ML20040H218
| ML20040H218 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 04/19/1978 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML18023A006 | List: |
| References | |
| NUDOCS 8202170453 | |
| Download: ML20040H218 (7) | |
Text
_.
=,.. - -
g
. - p.,
f-Q
/
ENCLOSURE 1 4
TOPICAL REPORT EVALUATION i
APR 191978 Report Titles: An Improved Progr'am for Thermal-Hydraul.ic Analysis of Rod Bundle Cores - WCAP-7955 - June 1973 Application of the THINC-IV Program to PWR Design -
WCAP-8054 - September 1973 Originating Organization: Westinghouse Electric Corporation Reviewed by: Reactor Systems Branch 1.0 Sumary of Tooical Reoorts THINC-IV is i hydraulic computer program designed to calculate flow conditions in multiple parallel channels such as a reactor core under steady-state conditions.
Like its predecessors THINC-I and THINC-II, the code accounts for flow interaction along the channels because of mismatches in hydraulic resistances and/or thermal mixing.
An important innovation
.in THINC-IV, aside from changes in numerical techniques, is the inclusion of inertia terms in the lateral momentum equation that determines the, cross ficw between channels.
Heretofore, these terms have been omitted and the effective lateral resistance lumped into a lateral loss coefficient.
THINC-IV solves a boundary value problem which means that local flow disturbances can prcpagate their effects upstream.
Comparisons with several types of data have been perfor:ned to verify the THINC-IV calculations. These data, obtained from different thermal-hydraulic tests ranging from 3x3 red bundles up to reactor cores, include gross flow parameters such as pressure drop and local parameters such as flow and quality in subchannels.
In addition, studies were presented to demonstrate the sensitivity of several thermal-hydraulic models used in THINC-IV.
THINC-IV is used for thermal-hydraulic design calculations'of Westinghcuse cores during steady-state and transient events that can be analyzed cen-servatively as pseudesteady-state prcblems. A set of three separate THINC-IV calculacions 'are performed for these evaluations. The first, a
~
core-wide analysis, treats each assembly as a flow channel anc accounts for gross flow redistribution based on the radial power distribution in the core and the assumed inlet flow.
For desig'n calculations, the flow to the l
hot assembly is reduced to-account for inlet plenum effects and the hot assembly is surrounded by assemolies of equal power.
No turbulent mixing between assemblies is considered.
8202170453 920212 PDR ADOCK 05000445 j
O PDR
8
/
g
[
(
~
~
The second step divides the hot assembly into four flow channels using the axial flow conditions from the first step as boundary ' conditions. The third step examines the hot channel flow conditions using axial flow distributions from step 2 as boundary conditions to a detailed analysis of the hot quadrant of the hot asser.bly.
The hot channel accounts for manufacturing tolerances and power distribution uncertainties.
Inter-channel enthalpy, mixing is considered in step 3..
2.0 Sumary of Staff Evaluation 2.1 Basic Ecuations' The three conservation equations--mass, energy, and momentum--are solved to-determine the flow conditions in a three-dimensional array where axial gradients are dominant. A lateral mcmentum equation is included which accounts for changes in lateral momentum of fluid crossing the side faces of a control volume (sV ) and of fluid crgssing the axial faces of a control 2
volume (sUV).
Experience with COBRA-IIICll) has shown that the latter cross product term ( UV) is important and may even dominate the crossflow solution.
Turbulent mixing is incorporated only in the energy equation. There is no turbulent mixing of momentum, which is a second-order effect that should not greatly change the solution.
The lateral momentum equation in THINC-IV is different than that used in
'THINC-I which equated lateral pressure differences to irreversible lateral friction losses only.
A large lateral loss coefficient is used in THINC-I to compensate for the lack of acceleration terms. A comparison of THINC-I and THINC-IV calculations for problems with 10-20% flow perturbations
' indicates significant differences with the THINC-I calculations exhibiting a more damped flow behavior along the channel. THINC-II has no lateral pressure drop model (zero resistance) and calculates flow perturbations similar to those obtained with THINC-IV.
Based on these comparisons, the THINC-IV calculations show that an open lattice flow situation does not support significant lateral pressure gradients.
The conservation equations are solved by application of a small perturbation technique that assumes that the axial gradients are dominant. Two sets of equations containing unperturbed and perturbed / unperturbed parameters are put'in a finite difference form using centerad) differences around a computational cell similar to the MAC scheme U.
Based on our review, the formulation of the finite difference ecuations and the precedurs used to solve them is numerically correct and procerly aopliec(1).
The main assumption of small perturbations to a basically one-dimensional c
flow must be maintained if accurate predictions are to be obtained. This i
should not be a limitation for PWR applications where axial flow perturbations of less than 20% are observed.
8 O
9 m
'~
...[.-.____.,.
~
~
(
k.
3-4 2.2 Correlations The axial friction factors are evaluated using the Novendstern-Sandberg correlations for isothermal, norti'sothermal, and two-phase flow conditi
'These correlations are based on a survey' study of pressure drop data (ons.
c).
During our review, it was noted that the nonheated thimbles were treated the same as heated surfaces in evalu_ating the axial friction factor.
- i Westinghouse reprogramed the friction factor model1to apply the isothermal friction factor to the unheated thimble walls. Similarly, it was noted t* at the nonisothermal friction factor model used the crud-metal interface temperature rather tha'n the crud-surface temperature. This model has been modified to use the
- nore appropriate crud-surface temperature.
The lateral resistance factor and the void fractiorJ correlations are part
.of the THINC-IV program and are discussed in.Ref.t3.
Sensitivity studies were performed using maximum anticipated perturbations of these correletions(7).
The results of these studies showed a negligible effect on the calculated minimum DNB ratios.
As a result, the staff finds these correlations acceptable for these hydraulic calculations.
The thermal mixing used in the THINC calsulations can vary becaus~e of the fuel assembly design.
Therefore, this thermal mixing is considered an input variable which should be justified for each application.
Similarly,'
hydraulic losses due to spacers and end fittings should be treated as input variables to the.THINC analysis.
2.3 Comcarisons with Ex erimental Data
~
Data obtained at a number of operating PWR plants were cdmpared with THINC-I and THINC-IV predictions of assembly exit temperatures (3,4). Several observations can be made about these ccmparisons. The most significant is that the THINC-I.and THINC-IV predictions are indistinguirinable. This is attributed to the fact that the predictions are best estimater and do not include an inlet plenum flow maldistribution penalty used in design calculations and the exit ficw conditions are all subcooled se that no ficw perturbations are caused by two-phase flow effects. - A,s a result, these comparisons cannot verify which code is more appropriate for calculating core-wide flow conditions.
Both codes overpredicted the exit temperatures in the highest pcwered assemblies, however, they un'derpredicted the enthalpy rise in slightly
~
lower powered assemblies by as much as 10'.'.
Although the results indicate c
that the calculations are conservative for,the highest pcwered assemblies, the staff feels that the tests were not sufficiently sensitive to c:nsider these results conclusive.
A more significant study was cerformed at Westinghouse using two Idxla red M cles in a 1cw :res: rs s " :'e-:hase flow icco.
Flew was res ri: tac a
- ne in'.e
- One f :36 n:L _:iss in: Fi:: -tu:e :enerses.ars 2:s 1:r:ss the two assemblies of different axial -locations.
In ene extrame case tne inlet ficw was totally blocked in one of the assemblies. 'A ::m;arison of L
4 s
7
~
~.
7, the average flow in each assembly compared favorably with the THINC-IV and THINC-II predictions, however, the THINC-I prediction was grossly in error.
The flow recovery predicted by THINC-IV lagged behind the data somewhat more than the'THINC-IIprediction.for the assembly with'the blocked inlet.
This response is-attributable to"the basic differences,in the two models, l
~
~
A more detailed THINC-IV analysis of the blocked inlet test treated the subchannels in each rod bundle individually.
A comparison with t
measured results showed that THINC-IV could not predict the axial flow gradient very well at the interface between red bundles. A separate analysis was performed by Westinghouse with a code that included the effects of momentum transfer between channels. This analysis predicted the fine flow structure at the rod bundle interface. This is consistent with the lag -in, flow recovery observed in the THINC predictions.
If the model included mcmentum interchange, the recovery would be improved because of the~ siignificant axial flow gradient at the interface. As noted in Section 2.1, this deficiency in th' model is not considered important for e
the less severe flow conditions encountered in PWR design calculations.
A THINC-IV prediction of jetting at a fuel assembly in1 t)was compared to,
9 subchannel flow data obtained at Combustion Engineeringle. The comparison showed excellent agreement for a very extreme flow test. The staff feels that these comparisons for such severe flow condit. ions are a good verification of the basic hydraulic model employed in THINC-IV.
Subchannel exit temperature da'ta were obtained from a 5x5 red bundle configuration with the inner nine rods having a 20% higher linear pcwer.
A comparison between these data and THINC-IV and THINC-II predictions shew reascnably good agreement ( 5%);.however, there was practically no difference between the two THINC calculations. As a result,.the staff does not consider these data a significant verification of the THINC-IV program.
Subchannel flow conditions (mass velocity and quality) were measured at the exit of selected subchannels in a 3x3 test section operating at low mass 6 lb/ft -hr) and quality (0 to 20%) at 1000 psi.
THINC-IV 2
velocities (n.10 and THINC-II predictions of these tests are not in goed agreement with the data (>10% variation); hcwever, the test conditions are not typical of ?WR flow conditions because of the presence of corner and side channels in
-the test section and high void fractions present.
For this reason, :ne staff.did not give. weight to these comparisons.
2.4 Desian Calculations The design calculations are performed in t ee steps. The first step examines the core-wide effect and treats each assembly as a flow channel and does not consider thertal mixing between channels.
The omission of ther al mixing in these calculations.is censervative because it neglects a e:nanism f:r reducing :ne en:ha.:y r,sa n tre not assa ciy.
i e-g yr m
~
.._. _ __. : _ _ c
,. _. _ g _ _..
L. (..
s The inlet flow rate to the-hot assembly is reduced fod the core-wide
~
calculations-to simulate inlet plenum maldistributions'. Although this
=isfan, input va'riable.. sensitivity studies were performed to demonstrate that~ the inlet conditions do not"have a significant effect on the
' calculated minimum DNS, ratios.
Similarly, sensitivity studies were performed to demonstrate the effect df core exit pressure distributions on the DNB calculations.
Recent experimental data have shown tnat these exit pressures are sufficiently large-and should be ac. counted for in the design calculations.
The second step in the calculation chain examines the' flow conditions in each of the four quadrants of the hot assembly.
This step is only needed if. the hot assembly is not located at the center of core where symmetry
. applies.
The third step examines the flow c6nditions in the hot channel by simulating each of the subchannels (flow area defined by four adjacent rods) in a
. quadrant of the hot assembly. Sensitivity studies showed that 20 axial segments are sufficient for the design calculations.
The design calculations performed to demonstrate plant performance include the following assumptions:
a.'
reduced effective core flow rate i
b.
reduced flow area in the hot channel c.
increased inlet temperature i
d.
increased local and bundle power generation e.
reduced flow in hot assembly f.
increased core power g.
reduced system pressure The combination of these assumptions results in conservative calculations of flow conditions at the luration of minimum CNB ratio.
Using typical design paradeters, Westinghouse calculated hot channel flow conditions and. minimum CNS ratios with THINC-I, -II, and -IV.
Although the flow rates along the hot channel varied considerably for the three calculations, the effect on minimum CNB ratio was quite modest.
The CNBR calculated with l
THINC-IV was higher than that calculated with THINC-I (large lateral l
resistance) and about.the same as that calculated with THINC-II (zero lateral resistance).
3.0 Staff Position The THINC-IV computer program is a reascnably formulated evaluation model for calculating the steady-state flow. conditions in a system of.ultiple
- arailsi
- hanr.els.
Favorable c:.:aris:r. :# 'M:. C-IY results witn ca a fr:m i
fi:w :as s wi:n severe cle:Ra:a arf :na ::r:sr.a-ive cissi:n :f in er-assembly thermal mixing are the major bas'es f:r finding -his code 32:e; a:le.
l l
t i
O
s s
=
( {.
Because of the small but finite effects of uncertainties in the hydraulic submodels used in THINC-IV, the staff considers that the code itself is more representative of a best-estimate analytical technique.
However, it is
. noted that the design ~calculatic.ns are performed using conservative assumptions with. respect to plant operating conditions, fuel fabrication tolerances, and power peaking ur.e rtainties.
As a result, the staff considers that the THINC-IV calections, using these input assumptions, are conservative.i We find that the THINC-IV code as described in WCAP-7956 (with the modifications 'noted in Section 2.2) is acceptable for performing steady-state core hydraulic calculations as described in WCAP-8054 (with the conservative assumptions noted above).
These calculations should be limited to conditions of single phase or homogeneous two-phase flow (such as bubble flow regime) because of the models used for lateral flow diversion and mixing.
1 m
C s'
j l
l l
l t
-=.--s-
-?
'( (
~
4.0 References 1.
C. W. Stewart, A Technical Rev.iew of the THINC-iV Thermal-Hydraulic Code,. FATE-75-110 (April 1975P.
2.
J. E.' Welch, et al.,'The MAC Method, A Computing Technique for Solving Viscous, Incompressible, Transient Fluid Flow Problems Involving Free Surfaces, LA-3425 (November 1965).
3.
H. Chelemer, et al., THINC-IV - An Improved Program for Thermal-Hydraulic Analysis' of Rod Bundle Cores WCAP-7956 (June 1973).
4.
T. Burke, et al., Analysis of Data from the Zion (Unit 1) THINC Verification Test, WCAP-8453 (December 1974).
5 ~. R. Marshall and R. Latendre, Influence of Inlet Geometry en Flow in
.the Entrance Region of a Nuclear Reactor Rod Bundle, ASME Paper 68WA/MT-34(1968).
E. Novendstern and R. Sandberg, Single Phase, local Boiling)and Bulk 6.
Boiling Pressure Drop Correlations, WCAP-2850-L (April 1966.
7.
Letter frem C. Eicheldinger (W) to J. Stolz (NRC) dated June 29, 1977 (NS-CE-1473).
5.0 Additional References used in the Review 1.
Letter from C. Eicheldinger (W) to D. Vassallo (NRC) dated March 10, 1975 (NS-CE-577).
2.
Letter frem C. Eicheldinger (W) to D. Vassallo (NRC.) dated April 25, 1975 (NS-CE-626).
3.
Letter from C. Eicheidinger (W) to D. Vassallo I(NRC) dated July 30, 1975 (NS-CE-730)..
4.
Letter from C. Eicheldinger (W) to 0. Vassallo (NRC) dated November 19, 1975 (NS-CE-844) 5.
Letter frem C. Eicheidinger (W) to D. "Vassallo (NRC) dated June 11, 1976 (NS-CE-1100).
c 6.
LetterfromJ.Cuta'(PNL)toS~.IsraellbC)datedMarch3,1976.
7.
Note from J. Shefcheck (W) to S. Israel (NRC) dated March 15, 1976.
,.