ML20245H484
| ML20245H484 | |
| Person / Time | |
|---|---|
| Site: | Oconee, Mcguire, Catawba, McGuire, 05000000 |
| Issue date: | 08/09/1989 |
| From: | Tucker H DUKE POWER CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 8908170124 | |
| Download: ML20245H484 (4) | |
Text
pr 33 3y a
T hj W '
- i.
~ a 0.t f*7-
. DUKE POWER C>OMPANY C*
P.O. B03c 33189 I
[*'
CHARLOTTE, N.C. 28242
- HAL B. TUCKER '
ret u ssore
. viam enemment -
(704) 373-4531 3
pers.um enuoverson i
i
' August;9,j.1989 :
1 Document' Control Desk' l
U.:S. Nuclear Regulatory Commission-
' Washington,LD.C. 20555 j
Subj ect:- Oconee Nuclear Station.
Docket. Numbers 50-269, -270, and.-287 McGuire Nuclear Station
-)
Docket Numbers 50-369 and -370' Catawba Nuclear Station I
' Docket Numbers 50-413 and -414 Additional Information On DPC-NE-3000,
)
~" Thermal-Hydraulic Transient Analysis Methodology"
=!
On July 18.and 19, 1989, the reviewers of'the subject Topical Report and the'
~
authors.'had a telephone conversation regarding~RETRAN once-through steam in,
- generator; heat transfer modeling. The attached discussion serves to formalize.
and document.the information exchanged during that telecon.
i If there'are any questions, please call Gregg Swindlehurst at=(704) 373-5176 or
~
Scott Gewehr at.(704) 373-7581.
-Very truly.yours, 1
H.'B Tucker i
r SAG 181/lcs
- xc:
Mr. D. S. Hood Mr. S. D. Ebneter Office of Nuclear Reactor Regulation Regional Administrator U. S. Nuclear' Regulatory Commission U. S. Nuclear Regulatory Comm.
Washington, D.C.
20555 101 Marietta St., NW, Ste. 2900..
Atlanta, Georgia 30323 1
Mr. D. Katze Mr. W. T. Orders Office of Nuclear Reactor Regulation NRC Resident Inspector i osri$-.
.U. S. Nuclear Regulatory Commission Catawba Nuclear Station A
Washington, D.C.
20555'
.{
1 88a 00 Mr. L. A. Wiens Mr. P. K. VanDoorn o
Office of Nuclear Reactor Regulation Senior Resident Inspector w
U. S. Nuclear. Regulatory Commission McGuire Nuclear Station gg Washington, D.C.
'20555
- Oc jM Dr. K. Jabbour Mr. P. H. Skinner
.g7 Office of Nuclear. Reactor Regulation NRC Resident Inspector sg
'I o.o U. S. Nuclear Regulatory Commission Oconee Nuclear Station W
t 1*1L Washington, D.C.
20555 j
i a
s l
I o
August 4, 1989 Duke power Response to NRC Questions l
l Regarding Steam Generator Heat Transfer Modeling with the RETRAN Code On July 18 and 19, 1989 the NRC contractor, International Technical Services, asked several questions regarding the capability of the RETRAN code to simulate heat transfer on the secondary side of the B&W once-through steam generator. The following discussion is intended to respond to these questions by providing an overview of modeling, validation efforts, observed code /model limitations, and the approach for accommodating these limitations.
The Duke power Oconee RETRAN model nodalizes the steam generator secondary tube bundle region with a vertical stack of homogeneous volumes. The steam generator downcomer is modeled as a separated bubble rise volume.
- ne adequacy of the nodalization has been addressed by sensitivity studies which were submitted in response to a previous NRC question.
The validation of the steam generator modeling, including secondary heat transfer, consists of both steady-state and transient comparisons to reference data.
The steady-state comparisons include the axial variation of primary and secondary temperatures in the tube bundle region, and comparisons to steam generator secondary mass.
Although the axial temperature profiles are not exactly matched in the RETRAN prediction, the boiling length and the general shape of the temperature profile at full power are reasonably well-predicted, and the amount of superheat at the exit matches plant data.
The secondary masc inventory predicted by RETRAN is close to other code-predicted reference values, although no actual plant data exists. The capability of the RETRAN code to achieve an initialization that compares this well to the reference data is noteworthy, and is as good or better than similar codes.
Validation of the steam generator heat transfer modeling during transient conditions is based on the comparisons to plant transient data as described in Chapter 4 of DPC-NE-3000.
The plant transients selected include a wide spectrum of heat transfer conditions and phenomena, including the following.
The pressurization /depressurization cycle that occurs following a e
turbine trip, from several different pre-trip conditions [4.1.1, 4.2.1, 4.2.2, 4.2.3, 4.4.1, 4.6.1]
Steam generator dryout due to loss of feedwater [4.1.1]
e Lepressurization and overcooling due to excessive steaming following e
reactor trip [4.2.1]
A severe overfeed following reactor trip [4.2.2]
e Steam generator depressurization and dryout due to excessive e
steaming, followed by a rapid overfill at low pressure [4.2.3]
e A transition to natural circulation [4.3.1]
Steady-state natural circulation at various power levels [4.3.3]
e A power maneuver resulting from a decrease in feedwater flow [4.5.1]
e A power maneuver resulting from an increase in steaming rate [4.5.2]
e 1
=-
e By comparing the performance of the RETRAN model to a large set of data, it is possible to draw conclusions about the capability of the model to simulate the integrated response of the steam generator. Accurate simulation of steam generator pressure and level, the dominant parameters with respect to heat transfer, and the resulting cold leg temperatures, indicate the quality of the heat transfer prediction. As shown in Chapter 4 of DPC-NE-3000, RETRAN does compare reasonably well with the plant data over a wide range of transient conditions.
It is apparent from these benchmark analyses that the rapid secondary transient resulting from a turbine trip is a good test of the code /model. For this transient RETRAN predicts too much heat transfer due to the lack of an unequal phase velocity model and a consequent overprediction of the boiling length. This results in lower cold leg temperatures for a period of time immediately following the trip. Provided that the model is initialized with a correct inventory, and that the feedwater boundary condition is characterized appropriately, this overprediction of the heat transfer rate will be limited in duration and magnitude. As such there is only a limited impact on the simulation results. This can be confirmed by the predictions of pressurizer level, which indicate the net effect of primary coolant expansion or contraction due to changes in steam generator heat transfer.
In all cases the comparisons between RETRAN predictions and plant transient data show that the pressurizer level trends are correctly simulated, and in most of the tenchmarks the agreement is very good.
The' capability of the code /model to predict natural circulation.is also demonstrated in the validation analyses. The RETRAN simulations of a transition from forced to natural circulation, and several steady-state natural circulation data, indicate a small overprediction of the natural circulation flowrate. The transition to natural circulation is predicted very well.
The difference in the steady-state comparison is a combination of several factors which affect the delicate balance between the loop density distribution, which is strongly influenced by the steam generator heat transfer profile, and loop frictional losses at low flow conditions.
Uncertainty in the core power level is also a contributor. The integrated effect as predicted by the code /model is a very reasonable comparison with the plant data.
The benchmarks presented in Chapter 4 did not include any comparisons to scaled integral test facility data or to separate effects tests. The following summarizes the rationale for not including such validation work, and why the plant data which was used is sufficient.
Scaled facility test data has typically focused on LOCA phenomena, for which plant-scale data is unavailable. As stated in DpC-NE-3000, the Oconee RETRAN model is not intended for simulating LOCAs. Another limitation of test facility data is that the impact of scaling on the validity of the data must be addressed.
This can be very difficult if consideration is being given to modify the code l
or model based on the scaled data. Separate effects tests are mainly useful l
for qualifying particular aspects of a code, such as phase separation models or critical flow models. These data are therefore most useful in the code development process, or to serve as a basis for coding modifications. The validation of the Oconee RETRAN model has not identified a need to modify the RETRAN code. The perception that plant transient data is not of high enough quality for code /model validation is not supported by our experience.
Provided that a large and broad database can be assembled, and provided that 2
.s
, < i sm 7( ?-
4 generally good agreement is obtained without adjusting the code /model to correct.significant mismatches between predictions and data, then plant data proves'very sufficient. If the database was limited in scope and quantity, and.if. major code /model adjustments were necessary, then questions regarding the broader capabilities of the simulation model would be warranted. The overall agreement between RETRAN predictions and plant data varies from' 4
reasonable'to. excellent. 'The data trends are predicted, and although some differences in timing and magnitude do' exist, no phenomena were missed. No tuning of the code and very limited adjustments to the model as detailed in DPC-NE-3000 resulted. Achieving this level of quality without good modeling of steam generator heat transfer would be impossible since the steam generator performance determines the response of most of the parameters of interest.
The validation efforts focused on the macroscopic integrated performance of the code and model by comparison to steady-state and transient plant data.
Based on the validation results presented, it can be concluded that the oconee RETRAN model has been thoroughly exercised, including modeling of steam generator heat transfer, and that it performs as well or better than other.
similar codes. This statement is valid in the context of the capability to simulate non-LOCA transients, which is the scope of DPC-NE-3000.
The observed limitations discussed previously, regarding steam generator heat transfer modeling, can be accommodated in a conservative manner when reanalyzing an FSAR Chapter 15 transient. Many of the FSAR transients are unaffected by the observed limitations since these limitations occur primarily post-trip, and the pre-trip response is of interest in the FSAR. Nevertheless, the potential for the vaserved limitations to have an impact is recognized,'and appropriate compensating assumptions will be incorporated as necessary to ensure a conservative reanalysis of FSAR transients.
3
_ _ - _ _ _ _ _ - _ _ _ - _