United States Nuclear Regulatory Commission - Protecting People and the Environment

ACCESSION #:  9502030105

BW
B&W Nuclear Technologies
                                             3315 Old Forest Rd
                    January 27, 1995             P.O. Box 10935
                    JHT-95-12          Lynchburg, VA 24506-0935
                                        Telephone: 804-832-3000
                                        Telecopy : 804-832-3663
Document Control Desk
U.S.  Nuclear Regulatory Commission
Washington, D. C. 20555

Subject:  Report of Preliminary Safety Concern Related to Large Break
          LOCA ECCS Analyses

Gentlemen:

The purpose of this letter is to advise you of a potential safety concern
regarding the effects of initial conditions used for large break LOCA
analyses for ECCS evaluations of the B&W operating plants and to report a
condition of nonconservative data handling in the Evaluation Model that, 
when corrected, results in a change in peak clad temperature of greater
than 50 F.

This concern relates to the initial core flood tank (CFT) conditions of
pressure and inventory and their potential effects on peak clad
temperature (PCT) predictions.  Historically, the CFT conditions assumed
for calculations at the 2-ft core elevations have been based on the
minimum CFT pressure and liquid inventory.  These inputs, which represent
the range of Technical Specification limits plus instrument uncertainty,
were considered to be the most limiting initial conditions for the 2-ft
LBLOCA PCT analysis.  Recently, a sensitivity study performed with a new
unapproved Evaluation Model (BAW-10192), concluded that the most-limiting
2-ft PCT is calculated when the maximum CFT liquid inventory and minimum
pressure were input as the initial conditions.  When the initial
conditions of maximum CFT liquid inventory and minimum pressure were
applied in the currently approved Evaluation Model (BAW-10104), the peak
clad temperature change was found to be greater than 50 F.

The 2-ft LOCA linear heat rate analysis limits for all operating plants
require preliminary reductions ranging from approximately 0.3 to 1.3
kW/ft, depending on the plant and fuel designs, to ensure that calculated
PCT's are within the acceptance criteria contained in 10 CFR 50.46.  A
preliminary review of all core operating limits for plants for which BWNT
performs the power distribution (all operating B&W designed plants,
except the Oconee Units), with administrative reductions in the LOCA
limits appropriate for the fuel and plant in question, suggests that the
current operating limits will be unchanged by these reductions.  This
information has been transmitted to all operating plant utilities.  Duke
Power Company has performed an evaluation based on preliminary plant
specific LOCA kW/ft limits and has determined that a small reduction in
core operating limits would be necessary.

In accordance with 10 CFR 50.46, this letter documents NRC notification
of a PCT change in excess of 50 F.  A change in peak clad temperature in
excess of 50 F will result from the change in CFT initial conditions or
from correcting the nonconservatism in data handling.  The peak clad
temperature limit of 2200 F specified in 10 CFR 50.46.b.1 would be
exceeded.


Attached is a more detailed description of the concern, including a
technical justification for the position that no safety concern exists. 
Further analytical work will more formally document this conclusion.  A
plan for resolution of this concern is currently being discussed with the
B&W operating plant owners for their approval.  BWNT believes it is
prudent to inform the NRC at this time on the nature of the concern.  All
analyses are expected to be completed by August 1, 1995.  A final report
will be submitted to the NRC at that time.  At the present time, it has
not been determined that this matter is a reportable item under 10 CFR
21.

If you have any questions concerning this matter, please contact the
undersigned at 804-832-2817, or you may contact Mr.  Robert Schomaker at
804-832-2917.

                                   Very Truly Yours,

                                   J. H. Taylor, Manager
                                   Licensing Services

cc:  R. C.  Jones 
     Linda Gundrum


                Preliminary Safety Concern of Initial CFT
                   Conditions for LBLOCA ECCS Analysis

Introduction:

B&W Nuclear Technologies is in the process of evaluating a Preliminary
Safety Concern (PSC 5-94) related to the core flood tank (CFT) initial
conditions assumed in the large break loss-of-coolant accident (LBLOCA)
evaluation model (EM) calculations.  The LBLOCA calculations in question
are those performed according to the methods described in BAW-10104
Revision 5 for the 2-ft core power peak.  Since 1991, the initial CFT
conditions for all 2-ft LOCA limit calculations have used the minimum CFT
pressure and liquid inventory.  These inputs, which represent the lower
range of Technical Specification limits plus instrument uncertainties,
were considered to be the most limiting initial conditions for the 2-ft
LBLOCA PCT analysis.  A sensitivity study performed with a new
(unapproved) RELAPS/MOD2-based EM (BAW-10192), concluded that the most
limiting PCT is calculated when the maximum CFT liquid inventory and
minimum pressure are input as the initial conditions.  When these input
changes were analyzed with the current (approved) CRAFT2-based EM, the
PCT change was found to be greater than 50 F.  In fact, the LOCA linear
heat rate limits (LHRs) had to be reduced by 0.3 to 1.3 kW/ft, depending
upon the plant and fuel designs, to calculate PCTs that do not violate
the 2200 F acceptance criteria contained in 10 CFR 50.46.

Analyses were begun to redefine LOCA LHRs, given the effect of the more
limiting CFT inputs on results for the 2-ft elevation.  In the course of
these studies, it was further determined that the process by which fluid
enthalpies have been passed from the CRAFT2 blowdown calculation to the
THETA1B heatup analysis may have led to nonconservative conditions being
used during a short period of the THETA1B calculation.  Investigation has
shown this data transfer process to affect the fluid enthalpies used at
all core elevations, but the effect upon peak cladding temperature is
significant only at the 2-ft elevation.  Moreover, the sensitivity of the
PCT results at the 2-ft elevation that had been attributed solely to the
CFT inputs was found to include both

                                    1


the effects of the CFT inputs and the fluid enthalpy transfer.  At this
point, the CRAFT2-to-THETA1B data transfer has been revised to assure
that conservative fluid enthalpies are used in calculating clad surface
heat transfer, and a fresh set of analyses are underway to reestablish
the limiting core linear heat rates.

Background:

In 1991, the B&W Owners Group Analysis Committee directed BWNT to use the
most conservative set of CFT liquid level, pressure, and line resistance
inputs to envelope possible plant conditions (Technical Specification
limits plus instrument uncertainty) for all subsequent LBLOCA analyses. 
BWNT performed a sensitivity study for the 6-ft elevation and determined
that the maximum pressure, minimum liquid level, and minimum line
resistance produced the highest PCT.  A CFT line resistance study
performed for the 2-ft elevation showed that the minimum pressure and
maximum resistance produced the highest PCT for the 2-ft cases.  The
minimum inventory was used because the downcomer was not completely
filled before the CFT emptied, and this condition limits the core inlet
flooding rate.  These input changes were incorporated into the LOCA limit
analyses performed after 1991, for Mark-B9, Mark-B10(OL), and Mark-B9A.

In November 1994, a sensitivity study performed using the RELAP5/MOD2-
based evaluation model (EM) for the Mark-B11 fuel design revealed that
the maximum CFT liquid inventory would produce the highest PCT.  It was
found that the combination of minimum gas volume and minimum pressure
would result in the lowest CFT flow during the adiabatic heatup period. 
This low flow decreased the liquid remaining in the reactor vessel lower
plenum and extended the time period to refill the lower plenum.  The net
result was an increase in the adiabatic heatup period by approximately 20
percent.  For a PCT defined by the ruptured segment near the onset of
core recovery, the PCT was found to be significantly higher than would be
produced with the minimum liquid volume.  On this basis, Preliminary
Safety Concern (PSC) 5-94 was written, related to the anticipated PCT
variation associated with the input of the minimum versus

                                    2


the maximum CFT liquid inventory in the LBLOCA emergency core cooling
system (ECCS) analyses of record.

Discussion:

The CRAFT2-based EM LOCA LHR limit at the 2-ft elevation is ruptured-node
limited.  A 2-ft analysis was performed with CRAFT2 for the Mark-B10(OL)
fuel design.  This analysis confirmed the RELAP5/MOD2 results that
prompted PSC 594.  During the initial investigation of this PSC, the
generic 177-FA lowered loop (LL) LOCA LHR limit for the Mark-B9 fuel had
to be reduced by 1.3 kW/ft to accommodate the PCT increase in the 2-ft
analysis.  This reduction consisted of two components.  The adverse
effects from the CFT parameter changes were compounded by an observed
variation in the CRAFT2 core path inlet enthalpy supplied to THETA1-B for
the fuel pin thermal analysis.  The enthalpy was supplied to THETA1-B on
a coarse data frequency (one point every 0.5 seconds), and was subject to
high frequency oscillations that corresponded to the changes in
instantaneous flow direction.  The calculated PCT was found to be
sensitive to the enthalpy sampling because a large enthalpy difference
existed between the two nodes surrounding the 2-ft elevation.  For the
previous Mark-B9 LOCA analysis, the coarse data frequency had resulted in
the use of a nonconservative enthalpy, which led to improved heat removal
and a higher 2-ft LOCA LHR limit.  The previous Mark-B9A and Mark-B10(OL)
limits were reviewed with respect to this nonconservative enthalpy used
for the hot pin calculations.  The review concluded that the
nonconservatism had probably been contained in those results.  All cases
that could require reanalysis were identified, and a schedule for
resolution was prepared.

Several Mark-B8 and Mark-B8A 2-ft LOCA limits were included as candidates
for reanalysis used nominal CFT conditions as the licensing basis. 
Therefore, the 2-ft LOCA limits are not subject to the CFT input change,
however, the analysis is subject to the enthalpy data transfer
nonconservatism.  The adjustment related strictly to the Mark-B8 enthalpy
contribution is expected to be sufficiently small, such that no change is
needed to core operating limits.

                                    3


At the start of the reanalyses, other aspects were identified relative to
the core path inlet enthalpy calculated by CRAFT2.  In addition to the
high frequency oscillations, it was observed that the enthalpy could also
be skewed in the nonconservative direction by a nonhomogeneous treatment
of the core nodes following total dryout with subsequent return to two-
phase conditions.  The LBLOCA blowdown model is constrained to
homogeneous flows calculated by homogeneous node conditions.  So long as
the nodes remain continuously two-phase, the homogeneous conditions are
correctly calculated.  After dryout and return to two-phase conditions,
however, the homogeneous condition is not met.  Under these conditions,
CRAFT2 allows the node to separate the steam and liquid phases.  Figure 1
shows the code model with the separate interpretations of the nodal
conditions.  Because of the flow path connections, this configuration can
artificially cause the inlet flow path enthalpy to be lower than the
nodal homogeneous enthalpy.  Since this enthalpy is transferred to
THETA1-B, it can cause nonconservative conditions to be used for the hot
pin analysis.

This nonhomogeneous behavior was found to occur in all LBLOCA analyses,
regardless of the peak power elevation.  The timing and duration of the
condition varied as a function of core height.  This discovery forced an
investigation into the cause and led to an expanded scope of review to
validate all LOCA LHR limits.  Approximately two weeks were needed to
evaluate the cause and effects of these conditions.  CRAFT2 input model
changes were made to preclude the possibility of the nonhomogeneous
condition.  A temporary CRAFT2 code version was created to force the flow
path conditions to be consistent with the homogeneous nodal conditions. 
This work concluded that no change was required in the 4-ft through 10-ft
elevation LOCA limits.  The PCTs calculated for these elevations may
increase, but the increase is less than 20 F.  All 2-ft elevation cases,
however, can have additional nonconservative heat removal due to this
lower enthalpy supplied to THETA1-B.

The work performed to date supports the validity of a simple change in
the enthalpy data transfer to the THETA1-B analysis, without input model
changes or a new CRAFT2 code version.  The current analyses will continue
to calculate the CRAFT2 blowdown transient without any changes.  The
inlet enthalpy from

                                    4


the CRAFT2 analysis will be adjusted before the input is supplied to
THETA1-B.  The inlet enthalpy will be conservatively set to envelope the
upstream homogeneous nodal enthalpy based on the filtered flow direction. 
THETA1-B will be run with this conservative enthalpy to determine a PCT.

After the initial 2-ft LHR limit reduction estimates were available, BWNT
performed preliminary evaluations of the current fuel cycles (except the
Oconee Units) to determine if the operating limits remained valid given a
1.3 kW/ft decrease in the 2-ft Mark-B9 and Mark-B9A LOCA LHR limits.  The
preliminary results indicated that no changes to the operating limits
were needed for the present fuel cycles.  Duke Power Company performs the
power distribution analyses for the Oconee units, and preliminary review
by Duke concluded that the Oconee operating limits will be more
restrictive.  To minimize the effect on the Oconee units, additional
analyses using less restrictive, plant-specific CFT inputs have been
included as a part of the evaluation scope.  The expected reduction in
the 2-ft LOCA limit is expected to be approximately 0.3 kW/ft, which will
result in a very small change in the core operating limits.

Evaluation Schedule and Plant Operation:

The LBLOCA analyses required to determine the new LHR limits and their
impact on operating limits are scheduled to be complete by August 1.  A
final report will be sent to the NRC upon completion of all of this work. 
In the interim, plant operation is justified based on the preliminary
assessment that the 2-ft LOCA LHR limit, with a 1.3 kW/ft reduction for
Mark-B9 and Mark-B9A fuel types, will not lead to changes in the current
core operating limits for any plants that BWNT performs the power
distribution analyses.  The operating limits will prevent the core
maximum LHR from reaching a power level that could violate the 10 CFR
50.46 criteria.  Duke Power Company performs the power distribution
analyses for the Oconee Units, and plant specific analyses have
determined that operation of the Oconee Units is justified with
implementation of more restrictive operating limits.  At this time in the
evaluation, these bases support the expectation that the subject PSC and

                                    5


adjustment of enthalpies applied in the heatup analyses will not
represent a substantial safety concern.

Summary and Conditions:

The adjustment of the traditional CFT liquid inventory input assumption
and nonconservative enthalpy specified to the hot pin thermal analysis
will produce changes in the LBLOCA peak cladding temperature in excess of
50 F.  Reductions in the LOCA LHR limits of 0.3 (Oconee specific) to 1.3
kW/ft (Generic 177-FA RL and LL) are required to continue to meet the 10
CFR 50.46 acceptance criteria of 2200 F.  The majority of the reduction 
is related to the CFT inventory input and its effect upon the adiabatic
heatup time.  The nonconservative enthalpy used in the hot pin analysis
can cause PCT changes that are difficult to quantify without separate
analyses isolating the coarse data frequency from the nonhomogeneous
effects.  The most conservative CFT liquid inventory input will be used
in analyses that conservatively adjust the enthalpy transferred from the
CRAFT2 results to THETA1-B for the hot pin analysis.  The new 2-ft LOCA
LHR limits will be used in power distribution analyses to determine if
any changes are needed to the core operating limits.  The required
analyses are in progress.  The results of the LOCA analyses, including
any potential changes to core operating limits, are scheduled to be
complete by August of this year.

                                    6


Figure 1. "CRAFT2 Core Noding Representation" omitted.


POWER REACTOR                           EVENT NUMBER: 28284

FACILITY: DAVIS BESSE    REGION: 3      NOTIFICATION DATE:  01/26/95
UNIT:  [1] [ ] [ ]       STATE : OH     NOTIFICATION TIME:  15:22 [ET]
RX TYPE: [1] B&W-R-LP                   EVENT DATE:         01/26/95
                                        EVENT TIME:         15:16 [EST]
NRC NOTIFIED BY: MELSSEN                LAST UPDATE DATE:   01/26/95
HQ OPS OFFICER: CHAUNCEY GOULD 
                                                  NOTIFICATIONS
EMERGENCY CLASS: NOT APPLICABLE 
  CFR SECTION:                          MARK RING                RDO 
(illegible) 50.72(b)(1)(ii)(B) OUTSIDE DESIGN BASIS

--------------------------------------------- ---------------------------
UNIT  SCRAM CODE  RX CRIT  INIT PWR  INIT RX MODE  CURR PWR CURR RX
MODE 
 L        N         Y         100   POWER OPERATION  100  POWER OPERATION

                               EVENT TEXT

THE ALLOWABLE LINEAR HEAT RATE LIMITS FOR THE LOCA ANALYSES ARE
NON-
CONSERVATIVE AT THE 2 ft.  CORE ELEVATION (SEE RELATED EVENT 28286).

THE FOLLOWING IS THE TEXT OF A FACSIMILE SENT BY THE LICENSEE:
"THE LOCA ANALYSIS FOR DAVIS-BESSE PERFORMED BY B&W NUCLEAR
TECHNOLOGIES
(BWNT) DETERMINES THE ALLOWABLE LINEAR HEAT RATE LIMITS AS
PRESENTED IN
THE (illegible) OPERATING LIMITS REPORT (COLR).  BWNT HAS DETERMINED THAT
ANALYTICAL RESULTS FOR THE 2 ft. CORE ELEVATION ARE NON-CONSERVATIVE. 
THE NON-CONSERVATISM IS LINKED TO OPERATION OF THE CORE FLOOD TANKS
AT A
MAXIMUM PERMISSIBLE LEVEL AND A MINIMUM PERMISSIBLE PRESSURE.  THIS
RESULTS IN THE MOST RAPID CORE FLOOD TANK PRESSURE DECREASE AND THE
LOWEST RATE OF CORE FLOOD INJECTION.  ONLY THE 2 ft. CORE ELEVATION IS
SENSITIVE TO THE RESULTS OF REDUCTION IN FLOW RATE.

ADDITIONALLY, THERE IS A COMPUTER CODE INTERFACE CONCERN BETWEEN
THE
CRAFT2 AND THETA1-B COMPUTER CODES USED IN THE LOCA ANALYSIS, WHICH
ARE
PART OF THE NRC APPROVED 10 CFR APPENDIX K LOCA MODEL.  THE
COMBINATION
OF NODING, (illegible) DIRECTION CHANGES DURING REACTOR VESSEL
BLOWDOWN,
AND A 0.5 sec. CODE OUTPUT SAMPLING CAN RESULT IN NON-CONSERVATIVE
FLUID
ENTHALPY BEING INPUT TO THE THETA1-B COMPUTER CODE.  BASED UPON
INFORMATION OBTAINED FROM BWNT JANUARY 26, 1995 THE COMBINED
EFFECTS WILL
RESULT IN A CALCULATED PEAK ADDING TEMPERATURE (PCT) CHANGE OF
OVER 50
1/2F AT THE 2 ft. ELEVATION.  THIS WILL ALSO CAUSE PCT TO EXCEED THE 2200
1/2F LIMIT FOR THE 2 ft. ELEVATION AS STIPULATED IN 10CFR50.46(b) (1). 
HOWEVER, A REDUCTION IN THE 2 ft. (illegible) LINEAR HEAT RATE OF 1.5
kw/ft IS SUFFICIENT TO MAINTAIN PCT WITHIN ESTABLISHED LIMITS.

THE B&W FUEL COMPANY HAS REVIEWED, THE DAVIS-BESSE CORE OPERATING
LIMITS
FOR THE PRESENT CYCLE AND HAS CONCLUDED THAT POWER TILT, POWER
IMBALANCE
AND CONTROL ROD INSERTION LIMITS AS PROVIDED IN THE TECHNICAL
SPECIFICATIONS ARE NOT ADVERSELY IMPACTED.  HOWEVER, COLR
ALLOWABLE
LINEAR HEAT RATES ARE 


ADVERSELY AFFECTED AND WILL NEED TO BE REVISED.

WHILE TECHNICAL SPECIFICATION LIMITS WILL CONTINUE TO BE ADEQUATE
DURING
NORMAL OPERATION, THE COLR LIMITS HAVE A POTENTIAL IMPACT ON PLANT
OPERATION DURING UPSET CONDITIONS (e.g., DROPPED CONTROL RODS) WHICH
AFFECT CORE FLUX DISTRIBUTION.  IN THESE CASES THE COLR LIMITS ARE
ADMINISTRATIVELY COMPARED TO MEASURED CORE POWER DISTRIBUTION. 
THE PLANT
SOFTWARE WHICH IS USED TO MAKE THIS COMPARISON HAS BEEN REVISED TO
IMPOSE
A 1.5 kw/ft PENALTY AT THE AFFECTED ELEVATIONS.  THIS WOULD POTENTIALLY
REQUIRE A GREATER CORE POWER REDUCTION DURING UPSET CONDITIONS. 
THIS
ACTION WILL ENSURE THAT IF A LOCA SHOULD OCCUR DURING AN UPSET
CONDITION,
PCT WILL REMAIN WITHIN ACCEPTABLE LIMITS." 
END OF TEXT

THE LICENSEE HAS INFORMED THE RESIDENT INSPECTOR.


POWER REACTOR                           EVENT NUMBER: 28286

FACILITY: OCONEE         REGION: 2      NOTIFICATION DATE:  01/26/95
UNIT:  [1] [2] [3]       STATE : SC     NOTIFICATION TIME:  17:50 [ET]
RX TYPE: [1] B&W-L-LP [2] B&W-L-LP,     EVENT DATE:         01/26/95
         [3] B&W-L-LP                   EVENT TIME:         17:00 [EST]
NRC NOTIFIED BY: BURCHFIELD             LAST UPDATE DATE:   01/26/95
HQ OPS OFFICER:  TIM McGINTY
                                                  NOTIFICATIONS
EMERGENCY CLASS: NOT APPLICABLE 
  10 CFR SECTION:                       GEORGE AL BELISLE             RDO
(illegible) 50.72(b)(1)(ii)(B) 
OUTSIDE DESIGN BASIS

UNIT  SCRAM CODE  RX CRIT  INIT PWR  INIT RX MODE  CURR PWR CURR RX
MODE 
 1        N         Y         100   POWER OPERATION  100  POWER OPERATION 
 2        N         Y         100   POWER OPERATION  100  POWER OPERATION
 3        N         Y         100   POWER OPERATION  100  POWER OPERATION

                               EVENT TEXT
-------------------------------------------------------------------------
THE ALLOWABLE LINEAR HEAT RATE LIMITS FOR THE LOCA ANALYSES ARE
NON-
CONSERVATIVE AT THE 2 ft. CORE ELEVATION (SEE RELATED EVENT 28284).

THE LICENSEE PROVIDED THE FOLLOWING REPORT BY FACSIMILE (SEE RELATED
EVENT 28284):
"ON JANUARY 26, 1995 AT 5:00 PM, OCONEE NUCLEAR STATION (ONS) RECEIVED A
FAXED LETTER FROM B&W NUCLEAR TECHNOLOGIES (BWNT) WHICH INDICATES
THAT AN
ERROR EXISTS IN THE ECCS EVALUATION MODEL.  (Illegible) ERROR DEALS WITH
THE (illegible) INITIAL CFT (CORE FLOOD TANK) LEVEL (illegible) THE
MANNER IN WHICH ENTHALPY IS TRANSFERRED BETWEEN 'EM' CODES. BWNT
HAS
ESTIMATED THAT A REDUCTION IN THE LOCA LIMIT AT THE 2 FOOT ELEVATION
OF
UP TO 1.3 kw/ft MAY (illegible) NECESSARY TO ASSURE THAT THE 2200 1/2F
PEAK CLADDING ACCEPTANCE CRITERION 10CFR50.46(b)(1) REMAINS SATISFIED. 
THIS ESTIMATE IS BASED ON A GENERIC (illegible) ANALYSIS FOR THE LOWERED-
LOOP B&W PLANTS.

DUKE POWER HAS REQUESTED THAT BWNT PERFORM A PLANT SPECIFIC
ANALYSIS FOR
(illegible).  PRELIMINARY ESTIMATES INDICATE THAT THE LOCA LINEAR HEAT
RATE REDUCTION FROM THIS PLANT SPECIFIC ANALYSIS SHOULD BE IN THE 0.3
kw/ft (illegible).  DUKE POWER'S NUCLEAR ENGINEERING STAFF WILL USE A
BOUNDING REDUCTION IN THE LHR (LINEAR HEAT RATE) LIMIT OF 1.3 kw/ft TO
CONSERVATIVELY ESTABLISH NEW OPERATING LIMITS ON AXIAL IMBALANCE
FOR ALL
(illegible) UNITS.  THIS ACTION WILL ENSURE THAT THE ACCEPTANCE CRITERIA
OF 10CFR50.46 WILL REMAIN SATISFIED IN THE EVENT THAT, WHEN THE 
BWNT ANALYSES (illegible) COMPLETED, A REDUCTION IN THE LHR LIMIT AT THE
2 FOOT ELEVATION ACTUALLY OCCURS. 
END OF TEXT

THE LICENSEE HAS INFORMED THE RESIDENT INSPECTOR.
-------------------------------------------------------------------------
*** END OF DOCUMENT ***

Page Last Reviewed/Updated Thursday, March 29, 2012