ML20215N690
| ML20215N690 | |
| Person / Time | |
|---|---|
| Issue date: | 11/04/1986 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20215N683 | List: |
| References | |
| NUDOCS 8611070164 | |
| Download: ML20215N690 (11) | |
Text
.
ENCLOSURE SAFETY EVALUATION REPORT OF TOPICAL REPORT BAW-10155, "F0AM2 - COMPUTER PROGRAM TO CALCULATE CORE SWELL LEVEL AND MASS FLOW RATE DURING SMALL-8REAK LOCA" TOPICAL REPORT EVALUATION 1.0 -INTRODUCTION Babcock & Wilcox (B&W) proposes to use the F0AM2 thermal-hydraulic computer code for analysis of postulated loss-of-coolant accidents (LOCAs) in future licensing actions.
Although there are many aspects to evaluate in a postulated LOCA, this code is concerned with the water swell level in the core of the reactor.
The code, as described in the B&W topical report BAW-10155 (Ref. 1), determines whether, at any given time in a LOCA, the water inventory in the core is sufficient to cover the entire active core with a water-steam mixture.
If it is determined that the core is only partially covered with.the water-steam mixture, the code will calculate the water swell level (i.e., the increase.in the volume of water-steam due to void formation) and the steam generation rate.
If it is determined that the core is completely covered and the water inventory is sufficiently large to cause the water-steam mixture to flow through the vent valves of the reactor, the code will, on the user's request, cal'ulate the mass' flow rate of the reactor c
core and steam generation rate.
In addition, FOAM 2 calculates the sur-face temperatures of the cladding along the entire length of the fuel rod for the average channel and above the water swell elevation for the hot channel.
The code also determines the elevation where the channel dryout occurs for the average channel.
I B611070164 861104 PDR TOPRP ENVDW C-PDR
, The staff's review is based on the information provided in BAW-10155 and in B&W's responses to the review questions (Ref. 2).
The results of the review are discussed below.
2.0 EVALUATION OF GOVERNING EQUATIONS There are two principal equations used in the code:
(1) an equation that is used to model the formation and rise of bubbles in the core of the reactor; and (2) an energy equation that is used to calculate the enthalpy distribution and the swell water level.
The Wilson-bubble rise model (Ref. 3) used for modeling the two phase slip was derived for the pressure range of 150 to 600 psig.
This model has also been shown to be applicable at higher pressures..At the staff's
. request, B&W compared the void fraction predicted by the F0AM2 Wilson correlation-with test data.
T'he data included the void fraction measured at 1000 and 2000 psia from an 18-inch diameter vessel (Ref.
13), and the void fraction measured at 2000 psia from the test vessels with diameters of 2.5 and 18 inches (Ref. 14).
The comparison in-dicates that at higher pressures (1000 and 2000 psia), the F0AM2 Wilson prediction is in good agreement with the test data.
Based on the review of data, the staff concludes that the use of the i
Wilson model in FOAM 2 is acceptable.
However, use of the Wilson model for pressures less than 150 psig must be justified.
The energy equation used to calculate the enthalpy distribution and the swell water level is based on the balance of the heat input and enthalpy,
-v-s
' r.
and kinetic and potential energy increases.
The kinetic and potential energy terms were neglected in FOAM 2.
The staff finds that this as-sumption is acceptable since these two terms have a small effect on the overall results compared with the enthalpy term.
Furthermore, neg-f lecting these terms results in higher calculated enthalpy (higher boil-off rate) and thus lower calculated swell water level (which is a more conservative result).
B&W assumes that the coolant behavior inside the core is steady-state.
Steady-state energy balance equations are used to calculate the axial enthalpy distribution in the core.
This assumption restricts the ap-plicability of FOAM 2 to the slow' blowdown phase and the reflood phase of small and large break LOCAs.
The applicable ranges proposed'by B&W.
are: the reactor system flow $-50 lbm/sec for a small break LOCA and the reactor system flow 5 200 lbm/sec for a large break LOCA.
The staff l
finds that the proposed applicable reactor system flow is small (less than 1% of the rated flow for al typical B&W plant).
The assumption ~of steady-state energy' balance equations is consistent with the assumption in-cluded in the codes used by the industry.
The staff concludes that the
~
i proposed ranges are acceptable.
However,-justification must be provided i
l Por reactor system flow rates exceeding these ranges.
3.0 EVALUATION OF CALCULATIONAL METHODS
]
FOAM 2 consists of two distinct series of calculations:
the hydraulic analysis and the thermal analysis.
Both enalyses are performed at each time of interest in the reactor transient.
, 3.1 Hydraulic Analysis In tne hydraulic analysis, FOAM 2 calculates either the swell water level and mass flow rate through the core corresponding to the given core equivalent water level, or the water-steam flow through the vent valves depending upon whether the core is covered with the water-steam mixture.
To determine whether the core is covered with the water steam mixture at the time under consideration, the minimum mass flow rate which is required to cover the core is calculated by using energy balance equations.
Based on the minimum required flow rate, a corresponding mass inventory in the core is calculated.
The calcu-
-lated mass inventory is then compared with the input water inventory
'obtained from the CRAFT 2 code (Ref. 11) calculetion.* If the input value is smaller, the core is defined as uncovered; otherwise, it is covered.
When the core is determined to be uncovered with the water-steam mixture, the swell water level and mass flow rate are determined by using an interactive process as follows.
An initial core mass flow e
is assumed. The axial enthalpy and quality distribution of the water-steam mixture are calculated using the flow rate.
Based on the calculated axial enthalpy and quality distribution, the swell water level and the equivalent water level (EWL) are calculated.
The cal-culated EWL is compared with the'EWL corresponding to tho input mass inventory (which is determined by the CRAFT 2 code).
If the error is larger than a preset :orevergence criterion, a new value of. mass flow
. rate is estimated in'the code, and the process is rsveated until convergence is obtained.
When the calculations indicate that the core is covered with the water-steam mixture, FOAM 2 calculates the amount of water-steam flow through the vent valves.
Tne flow through the vent valves is a function of pressure drop across the particular valve.
In the iteration, the maximum equivalent water level (or water inventory) without overflow is calculated.
The input water inventory from CRAFT 2 is compared with this value.
The difference between these two terms determines whether there is an overflow through the vent valve and the amount of overflow, if any.
The staff has reviewed the calculational procedures and finds that the equations used for the calculations are correct and the method
.of approach is adequate and acceptable.
However, since the overflow rate depends on the correlation between the vent valve flow and pressure drop across the valves, use of the correlation must be justified for vent valves different from the existing vent valve type.
3.2 Thermal Anclysis In the thermal analysis, the F0AM2 code calculates the elevation of critical heat flux (CHF) and the cladding surface temperature distributior..along the entire core length for the average channel.
. For the hot channel, FOAM 2 calculates the cladding surface temp-eratures above the swell water level.
The Macbeth CHF correlation is used for the prediction of the CHF location.
Since the Macbeth CHF correlation has an applicable mass flux range down to about 0.4% of rated flow for a typical 177-assembly B&W plant, the staff concludes that its application in F0AM2 is acceptable.
The heat transfer co-efficient used for calculation of the. cladding surface temperature is based on the McAdams correlation.
As indicated in Reference 5, the correlation overpredicts the convective heat transfer coefficient, and thus yields nonconservative cladding surface temperatures.
- However, ORNL~ data (Ref. 6) indicate that the radiation heat transfer from the fuel rod to the superheated steam is significant.
The FOAM 2 heat transfer calculation does not include radiation heat transfer.
- Also, the cladding temperature predicted by FOAM 2 is higher and conservative when compared with the ORNL data (Ref. 6).
The staff, therefore, concludes that the F0AM2 calculation of cladding surface temperature is acceptable because the conservatism of neglecting' radiation heat transfer overrides the nonconservatism of using the McAdams corre-lation.
4.0 FOAM 2 VERIFICATION B&W compared the F0AM code (Ref. 4), an earlier version of FOAM 2, pre-dictions with data from experimental tests by four companies:
(1) General Electric Company (GE) (Ref. 7); (2) Hitachi Corporation (Ref. 8); (3) the Westinghouse Electric Company (Ref. 9), and (4) B&W's Alliance Research Center (Ref. 10).
The GE, Hitachi, and Westinghouse tests resulted in data on
~7-collapsed liquid levels, whereas.the B&W test resulted in data on swell water level and fluid temperature.
The basic models of the FOAM and F0AM2 codes for the swell water level calculations are the same.
B&W's comparison of the calculated minimum water level required for covering the core with the GE data indicates that the maximum deviation of the F0AM prediction for all data points for the case at 100 psia was about 0.6 ft. in terms of the collapsed liquid level (test channel is more than 10 ft. high).
A similar result was obtained in the com-parison with. data at atmospheric pressure.
F0AM consistently under-predicts the collapsed liquid level required to cover the core.
The maximum underprediction is up to one foot in liquid level.
The deviations between the predicted and measured water level data are within the measurement uncertainty.
Since F0AM generally underpredicts the data in the low pressure range (which is nonconservative), and the Wilson bubble rise model was derived for pree,sures higher than 150 psig, the staff concludes that justification has to be provided for FOAM applications at pressures below 150 psig.
The Hitachi test is similar to the GE test, except more experimental points were obtained and less deviation was observed in the data.
B&W compared the data from the Hitachi test with the FOAM results (Ref. 1). The com-parison shows similar agreement as in the GE case.
Since the test was performed at atmospheric pressure, F0AM also underpredicts the minimum required liquid level for the core covering at press res less than 150 psig as. indicated by the comparison with the GE data.
e
,. 'Unlike the GE and Hitachi tests, the Westinghouse test is a slow-transient test rather than a steady-state test, and the tests were performed at two
~
pressure levels (400 psia and 100 psia).
The F0AM estimates are in good agreement with the test data at 400 psia. There is almost no deviation between the FOAM results and the data.
The data comparison at 100 psia shows some slight deviations.
The maximum deviation is about 0.64 ft. in collapsed liquid level for the case at the 10 ft elevation, and a maximum deviation of about 0.3 ft for the cases at the~8 and 6 ft. elevations.
Even though these errors are small and within the error bands of the test 8
data, the FOAM calculations consistently underpredict the minimum collapsed liquid level which is required to cover the core and support trie conclusion drawn from the GE data about the FOAM limitations at low oressures.
The F0AM estimated swell water levels are in' reasonably good agreement with the results of the B&W tests (at l'80 psia). The F0AM predictiors lie within the scattering band of the data points.
Also, the F0AM code predicts higher fluid temperatures at upper elevations than the B&W test data.
This makes the code more conservative for the case tested.
In summary, B&W compared the F0AM code estimates with the B&W test data on swell water level and fluid temperature and with the GE, Hitachi, and Westinghouse data on collapsed liquid level.
The comparisons indicated that at higher pressures (180 and 400 psia), the F0AM prediction is in good agreement with the test data.
Data comparisons at lower pressure indicate that the FOAM code consistently underpredicts the liquid inventory required to cover the core.
Since F0AM underpredicts the data
i at low pressures and the Wilson. bubble rise model was derived for pressures higher than 150 psig, the staff concludes that justification is required for application of FOAM.at pressures lower than 150 psig.
It should be noted that although the basic models of the F0AM and FOAM 2 codes for the swell water level calculations are the same, there are some differences in the numerical iteration scheme.
The FOAM 2
~
numerical convergence is largely basad on the difference in the cal-culated and actual equivalent water level data, whereas, the F0AM con-vergence is based on the inlet and outlet flow rates.
B&W compared the calculated results for swell water level from the F0AM and F0AM2 codes (Ref. 2) and showed that the~ codes yielded essentially the same results.
Therefore, the staff concludes that the swell water level benchmark results'and the staff's review results, although based on the FOAM code, are applicable to the FOAM 2 code.
5.0 CONCLUSION
Based on its review, the staff concludes that topical report BAW-10155 is~ acceptable for referencing in licensing actions for calculating the swell water level and mass flow rate during the slow blowdown and reflood i
phases of the large and small break loss-of-coolant-accidents.
- However, in order to resolve concerns regarding the modeling assumptions, the.fol-lowing restrictions are imposed on the use of the FOAM 2 code:
4
., _ _ _ ~
4
' 1.
The code'can be used for core conditions with core flow rates less than or equal to 50 lbm/sec and 200 lbm/sec fo'r the small and large break LOCAs, respectively.
However, justification must be provided for flow rates exceeding these ranges.
2.
To assure correct vent valve flow calculations,~the correlation between the vent' valve volumetric flow and the pressure drop across the valves used in calculations must be justified for vent valves different from the existing vent valve type.
3.
Since the F0AM2 prediction is nonconservative as compared with the test data at low pressure (100 psia) and the Wilson bubble rise model was derived for pressures higher than 150 psig, use of FOAM 2 at pressures less than 150 psig must be justified.
~
i e
i D
~. _,, _, _. _. _ -.. _, _ _,. -.
,,,.m_
m
~
?
j i
- 6. 0 REFERENCES
?
1.
BAW-10155, "F0AM2 - Computer Program to Calculate Core Swell Level and Mass Flow Rate During Small-Break LOCA," dated November 1982.
2.
Letters dated July 6 and August 15, 1984 from J. Taylor (BAW) to C. O. Thomas, (NRC).
3.
J. F. Wilson, R. J. Grenda, and J. F. Patterson, "The' Velocity of Rising Steam in a Bubbling Two-Phase Mixture," ANS Transactions, Vol. 5, pp. 151-152, 1962.
4.
B. M. Dunn, C. D. Morgan, L. R. Cartin, "Multinode Analysis of Core Flooding Line Break for B&W's 2568-MWt Internals Vent Valve Plant," BAW-10064, Babcock &~Wilcox, Lynchburg, Virginia, 1973.
5.
N. K. Savani, J. R. Paljug, and R. J. Schoma(er, "B&W's Small-Break LOCA ECCS Evaluation Model," BAW-10154P, Baocock & Wilcox, Lynchburg, Virginia, 1982.
6.
T. M. Anklam, "0RNL Small-Break LOCA Heat Transfer Test Series I:
Rod Bundle Heat Transfer Anaysis," NUREG/CR-2052, Oak Ridge National Laboratory, Oak Ridge, Tennessee, August 1981.
7.
Amendment No. 12, W. H. Zimmer Nuclear Power Station PSAR, Docket No. 50-358, April 1970, Cincinnati Gas & Electric Company, Cincinnati, Ohio.
8.
Ogasawara, et. al., " Cooling Mechanism of the Low Pressure Coolant Injection System of Boiling Water Reactors and Other Studies on the Loss-of-Coolant Accident Phenomena," Hitachi Research Laboratory, Hitachi LTD., Japan,1973.
9.
J. P. Cunningham, O. J. Mendler, and R. W. Steer, " Core Uncovering Test at 100 and 400 psi, Westinghouse Electric Company Report No. DT-T&H-517, March 1973.
10.
S. I. Abdel-Khalik and R. T. Bailey, "An Investigation of the Level Swell Core-Cooling Mechanism for a PWR,: AICHE HT&EC Division, 1976 National Heat Transfer Conference, 1976.
11.
BAW-10092, Rev. 3 " CRAFT 2-FORTRAN Program for Digital Simulator of a Multinode Reactor Plant During Loss of Coolant," October 1982.
12.
A letter from R. Schomaker (BAW) to S.' Sun (NRC), Dated September 12, 1986.
13.
J. F. Wilson, W. E. Littleton, H. P. Yant, and W. C. Meyer, " Primary Separation of Steam from Water by Natural Separation," ACNP-65002 (April 15, 1965), Published by Allis Chalmers AED, Milwaukee.
14.
N. Zuber, F. W. Staub, G. Bijwaard, and P. G. Kroeger, " Steady State
.and Transient Void Fraction in Two Phase Flow Systems," EURAEC GEAP-5417 (January 1967), Published by General Electric APED, San Jose.