 but had an increasing effect as decay ratio decreased

.{up to a 0.27 DR difference with a

HZ filtered DR of 0.88).

The

.3 HZ filter had.a significant effect on ANNA DR output at less than

OR and would not have been fully effective for TS required stability monitoring had WNP-2 been previously in the TS defined region of instability.

As a result of the evaluation, SNP recommended, and the licensee accomplished, a change of the 0.3 HZ filters to 5 HZ.

The ANNA hardware and software were not safety grade.

ANNA does not provide an audible alarm; for this reason, the licensee has modified their procedures to have a dedicated ANNA operator to monitor ANNA whenever operating above

/o power and flow and below 50 / power and flow.

5.4 Evaluation of Licensee Event Investi ation The licensee initiated two efforts to investigate the August 15, 1992 event:

'a technical analysis and a root cause analysis.

Technical analysis of the event has been completed.

However, efforts to further define the core instability region and other longer term corrective actions were continuin At the conclusion of this inspection, the licensee's root cause analysis had not been completed.

A preliminary executive summary of this analysis was reviewed by the AIT.

5.4. 1 Technical Anal sis The licensee presented an initial technical assessment of the event to the AIT on August 17, 1992.

The assessment was flawed in that it did not consider that all but two of the LPRH data recordings had been attenuated by about

/o (see Section 5.3.6).

The licensee used data from one of these LPRHs, located at core position 32-09A, to represent the hot channel (worst case) for analysis purposes.

The licensee analyzed LPRH 32-09A data and determined that no Operational Hinimum Critical Power Ratio (OHCPR) limits had been violated.

The licensee did not attempt to explain why the peak oscillations appeared to take place on the core periphery, instead of at the center of the core where the highest neutron flux density was located.

The licensee later corrected and reanalyzed the LPRH data for the correct hot channel at the center of the core.

The licensee's reanalysis identified that the OHCPR limits were exceeded, but not the Safety Limit CPR.

The attenuation of LPRH data appears to'ave been identified in parallel by the AIT and licensee staff.

Other aspects. of the licensee's initial analysis appeared to be acceptable.

The licensee continued the technical assessment of the event through the second week of the AIT's site visit with assistance from their fuel contractor, Siemans Nuclear Fuel (SNP).

Short term corrective actions initiated by the licensee's technical investigation included:

Core stability was analyzed for a variety of different power distributions near the critical power/flow point where recirculation pumps are shifted to fast speed.

- Stability comparisons were made between the August 15, 1992 event and previous startups.

- Operational limits were established for plant parameters that effect reactor stability in order to reduce the likelihood of instability.

I Attenuation effects of filters on the LPRHs were analyzed.

The 0.3 hertz filters on the LPRH inputs to ANNA (Advanced Neutron Noise Analysis) were replaced with 5.0 hertz filters.

- Procedures were revised to include the cautions and limitations that resulted from these analyses.

The AIT concluded that the licensee's technical investigation adequately determined the immediate ca'uses of this even.4.2 Root Cause Anal sis The licensee's preliminary root cause analysis executive summary adequately identified what happened during this event; however, it did not develop an explanation why these events were not prevented by the licensee's staff.

The AIT identified several areas that were not addressed in the executive summary:

Inadequate implementation of BWROG guidance on instabilit Ineffective operator training and STA/SNE qualification regarding instabilit Failure of the crew to compensate APRH alarms/trips for high flux peaking factors (see Section 6 for further discussion).

E Failure to identify technical assumptions used in safety analyses and core stability analyses, and incorporate this limitations into operating procedures.

5.5 Assessment of the Licensee's Emer enc Notification Process PPH 13. 1. 1 "Event Classification,"

was the licensee's implementing procedure for emergency event classification.

PPH 13. 1. 1 contained no specific direction on classification of a power oscillation event.

Kowever, PPH 13. 1. 1 defined a Notification of Unusual Event (NOUE)

as

"A condition at the plant or its surroundings, that threatens the normal level of plant safety, or an event where an increased awareness on the part of plant operating staff is warranted'

"

Based on this information, the Shift Hanager declared a

NOUE due to "core instability."'he AIT determined that the licensee's event classification was proper for the August 15 power oscillation event, and that sufficient guidance was provided in the licensee's procedures for this determination.

All notifications appeared to be prompt and informational.

The AIT concluded that termination of the event when plant conditions were stable, with the operators executing a

controlled cooldown of the plant, was conservative.

6.

Descri tion of Flow Biased ATWS Tri Set oint Errors During a review of the data provided by the licensee, the AIT noted that the thermal limits printout at 0153 on August 15, 1992 indicated that the core T-factor (a measure of flux asymmetry defined by TS 3.2.2)

was calculated.to be 0.793.

However, in response to this, the licensee did not perform the flow biased setpoint adjustments or APRH gain adjustments as required by TS 3.2.2 whenever the T-factor was less than 1.

In addition, five other thermal limits printouts, between 0053 and 0300 on August 15, 1992, also indicated that the adjusted flow biased setpoints should have been in effect.

The lowest value calculated for the T factor was 0.746, just prior to the closure of the A loop recirculation flow control valv The AIT's review of licensee records indicated that PPH 7.4.2. 1,

"Power Oistribution Limits," had not been performed on August 14 nor August 15.

(PPH 7.4.2. 1 was the licensee's implementing procedure for determining margin to the thermal limits, for adjustment of the flow biased setpoints discussed above.)

The STA stated that TS 3.2.2 had always been interpreted by the Supply System staff to allow 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> after achieving 25/ power for performance of PPH 7.4.2. 1.

The STA also stated that although the computer printouts indicated that the APRH flow biased setpoints required adjustment, this adjustment could also be delayed for 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />.

The AIT concluded that this was an incorrect interpretation of the TS by the licensee.

PPH 3. 1.2,

"Reactor Plant Cold Startup," step 5.8.8 stated "Notify the STA to verify APLHGRs, LHGRs, MCPRs and APRH setpoints."

The CRS stated that he signed the step off because he notified the STA.

The Operations Hanager stated that he agreed with the CRS, and that the responsibility for completing this ste'p rested with the STA.

The AIT concluded that this was a poorly written step that could have provided much more clear delineation of responsibility 'for completing a Technical Specification requirement.

Step 5.8.8 was rewritten by the licensee prior to completion of the AIT inspection to require the STA's signature in PPH 3. 1.2 that 7.4.2.

1 has been satisfactorily completed prior to exceeding

%%d power.

The AIT calculated that the proper scram setpoint should have been 43.6 percent power, and the rod block/APRH Upscale alarm setpoint should have been 36.9 percent.

The instrumentation for the flow biased scram inserts a six second time delay to simulate the thermal time constant of the fuel, and therefore did not provide scram protection for a power oscillation event.

The flow biased rod blocks and APRH Upscale alarms do not have a time delay.

These trips apparently would have been received earlier in the event had the setpoints been lowered, providing the operators with earlier information to diagnose the onset of power oscillations.

In addition, as a result of these alarms, operators might not have increased reactor power to 36.4 percent, due to the close proximity to the rod block trip poin APPENDIX A Notes (1)

These numbers were estimated in a coarse manner from'the following facts derived from Fig.

1:

~

LPRH 32-17-C average value was 19 W/cm'

LPRH 32-17-C peak value at 02:59.00 was 21 W/cm'

LPRH 32-17-C peak value at 02:59.30 was 24 W/cm'

There were 15 oscillations in the 30 seconds between those measurements, indicating a

2 second oscillation period (0.5 Hz).

~

Therefore, the decay ratio was estimated to have been

(24 19) xs y 06 (>>->>I (2)

To perform these calculations, both the VIPRE and XCOBRA-T code were driven with an average core power oscillation of 30 N of rated power (conservative with respect to the observed 26 /).

The core pressure drop was forced to follow the measured pressure drop; this resulted in a peak-to-peak flow oscillation of 26 %.

The calculated CPR change was 0. 17 for the YIPRE code and 0.18 for the XCOBRA-T code.

For the conditions at the time of the instability, the HCPR at the hot channel (32-33)

was 1.946, the OLHCPR was 1.795, and the SLHCPR was 1.07.

Therefore, the HCPR during the osci,llations was 1.78 (= 1.946-0. 17), which was less than the OLHCPR but not less than the SLHCPR.

(3)

Control rod sequencing procedures at four other BWR sites were reviewed to determine if practices at WNP-2 were abnormal.

Discussions with resident inspectors and licensee staff identified some differences in startup strategies; however, most differences could be attributed to different plant design or operating limits.

Two major differences between sites include variable speed recirculation pumps, and the restriction that WNP-2 placed on operating recirculation pumps in fast speed.

However, procedures from the four other sites consistently followed two operating strategies that were consistent with WNP-2 procedures:

- A qualified reactor engineer could authorize changes to the control rod withdrawal sequence without peer revie Neither quantitative guidance nor requirements were specified to control reactor power distributions during startup.

The AIT concluded that the control rod sequencing procedures in use at WNP-2 were consistent with other licensee The utility used a training update system to incorporate information from operating events, licensee event reports (LERs), problem evaluation requests (PERs),

design changes, and generic information.

Coordination with the Operations department was conducted primarily through -the Operations Procedures Supervisor, who reported-directly to the Operations Hanager for development and modification of procedures.

The licensee also described a secondary strategy for developing training material that incorporated information from industry experience and various industry committees, including the BWR Owners'roup.

This strategy was an informal process in which information from industry committees was distributed according to a standard distribution list'or the specific topic.

Personnel receiving the information decided how to act upon it.

The Harch 18, 1992 BWROG letter was distributed according to this list, and went to the Hanager, Reactor Engineering, the Hanager, Nuclear Engineering, the STA trainer, and the former Supervisor of shift nuclear engineers, who was the SNE/STA on shift at the time of the event.

LAPUR calculations (See Appendix E) indicate that the proposed startup path (case-8 control rod pattern)

should, result in stable operation.

The maximum decay ratio calculated by LAPUR during the proposed startup is 0.3, and it corresponds to the operating point just following the closure of the flow control valve (FCV);

Stability calculations for future startup calculations will be performed by the STAIF code.

STAIF is a frequency-domain code with 1-D neutron kinetics and multiple thermohydraulic channels.

STAIF validation has not been reviewed by NRC.

SNP has proposed to submit validation documentation to the NRC in June 1993.

The AIT performed a set of audit calculations of STAIF analyses using the LAPUR code.

This audit has resulted in good agreement between the two codes (between

/o and 20 I mismatch in calculated decay ratios).

STAIF calculations for the August 15 event resulted in a decay ratio of 0.86, which is lower by 15 l to 20 X than the actual decay ratio during the event.

These analyses were as follows:

Cycle 8 Fuel Cycle Design Report - Harch 1992 (Provided the results of fuel cycle design calculations, design core loading, projected control rod patterns, and evaluations of shutdown margin, hot excess, and thermal margins for the core design)

Cycle 8 Reload Analysis - Harch 1992 (Summarized the results of various safety analyses for the cycle

reload including the fuel mechanical design analysis, the thermal hydraulic design analysis, the nuclear design analysis and the anticipated operational occurrences analyses)

Cycle 8 Plant Transient Analysis Report - June 1992 (Provided the results of the fuel vendors evaluation of postulated system transient events during cycle 8 operation)

Cycle 8 Startup and Operations Letter Report - June 1992.

(Provided the neutronics information necessary for cycle 8 startup including the initial estimated critical position and target control rod patterns)

(8)

The operators scrammed the reactor at 0300 on August 15, 1992.

The NRC resident inspector was called at home at approximately 0315, and was given an initial briefing by the Shift Manager.

At 0320 an Unusual Event was declared by the shift manager.

Notification of the emergency response team was made at 0327, and the Emergency Operations Facility Communications Center (EOFCC)

was notified via facsimile at 0335.

These notifications initiated automatic pagers to notify the emergency

'esponse team.

The EOFCC duty officer then notified the state, local, and other federal agencies by facsimile or telephone.

All state, local and appropriate federal agencies were notified within 15 minutes of the declaration.

At 0355, the licensee notified the NRC Headquarters Duty Officer via the Emergency Notification System (ENS).

The licensee terminated the Unusual Event at 0430 on the basis that the plant was in a stable shutdown condition.

The EOFCC was notified that the UE was terminated at 0440.

The EOFCC, in turn, made all of the appropriate notifications.

The NRC was nbtified via ENS that the UE was terminated at 044 APPENDIX B Detailed Sequence of Events The following table gives the sequence of events as reconstructed by the AIT through interviews, calculations, and review of licensee records:.

Time Hours 2355:00 August

0053:14 August

0256:14 Descri tion of Events Reactor Power held at 34 %, for Turbine Bypass Valve Testing and control rod drive timing testing of six control rods.

A power distribution limits printout (known as."HON Run") was obtained.

Core thermal Power was 37. 1 percent, and total core flow was 30.5 percent.

CHPF was 3.247, HCPR was 1.864, APF was 1.46, the radial peaking factor (RPF)

was 1.59, and the T factor was

.870.

Commenced shutting RRC-FCV-60A. Reactor power was at 36.4

%,

and total core flow was at 30.5

%.

The reactor was on the 74.9

% rod line.

A HON run taken just prior to the valve movement indicated CHPF was 3.782, HCPR was 1.801, APF was 1.59, RPF was 1.63, and the T factor was

.746.

Xenon reactivity worth was approximately 2.4

% dk/k.

Nearly all shaping control rods had been fully withdrawn from the core.

0258: 18 Onset of ower oscillations from APRH recorder.

0259:03 0259:49 0259:59 0300:28 (est.)

0300:38 FCV-60A Fully Shut.

The reactor was operating on the 74.2 percent rod line.

(33.5

% reactor power,

%

core flow) Operators noted LPRH downscales randomly li htin across full core dis lay.

Rod Block Alarm received (APRH Flow Biased, first of 17 received at two second intervals).

At this time APRHs were oscillating between 23 and

% power peak-to-peak every two seconds.

APRH Upscale Alarm received (Flow Biased, first of 5 received over the next 38 seconds

.

Operators evaluate power oscil'lations, Shift Hanager determined that a reactor scram was required due to ower oscillations )10

%

eak to eak.

Hanual Reactor Scram.

All control rods verified inserted, and reactor power verified at 0.

The turbine tripped, and electrical power supply shifted automatically to the startup transformer from the main turbine enerato Time Hours 0300'44 0300:51 0301:23 0311 0320 0337 0355 0430 0747'escri tion of Events Reactor Water Level decreased to +13", operators entered Emer ency 0 eratin Procedures.

Reactor Water Level was at its lowest level (-15")

o'

the wide ran e level instrumentation.

Reactor Water Level increased to +13" via Feedwater Level Control System in Automatic mode, and operators maintain level >+13".

The Shift Manager called the NRC Resident Inspector at home, and briefed the resident ins ector on the event.

The Shift Mana er declared an Unusual Event.'.

REA-RIS-19 (Containment LOCA Radiation Monitor) levels increased from 60 c m at 0300 to a peak of 1500 c m.

Operators informed NRC Headquarters Operations Officer via ENS of the Unusual Event and Engineered Safety Features Actuation.

Unusual Event terminated based on stable plant conditions.

Reactor coolant sample isotopic analysis revealed normal levels of all isotopes.

The APRM flow-biased Rod Blocks and upscale alarms were received because the setpoint for these trips was approximately 49.5

%,

and the peak power during the oscillation event appeared to fall within the instrument tolerance of the APRM flow biased trips (the peak power noted on any of the APRM channels was 48.75).

The increase in containment LOCA radiation monitors appeared to be due to the venting of the scram discharge volume to the equipment drain APPENDIX C

Stability Calculations for the August 15, 1992 Event

'Using the best estimates available for power, flow, and power distribution, LAPUR predicts a corewide decay ratio of 1.05 for the conditions of the oscillations.

To achieve this agreement with the data, it was required to model in detail the mixed core conditions present at the time in WNP-2.

Tables C. 1 and C.2 summarize some of the results from these LAPUR runs.

Two conclusions can be reached from the results in Tables C. 1 and C.2:

(1)

The =hot channel was most probably stable, but not with much margin (decay ratio 0.83).

A single channel thermohydraulic instability was not likely, but the AIT could not discard it without analyses because of the highly peaked radial power in this event.

(2)

Even though the event data shows that the instability was clearly of the in-phase or fundamental mode, the LAPUR analyses indicated that the out-of-phase mode was. fairly unstable too.

From these LAPUR results, the AIT could determine which of the two modes would likely dominate.

Therefore, other startup events with these highly peaked power distributions may resultin out-of-phase oscillations.

The AIT performed some sensitivity analyses to identify the root cause of this instability.

As stated before, the root cause was the highly peaked radial and axial power distribution.

The other contributing factor was the mixed-core characteristics present in WNP-2 at the time of the event.

For example, the AIT performed a

LAPUR run with exactly the same input conditions, but changing all 9x9-9X fuel for 8x8 fuel.

The resulting core and hot-channel decay ratios were reduced by 20

/o to 0.87 (core)

and 0.67 (hot channel);

The AIT performed another LAPUR run similar to the one above, but substituting all 8x8 fuel by 9x9-9X.

In this last case, the core decay ratio was 0.95 and the hot channel decay ratio was 0.73.

Therefore, the AIT concluded that the mixed core is 10 / worse than a full core of 9x9-9X fuel.

Table C.3 summarizes these results.

Based on the LAPUR studies, the AIT performed some more sensitivity studies to define a best-estimate region where WNP-2 should have been unstable under the August

startup conditions.

This exclusion region is based on the radial and axial power distributions observed on August 15.

Figure C. 1 shows the LAPUR-calculated unstable region based on the August

power distributions.

In this figure, the August

instability event condition is shown along with the startup path that the reactor operat'or followed.

Th'e AIT concluded from these best-estimate LAPUR calculations that:

(1)

The operating point where the oscillations were observed is within the unstable region for the in-phase (or core wide) oscillation mode, and barely outside the out-of-phase (or regional) oscillation mode.

Note that these are best-estimate, after-the-fact calculations.

If these had been predictive-type calculations, the AIT would have added some conservatism (at least a 20 / factor for the calculated decay ratio)

so that it would be predicted that this operating point could have oscillated in either the in-phase or the out-of-phase mode (2)

The operating point at 36% power and 30% flow, before the flow control:

valve was closed, was barely stable; therefore, should the ANNA decay rat'io monitoring system had been operational, it would have displayed a

large (about 0.9) decay ratio before the pump upshift operation was started.

Table C. 1 LAPUR-calculated channel decay ratios for nominal 8/15 conditions Region number Number Bundles

9x9-9X

9x9-9X

9x9-9X 125.

SxS 121 SxS

SxS 350 mixed Relative Power 136%

162%

187%

106%

134%

156%

62%

Relative Flow 98%

92%

87/

104%

'01%

97%

101%

Channel Decay Ratio 0.46 0.61 0.83 0.04 0.18 0.41 0.0 Table C.2 LAPUR-calculated out-of phase decay ratios for nominal 8/15 conditions Assuming Eigenvalue separation between first azimuthal and fundamental mode is

-$0.5

-$ 1.0 First azimuthal (out-of-phase)

mode decay ratio 1.07 1.00 0.92

Table C.3 Comparison of stability of mixed core versus single-fuel cores Core Fuel Hixed Sx8 9x9-9X Core decay ratio 0.87 0.95 Hot-channel deca ratio 0.83 0.67 0.73 Out-of-phase deca ratio 0.79 0.89 Assumes

-$ 1.00 eigenvalue separation

WNP-2 8/15 STARTUP CONDITIONS

LAPUR-consta, estImat nt out.

ed IIn of-Pllg e of se DR~

p

0 w

E I:I

<0

--- ----tA'PU cons R-- eat: Iln t.ant.

I n aCed--.

-phase I-I-ne--off---

DR=I

25

20

35

<0

F LOW g%)

55 Figure C. 1 Best-estimate lines of constant decay ratio=1.0 for actual conditions of 8/15 event, assuming constant power distribution

APPENDIX D Stability Calculations for the August 30, 1992 Startup The AIT performed. a series of LAPUR calculations to determine the stability of the new startup path proposed by WNP-2.

This new startup path was designed to minimize the axial and radial peaking factors by selecting a nonaggressive control rod pattern.

Some of the parameters that define this proposed startup are shown in the predicted NON output shown in Figure D. 1.

Based on these parameters and the -predicted full 3-D power distribution supplied by the licensee, the AIT performed a series of LAPUR calculations to determine the stability of the proposed conditions.

The results are summarized in Table D. 1 and Figure D.2.

The maximum calculated core decay ratio during the proposed startup is 0.3, which indicates a significant margin to instability limits.

Table D. 1 LAPUR calculated decay ratios for proposed.

WNP-2 startup Power (%)

30%

33%"

30.3%

Flow (%)

30%

30%

26.3%

Core decay ratio 0.25 0.28 0.30 Out-of-phase decay ratio'.08 0.12 0.16

'ssumes a -$ 1.0 eigenvalue separation

'A Case

BEST ESTIHATE ¹7 MITH FCV CLOSED ANO FLAT INITIAL GUESS CORE PERFORHAHCE LOG SHORT EDIT PREDICT CALCULATIOH CALCULATION TYPE :

NORHAL COHVERGEHCE

TIGHT SYNNETRY :

FULL CTP CALCULATIOH

HEAT BALANCE SAV$-92JUN25.162052 M2C8F1 T'L'0 TSSS BOC STATE CONDITIONS FLOM RATES CORE PARAHETERS NUCLEAR LIHITS LOCATION GHME 336.40

'MT CHEO 0.2333 CNPF 3.203 25-26-07 GHMT 1008.0 (30.3")

'MTSUB CAEO 0.1346 CHFLCPR 0'88

20 PR 962.5 MTFLAG CAOA 0.0445 CNAPRAT 0.329 25-36-07 DHS 32.00 MFM CAVF 0.2858 CHFDLRX 0.344 25.26.07 MT 28.54 (26.3

)

'4D CAPO 14.9098 CHFDLRC

,1.134 25-25-07 PRAT IO 1.048 RML ER 1.052 CDLP

.0.8448 P-PCS

~ 4.73 25-28.08 ERATIO 1 '04 TARGET F 048 DPCC 4.7782 P.PCFC 4.73 25'8 08 CYCLE EXPOSURE 513.5 NMD/HTU KEFF 1.0089 P-PFPR-8'9 25-26 07 LOCATION

2

4

6

8

10

12 AXIAL REL PQMER 0.31 0.80 1 ~ 14 1.34 1.31 1.22 1.14 1.13 1.12 1.03 0.89 0 ~ 58 REGION REL POMER 0.92 1 ~ 04 0.92 1.04 1.20 1.04 0.92 1.04 0.92 RING REL POMER 0.83 1.33 1.05 1.05 1.02 1.26 1.08 0.66

~

' 'N11tt1%'t*01 CONTRQL ROQ QATA o1'Pet ~ 0 ~ %tete

59

51

4339--

35.-

3127--

23--

19 -.

11

03

/

08

00

08

06

00

06

10

06

14

22

18

42

16--.40

00

00

08

22

30

~

OQ o

--

00

--

'00

--

04 38 "

00

--

00

--

00

30

38

00

00

00

38

18

16

42

42

50

06

00

06

08

~.

38

. ~

18

50

58,

55

47

-- 39

-- 35

-- 27

-- 23

- ~

15

07

58 Figure D. 1 Proposed conditions for next startup.

Two pumps at minimum speed, one flow control valve close 'LI

WNP-2 PROPOSED* STARTUP CONDITIONS

p

0 w

E R

~ 30

<0 DR= 0

~ 28

25

DR=

g

~ 25

25

35

45 FLO%'%)

55

Figure D.2 LAPUR calculated decay ratio for proposed startup path k

APPENDIX E

Stability Calculations for the August 2, 1992 Startup On August 2nd, the WNP-2 plant was started up following a less aggressive rod withdrawal sequence than on the August 15th startup that resulted in unstable power oscillations.

This August 2 startup can be used as an example that the Cycle 8 core can be operated safely without startup instabilities if care is taken to minimize the radial and axial peaking factors.

Figures E. 1 and E.2 provide a comparison of the control rod patterns and limiting parameters for the two startups.

As it can be observed, the August 2 startup reached a

smaller maximum core peaking factor (3.468 compared to 3.856 for August 15),

significantly lower core-average axial peaking factor (1.29 compared to 1.62 for August 15)

and lower CPR ratio (0.826 compared to 0.922 for August 15).

All these conditions favored the stability of the August 2 startup.

The August 2 operating conditions were reached by not withdrawing 39-30 and symmetric rods (12 notches compared to 28 notches for August 15)

as much, and compensating the additional reactivity by pulling more rods all over the core.

The August 2 control rod pattern results in a more uniform radial and axial power profiles.

SNP estimated, using the STAIF frequency domain code, that the in-phase-mode decay ratio for the August 2 startup was 0.78 (compared to 1.01 for the August 15 startup).

The improvement in decay ratio value is mostly due to control rod pattern and the resulting power distribution.

If the August 15 startup had been run with the August 2 control rod pattern, STAIF estimated that the decay ratio would have been 0.68 (the August 15 startup was performed at a

lower rod line that the August 2 startup).

WNP-2 WK-9232 92AUG02-09.54.20 224 MWD/MTU PREDICT CALCULATIO:

CORE PERFORMANCE LOG -- SHORT EDIT PREDICT CALCULATION CALCULATION TYPE

NORMAL CONVERGENCE 'IGHT CTP CALCULATION

HEAT BALANCE SAV$-92JUN25-162052 SYMMETRY: FULL W2C8F1 TWO TSSS STATE CONDITIONS GMWE 392.66 GMWT 1194.1 (35.

PR 963.3 DHS 34.80 WT 29.64 (27

'RATIO 1.060 ER 1. 052 ERATZO 1.004 CYCLE EXPOSURE LOCATION AXIAL REL POWER REGION REL POWER RING REL POWER APRM GAFS More FLOW RATES WT 29.6 9~o)

WTSUB

85 WTFLAG,

WFW 4.80 3~o)

WD 7.18 CORE PARAMETERS CMEQ 0.2934 CAEQ 0.1562 CAQA 0..0527 CAVF 0.3273-CAPD 17.6625

'WL

"'35.7981 CDLP-0.82'42 4.8673 1.0079 NUCLEAR LIMITS CMPF '.468 CMFLCPR 0.826 CMAPRAT 0.422 CMFDLRX 0.441 CMFDLRC 1.228 P-PCS P-PCFC P-PFPR-4. 32-4. 32 7 ~ 3 2

7

9.

1.07 1.00 0.97 0.99 0.99 1.05 0.86 1.06 0.88 1.28 1.04 0.47 0.66 TARGET 1. 04 8 DPCC 223.4 MWD/MTU KEFF-1

3-

5 0. 54 1. 19 1. 29 l.'26 1. 16 0.90 1.09 0.88 1.07 1.28 0.82 1.37 1.20 1.29 1.10 0.65 0.69 0.68. 0.66 0.67 LOCAL 35-36 35-3 E, 35-3 E 35-36 35-36, 17-28 17-28 35-36

0.92 0.

CONTROL ROD DATA **+**********

59

51

,47

39

00

27 00,

'3

15

.

'03

00

16

08

36

--

16

00

08

36

10

18

10

18

26

00

--

08

--

00

--

12

--

00

--

08

--

16

--

00

26

16

00

00

08

00

12

08

36

08

00

12

00

34

42

50

34

42

50

00

59

51

43

00

31

27

19

11

03

Fig E. l.

Summary of core state for the August 2nd startup condi tions with one flow control valve closed and both recirculation pumps at minimum 'spee WNP-2 WK-9234 92AUG15-03.00.14 508 MWD/MTU TRIGR=BACKUP REV=FE CORE PERFORMANCE LOG -- SHORT EDIT MON CALCULATION CALCULATION TYPE

NORMAL CONVERGENCE:. TIGHT CTP CALCULATION

HEAT BALANCE SAV$-92JUN25-162052 PPLXBU SYMMETRY

QUARTE; W2C8F1 TWO TSSS B'TATE CONDITIONS GMWE 336.40 GMWT 1119.0 (33.

PR 963.3 DHS 35.81 WT 29.02 (26.

PRATIO 1.216 ER 1. 052 ERATIO 1.004 CYCLE EXPOSURE FLOW RATES WT 29.0 74)

WTSUB 38.94 WTFLAG

WFW 4.26 74)

WD 7. 63 TARGET 1. 048 508.6 MWD/MTU CORE PARAMETERS CMEQ 0.3394 CAEQ 0.1457 CAQA

=0.0494 CAVF 0.3450 CAPD 16.5517 RWL 35.8187 CDLP-0.8448 DPCC 4.7954 KEFF 1.0059 NUCLEAR CMPF

.

CMFLCPR CMAPRAT CMFDLRX CMFDLRC P-PCS P-PCFC P-PFPR LIMITS 3.859 0.922 0.441 0.460 1.367-3.63-3.63-7.07 LOCAT.

35-26 35-36 35-26*

35-26 35-26 33-18-33-18 3 5-2 6.

LOCATION AXIAL REL POWER REGION REL POWER RING REL POWER APRM GAFS

.

More

2

4 0. 78 1. 62 1. 60 1. 42 0.87 1.06 0.86 1.07 0.96 1.61 1.41'.37 0'.95 0.96 0.94 1.00

1. 26 1. 50 1. 09 0. 93

7

9

11 1;

1.14 1.01 0.89 0.79 0.66 0.51 0.,

1.06 0.84 1.01 0.87 1.26 0.97 0.42 0.94

• • • +*********

CONTROL'OD DATA *************

59

51

43

35 00

27

23

15

07

02

00

00

- 38

00

00

00

--

-

16

00

--

00

28

28

00

28

--

00

00

--

00

~ 00

10

18

00'6

30

00

38

10

18

26

~ 30

46

54

00

= --

00

--

00

--

00

--

00

--

00

46

54

59

51

43

00

31

27

19

11

03

CONTROL ROD SEQUENCE CONTROL RODS SYMMETRIC A-2 CONTROL ROD,DENSITY

0.267 Figure E-2.

Summary of core state for the August 15th startup conditions with one flow control valve closed and both recirculation pumps at minimum spee APPENDIX F

Review of Previous C cle 8 Startu Reactivit Anomalies On July 4, 1992, the licensee conducted two startups.

During both of these startups the core was critical on a rod significantly. before the expected rod.

On July 4, 1992, which was the initial post-refueling startup, the reactor went critical ten rods before the expected critical position in one half of the core.

Eight of those ten rods were lower worth rods.

The Hanager, Reactor Engineering, and personnel from the fuel vendor, were present to observe this startup.

Based on the judgement of the personnel present, the decision was made that no abnormal conditions existed.

The startup continued based on the recommendation

'from the Hanager, Reactor Engineering.

Based on reload analyses calculations SNP had predicted that criticality would be reached when pulling rod 18-,15, which is 12 rods behind rod 26-07 (the actual critical rod) in the startup sequence (see Fig.

F. 1).

SNP now estimates that the cold critical calculations were in error by approximately 0.5

/o ZK/K (or '5 mK); they attribute the error, in part to the fact that Cycle 7 was a very short cycle and the residual gadolinium concentration was hard to predict.

SNP presented some historical data (see Fig. F.2) which indicated that most of their cold-critical k-effective calculations lie within a band that is 1.0

%%u ZÃ/K wide.

Therefore, the licensee concluded that the 0.5 X ZÃ/K error was not unusual.

The TS permitted a maximum cold-critical k-effective error of 1.0 / dK/K.

A secondary effect of mispredicting the.rod at which criticality occurs is that rod 26-07 is a high worth rod that, when withdrawn, results in a large'lux perturbation.

This high worth resulted in the flux tilt that was observed during the startup.

As a general rule, it is preferable that criticality be reached with a low-worth rod to'inimize flux tilts.

The Shift Hanager submitted a Problem Evaluation Request (PER) to have the problem formally evaluated.

The Explanation and Proposed Resolution, provided by the Hanager, Reactor Engineering, provided. several factors that contributed to what was termed

"a situation of uncommon flux distribution during the initial critical."

The factors included a high worth rod being pulled in proximity to SRHs and IRHs, which made the flux appear higher in the vicinity of the rod being pulled.

The PER further explained that control room personnel should expect this type of response whenever high worth rods are pulled resulting in an asymmetric rod pattern.

If this situation created difficulty to Operations personnel, then banking high worth rods together with their symmetric counterparts was suggested within the constraints of the Rod Sequence Control System.

While this event was not the focus of the AIT, consideration of this event as it related to the core power instability event was specified in the AIT charter.

The AIT could not find any technical nexus-between the early criticalities and the power oscillation events.

The AIT noted that the Shift Hanager was concerned enough to generate the PER, and should be comme'nded for pursuing resolution of this anomal The AIT was concerned, however, with the licensee's position that early criticalities, even with high worth rods, were to be expected.

The focus of

. the PER-provided solution was on the flux tilt, and not on the early criticality.

Furthermore, rather than direct that the shift staff stop and ensure a reactivity anomaly is well understood, the Reactor Engineering Department provided a means to get around flux anomalies by pulling banked rods.

How this would correct for the early criticality was not explaine.0 I

I I

I I

I I

I I

I I

I I

I I

I

~

~

I I

)0rea'(c f~

CP f

) fCn/

/SD'r 1.4 1.2 hC D

1.0 cH 0.8 0.6 Weegu~god gg-j)7 aV ~yPcg gg c p'r Vie&

period/ /O8 <ec 0.4 0.2 0.0 I

I I

I I

I I

~

I I

I c) 8

~ g R

o 5 4

~~

~c Ch

~

W W

W Ol cv cv ev S

CV sh

&

'

Control Rod Fig. F.l.

Control rod viorths for the selected startup viithdrawal sequence.

Criticality was predicted for rod 18-15, but Mas reached at rod 26-07, a 0.5',l 8</K erro NICROBURN-8 CRLCULRTED COLD CRITICRL RNRLYSIS'anuary 1992 1.022 1.020 1.018 1.016

1.010 o

1.012 I

1.010 I

1.008 0 HNPl p

GGl

KS1

+ KS2 x 501 TARGET

0S C0 C3 1.006 1.000 1.002 1.000

p 0h +

0 0.998 C3 0.996 0.990 0.992 0.990 0.988 0.986-10 00

1000 2000 3000 5000 5000 6000 7000 8000 9000 10000 11000 12C CycLe Exposure (NHd/NTU)

Fig.

F.2.

Historical cold-corrected k-effective values predicted by SNF, showing a

10% ZE/K bounding error ban APPENDIX G

Survey of Rod Sequence Practices at Other BWRs LA SALLE:

Procedures reviewed:

LAP-100-13, "Control Rod Sequence Package Preparation, Review, and Implementation" LTP-1600-2,

"Guidelines For Control Rod Sequence Development" A Qualified Reactor Engineer (QRE)

may authorize changes to the sequence, and documents approval on LAP-100-13 Attachment F.

A new sequence, or major revision (not defined)

must be approved by two QREs, including the lead QRE or designee.

No prescriptive guidance regarding peaking factors or power dist'ributions during startup were identified.

SUSQUEHANNA:

Procedures reviewed:

AD-QA-138, "Control of Core Reactivity Changes At SSES" RE-2TP-013,

"Power Ascension and Shaping Using Control Rods" A Qualified Reactor Engineer (QRE)

may authorize changes to the sequence, and documents approval on forms AD-QA-138-10 and AD-QA-138-2 (page 13).

Procedure specifically identifies this change method for setting rod pattern on startup.

A new sequence must be approved by two QREs, including the lead QRE or designee.

No prescriptive guidance regarding peaking factors or power distributions during startup were identified.

HOPE CREEK Procedures reviewed:

HC.RE-IO.ZZ-001(Q),

"Core Operations Performance" RE.FH.ZZ-OOI(Q), "Guidelines for Control Rod Movement Power Operation" A Qualified Reactor Engineer (QRE)

may authorize changes to the sequence, and documents approval on approved forms. Prescribed control rod patterns are followed to the 60M rod line, with 55/ total core, flow (variable speed recirculation pumps).

Flux shaping is then performed - only limits are Technical Specification thermal limits.

"I No prescriptive guidance regarding peaking factors or power distributions during startup were identifie GRANO GULF:

Procedures reviewed:

17-S-02-400,

"Control Rod Sequences and Movement Control,"

A gualified Reactor Engineer ((RE)

may authorize changes to the sequence, and documents approval on a Hovement Tracking Sheet (Attachment II or III).

Page 9 gives the (RE specific (broad) authority to modify or amend the sequence.

A new sequence, or major revision (not defined)

must be approved by two gREs, including the Reactor Engineering Superintendent or designee.

No prescriptive guidance regarding peaking factors or power distributions during startup were identified.

DUANE ARNOLD Procedures reviewed:

RCP-DI-¹5 A gualified Reactor Engineer ((RE) must revise the entire pull sheet to perform any sequence modification.

This revision is then approved by reactor engineering group leader, and operations.

Any major events (startup, sequence exchange, etc.)

are discussed with their GE refueling representative.'):

Duane Arnold has variable speed recirculation pumps, and typically does not pass above the 70/ rod. line while near the flow instability region.

No prescriptive guidance regarding peaking factors or power distributions during startup were identifie a ~

APPENDIX H AUGMENTED INSPECTION TEAN, CHARTER WASHINGTON NUCLEAR PLANT 2 POWER OSCILLATIONS ON AUGUST 15, 1992 The Augmented Inspection Team (AIT) is to perform an inspection to accomplish the following:

E 1.

Develop a complete'equence of events and description of the power oscillations which occurred on August 15, 1992.

In addition, develop a

sequence of events and description of the plant reactivity anomalies which occurred during the plant startup from the most recent refueling outage.

Include in the sequence of events operator actions, decision making and communication'eading up to the decision to trip the reactor.

2.

Identify those equipment failures human performance errors, procedural deficiencies, and quality assurance deficiencies that contributed to these events.

In making this analysis, specifically include the following:

a.-

Determine what procedures the licensee had implemented.to avoid the mitigate power instability, and assess the adequacy of these procedures.

b.

Review the core fuel loading and rod patterns in effect during the event.

Review the most recent core reload safety analysis to determine whether the predicated areas of flow instability were properly calculated.

Assess whether these areas were contributing causes for the events.

c.

Review the operating status of the recirculation flow control system.

Assess whether the operation of the components in this system were contributing causes foe the events.

d.

Review the operating status of the turbine control system.

Assess whether the operation of the components in this system were contributing causes for the events.

e.

Assess the operating crew's performance during and subsequent to each event.

Ascertain the control room organization and staffing level during the event and the role of the STA before, during and after the event.

3.

Evaluate the effectiveness of the licensee's investigation of these

. events.

4.

Determine the root causes of the power oscillation event, from equipment, personnel, and organizational perspectives.

5.

Assess the adequacy of the licensee's notification of other agencies of both of these events.

Inspection Procedure 93800,

"Augmented Inspection Team Implementing Procedure" and Hanual Chapter 0325,

"Augmented Inspection Team" provide additional administrative guidance with will be used by the AI APPENDIX I 1.

PERSONS CONTACTED Washin ton Public Power Su

S stem WPPSS D.

• A.

• J

• J

• C

• J

• G D.

D.

• L W.

R.

W.

L.

S.

• R.

• J W.

D.

J.

S.

C.

G.

D.

D.

R.

S.

D.

G.

D.

R.

R.

D.

G.

W.

L.

V.

W.

H.

C.

C.

L.

L.

T.

D.

L.

H.

L.

L.

L.

C.

A.

J.

E.

L.

J.

H.

W.

K.

0.

R.

L.

A.

D.

H.

J.

R.

C.

R. 'H.

B.

J.I.

G.

R.

J.

Je J.

Sieme ution ion Hazur, Managing Director Oxsen, Deputy Managing Director Parrish, Assistant Managing Director, Operations Baker, WNP-2 Plant Manager Powers, Director, Engineering Gear hart, Director, Quality Assurance Sorensen, Regulatory Program Hanager Larkin, Engineering Services Manager Whitcomb, Nuclear Engineering Manager Harrold, Assistant Plant Manager Shaeffer, Operations Manager Webring, Technical Manager Sawyer, Shift Manager, Operations Grumme, Manager, Nuclear Safety Assurance Washington, Manager, Nuclear Safety Engineering Romanelli, Manager, Communications Britton, Public Affairs Officer Estes, Control Room Supervisor Strote, Control Room Supervisor Rhoads, Manager, Operations Event Analysis and Resol HcKay, Manager, Licensed Operator Training Halbfoster, Chemistry Manager Wooley, Procurement QA Manager Herhar, Operations Procedures Supervisor

'Atkinson, Reactor Systems Supervisor Vosburgh, Safety Analysis Supervisor Kirkendall, Plant Support Engineering Supervisor Moore, Control Room Operator Westgard, Control Room Operator Hughes, Control Room Operator Torres, Principle Engineer, Reactor Engineering Talbert, Station Nuclear Engineer, Shift Engineering Skeen, Principl'e Core Analyst Engineer, Fuels DeBatista, Principle Procurement QA Engineer Simons, Principle Procurement, QA Engineer Nowack, Operating Experience Engineer Huth, Engineer, Operations Event Analysis and Resolut'kins, System Engineer Freeman, System Engineer ns Nuclear Fuel Cor oration J.

L.

" R.

C.

A.

Morgan, Vice, President, Engineering Federico, Manager, BWR Fuel Engineering (BWRFE)

Copeland, Manager, Reload Licensing Volmer, Manager, QA Reparaz, Manager, Fuel Design

J.

Ingham, WNP-2 Reload Team Leader, BWRFE D. Pruitt, Engineer, BWRFE P.

Wimpy, Senior Engineer

'.

Nelson, Senior gA Engineer S. Jones, Senior Engineer J. Maryott, Staff Engineer General Electric GE H.

C. Pfeffler, Licensing Project Manager C.

R. Boznak, Site Service Manager D. A. Salmon, Senior Engineer Bonneville Power Administration BPA

• J.

R. Lewis, Director, Division of Nuclear Projects A. J.

Rapacz, Project Representative Institute of Nuclear Power 0 erations INPO P. Huffmeier, Events Analysis Nuclear Re ulator Commission N. Conicello, Project Manager, NRR J.

Wechedburger, Regional Coordinator and Policy Analyst, OEDO N. Hunnemuller, Operator License Examiner, NRR M. Peck, Acting Resident Inspector, LaSalle Station Others

• E. Smith, Tri City Herald

• Attended AIT exit meeting on August 29, 199 APPENDIX J Shift Staffing During the Event The WNP-2 control room was staffed by two licensed senior reactor operators, the Shift Hanager (SH)

and the Control Room Supervisor (CRS); three licensed reactor operators, known as Control.Room Operators (CRO);

and a Shift Technical Advisor (STA).

This manning met the staffing level required by the.

Technical Specifications (TS).

The STA was certified for his position by the licensee.

The operating crew was performing the following duties:

POSITION FUNCTIONAL RESPONSIBILITIES CRS CRO¹1 The SM came on watch at 2230, and was responsible for providing the operating crew overall direction and control.

He was in the control room during the event near the CRO's desk.

The CRS came on watch at 0230, and was responsible for directly supervising the plant startup.

He was in the control room near panel P-603 (which contains the full core display and control rod selectors)

during the event.

CRO¹1 came on watch at about 1930, and was operating the reactor recirculation system flow control valve (FCV) just prior to the event.

CRO¹2 CRO¹2 came on watch at 0230, and was coordinating containment inerting with nitrogen from a control room back panel.

After the reactor scram, he assisted the other operators with the plant shut down.

CRO¹3 CRO¹3 came on watch at 1830, aqd had been performing control rod manipulations at panel P-603 before the event.

During and after the event, he monitored balance of plant systems, including the feedwater and condensate systems.

STA/SNE 'he STA came on shift at 2230.

The STA was also qualified as a

Shift Nuclear Engineer (SNE),

and is responsible for monitoring and controlling the reactor's neutron flux profil APPENDIX K Partial List of Procedures and Corres ondence Reviewed 1.

PPH 3. 1.2,

"Reactor Cold Startup" 2.

PPH 9.3.9,

"Plant Power Haneuvering"

'.

PPH 4. 12.4.7,

"Unintentional Entry Into Region of Potential Core Power Instabilities" 4.

PPH 5. 1. 1,

"RPV Control (Non-ATWS)"

5'.

PPH 3.3.1,

"Scram Recovery" 6.

7.

r Letter dated September 12, 1988.

The licensee's response to NRCB 88-07.

Letter dated Harch 3, 1989.

The licensee's response to NRCB 88-07 Supplement 1.

8.

NRC Inspection Reports 50-397/88-37 and 89-11, which included review and closeout of the licensee's responses to NRCB 88-07 and NRCB 88-07 Supplement 1.

9.

. Administrative Procedure 1. 10.4,

"External Operating Experience Review,"

which contained the licensee's formal program for review of external operating experienc APPENDIX L Definitions of Core Ph sics Parameters fuel bundle power. that would CPR =

cause de arture from nucleate boilin DNB hottest operating fuel bundle power CHFLCPR =

o eratin limit CPR actual minimum operating CPR CHPF

=

or hottest fuel in ower average fuel pin power (axial x radial x pin) peaking factors at the -location of the hottest pin

hw