ML16258A104
| ML16258A104 | |
| Person / Time | |
|---|---|
| Site: | South Texas, Perkins  |
| Issue date: | 09/12/2016 |
| From: | Richards D South Texas |
| To: | Lisa Regner Plant Licensing Branch IV |
| References | |
| MF2400, MF2401 | |
| Download: ML16258A104 (20) | |
Text
NRR-PMDAPEm Resource From: Richards, Drew <amrichards@STPEGS.COM>
Sent: Monday, September 12, 2016 12:56 PM To: Regner, Lisa Cc: Harrison Albon; Blossom, Steven
Subject:
[External_Sender] FW: SNPB-3-13 follow up - CCFL Attachments: SNPB-3-13 follow up - CCFL - revised.docx
- Lisa, See attached for your review.
- Thanks, drew 361.972.7666
Original Message-----
From: Vaghetto, Rodolfo [1]
Sent: Monday, September 12, 2016 11:22 AM To: Ernie Kee <erniekee@gmail.com>; Harrison Albon <awharrison@STPEGS.COM>; Blossom, Steven
<sdblossom@STPEGS.COM>; Richards, Drew <amrichards@STPEGS.COM>
Subject:
Re: SNPB-3-13 follow up - CCFL
The revised version of the SNBP 3-13 is attached 1
Hearing Identifier: NRR_PMDA Email Number: 3058 Mail Envelope Properties (d294a08b271d442084723fc52cbebc39)
Subject:
[External_Sender] FW: SNPB-3-13 follow up - CCFL Sent Date: 9/12/2016 12:55:51 PM Received Date: 9/12/2016 12:56:07 PM From: Richards, Drew Created By: amrichards@STPEGS.COM Recipients:
"Harrison Albon" <awharrison@STPEGS.COM>
Tracking Status: None "Blossom, Steven" <sdblossom@STPEGS.COM>
Tracking Status: None "Regner, Lisa" <Lisa.Regner@nrc.gov>
Tracking Status: None Post Office: CMBXEXCH01.CORP.STPEGS.NET Files Size Date & Time MESSAGE 682 9/12/2016 12:56:07 PM SNPB-3-13 follow up - CCFL - revised.docx 1022967 Options Priority: Standard Return Notification: No Reply Requested: Yes Sensitivity: Normal Expiration Date:
Recipients Received:
Counter-Current Flow Limitation (CCFL) Correlation CCFL is a high-ranked phenomenon that requires empirical support; the STP LTCC EM model CCFL correlation uses bounding values found by examining a large parameter space that covers the conditions encountered in the LTCC simulations.
The simulations performed with the LTCC EM show that, during the post-core blockage phase, conditions for counter-current flow limitation (CCFL) at the top of the core may occur (see Appendix A). During this phase, vapor produced leaves the core by flowing upward through the core outlet, while liquid water (reaching the top of the core through alternative flow paths) moves downward toward the core. Conditions that affect the CCFL at the top of the core include the liquid and vapor velocities, the liquid and vapor properties, and the geometry.
This condition is considered one of the most important phenomena during the Post-Core Blockage LTCC phase as it affects the behavior of the liquid entering the top of the core during this phase and, subsequently, the core coolability.
The response is divided into four sections:
Model
Description:
First, a general description of the correlations available in RELAP5-3D and their formulation is provided.
Assumptions and Boundary Conditions: Description of how the models are used in the LTCC EM, and the assumptions made on the CCFL coefficients is then described, to confirm the appropriateness of the boundary conditions adopted in the LTCC EM for the simulations performed.
Code Validation: At the end, a list of validation cases performed by the code developers is included.
==
Conclusions:==
Final conclusions are stated at the end based on the evidence provided in the response.
Model Description RELAP5-3D code includes three forms of the CCFL correlation - Wallis form [1], Kutateladze form, and Benkoff [2] form - a form in between the Wallis and Kutateladze forms and used in TRAC-PF1 code.
A description of these models and their formulation in RELAP5-3D is provided below.
The CCFL correlations can be expressed in the following general equation [2]:
H 1/g 2 + mH 1/f 2 = c (1) where Hg is the dimensionless vapor/gas flux, Hf is the dimensionless liquid flux, c is the vapor/gas intercept (value of H 1/g 2 when Hf = 0, i.e., complete flooding), and m is the (negative) slope, that is the vapor/gas intercept divided by the liquid intercept (the value of H 1/f 2 when Hg = 0). A typical plot of H 1/g 2 versus H 1/f 2 is shown in Figure 1. The dimensionless fluxes have the form 1/ 2 g
H g = jg (2) gw( f g )
1/ 2 f
H f = jf (3) gw( f g )
where j g is the vapor/gas superficial velocity ( g vg ), j f is the liquid superficial velocity ( f v f ), g is the vapor/gas density, g is the liquid density, g is the vapor/gas volume fraction, f is the liquid volume fraction, g is the gravitational acceleration, and w is the length scale and is given by the expression w = D1j L (4) where is a user-input constant. Also, D j is the junction hydraulic diameter and L is the Laplace capillary length constant, given by 1/ 2 L= (5) g ( f g ) ,
where is the surface tension.
In Equation (4), can be a number from 0 to 1. For = 0 , the Wallis form of the CCFL equation is obtained; and for = 1 , the Kutateladze form of the CCFL equation is obtained. For 0 < < 1 , Bankoff form in between the Wallis and Kutateladze forms is obtained.
Figure 1. Plot of H 1/g 2 versus H 1/f 2 for a typical CCFL correlation.
First, according to Bankoff et al.s guidelines, coefficients , c, and m in the generalized CCFL model above (Equations (1) through (5)) can be estimated based on experimental data:
= tanh( kc D j ) , (6) m =1, (7) 1.07 + 4.33 x 10 D D < 200 3 *
- c= (8) 2 D* 200 where the critical wave number kc = 2 / t corresponds to the maximum wavelength that can be sustained on an interface of length t (the plate thickness), and is the perforation ratio (fraction of plate area
occupied by holes); and D* is a Bond number defined as 1/ 2 g ( f g )
D* = n D , (9) with n the number of holes and D the plate diameter.
The formulation of the coefficients listed in Equations (6) - (9), and the models adopted in RELAP5-3D, is similar to the approach adopted in TRACE computer code and in the TRACE code manuals in regard to the CCFL [11].
Assumptions and Boundary Conditions Figure 2 presents various CCFL experimental data for a perforated plate with different hole diameters (d),
plate thicknesses (t), number of holes (n), and pitches (p) in the graph of j1/g 2 vs. j1/f 2 .
Figure 2. CCFL data for various conditions (reproduced from [4])
The data indicates that the number of holes is the most critical factor determining the accessibility area, while other factors are relatively less important.
As the number of holes in the plate increases, the experimental data tend to shift upward (i.e., toward larger c) maintaining the same slope (m ~ 1). This is the case for the data produced by Bankoff [3] (n=2, 3, and 15), Sobajima [6] (n=25), and Kokkonen [8] (n=52).
In the same figure, the condition investigated by Wallis (vertical pipe, sharp-edged, c=0.725, m=1) is also shown (experimental points not included).
It can be noted that for perforated plates with large number of holes, experimental data may be bounded by the conditions (c=1, m=1). This condition is the one adopted in the LTCC EM.
Additional sensitivity analysis is conducted to include limiting cases for the CCFL. In particular, the coefficients proposed by Wallis (developed for vertical sharp-edged pipes, c=0.725, m=1) to account for edge effects) appear to be the most limiting conditions, as shown in Figure 2.
A list of conditions applied to the simulations performed is included in Table 11.
Table 1. CCFL Conditions used (LTCC EM and Sensitivities)
Case c m Notes Wallis - Smooth-LTCC EM - Base 0 1 1 edges. Bounding condition Bankoff - STP Sensitivity 1 0.038422 2 m Geometry Wallis - Sharp-Sensitivity 2 0 0.725 1 edges. Worst Condition The conditions calculated with Equations (6) - (9) for the STP geometry (Sensitivity 2) put in evidence the following:
Coefficient is close to zero. This can be interpreted as a negligible effect of the surface tension for this geometry. Subsequently the Bankoff model becomes similar to the Wallis model.
The value of intercept c is larger than the condition proposed by Wallis for smooth-edged pipes, confirming that the conditions assumed for the LTCC EM - Base are bounding the expected phenomenon in the STP top core geometry.
The Sensitivity 2 was included in the list of sensitivities since it includes the worst conditions for the CCFL, by minimizing the accessibility area. Although these conditions do not realistically represent the geometry under consideration, the simulation results show sufficient liquid flow through the top of the core, maintaining adequate core cooling.
Figure 3 graphically represents these condition together with the available experimental data.
1 Sensitivities are conducted starting from the 16 HLB scenario included in the LTCC EM, by simply changing the CCFL coefficients at the core outlet junction.
Figure 3. CCFL Coefficients - LTCC EM and Sensitivities The line representing the conditions used in the LTCC EM (c = 1 and m=1) it can be regarded as a conservative choice for medium and high number of holes, minimizing the accessibility area.
The conditions used in the Sensitivity 2 (c= 0.725, and m = 1) recommended by Wallis [1] for the sharp-edged pipes - is observed to cover all the CCFL data, and thus, this line can be regarded as the most conservative condition (with the smallest range of the accessibility of CCFL).
The PCT (Figure of Merit) as a function of the time during the Post-Core Blockage LTCC for the cases listed in Table 1 is plotted in Figure 4. The figure shows that under the conditions imposed in the LTCC EM (c=1, m=1), and also under worst conditions of the Sensitivity 2 (c=0.725, m=1), the PCT is maintained well below the limit of 800 °F.
Figure 4. Simulation Results - PCT Code Validation Two separate effects tests are used in the RELAP5-3D code assessment, for the CCFL model. These tests and the outcome of the code comparison is briefly discussed below. Additional detail on the facility description, RELAP5-3D model and results is provided in Appendix A and B respectively (see also volume 4 of the RELAP5-3D users manual).
Dukler-Smith air-water flooding CCFL in vertical tube (See Appendix B or Section 4.15 in manual Vol.3 for the details.)
Upper plenum test facility (UPTF) Test 6, Run 131 downcomer CCFL (See Appendix C or Section 4.16 in manual Vol.3 for the details.)
The Dukler-Smith air-water flooding experiments assessed the Wallis CCFL model. Semi- and nearly-implicit predictions were in reasonable agreement with the Dukler-Smith experiment data for countercurrent air-water flow in a single tube over the range of air flows from 0.0126 to 0.126 kg/s. The assessment also shows that the Wallis correlation is implemented correctly.
UPTF downcomer CCFL Test 6, Run 131 was used to compare the relative performance of the annulus and pipe components for simulating the refill of the lower plenum during a loss-of-coolant accident. The two components are similar except that all the liquid is placed in the film, with no liquid allowed in drops, in the annulus component when in the annular-mist flow regime. Liquid is allowed in both the film and drops in the annular-mist flow regime in the pipe component. Both the semi- and nearly implicit calculations were judged to be in reasonable agreement with the measured liquid level data for UPTF Test 6, Run 131. The calculated refill was similar to that observed in the test. The RELAP5-3D calculation in which the downcomer was modeled with annulus components was in better agreement with the measured results than when pipe components were used. The flow regime model in the annulus component, which puts all the liquid in the film, resulted in a better prediction of the lower plenum refill for the UPTF test.
The pipe component provided a conservative prediction of the amount of liquid in the lower plenum.
Conclusions The CCFL conditions is identified as one of the most important phenomena during the Post-Core Blockage Phase. Due to the relatively high upward vapor velocity at the top of the core right after core blockage, conditions for CCFL may be expected in this region. CCFL closure relationships are available in RELAP5-3D. Assessment of the code capabilities and in predicting this phenomenon and the correctness of the use of such correlations was conducted and described. The assessment has confirmed that:
- 1) RELAP5-3D includes models for predicting the CCFL phenomenon. These models and the implementation/formulation approach adopted by the code developers are similar of the ones included in the TRACE computer code.
- 2) The code validation performed by INL confirmed that the correct implementation of these correlations.
- 3) The use of these correlations and the related coefficients in the LTCC EM is confirmed to be adequate for the for geometry and the conditions under consideration. In particular, the coefficient included in the LTCC EM are bounding experimental data for perforated plates.
- 4) Sensitivity study is conducted to confirm that margin to the boundary conditions included in the EM is accounted.
The proposed LTCC EM is adequate to predict the expected phenomena of CCFL at the top of the core for the conditions expected during the Post-Blockage LTCC.
References
[1] G. B. Wallis, One-dimensional Two-phase Flow, McGraw-Hill, New York, 336-345, 1969.
[2] S. G. Bankoff and S. C. Lee, Multiphase Science and Technology, Vol. 2 (Edited by G.F. Hewitt, J.M.
Delhaye, N. Zuber), Chapter 2, A Critical Review of the Flooding Literature, Hemisphere, New York (1985).
[3] S. G. Bankoff, R. S. Tankin, M. C. Yuen and C. L. Hsieh, Countercurrent Flow of Air/Water and Steam/Water Through a Horizontal Perforated Plate, Int. J. Heat Mass Transfer, 24, 8, 1381-1395 (1981).
[4] H. C. No, K.-W Lee, C.-H. Song, An Experimental Study of Air-water Countercurrent Flow Limitation in the Upper Plenum with a Multi-hole Plate, Nucl. Eng. Tech. 37, 557-564 (2005).
[5] H. M. Lee, G. E. McCarthy and C. L. Tien, Liquid Carryover and Entrainment in Air-water Countercurrent Flooding, EPRI report, NP-2344 (1982).
[6] M. Sobajima, Experimental Modeling of Steam-Water Countercurrent Flow Limit for Perforated Plates, J. Nuclear Science and Technology, 22, 9, 723-732 (1985).
[7] G. P. Celata, N. Cumo, G. E. Farello, and T. Setaro, The Influence of Flow Obstructions on the Flooding Phenomenon in Vertical Channels, Int. J. Multiphase Flow, 15, 2, 227-239 (1989).
[8] I. Kokkonen and H. Tuomisto, Air/Water Countercurrent Flow Limitation Experiments with Full-Scale Fuel Bundle Structures, Experimental Thermal and Fluid Science, 3, 581-587 (1990).
[9] J. Zhang, J. M. Seynhaeve and M. Giot, Experiments on the Hydrodynamics of Air-Water Countercurrent Flow Through Vertical Short Multitube Geometries, Experimental Thermal and Fluid Science, 5, 755-769 (1992).
[10] C. L. Tien, K. S. Chung, and C. P. Lin, Flooding in Two-Phase Countercurrent Flows, EPRI NP-1283, December 1979.
[11]. ML071000097, TRACE V5.0, Theory Manual - Fields Equations, Solution Methods, and Physical Models
Appendix A. Dukler-Smith Air-Water Flooding (Section 4.15 in Manual Vol.4)
Dukler and Smith [A1] conducted a simple flooding experiment at the University of Houston to study the interaction between a falling liquid film with an upflowing gas core. A RELAP5 model for the Dukler-Smith countercurrent flow limitation (CCFL) test facility was developed for earlier assessments using these experiment data. The work of Riemke [A2] and Davis [A3] is representative of these earlier assessments. This work draws heavily on the earlier assessments, basically repeating the assessment for the latest code version.
Experimental facility description A schematic of the Dukler-Smith experiment facility is shown in Figure A-1(a). The flow system consisted of a 1.52-m (5-ft) length of 0.051-m (2-in.) inner diameter Plexiglas pipe used as a calming section for the incoming air, a 0.305-m (1-ft) diameter section of Plexiglas pipe for both introducing the air to the test section and removing the falling liquid film, a 3.96-m (13-ft) test section consisting of 0.051-m (2-in.) diameter Plexiglas pipe, and an exit section for removing the air, entrainment, and the liquid film flowing up. Measurements were taken of the pertinent flow rates, pressure gradients, and the liquid film thickness over a wide range of gas and liquid flow rates in the flooding region. The liquid film upflow, downflow, and entrainment rates were determined by weighing the liquid flow for a fixed period of time (see discharge lines to weigh tanks labeled B in Figure A-1(a)). Most of the instantaneous measured parameters oscillated once quasi-steady state conditions were reached, and it was necessary to time-average these parameters. Dukler and Smith indicate that the CCFL process is basically an unstable process that is driving the oscillations. In the RELAP5-3D simulations the air and water flow predictions at the measuring point also showed oscillations. The predictions were averaged over 30 s for purposes of showing comparisons to the data.
Input model description The experiment was modeled using the nodalization shown in Figure A-1(b). The air injection is specified by a time-dependent junction (Component 102) to match the experiment value. The homogeneous (single-velocity momentum equation) option was specified at the air injection point (Junction 10103) to prevent liquid from flowing down the air injection pipe. The falling liquid film drained through Junction 10102.
Inlet liquid flow rate was also specified by a time-dependent junction (Component 106) to match the measured value. The falling liquid film drained through a time-dependent junction (Component 195) where the outlet flow was set by a control system to maintain a fixed level in Pipe 190. A pressure of 0.104 MPa was specified for the drain tank (Component 200). Pressure in the test rig was controlled to 0.1 MPa by a time-dependent volume (Component 110).
Dukler discussed more than one CCFL correlation, but the one that appeared to be best for his test is a Wallis [A4] form of the correlation: j1/g 2 + mj1/f 2 = c , where j is defined as the non-dimensional superficial velocity, subscripts g and f refer to gas and liquid respectively, m is the slope, and c is the gas intercept constant. This correlation was found to be reasonable for air/water systems where standing waves appeared on the surface of the liquid film. Dukler found this to occur in his experiment. Wallis4.15-4 indicated that m = 1 and C varied between 0.88 and 1.0 for small diameter round tubes. The RELAP5-3D input model activated the CCFL model at the junction between Components 104 and 105. The CCFL input data for this junction used the following values: junction hydraulic diameter = 0.0508 m, flooding correlation form =0.0 (Wallis CCFL form), gas intercept c = 0.88, and slope m = 1.0.
(a) (b)
Figure A-1: (a) Schematic of the Dukler-Smith Air-Water Test Facility (from [11]) and (b) nodalization diagram for the Dukler-Smith test facility (from RELAP manual)
Data Comparisons and Results The requested time step size for both the semi-implicit and nearly-implicit advancement scheme calculations was 0.005 s. Calculations were run with values of liquid and air injection flows consistent with the data. Figure A-2(a) compares the calculated liquid downflow rates with data at the given liquid injection flow rate. Good agreement with the data is observed in the predictions with both the semi-implicit and nearly-implicit advancement schemes. At the higher liquid injection rates, the calculated liquid downflow was less than the data, indicating more of the injected liquid was entrained and exited through the top. A possible reason for the under-calculated liquid downflow is that the values for the gas intercept and slope do not fit the data. As shown in Figure A-2(b), the semi-implicit calculated results are in excellent agreement with the flooding correlation of Wallis as the x-intercept (square root of superficial vapor velocity) for the predictions is 0.88 and the slope is 1.0. Thus the code is working properly based on the intercept and slope values input to the model.
(a) (b)
Figure A-2. Comparison of RELAP5-3D predictions to Dukler-Smith data: (a) liquid mass flow vs. gas mass flow and (b) superficial liquid velocity vs. superficial gas velocity Conclusions and Assessment Findings RELAP5-3D predictions are in reasonable agreement with the Dukler-Smith experiment data for countercurrent air-water flow in a single tube over the range of air flows from 0.0126 to 0.126 kg/s. The assessment also shows that the Wallis correlation is implemented correctly.
References
[A1] A. E. Dukler and L. Smith, Two Phase Interactions in Counter-Current Flow: Studies of the Flooding Mechanism, NUREG/CR-0617, January 1979.
[A2] R. A. Riemke, Countercurrent Flow Limitation Model for RELAP5/MOD3, Nuclear Technology, Vol. 93, pp 166-173, February 1991.
[A3] C. B. Davis, Validation Report: RELAP5-3D Flooding Model, Code Version 1.3.5, R5/3D-01-05, October 2, 2001.
[A4] G. F. Hewitt and G. B. Wallis, Flooding and Associated Phenomena in Falling Film Flow in a Tube, UKAEA Report AERE-R 4022, 1963.
Appendix B. UPTF Downcomer CCFL Test 6, Run 131 (Section 4.16 in Manual vol.3)
Experiments were performed in the Upper Plenum Test Facility (UPTF) to obtain full-scale data on downcomer/lower plenum refill behavior during a loss-of-coolant accident initiated by a large break. The experiments provided a counterpart to testing that was done previously in scaled facilities.
Code Models Assessed The relative performance of the annulus and pipe components for simulating the refill of the lower plenum during a loss-of-coolant accident was compared. The two components are similar except that all the liquid is placed in the film, with no liquid allowed in drops, in the annulus component when in the annular-mist flow regime. Liquid is allowed in both the film and drops in the annular-mist flow regime in the pipe component.
Experiment Facility Description UPTF is a full-scale model of a four-loop 1300-MWe pressurized water reactor (PWR), including the reactor vessel, downcomer, lower plenum, upper plenum, and coolant loops. Simulators are used to represent the core, primary coolant pumps, steam generators, and containment. A schematic view of the test facility is shown in Figure B-1. Key dimensions are presented in Figure B-2. The test vessel, core barrel, and internals are full-size representations of a PWR, with four full-scale hot and cold legs that simulate three intact loops and one broken loop. Figure B-3 shows the positions of the four loops relative to the downcomer.
Figure B-1. Schematic of the UPTF.
Figure B-2. Key dimensions of the UPTF Figure B-3. Cross-section of the UPTF reactor vessel.
Test 6 [B1, B2] was a quasi-steady experiment that was carried out to obtain full-scale data on downcomer/lower plenum refill behavior. Predetermined steam and emergency core cooling (ECC) water flow rates were injected into the system to determine the penetration of ECC water into the downcomer and lower plenum as a function of steam flow up the downcomer. Run 131 was selected for analysis. The system was initially filled with slightly superheated steam at about 2.5E5 Pa. The test was initiated by starting the steam flow from the core and steam generator simulators. ECC injection into the cold legs of the three intact loops began about 12 s later. The temperature of the ECC water was initially near saturated conditions. However, the steam flow caused the pressure to increase during the test, which caused the subcooling to increase. About 1 kg/s of nitrogen was injected along with the ECC to simulate noncondensable coming out of solution. The injection flow rates were terminated near 80 s. The pump simulators were closed during the test so that all the injected steam had to flow through the downcomer.
Boundary conditions for Run 131 are summarized in Table B-1.
Table B-1. Summary of Test 6, Run 131 boundary conditions.
Input Model Description The RELAP5-3D nodalization used to simulate Run 131 is shown in Figure B-4. The model explicitly represented all four coolant loops. The break (Component 505) connected Loop 4 to the containment simulator (Component 599). The downcomer was divided into two halves, with Components 111 and 112 connected to the broken side (Loops 1 and 4) while Components 121 and 122 were connected to the intact side (Loops 2 and 3). These downcomer flow paths were connected in crossflow using single and multiple junctions (Components 118 and 119). The lower plenum was divided axially into two control volumes (Components 150 and 160), each containing approximately the same fluid volume. The core and hot legs were combined into a single volume (Component 180). A time-dependent junction (Component 198) supplied steam flow to the core. The ECC and nitrogen flows were supplied by time-dependent junctions (Components 398, 498, and 698).
Standard code options were applied except that the choking model was turned off at all junctions except for the break because of the low pressure at which the test was conducted (in order to prevent unphysical choking at the other junctions).
Figure B-4. RELAP5-3D nodalization for UPTF Test 6.
Data Comparisons and Results Three RELAP5-3D calculations were initially performed, each with a requested time step of 0.01 s. The first two calculations used the semi-implicit numerical scheme. The downcomer was modeled with four annulus components (111, 112, 121, and 122) in the first calculation and four pipe components in the second calculation. The third calculation was identical to the first one except that it used the nearly-implicit numerical scheme. Figure B-5 shows the calculated pressure in the downcomer during the test.
The initiation of steam flow, which corresponds to 0 s on the figure, caused the pressure to increase. The pressure increased again near 14 s when ECC reached the break, which reduced the volumetric flow out the break. The termination of steam and ECC flows near 80 s caused the pressure to decrease. Measured results are not presented because only limited data for UPTF Test 6 are publicly available.
Figure B-5. Calculated downcomer pressure in UPTF Test 6, Run 131.
The capability of the code to calculate the refill of the lower plenum is illustrated in Figure B-6, which shows measured and calculated collapsed liquid levels. The calculated values were obtained from the total liquid volume in the lower plenum and converted to liquid levels after accounting for the curvature of the lower head as well as the internals in the lower plenum. This method accounts for the varying flow area as a function of height and thus allows a more realistic indication of the liquid level than the traditional collapsed liquid level, which is obtained as the liquid volume fraction times the height summed over the number of volumes. The discussion will initially concentrate on the first calculation, which used annulus components and the semi-implicit numerical scheme. The gradual increase in the calculated liquid level prior to 15 s was due to the accumulation of droplets that were formed by condensation of the injected steam from the core. ECC first reached the lower plenum at 15.8 s. The liquid level increased relatively rapidly until 40 s, when the rate of increase decreased significantly until the steam and ECC injection ended near 80 s. The ending of the injection caused the pressure to fall as shown previously in Figure 4.16-5. The subsequent flashing in the lower plenum caused a reduction in the liquid level as the steam produced carried liquid from the lower plenum to the break. The calculated behavior was generally similar to the experiment except that the water began to reach the lower plenum about 6 s earlier than in the test and the level decrease after 80 s was much more pronounced than in the test. The calculated and measured rates of level increase were similar during the refill period. The increase in the indicated level at the start of the test is attributed to the effect of the steam flow on the differential pressure taps, rather than the presence of actual liquid because ECC flow did not begin until about 12 s. The different flow regime model used in the pipe component resulted in a delay in the liquid reaching the lower plenum and a substantially slower rate of refill. The calculated results with the nearly- and semi-implicit numerical schemes were similar.
Figure B-6. Measured and calculated collapsed liquid level in the lower plenum in UPTF Test 6, Run 131.
Figures B-7 and B-8 show the calculated mass flow rate and fluid density in the broken cold leg, respectively. The figures indicate that ECC first reached the break near 14 s. Thereafter the mass flow and density increased substantially due to the bypass of ECC. More bypass was initially obtained in the calculation with the pipe component. The flow rate and density also increased substantially in the two annulus calculations when the injection flow rates were terminated near 80 s. The flashing in the lower
plenum that was caused by the pressure decrease caused liquid to be entrained from the lower plenum to the break, resulting in an increase in the flow rate and density in the broken loop.
Figure B-7. Calculated break mass flow in UPTF Test 6, Run 131.
Figure B-8. Calculated fluid density in the broken cold leg in UPTF Test 6, Run 131.
An additional sensitivity calculation was performed to investigate the effects of the lower plenum nodalization. In this sensitivity calculation, the lower plenum was modeled with one control volume (Component 150) rather than the two volumes used previously. Figure B-9 shows that the initial refill of the lower plenum was similar with both models. However, the refill of the lower plenum slowed earlier
when the single volume was used. The two calculations bracketed the data between 40 and 80 s, with the single volume lower plenum under predicting the level and the two-volume model over predicting it. The sensitivity calculation demonstrates that the total amount of liquid stored in the lower plenum at the end of the refill period depends on the nodalization.
Figure B-9. The effect of nodalization on collapsed liquid level in UPTF Test 6, Run 131.
Conclusions and Assessment Findings The RELAP5-3D calculations are judged to be in reasonable agreement with the measured liquid level data for UPTF Test 6, Run 131. The calculated refill was similar to that observed in the test, but started about 6 s earlier.
The RELAP5-3D calculation in which the downcomer was modeled with annulus components was in better agreement with the measured results than when pipe components were used. The annular mist flow regime model in the annulus component, which puts all the liquid in the film, resulted in a better prediction of the lower plenum refill for the UPTF test. The pipe component provided a conservative prediction of the amount of liquid in the lower plenum. The semi- and nearly-implicit calculations produced similar results. The liquid inventory in the lower plenum at the end of the refill period depends on the nodalization.
References
[B1] J. Liebert and P. Weiss, UPTF Experiment Effect of Full-Scale Geometry on Countercurrent Flow Behaviour in PWR Downcomer, Proceedings of Fourth International Topical Meeting on Nuclear Reactor Thermal-Hydraulics, NURETH-4, Karlsruhe, F.R.G., October 10-13, 1989, Volume 1, pp. 67 -
74.
[B2] H. Glaeser, Downcomer and tie plate countercurrent flow in the Upper Plenum Test Facility (UPTF), Nuclear Engineering and Design, 133 (1992), pp. 259-283.