ML20202G262

Topical Rept Evaluation Accepting Exxon Nuclear Co Large Break ECCS Evaluation Model,Exem/Pwr,For Ref in License Applications

ML20202G262
Person / Time
Issue date:	07/08/1986
From:	NRC
To:
Shared Package
ML20202G180	List: ML20202G175 ML20202G262
References
NUDOCS 8607150352
Download: ML20202G262 (33)

Text

.

SAFETY EVALUATION OF EXXON NUCLEAR COMPANY'S LARGE BREAK ECCS EVALUATION MODEL EXEM/PWR

1.0 INTRODUCTION

In References 1 through 15, Exxon Nuclear Company (ENC) provided a description of updates to its large break ECCS Evaluation Model (EM) for PWRs. This model, which is based upon the currently approved ENC WREM-IIA PWR ECCS EM, Reference 16, is referred to as EXEM/PWR.

The model updates of EXEM/PWR are shown on Table 1.

This safety evaluation report presents the staff's findings on these topical reports. Specifically, this safety evaluation report addresses the compliance of EXEM/PWR to the requirements of Appendix K to 10 CFR 50. Each of the model changes is discussed separately below.

2.0 EVALUATION 2.1 Fuel Rod Model-R0DEX2 Code The RODEX2 code is documented in Reference 7. The RODEX2 code is based upon the previously approved GAPEX code, Reference 17. As part of the EXEM/PWR model, ENC uses the RODEX2 code to provide the initial fuel stored energy and fuel rod internal pressures utilized as inputs to various portions of the evaluation model.

8607150352 860708 PDR TOPRP EMVEXXN C PDR

. - The staff has previously reviewed and approved the R0DEX2 code for LOCA applications. The evaluation of this code is contained in Reference 18. Specifically, the staff found that the RODEX2 code satisfies the requirements of Appendix K, section I.A.1.

2.2 Clad Swelling and Rupture Model In Referenge 6, ENC proposed a revised clad swelling and rupture model. This model, which includes the data base of NUREG-0630, Reference 19, is used in the RELAP4 and T00DEE2 codes.

The staff has previously reviewed this model for compliance with section I.B. of Appendix K. As documented in Reference 20, the staff found this revised model to meet those requirements.

2.3 Modified Fuel Rod Model Reference 3 describes updates made to the RELAP4-EM code to make its fuel models consistent with the approved RODEX2 fuel performance code.

The RELAP4-EM code is used by ENC in the PWR blowdown analysis. The updates include gap conductance, internal rod pressure, fuel conduc-tivity, and radial power distribution.

e

, . 2.3.1 Fuel Cladding Gap Conductance The fuel cladding gap conductance consists of three , components, gas conduction, fuel cladding solid contact, and radiation, for steady-state heat transfer. The formulation of gap conductance is identical to those approved for RODEX2. Therefore, the steady-state gap con-ductance is acceptable for RELAP4-EM.

For transient heat transfer, the RODEX2 code does not have any transient gap conductance model. However, ENC provides correlations among steady-state gap conductance, transient linear heat generation rate, and fuel cladding contact pressure, which enable the model to extend the calculation into the blowdown regime. ENC has also done a sensitivity study to demonstrate that the correlations result in a conservative fuel stored energy. Since the blowdown duration is relatively short and the transient heat transfer has been shown conservative, the staff concludes that the gap conductance model is applicable for transient calculations performed with RELAP4-EM.

2.3.2 Rod Internal Pressure The rod pressure expression is identical to the one in the approved RODEX2 code. Therefore, this expression is acceptable for use during blowdown transient in RELAP4-EM.

- - . . .. ,. , ~ . .

_m, .-

. l

.. To calculate plenum temperature, the analysis uses a plenum temperature l

slightly higher than the coolant temperature during steady-state as in the approved RODEX2 code, and a plenum temperature evaluated based on

)

the approved BULGEX code (Reference 21) for the transient. The staff thus concludes that the plenum temperature calculations are acceptable for RELAP4-EM.

2.3.3 Fuel Conductivity and Radial Power Distribution The fuel conductivity and radial power distribution are entirely the same as in the RODEX2 code. Therefore, these are acceptable for use in RELAP4-EM.

2.4 REFLEX Leakage Flow Model -

The currently approved ENC REFLEX code, Reference 16, which is used for calculating the reflooding phase of a LOCA, does not consider leakage flow paths from the upper plenum to the downcomer. ENC has proposed a modification to the REFLEX noding to account for this leakage path. ENC has stated that this model will be utilized only when the leakage flow path can be characterized and justified. Sensi-tivity studies, documented in Reference 1, have been performed and show that this model change results in only a small reduction, j approximately 20*F, in peak cladding temperature.

., Inclusion of a leakage flow' path in REFLEX will result in a more representative model of the plant configuration. In fact, the leakage flow path is already included in the blowdown model. Thus, the staff finds this model change to be acceptable provided that the leakage flow path can be well characterized. Justification for use of this model change must be provided as part of the plant application studies performed to demonstrate compliance with 10 CFR 50.46.

2.5 Split Break Model Currently the REFLEX code only simulates a guillotine break config-uration with a discharge coefficient of 1.0. This assumption is conservative for split breaks and guillotine breaks with discharge coefficients less than 1.0. As part of EXEM/PWR, the REFLEX code has been modified to allow modelling of split breaks and guillotine breaks with smaller discharge coefficients.

l 1

For modelling of split breaks, the REFLEX code has been modified to allow the fluid streams from the downcomer and steam generators to mix before leaving the break. A junction is then used to simulate the break path to containment.

I

. - 1 Double-ended guillotine breaks with smaller discharge coefficients are simulated with the current REFLEX noding scheme. However, to account for the smaller discharge coefficient, an equivalent K-factor is used to simulate the increased break res~istance.

The staff has reviewed these model changes and the assumptions utilized and finds them acceptable.

2.6 REFLEX Core Outlet Enthalpy Model l

The currently approved REFLEX model uses a constant value for the core exit enthalpy. The core exit enthalpy used is determined at the upper plenum pressure and the fluid temperature corresponding to the steam generator secondary side saturation temperature. The core exit en-thalpy model has been upgraded such that fluid enthalpy is calculated based upon an energy balance performed for the core.

The revised core outlet enthalpy model accounts for energy added to the fluid below the quench front, stored energy release as the quench front progresses, and energy added to the fluid above the quench front. To demonstrate the appropriateness of the model, ENC performed benchmarks of FLECHT tests 34711, 34610, and 31922, Reference 22. These benchmarks showed good agreement to the data.

Based upon the benchmarks performed, and a detailed review of the equa-

, tions utilized, the staff has concluded that this model is acceptable.

I

2.7 Steam Cooling Model Section I.D.5 of Appendix K to 10 CFR 50 requires that a steam cooling model be utilized to predict heat transfer coefficients when flooding

. rates fall below one inch per second. In addition, the steam cooling model must take into account the effect of flow blockage relative to both local steam flow and heat transfer.

Exxon developed, as part of their currently approved ENC WREM-IIA PWR ECCS evaluation model, a steam cooling model which fully complied

with these requirements. However, recent experimental data in Ref-erences 23 and 24 have shown that the currently approved Exxon steam cooling model is overly conservative. As a result, Exxon developed, and submitted as part of EXEM/PWR, a revised steam cooling model.

The revised steam cooling model calculates an equivalent steam flow for use in the T00DEE2 (Reference 25) energy solution which assures that superheated steam exits the core. This flow rate includes the effect of blockage by using the currently approved flow divergence model of the ENC WREM-IIA PWR ECCS evaluation model.

The rod surface heat transfer coefficients are calculated by the following method. First, local unblocked heat transfer coefficients l l

are calculated using an appropriate reflood heat transfer correlation i

. for the fuel modeled, j l

. . - . . - . - . -, ,, ,.,_._v.

The local heat transfer coefficients are then modified to account for the effect of blockage on mass flux and hydraulic diameter. In addition, the heat transfer coefficients are adjusted to account for the effect of increased turbulence and breakup of entrained liquid droplets downstream of the blockage. The net effect of these modi-fications is a decrease in heat transfer downstream of the flow blockage relative to that which would be obtained in an unblocked core. Calculations performed by Exxon with the revised steam cooling model indicate that peak cladding temperatures are approximately I 15 F higher relative to that which would be obtained using the un-blocked ENC-2 FLECHT coefficients.

The staff has reviewed the revised steam cooling model and finds it acceptable. Recent experimental data in Reference 23 and 24, obtained with flooding rates below one inch per second, indicate that the effect of blockage is to enhance heat transfer, relative to an unblocked fuel assembly, downstream of the blockage plane. Since the revised EXXON steam cooling model predicts decreased heat transfer, the staff finds that the effect of flow blockage on local steam flow and heat transfer has been treated conservatively. Thus, the revised steam cooling model fully meets the requirements of Section I.D.5 of Appendix K to 10 CFR 50.

2.8 f!evised Radiative Heat Transfer Model The currently approved ENC radiative heat transfer model is used during the non-convective refill period prior to the start of core reflood and r ge-wr--- w- -

, -.i.- +-- -- + - - - - - * +

• s.-, - --- + - -

.. is described in References 26 and 27. Heat transfer during this period is primarily by radiation from the hot rod to surrounding cooler rods and guide tubes witnin the fuel assembly. The ENC radiative heat transfer model is applied in the T00DEE2 code. The model effectively lumps the surrounding fuel rods and guide tubes into an equivalent heat sink in order to determine the radiative heat loss from the hot rod to its surrounding on a transient basis. Verification of the ENC I

J model was provided by comparison of T00DEE2 and HUXY, Reference 28, code predigtions. HUXY is the approved ENC heating code for BWRs and explicitly considers rod-to-rod radiative heat transfer on a transient basis.

The approved model was verified for use on Westinghouse plants with 15x15 and 14x14 fuel lattices. As part of the EXEM/PWR updates, additional verification for the model was provided to extend its applicationto14x14CombustionEngineering(CE)and17x17 Westinghouse (W) plant fuel lattice geometrics. Detailed information on the veri-fication performed is provided in Reference 4. As part of the additional information, the methodology employed to calculate the required T00DEE2 input was provided.

ENC provided the results of benchmarks of the T00DEE2 comparisons to the HUXY code for 14x14 CE,15x15 W, and 17x17 W fuel lattice geo-metrics. The results showed that the T00DEE2 codes conservatively overpredicted hot rod temperatures, as compared to the HUXY calcu-lations, for all the fuel lattice geometries examined, l

'- The staff reviewed the calculational techniques employed by ENC to calculate the T00DEE2 inputs and found the approach utilized to be reasonable. Additionally, ENC has demonstrated that these input parameters, when utilized in the T00DEE2 code, results in conservative cladding temperatures with respect to the HUXY code. As a result, the staff concludes that the proposed extension of the radiation heat transfer model in T00DEE2 to other fuel lattice geometrics is acceptable.

2.9 Revised Carryout R' ate Fraction (CRF) and Heat Transfer Coefficient Correlations ENC proposed several CRF and heat transfer coefficient correlations as part of EXEM/PWR while the evaluation model was undergoing staff review. The initial EXEM/PWR submittal of Reference 1 provided new CRF and heat transfer coefficient correlations for 17x17 fuel rods. These correlations were based upon the FLECHT SEASET data of Reference 22.

Subsequently, ENC provided, in Reference 2, revised correlations in order to expand their applicability to both the 15x15 and 17x17 fuel rod designs. These correlations were based upon the FLECHT SEASET data of Reference 22 and the FLECHT data of References 29 and 30. As a result of staff questions concerning the adequacy of these correlations, ENC recorrelated the FLECHT 15x15 and 17x17 data and in Reference 31 proposed new correlations.

In addition to the correlations developed from the FLECHT data, ENC

. obtained reflood heat transfer data for their fuel. design using their

i

. - 11 _

Fuel Cooling Test Facility (FCTF). These data, and the resultant cor-relations, were provided in References 8 and 12. In Reference 13, in l

response to staff questions, ENC stated that only the FCTF-based cor- 1 relations will be used for future licensing analyses performed with the EXEM/PWR model.

Accordingly, the remainder of the evaluation provided in this section will only discuss the staff's conclusions with respect to the FCTF-based correlations and ENC methods for applying these correlations to i- reactor conditions. The staff has noted, in Reference 4, several concerns with the ENC correlations of Reference 1 and 2. Also, as a result of ENC's response, Reference 13, to staff questions, the staff has not reviewed the correlations of Reference 31. Thus, as part of the staff's approval of the EXEM/PWR evaluation model, the staff prohibits the use of the correlations in References 1, 2 and 31.

The remainder of this section is organized as follows. Section 2.9.1 describes the staff's evaluation of the FCTF and the resultant data.

The correlations developed from this data are evaluated in Section 2.9.2.

The methods utilized by ENC to apply these correlations to reactor simulations are described and discussed in Section 2.9.3.

2.9.1 FCTF Description and Test Data ENC performed a series of reflood heat transfer tests at its FCTF. The i

purpose of this testing was to obtain a reflood heat transfer data base

. for ENC's 17x17 PWR fuel design, including the effccts of production spacers and the upper tie plate, during simulated LOCA conditions. A description of the FCTF testing program and the resultant data is pro-vided in References 8 through 11.

The FCTF is a closed system which provides known, definitive boundary conditions for the test fuel assembly. Key features include a steam loop, a reflood injection system, a water treatment system, a steam supply system, an AC power supply and a prccess control / data acqui-sition system. Mass flow, temperature, and process instrumentation ,

is used to measure all mass and energy flows into and out of the test vessel.

The test assembly utilized consisted of fifty-seven (57) electrically heated rods, four (4) simulated guide tubes, eight (8) grid spacers, and an upper tie plate. The heater rods utilized stainless steel cladding.

]

The pitch of the heating coil within the rod was varied axially to give a chopped cosine power distribution with a 1.66 peak-to-average power ratio. The test assembly simulates one-fourth of a full scale ENC 17x17 fuel rod. The test assembly wa's mounted in a cylindrical low mass hcusing. Thermocouples located in the heater rods measured rod temperatures during the test.

In reviewing the FCTF, the staff concentrated its examination on the provisions taken by ENC to eliminate problems experienced in the early

. FLECHT tiests. Specifically, these concerns were related to fallback of liquid from the upper plenum and vessel housing effects.

l l

s The first concern relates to the deentrainment of droplets in the upper plenum and the subsequent fallback of liquid into the test assembly. While such deentrainment and fallback is expected in a PWR during a LOCA, the small scale involved in tests such as those performed in FCTF and FLECHT prohibit proper modelling of these effects. Thus, the FLECHT experiments were designed to minimize the accumulation of liquid in the upper plenum.

Within FCTF, a drain line was employed in the upper plenum to prevent accumulation of liquid on the upper tie plate. This drain line was located below the upper tie plate. ENC provided an example of the upper plenum delta pressure for one of the tests; minimal accumu-lation of liquid on the upper tie plate was shown. Additionally, ENC stated that visual observations with the upper plenum sight glasses during testing also indicated minimal water buildup. Based upon the physical location of the drain line and the ENC observations, the staff has concluded that adequate provisions were taken in FCTF to prevent fallback of liquid from the upper plenum.

There are two concerns related to vessel housing effects. First, since the housing temperature is colder than the heater rods, additional radiative heat transfer, which is not typical of reactor conditions, may occur during the experiments. In response to staff concerns, ENC provided additional information, Reference 9, which addressed this effect.

l

_ 14 _

In order to ensure that radiative heat transfer to the housing did not have a significant effect on the observed heat transfer coeffi-cients, ENC' excluded data from the outer two rows of test rods, which had a direct view to the vessel housing, in developing the heat

,, transfer coefficients. Additionally, ENC examined the radiative view factors for the rods correlated and concluded that the overall heat transfer coefficient for the hottest rods was affected by less than 2% due to radiation to the vessel housing. For the worst case rod in the-test bundle, radiation to the housing would have affected heat transfer by less than 11%.

Within Reference 11, ENC provided the results of FCTF tests wherein only the housing temperature was varied. While variations in peak cladding temperatures were seen in these tests, the relative differ-ences were less than 40 F and did not show a consistent pattern with respect to housing temperature. Based upon the FCTF test results, a review of the ENC heat transfer calculations, and the similarity of these results to those obtained in the FLECHT tests, the staff has con-cluded that the FCTF test results were not significantly affected by radiative heat transfer to the vessel housing.

During the earlier FLECHT tests, e.g., References 30 and 32, it was observed that energy elease from the vessel housing had a significant effect on quench front propagation and heat transfer coefficients versus time. Ideally, the bundle flow area for the experiment should match reactor bundle flow areas and the heat transfer to the fluid

from the vessel housing wall should be minimized. In this manner, fluid conditions representative of those expected in a reactor under similar reflood conditions would be achieved in the test bundle. The earlier FLECHT test bundles had excess flow area between the bundle and the vessel housing wall. To obtain appropriate fluid conditions, the earlier FLECHT tests heated the vessel housing such that the energy release from the housing approximates the energy addition per unit cell flow area that would be obtained by heater rods. This

~

method was found acceptable during the staff's review of the FLECHT experiments.

In the more recent FLECHT and FLECHT SEASET tests, steps were taken to minimize the impact of housing energy release. Specifically, inserts were used to minimize the excess flow area of the test bundle, and a low mass vessel housing was used to reduce energy release to the fluid.

In Reference 33, it was shown that these actions essentially eliminated the eff'ect of the vessel housing energy release on the test results.

For the FCTF, ENC also utilized inserts to reduce excess flow area and a low mass vessel housing. However, within Reference 10, ENC presented the results of FCTF tests which showed that the quench front propa-gation was affected by the vessel housing temperature. The staff requested that ENC provide further information on the effect of housing temperature; ENC's response is contained in Reference 11.

I

l -

I Within their response, ENC noted that they did not specify a housing temperature criteria for the FCTF tests. The response also noted that, like the earlier FLECHT test results, the housing temperature affected quench time and heat transfer coefficients versus time, but not the peak cladding temperature. To determine what housing temperature would have been necessary during the FCTF testing to provide appropriate l

energy release per unit flow area, ENC performed an energy balance on the vessel housing assuming a 3 foot quench elevation. The 3 foot quench elevation was utilized by ENC as it bounded the rod quench elevation at the time of peak cladding temperature. The results of these calculations were then compared to the actual initial housing temperature achieved during the FCTF tests. In approximately 12 of the 17 tests used for correlation development, the actual housing tem-peratures exceeded the " required" temperatures which were calculated.

Thus, ENC concluded that although initial housing temperatures were not specified for the FCTF tests, the values actually achieved were, in general, conservative.

The staff reviewed ENC's conclusions with respect to housing energy release. For the FLECHT tests, housing temperatures were specified based upon the 6-foot elevation quench times. If ENC had used a 6-foot elevation quench time, the " required" housing temperatures would increase. Increased " required" housing temperatures would affect the acceptability of the FCTF initial conditions.

• j

. The staff was unable to conclude that the ENC's use of only a 3-foot quench elevation to set " required" housing temperatures is appropriate.

In applying an ECbS evaluation model,Section I.A of Appendix K to 10 CFR 50 requires examination of power shapes other than the chopped cosine power shape used in the FCTF tests. Thus, it is not clear that a 3-foot quench elevation would bound the actual quench elevation at the time of peak cladding temperature for other power shapes. Ad-ditionally, 10 CFR 50.46(b)(2) requires a demonstration that local cladding oxidation not exceed 17% of the cladding thickness. To demonstrate compliance with this requirement, the peak cladding tem-perature is not the important item. The overall cladding temperature transient, which is dependent upon the overall time dependent nature of the heat transfer coefficient, is critical. Thus, the staff finds the FCTF testing did not appropriately consider the effect of housing energy release.

In researching the effect of housing energy release on quench front, the staff noted the results of the previous FLECHT tests of Refer-ence 30. Within that report, the results of FLECHT tests were pre-sented wherein only the housing temperature was varied. Those tests were then compared to a series of tests wherein only rod power was varied. It was shown that quench time was primarily determined by total energy release, regardless of whether the energy release is from stored energy in the housing or generated energy from the rods.

Thus, the staff has concluded that the effect of housing energy

release for the FCTF tests can be accommodated by adjusting the rod

! power utilized in the calculation of quench front movement. The specific method to be utilized is described in Section 2.9.3.

To further assure the suitability of the reflood heat transfer data from the FCTF, the staff requested that ENC discuss the effects of parametric variations in the test parameters on important reflood variables .such as quench front, CRF and heat transfer coefficient.

The trends exhibited by the FCTF tests were also to be compared to those obtained from the FLECHT facility. ENC provided this assessment in Reference 10.

In comparing the parametric trends, ENC noted that, in general, the variations noted in the FCTF tests were similar to the FLECHT tests.

Additionally, it was seen that the FCTF tests had slightly higher heat transfer when compared to the FLECHT tests. ENC concluded that the increased heat transfer primarily resulted from the simulation of the upper tie plate which resulted in greater bundle retention. As the upper tie plate would be expected to' result in better fluid retention due to the enhanced droplet deentrainment, the staff concludes that the enhanced heat transfer is expected.

Finally, the staff examined the experimental error inherent in the ,

1 facility. For the tests used in correlation development, ENC stated ,

1 l

._ _. , _ _ _ _ . . _ _ _ _ . _ . _ . _ . _ _ _ , _ . _ . _ , _ _ _ . _ _ _ _ _ . _ _. - . _ , _ . . ~ . , _ .

- that mass closure was obtained within 5%. With respect to error in heat transfer coefficient, ENC reported that, after 34 seconds, the error was 6%. Since these values are similar to those obtained during the FLECHT experiments, the staff finds the experimental errors acceptable.

In summary, the staff finds that the FCTF is adequate for obtaining reflood heat transfer data. While the effect of housing energy release was not considered during the FCTF testing, the staff has concluded that compensation for this effect can be accomplished by modifying the use of the correlations. Thus, the staff finds that the experi-mental data obtained is appropriate and can be used for developing correlations to comply with Sections D.3 and D.5 of Appendix K.

2.9.2 Correlation Development and Benchmarks A total of 50 tests were conducted as part of ENC's testing program.

From these tests, ENC excluded those tests which did not meet certain specified criteria. As a result of this screening,11 tests were used for heat transfer correlation development; 14 tests were used for CRF correlation development; 44 tests were used for quench front correlation development. The range of applicability for the corre-lations is provided on Table 2. The details on the development of these correlations, and their benchmarks, are provided in Refer-

. ence 12.

l

l l

The quench and heat transfer correlations were based upon all available thermocouple data for the tests selected. In developing the quench front correlation, all data which indicated decreasing quench times with increasing rod elevation, due to a " top down" quenching phenomena, were excluded. Heat transfer data from the outer two rows of rods were excluded in developing the heat transfer coefficients. The CRF data was obtained by measuring coolant mass effluent at the bundle exit. As initially developed, all the correlations were best estimate and were empirical in nature.

As a result of staff concerns about the conservation of initial set of correlations developed by ENC and reported in reference 34, ENC mod-ified their correlations and provided the revised forms in Reference 12.

With respect to the CRF correlation, the staff was concerned about possible storage of liquid in " dead" volumes, i.e., volumes not directly located in an exit fluid flow path, and requested that ENC examine what CRF would be obtained by using bundle mass storage data. As a result of this examination, ENC increased the CRF calculation by 15%. For the heat transfer coefficient correlation, ENC modified the T0FF term in order to assure conservative heat transfer coefficients for all core  :

)

elevations.

]

1

'~

To assure the adequacy of the ENC correlations, the staff reviewed nu-merous comnarisons to the FCTF data. These comparisons were presented in Figures 3.1.1 through 3.3.136 of Reference 12. The staff examined these benchmarks and concluded the following: '

a. The quench front, and hence the quench velocity, correlation is acceptable over the parameter ranges given in Table 2. The correlation provided a con-servative, or best estimate, fit for 13 of the 14 tests benchmarked.

b. The modified CRF correlation, provided a conservative, or best estimate, representation for 12 of the 14 tests even when bundle storage was used to calculate CRF. Thus, the staff finds the correlation to be acceptable.

c. The heat transfer correlation was conservative, or best estimate, for 93 of the 112 data points ex-amined. Additionally, no inherent biaseg for any elevation or parameter variations were noted. Thus, the staff finds the heat transfer correlation ac-ceptable.

O Based upon the foregoing, the staff has concluded that ENC has utilized appropriately conservative methods in selecting the data for use in #

correlating the FCTF data. Also, the correlations have been demonstrated to conservatively represent the FCTF data. Therefore, the staff concludes

> that the correlations provided in Reference 12 meet the requirements of Sections D.3 and D.5 of Appendix K.

i 2.9.3 Application of Correlations The correlations described above have been developed based upon a q certain parametric range of conditions and for constant flooding rates.

During a plant simulation, reflood rates and system conditions will vary continuously during the celculation and may exceed the correlation

'ange. This section discusses ENC's methods for applying these corre-lations to reactor calculations.

As shown in Table 2, the FCTF-based correlations are valid over a core elevation range of 4.0 to 10.3 feet. For core elevations below 4.0 feet, ENC stated they will utilize the heat transfer coefficient calculated at the 4.0 foot elevation. As this approach results in conservatively low heat transfer coefficients at these elevations, the staff finds the approach acceptable. The heat transfer co-efficients above 10.3 feet will be calculated by adjusting power, in accordance with the Z-equivalent methodology of Reference 5, such that the heat transfer coefficient would be calculated based J

i

upon an effective 10.3 foot elevation. As described later in this section, the staff finds the Z-equivalent methodology to be acceptable.

Hence, the approach used by ENC for calculating heat transfer co-efficients for elevations in excess of 10.3 feet is acceptable.

n ENC also intends to restrict the usage of the correlation to the range of parameters tested. That is, if calculated plant conditions exceed the range specified in Table 2, ENC will use either the upper or lower bouno, as appropriate, for the parameter. ENC examined the possibility of calculated plant conditions exceeding the correlation ranges. In general, it was noted that the ranges are expected to lie within the correlation boundaries. Where conditions may exceed the correlation boundaries, either the parameter has only a small effect on heat transfer coefficients, as in the case of initial rod temperature, or will be conservatively treated, as in the case of reflooding rates exceeding the correlation boundary. Thus, the staff finds ENC approach to be acceptable.

In applying the various correlations, ENC uses the REFLEX code, Ref-erence 16, to calculate quench front movement, reflooding velocity and the carryout rate fractions. These parameters are all calculated based upon average core conditions. The calculated quench front movement and reflooding velocity are then input to the T00DEE2 code, which is used to predict the hot rod heat transfer coefficient and cladding tem-perature transient.

l l

.- As discussed in Section 2.9.1, ENC did not specifically account for housing energy release during the FCTF tests. However, the FLECHT tests demonstrated that the effect of housing energy release could be accounted for by adjusting rod power. In the FCTF, there was approximately 10% excess flow area for the bundle. To i

ensure that the quench front movement and CRF properly accounts for housing energy release effects, the staff requires that rod power used in the REFLEX calculation be increased by 10% over that calcu-lated for the average core. In this manner, appropriate quench front movement and CRF would be calculated for use in the hot rod calculation.

No modification is necessary for the T00DEE2 calculation as the heat transfer correlation already explicitly accounts for the location of the quench front.

To apply the correlations to the varying reflooding rates expected during a reactor simulation, ENC uses an effective reflooding rate.

The method used by ENC to calculate the effective reflooding rate is provided in Reference 12. During the staff's review of the ENC method for applying the various correlations, additional justification

~

for the effective reflooding velocity method was requested. Via Reference 35, ENC provided comparisons of the predicted heat transfer coefficients for the variable flooding rate FLECHT tests, FLECHT runs 32333 and 32335, using three different methods of defining effective reflooding rates. In addition, comparisons of the predicted heat transfer coefficients for the D.C. Cook 2 flooding rates, from Reference 36, were provided for these three different approaches.

The three different methods studied by ENC were: 1) the effective flooding rate velocity as defined by EXEM/PWR; 2) the time shift method of WCAP-7665, Reference 32, with scaling parameters; and,

3) the time shift method of WCAP-7665 without scaling parameters.

The time shift is defined such that the total amount of water in-jected in the bundle with variable flooding rates will equal the instantaneous flooding rate multiplied by the real time plus the time shift. WCAP-7665 further adjusts this time shift with scaling parameters, It should be noted that the second approach is that used in the currently approved ENC WREM-IIA model, while the third method is the same as that approved in another vendor's ECCS Evalu-ation Model.

Examining the comparisons of the EXEM/PWR methodology to the two other methods, it was seen that the EXEM/PWR methodology calculated later quench times than the two other methods. Similar heat transfer coefficients were obtained using all the methods, with the EXEM/PWR methodology yielding heat transfer coefficients which fell between the two other acceptable approaches. In addition, the EXEN/PWR methodology predicted conservative heat transfer coefficients relative to those obtained frcm FLECitT tests 32333 and 32335. Addi-tionally, as part of Reference 12 ENC presented comparisons to three variable flooding rate tests from FCTF. These tests showed the ef-fective flooding rate methodology to be conservative. Thus, the staff finds the effective reflooding rate used by ENC to apply the FCTF correlations to reactor conditions to be acceptable.

.. ENC also proposed a method for applying the FCTF-based correlations to axial power distributions different from the 1.66 chopped cosine axial power distribution used in the FCTF. This method is also used to account for differences in fuel length, fuel rod diameter and coolant channel flow area. Adjustments are made in both the REFLEX and T00DEE2 codes when this methodology is used.

Within the REFLEX code, differences in axial power distribution and fuel rod geometries are accounted for by modifying the initial core average values for QMAX and TINIT. The adjusted parameters are then used in the quench front and CRF correlations. The specific method employed by ENC to define these parameters is documented in Reference 5.

l These adjustments may be viewed as resulting in a REFLEX prediction for fuel rods with an FCTF power distribution.

The REFLEX-calculated core reflooding rates, quench front propagation and CRF are the input to the T00DEE2 code for calculating the hot rod temperature transient. The values for QMAX and TINIT input to the T00DEE2 code are adjusted in the same manner as that for REFLEX ex-cept that hot rod initial values are used. The adjusted values are then directly used in the heat transfer coefficient correlation. An equivalent elevation, based upon conserving integral power, is defined for applying the T00DEE2 calculated heat transfer coefficient to the l fuel rod. That is, if the equivalent FCTF rod at 8 feet has the same l

1 1-i 27 -

)

~

l integral power as the simulated reactor rod at 9 feet, the T00DEE2 calculated heat transfer coefficient at 8 feet would be applied for predicting the 9 foot elevation cladding temperature transient. The

}. specific method used for conserving integral power is discussed in l Reference 5.

To demonstrate the appropriateness of their model, ENC benchmarked

, data from the FLECHT skewed profile low flooding rate heat transfer tests 11428, 14331 and 16110. These data were obtained from Refer-ence 37. The proposed method yielded lower heat transfer coefficients, and thereby higher cladding temperatures, than observed in the FLECHT tests. Thus, the method was shown to be conservative. In addition to

evaluating the infomation provided by ENC, the staff examined some of the FLECHT data to further assure that the ENC methodology is conser-i vative. Comparisons were made between the FLECHT cosine tests 02414 ,

! and 03113 and the skewed power shape tests 15305 and 11003 using the l proposed ENC method. These data comparisons showed the ENC method to l be conservative. Thus, the staff finds the ENC Z-equivalent method- .

l j ology to be acceptable.

Within Reference 14 ENC provided additional justification for util-izing the FCTF data to 14x14, 15x15 and 17x17 PWR fuel designs different from the specific fuel rod design used in the FCTF tests. ,

I In essence, ENC proposes to use the Z-equivalent methodology, discussed

above, to account for the differences in fuel rod designs. ENC used i

j -

the FLECHT overlap tests from the 17x17 FLECHT SEASET, which were 4

~ - . . _ - - . . _ . . , - - . - - - _ . - - . - , - , _ . _ - _ - - . , ,.,-,..,.._.m-m_-.___-y.~.---,-,_._m._,

O

] performed to provide a comparison to the 15x15 FLECHT tests, to demon-strate the extension of the Z-equivalent methodology. The overlap tests rescaled the 15x15 FLECHT test conditions to preserve both generated power and stored energy per unit flow area. The specific method utilized for scaling the FLECHT overlap tests is essentially i equivalent to the ENC Z-equivalent methodology. The comparison of 1

the data indicates that the scaling approached used results in good l agreement between the tests.

I l

) The staff reviewed the ENC arguments for extending the FCTF data to i

l other fuel rod designs. Because the FCTF data generally confirmed the i same trends as the FLECHT data, extension of the FCTF data with the Z-equivalent methodology, based upon the FLECHT tests, is acceptable.

The staff also notes that ENC has committed, in Reference 15, to per-form a series of 15x15 FCTF reflood heat transfer tests in 1986. This data will then be used by ENC to confirm the validity of extending the 17x17 FCTF-based correlations. Based upon this comitment, the staff finds the ENC proposal to be acceptable.

1 l

In suninary, the staff finds the proposed ENC methods for applying the FCTF correlations to reactor simulations to be acceptable.

1 I

l 3.0 Application of EXEM/PWR As part of the staff review of EXEM/PWR, the staff requested ENC to evaluate the impact of using EXEM/PWR on previous break spectrum and power shape studies performed for licensing calculations. ENC's re-sponses were provided in Reference 4. With respect to defining the limiting break size, it was stated that ENC will justify the limiting break size from the LOCA/ECCS spectrum for use with the EXEM/PWR model on a plant specific basis either by justifying the validity of previous spectrum results or performing additional spectrum calculations. The staff finds this approach acceptable to meet the requirements of Sec-tion C.1 of Appendix K.

With respect to power shape studies, ENC stated that since the data bases were similar for developing the heat transfer correlations, they expected that application of the EXEM/PWR would yield similar results to those previously obtained by either ENC or the NSSS vendor. Thus, the response implies that ENC does not feel additional calculations are required. The staff finds this response unacceptable as the l

power shape effect on peak cladding temperature is not just dependent on the heat transfer correlation, but rather the evaluation model as a l whole. Therefore, the staff requires that ENC document compliance with Section I.A of Appendix K when applying EXEM/PWR to licensing calcula-tions for a specific plant.

l

.- 4.0 Conclusions In summary, the staff finds that the EXEM/PWR ECCS Evaluation Model, as documented in References 1 through 15, fully complies witn the requirements of Appendix K to 10 CFR 50 subject to the following limitations:

i

a. ENC is prohibited from using the CRF and heat transfer correlations of References 1, 2 and 31 in future analyses performed with EXEM/PWR. Only the correlations of Reference 12, applied in conformance with item b. below, is accepted as part of the EXEM/PWR EM.

b. In performing REFLEX calculations, ENC shall increase the average core rod power, used in the quench front and CRF correlations, by 10%.

i In addition, when EXEM/PWR is applied to plant-specific licensing calculations, ENC must:

a. Provide justification that the leakage flow path from the upper plenum to the downcomer is well characterized if the REFLEX leakage flow model is used.

b. Provide justification for the limiting break size analyzed, or provide additional break spectrum calculations, and

c. Demonstrate compliance with Section I.A of Appendix K to 10 CFR 50.

This model should become effective no later than 90 days after receipt of this SER by ENC', That is, all submittals demonstrating compliance with Appendix K using the ENC model after that time should use the model des-cribed in this report. During the 90 day period, any applicant or licensee requiring an Appendix K ECCS analysis may elect to submit an analysis based upon the previously approved ENC WREM-IIA model, with consideration of the effect of NUREG-0630, or otherwise submit analyses conforming to this SER.

i l

1

E l

,o l

TABLE 1 l ECCS Model Updates of EXEM/PWR Fuel Rod Model - RODEX2

- Stored Energy

- Fission Gas Release Blowdown Model - RELAP4-EM Code

- NUREG-0630 Clad Rupture / Blockage Model

- Modified Fuel Rod Model Reflood Model - REFLEX Code

- Leakage Flow from Upper Plenum to Downcomer

- Split Break Model

- Core Outlet Enthalpy Model -

- Revised Carryout Rate Fraction Correlation Heatup Model - T00DEE2 Code

- Revised Steam Cooling Model

- NUREG-0630 Clad Rupture / Blockage Model

- Revised Radiative Heat Transfer Model

- Revised Reflood Heat Transfer Correlation

TABLE 2 Range of Applicab'ility For REFLOOD Correlations E

Pressures, psia 19.4 -

40.

Reflood Velocity, in/sec 0.59 -

1.77 Subcooling. *F 20 -

134 Initial Peak Rod Temperature, 'F 1054 -

1650 Maximum Heat Flux, kw/ft 0.39 -

0.68 Elevation,'ft 4.0 -

10.3 e