ML20211K768

Evaluation of Fuel Pin Failure Timing in Boiling Water Reactors

ML20211K768
Person / Time
Site:	Grand Gulf
Issue date:	07/30/1999
From:	Knudson D, Schultz R IDAHO NATIONAL ENGINEERING & ENVIRONMENTAL LABORATORY
To:	NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
Shared Package
ML20211K719	List: ML20211K717 ML20211K749 ML20211K768
References
CON-FIN-W-6095 NUDOCS 9909080065
Download: ML20211K768 (14)

Text

'

EVALUATION OF FUEL PIN FAILURE TIMING IN BOILING WATER REACTORS D.L.Knudson R. R. Schultz July 30,1999 Idaho National Engineering and Environmental Laboratory Idaho Falls, ID 8341S Prepared for the U. S. Nuclear Regulatory Commission Office of Nuclear Reactor Research Washington, D. C. 20555 Under DOE idaho Operations Office Contract DE-AC07-94lD13223 Job Code Number W6095 DR DO K 05 0 16 P PDR

SUMMARY

The Boiling Water Reactor Owners Group (BWROG) recently completed analyses to determine the minimum time from reactor accident initiation to the first fuel pin cladding failure in a boiling water reac-tor (BWR). The BWROG submitted their results to the Nuclear Regulatory Commis ion (NRC) in an effort to further the development of NUREG-1465, which provides a basis for a revised and more realistic source term for light water reactors (LWRs). The NRC then sponsored this report to evaluate the BWROG submittal, which predicted that the first BWR fuel pin failure could occur no earlier than 121 s after acci-dent initiation. .

]

i The BWROG submittal was based on a set of the most limiting conditions (relative to the timing of i fuel pin failure). Specifically, the BWROG approach was based on an analysis of the most limiting design basis accident (DBA), the most limiting fuel design, and the most limiting plant geometry. The basic  !

approach used in this evaluation is identical to that BWROG approach. Specifically, two DBLOCA tran- 1 sient calculations were completed using SCDAP/RELAPS and FRAPCON-3 computer codes to determine the minimum time from reactor accident initiation to the first fuel pin cladding failure. Those calculations 1 were completed for conditions near the beginning of life (BOL) and conditions at the end of life (EOL) to cover fuel pin power, temperature, and pressure variations that affect the failure time. The results indicate {

that the high power /high stored energy conditions near the BOL will lead to an earlier failure than the EOL l I

high pin pressure conditions. Specifically, a minimum time to failure of 152 s was calculated for condi.

tions near the BOL, which is in close agreement with the BWROG result of 121 s. Relatively minor differ-ences between the results calculated here and the BWROG results are expected to be primarily due to differences in initial conditions and assumptions regarding transient progression. Consequently, it appears that the BWROG calculation of 121 s is a reasonable estimate for the earliest time for fuel pin failure in a l BWR. I i

iii i

i

CONTENTS SUMM ARY ... .. . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii 4

FIGURES . ... ... ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . .. vi TAB LES ..... .. . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vn

.1. INTRODUCTION..... . .. . . ... .. . . . . . . . . . . . . .. ...... . . . . . . . . . . . . . . . . . . . . . . . . .1

2. AN A LW1C A' APPR

- OA CH . . . .. .. . . . . . ... . . .. .. . . . . . . .... .. ..... .... ...... ..... . . . . . . . . . .. . . I 2.1. Accident Conditions.... .. ....... . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......2 2.2. FU eI & Si g n . .. ... .... . . .. ..... ... . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . .

2.3. Plant Geometry ......... ....... ... .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... .. ... ..... 2

3. SCDAP/RELAPS MODELING.. ..... . . .. ... ......... . .... . . . ........................................2 3 1

4. FRAPCON 3 MODELING.. . .. . .. .. . . . . . . . . . . . . . . _ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5

5. ESETS . ... .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

6. CONCLUSIONS .... . ...... . .. ... . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . 8

7. EENCES.... .. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......8 I

l i

i l

l 1

r i

c V

FIGURES

1. VY (three channel) core nodalization with RELAPS/ MOD 3.1 fuel hea't structures and core bypass.. ?

2. Core nodalization with SCDAP structures and interstitial flow areas in two channels.. ............4

3. Power / exposure relationship for gel 1 fuel.... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 6

4. Cladding hoop strain at the axial location of failure for the peak fuel pins.... ....... .. . .. . ... ........ ,,,.. g l

l I

a

?

l f

1 i

l Vi

l TABLES

1. SCDAP/RELAP5 core components for a two channel radial division. . . . .5

2. Timing of fuel pin cladding failure.. ... . . . . . . .. ... .. 7 i

t

)

1 l

l l

Vii

EVALUATION OF FUEL PIN FAILURE TIMING IN BOILING WATER REACTORS l

1. INTRODUCTION The minimum time from reactor accident initiation to the first fuel pin cladding failure is an important factor in determining the response necessary to prevent a radiological release. Analyses were previously completed to calculate the minimum time to fuel pin failure in pressurized water reactors (PWRs).3 Results from that work were incorporated into NUREG-1465, which provides a basis for a revised and more real-istic source term for light water reactors (LWRs).2 Similar analyses for boiling water reactors (BWRs) were not available at the time NUREG-1465 was published, although it was acknowledged that the time to BWR fuel pin failures was expected to be significantly longer than the time to PWR fuel pin failures. In the absence of BWR-specific analyses, NUREG-1465 promoted a conservative application of the PWR timing results.

The Boiling Water Reactor Owners Group (BWROG) recently completed BWR-specific fuel pin fail-ure analyses and submitted their results to the Nuclear Regulatory Commission (NRC)in an effort to fur-ther the development of NUREG-1465.3 The NRC then sponsored this report to evaluate the BWROG submittal, which predicted that the first BWR fuel in 7 failure could occur no earlier than 121 s after acci-dent initiation. This report documents analyses that were designed to independently check the BWROG results. The independent check relied on the use of SCDAP/RELAPS and FRAPCON-3 computer codesO to calculate the minimum time to BWR fuel pin failure. SCDAP/RELAP5 was used to simulate plant tran-sient thermal-hydraulics and core behavior including calculation of the fuel pin cladding failure. FRAP-CON-3 is a detailed fuel pin performance code with high burnup capabilities that was used to establish initial (steady-state) fuel pin conditions needed by SCDAP/RELAP5. Details associated with the approach that was used to complete this evaluation are outlined in Section 2. Pertinent aspects of SCDAP/RELAPS and FRAPCON 3 models are discussed in Sections 3 and 4, respectively. Results are given in Section 5 and associated conclusions are summarized in Section 6. I

2. ANALYTICAL APPROACH The time from reactor accident initiation to the first fuel pin failure is a function of the severity of the accident conditions, the characteristics of the fuel, and other plant specific design features. Instead of ana-lyzing all possible combinations of those factors, the BWROG recognized the_value of a fuel pin failure l analysis that could be applied to all BWRs. Therefore, the BWROG developed a " generic" analysis by incorporating a set of the most limiting conditions (relative to the timing of fuel pin failure)into their eval-uation. Specifically, the BWROG approach was based on an analysis of the most limiting design basis accident (DBA), the most limiting fuel design, and the most limiting plant geometry. As noted by the BWROG, the associated (limiting) results will be conservative relative to results that would be obtained for most plant-specific conditions. The basic approach used in this evaluation is identical to that BWROG l analytical approach. The most limiting conditions, as used in the BWROG analysis and in this evaluation, are discussed below.

I

1 1

I l

2.1. Accident Conditions )

l A review of current reactor Final Safety Analysis Reports (FSARs) was included m NUREG-1465 for I the purpose of identifying all DB As with a potential for fuel pin failure. Results from that review indicated that the shortest time from accident initiation to the first fuel pin failure was consistently associated with design basis loss of coolant accidents (DBLOCAs). For that reason, a DBLOCA was assumed without emergency core cooling system (ECCS) injection. A BWR DBLOCA consists of a double-ended break in the suction side of one of two recirculation lines, which represents a limiting condition because break flow is maximized. Analysis of a DBLOCA without ECCS injection will lead to early core uncovery and heatup, and the potential for a corresponding early failure of the fuel pin cladding.

2.2. Fuel Design The most limiting fuel design consists of fuel with the highest peak linear heat generation rate (PLHGR), the highest stored energy, and the highest internal pressure. General Electric (GE) evaluated a number of different BWR fuel designs (including GE fuel types 8,9,10,11, and 12 along with Siemens 8x8 and 9x9 fuels) to identify the most limiting type.3 Results from that evaluation indicated that Gell fuel is the most limiting. There was no attempt to verify this result. Instead, this evaluation (like the BWROG analysis) was completed assuming a gel 1 fuel loading. (GElI bundles are configured in a 9x9 fuel pin array with two large centrally located water rods that occupy the space that would otherwise accommodate seven fuel pins.)

2.3. Plant Geometry The most limiting plant geometry includes the combination of the smallest reactor pressure vessel (RPV) and the largest recirculation line. That combination is limiting because the initial water inventory i decreases with RPV size and rate of depletion of that inventory increases with the size of the recirculation I line break. Early core uncovery associated with such a geometry will lead to early heatup and failure of the fuel pin cladding. Based on designs associated with currently operating BWRs, the most limiting combina-tion consists of a BWR with a RPV diameter of 205 in and a recirculation line diameter of 28 in.

The Vermont Yankee (VY) BWR has the specified limiting geometry and an existing RELAP5 MOD 3.16input deck was available. For that reason, the VY RELAPSMOD3.1 input deck was used as a basis for the SCDAP/RELAP5 model as discussed below.

l 1

3. SCDAP/RELAP5 MODELING l I

The SCD AP/RELAPS model used in this evaluation was based on an existing RELAP5 MOD 3.1 input deck for the VY BWR and Gell fuel specifications. The RELAP5 MOD 3.1 input deck and the fuel spec- l ifications have proprietary classifications. Both VY and GE agreed to allow use of their proprietary data in this evaluation only, provided that the data are not disclosed to any third party. Consequently, general information concerning the VY input deck and the gel 1 fuel loading are discussed in this report. How-ever, specific proprietary information is not given.

2  !

l I

i As a first step, the VY RELAP5/ MOD 3.1 input deck had to be converted to run under the urrent ver-sion of SCDAP/RELAP5. That conversion, which was necessary to invoke models capable of simulating fuel pin failure, basically focused on replacing the RELAP5/ MOD 3.1 core with a SCDAP/RELAP5 core.

The VY core consists of a total of 368 fuel bundles producing approximately 1633 MWt. Each VY bundle is configured in an 8x8 fuel pin array with one large centrally-located water rod that occupies the space that would otherwise accommodate four fuel pins. OE limiting specifications included a core producing 1880 MWt from a total of 484 fuel bundles. Each GElI bundle is configured in a 9x9 fuel pin array with j two large centrally-located water rods that occupy the space that would otherwise accommodate seven fuel pins. (The VY RPV geometry will apparently accept the larger gel I core.)

The VY RELAP5/ MOD 3.1 core model consisted of three heated channels and a bypass flow path as shown in Figure 1. The four hottest (highest powered) fuel bundles were loaded into Volume 162 (V162),

248 average bundles were loaded into Volume 142 (V142), and the remaining 116 peripheral fuel bundles were loaded into Volume 122 (V122). All interstitial flows and water rod flows were lumped into (bypass)

Volume 102 (V102).

l 9.

Y2%

n n h n JI69 J149 J129 J109 I

1

. I.

Vl62 V142 V122 V102 fuel pins f

I I , .

n JI il dl db Jl60 2 Jl40 2 J120-2 Jl001 33 " I J 140-1 1241

. i

!I '

,Vl60 ' V120 V100

! V140l ,. I a

J036 a

J034 o

1032 h I 737 - from CRGTs j

, J031 j

V012 l l

k, l

Figure 1. VY (three channel) core nodalization with RELAP5/ MOD 3.1 fuel heat structures and core bypass.

t

! 3 t )

v: ..

9 The three VY heated channels were collapsed into a two channel model in this evaluation as shown in Figure 2. That simplification was made because (1) three channels were not needed in order to calculate the minimum time from reactor accident initiation to the first fuel pin failure (which will always occur in a peak pin in a hot bundle) and (2) the supplied data did not include details needed to establish a three chan-nel model for a gel 1 ' core. Each channel consisted of a bundle flow path and an associated interstitial flow path. Volumes 162 (Vl62) and 164 (V164) represent hot bundle and hot interstitial flow paths, respec-

~ tively. Volumes 142 (V142) and 144 (V144) represent average bundle and average interstitial flow paths,

' respectively. Four hot (high powered) bundles and 480 average bundles were assumed.

T.

n a o. a 1169-1 1169-2 11491 J 149-2 N ., , .

.[~

.s 3

y.. 4 ,A

.g. .  %.

.g. 'L

.b ] '.

pipns wer r ds V162.kVid jVl42 V144

" -h channel boxes and control blades

-) (:- .

T L

y.

g. -  %

.g= 4  %

5 $  %.

. o a o o 1160-1 J150w2 1140-1 JI501 V160 Vl50 V140 o o o o J036 J035 J034 from CRGTs '~ V012 q,:

Figure 2. Core nodalization with SCDAP structures and interstitial flow areas in two channels.

' The individual components associated with the SCDAP/RELAP5 core model are listed in Table 1. As

- indicated, each channel included a number of fuel pin, water rod, and channel box / control blade compo-

- nents. GE specified radial peaking factors were applied to distinguish between peak fuel pins, average fuel

. pins in the hot bundles, and average fuel pins in the average bundles. A single peak fuel pin was assumed

.'in each hot bundle as indicated in the table.

4 1,

Table 1. SCDAP/RELAPS core components for a two channel radial division.

Components Radial Region Number of Bundles Chann Boxes /

Peak Pins Average Pins' Water Rodsb Hot channel 4 4 292 8 4/1 Av'erage channel 480 0 35,520 960 480/120

a. Average pins in the hot channel and average pins in the average channel differ in accordance with GE spectfied radial peaking factors.

b. Two water rods in each bundle occupy a total of seven fuel pin positions.

After the SCDAP/RELAP5 two channel core model was developed and incorporated with other VY RELAPS input, a steady state calculation was performed to determine hydrodynamic conditions. This cal-culation was needed to establish acceptable operating conditions for the GElI core, boundary conditions for FRAPCON-3 models (see Section 4), and initial hydrodynamic conditions for the DBLOCA transient.

A number ofinitial conditions and assumptions were required to complete this calculation including a core power of 1880 MWt.

• a RPV dome pressure of 1043 psia (which is approximately 17 psi above the VY RELAPS/ MOD 3.1 pressure),

• a RPV water level of 162 in above the top of the enriched fuel (which is consistent with the VY RELAP5/ MOD 3.1 level).

= a mass flow per pin equal to the VY RELAP5/ MOD 3.1 mass flow per pin,

= a core exit quality of 0.14 (which is consistent with the VY RELAP5/ MOD 3.1 model),

a feedwater and steam flows consistent with the core power, and a a feedwater temperature of 377.5 F(which is above the VY RELAP5/ MOD 3.1 feedwater temper-ature of 373.1 F).

Results from the SCDAP/RELAPS hydrodynamic steady state were used as boundary conditions for FRAPCON-3 calculations as outlined in Section 4. Resulting FRAPCON-3 fuel pin temperatures and pres-sures were then used to achieve a final SCDAP/RELAP5 steady state. Specifically, fuel pin thermal con-ductivities, gap conductances, and heliom inventories were iteratively adjusted until SCDAP/RELAP5 steady state fuel pin results agreed with FRAPCON-3 fuel pin temperatures and pressures. The resulting

' SCDAP/RELAP5 model was then used to complete transient calculations as discussed in Section 5.

4. FRAPCON-3 MODELING Three different FRAPCON 3 models were needed in order to initialize the SCDiP/RELAP5 model.

Specifically, models were needed for peak fuel pins, average fuel pins in the hot bundles, and average fuel pins in the average bundles, which is consistent with the SCDAP/RELAP5 core nodalization shown in Figure 2. In all cases, model input was primarily derived from GE specifications for gel 1 fuel. Because those specifications are proprietary, model inputs cannot be discussed. It is wonh noting, however, that GE 5

specifications for the axial power profile, radial peaking, PLHGR, and burnup were some of the most sig-nificant parameters.

Hydrodynamic boundary conditions (including coolant pressure, temperature, and mass flow) from an initial SCDAP/RELAPS steady state calculation were also used as input in the FRAPCON-3 models (see Section 3). FRAPCON-3 calculated fuel pin response as a function of time from the beginning of life (BOL) to the end of life (EOL). Fuel pin temperature and pressure results were then used to initialize the SCDAP/RELAPS model prior to any transient initiation as discussed below.

5. RESULTS Two DBLOCA transient calculations were completed using SCDAP/RELAPS with corresponding FRAPCON-3 core temperatures in order to determine the minimum time from reactor accident initiation t'o the first fuel pin cladding failure. In the first calculation, the transient was initiated near the BOL. In the second calculation, the transient was initiated at the EOL. Two calculations were needed because GE spec-ifications indicate that the subject fuel pins are operated at a constant power for a relatively short period of time. Thereafter, power decreases linearly as exposure increases as illustrated in Figure 3. Consequently, power and stored energy are high near the BOL while pin pressure is relatively low. Conversely, pin pres- 1 sure is high at the EOL while power and stored energy are relatively low. Completing two calculations {

allowed evaluation of the full range of fuel pin power, temperature, and pressure variations that affect the l failure time.

l

' knee' n

E i8 i 6

t 3

i Exposure (GWd/MRI)

Figure 3. Power / exposure relationship for Gell fuel.

6

c i;

l_'.-

Both calculations were completed using a common set of initial conditions and assumptions including l . all previously outlined steady state initial conditions and assumptions, l

' an assumed SCRAM decay power that was based on a normalization of the VY RELAPS/ MOD 3.1 -

SCRAM response, l

L an assumed feedwater flow coast. down over 4.5 s (which is consistent with the VY

_ RELAPS/ MOD 3.1 coastdown),

a an assumed main steam isolation valve (MSIV) closure in 6 s (which is consistent with the VY RELAPS/ MOD 3.1 MSIV closure),

an assumed failure that leaves the broken loop recirculation pump discharge valve open,

. an assumed constant containment (drywell) pressure of 16.5 psia, and

• an assumed rupture strain of 0.18.

The calculation near the BOL was actually completed for conditions associated with the power / expo-sure ' knee'shown in Figure 3. At that point, power and stored energy are at maximums. Furthermore, the pin pressure is as high as possible before power begins to drop, The calculation at EOL was completed for ,

conditions associated with the time that the GE specified bumup was actually achieved. The timing of fuel I pin cladding failure for these two calculations, relative to transient initiation, is given in Table 2. Those i results indicate that the minimum time from reactor accident initiation to the first fuel pin cladding failure is 152 s.

Table 2. Timing of fuel pin cladding failure.

Calculation Time of Fuel Pin Cladding Failure Relative to Transient Initiation (s) ]

Near BOL* - 152

@ EOL 252

a. Corresponding to conditions associated with the power / exposure ' knee' shown in Figure 3.

In both calculations, break initiation led to a rapid depressurization of the RPV and an early core uncovery. Thereafter, decay heat removal was reduced and fuel pin temperatures and pressures began to increase, which led to ballooning of the fuel rod cladding as indicated in Figure 4. Ballooning continued until the assumed rupture limit was reached. As clearly indicated in the figure, the high power /high stored energy conditions near the BOL resulted in an earlier failure than the EOL high pin pressure conditions. ,

l

' The cladding failure time of 152 s for conditions near the BOL appears to be in close agreement with I I

the BWROG result of 121 s, given the numerous differences that exist between the two analyses. Obvi-ously, the use of different computer codes, each with their own assumptions and limitations, will introduce some discrepancy However, it is believed that differences in initial conditions and assumptions regarding transient progression are also important. Although apparently reasonable initial conditions and assump- )

tions were made in this evaluation, it is not currently known how all of those assumptions compare with the treatment in the BWROG analysis, i

7 l

\

l 0.25 . .

, . . , l pin failure ,

0.20 -

h -

.E hoop-90) near BOL -

C O hoop-901 O EOL

{ 0.15 8

.c -

E g 0.10 - -

a o .

l 0.05 - -

0.00 W - '< ' '

O.0 100.0 200.0 300.0 l Time (s)

Figure 4. Cladding hoop strain at the axial location of failure for the peak fuel pins.

6. CONCLUSIONS Two DBLOCA transient calculations were completed using SCDAP/RELAP5 with corresponding FRAPCON-3 core temperatures in order to determine the minimum time from reactor accident initiation to the first fuel pin cladding failure. Those calculations were completed for conditions near the BOL and con-ditions at the EOL to cover fuel pin power, temperature, and pressure variations that affect the failure time of Gell fuel. The results indicate that the high power /high stored energy conditions near the BOL will lead to an earlier failure than the EOL high pin pressure conditions. Specifically, a minimum time to fail-ure of 152 s was calculated for conditions near the BOL, which is in close agreement with the BWROG result of 121 s. Relatively minor differences between the results calculated here and the BWROG results are expected to be primarily due to differences in initial conditions and assumptions regarding transient

. progression. Consequently, it appears that the BWROG calculation of 121 s is a reasonable estimate for the earliest time for fuel pin failure in a BWR.

i

7. REFERENCES

1. K. R. Jones, N. L. Wade, K. R. Katsma, L. J. Siefken, and M. Straka Timing Analysis of PWR Fuel Pin failures, NUREG/CR-5787, Idaho National Engineering and Environmental Laboratory, September 1992.

2. L. Soffer, S. B. Burson, C. M. Ferrell, R. Y. Lee, and J. N. Ridgely, Accident Source Tennsfor Light-Water Nuclear Power Plants, NUREG-1465, U. S. Nuclear Regulatory Commission, February 1995.

8

t l

3. W. K. Hughey (Entergy Operations, Inc.) letter to the U. S. Nuclear Regulatory Commission, "Sub-mittal of BWROG Report - Prediction of the Onset of Fission Gas Release from Fuel in Generic BWR Application of NUREG-1465 Source Terms for Grand Gulf Nuclear Station Rebaselining Study,"

GNRO 97/00034, May 6,1997.

1

4. SCDAPIRELAPS Development Team, SCDAP/RELAP5/ MOD 3.2 Code Manual, User's Guide and Input Manual. NUREG/CR-6150, Volume 3, Revision 1, Idaho National Engineering and Environ-mental Laboratory, July 1998.

5. G. A. Bema, C. E. Beyer, K. L. Davis, and D. D. Lanning, fRAPCON-3: A Computer Code for the Calculation of Steady-State, Thermal-Mechanical Behavior of Oxide Fuel Rods for High Burnup, NUREG/CR-6534, Volume 2, December 1997.

6. C. F. Slater et al., " Development and Developmental Assessment of RELAP5/ MOD 3.1," presented at the RELAP5 Iriternational Users Seminar, Boston, MA, July 6-9,1993.

9 o

l I

l 9

k_