ML20212G319
| ML20212G319 | |
| Person / Time | |
|---|---|
| Issue date: | 09/30/1999 |
| From: | Ishii M, Soon Kim, Richard Lee, Revankar S, Tinkler C, Wu Q, Zhang G NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES), PURDUE UNIV., WEST LAFAYETTE, IN |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-L-1990 NUREG-CR-6510, NUREG-CR-6510-V01, PU-NE-96-2, NUDOCS 9909290193 | |
| Download: ML20212G319 (190) | |
Text
{{#Wiki_filter:_ 7 - - - - - - - - - - - - - - - - - - - _ NUREG/CR-6510, Vol.1 PU NE-96/2 l Corium Dispersion in L' Direct Containment Heating F Separate Effect Experiments With Water and Woods Metal Simulating Core Melt for Zion Reactor Conditions Purdue University M
-s/
M L [')DY , ! I # ' U.S. Nuclear Regulatory Commission r / Omce of Nuclear Regulatory Research ! j Washington, DC 20555-0001 ,,,,/
] 990930
< 11 PDR
NUREG/CR-6510, Vol.1 PU NE-96/2 l'E Corium Dispersion in Direct Containment Heating [ Separate Effect Experiments With Water and Woods Metal Simulating Core Melt for Zion Reactor Conditions Purdue University M
/
=
U.S. Nuclear Regulatory Commission ,7 '\ i Office of Nuclear Regulatory Research i Washington, DC 20555-0001 ,,,,, n ! L 18 " PDR i -
AVAILABILITY NOTICE Availability of Reference Materials Cited in NRC Publications NRC publications in the NUREG series, NRC regu- NRC Public Document Room I lations, and Title 10, Energy, of the Code o/ Federal 2120 L Street, N.W., Lower Level Regulations, may be purchased from one of the fol- Washington, DC 20555-0001 lowing sources: < http://www.ntc. gov /NRC/PD R/pdr1.htm >
" ~
- 1. The Superintendent of Documents U.S, Govemment Printing Office Microfiche of most NRC documents made publicly RO. Box 37082 available since January 1981 may be found in the Washington, DC 20402-9328 Local Public Document Rooms (LPDRs) located in
< http://www. access.gpo. gov /su_ docs > the vicinity of nuclear power plants. The locations 202- 512 -1800 of the LPDRs may be obtained from the PDR (see 2, previous paragraph) or through:
The National Technical Information Service Springfield, VA 22161 -0002 <http://www.nrc. gov /NRC/NUREGS/
<http://www.ntis. gov /ordernow> SR1350/V9/lpdr/html>
703-487-4650 Publicly released documents include, to name a The NUREG series comprises (1) brochures few, NUREG-series reports; Federal Register no-(NUREG/BR-XXXX), (2) proceedings of confer- tices; applicant, licensee, and vendor documents ences (NUREG/CP-XXXX), (3) reports resulting and correspondence; NRC correspondence and from international agreements (NUREG/lA-XXXX), internal memoranda; bulletins and information no-(4) technical and administrative reports and books tices; inspection and investigation reports; licens- [(NUREG-XXXX) or (NUREG/CR-XXXX)], and (5) ee event reports; and Commission papers and compilations of legal decisions and orders of the their attachments. Commission and Atomic and Safety Licensing Boards and of Office Directors' decisions under Documents available from public and special tech-Section 2.206 of NRC's regulations (NUREG. nicallibraries include all open literature items, such XXXX), as books, journal articles, and transactions, Feder-al Register notices, Federal and State legislation, A single copy of each NRC draft report is available and congressional reports. Such documents as free, to the extent of supply, upon written request theses, dissertations, foreign reports and transla-as follows: tions, and non-NRC conference proceedings may be purchased from their sponsoring organization. Address: Office of the Chief Information Officer Reproduction and Distribution Copies of industry codes and standards used in a Services Section substantive manner in the NRC regulatory process U.S. Nuclear Regulatory Commission are maintained at the NRC Library, Two White Flint Washington, DC 20555-0001 North 11545 Rockville Pike, Rockville, MD E-mail: < DISTRIBUTION @nrc. gov > 20852-2738. These standards are available in the Facsimile: 301- 415- 2289 library for reference use by the public. Codes and A portion of NRC regulatory and technicalinforma- standards are usually copyrighted and may be tion is available at NRC's World Wide Web site: purchased from the originating organization or, if they are American National Standards, from-
<http://www.nrc. gov >
American National Standards Institute A!! NRC documents released to the public are avail- 11 West 42nd Street able for inspection or copy:ng for a fee, in paper, New York, NY 10036-8002 microfiche, or, in some cases, diskette, from the < http://www. ansi.org > Public Document Room (PDR): 212-642-4900 DISCLAIMER This report was prepared as an account of work sponsored by any legal habarty or responsibahty for any third party's use, or the an agency d the Unned States Government. Neither the United results of such use, of any information, apparatus, product, or States Govemment ror any agency thereof, nor any of their em- process disclosed in this report, or represents that its use by ployees, makes any warranty, expressed or imphed, or assumes such third party would not infnnge pnvately owned nghts.
NUREG/CR-6510, Vol.1 PU NE-96/2 Corium Dispersion in Direct Containment Heating Separate Effect Experiments With Water and Woods Metal Simulating Core Melt for Zion Reactor Conditions Manuscript Completed: July 1999 Date Published: September 1999 Prepared by M. Ishii, Q. Wu, O. Zhang, Purdue University R.Y. Lee, C.O.Tinkler, Nuclear Regulatory Commission Schoolof Engineering Purdue University West Lafayette,IN 47907 R.Y. Lee, NRC Project Manager Prep red for Divisi:n of Systems Analysis and Regulatory Effectiveness Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Wcshington, DC 20555-0001 NRC Job Code L1990
,- m W..... )
l I NUREG/CR-6510, Vol. I bas been reproduced from the best available copy. 1 1 l T
ABSTRACT The research conducted at Purdue University addresses corium dispersion during a Direct Containment Heating (DCH) scenario in a severe nuclear reactor accident. In a DCH event, the degree of corium dispersion has a large effect on the resultant containment pressurization. In view of this, a separate effect test program on the corium dispersion mechanisms in the reactor cavity and on the subcompartment trapping mechanisms was initiated in 1992 at Purdue University under the direction of the US Nuclear Regulatory Commission. The four major objectives of this study are: (1) to perform a detailed scaling study using the newly proposed step-by-step integral scaling method, and to evaluate existing models and correlations for droplet entrainment, particle size and size distribution and particle trapping, (2) to design and construct a 1/10 scale Zion reactor model, and to perform carefully scaled experiments using air-water and air-woods metal to simulate the prototypic steam and core melt, (3) to develop reliable mechanistic models for the corium dispersion and transport in the accident scenario, which can be used to predict the liquid and gas blowdown, entramment, droplet size, liquid carryover to the containment, and the subcompartment trapping, and (4) to use the models to perform stand alone calculations for the prototypic conditions. In this report, efforts are focused on the first two objectives, whereas the modeling study is covered in a separate report. iii NUREG/CR-6510
CONTENTS ABSTRACT..................................................................................................................iii CONTENTS..................................................................................................................v LI ST O F TAB L ES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . LI ST OF FIGURES . ... . .. . . . . .. . .. . . . . . . . . . . . .. .. . . . . . . . . . . . . . . .. . . .... . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . NOMENCL ATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . EXEC UTIVE S UMMARY . . . . . . . . . . . . . . . . . . ... . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . ACKNO WLEDGMENTS . . . . . .. . .. . . . . .. .. .. . . . . . . . . . . . .. . . .. . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . . , J
- 1. INTRO DU CTI ON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
- 2. SCALING OF DCH PHENOMENA .... .. .. ............... ... ................... ............ ........ ..... . . 4 2.1. Stepwise Integral Scaling Method .................................................... ............... 4 2.2. Scaling Study of Corium Dispersion in DCH ........................................ ......... . 5 2.2.1. Initial Corium Jet Break-up.............................................. .................... 5 2.2.1.1. Single Phase Jet Breakup ..... ....................................... ......... 6 2.2.1.2. Two-phase Jet Breakup..... .......... ........ ....... ... .. .. . . .............. .. . . 6 2.2.1.3. Jet Break-up Droplet Diameter............................................... 7 2.2.1.4. Secondary Disintegration of Jet Droplet................................. 7 2.2.1.5. Corium Drop or Jet Impingement Phenomena...................... . 8
- 2.2.2. Corium Spread-out Over Cavity Wall ................................................... 8 2.2.3. Estimate of Cavity Conditions .............................................................. 9 2.2.4. Flow Regime in Cavity .................................... .... ............... ............ .... 10 2.2.5. Corium Entrainment and Droplet Size................................................. 11 2.2.6. Time Constant of Various Phenomena................................................ 12 2.3. Application of the Scaling Study ....................................................................... 12 2.3.1. Prototypic Case . ... . . . . . . . . .. .. . .. .. ... .. . . . . .. . .. . . .... ... ..... ... . . . . . .. .. . . . .... . . .. . . . . . .. . . . 13 2.3.2. Scaled Experiments . . . . ... .. . . . . . . . . .. . ... .. . .. . . . . . . .. . . .. . . .. .. . . . . . . . . .. .... . . . .. . . . . . . . . .. 14 3 . TES T FA CILITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . .
3.1. Low Pressure Test Facility ........ ...................... ..... .. ............. .......... . ................ 19 3.2. High Pressure Test Facility.............. ............................... ... . ......... .... ... .. . ... ...... 3 6
- 4. INSTRUMENTATION AND MEASUREMENT................................................. . 40 4.1. Droplet Size, Size Distribution, and Mass Flux Measurements........................ 40 4.2. Liquid Film Thickness Measurement ............................................ ................. 53 4.3 Gas and Liquid Velocities........................................... ........... . .. . ............. ....... 61 4.4. Pressure and Temperature Measurements.......................... ............................ 64 i 4.5. Flow Visualization .... ..... . . .. . . . . . .. ...... . .. . ... .... .. .. . .. . . . . .. .. . ... . . . . .. . ... . . .. . . . . . . ... . . .
! ~ 4.6. Test Control and Data Acquisition System .............................. ...................... 64 i i v NUREG/CR-6510 u_ __
- 5. EXPERIMENTS AT 1.4 MPa VESSEL PRESSURE. ....... ........ ....... .. ............. . 65 5.1. Standard Air-water Tests . . . . . . . . .. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .. . . . . . . . . . . . . .
5.1.1. Liquid Blowdown History and Flow Visualization................. ............ 65 5.1.2. Gas Velocity and Water Film Velocity........ ....................................... 65 5.1.3. Water Film Thickness in the Cavity ...... .. .................. . ..... .............. 67 5.1.4. Droplet Size Distribution.................... ... ......... .......... ................ ........ 6 8 5.1.5. Water Droplet Mass Flux .. . . .... .... .. . . .. . . . .. ... . ... ... . . ... ... . . ... . .... . .. . . . . ... . . . . 6 8 5.1.6. Liquid Carryover to the Containment / Exhaust Chamber........... ......... 69 5.2. Standard Air-woods Metal Tests............................... ............... .................... 92 l 5.2.1. Blowdown History . . . . . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.2. Gas and Liquid Velocities .. ... ..................... . ........... ........................ . 93 t 5.2.3. Film Thickness in the Cavity...... ............................ . ....... ................. .. . 93 5.2.4. Droplet Size, Size Distribution, and Mass Flux................................... 93 5.2.5. Liquid Carryover to the Upper Containment...................... .............. 94 5.3. Experimental Results of Parametric Studies ................. .............. .................100 5.3.1. Gas Discharge Nozzle Size Effects ........................ ........ ..................100 5.3.2. Liquid Inventory Effects............................... ........................ ........... 100 5.3.3. Subcompartment Trapping .......................................... ................... .. 101 1
- 6. HIGH PRES SURE EXPERIMENTS ...................................................... ... ........... 108 6.1. Water Test at 6.9 MPa Vessel Pressure..... ...................................................109 6.2. Water Test at 14.2 MPa Vessel Pressure....................... ....................... ........ I 10
.......111 6.3. Woods Metal Test at 14.2 MPa Vessel Pressure .................... ............
6.4. Woods Metal Test at 10.4 MPa Vessel Pressure .............. .............................I12 6.5 Water Tests with Different Nozzle Sizes............................. ............... ...... ..113
- 7.
SUMMARY
AND CONCLUSIONS...... .................................................... ......... 166 REFE REN CES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . NUREG/CR-6510 vi
LIST OF TABLES Table Page 2.1 Input parameters used in Table 2.2.... ............................ . ............ .... ......... . .... 17 2.2 Sample calculation for various parameters in corium dispersion...........................18 3.1 Summary ofinitial and geometry conditions of SNL, ANL, and Purdue DCH tests ................................................................. 20 41 Instruments and the parameters measured ........................................................... 42 5.1 Film flow parameters before gas blowdown........... .............................................. 85 5.2 Time-averaged film thickness after gas blowdown...... ........................................ 85 5.3 Parameters measured after gas blowdown with different gas nozzle sizes ..........102 5.4 - Liquid inventory effects on measured parameters .............................................. 102 6.1 Rupture disc selections and the initial pressure conditions in the gas tank and the test vessel ..................................................................... 108 6.2 Results of the nozzle size effects on the dispersion parameters .... ...................... I15 7.1 Comparison of the major results under different vessel pressures.......................167 7.2 Comparison between experimental data and preliminary estimation ...................168 l l l vii NUREG/CR-6510 o_ .I
LIST OF FIGURES Figure . Page i 1.1 Schematic diagram of the DCH accident for Zion reactor geometry...................... 2 3.1 Schematic of the PU low pressure test facility............... ..................................... 21 3.2 Test vessel and discharge line ......................................... ................................... 22 3.3 - Cavity geometry (1/10 scale) .... ........... .. ...... ................... . . ... . . . .. ... .... ..... .. . ... ... .. ... 23 3.4 Dimensions of the test cavity . .............. .......... ............... . . ....... . . .... ...... .. . ...... ....... 24 3.5 Dimensions of the test loop ...... ........... .................... .... . .. ..... ... . .. ...... .... ...... .. .... .. 27 3.6: Plan view of subcompartment (level 1) ............................................................... 28 3.7 Plan view of subcompartment (level 2) ........................................................ ...... 29 3.8 Plan view of subcompartment (level 3) ............................................................... 3 0 3.9 Semi-cross sectional view of the subcompartment............................................... 31 3.10 Cross sectional view of the containment / exhaust chamber...................... ............ 32 3.11 Plan view of the upper containment (level 4)....................................................... 33
~
3.12 Semi-cross sectional view of the upper containment....................... .................... 34 I 3.12a Top view of the upper containment..................................................................... 3 5 1 3.13 ; Schematic of the PU high pressure test facility.................................................... 37 3.14 Liquid and gas discharge system and supporting frame........................................ 38 3.15 Test cavity for high pressure experiments............................................................ 39 4.1 Instrumentation in the test facility ............ .... ... ............... . . .. .. ... ... .. .... .... .... ..... . .. . .. 41 1 4.2 Droplet sampling and sizing system .................................................................... 43 4.3 Droplet sampling probe .. . .. . .. .. . . .. . .. . . . . .. . . . . .. . . . . . .. . .. . .. . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.4 . Water droplet collection mechanism ................................................................... 46 4.5 Woods metal droplet collection mechanism................................................ ........ 47 4.6a Calibration results on droplet size distribution measurement at j, = 35 m/s and j r = 0.04 m/s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 8 4.6b Calibration results on droplet size distribution measurement at j, = 60 m/s and j r = 0. 04 m/s . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 9 4.6c Calibration results on droplet size distribution measurement atj, = % m/s and j r = 0.04 m/s . . . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 5 0 4.7 Digitized droplet image from a water droplet sample .......................................... 51 4.8 Needle-type conductivity pmbe for water film thickness measurement................ 55 4.9 Signal conditioning circuit of the conductivity probe........................................... 56
~ 4.10 Calibration system for film thickness probe ......................................................... 57 4.11 . Calibration curve for water film thickness probe ................................................. 58 4.12 Woods metal film thickness probe....................................................................... 5 9 4.13 Calibration of the liquid metal film thickness probe ................................ ............ 60 5.1 Water and air blowdown transients (1.4 MPa vessel pressure) ................... ........ 69 5.2 Cavity pressure transient (air-water test,1.4 MPa vessel pressure).......... ........... 70 5.3 Flow visualization of the waterjet (1.4 MPa vessel pressure).............................. 71 5.4 Flow visualization of the water film spreadmg in the cavity
( 1.4 MPa vessel pressure) .... .. . . ... . . . . ...... .. . . . ... . . . ... . ... . . ... ..... . . . . . . . . . . . . . . . .. . . . . . .. . . .. . . . . . . 72 NUREG/CR-6510 viii
Figure Page 5.5 Flow visualization of water film flow in cavity chute (1.4 MPa vessel pressure)... 73 5.6 Flow visualization of the water carryover (1.4 MPa vessel pressure) ................... 74 5.7 Gas velocity in the cavity (air-water test,1.4 MPa vessel pressure)........... .......... 75 5.8 Gas velocity profile in the cavity pedestal section (1.4 MPa vessel pressure)...... . 76 j 5.9 Gas velocity profile in the cavity chute section (1.4 MPa vessel pressure)............ 77 5.10 Water film velocity measured by hot film probes (1.4 MPa vessel pressure)......... 78 5.11 Water film thickness transient at CH1 (1.4 MPa vessel pressure)......................... 79 5.12 Water film thickness transient at CH2 (1.4 MPa vessel pressure)......................... 80 5.13 Water film thickness transient at CH3 (1.4 MPa vessel pressure)... ..................... 81 5.14 Water film thickness transient at CH4 (1.4 MPa vessel pressure)..................... .. 82 ) 5.15 Water film thickness transient at CH5 (1.4 MPa vessel pressure)......................... 83 5.16 Water film thickness transient at CH6 (1.4 MPa vessel pessure)........... ............. 84 5.17 Water droplet size distribution at cavity exit in time window #1.................. ..... . 86 5.18 Water droplet size distribution at cavity exit in time window #2.................... ..... 87 5.19 Water droplet size distribution at cavity exit in time window #3 .......... ..... ... .... 88 5.20 Water droplet size distribution near cavity floor in time window #2............. ....... 89 5.21 Water mass collection in different time window............................................... .. 90 5.22 Water droplet mass flux distribution at cavity exit .............................. ..... . ........ 91 5.23 Air-woods metal blow-down history (1.4 MPa vessel pressure)....... ................... 94 5.24 Cavity pressure transient (1.4 MPa vessel pressure, WM test) ............................. 95 5.25 Average cavity gas velocity transient (1.4 MPa vessel pressure, WM test)........... 96 5.26 Woods metal film thickness signals at CH1 and CH6 (1.4 MPa vessel pressure) .. 97 5.27 Woods metal droplet size distribution at cavity exit (1.4 MPa vessel pressure)..... 98 5.28 Woods metal droplet size distribution in containment (1.4 MPa vessel pressure).. 99 5.29 Effects of gas nozzle size on entrainment in cavity (1.4 MPa vessel pressure)....103 5.30 Effects of gas nozzle size on carryover (1.4 MPa vessel pressure) .....................104 5.31 Effects of water inventory on entrainment in cavity (1.4 MPa vessel pressure)...105 5.32 Effects of water inventory on carryover (1.4 MPa vessel pressure) ....................106 5.33 Schematic of the deflector in subcompartment trapping tests.............................107 5.34 Effects of deflector angle on water carryover (1.4 MPa vessel pressure)............107 6.1 Pressure transient in the test vessel (6.9 MPa water test).................................... I16 6.2 Water and gas mass flow rate transients (6.9 MPa water test) ............................ I17 6.3 Cavity pressure transient (6.9 MPa water test) ................................................... I 18 6.4 Gas velocity in the test cavity (6.9 MPa water test) ............................................ I 19 6.5 Water film thickness transients (6.9 MPa water test) .........................................120 6.6 Size distribution of the droplets at cavity exit (6.9 MPa water test) ...................121 6.7 Pressure transient in the test vessel (14.2 MPa water test)............ ....................122 6.8 Water and gas mass flow rate transients (14.2 MPa water test)..........................123 6.9 Cavity pressure transient (14.2 MPa water test) ................................................ 124 6.10 Gas velocity in the test cavity (14.2 MPa water test) .........................................125 6.11 Typical water film thickness transients (14.2 MPa water test)............................126 6.12 Size distribution of the dropletsat cavity exit (14.2 MPa water test)...................127 6.13 Pressure transient in the gas tank (14.2 MPa woods metal test)...................... ..128 ix NUREG/CR-6510
m Figure Page 6.14 Liquid and gas mass flow rate transients (14.2 MPa woods metal test)...............129 6.15 Cavity pressure transient (14.2 MPa woods metal test)......................................130 6.16 Gas velocity in the test cavity (14.2 MPa woods metal test)............................ ..131 6.17 Liquid film thickness transients (14.2 MPa woods metal test)........................... .132 6.18 Size distribution of droplets at cavity exit (14.2 MPa woods metal test).......... ..133 6.19 Gas tank pressure transient in woods metal test (3.5 cm nozzle, Po=10.4 MPa)..134 6.20 . Liquid and gas discharge rate (3.5 cm nozzle, Po=10.4 MPa) .............................135 6.21 Cavity total pressure transient (3.5 cm nozzle, Po=10.4 MPa) .................... .......136
~ 6.22 Gas velocity in the cavity (3.5 cm nozzle, Po=10.4 MPa)................... ................137 6.23 Droplet size distribution at the cavity exit (3.5 cm nozzle, Po=10.4 MPa)...........138 6.24 Tank pressure transient in 6.9 MPa water test with 4.0 cm nozzle......................139 6.25 Tank pressure transient in 10.7 MPa water test with 4.0 cm nozzle....................140 6.26 Tank pressure transient in 14.2 MPa water test with 4.0 cm nozzle....................141 6.27 Tank pressure transient in 6.9 MPa water test with 5.0 cm nozzle......................142 6.28 Tank pressure transient in 14.2 MPa water test with 5.0 cm nozzle....................143 6.29 Water and gas discharge rate in 6.9 MPa test with 4.0 cm nozzle.......................144 6.30 Water and gas discharge rate in 107 MPa test with 4.0 cm nozzle......................145 6.31 Water and gas discharge rate in 14.2 MPa test with 4.0 cm nozzle.... .... ...........146 6.32 Water and gas discharge rate in 6.9 MPa test with 5.0 cm nozzle.............. ........147 6.33 Water and gas discharge rate in 14.2 MPa test with 5.0 cm nozzle.............. ... ..148 6.34 Cavity pressure transient in 6.9 MPa test water test with 4.0 cm nozzle .............149 6.35 Cavity pressure transient in 10.7 MPa test water test with 4.0 cm nozzle ...........150 6.36 Cavity pressure transient in 14.2 MPa test water test with 4.0 cm nozzle ...........151 l 6.37 Cavity pressure transient in 6.9 MPa test water test with 5.0 cm nozzle .............152 6.38 Cavity pressure transient in 14.2 MPa test water test with 5.0 cm nozzle ...........153 6.39 Cavity gas velocity transient in 6.9 MPa test water test with 4.0 cm nozzle........154 6.40 Cavity gas velocity transient in 10.7 MPa test water test with 4.0 cm nozzle ......155 6.41 Cavity gas velocity transient in 14.2 MPa test water test with 4.0 cm nozzle ......156 6.42 Cavity gas velocity transient in 6.9 MPa test water test with 5.0 cm nozzle........157 6.43 Cavity gas velocity transient in 14.2 MPa test water test with 5.0 cm nozzle ......158 6.44 Water droplet size distribution in cavity (6.9 MPa test,4.0 cm nozzle)...............159 6.45 Water droplet size distribution in cavity (10.7 MPa test,4.0 cm nozzle).............160 6.46 . Water droplet size distribution in cavity (14.2 MPa test,4.0 cm nozzle).............161 6.47 Water droplet size distribution in cavity (6.9 MPa test,5.0 cm nozzle).............. 162
)
6.48 Water droplet size distribution in cavity (14.2 MPa test,5.0 cm nozzle).............163 6.49 Effects of vessel pressure and nozzle size on Dym in cavity for water tests ......... 164 6.50 Effects of vessel pressure and nozzle size on water carryover.............................165 NUREG/CR-6510 x j
m NOMENCLATURE English Greek A Area a void fraction Cr friction factor S fihn thickness Cd dynamic wave velocity e entrainment rate d diameter (droplet) dynamic viscosity D diameter p density Eu Euler number o surface tension, lumped parameter f friction factor t time constant F fraction Subscripts g; - gravitational acceleration 0 initial condition, zero order value l h fdm thickness, height b blowdown
- j. superfacial velocity B jet breaks M mass e cavity, corium N viscosity number er critical value p pressure d dispersed phase P pressure e entrainment, cavity exit R universal gas constant
- f. liquid phase Re Reynolds number g gas phase t time h hydraulic or break (diameter)
T temperature i- initial, imp'mge u velocity ! im impact v velocity J jet V volume, velocity : r relative value w width ; si sphericallimit value (droplet) ; We Weber number t choking condition : x space coordinate y vessel condition i g NUREG/CR-6510
EXECUTIVE
SUMMARY
In Direct Contamment Heating (DCH) accidents, one of the fundamental factors of containment heating and pressurization is the degree of the molten corium dispersion. The degree of the heat transfer and chemical reactions that may lead to containment over-pressurization are proportional to the available surface area of the molten corium. If the corium is highly dispersed and the resultant aerosol particle size is very small, the risk to containment failure can be very high. Thus, it is important to investigate the corium dispersion process concerning the mean corium droplet size as well as the dispersal fraction. To understand the corium dispersion phenomenon, a separate effect test program on the corium dispersion mechanisms in the reactor cavity and the subcompartment trapping mechanisms was initiated in 1992 at Purdue University under the direction of the US Nuclear Regulatory Commission. The four major objectives of this study are: (1) to perform a detailed scaling study using the newly proposed step-by-step integral scaling method, then to evaluate existing models and correlations for droplet entrainment, particle size and size distribution, and the particle trapping, (2) to design and construct a 1/10 scale Zion reactor model, and to perform carefully scaled experiments using air-water and air-woods metal to simulate the prototypic steam and core melt, (3) to develop reliable mechanistic models for the corium dispersion and transport in the accident scenario, which can be used to predict the liquid and gas blowdown, entramment, droplet size, liquid canyover to the containment. and the subcompartment trapping, (4) to use the I models to perform stand alone calculations for the prototypic conditions. In this report, efforts are focused on the first two objectives, whereas the modeling study is covered in a separate report. For the scaling study, the major mechanisms of the corium dispersion and transport are identified. These are the liquid jet breakup, liquid film flow before gas blow down, droplet entramment, and the subcompartment trapping. Non-dimensional groups and characteristic time constants are obtained to scale the accident transient. Based on the study, the 1/10 scaled test facility is designed and evaluated to preserve the major phenomena occurred in the prototypic accident. Various instruments are used to capture the blowdown transient in a time interval of 10 seconds. Parameters such as the vessel pressure, liquid and gas mass flow rate, liquid film thickness and velocity in the cavity, entrainment rate and droplet size distribution in the cavity, and the total liquid carryover in the contamment are obtained in great details. The tests are carried out under different initial conditions. According to the pressure conditions, these tests can be grouped into two categories, i.e., the 1.4 MPa low pressure tests and the high pressure tests with a vessel pressure ranging from 6.9 MPa to 14.2 MPa. In the 1.4 MPa low pressure tests,3.5 cm liquid discharge nozzle is used. The gas supply comes from a separate 10.16 cm pipe line that provides sufficient gas velocity in the cavity to preserve the prototypic entrainment phenomenon at a vessel pressure of 6.9 MPa. Three kinds of tests are included in the low pressure category. These are the standard air-water tests, the standard air-woods metal tests, and the parametric tests. In xiii NUREG/CR-6510
r i the standard tests, 7 liters of water or woods metal is used to simulate the core melt l (about 40% of the total core mass). With the measurements and flow visualization, l detailed information regarding the dispersion phenomena are obtained. These include (1)
- liquid discharge and liquid jet spread-out upon the impingement on the cavity floor, (2) liquid film motion and transport, (3) liquid entrainment, entrained droplet size distribution and mass flux, (4) gas velocity and velocity profiles, (5) liquid trapping in the subcompartment, and (6) liquid carry over into the containment / exhaust chamber. In the case of water tests, it is found that about 15% of the liquid flows out of the cavity before gas blowdown without dispersion. In the cavity,43% of the discharged water is entrained into droplets with a volume median diameter of 406 m. After the tests,2.7% of water is ;
recovered in the upper containment and the rest deposits in the subcompartment. No i liquid is observed in the test cavity. For the woods metal tests, due to the low liquid jet j velocity (11m/s), the liquid film flow cannot reach the cavity exit before gas blowdown. l The discharged woods metal is thus entirely subjected to the entrainment process. About l 11% of liquid metal is entrained into droplets in the test cavity with a volume median diameter of 1.1 mm. After the tests,1.6% of woods metal is recovered in the upper containment and 8% remains in the cavity in the form of thin crusts on the cavity walls. In the parametric studies, the effects of gas discharge nozzle size (or cavity gas velocity), water inventory, and subcompartment trapping conditions on the liquid dispersion process are obtained. The increase in gas velocity significantly intensifies the liquid entrainment in the test cavity, but has negligible influence on the liquid film motion in the cavity. Larger j liquid inventory tends to enhance the entrainment rate in the test cavity due to the increase i in the Reynolds number of the liquid film flow. However, the liquid canyover to the containment remains constant regardless of the changes of discharged liquid volume for the test range. The subcompartment trapping experiments demonstrate that the trajectory dependent phenomenon is predominant in determining the effectiveness of liquid entrapment. In the high pressure DCH tests, the water tests are conducted at the vessel pressure of 6.9 MPa and 14.2 MPa, and the woc <is metal tests are performed at 10.7 MPa and 14.2 ) MPa vessel pressures. For woods metal tests,38 % of the discharged liquid is entrained into droplets in the cavity with a volume median diameter of 857 m. About 3.4 % of the l 56 kg woods metal is dispersed into the upper containment chamber. Compared to the l previous 1.4 MPa experiments, the entrainment under high vessel pressure is more intensive and the resultant droplet size is smaller. In the case of the 14.2 MPa water tests, it is found that about 5.3 % of the total water is dispersed into the upper containment. In the cavity,26 % of the water is entrained into droplets with a volume median diameter of 151 m. As the vessel pressure drops to 6.9 MPa, about 36 % of the discharged water is dispersed into droplets with a volume median diameter of 184 m, and the dispersed liquid fraction in the containment is only 2.4 %. A comparison of the data under different vessel pressures indicates that the liquid transport process is much faster as the vessel pressure increases. Consequently, the entrainment time becomes shorter, which tends to reduce the fraction ofliquid dispersion in the test cavity. However, as a result of vessel pressure rise, the gas velocity increases, which in turn enhances the entrainment process. These two NUREG/CR-6510 xiv
competing factors determine the degree of the entire liquid dispersion transient. In general, as the vessel pressure increases in the high pressure tests, the results show an increase of the dispersed liquid carryover to the upper containment. In the high pressure tests with large nozzle sizes, in spite of the diversity of the entrainment fraction in the test cavity, more carryover is obtained as the gas discharge rate increases. This phenomenon implies that the droplet re-entrainment and transport in the subcompartment are the key factors in determining total carryover to the upper cortainment More theoretical efforts are therefore required to analyze droplet transport from the subcompartment to the upper contamment building. The experi vntal results also indicate that corium dispersion mechanisms identified in the scaling stou; are adequate. According to a preliminary estimation based on the scaling study, the entrainment rate correlation and the droplet size distribution model proposed in the scaling study are accurate for both air-water and air-woods metal tests. This conclusion is encouraging for further detailed theoretical modeling based on the identified mechanisms and correlations proposed in the scaling study.
)
xy NUREG/CR-6510 l l
ACKNOWLEDGMENTS The authors would like to express their gratitude to Mr. D. Zheng and Mr. K. Sato for their participation in the high pressure tests, to Mr. P. O'Brien for Ids help in assembling the subcompartment, and to Mr. B.E. Hoeffer who did part of the tedious cleaning work j after woods metal tests. Sincere appreciation is also due to Dr. F. Eltawila for his valuable discussions and support to this project. The authors extend their formal thanks to the personnel at the Nuclear Engineering Machine Shop of Purdue University, particularly to Mr. Don Bower and Mr. Bob Sanders. This work was funded by the US Nuclear Regulatory Commission under the contract ; number ofNRC FIN L1990. ' 1 I l I NUREG/CR-6510 xvi
- 1. INTRODUCTION Due to the occurrence of the small break loss-of-coolant-accident at TMI-2 that led to partial core damage, interests in low probability severe accidents have increased significantly. The Direct Containment Heating (DCH) accident is this kind of accident scenario identified in Zion and Indian Point Probabilistic Safety Studies [1,2] in the early eighties. Since then, extensive studies have been condt":ted both experimentally and analytically. As one of the main objectives of this report, the hydrodynamic characteristics of the corium dispersion phenomena in the reactor cavity during DCH au:idents is studied experimentally. A l/10 scale Zion reactor model is designed and constructed to carry out the accidental blowdown transient using water and liquid woods metal to simulate the prototypic core melt. The detailed experimental results of the corium dispersion and transport constitute a solid data base for the theoretical modeling, which is covered in a separate report.
The DCH in severe nuclear reactor accident is a postulated, beyond design, low probability scenario, (Fig.1.1), which may lead to comainment over-pressurization after PWR core melt-down and pressure vessel failure. A typical accident scenario, postulated to demonstrate the whole DCH transient, begins with a loss of the off-site power and the resultant turbine trip and core shut-down. With a followed diesel generator failure, the station black-out occurs. Due to lack of electrical power, the main feed-water pumps stop running, and the Emergency Core Cooling System (ECCS) can not be activated. Thereby, the accumulated decay heat of the reactor core over-pressurizes the primary coolant system, resulting in the actuation of the relief valves on the pressurizer and thus a continuous release of the reactor coolant. Once the core is uncovered by the coolant, the over-heating and melting of the core are unavoidable. Due to gravity effects, the molten corium may migrate to the lower head of the reactor pressure vessel, inducing a failure at the instrumentation penetration of the vessel bottom. Driven by the high vessel pressure, the core melt is ejected into the reactor cavity, followed by a rapid blowdown of the primary steam inventory. In the cavity, the high speed steam stream entrains part of the discharged corium into the containment building in the form of fine aerosol particles, which can greatly enhance the chemical reactions and the heat transfer between the hot corium and the blowdown steam. Consequently, the containment atmosphere will be heated and pressurized by the intense heat transfer, the chemical reaction induced heating, and the possible hydrogen combustion. If the pressure exceeds the design limit, the containment may be damaged, resulting in release of fission products to the environment. The above described accident scenario is the so-called Direct Containment Heating (DCH), which was first identified in the Zion [1] and Indian Point [2] Probabilistic Safety Studies. Previous studies [3-14] indicate that the most significant factor affecting the direct containment heating is the degree of the molten corium dispersion. 'lhis is because the heat transfer and chemical reactions which may lead to the containment over-pressurization are basically proportional to the available surface area of the molten 1 NUREG/CR-6510 ,
. o.4 G.',,. p .t w 5 4 D ~.- ? en "L,
g * * .f 7" * , e e g
- j. .
,.e t containment 9 e .-
u-
. v H -
[ .~
- Corlun! particles ,e "t
i 4:
. polar crane e s.,
n .L
.. r. -
7;,l v,.
;. , ~. ;-
p, -
. :.%..:,c -:
~
- . . . s, 17 ,:
n $ IJ "i 'yQ
'Ey Lpc ' '.';
;j @, i: , ,
po: :s
- ax- p_ . %w s l
I e G
^
l: '
? /~h n k [
I f Q - s L. ry c. 4
- za .,.
- ..e. .. ~.
i+ >.' A V S i .C
+
'~ , ..s 9 :- 3?- .
- +s >-
zs;$r., T'+ff'.",,l . G2 lgf 'C'Dx+&%?.4)2M) W y+g.%z g M:S
'^
yggs MZ
^
AM. cavity
~ ~'
W: Sent rahle # ',~ . . $$ preunrl7pr Apft: 457sF37&
.ar.2ess car.2e>2r34 1
1 1 1 Figure 1.1, Schematic diagram of DCH accident for the Zion reactor geometry l NUREG/CR-6510 2
corium. If the corium is highly dispersed and the droplet size is very small, the resultant pressurization of the containment could be large. In spite of the importance of the corium dispersal mechanisms and droplet size distribution for DCH analysis, existing correlations cannot be applied reliably to assess the degree of containment pressurization. Two main reasons for this deficiency are: (1) mechanisms of corium dispersion are not well I understood or mod &d, and (2) no scale-up capability exists for available droplet
- entrainment correlatie .
One of the impcrtant process during the corium dispersion transient is the subcompartment trapping. The structures such as the seal table room, which stands in the direct path of the debris flow have major effects on the ability of the debris to remain airborne in and out of the subcompartment. The significant influence of the structures on the debris dispersal has been reported [10,11]. The available correlations on droplet dispersion process are mainly based on the droplet entrainment process in annular flow. The mechanism ofliquid film disintegration has not been considered in the process. The effect of the film ejection and disintegration near the exit of the cavity may have a significant impact on the total subcompartment trapping. All these factors should be understood for an accurate assessment of DCH. Wihut the flow obstacles, the l dispersion is mainly controlled by the diffusion process. However, in the presence of the flow obstacles / deflectors, the subcompartment trapping may be quite different depending on the characteristic of the obstacle. To study both experimentally and analytically the corium dispersion mechanisms in the i reactor cavity and subcompartment trapping, the four major scopes of the Purdue University DCH project are identified u: (1) To perform a detailed scaling study with the newly proposed stepwise integral scaling method [15] to identify the mechanisms that govern the transient of corium dispersion l and to evaluate the existing physical models for each individual mechanism [16]. l (2) To design and conduct simulation experiments using water-air and woods metal-air in a 1/10 linear scale Zion reactor model, collecting dispersion information such as the film flow transient in the cavity, the entrained droplet size and size distribution, and the dispersion fraction of the total discharged liquid. (3) To develop or select reliable mechanistic models and correlations for the corium dispersion phenomenon. (4) With the models to carry out stand alone calculations under prototypic conditions. In this report, the scaling efforts and the simulation experiments are presented. The modeling part of the project will be included in a separate report. 1 l l 3 NUREG/CR-6510
I
- 2. SCALING OF DCH PHENOMENA l
l Specific scaling criteria focused on key phenomena and important mechanisms are developed by the step-by-step integral scaling method described below. This method was proposed by Ishii [16], which will be applied here to the corium dispersion problem in the reactor cavity during the DCH accident. 2.1. Steowise Integral Scalina Method The first step is to identify the system and initial and boundary conditions for the problem. This is followed by the selection of the major subsystems and identification of the interfacial conditions between them, then the scaling study is carried out for each subsystem separately. He second step is to identify the key transfer processes and potentially important mechanisms, which should be ranked in terms of the relative i importance and order of events. The possible bifurcation phenomena and feedback mechanisms should also be identified. The bifurcation phenomena highlights the changes in transfer mechanisms whereas the feedback mechanisms focus on the coupling effects between different transfer processes. Based on the above study, the sequence of the scaling study in terms of the transfer processes and mechanisms is determined. The actual scaling analysis is carried out by starting from the most dominant process and considering the various mechanisms which can cause that transfer process. The step-by-step approach ! is used here by considering only one mechanism at a time. The analysis should start from the subsystem boundary and upstream event. First, the most possible mechanism is chosen. The integral rate process equations, integral balance equations, and boundary conditions are identified for a particular mechanism together with the transition criteria for the mechanism. He first set of scaling criteria are obtained by non-dimensionalizing the integral response from the balance equations or the rate equations. The next step is carried out by considering the second mechanism. The transition criterion between the first and second mechanisms should be evaluated if the bifurcation is possible. Also the coexistence of the first and second mechanisms should be considered, for which case the relative importance between them is evaluated using the scaling parameters. It is also important to obtain the characteristic time constant for each mechanism. It will give an estimate of the time required to complete the transfer procem within the subsystem by that particular mechanism. I By continuing the above step-by-step scaling analysis, a whole set of scaling criteria, transition criteria and characteristic time constants are obtained for the most dominant transfer process. If other processes are considered to be important, then similar steps are repeated. Following the above approach, it is possible to evaluate the relative importance of various effects and mechaniama between the prototype and the scaled-down or simulant
-system. For this purpose the scaling criteria and time constants are used. In order to l
NUREG/CR-6510 4
obtain quantitative results, it is necessary to give initial and boundary conditions for bcith systems. Using these, the actual values of the scaling parameters and time constants are calculated. Then the similarity as wel! as the scale distortions between the systems are analyzed. It is essential that the model system should reproduce dominant phenomena and key mechanisms that are similar to those expected in the prototype. From this requirement, the desirable experimental conditions for scaled-down model or simulant model can be developed. When the uncertamty for the dominant mechanism in the key transfer process is large, the separate effect test focusing on that particular phenomenon is indicated. -It is also noted that the experimental data from the model system are indispensable to test the underlining assumptions and physical models used to develop the scaling criteria. 2.2. Scaline Study of Corium Discersion in DCH The stepwise integral scaling method explained above is applied to the corium dispersion in the reactor cavity in DCH. The subsystem is considered to be the reactor cavhy with the reactor vessel break as the upstream boundary. The previous studies for the DCH problem indicated that the most significant factor affecting the contamment heating and pressurization was the degree of the molten corium dispersion. This is because the heat transfer and chemical reactions which may lead to the containment over-pressurization are basically proportional to the available surface area of the molten corium. Mr the scaling study, therefore, the corium dispersion is taken as the most important phenomenon to be studied in detail in the reactor cavity. For the molten corium dispersion and corium transport in the reactor cavity, the following four mechanisms are critical: (1) Corium discharge and coriumjet disintegration (2) Liquid corium spread-out upon impact of the jet or droplet (3) Liquid mass transport due to inertia, pressure and shear force (4) Entrainment and drop formation by streaming gas These mechanisms are studied by using the step-by-step integral scaling method, starting from the upstream event. 2.2.1. Initial Corium Jet Breakup The molten corium jet can disintegrate into droplets after the discharge from the reactor vessel break. The corium discharge can be in a form of a single phase jet or two-phase jet due to the depressurization or punch-through of the gas flow over the liquid corium surface in the reactor vessel. Several possible mechanisms of jet disintegration length are discussed below. The comparison of the jet disintegration length with the height below the reactor vessel gives the base to determine whether the molten corium disintegrates before impinging on the floor or not. Obviously, if the breakup length, Ls, is much smaller than the height, then it is expected that the jet disintegrates into small droplets during the vertical downward motion below the break. Hence it is important to know the breakup length. 5 NUREG/CR-6510 I
2.2.1.1. Single Phase Jet Breakup The single phase liquid jet can breakup in two major modes depending on the relative velocity between the liquid and gas phases, (Obot and Ishii, [17]). The first mode is due to the hydrodynamic instability of thejet surface. For a relatively small gas Weber number within the range of We,, a 8# ' ' < 33 , (2-1) a thejet breakup length Lei is given by 18' = 595 ## # (2-2) D,, y a where v,, va, and D; are the relative velocity, jet velocity, and jet initial diameter. The second mode ofjet surface breakup is due to relative motion. When the relative velocity is high, We, > 3.5, the breakup length can be significantly reduced [17] and is given by
' b = 1110 #'## (We,,)-"* . (2-3)
D, y a These correlations indicate that the capillary number, prvo/a, and gas Weber number scale the jet breakup length. The time constant is Lai/o v. I i 2.2.1.2. Two-ohase Jet Breakup ! When the jet consists of two phases due to either the gas blow through or gas generation the effect of the void fraction should be considered, Denten and Ishii [18]. In this case b = 595 #' ' "> 1 (2-4) ia< D, 0.854, for i We,,a #') # < 3.5 (2-5) a Where j; = a;vy + (1-as)v o. as and vp are the void fraction and gas velocity at the jet discharge point, respectively. For higher relative velocity, We, > 3.5, NUREG/CR-6510 6
C i # "' s2 = 1110 1 (We,,)~"* (2-6) D, a < 0.854 These correlations indicate that the existence of void accelerates the jet disintegration significantly. 2.2.1.3. Jet Breakuo Droolet Diameter j l The second phenomenon of importance for the corium jet disintegration is the j resultant droplet size. For the DCH problem the droplet size is the key in determining the J degree of chemical reaction, heat transfer and corium transport. In the following, the
. droplet size from the disintegrating . jet is discussed in terms of the primary jet disintegration and subsequent droplet disintegration. The correlations given below are applicable both for the single and two-phase jet disintegration. By modifying the annular droplet size correlation of Kataoka, Ishii and Mishima [19] the mean droplet -size in the i disintegrated jet flow (De Jarlais, Ishii, and Linehan, [20]) is given by r S V3 d = 0.028
# #L P,v 2 (Re,)"'<Pt , (2-7) r and the maxunum size is given from the maximum log normal distribution as do. = 3.13d.
This criterion roughly corresponds to a Weber number of 12. 2.2.1.4. Secondary Disintegration of Jet Droolet When the jet disintegrates, the initial droplet size is given by the above correlation. > However, these initial droplets can often be relatively large and may not be stable. The droplets from the jet can further disintegrate under two conditions. These are: (1) Existence of extremely high gas turbulence, I (2) Exceeding the sphericallimit. During the molten corium discharge phase, the first mechanism is unlikely. Under extreme conditions such as the sudden exposure of droplets to shock waves, the disintegration Weber number can be as low as 2 or 3. However, the condition around the disintegrated jet before the gas blow down phase does not meet such extreme gas turbulence requirement. The second mechanism indicates that if the initial droplet size given by the above correlation far exceeds the spherical stable drop size limit, then droplets will further disintegrate to reach this stable limit. The spherical limit is given by d,, = 4 (N,)"' (2-8)
<8bP 7 NUREG/CR-6510
e where the gas viscosity number, Ng, is defined by N, = #' (2-9) (p,adatgap)n 2.2.1.5. Corium Droo or Jet Imoinnement Phenomena in the above, the criteria for the jet breakup length and resultant droplet size are discussed. The next phenomenon ofinterest is the vertical impingement of the intact jet or the disintegrated jet in the form of droplet flow. The main question here is whether the impinging liquid mass spreads out coherently over the cavity well or bounces back into the cavity space. The phenomena after the drop orjet impingement at the cavity floor can be scaled by the impact Weber number defined by We ,a #' (2-10) a where d is the drop or jet diameter. For We m, > 80, the drop or jet will spread out as a liquid film due to large inertia overcoming the surface tension effect to recover. When We < 30, the droplets bounce back after the impingement, Bolle and Moureau [21]. l 2.2.2. Corium Spread out over Cavity Wall Under a prototypic condition the above criterion for the corium spread out is most ; likely satisfied. Then it is important to know the thickness and velocity of the molten ' corium film. For this purpose, several different length scales for the film thickness are considered below. From the continuity relation, and assuming that the magnitude of the velocity remains constant during impingement, the initial film thickness 6, at the vicinity of , the impingement point is given by 1 = 025. (2-11) D, , The order of the lower limit of the corium film thickness is estimated by the symmetric spread of the liquid film with the constant liquid film velocity. The length scale for the extent of the film spread is taken as the hydraulic diameter of the cavity, Dn. Then ) i D>' I S,,;, = . (2-12) ; 4D, ; l It is noted that instead of D3, other length sca.es may be used here, however for the present study Da is satisfactory; , NUREG/CR-6510 8
The maximum thickness is obtained by assuming the entire corium mass accumulated on the floor. Thus
- 6. = PM" A, , (2-13)
I where R and An are the total corium discharge mass and cavity floor area. Another reference film thickness can be obtained by assuming that the all molten corium spreads over the entire cavity surface. In this case 6, = M" = M" (2-14) P4 t trpfD,L, , where Aw and L are the cavity wall area and cavity length, respectively. 2.2.3. Estimations of Cavity Conditions In order to evaluate the molten corium dispersion in the reactor cavity, it is necessary to specify the global cavity conditions in terms of the liquid and gas flows. These parameters can be estimated from the geometry and boundary conditions. In the above, several length scales for the corium film thickness have been obtained. Typical values for these are calculated in Table 2.2 by considering the following reference conditions. In the following analysis, the typical film thickness of Sr = S i is assumed over the entire cavity wetted perimeter. Then from the continuity condition, the film velocity is given by Djv# v' = . (2-15) 4 D,6, The film Reynolds number is defined by Ref = #' ' * = # '. (2-16) Pr Pt If the film is only on the floor of the cavity, the velocity v rand Ret should be about x times the values given above. Because of the relatively high initial fihn velocity (in the order of 40 m/s), the liquid corium can climb up the side walls and may even cover the ceiling of the cavity. The actual values should be somewhere between them. The gas flow conditions are estimated along the analysis proposed by Henry [22]. By assuming the choked flow for the gas discharge following the corium discharge, the pressure at the throat is given by P = 0.6P, , (2-17) and the choked flow velocity by 9 NUREG/CR-6510
{ ' I J RT v, = . (2-18) With the ideal gas at the isothermal condition p, = 0.6p, . (2-19) 4 i The unlimited isothermal expansion can give the maximum velocity of twice the discharge velocity. The cavity gas initir.1 velocity before entrainment can be given from the continuity relation and the cavity pressure. Thus 0.6 P, ' A , ' R T v,. = - (2-20) P,, gA, aMo where A, and A. are the vessel break area and cavity flow area, respectively. Peo is the ) cavity initial pressure. This equation indicates that the cavity gas velocity is a strong i function of the cavity pressure. l 2.2.4. Flow Regime in the Cavity When the molten corium spreads out in the reactor cavity, three different two-phase flow regimes are possible. These are stratified wavy flow, slug flow and annular flow. In : order to identify the most possible flow regime, the transition criteria between these regimes are examined. The onset of slugging from horizontal stratified wavy flow is given by the following form (Wallis and Dobson, [23], Iishima and Ishii, [24], Taitel and Dukler
,[25]): '
v, 2 0.5]Apg h, / p, (2 21) In the Taitel and Dukler correlation, the constant of 0.5 is replaced by a function of the void fraction which approaches unity as the relative film thickness becomes smaller. The transition to annular flow either from the stratified wavy or slug flow is ! determined by the entrainment process. Thus if the gas (or relative) velocity exceeds the j onset of entrainment velocity, the entrainment of wave crests or' liquid slugs occurs. In case of slug flow, this leads to the elimination ofliquid slugs. Since part of the entrained , droplets are deposited on the wall surface, the onset of entrainment leads to the formation j of annular flow with liquid wetting the whole surface. However, due to the gravity effect ; it is expected that the liquid film thickness at the cavity floor is much larger than those at j the side wall and ceiling. The criterion for the annular flow transition is discussed below l in terms of the entrainment process. l NUREG/CR-6510 10
2.2.5. Corium Entrainment and Droplet Size The most likely flow regime in the cavity is the annular film flow. In this case the droplet entminment becomes the most important mechanisms to disperse the molten corium mass. The onset of entrainment, entrainment rate and entrained droplet size are discussed below. The onset of entrainment criteria is given (Ishii and Grolmes, [23]) by
#' ' b 2 N',' (2-22) ,
a Pt 1 where the viscosity number is defined as N, = #' (2-23) {pg a.jalgbp),,2 1 The entrainment rate from the film is given (Katoaka and Ishii [26]) by r 3 026
* = 6.6 x 10-7@9 W e) " b ,
@M Pt
t where Ref= , We = #' ' * (2-25) Pt a g Ps , Here Da is the hydraulic diameter of the cavity. The initial droplet size is i r 3 -1/3 r 32/3 d = 0.028 # 2 P, #' 2 Re7' Re /3 (2-26) Pave gPt>
t However subsequent disintegration may occur if We, a #8 ' '# > 12 (2-27) a It is noted that drop size can be as small as Wea = 1.7 ~ 2.5 if very high gas turbulence exists. Under the reactor cavity condition, the criterion of the Weber number at 12 is more likely than the later criterion. Furthennore, for the droplet to disintegrate beyond the initial entrainment drop size, a sufficient interaction time should exist between drops and the gas flow. 11 NUREG/CR-6510 l l 2.2.6. Time Constant of Various Phenomena Several time constants are important to the analysis of the corium dispersion in the ! reactor cavity. These are listed below. Corium Discharge Time: ry= M'*"""" . (2-28) xpfD]v, I4 Primary System Blowdown Time: Y"' r,, = . (2-29) 0.6xDjv, I4 L Corium Film Transport Time: r f= -- ' . (2-30) v, S Corium Entrainment Time: r, = tAl . . (2-31) e The latter two time constants are particularly important in determining the dominant transport mode of the corium outside the reactor cavity. 1 2.3. Aoplication of the Scaling Study j The above scaling study is applied to the following typical reference conditions, (see Table 2.1). Cavity geometry: Zion Reactor Vessel Break Size: 35 cm in diameter Molten Corium Mass: 54 tons j Vessel Pressure: 7 MPa (1000 psia) Cavity Pressure: 0.15 MPa (22 psia) Relative to the above prototypic conditions, several models and simulation experiments l are considered. At full scale and full pressure, the molten corium-steam system is i estimated. For the 1/10 linear scale-down model, the following cases were considered for ; sample calculations. Corium-Steam: fullpressure,7 MPa l Thermite-Steam: full pressure,7 MPa . Water-Air: reduced pressure,1.4 MPa, (200 psia) to 0.1 MPa Woods Metal-Air: reduced pressure,1.4 MPa, (200 psia) to 0.1 MPa The break size for the liquid discharge is geometrically scaled, thus it is 3.5 cm diameter. However, in order to evaluate the parametric effects of the gas flow as well as to compensate for the reduced pressure in the vessel, several enlarged flow area for the gas jet with the increment of twice, three times, five times, and seven times are considered in the sample calculations (Table 2.2). NUREG/CR-6510 12 I l l 2.3.1. Prototypic Case When the reference conditions were applied to the phenomenological models used in the scaling study, the following results were obtained (Table 2.2). The molten corium is I discharged at the velocity of 43 m/s. The disintegration length for the single and two-phase jets are 6.5 m and 1.7 m respectively. Thus, the corium jet tends to disintegrate before reaching the floor of the cavity in the case of two-phase flow. A void fraction of 0.5 is used for this prediction. Since the jet velocity is high, the second mode of disintegration due to the relative motion is applicable. The resultant droplet mean diameter is 4 nun with the maximum size of 12.4 mm. The ultimate spherical limit is 1.2 mm. The value of the impingement Weber number far exceeds the spreading limit of 80. Thus both the coherent jet and droplets spread out upon impingement on the floor and form a corium liquid film rather than bouncing back and form a dispersed droplet flow. This indicates that most of the mass that is discharged as a jet and mostly disintegrated into droplets reforms a coherent liquid film upon impingement on the floor. Therefore, for the corium dispersion in the cavity, the liquid film entramment becomes the most important mechanian. The duration of the entramment depends on the liquid film residence time in the cavity. Hence the liquid film motion and transport out of the cavity is also important. For estimating the film motion, the film thickness and velocity are essential. Several reference values for the film thickness are given below. Initial Thickness: Si = 8.75 cm Mmimum Thickness: Smm = 1.16 cm Maxunum Thickness: Smx = 19.56 cm Whole Wall Static Spreading: S = 4.04 cm The initial film velocity is very high at about 40 m/s. Furthermore, complicated three dimensional motion and mixmg due to the geometry of the cavity are expected. Because of these, some of the liquid can climb up the side walls and may even flow over the ceiling of the cavity. In the subsequent analysis, a typical film thickness of 8.75 cm is assumed over the entire cavity wetted perimeter. This value is chosen in view of the initial thickness and the average between the minimum thickness and whole wall static spreading thickness. The conesponding average fihn velocity is 5.24 m/s in the axial direction from the continuity relation. The actual flow should have a very complicated three dimensional pattern. If most of the liquid flows only over the cavity floor then the velocity is about 15 m/s. With the film velocity of 5.24 m/s, the fihn residence time is in the order of 4 seconds. The jet discharge time for 54 tons of the molten corium from the 35 cm break is about 1.8 seconds. When the liquid starts to flow as a film, three different regimes are possible as discussed previously. The stratified to slug flow transition criterion gives the required gas velocity of 300 m/s, which is about twice the expected gas velocity. However, the more important transition is that to the annular flow, which is determined by the entramment process. As shown below, for the case of the sample calculation, the onset of entrainment velocity is exceeded by the expected gas velocity. Hence these two transition criteria 13 NUREG/CR-6510 I 1 indicate that the most possible flow regime in the cavity is annular flow with a thicker fihn ( at the bottom of the cavity and a thinner film on the sides and ceiling. The entrainment ofliquid from the film is govemed by the relative velocity and film Reynolds number. The minimum relative velocity required for the onset of entrainment is given by v,in Table 2.2(c). The steam velocity of 131 m/s in the cavity far exceeds the onset of entrainment velocity of 63 m/s at the assumed cavity pressure of 0.15 MPa. Thus significant entrainment of the film into droplets is expected. The calculated entrainment rate is 9.018 g/cm2 s, for which, the characteristic time is 7.44 seconds. This value should be compared with the film residence time of 3.6 seconds. The two characteristic time constants indicate that the fihn transport and entrainment mechanisms are of the same order of magnitude. A little less than one half of the molten corium is expected to be entrained by the streaming gas, and the remaining mass should be discharged from the i cavity to the lower compartment as a liquid film. The mean droplet size from the entrainment is 4.3 mm, which is rather large, and subsequent disintegration due to the relative motion between the droplets and gas flow may be possible. When the free stream gas velocity and Weber numbers of 12 to 2 are used, the droplet stability criterion gives a diameter of about 0.64 to 0.11 mm. The very i high entramment rate shown above will certainly slow down the gas flow, especially in the boundary layer region, since the entrainment process and subsequent acceleration of droplets requires considerable momentum transfer from the gas to liquid. When the onset i of entrainment gas velocity is used as a mean gas velocity in the droplet boundary layer, the criterion gives the droplet the size of 2.4 to 0.4 mm. Thus it is expected that the sizes of droplets are in the range of 0.4 to 4.3 mm. The significant effects of the cavity pressure on the gas velocity and the entrainment . process should be noted. For example, if the cavity pressure is 0.3 MPa, twice the pressure in the sample calculation, the entrainment rate is reduced by a factor of 2.35. The corresponding characteristic time for entrainment is 17.5 seconds, which is much larger than the film residence time of 3.6 seconds. In that case the corium dispersion is considerably reduced. , The break size has also very strong effects on the corium dispersion. The sample calculations are carried out by assuming the diameter of the break to be 0.35 m. The increase in the break size shortens both the corium and gas discharge time. However, the most important effect is on the cavity gas velocity. For example, a diameter of the break i that is twice as large leads to a four times larger gas velocity and a nearly four times larger entrainment rate. In that case, the entrainment becomes the dominant corium transport process. 2.3.2. Scaled Experiments It should be also emphasized that the above results are obtained from the existing phenomenological models based on experimental data far from the prototypic DCH conditions. Most of the database is obtained in standard air-water systems with a l l NUREG/CR-6510 14 relatively small hydraulic diameter of 1 to 2 cm. Only the onset of entrainment criterion has a relatively larger data base with Da ranging from 1 to 15 cm. Hence, there is a great uncertainty in predicting the corium dispersion in the DCH problem. Two main reasons for this deficiency are: (1) mechanism of corium dispersion are not well understood, and (2) large uncertainty in the scale up capability of the available droplet entrainment correlations. In view of the above, well scaled and focused separate effect experiments on the corium dispersion phenomena may be required. These separate effect experiments should be focused on understanding the mechanisms of liquid dispersion and establishing a data base which can be used to develop phenomenological models applicable to prototypic ! conditions. For this purpose, several scaled down experiments and simulation experiments are evaluated as a demonstration of the bottom-up scaling method. The results are summarized and compared to the hypothetical prototypic conditions in Table 2.2. In the sample calculations, three different cases are considered. The most important phenomena ofliquid entrainment and droplet size are discussed below. i) Corium-Steam (1/10 scale, full 7 MPa pressure) 2 The entrainment rate is 1.27 g/cm s and the droplet size is 1.36 mm. The entramment rate is roughly 1/7 of the prototypic case, thus the entrainment time is 5 second, which is comparable to the realistic case. The liquid film transport velocity is essentially the same, at 5.2 m/s, thus the fihn transport time is reduced by a factor of 10 which is the linear scale rate. Hence the relative significance of the entrainment is reduced by a factor of 7 in this system. This is a significant scale distortion. The droplet size is reduced by a factor of 3. In comparison the system dimension is reduced by a factor of 10. Thus there is a scale distortion in the surface area by a factor of 3. ii) Water-Air (1/10 scale, reduced pressure 1.4 MPa,5 times break area for gas) The entrainment rate is 1.92 g/cm's and the droplet size is 0.14 mm. The entrainment rate is about 1/5 of the prototypic case, and the entrainment time is 0.45 second. Here the smaller density of the water has a very stmng effect. The entrainment time is reduced by a factor of 17. The liquid film transport velocity is 6.6 m/s, which is comparable to the prototype. Hence the film transport time is reduced by a factor of 12.5. Therefore, the ratio of the entramment time to the transport time is distorted by ; a factor of only 1.3. The entrainment is slightly accelerated in this system, but overall I the agreement is good. The droplet size is reduced by a factor of 30, and the system dimension is reduced by a factor of 10. However, this distortion can be elimmated if the break area for the gas is two times the scaled value. In this case the droplet size is 0.48 mm, which is about 1/10 of the realistic case. Then the geometrical scales are well matched between the system scale and the internal scale (droplet size). iii) Woods Metal-Air (1/10 scale, reduced pressure 1.4 MPa,5 times break area for gas) 2 The entrainment rate is 1.48 g/cm s and the droplet sizes 0.97 mm. The entrainment rate is roughly 1/6 of the prototypic case. The corresponding entrainment time is 4.72 seconds, which is comparable to the reactor case. The liquid transport velocity is 2.34 15 NUREG/CR-6510 m/s, hence the liquid film transport time is reduced by a factor of 4.5. Therefore, the ratio of the entrainment time to the transport time is increased by a factor of 2.8. This implies that the relative importance of the entrainment is reduced. The droplet size is reduced by a factor of 4.4, which should be compared to the linear system scale down factor of 10. Hence the internal surface area is reduced relative to the system surface area by a factor of 2. It is noted that this system behavior is similar to the 1/10 scale corium-steam system. Furthermore, by reducing the break area for gas flow, the droplet size can be increased to about 1.9 mm, which is much closer to the physical ' dimension of the droplet in the reactor case than that in the water-air system. The above sample calculations demonstrate the usefulness of the bottom-up scaling method in evaluating various possible experimental conditions. It is noted, however, that the numbers obtained for various parameters are based on the best available phenomenological models and correlations. As mentioned above, the data base for these correlations at the prototypic conditions is missing. Hence the verification of both the phenomena and correlations at conditions similar to the reactor conditions is necessary. Furthermore, the present discussion has been limited to the hydrodynamic effects in the corium dispersion. The effects of water in the cavity and the solid materials in the molten corium have not been addressed here. Both of these effects can have a significant influence on the corium dispersion phenomena, which should also be evaluated by further research. However, the methodology of the bottom-up scaling and its effectiveness in analyzing the key phenomena are clearly demonstrated. NUREG/CR-6510 16 ] Table 2.1, Input parameters used in Table 2.2 ZION-S2D,1/1 ZION-S3D,1/10 (1) Corium,54 ton,7.02 m'. (1) Corium,54 kg,0.00702 m'. 3 pr = 7676 kg/m pr = 7676 kg/m' or = 0.975 N/m o r= 0.975 N/m r = 4.41x10 Pa s pr = 4.41x10 Pa s (2) Saturated steam in vessel,600 K. (2) Saturated steam in vessel,600 K. (3) Saturated steam in cavity. (3) Saturated steam in cavity. p, = 1.155 kg/m' p, = 1.155 kg/m' , = 1.28x10-5 Pa s , = 1.28x10-5 Pa s (3) P, = 6.2 MPa, P = 0.2 MPa. (3) P, = 6.2 MPa, P, = 0.2 MPa. (4) Discharge hole size: 0.35 m. (4) Discharge hole size: 0.035 m. SNL/IET-1, R,1/10 ANL/IET-1RR.1/40 (1) Thermite,43 kg,0.0103 m'. (1) Thermite,0.71 kg,1.71x10" m'. pr = 4370 kg/m pr = 4370 kg/m' or = 0.975 N/m or= 0.975 N/m r = 3.26x10-' Pa s r = 3.26x10 Pa s (2) Saturated steam in vessel,600 K. (2) Saturated steam in vessel,600 K. (3) Saturated steam in cavity. (3) Saturated steam in cavity. p, = 1.155 kg/m' p, = 1.155 kg/m' , = 1.28x10 5 Pa s = 1.28x10 Pa s (3) P, = 6.2 MPa, P = 0.2 MPa. (3) P, = 6.2 MPa, Pe = 0.2 MPa. (4) Discharge hole size: 0.035 m. (4) Discharge hole size: 0.009 m. PU/ WATER,1/10 PU/WM,1/10 (1) Water,7.02 kg,0.00702 m'. (1) Wood metal,56.2 kg,0.00702 m'. pr = 1000 kg/m' pr = 8000 kg/m' o r= 0.07 N/m or = 0.44 N/m r = 0.959x10 Pa s r = 2.30x10 Pa s (2) Air,293 K. (2) Air,293 K. (3) Air in cavity. (3) Air in cavity. p, = 1.71 kg/m' p, = 1.71 kg/m' , = 1.84x10 5 Pa s , = 1.84x10 5 Pa s (3) P, = 1.4 MPa, P. = 0.15 MPa. (3) P, = 1.4 MPa, Pe = 0.15 MPa. (4) Discharge hole size: 0.035 m. (4) Discharge hole size: 0.035 m. 17 NUREG/CR-6510 1 Table 2.2, Sample calculation for various parameters in coriurn dispersion (a) Jet disintegration and impingement CASES ve Lu L2b d d., da B We, We., 4 (m/s) (m) (m) (mm) (mm) (mm) (10 ) ZION-S2D,1/1 39.6 6.5 1.7 4.0 12.4 1.2 0.18 550 4.9 ZION-S3D,1/10 39.6 2.0 0.5 1.3 3.9 1.2 0.18 65 1.5 SNUIET-1, IR,1/10 52.4 1.5 0.4 0.9 2.7 1.5 0.18 120 1.1 ANUIET-IRR,1/40 52.4 0.8 0.2 0.4 1.4 1.5 0.18 30 0.54 PU/ WATER,1/10 50.0 0.7 0.2 0.1 0.2 1.5 0.69 2100 0.26 PU/WM,1/10 17.7 1.8 0.5 2.0 2.0 6.3 0.09 43 1.2 1 (b) Film spreading CASES Si S., S. , 6, vt Ret (cm) (cm) (cm) (cm) (m/s) (10') ZION-S2D,1/1 8.75 1.16 19.6 4.04 5.24 31.9 ZION-S3D,1/10 0.88 0.12 1.96 0.40 5.24 3.19 SNL/IET-1, IR,1/10 0.88 0.12 2.73 0.33 6.95 3.26 ANUIET-1RR,1/40 0.22 0.03 0.72 0.09 7.10 0.86 PU/ WATER,1/10 0.88 0.12 1.95 0.40 6.63 2.42 PU/WM,1/10 0.88 0.12 1.95 0.40 2.34 2.85 (c) Entrainment CASES v, y, e(g/ t d d dw. do, Re g We 2 5 (m/s) (m/s) cm s) (s) (mm) (mm) (mm) (mm) (10 ) (10 ) ZION-S2D,1/1 131. 63.0 9.02 7.44 4.30 13.5 0.64 0.11 31.3 10.2 ZION-S3D,1/10 131. 63.0 1.27 5.27 1.36 4.26 0.64 0.11 3.13 1.02 l SNUIET-1, IR,1/10 131. 56.1 0.89 4.30 1.35 4.22 0.64 0.11 3.13 0.86 j ANUIET-lRR,1/40 139. 56.1 0.32 3.09 0.62 1.95 0.57 0.10 0.83 0.24 l PU/ WATER,1/10 Dj = 3.50 cm 21.8 13.3 0.10 8.94 1.21 3.79 2.14 0.36 0.54 0.26 4.95 cm 43.6 13.3 0.35 2.78 0.48 1.50 0.36 .060 1.07 1.02 6.06 cm 65.4 13.3 0.75 1.17 0.28 0.88 0.14 .024 1.61 2.30 l 7.00 cm 87.1 13.3 1.27 0.69 0.19 0.60 0.08 .013 2.14 4.10 l 1 7.83 cm 109. 13.3 1.92 0.46 0.14 0.44 0.05 .008 2.68 6.40 8.57 cm 131. 13.3 2.69 0.33 0.11 0.35 0.03 .005 3.21 9.92 PU/WM,1/10 Dj = 3.50 cm 21.8 43.4 0.08 92.6 - - - - 0.54 0.08 4.95 cm 43.6 43.4 0.27 25.7 - - - - 1.07 0.33 6.06 cm 65.4 43.4 0.58 12.1 1.91 5.98 0.78 0.13 1.61 0.73 7.00 cm 87.1 43.4 0.98 7.13 1.30 4.08 0.43 .071 2.14 1.30 7.83 cm 109. 43.4 1.48 4.72 0.97 3.03 0.27 .045 2.68 2.04 8.57 cm 131. 43.4 2.08 3.37 0.76 2.37 0.19 .031 3.21 2.93 NUREG/CR-6510 18
- 3. TEST FACILITY The Purdue University 1:10 scale DCH test facility (Fig. 3.1) is designed to perform separate effect experiments on corium dispersion based on the stepwise integral scaling study by Ishii et al. [16,28]. Air-water and air-woods metal are used in the experiments during separate runs to simulate steam and the molten corium. The following prototypic reference conditions are observec Reactor vessel break size: 35 cm
' Molten corium mass: 54,000 Kg Reactor vessel pressure: 7.0 MPa Geometry: Zion PWR ; i These reference conditions were also used in the SNL and ANL counterpart integral effect experiments. The SURTSEY test facility at SNL is a 1:10 linear scale model of the Zion Nuclear Generating Station. The break size is 3.5 cm in diameter, simulating the l ablated hole on the RPV bottom wall. The COREXIT facility at ANL is a 1:40 scale model of the Zion reactor system. Different from both SNL and ANL tests, the Purdue University DCH experiments are designed and operated under a vessel pressure of 1.4 1 MPa (Low Pressure Test) to 14.2 MPa full pressure (High Pressure Test). Furthermore, water and woods metal (melting point: 75 C) rather than high temperature thermite are used in the tests to simulate molten corium. The initial conditions and geometric features of the ANL, SNL, and Purdue University DCH tests are summarized in Table 3.1. 3.1. Low Pressure Test Facility The schematic diagram of the low pressure test facility is shown in Fig. 3.1, which consists of four major parts, i.e., the liquid and gas delivery system, the test cavity, the subcompartment system, and the containment / exhaust system. The entire system occupies a space of 8 m x 7 m x 9 m. The liquid and gas delivery system consists of three 0.6 m' air tanks (200 psi maximum pressure capacity), a 2 hp air compressor, a manifold of air lines, and a test vessel. The test vessel is made of stainless steel with a volume of 13 liters and a halfinch thick wall (Fig.3.2). A 5.08 cm pipe with a 3.5 cm nozzle simulating the melted penetration connects the test vessel to the cavity located beneath the vessel. For air-water tests, a solenoid valve is installed in this line to trigger the water discharge (Fig. 3.2). In the latter l woods metal experiments, the soleuoid valve is replaced by a rupture disc to avoid malfunctions due to liquid metal freezing inside the triggering system of the valve. This solenoid valve is then relocated upstream of the test vessel. Once the valve is open, the pressure in the vessel rises and breaks the rupture disc to trigger the experiment. In addition, a melting pot is connected to the vessel and the entire liquid metal delivery line as well as the test vessel is heated up to 95 C with electrical heaters. The vessel pressure l and temperature are monitored with pressure sensor and thermocouple. The gas is ! 1 l ) 19 NUREG/CR-6510 2. 4 Ps labd , . 2 N4 253 4 5 )t s . ) 0 1 M t ePC 1 0 c t 1 m%% u s 1 s70 d e ( do 2,1,0 0 nr0 ei0 0 t e p a3 u oin3 o H dr wB S ( C P u
- 4. P s 0%%
403 ir475 1 5 D 1 M 551 a1 23 e u dr - u a 4 P 2. P N425 5 2 4 3 d 1 M 1 0 n a 0 ) ) t c 1 u r d I a e 0 4 nr0 L ( e 9. P 0. a t 0 N629.5 5 ei0 a 3 0 N dr u 6M 7 w 3 0 3 p o A , P u ( 5 1 L a
- 4. P r475 N
S f 1 M ia1 23 o ) s s 0 ,02 n 4 1 lAlA005 m '0' 7 a21 x8 9 1 20 i o 1 ( 7. 0 % te6 0 5.N 30 5 0 L 3%2 1 t s t. 0 i N 4 6 d A 3 n o c ) y 0 e ,r t r 1 1 3 FC0 0 m 5 45 8 20 8 4 e ( L 4%%3 21 2 a 2 3. e6 t s 0 3 9N0 8 3 3 m N 5i o e S g t d S n S a P r, % P 0O 0 0 l i a N n 0 0,U 2 60 40 m.6 a2 e63 005 00,ir37 a 3 0 6 i t o 4 51%Z r, 2 t s 3 7 7 2 i n iZ 6% f 5 o 1 y ) r ) a )K n ) ) a ) 'm K P m ( ( - m ni o (. e no er o tn M')m e (c ot c . mi u m )i t gi a r . i( t u Kops ssf a i=a er (e izs ui t loo s a. S (s -_ = umk s sua vp v a ma s em. = el e r or e o . ) 1 m oc (m - a c pvb fr ct K g 2 ( l e t. . t r e b s a. a a s a s a Wy T M G : i t t le V - v a P M R C C o 3EenW&5o 1 J call l l % l. . ----...a - - - ~ ~ containment relict tak e - ,-.w.s be.,p.. ** ass line .,w,.,# air vent: isolation t al e s gNlrx ! e " =, , ' , 4 i =~ i ruP.ture disk charpine tank j QN~j 3 charging s alvd\ <).. P 4 N..YN. 42 Ml N tanks nw
- L P~ '
3 3 < seal tablg \ . [ ' El3Lrg p %q j reactor ~~~ cart cmits ~~/ % 4 > + um;wwx --- ~no. m 'imm,. - a,hzer;;~c,wi%;. zwc::- r/&g, !v3 i Fig. 3.1 Schematic of the PU low pressure test facility 21 NUREG/CR-6510 To Air Tank i , Liquid Charge Port i j i i i Test Vessel,13000 cm I ! Pressure & Temperature w . \ MeasuringPon ~ l 5.08 cm ID, S.S. Pipe # l i.~ r 4 ,' - Rupture Disc or . . Solenoid Valve 5 ;e l/ . m ? t .- 3.5 cm Discharge Nozzle t i f Fig. 3.2 Test vessel and discharge line NUREO/CR-6510 22 269 m . D=518 mml 1 i SCm { RPV
- 28 a,
/ 0 i G i g I n, i s* .' I l 4 221 mm 517 7 mm l 391M mm 937.5 mm l 1196.5 men (a) Dimensions of the Zion reactor cavity (1/10 scale) i #2 7.5* RI "I _ 517.7 mm 9374 m:n (b) Floor plan view of the modified cavity (shaded area,1/10 scale) Fig. 3.3 Cavity geometry (1:10 scale) 23 NUREG/CR-6510 l l l l ) l l f^ %s l { :wn.ht I i.7.!j prusurew p CMify MII .VCSSbNNE 4-mem /> i g s. -- 3.-trirn ,7 l "Y s w'/ % y / / '- ~u ' \ C, !/ \\s\\ \" \\ \ l- # F1& sun, ~CD{I48 I c, \$g\ g:/---~.- 1 g J 8 i ]I ~ ! 'f \, \, g\ Yg ' as \\ J s , -1 ,- .. dischaq;t ~ j t g 9 ,y, gh.j j / g[. " ! J60/sIcd i El O' b; *l O .
- p a c . a 8
w ,- 4 '\ ,'
- E C?.Vity -
a v 4 4 D ~ l gfM%s, . , ,{/ 34 !CH3 i CH2 CH1 I i 1197 mm 1 Fig. 3.4 Dimensions of the test cavity NUREG/CR-6510 24 3 supplied from three 0.6 m air tanks pressurized to 1.4 MPa with a 2 hp air compressor. Since the gas supply system is designed to operate at a pressure (1.4 MPa) much lower than the prototypic reactor vessel pressure (6.2 MPa), an auxiliary gas discharge line of 10.16 cm ID is thus installed, which bypasses the test vessel and leads directly to the test cavity, providing the correctly scaled gas velocity in the cavity. A 3-inch solenoid valve is equipped in this auxiliary line to control the gas blowdown. It opens right after the completion of the liquid discharge. The test cavity is designed based on the Zion reactor cavity dimensions. Fig. 3.3 (a) shows the 1:10 scaled dimensions of floor plan view of the prototypic reactor cavity. For the ease of construction and material efficiency, the floor plan of the cavity is slightly modified as shown in Fig. 3.3 (b). The circular part of the prototypic cavity is replaced by extended side walls. However, the cross sectional area of the cavity is scaled properly, j and the area difference caused by this modification is less than 1%. In Fig. 3.4, the cross ' sectional view of the test cavity is shown, where the bottom of the RPV is replaced by a flat plate, which isolates the test cavity from the subcompartment. The liquid and gas discharge lines penetrate through this plate to simulate the prototypic vessel break. In order to access the inside for the sake of cleaning after woods metal tests, the test cavity is constructed in two sections: the chute section and the pedestal section. These two sections are connected by a pair of rectangular flanges as shown in Fig. 3.4. For the purpose of flow visualization, transparent polycarbonate glass (Lexan) is used to build the test cavity. One inch thick Lexan sheets are employed for the bottom plates, the floor of the riser section, and the end flanges. The side wall plates are three quarter inch thick. During the course of construction, all plates are dry assembled using screws in the comer area to ensure proper fitting. Finally, the cavity is glued together with these l screws as guide posts. A special epoxy, polyurethane adhesive (Hanel Plastic Company), I is applied to cement the joints. On the bottom plate and the side wall of the cavity, instrument penetrations are prepared for various measurements. These include the film thickness probe ports, hot film probe fittings, pressure taps, thermocouple joints, and isokinetic sampling probe insertions. The entire test cavity is mounted on a cart made of steel angles. This cart serves as a solid support when the cavity is installed to the subcompartment. For maintenance, the cart also serves as a conveyor vehicle for the cavity. The cart sits on rails so that the cavity can be easily aligned to the installation ports. To firmly hold the test cavity on the cart without hindering the instrumentation penetrations, a complex mounting frame is carefully designed. When the cavity is mounted to the subcompartment, the cart is raised from rails, and is then provided with a direct ground support. The subcompartment of the test model is also scaled to the Zion geometry. However, for the case of construction, the outer shape of the subcompartment structure is octagonal rather than circular as in the prototypic case. The interior parts of the subcompartment are designed to model the refueling canal, steam generators and circulation pumps, reactor pressure vessel, and the seal table. The steam generators and circulation pumps near the cavity exit are modeled with wood cylinders, while the reactor vessel and other 25 NUREG/CR-6510 \ structures are modeled by rectangular boxes occupying their respective volumes. On the ! ceiling of the subcompartment, four air vents are provided to connect the containment / exhaust dome located on top level. The bottom floor is built with 1.5 angle slope so the liquid accumulated on the floor can be drained and collected. On the side walls of the subcompartment, one visualization window and two transparent entrance doors are installed. The entire structure is elevated about 1.3 m above the ground level, as shown in Fig. 3.5, so that the test cavity can be easily moved to the mounting location. Three instrument ports are placed near the cavity exit for isokinetic sampling probes. In Fig. 3.5, the subcompartment is shown with various level dimensions. In Figs. 3.6 to 3.8, the plan view of the three levels of the subcompanment are shown. The semi-cross sectional view of the subcompartment assembly is shown in Fig. 3.9. In Table 3.1, the geometric parameters of the Purdue University facility are compared with the SNL and ANL facilities. The structure materials used for the construction of the subcompartment are marine grade plywood and chemically treated wood beams. For ease of cleaning after woods metal tests, all the interior components and walls are covered with 1/32 inch thick stainless steel sheets. The comers and joints of the subcompanment structure walls are sealed with siliconized acrylic caulk. Outside the subecmpartment, a steel beam enforcement suucture is mounted to support against the internal pressure. Moreover, three light fixtures are placed inside the subcompartment for illumination. Above the'subcompartment on the second floor of the laboratory building, a square shaped containment / exhaust chamber is built with marine grade plywood (Fig. 3.10). Its inside walls are also lined with thin stainless steel sheets. Here, an accurate geometric scaling of the upper containment is not considered important because the main purpose of this facility is to study the hydrodynamic aspects of the liquid entrainment in the cavity and the liquid entrapment in the subcompanment. Nevertheless, the air vents and the seal table exit that connect the subcompartment to the upper containment are exactly scaled (Fig. 3.11). To entrap the dispersed liquid droplets, several wood partitions are erected inside the containment to diven the gas-droplet flow as shown in Fig. 3.12a and 3.12b. In such a way, the droplet carryover would deposit on the partitions, which allows more accurate estimation to the liquid carryover in the blowdown tests. The fine droplets that can survive the entrapment of the partitions will be eventually captured by a meshed screen gas / liquid separator equipped at the exit. On the side wall of this chamber, a transparent window is built for flow visualization purpose. A entrance door is also equipped for the collection of the total carryover after each test. NUREG/CR-6510 26 _ _ . . _ . . . =
- sw .m
.4 ~'C nun. . . 6 .._.y._.. . 5 . 1"FihitsDiiiiddWi _ .f.. : j se s .n.= 4 + ....j-I - E f e N . .E y '., E_ , Yq , -, t. l ~ . b Ex.:.ir. ment:xhaust .1 l
- t. . !
O dCel 4 [ gc ,rr: ro r \ rver- l ^ f Level 3 EN) Li QQ%5$$$55555!#25'ikl{Q$kf-
- t t
a b 1 g g Pf ::. e s : ; .- z
- y. 5 !: 3 1,
E I! 5, - , , r. - . y, m v c i. v s- r. . g-- ., -~~ w . i I B : _l ^! . %.i ; Subcompximer.t O g _ Levi.1 e e. t j _ G[ \'\ .. -. g > m = : f; J - Ei S: 6 Test cavity i r: e 4 ,: 2 I NYns3 c.w @T i _I , .1 . . d1, . ?""e"I .r ' Nk '.. ~ < ~:.- ;. =.: , Fig. 3.5 Dimensions of the test loop 27 NUREG/CR-6510 fY' I
- sco mn-1220 m:r.
- n. .
.R ,. . .. Q. ,.-.=.. .... , / ..[ ' Drain .\ -'N.- e . Q- g i s. *t - lll) .,a -* e- . .. .. . . . . . .. . . . . . .. ..
- g. ,
f:,,, ; ;.1 ~, ,5 g. 71 ~ .9:. , - t9,9 .
- M: -
.2 . f ,f g f ? * ). ,' ~* p + > g3 g : ~gg ,Y p& ?, ;' ; >; O -:Jt19.nfrL1i, , _ . . +. ..
- w , . ..
*Z (;: '. i,, ({?;i.ll- ,eyU ',h 7c'} J ; 3*' 7 .g. . i. ,{ ,'j ..J _ .gi ~ l Q , : ^ +- ~- . n. .. .... [ e , -. ;j b/ . ,x .. s , , .:p :: . ; a, : . %o . . . , . . I
- v. . s*** .y ."?*- ;3 ;.,;,. 2
. a i l Fig. 3.6 Plan view of subcompartment (level 1) i l I NUREG/CR-6510 28
- m rn - -
i220 mn y
- u. ., , :--M1 ..
[Ww.: .. .f . l ip b ,- rc " - .s ai .- 3-
- c. , < . 1;, -
i . , _ m. , s , ~ . ,;:.
- w*
s. , e ; , ., 3. s . Y 3 g,.~. .\ f %- ,,_ j - ~ (.4 j.lg: -_, ;
- . lh,,s.,9,4 .'. . * .$.+t f.ca " t - u, m.,.
m m 19 <- : ... w t: . . 1 .u. u4 d w fr g: -~s~ R RO.so h_ . . . q. q _f.
- y. f< .l.k
- 7- >
s.+:h. .wn :- 3 .<. . ;m9mq . ..,ge r- .T.v. E 1 l '. . ;ces <- 4 t.'- 50 - v . .+*- 4:: ,' .,- a .,.. ya. yA g e.s V:a.3.,s. ..r.3.. v ~< c . ~ ~ i . e
- y- m- -
4 .:;3 T og7,g .u: . . ,?.z.. . i. . , py . .. < * < g' , % . . t Qj ,. ), . U"s [4 .. ..-, . ,_ .)h ;1 '3 . .,l *+ y~.,. . . n,,. a .o. c. ~. . ffy -. . ? q~w.L., , y;p ?:: b . ~ .,m*' :'?'.f ' .. .'iWc myn '-- 1.?I. .e . .,e . _ .
- k,.. ' ?p;,L.,. ,% .; (. < J n ::.nl- . . .,. / l
.,,~.r. ,* au;~ , , . . . t h "'
- 9. ~- t ,.f,..<.e.+, ,. . . N t.,
smww. e.. . ./ ..e,.. ,p.8 - f c s . r. . .w r. , .m.. . .. .w- - . .m! - ? ,)_ vjy.:, .. r k.g.;' ,c., .* g, K T? r..~..T , *... 9, .i,?[}ii:,C,+,5..' . j . , . f .; - . I - /7 / ,Aa' * ..ph 8h 'h" 'e e. .%d p -ey* 'd
- 4 $ /if '
. . I. . ,k, ,.'....i~.$. , , aE.[~4 .v [ =4,. % ,w vw na - N Ws \.V l?.'l$.,'3 :' YiW - v-w \.$,Yy . ~Y,.,ne'fh.15 ) ,s v ~' .,e F.,.- n .C nL- y y p .s ;9. . p.. . b - : ya.,; y.- .: :;,= - 4 Fig. 3.7 Plan view of subcompartment (level 2) 29 NUREG/CR-6510 i m;. .:. .f-. , , .
- p. : $b~.9 ,
2.. . Drain s k s. Q g x M~r ly_ ' .. , . 4,. *$' hin: N. - ^ ^ ',. _ .p EYlEi 92 ? f2) ' b 'T1
- s:al table ex;t .
jy. dg RPV g, S,, lN $9.8 cry J .,, ,.. . . .J;dO ~' ! h ,l d!.:.:..
- 7. ).
h:- , h; 0 .:., y p.1 IEi] ig$. !? : -- J v, . ,, , ,4 : ; aqy i . pr z,V i ,; , s f o , ('# Mm.m.7: , ., , . .gg., :^;r . {f.: 3 .......... ,,;.. , 3 i f Fig. 3.8 Plan view of subcompanment (level 3) / i NUREG/CR-6510 30 i i ~ ' il SG: Steara G:nerator ST: Scal'Iab;c Reom CP: Circulatw. Par.p FC: Tuel Canm1 RPV: Reactor Pressure Wml ***==-a,,,,,~~.., / i-h' ' _ , - bNJ"7~ 'N f --#*" sa , 1 i ' i e i E ! y ej sT SG' r en n g s M ' i hg f b + gy . s ,,, q !p ; ti., \ - n4 ~ a m* p .n%. - ?? i N $3 , #}%'a 'k$$[fik hi ifbdN!Nh;.:v%W/ ' s I[' $f3hhf'1 ~:s. 4 mnsv3 %: ' J .l 47' l L; i a 1 l ' , i .- i % r l
- l. _
Fig. 3.9 Semi-cross sectional views of the subcompartment ; 31 NUREG/CR-6510 4 l e v e L ~ ~ , l y _E r e b m leg a h - c t s u - a l h l x I4 e , / t n = e m i n t a m m n - m o m _ m m c - 0 7 4 0 9 e 4 l 1 8 ht 1 _ 9 2 f l o _ w I % i e , v _ l a _ n _ o ~ i t c e l CE s s . s _ o r _ _t i _. C - <I. 7, 0 1 { _ 3 g i F H RS r e e l f d n o c e S I@E E M M S R @ b- ,I n -d e..
- u
$p$$, .h,. . 4 a n E E a Seal table exit u
- e. eses mi 9
E - ae LE y .ge 9:: n ,ggas t,_ _ f' o.0965 a7 g/ > ' P7. - y d ifd Efd le: 2890 mm Fig. 3.11 Plan-view of the upper containment (level 4) I 33 NUREG/CR-6510 i 5 Y\ k !! i A \\\\ \\ \. \s Xtx - \ , i T y w r-- - ] s[9 .! 1 Tx\= , , , J" - 4 .] 4 \ \ l c . / - 1yNa 1 s .! e N \ i \ . \ I J D i NUREG/CR-6510 34 s:'d l p-s s , 2 - 'l / / exhaust sQ -, / / / / r upper ceih.ng . 1 ? f % 1 h:$$ h%j hNj f:N'j l i i Level 5 1 Fig. 3.12b Top view of the upper containment i l 1 l 35 NUREG/CR-6510 ) l l 3.2. High Pressure Test Facility The high pressure test facility (Fig. 3.13) is rebuilt based on the low pressure test setup. The original gas and liquid delivery system were designed for the pressure up to 1.7 MPa. Therefore, for high pressure experiments, some necessary modifications should be made. These modifications include cavity enforcement, replacement of the gas and liquid delivery system, and the construction of a heavy duty test vessel supporting structure. The new liquid and gas delivery system consists of a reactor test vessel, a gas tank, a bypass line, and two rupture discs for test initiation. These components are all made of carbon steel or stainless steel with pressure rating up to 18 MPa. The 16 liter test vessel is made of stainless steel (Fig. 3.14). A schedule 80 stainless steel pipe of 5.08 cm OD connects the vessel bottom to the test cavity, and a 3.5 cm nozzle simulating the reactor vessel break is mounted at the top of the pipe. Below the nozzle, a rupture disc (manufactured by LaMott Corporation ) isolates the test vessel from the cavity. This rupture disc breaks only when the vessel pressure exceeds the designed value. Gas is then 3 supplied to the vessel from a 0.5 m steel gas tank through a 7.62 cm stainless steel pipe. At the top of the tank as shown in Fig. 3.13, another rupture disc is installed, which breaks if the pressure difference between the gas tank and the test vessel exceeds the disc i burst limit. In the experiment, the gas tank is pressurized to the working pressure, while the test vessel is charged to a desired pressure through a 0.9 cm bypass line such that both rupture discs do not break. The test is initiated by opening a small relief valve (0.64 cm ID) to reduce the test vessel pressure. As the vessel pressure drops, the pressure ' difference between the gas tank and the test vessel builds up. Once the first rupture disc at the top of the gas tank can no longer stand the pressure difference, this disc breaks. ! Instantaneously, the high pressure gas in the tank pushes to the test vessel and ruptures I the second disc to start the blowdown process. The crucial factor here is to choose the l right combinations of the two rupture discs to prevent premature disc breaks. Rupture disc selection details will be presented later in the test procedure. 1 I During the blowdown transient, tremendous thrust force is exerted on the bottom of the test vessel. To hold the vessel rigidly, a strong support frame anchored to the ground j is constructed with 10.16 cm x 10.16 cm steel I-beams and 10.16 cm x 15.24 cm steel channels (Fig. 3.14). On the other hand, the gas tank is welded to a halfinch thick steel plate, which is anchored on the concrete floor. In woods metal experiments, additional ; steps are taken to heat the test vessel as well as th: 5.08 cm discharge line up to 95 C i with electrical belt heaters, keeping the charged woods metal in liquid state. Furthermore, different from the previous low pressure experiment, the high pressure test j is performed with nitrogen gas instead of air to simulate steam. Commercial 42 MPa N2 cylinders are used to pressurize the system. For flow visualization purpose, the original test cavity for low pressure tests is made of transparent polycarbonate (Lexan) board. Due to the large surface area, the cavity can only hold a pressure up to 170 kPa. In high pressure tests, it is estimated that the cavity NUREG/CR-6510 36 pressure can be as high as 280 kPa. Therefore, cavity enforcement appears to be necessary. The enforcement is done by armoring the cavity with 0.64 cm thick steel plates, which makes the cavity strong enough to withstand a pressure up to 350 kPa (Fig. 3.15). Unfortunately, this modification causes flow visualization to be impossible. The whole structure is bolted on a steel cart, and thus tne cavity can be easily detached from the subcompartment for maintenance sersice. 1.19 waled ('onditivas. eshatist (W una Lit.aie. 7.0 lia e N thca',. Size i 2 sui , et stilal11111elli acliel s ab e by -pass li !Ie Nf.1% l s,
- f. j
/ - = =m mutation valve.N 4 - Q4-- N -l Y. ~ 1 ruptine disk _ y } l charping tank ....,'s I, ( i chargitte vah e\s ~.__, x a i \ 43stlG~~~l>H9: N ,, tanks seal table g\ l . _ _ . _ _ i \ ruplurg diskf4p / reactor earl cavity _ . _ . ' -- ; tw .c. '..'3 , I'd g.;;, ~m ., s " s 7C-G'.' 7 W"'?.Ck 's - ~ Fig. 3.13 Schematic of the PU high pressure test facility 37 NUREG/CR-6510 l l '"*" \ 1 ~~*====*=.. ' a.% u_a, _ K'. 1 ~> -e _ ' ,6 k u,,, , _ eA..< v wu, m, Se , y ,. -- < - dwbtMet 4 "N. SV 2* A ,1 'l w ; Gm*uinkW e a.- A' 7 O e , \ y &_ f ssr s Fig.314 LI 6 ed ** B disch8E8 systetu *d supPordng h## gpy,GICPM 38 1 memmi "h. * , 11 s..pa.r.n. '< ; ,o . s9 ,y#. Wi- Q_~.. $ ..,= = =.m - ,. c. - - ~. .r c ,p;r7 7. -'.. =:p m y-n., b . .. : a s- { f. . 1
- n. a s
* ' .s* a. t . __ 1 - : I: le u.pd I E: es an strid sieet ;iaic l- - d: 1.2 em tha.k stunleu stee: plate E: 1 :~ e:n OD :.t.culen ueel bar Fig. 3.15 Cavity enforcement 39 NUREG/CR-6510
- 4. INSTRUMENTATION AND MEASUREMENT Various probes installed in the test facility (shown in Fig. 4.1) provide data covering the entire test transient. A complete list of instruments and the parameters measured are given in Table 4.1. In terms of their functions, the instruments are classified into six categories, which include droplet size distribution and mass flux measurements, liquid film thickness measurement, gas and liquid velocity measurements, pressure and temperature measurements, flow visualization, and data acquisition system.
4.1. Droplet Size. Size Distribution. and Mass Flux Measurements The droplet size, size distribution, and droplet mass flux at the cavity exit are the most important parameters in the DCH separate effect experiment. These parameters are measured with an isokinetic droplet sampliug probe along with the image processing system. The schematic diagram of this system is shown in Fig. 4.2, including an isokinetic probe for droplet sampling, the droplet collection mechanism, an image processing system, and the locally developed software for droplet counting and sizing. Isokinetic sampling probes have been widely used for velocity or mass flux measurements in single-phase and two-phase flows [29] as well as in solid particle sampling and sizing [30]. However, its application to droplet sampling and sizing has not been explored due to the difficulties in droplet collection and possibh measuring distortion caused by droplet deposition, break-up, and coalescence. Traditionally, the isokinetic probe consists of two thin-wall tubes, the larger sampling tube and the smaller pressure tube (Fig. 4.3). For incompressible single-phase flow, fluid velocity is related to the dynamic pressure by Bernoulli equation: 1 -pv2 , p _ p. (41) 2 It suggests that any disturbance in the pre-existing local pressure would lead to the deviation in the measured velocity. If the disturbance is small, this velocity deviation is: v, - v. p, - p.
- v. pv, Therefore, if an isokinetic condition (p, = p.) is reached, the measured velocity will be equal to that of the undisturbed stream. This condition is essential for correct velocity measurements, especially in a two-phase flow system, where a slight deviation from the isokinetic condition may result in an appreciable error in the gas velocity measurement.
To maintain the isokinetic condition, the attached small tube in Fig. 4.3 supplies sufficient pressure at the tip of the probe set. For droplet sampling the gas velocity is not I NUREG/CR-6510 40 ( Table 4.1 Instruments and the parameters measured Instrument Parameter Pressure transducer Pressure Thermocouples Temperature Hot film anemometer Liquid film velocity ) Conductivity probe Water film thickness Capacitance probe Metal film thickness Pitot tube Liquid flow rate Pitot tube Local gas velocity Pitot tube Gas volumetric flow rate Isokinetic probe Droplet sampling I Isokinetic probe Droplet mass flux Camera Flow visualization VCR Flow visualization 41 NUREG/CR-6510 1 I t
- i
. .) 3 6 ; ' A: = {- 3 - - - . 3 j 7 Y[ (j 's l <J ,urf i j :. s I"- Y N, l' ,7 i 9 ' ^ 1 I fg [ .! %; ~'d 9.. J J 13 85'bd f
- r. -[_
w 6 ~% 16 I I ; 7F 1 11 5 _ ?- -7' I i ' .:; . -- u , . _.._g I
- 1. reputater 9. droplet collect:ng im
- 2. charging sah e 10. isek:netic protr 2 im:aang s ah e 11. therrmi niuple I J rehef vahc 12. heaters and insulators
.;1 -pass 3 ':ine 1.4 Ap measmrment line
- 6. wn:erfuoods meta! feed tank 14. film thickness prulv
- 7. oil! heavy syrup feed-tank li pitot tuk N. wienoid valve 16. dats ac<1uisition and eenuul .sp;en. )
{ 1 Fig. 4.1 Instrumentation in the test facility NUREG/CR-6510 42 drivere oi! inieciion J, ... 'I shutter "]lu l isokinetic probe xN _ _ ~ vent 1 j N ' .\ . ; g N soi.l film - N : [ , sampl.ing ' ~ window i F ( j . . . . .. ._. .. conlainer . . 'h p o rt a bl e i 1 #, droplet l image , - container l monitor . p j ; %; l i I ... e ! l { Y T IZi 8 \ g computer ; e i n u. .nm [,_ CCD imcoe hN [ ~ printer results grobber Fig. 4.2 Droplet sampling and sizing system 43 NUREG/CR-6510 F l ! i droplef flow samph,ng tube "maanwnnn --
- J-__
= Pm- 8.75 1,m. m-9.53'mm . T----~.-~ .m g Pw pressure tube l i, .I L Fig. 4.3 Droplet sampling probe l 4 NUREG/CR-6510 44 l [ a concern. If the deviation of the gas velocity is small, the motion of most droplets would not be affected due to the large inertia of the droplets, unless the size of the droplets are extremely small (sub m range). In present DCH tests, since the liquid droplet size is in the range of several hundred micrometers, as long as the deviation of the mean gas stream velocity is about the same order of magnitude as the turbulent fluctuation, the isokinetic condition has little effects on the droplet size measurements. This conclusion was l verified with a separate test loop [31,38] and the similar investigation was carried out by ; Paras et al. [32]. For this reason, the probe in present tests consists only of one tube of l 3/8 inches ID. The isokinetic probes are mounted parallel to the flow stream to direct the droplet flow into the collection complex. Careful measures must be taken to prevent breakup and coalescence during the collection process. One way is to expose an oil-coated plate perpendicular to the particulate flow. The thin layer of the oil film may, however, not be sufficient to prevent large liquid droplet break-up upon the impact. Another method is to freeze liquid droplets using cold gas, but it requires a relatively long freezing path, which makes the system impracticaly large and complicated. Rao [33] developed a method of using liquid nitrogen injection to freeze drops inside the tube. The drawback of this ! approach is that the frozen droplets and film may block the probe tube. I In the present system, the droplets extracted from the flow field are collected in the collection box by a falling oil film for water droplets and a pool of highly viscous syrup for woods metal particles. For the collection of water droplets, viscous oil (Amoco #200) is injected from the top of a cylindrical collection chamber, forming a falling film around half of the inner cylindrical wall (Fig. 4.4). Incoming droplets form the side window are captured in the oil film and then transported to a portable box located under the complex. Since the clean oil is supplied continuously, the problem of droplet coalescence encountered in the oil-plate technique could be essentially eliminated. To prevent further coalescence in the portable box, an electrical shutter is mounted between the probe and the collection chamber to control the sampling time interval so that the number density of the entrapped droplets is low enough to avoid droplet coalescence. In addition, the droplet disintegration is prevented by carefully selecting the properties of the oil. The oil should be immiscible and have high viscosity, a smaller surface tension, and a slightly smaller density compared to the liquid droplets. Under these conditions, the droplets would be stopped within a short distance from the oil surface inward without disintegration, and the penetration depth is as short as several droplet diameters. For heavier liquid-metal droplets, a pool of liquid is used for droplet collection because liquid metal droplets have a much longer penetration depth due to their large inenia. Fig. 4.5 shows the method used for woods-metal droplet collection. In order to effectively stop the droplets, viscous syrup is used as the collection liquid. The syrup is made by melting five parts (mass) of pure granulated sugar and mixing with one part (mass) of water. The symp is injected into the pool to a level above the sampling window. A shutter opens the window when sampling begins. Due to the gas impact pressure on the window area, no syrup can flow out of the pool in the course of sampling. 45 NUREG'CR-6510 m driver n ! cil injection f ,_. a f q shutter j', i L_ isokinetic probe \- ' -~ vent s - g 1_-_- 3, 1 j N j v.l filmoi s ! N sa mpl.ing i }[ - window I-i. i> 1 'l{} portable Q . . . . . . . . . , . . . container Fig. 4.4 Water droplet collection. mechanism NUREG/CR-6510 46 I I _ . _ _ _ __ 1 - _ - a v. injection of viscous fluid shutter ll (syrup) 1 '(driven from side) N _. , vonj isokinetic probe . \w s : I - G~ -j-. ' ]; m . l\_ . s - \ -( l- '; sampling window v ., I I portable V . . . . .. . . .. . . . . container Fig. 4.5 Woods metal droplet collection mechanism 47 NUREG/CR-6510 E - 5 2.0E-3 - C - Sid" probe data --- direct sampling data @ 1.6 E-3 - 7%L, s upper limit log-normal _ 3 V - i fitting function ,5 1.2 E-3 - E i j = 35 m/s . .o _ m 9 I f = 0.04 m/s _ 5 8.0E-4 - D : . Q ~ 5e 4.0E-4 - -i E - , .? y ''''' ' ' o > 0.0E0 250 500 750 1000 1250 droplet diameter D ( m) Fig. 4.6(a) Calibration results on droplet size distribution measurement at j, = 35 m/s and jr = 0.04 m/s. NUREG/CR-6510 48 y . - ::t. ' ;:: 4.0E-3 - - - 3/8" probe data _1 @. : j % --- direct sampling data : 3 3.2 E-3 - > upper limit log-normal - 4 [ T' .J. fitting function - c . , S 2.4 E-3 - E - J9 = 60 m/s .o 7- ' *C E ' Jf = 0.04 m/s - 1.6E-3 - o - o . - ~ 5e -8.0E-4 - E - - - - ' -"3 . o .#- ' ' ' > 0.0E0 - - - - - - - - ' ' ' 250 500 750 droplet _ diameter D ( m) { i I Fig. 4.6(b) Calibration results on droplet size distribution measurement at jg = 60 m/s andjr = 0.04 m/s. I I i 49 NUREG/CR-6510 T i I E 8.0E . , , , . , . , . :1 " 3/8" probe data O 6.0E-3 - g'Y direct sampling data _ 'f_ upper limit log-normal ] ] - ' fitting function c - ; .9 i y 4.0E-3 Jg = 96 m/s n . 'E j, = 0.04 m/s ' * = .= t - 2.0 E-3 a . c E .7' ~ E i_ - ,g o 0.0E0 g' ' - ' - ' - " - ' - > 500 100 200 300 400 droplet diameter D ( m) Fig. 4.6(c) Calibration results on droplet size distribution measurement at j, = 96 m/s i J . and jr = 0.04 m/s. 1 1 4 1 l NUREG/CR-6510 50 I *e "by.O..* e d e.o D e c , $ e. o o o ', N . g g' O ( .... 4 .O. -> , Fig. 4.7 Digitized droplet image from a water droplet sample 51 NUREG/CR-6510 l The woods-metal droplets that penetrated into the pool stop in the syrup and then solidify. It was found in the experiment that frozen droplets could be in nearly spherical shape or in irregular shape, depending on the temperature difference between droplets and the I collection liquid. . When the syrup temperature is slightly lower (~5*C) than the melting point of the woods-metal droplets, the shape of the droplets is near spherical without fragmentation. Too low temperature of the syrup leads to droplet breakup upon the impact due to the high gradient of the surface tension caused by the thermal gradient. On the other hand, high temperature of the syrup may result in droplet coalescence. Again, the opening duration of _the window is controlled with an electric shutter so that the droplet number density in the sample is limited to prevent further coalescence. Besides break-up and coalescence phenomena, the droplet deposition inside the probe tube should also be prevented. The way to evaluate the effects of direct deposition is to compare the measured droplet size distribution with the sample obtained without probe tube. In a test loop, the end of the horizontal pipe is fully open to a collection box shielded with a shutter covered window. Droplet samples from this configuration would not be distorted by the assumed deposition effects. Systematic comparisons are performed with different gas and liquid velocities for a various sizes of the probe tube. The results indicate that at relatively high gas velocities (>35 m/s), the change in droplet size distribution due to droplet deposition in the probe tube is insignificant. The diameters of the probes used are 7.9 mm (IJD = 32) and 11.1 mm (IJD = 24). Figs. 4.6 (a) to (c) present the test results with the probe tube of 7.9 mm in diameter under gas velocities of 35 m/s,60 m/s, and 96 m/s respectively. These results suggest that the size distributions measured with the probe are in good agreement with the distribution obtained from direct sampling. Moreover, the curves fit the upper-limit log-normal function proposed by Mugele and Evans [34] for the droplet size distribution in a droplet mist flow. This function describes the distribution of the volumetric fraction, f, of a droplet sample, dV ( dy &,w, (4,3) r ggs > y = In 3 (4-2) < D,3 -D>, k= , (4-5) D, where D, Dv., D , and x are the droplet diameter, volume median diameter, maximum drop diameter, and the distribution parameter, respectively. 4 and Dma needs to be determined from experimental data. After the collection, the droplet sample is analyzed by an image processing system I shown in Fig. 4.2. In this system, a CCD camera digitizes the droplet sample image from j NUREG/CR-6510 52 .I the bottom of the portable collection box. Each frame of the image contains droplets in an area of 3 mm2 as shown in Fig. 4.7, and the images are stored in a computer for analysis. A software is developed to measure the sizes of the droplets in the sample image and a total of about 1000 frames of pictures are needed for analysis in each test. This software processes droplet images at a speed of 6 frames per minute. In the test facility, four probes of 9.5 mm in diameter are provided for droplet sampling at different locations as shown in Fig. 4.1. One probe is installed at 7 cm above the cavity floor near the riser section designated as the low probe. Other three probes are located above the cavity exit in the subcompartment. These probes are designated as high probes and are aligned parallel to the cavity chute section. The normal distances between the probe center lines and the chute bottom wall are 0 cm (HL probe), 6.7 cm (HM probe), and 13.4 cm (HH probe). These three high probes can provide information on the spatial distributions of both the mean droplet size and the mass flux at the test cavity exit. The droplet mass flux is easily obtained by dividing the weight of the droplet sample with the sampling time duration. 4.2. Liauid Film Thickness Measurement The film thickness at the bottom floor of the reactor cavity is measured by two kinds ofimpedance probes. For the air-water simulating experiments, the probe measures the electrical resistance of water arourd the probe needles to obtain the information on water film thickness. For the woods-metal experiments, and due to its low electrical resistance, it is difficult to measure the liquid metal film thickness via its resistance. As an altemative, the probe is coated with insulating paint and the film thickness is measured j via the capacitance of the coating immersed in the liquid film. Six sets of probes are installed along the floor of the cavity (Fig. 4.1) to capture the transient film flow information. The water film thickness probe is made of two parallel conductive electrodes. The electrodes are stainless steel rods of 1 mm diameter and 5 cm height, placed I cm apart on a mounting plug as shown in Fig. 4.8. The selected geometry for probe spacing is based on a numerical analysis of the electrical field around the probe electrodes [35]. The measured electrical resistance is directly related to the instantaneous area-averaged water 2 film thickness in an area of about 5 cm around the two electrodes with 95% resolution. A high frequency carrier signal is applied as the driving power source in order to prevent polarization effects of the electrodes in water. With DC power supply, an ion layer is formed around the electrode adding significant impedance to the measurements [36]. This extra impedance is unstable and has slow time response characteristics (~10 ms). It is suggested that 50 kHz AC power source should be used to effectively eliminate this impedance which behaves like a capacitance [37]. Hence a carrier frequency of 80 kHz is chosen in the present measurements. The electronic circuit diagram for the output signal conditioning is shown in Fig. 4.9. In order to avoid common ground loop current, each pair of probes is electrically isolated with two transistor transformers. This is rather 53 NUREG/CR-6510 important for several probes functioning simultaneously since the interference between I probes would affect each other's reading. Calibration is carried out with a contact needle type probe attached to a micrometer (Fig. 4.10). It is necessary to calibrate the probes in-situ as the water conductivity varies from time to time. A set of special water cells with fixed heights are designed to correct the calibrations for every test. Typical calibration curve, plotted in Fig. 4.11, is linear with a linear correlation coefficient of 0.9999. Six probe sets (from CH1 to CH6) are installed along the cavity floor up to the exit. j Liquid woods metal is an electrical conductor with very small resistance. It is almost impossible to measure the film thickness via its electrical resistance because of the dominant unstable contact resistance. As an alternative, a capacitance type film thickness probe is developed (Fig. 4.12). Instead of bare electrode surface as the one used for water, one of the two electrodes is coated with electrically insulating paint (~ 0.07 mm )' thick). The coated electrode and the liquid metal film flow constitutes a capacitor, and its capacitance is determined by the coating area immersed in the liquid film. With high frequency current, this capacitor can be measured in term ofits impedance, and the same signal conditioning circuit used for water film thickness probes is applicable. Great care has been taken to chose a suitable carrier frequency to avoid the non-linear effects due to the resonance region caused by the RLC input loop. In this case, a carrier frequency of 300 kHz is selected to achieve a linear relation between the film thickness and its voltage output. The calibration is carried out with the contact probe shown in Fig. 4.10. The calibration curve (Fig. 4.13) presents a reasonable linear response to the film thickness range ofinterest with a linear correlation coefficient of 0.999. In contrast to the water film thickness probe, the capacitance probe offers local film thickness measurements around the periphery of the coated electrode. To magnify the output, the probe is designed in a plate form (Fig. 4.12) with large area of coating (1.3 cm wide, 5 cm high). Inside the electrode, an electrical sheet heater (KHLV-0504/5, OMEGA) is sandwiched between two copper sheets to keep the probe temperature above the melting point of woods metal (700C) such that the woods metal does not solidify on the electrode surface. The probe structure is shown in Fig. 4.12 with 0.8 mm in thickness and sharp edges. It is mounted parallel to the incoming film flow direction and the sharp edge would prevent the flow from climbing on the electrode. In the DCH experiments, the discharge of the molten woods metal is followed by high speed air blowing at room temperr.ture. The heat transfer from the probe to the air flow is so intense that the heater insid6 the probe can not mamtain the electrode temperature above the melting point of woods metal. The probe fails due to woods metal freezing on the electrode, resulting in a constant output, which indicates that the probe works only before the gas blowdown. With the measured liquid film thickness signals, several parameters besides the thickness transient can be obtained. These include the time interval for the film front to travel through the test cavity, the duration of the film flow in the cavity, the film front velocity, and the liquid wave velocity along the cavity floor. The time interval for the film front to travel through the cavity is important to determine the amount ofliquid that flows out of the cavity without entrainment. The film flow duration in the cavity roughly represents the entrainment time in the cavity. NUREG/CR-6510 54 1 l l electrodes e, , -a. .. %p! probe plug ,s w., +.+ ~ .,; 'l , _ 'h l S.S Wire ' , .: Lig : [j. 8 6. .'s id f.< c ' .j.:2;w@
- w. if R$?W'F;>ity NBeav bottom
- w:.::' .
I'Ml, , , :; . ~ c. .. p [o ' , - )r -? t.;
- 1;[::i2 ~ ^.- ? ',,
-:+ .t.s . : n -
- t.
u, & m,5 2
- .c, : M ,
~ %. ..ws h,; .. w w l i u Fig. 4.8 Needle-type conductivity probe for water film thickness measurement j l 1 1 a I 55 NUREG/CR-6510 1 a v .,*.3 u a a c~8 , '.,g * " g r e u , ;;, g' . . o ,. _ 'T *L. ,- . e * . a " 7, . a '=- ~. bs. i a 4 t',r. _* g . s , 4 t,,',_ 1-SQ 9 . ,i ,i =, . ,- ,. =, . r "~y 1I-8
- 4
= i Eu.i =m '~~ i "' .- a v ., r, a- ! 8' . ., ' .- _ a y= v R a
- t.
., ,.,. t e 1= . ,- .- e . '. .e='
- s i
= =. .
- p ,
ai n. , a' ~ .t = ,. ,',p , i . , .= ,.' e _ ge b . " ,. 4 3 g e _ * *
- g 4
.L. ,- P l = ' , , , ' -s. y I n =. = r, i t e a. R.a ' o , , v - , * *
- a.
s- , . t r bP i . t * >P . a ,= . c u P. d e =
- d4 r n
o c e - ht iI A - . . = . . f , a. o t i . u ',.. 1p , i cr - c a ,,. s i e . .. v* 5 , .*',a, , 7, i g n n i_ . = o ,s. . - N ,. ,,t v_, . , i , t ',a - a , i I -. d - *l , . ,_ n - . r o c =, u . , ',M e c c 'c l a l a . ._d . n g ,, 4 i . , a ' . S ",,3 *e =,, 9 ., 4 g i F h mt 2C$o3:nhu o c c I .1 I I! E. w. l , It g. micrometer E l 0 ) , 1 i w ' e e ; e V - Liquid )' p > ,l l .s.I. > , ,n u _ a film thickness probe Fig. 4.10 Calibration system for film thickness probes l I t 57 NUREG/CR-6510 f } i i 1 i 1 i 50 ~ l 40 - ! l - 30 - E 5-20 - 10 - 0 =' 4 5 0 1 2 3 V(volts) Fig. 4.11 Calibration curve for water film thickness probe NUREG/CR-6510 58 ym, f %. < -s / i qy / probt: ph:e J,. , q.b I . _ ' . _ ,..;ra pmbe pluu I . e..m. 7: ',it'i % ~ i!?h ,; - .>,. . NIk ..q ..: :. heater I ,/ \ I 2 elec: rode a h
- Coating o ; A-- -
l l 0.3 mm l l i f Fig. 4.12 Woods metal film thickness probe I 1 59 NUREG/CR-6510 I $0 - 40 - 30 - l G 1 9 . # 20 - 10 e 0 0 1 2 3 4 5 V(volts) Fig. 4.13 Calibration curve for liquid metal film thickness probe NUREG/CR-6510 60 4.3. Gas and Liauid Velocities l The gas velocity in the gas discharge line is measured with a Pitot tube as shown in Fig. 4.1. The Pitot tube is connected to DP cells manufactured by Validyne Engineering Corp. The measuring line is filled with water, which leads to fast response up to ikHz. J For the 1.4 MPa experiment, gas supply comes from the 10.16 cm auxiliary pipe. Since the gas flow is choked at the nozzle on the pipe end, the gas Mach number is high at the probe location and the compressibility of the gas flow need to be considered in computing the gas flow rate. A direct way to obtain the gas mass flow rate is to measure the local total pressure, static pressure, and temperature simultaneously, and to use the temperature measurement to calculate the local gas density. However, it is difficult to measure the local gas temperature due to the fast transient process. Therefore, a different approach is applied in the experiments, which is to compute the mass flow rate at the pipe inlet where the gas tank is located. Y A yP,, 1+Y~ M *, rh, = p,,V,, A = , (4-6) ]yRT, where Pio, Tio, and Min are the total pressure, temperature, Mach number at the pipe inlet, respectively. By assuming an isentropic process from the air tank to the pipe inlet, Pio equals the tank pressure Po and Tio is related to the tank temperature To in the form of I T, = T' . (4-7) 1 + I-I M 2 2 Therefore, the main task is to obtain the Mach number Mio. For this purpose, the pipe line is treated as an equivalent 10.16 cm pipe with a nozzle at the exit and the Pitot tube is j located close to the exit. Based on Fanno line results, Min is given by l - 1 1/2 ' -1/2 1 ~2 ' g y-1 y*2 , ' Po ' 1 -2 ' g y -1 y2 (4_g) M,,_y+1\ 2 L <Ps M _y+1< 2 > where M and P are the Mach number and the total pressure measured by the Pitot tube 1 near the exit, respectively. Because of choking at the nozzle with area A*, and by assuming isentropic process near the pipe nozzle, M is given by l l . _ r+i : ' 5 2v-o A 2 -=- 1 1+ y-1 M 2 , (4-9) A M;y +1 < 2 i_ where A is the area of the 10.16 cm pipe, which is equal to the inlet size in the experiments. Therefore, Eq. (4-8) can be simplified for Min at the pipe inlet with 61 NUREG/CR-6510 . 7+1 ' 3 2(7-o rp3 3 1 2 1+y-1 M ,2 = -- ;. (4-10) M,, _y + 1 R 2 > \ Ps A Consequently, the gas mass flow rate can be obtained from Eq. (4-6) if Po, To, and P can be measured. For the case of high pressure tests, the gas discharge and liquid discharge are from the same pipe as shown in Fig. 3.2. The vessel pressure, Pv, is monitored. During the short transient period, especially for the entrainment process, the vessel temperature remains constant Tv. Accordingly, the gas mass flow rate can be obtained in a way similar to the above presented. Using Eq. (4-9), the Mach number, M, at the pipe outlet near the discharge nozzle is obtained. The Mach number, Min, at the pipe inlet is given by { 2 (y+1)y2 4f7 ;_ (7,3)y2 l- M ,+ 7 ,3 In + y2 + 7,1 in = ' '(4-11) D yM 2 2y I I 21+I M[ 2 1+ I-I M 2 2 > _\ 2 >_ .\ _ Where f, L, and D are the average pipe friction factor, pipe length, and pipe diameter, respectively. Subsequently, the gas mass flow rate is obtained from Eq. (4-6) and (4-7). The gas velocity in the cavity is estimated in two ways. First, under the specific cavity pressure and temperature conditions, the average gas velocity in the cavity is computed from continuity equation by assuming a quasi-steady process. V, = - (4-12) A,p,, , p, = " , (4-13) Where Tc, pe, and Ae are the cavity temperature, cavity pressure, and the average cross section of the test cavity, respectively. The temperature is measured by a bare thermocouple. Due to its thermal inertial, the measured temperature decreases slower than the gas transient. Instead of the instantaneous' temperature, the lowest temperature value is used for the computation. Since the temperature unit is in Kelvin, the discrepancy is not significant. The second way is to directly measure the cavity gas velocity with local Pitot tubes, and the density is modified with the cavity pressure and temperature. These probes are placed on the cavity bottom floor, and their heights are adjustable in order to obtain the velocity profile information in the vertical direction. In air-water tests, water jet velocity is measured by a Pitot tube in the liquid discharge line, and the water film velocity in the cavity is measured by conical type hot film probes. In woods metal tests, however, Pitot tubes can not be used for the jet velocity I NUREG/CR-6510 62 1_______ _ - 1 measurement due to the blockage caused by woods metal solidification. Instead of direct measurement, the mean liquid jet velocity is indirectly obtained from the total discharged liquid volume divided by the nozzle area and the discharge time duration measured with a conductivity probe near the discharge nozzle. The other way is to calculate the jet velocity using the vessel pressure as the total pressure and considering the friction loss along the discharge pipe, A 2P* V, = j , , (4-12) pf < K,, + K, + D> where Kin and Kom are the minor loss factors at the pipe inlet and outlet, respectively. An is the cross-sectional area of the nozzle. The pressure in the test cavity is neglected compared to the vessel pressure. The results show very good agreement for different measuring methods. In the later phase experiments, Eq. (4-12) is used due to its simplicity. In the test cavity, liquid metal film velocity is estimated from the cross-correlation of the film thickness signals, because woods metal flow may damage hot film probes. The principle of the method is simple. Assuming the film thickness signals from the i-th and j-th probes are h i (t) and h;(t), the cross-correlation function Rg(t) is defined as Ry (r) = 1 h,(t)h,(t + r)dt (4-13) This function is t dependent. The transient time, im, of a wave between the two probes can be thus determined as Ry(t) reaches maximum. The most probable wave velocity Vrw is given by V=L1, 3 (4-14) r, where Ly is the distance between the two probes. A more sophisticated way is to perform a Fourier transformation of Ry(t), and then to find the peak frequency and phase angle to obtain the most probable velocity. However, no significant difference was observed in the course of data processing. With this method, the time window is important for the liquid film velocity calculation. If the time window covers the wave front, the film front velocity is always the solution, which can be easily obtained by the time lag of the front signal at the two probes. For a time window that excludes the wave front, the most probable velocity is usually higher than the front speed, because the wave riding on the film flow moves faster. Although these two velocities are not the actual mean film velocity, they characterize the film flow and supply a reasonable estimation for the film motion on the test cavity floor. 63 NUREG/CR-6510 4.4. Pressure and Temocrature Measurements I Variable reluctance type pressure transducers ( Model DP, Validyne Eng. Corp. ) are used in the test to measure the system pressure transients, which include the tank pressure, vessel pressure, cavity pressure, subcompartment pressure, and the static and q total pressures from Pitot tubes at different locations (Fig. 4.1). All the pressure sense lines are filled with water before the experiment to prevent signal fluctuations that could result from compressible air in the fast transient process. The air tank pressure is also monitored with strain gages manufactured by Dwyer Instruments Inc., while the liquid and air temperatures are measured by K-type thermocouples exclusively. ] 4.5. Flow Visualization A video camera (Sony Model 7x710 HI 8 mm) with high shutter speeds and a Nikon camera are used to record the flow phenomena in the cavity, the subcompartment, and the containment / exhaust chamber during separate runs. A high speed movie film camera (500 frames /second) is also employed for flow visualization purpose. Because the test j cavity walls are transparent in the low pressure tests, the liquid jet and the liquid film flow in the cavity before gas blowdown can be observed and recorded with proper light arrangement. The liquid flow from the cavity exit to the subcompartment is recorded l through the visualization windows on the subcompartment walls. Inside the containment / exhaust chamber, the pressure condition is not severe, which allows the video camera being mounted there to observe the liquid droplet flow arriving from the seal table as well as from the subcompartment vents on either side of the seal table. These images give us vital information about the liquid transport transient, which is helpful in the later phase modeling attempts. 1 4.6. Test Control and Data Acanisition System i I A timer / control circuit is designed and built to control the test sequence. It sends out precisely delayed signals to actuate the gas line solenoid valves and the droplet sampling shutter right at the end of the liquid discharge. All signals from the sensors are acquired with a high speed multiplexer A/D board and an IBM compatible personal computer (DELL 486D/50). The A/D board (AT-MIO-16F-5) along with multiplexer board (AMUX-64T) is manufactured by National Instruments. This combination gives high speed sampling capability with 32 differential ended analog input channels. Signals from six pressure transducers, six film thickness probes, two hot film probes, four thermocouples, four conductivity probes, and three time controls are acquired in each test. The data are taken in binary format for fast sampling rate. After each test the data sets are converted to text format and then reduced to dimensional form. The speed of data acquisition for each channel is 125 data points per second, and it takes 10 seconds to cover the entire blowdown history. NUREG/CR-6510 64 l I. i
- 5. EXPERIMENTS AT 1.4 MPa VESSEL PRESSURE
. At 1.4 MPa test vessel pressure, standard tests are performed with 7 liters of water or woods metal discharged from a 3.5 cm nozzle. The gas blowdown is from a separate 10.16 cm pipe directly connected to the test cavity to supply properly scaled gas velocity. Detailed parametric tests are also carried out regarding the effects of the liquid inventory, i gas nozzle size, and the degree of subcompartment trapping. 5.1. Standard Air-water Tests The standard air-water test is conducted with 7 liters of water at 1.4 MPa vessel pressure. Before the initiation of the experiment, pressure sense lines are filled with water in order to avoid the delay of pressure response due to the compressibility of air in the high pressure transient. All film thickness probes are in-situ calibrated. Camcorders with proper lighting arrangements are set to record status at the selected locations. The three 0.6 m' air tanks are pressurized to the working pressure by the compressor. The test vessel is charged with 7 liters water, and connected to the air tanks. Right after data acquisition begins, the simulated accident scenario starts with the liquid jet discharge actuated by a solenoid valve in the discharge line under the test vessel. At the end of the liquid discharge, air blowdown from the solenoid valves in the separate gas line is automatically triggered by the timer / control circuit. Meanwhile, controlled by the same timer / control circuit, the shutter behind the isokinetic droplet sampling probe opens for 0.1 seconds to allow the entrained droplets being collected in the droplet sampling l system. After the completion of the blowdown process, the container with the droplet sample is removed from the collection box to analyze the droplet size and size , distribution using image processing. The liquid mass dispersed. into the ! containment / exhaust chamber is collected and weighed immediately after the test. Three categories of tests are performed under standard conditions. These are the calibration tests for the time sequence control and adjustment, the normal tests for liquid dispersion and transportation, and the special tests for flow visualization, droplet flux distribution at the cavity exit, and gas velocity profile in the test cavity. A number of ; t tests are carried out under standard conditions and the detailed average results of the flow I parameters are presented with the relative discrepancies less the 10%. The accuracy of the time sequence is within 1%. 5.1.1. Liquid Blowdown History and Flow Visualization The typical air-water blowdown transient is shown in Fig. 5.1, which presents the mass flow rate ofliquid and gas obtained from Pitot tubes in the discharge lines. The test starts with water discharge, and the duration of the water discharge process is about 0.6 65 NUREG/CR-6510 seconds. The corresponding average liquid jet velocity is estimated to be 13 m/s. After the completion of water discharge, air blowdown begins and is maintained for about 6 seconds. In the cavity, a 135 kPa static air pressure peak from the Pitot tube located 2.5 cm above the cavity floor is observed as shown in Fig. 5.2. This peak occurs when liquid film and droplets are experiencing maximum shear stress, and thus corresponds to the liquid entrainment process in the test cavity. In the subcompartment the pressure rise is very small, about 1.4 kPa above atmospheric pressure as shown in Fig. 5.2. Flow visualization shows that the waterjet discharged from the nozzle is continuous with little expansion (Fig. 5.3). Once the jet impinges on the cavity floor, it spreads in all directions and climbs on the cavity walls in the form of film as shown in Fig. 5.4. The water film flow eventually merges in the cavity chute section on the bottom floor, and rushes into the subcompartment (Fig. 5.5). From the film thickness and velocity measurements near the cavity exit, the fraction of discharged water that flows out of the cavity before the gas blowdown is estimated to be 15%. This part of water is not subjected to the entrainment process. Once the gas discharge starts, typical annular flow pattern is observed as water film spreads all over the cavity walls, and the water in the cavity is then partially entrained into droplets and partially pushed into the i subcompartment as a film flow. Eventually, water film and large drops from the cavity impinge on the seal table floor and the subcompartment side walls. Part of the impinged water goes through the seal table room to the upper containment / exhaust chamber (Fig. 5.6). Meanwhile, some tiny droplets in the subcompartment are carried by the gas stream i into the containment / exhaust chamber via the air vents as shown in Fig. 5.6. l 5.1.2. Gas Velocity and Water Film Velocity The average gas velocity in the cavity is calculated from the mass flow rate (Fig. 5.1) ) in the gas discharge line through continuity equation as described in section 4. The local gas velochy in the cavity is measured with a Pitot tube located 2.5 cm above the floor in the pedestal section. The comparison between the local and average gas velocities in the cavity pedestal section, shown in Fig. 5.7, indicates that the initial local gas velocity is I nearly 150 m/s, about twice the average value. This disagreement leads to a fair I conclusion that the cross-sectional gas velocity profile in the cavity pedestal section is not ) uniform especially in the early stage of the gas discharge. Further evidence is obtained via direct measurement of the velocity profile during separate runs with a Pitot tube of adjustable height placed on the cavity floor,67 cm from the liquid nozzle center line. As i shown in Fig. 5.8, this gas velocity profile (initial maximum value) exhibits a typical velocity distribution resulted from jet impinging on a perpendicular wall. In the cavity chute section, however, the flat velocity profile shows a typical turbulent channel flow profile (Fig. 5.9), because the location is far away from the jet center and the elbow of the cavity ernances the mixing. Therefore, the local gas velocity in the chute section agrees with the average value derived from the mass flow rate in the gas discharge line. Whereas, due to the narrower cross sectional area of the chute compared to the pedestal portion, the average velocity in the chute is high with a maximum value of 180 m/s at the NUREG/CR-6510 66 beginning. This implies that a more intense entrainment process occurs in the chute than in the pedestal section. The water film velocities are measured with hot film probes mounted 0.5 mm above the cavity floor (Fig. 5.10). It is found that the local water film velocity varies between 6 m/s and 8 m/s in the pedestal section, and 4 m/s and 5m/s in the chute section. Water film front propagation velocity is also calculated, as suggested in section 4, by means of cross correlating the signals from two film thickness probes located in the direction of the film flow. The averaged film front propagation velocity is estimated to be 5 m/s to 6 m/s. 5.1.3. Water Film Thickness in the Cavity Fig. 5.11 to 5.16 show the film thickness signals in the air-water test at various locations on the cavity floor. In Fig. 5.11, the liquid and gas discharge transients are also presented. From the measurements of CH1 and CH6 (Fig. 5.11 and 5.16), the film flow in the cavity lasts about 0.8 seconds from the start of water jet discharge. Once the air blowdown begins, the liquid film is totally removed from the cavity within 0.2 seconds. Before gas blowdown, the water jet impinges on the cavity floor, resulting in a circular film flow. Part of the water flows along the cavity floor towards the chute section, and the rest climbs on the side walls and to the ceiling of the cavity. The forward moving film flow spreads all over the walls in the pedestal section, and finally merges at the bottom floor of the chute section, forming a thick water film (Fig. 5.4). As the water film moves from CH1 to CH6 along the floor, the film thickness increases as summarized in Table 5.1. The time averaged film thickness before gas blowdown is 0.5 mm at CH1, 0.7 mm at CH2, and 1.8 mm at CH3. In the chute section, the water flow remains on the lower wall as shown in Fig. 5.4, and the film thickness ranges from 3 mm to 20 mm, with an average of ~ 6 mm at the cavity exit. The liquid film front propagation is obtained from the film thickness signal transients. Since the location of each film thickness probe is known, the liquid film front location can be traced as a function of time. Table 5.1 shows the film front locations at different time. Fig. 5.11 and Fig. 5.16 indicates that the water film flow reaches the cavity exit (CH6 placed 5 cm from the cavity exit) in ~ 0.40 1 seconds after the water jet initiation. Since the total water jet discharge time is 0.6 seconds, the time duration for water film to continuously rush out of the cavity exit before the gas blowdown is 0.2 seconds. This part of water will not be subjected to the latter dispersion process and it amounts to 15 % of the total discharged water volume estimated from the film thickness and velocity measured at the cavity exit. After gas blowdown, the two-phase flow in the cavity is in the annular flow regime, and the time-averaged film thicknesses at different probe locations are summarized in Table 5.2. In the chute section, the film thickness on the floor decreases from ~ 5 mm at CH4 to 1.6 mm at CH6 near the cavity exit. Within 0.2 seconds, the water is completely swept out of the test cavity, and this time duration (0.2 s) is thus referred to as entrainment time. 67 NUREG'CR-6510 5.1.4. Droplet Size Distribution Droplet samples are collected from both the low level probe and the high level probes (Fig. 4.1). The droplet size distribution measured by the high probes is useful in analyzing the dispersion of the droplets into the subcompartment while the low probe provides information of droplet entrainment in the cavity pedestal portion. To avoid excessive droplet sampling that may lead to droplet coalescence in the collection oil, three consecutive time windows are chosen to limit the sampling duration of each probe. The time window is controlled with a shutter installed at the inlet of the sampling probe, and the shutter is accurately driven by the timer / control circuit. From the moment of gas blowdown, these three time windows are divided in time frame of 0-0.11 seconds as window #1, 0.11-0.22 seconds as window #2, and 0.22-1.22 seconds as window #3. Since the entrainment time is estimated to be 0.20 seconds, the first two windows adequately cover the main entrainment process. The additional third window is designated longer to cover the end and post entrainment process. The measured volumetric fractional distributions of droplet sizes are presented in Fig. 5.17-5.20, where D. and D32 denote volume median diameter and Sauter mean diameter respectively and the dotted line represents the upper limit log-normal distribution function [28]. Measurements indicate that the mean droplet size at the cavity exit decreases from time window #1 to window #2 and #3. The corresponding volume median diameters are 566 m in time window #1, 406 m in time window #2, and 228 m in time window #3. The volume median diameters measured from the low probe are 630 m in time window #2 and 399 mm in window #3, which are much larger than that from the high probes. In time window #1, flow visualization shows that liquid film are collected. Therefore, accurate droplet size can not be measured from the low probe in time window #1. 5.1.5. Water Droplet Mass Flux The isokinetic sampling probes are also used to measure the water droplet mass flux. In each time window, the droplets that travel through the 0.953 cm diameter sampling probe tube are captured on a dry cotton cloth, which is then weighed immediately after the test to get the droplet mass. The droplet mass flux is calculated by dividing the collected droplet mass with the time duration and the cross sectional area of the probe tube. Mass 2 2 flux measured with HH probe in time windows #1 and #2 are 17.2 g/cm s and 20.8 g/cm s respectively. In time window #3 the droplet flux is so small that it is negligible (Fig. 5.21). 2 Hence the time averaged droplet mass flux at the cavity exit is 19.0 g/cm s, and the corresponding average entrainment rate per unit area of the cavity wall is estimated to be 1.75 g/cm's. Mass collection at different locations across the cavity exit is also carried out to study the spatial liquid mass flow distribution. In Fig. 5.22, the total mass collections at HH, HM, and HL probe locations during the entire blowdown time span are plotted. Flow visualization shows that water film flows out of the cavity exit along the walls and the HL probe captures not only the entrained droplets but also part of the film flow. Consequently less water mass is collected as the probe moves away from the bottom wall NUREG/CR-6510 68 to the center line of the chute. With the droplet mass flux measured from the HH probe representing the average value at the cavity exit, it is estimated that roughly 43% of the total discharged water is dispersed into droplets in the test cavity. 5.1.6. Liquid Carryover to the Containment / Exhaust Chamber After the test, the cavity is completely dry and all the discharged water is blown into the subcompartment and the upper containment. Of the 7.0 kg discharged water, only 193 g (2.8 %) liquid is carried into the containment / exhaust chamber through the seal table exit and the four air vents, while the rest 97 % is entrapped in the subcompartment. The ligidd from the seal table exit to the upper containment is mostly in film form (Fig. 5.6), which amounts to 60 g (or 0.9 % of the total water inventory). Nearly 1.9 % of the liquid mass in the upper containment comes from the four air vents in the form of small droplets l following the air stream. 20.0 i i i ~ Liquid Flow _ 16.0 f - 'i!b - x *ca 12.0 - 3 . Gas Flow k ihy.;/ $ 8.00 $ f 2 ce o i \ / AV * % '-*s '-' - 4.00 --s., 1 I ' ' ' 0.00 O.0 2.0 4.0 6.0 8.0 Time ( s ) Fig. 5.1 Water and air blowdown transients (1.4 MPa vessel pressure) 69 NUREG/CR-6510 l i 140 , , . . . I 130 - - to e . . x v 2 120 - Cavity a w w . / . Q. Subcompartment 110 - - - 7 . ._._._._._._._. ._._ _.7 , -._._. 0.0 2.0 4.0 6.0 8.0 Time ( s ) Fig. 5.2 Cavity pressure transient (air-water test,1.4 MPa vessel pressure) NUREG/CR-6510 70 1 I }w . , 3- E. :- f E R
- ua:.. :
t, v E p 1-E - ' a a) Liquid jet discharge ., Fig. 5.3 Flow visualization of waterjet (1.4 MPa vessel pressure) 71 NUREG/CR-6510 8k : 1 by FiIm flow in the cavity ) i Fig. 5.4 Flow visualization of water film spreading in the cavity (1.4MPa vessel pressure) NUREG/CR-6510 72 t l < e ".a . r! ,-= . ,; , / c) Film flow in the chute section Fig. 5.5 Flow visualization of water film flow in cavity chute (1.4 MPa vessel pressure) 73 NUREG/CR-6510 -) 'r .. .y,,, ,;;. . . ,:a. ic ,,, .. , . . ,,,,'.',,;.'o.- - . 1 i? '?' .t y' ;; :.de .c. u,,qi".'l;;,' - ,i..~ f ' .d. ! l th. . ' .f?h.:u(. .. ,w., J. , s_e.a l._ t.a.b.l. e. e._k i t . ,,, . .,,, , J ..z,.,; '. . . ,$__. sir vent f., ij.? +h: s . d) Liquid carry-over to the containment l I Fig. 5.6 Flow visualization of water carryover in containment (1.4 MPa vessel pressure) I i l l NUREG/CR-6510 74 200 . . . i ^ 160 4 Local E ' 7 g 120 - i f.j - -8 . Average . > :g. I' ji $ 80 b' fA (fM,I ( G 4 i 1. > "b't s, , il l ' ' f Y[gl' lY j jt. ~i j/g[ifliI i'g o 40 - Nv.- - .. . 1 0 ' ' ' O.0 2.0 4.0 6.0 8.0 Time ( s ) Fig. 5.7 Gas velocity in cavity (air-water test,1.4 MPa vessel pressure) 75 NUREG/CR-6510 i j i ! I I l 200 , i 2 ~. 160 ' E . =o 120 - - O s. 3 m $ 80 - cn - x a Experimental data . m o Curve fitting ' - o 40 - O 0 25 50 75 100 Distance from the lower wall ( mm ) Fig. 5.8 Gas velocity profile in cavity pedestal section (1.4 MPa vessel pressure) l NUREG/CR-6510 76 l u_- ! r ....,....,....,... ,...., l '200 ,, = ,c = , ,, m 3 160 - average local ~ b '@ 120 e m i & 80 -- - x tt! - m 1 40 - _ 1 0 '''''''''''''''''''' l 0 25 50 75 100 125 Distance from the lower wall ( mm ) Fig. 5.9 Gas velocity profile in cavity chute section (1.4 MPa vessel pressure) 77 NUREG/CR-6510 r i i i i i i i i i i 28.0 - DCH0107.DAT-24.0 - I . 20.0 - $ 16.0 - g x ' g 12.0 $ 8 .00 - robe 1 4.00 - Probe ' 2 '- ' ' ' ' ' ' 0.00 1.30 1.35 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80 Time (sec) Fig. 5.10 Water film velocity measured by hot film probes (1.4 MPa s .ssel pressure) i i 78 NUREG/CR-6510 L; 4.0 . , 200 , i , , 8 - 180 static P in 4* gas pipe ---- 8 dP in 2* water pipe *...... ,' - 160 i a 3.0 - film thicka*** ,' _ g4o s . e
- - 120 2
100 0-E y E 2.0 - ~ - 80 c. *E . - 60 f r - 40 1.0 - ,.f r* /s/VArN,
- v i.t.y
- ,;g 20
.a, _ ,, e
- i -
0 , i g.s . . . . . . , . . . i 0.0 W . i ' ~' 20 1.5 2.0 2.5 t (s) l l i Fig. 5.11 Water film thickness transient at CHI (1.4 MPa vessel pressure) l 1 i 79 NUREG/CR-6510 ; 4.0 , 3.0 - Y - g2.0 - 1 1.0 - f . [' L - 4 0.0 ,,,,, 1.5 2.0 2.5 t(s) Fig. 5.12 Water film thickness transient at CH2 (1.4 MPa vessel pressure) . NUREG/CR-6510 80 r 1 i i - 20 - e 16 - J --.- 12 ' E E- - x 8 - 4 _ 0 . l . . , 1 . . . . t . . 1.5 2.0 2.5 ) t(s) ] Fig. 5.13 Water film thickness transient at CH3 (1.4 MPa vessel pressure). 81 NUREG/CR-6510 20 - i 6 i - 16 - 12 - E g . . x 8 - - 4 - - 1 . f . . i . . k l . . . . I i e 1.5 2., 2.5-t(S) Fig. 5.14 Water film thickness transient at CH4 (1.4 MPa vessel pressure) NUREG/CR.6510 82 n l I 20 - 6 i i - 16 - - 12 -- - T g . . I a - - . I o . 4 _ - h f 0 WY , . , , . . I 1.5 2.0 2.5 t(s) Fig. 5.15 Water film thickness transient at CHS (1.4 MPa vessel pressure) 83 NUREG/CR-6510 i 7.. 1 l 20 - > n s 16 - - 12 - - 1 J E - 1 E - 1 1 - 8 - 4 - - , 0 , , , , , , i , , , , i . . 1.5 2.0 2.5 t(s) Fig. 5.16 Water film thickness transient at CH6 (1.4 MPa vessel pressure) i NUREG/CR-6510 84 Table 5.1 Film flow parameters before gas blowdown. Probe No. CH1 CH2 CH3 CH4 CH5 CH6 Distance tojet center (m) 0.285 0.592 0.808 1.008 1.454 1.843 Time-ave. film thickness (mm) 0.5 0.7 1.8 2.1 5.1 5.8 Film front propagating time (s) 0.145 0.210 0.240 0.255 0.330 0.43 Table 5.2 Time-averaged film thickness after gas blowdown Probe No. CHI CH2 CH3 CH4 CH5 CH6 Distance tojet center (m) 0.285 0.592 0.808 1.008 1.454 1.843 Time-ave. film thickness (mm) 0.8 1.4 2.1 5.1 3.8 1.6 85 NUREG/CR-6510 I l l div/dD i i i i 4 0.0016 - - Experimental Data ~ Upper Limit Log Normal Distributior) 3 0.0012 - ,f s Cimtacteristic Diameter p_ f Omax = 2500 pm $ / g Dvm = $66p m d I L' ( Du = 275p m f .O ' O.0008 s s - ' sarnpiing ilme : wiratow n i E i s 2 N Fligh Probe i 0.0004 s, - '.-_ l _j- _- , 0.0000 500 1000 1500 2000 2500 > Droplet Diameter ( p m) l Fig. 5.17 Water droplet size distribution at cavity exit in time window #1 NUREG/CR-6510 86 t 1' div/i1D ' -0.0020 - - Experimental Data q 'E I - - Upper. limit Log Normal Distribution O' O.0016 - s - .b If[\ '3 , Cliaraeteristic Diameter m \ I Dinax = Moo sun 5 0.0012 - I s \ Dvm = 406 pm o ' '3 e 1 \ D s2 = 231 pm 0.0008 - sampi ng Time : Windnw s 2 ' $ I Higli Probe 0.0004 -t s i . 's - J 0.0000 I ' ' ' 'la ' 1 i La t la t u t-500 1000 1500 2000 2500 l Droplet Diameter ( p m) Fig. 5.18 Water droplet size distribution at cavity exit in time window #2 87 NUREG/CR-6510 I I l l l l ) i i dl,/dD 0.0032 -f Experimental Dat a i Upper limit log. normal rfiseribution E 0.0028 - s =2 - \ .P_ 0.0024 - \ - Characteristic Diameters b g vs ' Dmax= 850 p en _ 5 0.0020 -, g Dvm = 228p m .2 r D as = 1543m $ 0.0016 5 E S*'"Pli"9 tim *: *i"**8 3 .2 0.0012 High Probe ~ l h 0.0008 ' hs 1 l 0.0004 . 0.0000 ' - - ' - - ' - L_ - - ' - - - ' O 500 1000 1500 2000 2500 Droplet Diameter (pm) ; Fig. 5.19 Water droplet size distribution at cavity exit in time window #3 NUREG/CR-6510 88 n. div/dD i - i ' ' 6parimentni Data ,h 0.0012 - - - upper.Lirnil Log Normal Distribution - '. 5 $ 'i Characteristic Diameter i y Dmax = 2300 pm Dvm = 630p m , *O 0.0008 . g I \ N D 32 = 367p m Sampling time : wirxiow # 2 0 'l -' Low Probe > 0.0004 - g 's, I 1, I , , ~ ~ ~l i 0.0000 ' ' ' ' ' ' _) 500 1000 1500 2000 2500 Droplet Diameter ( pm) ' Fig. 5.20 Water droplet size distribution near cavity floor in time window #2 89 NUREG/CR-6510 i t i; I l l 30 - - High Probe (HH) window #1 . ) window #2 l ~ / E 20 -- o , v) I 'cn x v . x D
- =:
g 10 - window #3 ca 2 0 0.00 0.20 0.40 0.60 0.80 1.00 1.20 Time ( s ) Fig. 5.21 Water mass collections in different time window NUREG/CR-6510 ; 90 L _. 10.0 , , ~ , Io 8.0 - - y - . w - . C o 'g 6.0 - - .O - . *c - . 5 . . m 4.0 - - m - . g 2 - . 9 - . @. 2.0 - - i - - 0.0 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' O 5 10 15 20 Lower Wall Upper Wall Distance From Wall (cm) l l Fig. 5.22 Water droplet flux distribution at cavity exit i 91 NUREG/CR-6510 f 5.2. Standard Air-Woods Metal Test The standard air-woods metal test is conducted with 7 liters of molten woods metal at 1.4 MPa vessel pressure. Before the initiation of the experiment, pressure sense lines are filled with water in order to avoid the delay of pressure response due to the 3 compressibility of air in the high pressure transient. The three 0.6 m air tanks are i pressurized to the working pressure by the compressor. Solid woods metal pieces are ) melted in a melting pot located above the test vessel. The liquid metal is then charged i into the vessel by gravity force, maintained in liquid state by the electrical belt heaters j around the vessel. The solenoid valve in the liquid discharge line is replaced with a rupture disc, and the solenoid valve is relocated upstream of the pressure vessel to initiate the test. Before the test, the test vessel and the liquid discharge line are heated up to 95 C to prevent liquid metal from solidifying in the discharge system. Afterwards the pressure vessel is pressurized to 1.04 MPa to avoid gas punch-through that may lead to i distorted two-phase discharge. Once the solenoid valve located upstream the vessel is ) open, air comes from the 1.4 MPa air tanks to build up the test vessel pressure and ) I eventually to initiate the experiment by breaking the rupture disc. At the end of the liquid discharge, similar to air-water tests, solenoid valves in the gas discharge line are triggered l by the timer / control circuit to actuate the gas blowdown. Meanwhile, controlled by the I same timer / control circui', the shutter behind the isokinetic droplet sampling probe opens ) for 0.25 seconds to allow the entrained droplets being collected in the droplet sampling j system. After the completion of the blowdown process, the container with the droplet j sample is removed from the collection box to analyze the droplet size and size j distribution using image processing. The liquid mass dispersed into the l containment / exhaust chamber is collected and weighed immediately after the test. l Compared with air-water tests, woods metal tests are much more complicated and tedious. It is difficult to cany out such detailed tests as in air-water case. Difficulties are also encountered in flow visualization due to the opaque liquid film spreading all over the cavity walls. Nevertheless, the following important characteristics on liquid dispersion are obtained. 5.2.1. Blowdown History Woods metal jet velocity is determined by its discharge time duration (Fig. 5.23), which is measured with the signal span from the conductivity probe installed right below the liquid discharge nozzle (Fig. 2.1). The woods metal discharge time is measured to be 0.68 seconds and the average woods metal jet velocity is then calculated to be 11 m/s. Cavity pressurization is observed after gas blowdown with a pressure peak of 160 kPa as shown in Fig. 5.24. This pressure is higher than that in air-water tests (135 kPa) due to the larger density of the woods metal droplet that requires higher gas momentum to accelerate the liquid or droplets in the test cavity. Flow visualization at the cavity exit shows that almost no woods metal rushes out of the cavity before gas blowdown. NUREG/CR-6510 92 l l 5.2.2. Gas and Liquid Velocity In the gas discharge line, similar gas velocity transient as in the air-water tests is observed. This is because the gas flow is choked at the discharge nozzle and is solely ; determined by the air tank pressure. Thus the cavity average gas velocity is obtained in a way similar to the method used in air-water tests. However, the measurement oflocal gas velocity in the cavity is unsuccessful due to the blockage of the Pitot tube total pressure sense line by woods metal droplets. Fortunately the static pressure measurement is valid as shown in Fig. 5.25. Using this static pressure along with the total gas flow rate measured in the gas discharge line, the cavity average gas velocity is calculated (Fig. 5.26), and the maximum average velocity in the pedestal section in the early transient is about 75 m/s. The liquid film front velocity in the cavity is determined by the cross 1 correlation of the film thickness signals. The resultant woods metal film front velocity is about 6.5 m/s in the cavity pedestal portion and decreases to roughly 1 m/s to 3 m/s ir the l cavity chute section. After the initiation ofliquid discharge, the woods metal film front takes about 0.52 seconds to reach the cavity exit. However, it is observed that no woods metal flows out of the cavity before gas blowdown. 5.2.3. Film Thickness in the Cavity The capacitance film thickness probe gives meaningful readings before gas blowdown. After gas discharge, the 5 watts sheet heater sandwiched in the probe can not compensate for the tremendous heat loss to the gas stream. The probe then fails as a result of woods metal freezing on the electrode, showing a constant voltage output. Typical woods metal film thickness signals at CH2 and CH6 are shown in Fig. 5.27. ; Before gas blowdown, the film thickness ranges from 0.5 mm to 2.0 mm on the cavity pedestal floor. Near the cavity exit, the film thickness increases to about 20 mm. 5.2.4. Droplet Size, Size Distribution, rad Mass Flux In air-woods metal test, woods metal droplets quickly freeze inside the collection liquid and the problem of droplet coalescence in the collecting box is not a concern as discussed in section 4. Therefore, a single sampling time window is selected to cover the initial 0.25 seconds after gas blowdown. The size distribution of woods metal droplets collected at the cavity exit through the HH probe is shown in Fig. 5.28. The volume median diameter is measured to be 1.1 mm, which is more than twice the water droplet size. For droplet mass flux measurement, a time window duration of 0.25 seconds is used to adequately cover the entire entrainment process. The time averaged droplet mass flux 2 at HH is 44.1 g/cm s, and the corresponding entrainment rate per unit cavity wall area is 2 4.05 g/cm s. About 11% of the discharged woods metal is entrained into droplets in the test cavity. 93 NUREG/CR-6510 5.2.5. Liquid Carryover to the Upper Containment l l Only 1.6 % of the total 54 kg (7 liters) discharged woods metal is dispersed into the contamment/ exhaust chamber and ahnost all of it is discharged from the seal table exit. I Only extremely fine droplets are carried into the containment / exhaust chamber by the air stream via the four air vents. The size distribution of the droplets from the air vents is presented in Fig. 5.29, which shows that the volume median diameter is around 60 m. Nearly 24 % of the total mass is recovered in the seal table room,67 % is estimated to be trapped inside the subcompartment, and 7.4 % freezes on the walls of the test cavity. 1 800 i , , 6.0 signal from conductive probe m = 5.0 g v 600 - ^ - 3 [ 4.0 3 x 8 400 - - 3.0 pressure in gas discharge line Q. 2.0 j 200 - 5 -3 1.0 O O L y 1 9 0.0 2.5 5.0 7.5 10.0 Time ( s ) Fig. 5.23 Air-woods metal blowdown history (1.4 MPa vessel pressure) l l l NUREG/CR-6510 94 3-- 180 , , , , cavity static pressure 160 - cc CL . x . v 8 140 - 8 m -D . Q. 120 - 100 ' ' ' ' l 0.0- 2.0 4.0 6.0 8.0 10.0 l Time ( s ) l Fig. 5.24 Cavity pressure transient (1.4 MPa vessel pressure, WM test) i 95 NUREG/CR-6510 \ 200 . . , ,, 2 160 - - 1 . . .h120 - f l 2 h[b ' l { 80 - 1 J - .N - $ 40 - W 0 O.0 2.0 4.0 6.0 8.0 I Time ( s ) Fig. 5.25 Average gas velocity transient (1.4 MPa vessel pressure, WM test) NUREG/CR-6510 96 F i: , l 3 i j jet front jet end i t - so l J 5 . n 2 - . I l L t \ : n E ! 1 1 'o B E - 1, l v 5 v h i . V .c ; : 1 - 1 i j. $s . 10
- l
. CH2 i CH6 : j { 0 0 1 1.2 1.4 1.6 1.8 2 t (s) Fig. 5.26 Woods metal film thickness signals at CHI and Ch6 (1.5 MPa vessel pressure) 97 NUREG/CR-6510 I 1 l 9.0E-4 , E experimental measurement L c 7.5E-4 - --- upper-limit log-normal distribution g 'O ... - 6.0E 4 - / y D3 ,= 820 m .g J,/- D = 1120 m-Y ' 4.5E-4 - , ', , vm .o . .c , D ,,, = 3.5 mm g ' 5 3.0E-4 - ' N.,l _
- ~~. .
o ; ~ c , 1.5E l ..,,_ -E r ... _3_ - ! .. o l > 0.0 E0 O 1000 2000 3000 Droplet Diameter ( m) 1 Fig. 5.27 Woods metal droplet size distribution at cavity exit (1.4 MPa vessel pressure) l 98 NUREG/CR-6510 I 4 I j i 4 6 E' ' l C 0.016 . i O " charateristic diameters _ 0.012 - 8 D 32 = 46 m . D vm = 60 m' - l } D = 120 m - ',E 0.008 - max- 1 g 0 - 0 E 0.004 e E a - o -- ' ' ' ' ' ' ' > 0.000 O 30 60 90 120 150 180 Droplet Diameter ( m) Fig. 5.28 Woods metai dropht size distribution in containment (1.4 MPa vessel pressure) 1 99 NUREG/CR-6510 i 5.3. Exoerimental Results of Parametric Studies j 1 Parametric studw. are carried out for air-water tests under 1.4 MPa system pressure. The liquid dispersion phenomena are systematically investigated with different gas discharge nozzle sizes, water inventories, and subcompartment trapping conditions. These experimental results are valuable for analytical studies. t 5.3.1 Gas Discharge Nozzle Size Effect Four different gas discharge nozzle sizes are used in air-water experiments. With a fixed air-tank pressure (1.4 MPa), the change of gas nozzle sizes simulates different cavity gas velocities. Since the water discharge nozzle size of 3.5 cm is unchanged, the water film flow in the test cavity before gas blowdown is not affected by gas nozzle sizes. The parameters affected by cavity gas velocity are summarized in Table 5.3. As the gas nozzle size increases, both the cavity gas velocity Ve and static pressure Pc increase as expected. The entrainment time te, and the water film velocity Vr in the cavity chute section are not affected by gas nozzle sizes after gas blowdown. Here the water film velocity is estimated from the cross correlation between the film thickness signals of CH5 and CH6, because the hot film measurement gives large fluctuations after gas blowdown. These results indicate that the liquid film motion remains the same for different gas nozzle sizes. The volume median droplet diameters listed in Tab. 5.3 are obtained from the HH isokinetic sampling probe at the cavity exit in time window #2. It shows that droplets are smaller with faster cavity gas speed. According to the droplet flux measurement, more droplets are entrained in the test cavity as the gas discharge nozzle ( size increases (Fig. 5.29). With 4.95 cm gas nozzle, the fraction of entrainment in the l cavity is only 30%. Similar trend is observed for the fraction of water dispersed into the containment / exhaust chamber as shown in Fig. 5.30. As the gas nozzle size is reduced to 4.95 cm, the total percentage of the liquid carry-over drops to 2%. Nevertheless, the I liquid carry-over through the seal table room remains constant independent of the gas nozzle sizes in the studied range (Fig. 5.30). This implies that this fraction of liquid is mainly determined by the view factor between the cavity exit and the seal table opening. 5.3.2 Liquid Inventory Effects In the standard air-watcr test, 7 liters of water is used to simulate the core melt amount equal to 40% of the total core material. To investigate the liquid inventory effects on the corium dispersion phenomenon, air-water tests are undertaken with 4 and 11 liters of water. Table 5.4 summarizes the effects of water inventory on the measured i parameters. Before gas blowdown, the time needed for the liquid film front to reach the cavity exit from the jet center is almost constant (Tr ~ 0.4 s) in spite of water inventory changes. This is because the initial water momentum is solely determined by the system pressure rather than the water inventory. The water film velocity is faster with less water inventory due to the shorter water discharge time and therefore less liquid mixing in the NUREG/CR4510 100 test cavity. Moreover, greater water film thickness near the cavity exit is observed with respect to larger water inventory. For the case of 11 liter water inventory, nearly 30% of the discharged water flows out of the cavity before gas blowdown, while about 15% of water moves out of cavity for both the 4 and 7 liter water inventory cases. These results suggest that for the two smaller water inventory cases the secondary jet mixing effect in the cavity is less important for the film flow at the cavity exit before gas blowdown. After gas blowdown, The entrainment time duration (T.) is longer as the discharged water volume increases, but the liquid film velocity in the cavity chute section is unchanged. The volume median diameter of the droplets collected at the cavity exit 1 through HH probe in time window #2 increases slightly with larger water inventory due to the higher Reynolds number of the liquid film flow. More water is entrained into droplets in the cavity with respect to larger discharged water volume as shown in Fig. 5.31. Thus it is concluded that the effect of the liquid inventory on the entrainment rate is l much more significant compared to that on droplet size. However, the percentage of liquid carryover to the containment does not change with different volume of water as , / shown in Fig. 5.32. 5.3.3. Subcompartment Trapping The effect of subcompartment structures on liquid trapping is studied in air-water tests. A deflector plate, made of Lexon sheet (27 cm x 78 cm x 1.27 cm), is installed in the subcompartment above the cavity exit with the same height as the seal table floor and right in the path of the liquid flow (Fig. 5.33). The subcompartment structures that simulate steam generators, circulation pumps, and the seal table room are removed for ' these tests. The angle of the deflector plate is adjustable, and the liquid dispersion into the containment is measured at three different deflector angles, i.e.,0*,45 , and 90 . In Fig. 5.34, without the deflector and other obstructions, the fraction of the liquid dispersed into the containment / exhaust chamber is as much as 10%. This is a substantial j increase compared with that when steam generators, circulation pumps and seal table room are in place. With 0* deflector angle, the amount of dispersion is about 2%. This i fraction mainly comes from small droplets carried by the gas stream via the four air vents and matches the results in the standard test case. At 45 and 90* deflector angles, the dispersion fractions increase to 6% and 13% respectively. These results indicated that the trajectory dependent mechanism for droplet transport is important to determine subcompartment trapping. t 101 NUREG/CR-6510 l l Table 5.3 Parameters measured after gas blowdown with different gas nozzle sizes D,(cm) V,(m/s) P (KPa) Vr(m/s) t,(s) D. ( m) D32 (pm) 4.95 65 113 17.9 0.21 1200 660 6.06 72 117 19.2 0.20 810 470 7.83 80 123 18.9 0.20 540 260 10.16 90 134 19.4 0.20 406 231 D,: Gas discharge nozzle size, P,: Maximum cavity static pressure, Vr: Average water film velocity in the t: Entrainment time, l cavity chute section, D.: Volume median droplet diameter at cavity exit, i V,: Maximum area-averaged gas D:32Sauter mean droplet diameter at cavity exit. I velocity in the cavity ped:stal ( section, I Table 5.4 Liquid inventory effects on measured parameters water before gas blowdown after gas blowdown inventory (liter) tr(s) Vr(m/s) Fr hem., (mm) t,(s) Vr(m/s) D,(pm)
- 11. 0.42 3.1 30 % 4.71 0.25 19.2 420 7.0 0.40 4.0 15 % 4.68 0.20 19.4 406 4.0 0.40 6.5 15 % 1.95 0.15 18.8 402 Tr: Time for film to reach cavity exit Fr: Fraction of water out of cavity before gas blowdown, fromjet center, The other notations are the same as in Table 3.
Vr: Film velocity in cavity chute, NUREG/CR-6510 102 p: I i 0.50 , , .= ' $ 0.40 - .5 15 e - E .5 $C 0.30 r - o C 1 o ' '= i O } 'O.20 - 5.0 6.0 7.0 8.0 9.0 10.0 11.0 Gas nozzle diameter ( cm ) . Fig. 5.29 Effects of gas nozzle size on entrainment in cavity (1.4 MPa vessel pressure) 103 NUREG/CR-6510 ) i ) i I ) ....i ... .... .... ....,,...,.. , I 3.0 -
- 5. total
{ O . i b. u m o as 2.0 - ._5_r o O CD from seal table room h 1.0 - 9 g _ 2 8 a
- c. .
' ' ' t i i e . 0.0 ' 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 Gas nozzle diameter ( cm ) Fig. 5.30 Effects of gas nozzle size on carryover (1.4MPa vessel pressure) 4 i NUREG/CR-6510 - 104 - l m. O - .C. 0.40 - C .9. o - 2 t , i C .m 5 0.30 - aw ~ C w 0.20 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 Water inventory ( liter ) i Fig. 5.31 Effects of water inventory on entranunent in cavity (1.4 MPa vessel pressure) 105 NUREG/CR-6510 p j '.i I I 3.0 -- $ total y. u 5 - 5a 2.0 - ._57 . y O CD $c from seal table room 1.0 m - -v - 8 + B . O-0.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 4.0 . 5.0 Water inventory ( liter ) Fig. 5.32 Effects of water inventory on carryover (1.4 MPa vessel pressure) l I I l NUREG/CR-6510 .106 l 1 \. D , r i defleetor _-..) subeompartment cavity exit i Fig. 5.33 Schematic of the deflector in subcompartment trapping tests I i . i 14 - i ,i 12 E - O h10 _ . dispersion without deflector u v' - '5 8 - E D 6 - E , a 1E 4 ,. ' 2'- dispersion with deflector - 0-O 25 50 75 , Deflector angle ( degree ) Fig. 5.34 Effects of deflector angle on water carryover (1.4 MPa vessel pressure) 107 NUREG/CR-6510 l
- 6. HIGH PRESSURE EXPERIMENTS The high pressure experiments are performed with 7 liters of water or molten woods metal to simulate 40% of the core being melted. For woods metal tests, the vessel as well as the discharge line needs to be heated up to 95 C to keep the woods metal in liquid '
state. The system is pressurized with 42 MPa commercial nitrogen tanks in two steps. For the case of 14.2 MPa water tests, for instance, the gas tank and the test vessel are connected together with the bypass line as shown in Fig. 3.9, and they are initially charged to 11.8 MPa. Aftenvards, the test vessel is isolated from the gas tank by closing the bypass valve, meanwhile the charging process continues until the gas tank pressure reaches to the 14.2 MPa. The burst pressures for the rupture discs at the gas tank top and the vessel bottom are 3.5 MPa (disc #1) and 12.5 MPa (disc #2) respectively. With this kind of anangement, the system pressure combination can be maintained without breaking either one of the rupture discs prematurely. By opening the 0.64 cm relief valve linked to the test vessel, the experiment and data acquisition are initiated. About 2 seconds later, the vessel pressure drops to 9.0 MPa, and rupture disc #1 breaks because the pressure difference between the gas tank and the vessel reaches the disc burst limit (3.5 MPa). Instantaneously, the pressurized gas from the gas tank pushes into the test vessel via a 7.65 cm pipe and breaks the second rupture disc to start the blowdown transient. Similar procedures are applied to the 6.9 MPa water tests and the 14.2 MPa woods metal tests. However, appropriate rupture discs have to be selected with respect to the pressure conditions. In Table 6.1, the initial tank and vessel pressures are ! summarized along with the relevant rupture disc selections. Due to the complexity and l lengthy preparation, only two or three tests are performed under each condition to ensure consistency and repeatability, which are not sufficient for statistic error analysis. Therefore, the following sections cover only the representative test results. Table. 6.1 Rupture disc selections and the initial pressure conditions in the gas tank and the test vessel 4 Tests Tank P Disc Burst Pressure #1 Vessel Pressure Dise Burst Pressure #2 (MPa) (MPa) (MPa) (MPa) 6.9 3.5 4.8 6.3 6.9 MPa water test 10.4 5.5 6.3 9.0 10.4 MPa test 14.2 3.5 11.8 12.4 14.2 MPa water test 14.2 9.0 6.9 10.4 14 2 MPa woods metal test NUREG/CR-6510 108 7 6.1. Water Tests at 6.9 MPa Vessel Pressure In the 6.9 MPa water test,7 liters of water is employed to simulate the core melt and the vessel break size is 3.5 cm in diameter. This test is similar to the 1.4 MPa pressure water test because gas velocity in the cavity for the low pressure test is compensated by an auxiliary 10.16 cm gas line. The major difference would be the liquid discharge velocity. It is expected that the liquid film flow in the cavity at 6.9 MPa vessel pressure should be much faster and the entrainment duration is supposed to be shorter. The test . data is sampled in a 10 seconds time interval. Fig. 6.1 shows the vessel pressure transient during the blowdown process. The test . starts with the depressurization of the test vessel by opening the 0.65 cm relief valve as ] shown in Fig. 3.13. Within about 2 seconds, the vessel pressure drops to 3.4 MPa and rupture disc #1 breaks due to the pressure difference between the gas tank and the test vessel exceeds the burst limit of the rupture disc. A sharp rise of the vessel pressure is then observed in Fig. 6.1. Afterwards, the second disc at the vessel bottom ruptures instantaneously to initiate the blowdown process. The liquid and gas mass flow rate transients are plotted in Fig. 6.2, which are obtained directly from the pressure measurements in the vessel and the discharge line. The water discharge duration is estimated to be 0.048 seconds till the Pitot tube in the cavity senses gas pressure as shown in Fig. 6.3. The conesponding average waterjet velocity is calculated to be about 75 m/s. Inside the cavity, the pressure measured 5 cm above the bottom floor shows a 150 kPa pressure peak hs shown in Fig. 6.3. The area averaged cavity gas velocity (based on the average cavity hydraulic diameter, Dn=0.25 m) is obtained from the total mass flow rate and the cavity pressure. A comparison of the local and the average cavity gas velocities is made in Fig. 6.4. Similar to the previous low pressure test results, the local gas velocity is much higher due to the non uniform gas velocity profile in the cavity pedestal section. Typical water film thickness in the test cavity are plotted in Fig. 6.5, where CH2 and CH6 refer to the conductivity film thickness probes located at 59 cm from the jet center and at the cavity exit, respectively. It is found that all the discharged water is completely i swept out of the cavity within 0.202 seconds after gas blowdown. This time interval is ; thus referred to as entrainment time. Before gas blowdown, the film thickness in the cavity pedestal section is about 0.5 mm in average with a maximum of 1 mm. While the film thickness at the cavity exit is much larger, about 1.1 mm in average with a maximum of 6 mm. From the cross correlation of these film thickness signals, the average water film velocity in the test cavity before gas blowdown is estimated to be 50 m/s. The CH6 signal also indicates that roughly 0.032 seconds is needed for the liquid film front to i I reach to the cavity exit from the beginning of the water jet discharge. Since the jet discharge lasts 0.048 seconds, the duration of water film flow at the cavity exit before gas l blow-down is only about 0.016 seconds. During this time interval, the amount of water l that flows out of the cavity is estimated to be 12.6% of the total discharged liquid. This part ofliquid is not subjected to the dispersion process during gas blowdown. 109 NUREG/CR-6510 1 The water droplets entrained in the test cavity are sampled with a 0.95 cm isokinetic probe tube placed at the center of the cavity exit. The measured droplet size distribution is presented in Fig. 6.6, where dmx, dm, and d32 denote maximum droplet size, volume median diameter, and Sauter mean diameter, respectively, while the dashed line represents the upper-limit Log-normal distribution curve-fitting [28]. The results indicate that the volume median diameter of the droplets entrained in the cavity is 184 m, about half the size measured in the low pressure tests. The isokinetic sampling probe is also . l used to measure the water droplet mass flux. In such tests, the droplets that travels through the probe tube are captured on a piece of dry cotton cloth, which is then weighed immediately after the experiment to get the total droplet mass. A droplet mass flux of 20 i g/cm2s is then obtained by dividing the sample mass with the probe cross sectional area and the entrainment time. With this value representing the average droplet mass flux at the cavity exit, the entrainment rate per unit cavity wall area is estimated to be 1.84 g/cm2 s, and the total droplet mass entrained in the test cavity is roughly 33% of the discharged water. Again, this value is much smaller than that from the low pressure test (48%). The reason may be due to the much faster film flow speed in the cavity, which implies shorter entrainment duration and less entrainment fraction in the test cavity. After the test, all the 7 kg of discharged water is blown into the subcompartment and the upper containment chamber. However, only 2.4% of it is dispersed into the containment chamber via the seal table exit and the two front air vents. The rest 97.6% of water is trapped in the subcompartment. The side walls of the other two back air vents are dry and clean, and therefore the amount of droplets from the back air vents is negligibly small. Compared to the carryover in the 1.4 MPa tests,2.7%, the discrepancy is insignificant. This agreement suggests that the liquid film velocity in the cavity is not important in determining the total carryover though its influence on the droplet size and entrainment fraction in the test cavity is quite remarkable. It may be therefore safe to conclude that the carryover is mainly due to the re-entrainment process occurring in the subcompartment since it is almost independent of the dispersion parameters in the cavity. 6.2. Water Tests at 14.2 MPa Vessel Pressure The 14.2 MPa water test is performed with 7 liters of water discharged from a 3.5 cm nozzle. The main purpose of the test is to investigate the liquid dispersion phenomenon under the full prototypical reactor vessel pressure. A high liquid carryover is expected because of the much stronger gas flow that may result in substantial re-entrainment in the subcompartment. Typical blowdown history of the 14.2 MPa water experiments is shown via the vessel pressure transient in Fig. 6.7. Similar to the 6.9 MPa water tests, the two rupture discs break almost simultaneously to initiate water discharge. From the mass flow rate transient shown in Fig. 6.8, it is found that the water discharge is faster than that in the case of 6.9 MPa tests. The waterjet lasts only 0.036 seconds with an average speed of 94 m/s. Different from 6.9 MPa tests, the cavity pressure increases significantly up to 200 kPa after gas blowdown (Fig. 6.9). It suggests that the gas flow maybe choked at the cavity exit. The corresponding average cavity gas velocity shown in Fig. 6.10 reaches NUREO/CR-6510 110 I 132 m/s initially and drops to 55 m/s quickly as the cavity being pressurized. Because of the short entrainment duration, about 0.188 s, which will be discussed later, the peak gas velocity can be considered responsible for the entrainment process. This observation may greatly simplify the theoretical modeling of the corium dispersion phenomenon in the reactor cavity regardless of the latter phase choking or pressurization of the cavity. The water film thickness signals in Fig. 6.11 indicate that the average film thickness at CH2 and CH6 before gas blow-down are 0.6 mm and 1.2 mm respectively with an average film velocity of 83 m/s. The time interval for the water film flows out of the i cavity before gas blowdown is roughly 0.012 second, and the entrainment time is thus ) estimated to be 0.188 seconds. From these results the amount of water that flows out of j the test cavity before the dispersion process is estimated to be 17.1% of the total ! discharged liquid. In Fig. 6.12, the droplet size distribution at the cavity exit is presented where the volume median diameter is 587 m. About 26% of the discharged water is entrained into droplets in the test cavity, and the droplet entrainment rate per unit cavity 2 wall area is around 2.6 g/cm s. Similar to the 6.9 MPa tests, most of the liquid is trapped in the subcompartment. Only 5.3% of the 7.0 kg water is recovered in the upper ! containment. The wet area in the upper containment surrounds the seal table exit and the two front air-vents, and the two back air-vents are dry and clean. Most of the liquid carryover comes from the front air-vents and the seal table exit. These results also suggest that the liquid film transport is faster as the vessel pressure increases from 6.9 MPa to 14.2 MPa. Consequently the entrainment time becomes shorter, which reduces the fraction ofliquid dispersion. On the other hand, as a result of vessel pressure rise, the gas velocity increases, which in turn will enhar'ce the entrainment process. These two competing factors determine the degree of the whole liquid dispersion transient in the cavity. In general, however, more liquid is dispersed into the containment in the form of smaller droplets as the test vessel pressure increases, in spite of the fact that few droplets are entrained in the test cavity due to the shorter entrainment time duration. Similar to the conclusion drawn from the 6.9 MPa water test, it suggests that the re-entrainment in the subcompartment may be the main factor for the liquid carryover to the upper contamment. 6.3. Woods Metal Tests at 14.2 MPa Vessel Pressure Due to its complexity, only two tests are carried out at 14.2 MPa vessel pressure with a 3.5 cm discharge nozzle and 56 Kg woods metal. Before the test, woods metal is melted in an electrical heating pot located outside the main facility. Through a 2.54 cm pipe as shown in Fig. 3.13, the molten woods metal is then discharged into the preheated vessel that is insulated with fiber glass. To ensure the initiation of the test, the burst pressure of the rupture disc #2 is 10.4 MPa, about 3.8 MPa less than the working pressure. In Fig. 6.13, the gas tank pressure instead of the vessel pressure is presented to demonstrate the test transient, because the pressure sensor can not stand the heating imposed to the test vessel. The liquid discharge takes roughly 0.11 seconds as shown in 111 NUREG/CR-6510 Fig. 6.14, much slower than the water jet velocity under the same vessel pressure condition because of the large density difference. The corresponding jet velocity is 31 m/s. After the gas blowdown, the cavity static pressure increases to a maximum of 250 kPa as shown in Fig. 6.15 , which implies that the gas flow is also choked at the cavity exit similar to the water tests. The maximum initial average cavity gas velocity is 250 m/s (Fig. 6.16). In Fig. 6.17, the woods metal film thickness on the cavity floor is shown for CH2 and CH6 probes. It should be mentioned that the measurements are effective only before gas blowdown begins. For rapid gas flow, the thin heater sandwiched inside the probe plate can not compensate the tremendous heat loss to the fast gas stream. Hence, woods metal film freezes on the probe surface giving a constant output signal as shown in Fig. 6.17. Before gas discharge, however, the signals indicate that the maximum film thickness in the cavity pedestal section is about 2.3 mm. At the cavity exit, the film thickness increases up to 4.2 mm with an average of 1.4 mm, lasting roughly 0.044 seconds. With an average film velocity of 26 m/s, about 22% of the discharged metal rushes out of the cavity before entrainment occurs. Similar to the low pressure tests, after the gas blowdown, the liquid flow in the cavity is an annular mist flow. The woods metal freezes on the savity walls, forming a thin crust that amounts to 8% of the total discharged liquid metal. Through the 0.95 cm isokinetic probe, the droplet samples are collected with a pool of syrup heated up close to the woods metal melting temperature to prevent the hot droplet from disintegrating due to sharp temperature change. After the syrup solution cools down, the sample woods metal particles are separated out for analysis with the image processing system. Most'of the particles are irregular in shape, which are approximated to elliptical geometry in the size analysis algorithm. The size distribution shown in Fig. 6.18 indicates that the volume median diameter is 857 m, about five times larger than the water droplets under the same pressure conditions. By further weighing these droplet samples, it is estimated that nearly 38% of the discharged woods metal is entrained into droplets in the test cavity. In the upper containment,1.9 kg of woods metal (3.4% of the total discharge mass) is recovered after the test.' From the air vents, the carryover is mostly in the form of fine droplets with a volume median diameter of 60 m. Some small pieces of woods metal are found around the exit of the seal table room. This part of woods metal amounts to roughly 0.9%. In the subcompartment, a large fraction of the discharged woods metal is recovered around and in the seal table. The liquid metal I spreads around the subcompartment side walls, forming a halfinch thick crust. Inside the seal table, the solidified woods metal is found to be as thick as one inch in the corner areas. On the models of the steam generators and the pressurizers, small amount of woods metal is found due to the deposition of the dispersed fine droplets. Same observation is true for the two front air-vents. However, the side walls of the two back air-vents are clean, which suggests that the re-entrained droplet in the subcompartment enters the upper containment mainly through the front air-vents. NUREG/CR-6510 112 6.4. Woods Metal Test at 10.4 MPa Vessel Pressure One test is conducted with woods metal as the simulant under a vessel pressure of 10.4 MPa for a nozzle size of 3.5 cm. The burst pressure of the rupture disc #1, as shown in Table 6.1, is 5.4 MPa, which is about 4.8 MPa less than the working pressure. The breaking pressure of rupture disc #1 is 9.0 MPa and the vessel is initially charged to 6.3 MPa. By opening the relief valve connected to the test vessel as shown in Fig. 3.13, the experiment starts. As the vessel pressure decreases to se degree that the pressure difference between the gas tank and the vessel exceeds the burst limit of rupture disc #1, this. disc breaks and the high pressure gas in the gas tank rushes to the test vessel, breaking rupture disc #2 to initiate the blowdown transient. In Fig. 6.19, the gas tank pressure is presented to demonstrate the test transient. The liquid discharge takes roughly 0.24 seconds as shown in Fig. 6.20. The correspondingjet velocity is 28 m/s. After gas blowdown, the cavity static pressure increases to a maximum of 150 kPa as shown in Fig. 6.21, which shows large fluctuations in the early phase of the blowdown that probably resulted from the blockage at the probe tip with the solidified woods metal. The aversge maximum cavity gas velocity is 76 m/s (Fig. 6.22), which is obtained from the mass balance equation with the gas density adjusted to the measured cavity pressure. The droplets entrained in the test cavity are sampled with the isokinetic sampling probe directed to a pool of syrup solution that is heated up close to the woods metal melting temperature (75 C). After the solution cools down, the sample droplets are separated from the syrup solution and are then measured for size and size distribution using the image processing system. The size distribution shown in Fig. 6.23 indicates that the volume median diameter of the droplets entrained in the test cavity is 935 m, and the maximum diameter is about 2 mm. From these droplet samples, it is estunated that nearly 24% of the total discharged woods metal is entrained into droplets in the test cavity. In the upper containment,1.9 kg of woods metal (2.9% of the total discharged mass) is recovered after the test. Again, the carryover is mainly through the two front air-vents and the seal table exit. From the air vents, the canyover is mostly in the form of fine droplets with a volume median diameter of 60 m. Some small pieces of woods metal are found around the exit of the seal table room. This part of woods metal amounts - to roughly 1.0%. In the subcompartment, similar to the 14.2 MPa tests, most of the discharged woods metal is recovered around and in the seal table. The liquid metal spreads around the subcompartment side walls, forming a halfinch thick crust. Inside the seal table, the solidified woods metal is found to be about one inch thick in the comer areas. On the models of the steam generators and the circulation pumps, small amount of woods metal is found due to the deposition of the dispersed fine droplets. The same phenomenon is observed in the two front air-vents. However, the side walls of the two back air-vents and the neighboring area in the upper containment are clean, which suggests that the re-entrained droplets in the subcompamnent enter the upper containment mainly through the front air-vents. I13 NUREG/CR-6510 I I 6.5. Water Tests with Different Nozzle Sizes With 7 liters of water as the simulant, the tests are carried out with three different sizes of discharge nozzles (3.5 cm,4.0 cm, and 5.0 cm). The maximum nozzle size is limited by the discharge pipe diameter, which is 5.08 cm. For each nozzle size, three tests are performed each at 6.9 MPa,10.4 MPa, and 14.2 MPa, and the vessel pressure is increased step by step for safety reason. The main purpose of these experiments is to investigate the effects of the discharge nozzle size on the dispersion phenomena in DCH accident. In Figs. 6.24 to 6.28, the transients of the gas tank pressure are presented for different nozzle sizes and initial vessel pressures. From these data, the liquid and gas discharge rates are obtained as shown in Figs. 6.29 to 6.33. The corresponding cavity gas velocity transients are presented in Figs. 6.34 to 6.38. Here, the liquid discharge duration (to) plays a key role in determining the blowdown rate, which is estimated in two ways. The direct approach is to compare the time difference between the initiation of the tank pressure drop and the start point of the cavity pressurization (Figs. 6.39 to 6.43). By dividing the total discharged liquid volume with this time delay, the average liquid volumetric discharge rate is obtained. The other way is to calculate the liquid jet velocity based on the gas tank pressure with the assumption of single phase liquid blowdown (section 4). The results from the two different approaches agree with each other. As summarized in Table 6.2, the liquid discharge time (to) reduces significantly for large nozzle sizes (Dj), because to is inversely proportional to the area of the discharge nozzle. In the case of 5 cm nozzle under 14.23 MPa vessel pressure, the discharge time is as short as 0.030 seconds, and the liquid jet velocity is about 117 m/s. During the gas blowdown, the cavity pressure increases significantly (Figs. 6.39 to 6.43). For larger nozzle size or higher vessel pressure the cavity pressure may be much higher than 200 kPa, which is probably due to gas choking at the cavity exit. Initially, the cavity is under atmospheric pressure, hence the gas velocity in the cavity is very high. The maxin.um speed can reach 255 m/s in the tests with 5.0 cm discharge nozzle and under a vessel pressure of 14.2 MPa. As the cavity being pressurized, the blowdown gas accumulates, resulting in a sharp decrease in the gas velocity as shown in Figs. 6.44 to 6.48. The liquid film transport or the droplet entrainment duration in the cavity is very short, about the range the liquid film flow travels through the entire cavity length. The gas velocity for high pressure and large nozzle size tests is very high and the entrainment process completes far before the cavity pressure reaches its maximum or the choking condition. This. may simplify the analytical modeling for the corium dispersion phenomena in the high pressure injection cases. In Table 6.2, the major test results of the nozzle size effects on the dispersion parameters are summarized. The entrainment fraction in the test cavity increases generally as the nozzle size becomes larger or the vessel pressure is higher. Except in the case of 3.5 cm discharge nozzle,6.9 MPa test yields higher entrainment fraction in the cavity than that of the 14.2 MPa test. These results suggest there are two major factors NUREG/CR-6510 114 that may determine the entrainment fraction in the reactor cavity. The first is the gas flow speed. As the vessel pressure or discharge nozzle size increases, the corresponding gas speed in the cavity rises, resulting in larger entrainment rate and therefore more entrained droplets. On the other hand, higher vessel pressure or larger nozzle size causes faster liquid discharge and shorter residual time of the liquid film in the cavity. Subsequently, the entrainment duration in the cavity becomes shorter, resulting in less entrained droplets. These two factors compete against each other and subsequently determine the final entrainment fraction in the cavity. However, for a discharge nozzle size greater than 3.5 cm, the liquid film transport in the cavity becomes insignificant and higher entrainment fraction is observed for increasing nozzle size or vessel pressure. The measured maximum entrainment fraction in the cavity is about 48% of the total discharged water with 5.0 cm nozzle and 14.2 MPa vessel pressure. The size of the entrained droplets in the cavity decreases for larger size nozzles and higher vessel pressure. The minimum volume median diameter is 96 m, which is obtained in the test with 5.0 cm nozzle size under 14.2 MPa vessel pressure. Since the gas velocity increases drastically for larger nozzle and higher vessel pressure, the drag force on the liquid film interface rises proportional to the square of the gas velocity. Thereby, smaller droplets held by surface tension can be torn off the wavy crest. In Fig. 6.49, the dependance of the volume median droplet size are shown against the nozzle diameters. In the upper containment, as presented in Table 6.2 and Fig. 6.50, the liquid carryover increases monotonically with the increase of the nozzle size or the vessel pressure. In the pressure range of the test with a 4.0 cm nozzle size, the carryover shows a roughly linear dependence on the vessel pressure. Table 6.2 Results of the nozzle size effects on the dispersion parameters D, Po to Vy P.- V l Fa. d. d. Fa (cm) (MPa) (s) (m/s) (kPa) (m/s) (%) ( m) (pm) (%) 3.5 6.90 .048 140 147 55 33 184 700 2.4 14.2 .036 202 203 132 26 151 587 5.3 6.90 .053 106 162 96 21 173 588 2.6 4.0 10.7 .043 129 232 140 - 142 587 3.5 14.2 .034 148 280 181 30 128 450 5.4 5.0 7.00 .043 83 187 136 - 104 400 3.0 14.6 .030 117 353 255 48 96 242 5.7 l l 1 l ) i i j 115 NUREG/CR-6510 1; .. ! 1 i i l 8 Break size: 3.5 cm, Break pressure: 6.9 MPa, 6 - I A a $4 \ c 2 - 0 1 -2 3 4 , t( second ) Figure 6.1 Pressure transient in the test vs. sel (6.9 MPa water test) NUREG/CR-6510 - 116 r 200 20 Break size: 3.5 cm ) Break Pressure: 6.9 MPa - 1 n ^ .tc 150 - - 15 ^ ea x w a X V m x c j = N 100 - 4 3> - 10$- e l E .t, l o .t ~5 I g 3 50 - l 5 O l 1 ' ' ' ' ' 'I ''''''''''' ' ' 0 0 1 2 3 4 t ( second ) Figure 6.2 Water and gas mass flow rate transients (6.9 MPa water test) 117 NUREG/CR-6510 l 160 Break size: 3.5 cm, Break pressure: 6.9 MPa. 150 - / 140 - q n S v 130 -
- c. 1 120 -
110 - .1 100 '. 1 2 3 4 t( second ) Figure 6.3 Cavity pressure transients (6.9 MPa water test) l NUREG/CR-6510 118 200 . .
- Break size: 3.5 cm
- W Break Pressure: 8.7 MPa
- 't
.: local,1" above floor , , ^ 150 .. .... . . .... .. .. .. .. ....
- area averaged i
w 6 : 8 } 100 ....,.[.. : .. ... l
- p {
N . ao ' ? b T { h:: a 50 ., . N = - .) i 0------ -- -- -2 3 4 1 t ( second ) Figure 6.4 Gas velocity in the test cavity (6.9 MPa water test) 119 NUREG/CR 6510 I; 1 1 2 Break size: 3.5 cm Break pressure: 6.9 MPa n v h1 - = CH2 0 .-- . .- . r- : . . ,---^.- . .- 3 - 4 - n E3 - E Y_. 2 1 CH6 0 . . . ./ . k A. ^ ^^ . ~ . . . 1.8 2 2.2 2.4 t(second) Figure 6.5 Water film thickness transients (6.9 MPa water test) 1 1 NUREG/CR-6510 120 i . 3/8" isokinetic sampling probe . Dmax=700 m SE 3 - Dvm=184 m g o ' D32=126 p R [ Experimental data _f ' @ 4E-3 c . - i~ ' .o f s, Upper limit Log-normal fitting ' -'s / - 3m 3E 3 - I s
- u. , s g
s - E a 2 E-3 - J ' 's s l Q s > . s ['s - 1 E-3 -L l ,; .i -l . ; OE0 100 200 300 400 500 600 Droplet Diameter ( m) i Figure 6.6 Size distribution of the droplets at cavity exit (6.9 MPa water test) i 121 NUREG/CR-6510 15 i Break size: 3.5 cm Break Pressure: 14.2 MPa 10 - n S 2 v w 5-0 . . ... ......... ..... . 0 0.5 1 1.5 2 2.5 3 l t( second) Figure 6.7 Pressure transient in the test vessel (14.2 MPa water test) l 1 l l l i 122 NUREG/CR-6510 1 400 50 Break size: 3.5 cm Break Pressure: 14.2 MPa --40 ^ 300 - m 8 x .c e.o - 6 s --30 x= e 3 200 - \ $a E E --20 y .R ' 5 $> m a a <C O l ~3 100 - -10 ) 1 0 . . . . . . . 0 0 1 2 3 t ( second) Figure 6.8 Water and gas mass flow rate transients (14.2 MPa water test) 123 NUREG/CR-6510 l 250 l Break size: 3.5 cm Break Pressure: 14.2 MPa i 200 - n b v c 150 - l J 100 , , , At , , , , , , , , , 0 1 2 .3 i t( second ) Figure 6.9 Cavity pressure transients (14.2 MPa water test) 124 NUREG/CR-6510 4 i 140 - .
- Break size: 3.5 cm Break Pressure: 14.2 MPa 120 - - -- --- - - -
f l-3* 100 - . . v . . b - - 5- - - --- g 80 - o . y 6Q ..... .......... ... ....... .. . ................. 60 6 : g ................ .... ..............'............. U fQ . : 20 - - - - - 0 . . . 0-- 1 2 3 t ( second ) Figurc 6.10 Gas velocity in the test cavity (14.2 MPa water test) 125 NUREG/CR-6510 l F. l l i i 2 Break size: 3.5 cm 1.5 - BrerJc Pressure: 14.2 MPa .c c j 1- = 0.5 - CH2 g t_ , _ gq _ y m
- e. ~ _.
3 2- pq s 1-l CH6 t j UU - ~ P '-- -- 0 . . . . . . 1 0.5 0.7 0.9 1.1 1.3 1.5 t ( second ) Figure 6.11 Typical film thickness transients (14.2 MPa water test) i 126 NUREG/CR-6510 I l 1 . 3/8' isokinetic sampling probe . Dmax=587pm SE _ / s o _, Expen. mental data u s Dvm=151 km - I s I 3 D32=105 pm V @ 4E 3 - l \ I s f 3E 3 s Upper limit Log-normal fitting 3 2E-3 - f, W ' - > s4 s, . N ' ~ - 1 E-3 , ~ L ~ I 's N O E0 0 100 200 300 400 500 600 Droplet Diameter ( m) Figure 6.12 Size distribution of the droplets at cavity exit (14.2 MPa water test) 4 127 NUREG/CR-6510 i
- l 1
i 20 J j Break size: 3.5 cm I Break pressure: 14.2 MPa l 15 - l- - I e l l 2 10 - ! c I i 5- l I Liquidjet stans! l 0 . , , ) 6 3 4 5 t ( seconds ) Figure 6.13 Pressure transient in the gas tank (14.2 MPa woods metal test) NUREG/CR-6510 128 l 1000 40 Break size: 3.5 cm Break pressure: 14.2 MPa 800 - 30 2 7 ; # b 7 7 5 '600 - 5 a 5> - 20 g -E <0 e t 400 - E 2 a l 3 .sr o .a - 10 200 - i 0 ' . 0 3 4 5 6 t( seconds ) Figure 6.14 Liquid and gas mass flow rate transients (14.2 MPa woods metal test) 1 1 NUREG/CR-6510 J l 300 j ' Liquidjet starts Break size: 3.5 cm { I Break pressure: 14.2 MPa l i 1 I n 'I D v 2M - s I l I I I i 100 , , 3 4 5 6 t( seconds) Figure 6.15 Cavity pressure transients (14.2 MPa woods metal test) i l l l t i i NUREG/CR-6510 - 130 [:- l l ) t 150 . . . .
- : : Break size: 3.5 cm
- : : Break pressure: 14.2 MPa 120 .
90 ......... .., .. ......... .... . ..,............,........ . n . . . . se . . . . g . 60 : -- - - - 5 - --- - . .,. ......s....... 30 ...........s.... . ...s............ 0 3 3.5 4 4.5 5 ' 5.5 t ( seconds ) Figure 6.16 Gas velocity in the test cavity (14.2 MPa woods metal test) 131 NUREG/CR-6510 i .5 Liquidjet starts l l 4- l l 3- , M X2-1- CH2 l 2 I I I O , I l-l- l I I l. T -1 l 1 -l Gas jet starts CH6 1 I o . . .. ., .. . 3.8 3.9 4- 4.1 4.2 4.3 4.4 i i t( seconds) Figure 6.17 Liquid film thickness transients (14.2 MPa woods metal test)
- 4. 4 i
I i 1 1 NUREG/CR-6510 132 h' 1 E-3 . . . . , . . . . , . . . . . 3/8" isokinetic sampling probe Dmax=2400 m O 8E-4 - - - Dvm=857pm - y s o [' D32=637 pm , s = @ 6E-4 's s - j s . $ ; 'I ~ ~ E ,'- Experimental data 's s 20 - 2E 4 - s Upper limit Log normal fitting - I . t . / /. . . . t . . . . I . . . . t . . . . o 500' 1000 1500 ' 2000 Droplet Diameter ( m) Figure 6.18 Size distribution of the droplets at cavity exit (14.2 MPa woods metal test) 133 NUREG/CR-6510 l I i l i l 12 - . . . . . so . . . . . . . . !, . . . . . . h. . .,.. . . !, . Wo . j=3.$c. . . . D.. m .............. bds m.etc.1 teht, e 8- :-ii--i--i-M . . . . b . . . . . 2 V 4 . 1...... g f. .....,......... g . g . C I 4 ......... .. .....t.... ..s,...... . s. . ......J....,.... 2 *. i 0 2 3 4 5 6 7 8 i t (second) ! l' Fig. 6.19 Gas tank pressure transient in woods metal test (3.5 cm nozzle, Po=10.4 MPa) , I i NUREG/CR-6510 134 4 250 25 l l l Woods metal test,
- N : b. j=3.5 cm :.
200 - - - - - 20 t . . liquid flow rate : gas flow rate ; q x ' ;150 e x;-- - ; - -- -} -{ y - 15 6 a" p . . . . y o . . . . o e ~; 10 m , 100 -} - - - i g se- . 50 .- - 5 0 0 2 3 4 5 6 7 8 ) t ( second ) ) 1 Fig. 6.20 Liquid and gas discharge rate (3.5 cm nozzle, Po=10.4 MPa) i 135 NUREG/CR-6510 r I l l 1 160 . . . . .
- : : Woods metal test' 150 ; ;
; Dj=3;5 cm- - - :- 140 . q 130 - Q .R . S. . b o " 120 .- . 1..-- 110 , ' ' ~ - - - - 100 - 90 2 3 4 5 6 7 8 t ( second ) Fig. 6.21 Cavity total pressure transient (3.5 cm nozzle, Po=10.4 MPa) . NUREG/CR-6510 136 i l l l. i, 80 . . . . .
- : Woods metal test,
- : Dj=3:5 cm :
60~ - -- ';1;- .: q . ~$, 40. -- ::--- -l- -- > -- I 20 - '
- s. n.
.0 2 3. 4 5 6 7 8 t ( second ) Fig. 6.22 Gas velocity in the cavity (3.5 cm nozzle, Po=10.4 MPa) .p 137 NUREG/CR-6510 r.- i e g.e s s . 3/8'. isokinetic sampling probe - Dmax=2150 m _ 1E 3 - - '- Dvm=935 pm O - s s / D32=788 km ._!_ sExperimental data ' s .6 ', ve l 's _ Upper limit Loa normal fitting d- l - e SE 4 - Ns E i s a u A / \ . l Q ' , s ''~**' k OEo o 500 1000 1500 2000 2500 Droplet Diameter ( m) Fig. 6.23 Droplet size distribution at the cavity exit (3.5 cm nozzle, Po=10.4 MPa) 1 i l 138 l NUREG/CR-6510 'I I i l l I l r l l ~10 . . .
- : WaterIest, Dj=4.0cm 8 ........................
- n. 6 .
g . ~ 2 . y . . . 4 ............. ...... ........... ...... ............... 2 - - 0 6 7 8 9 10 t (second) Fig. 6.24 Tank pressure transient in 6.9 MPa water test with 4.0 cm nozzle I l 139 NUREG/CR-6510 12 00 . . . . . . _ i ; Water test, Dj=4.0 cm P0=10.70 MPa - 10.00 .~ . . . . r 8.00 . . . -. . . qq . . . . . . i 2 * *
- 2 6.00 : .---
a . 4.00 :- .-;. . .: . . . - . . .: . . . . . '.r 2.00 '. -:-:-:.....-.:... l I 0.00 , 0 1 2 3 4 5 6 7 i 1 l t (second) Fig. 6.25 Tank pressure transient in 10.7 MPa water test with 4.0 cm nozzle NUREG/CR-6510 140 o 20 . , . . ; Water test; Dj=4.0cm . 15 .....................>.......... 48 '% . . . g . .
- g. . . . .
. y 10 ...... ....;....... ..;....... v . . ..z..........z.......... w . 5 ~.-- 0 0 2 4 6 8 10 t (second) Fig. 6.26 Tank pressure transient in 14.2 MPa water test with 4.0 cm nozzle 141 NUREG/CR-6510 1 2 is ,r I 8 . . . .
- ; ; Water test, Djv5.0cm 7 .
6 ', >. -<.. <. 5 , . . .
- n. . .
g g4 . 4 . . . . i 3 . 2 . 1 . . 2 4 6 8 10 0 t (second) i I i Fig. 6.27 Tank pressure transient in 6.9 MPa water test with 5.0 cm nozzle , i i 1 - NUREG/CR-6510 L 142 15 . . . , , - - Water test, Dj=5.0 cm. 12 m 9 ................... ..................................... g .. . . . . c .. . . . g n= ..r..- 6 . . . . . -c.3 .- . . . 0 0 1 2 3 4 5 6 t (second) Fig. 6.28 Tank pressure transient in 14.2 MPa water test with 5.0 cm nozzle F 143 NUREG/CR-6510 i i i 40 1 150 . 4
- : WaterIest, Dj=4.0cm l
. . l . g . , m. 120 - ------ - --l--- - ------l--------------l------------- 32 m. g . . g . y o . 24 ee 1 c- 90 -- -- S 3: . o . v. . o c . . c %s . . m A ---------- - ---- -- ------------'-------------- 16 n O 60 4 '. '. > E l ) M . O j y . . 8 l 30 -- -- - --~; ----{-----------l-------------- . l 0 0 7 8 9 10 1 6 t (second) l Fig. 6.29 Water and gas discharge rate in 6.9 MPa test with 4.0 cm nozzle 144 NUREG/CR-6510 cc I 200.00 . . . 40 Water test, Dj=4.0 cm
- : ; P0=10.70 MPa Q 150.00
-----b ----i--- % - - l- - - - - - -l- - - - - - -l - - - - - - - l - - - - -30 - - g2 6 . . 6 w . . . . . o . . g 3 . . . . 20 5 8 100.00 -- --- :-- --- 4-- 4-- . c:: . . E a, . . . . . . . E o . . . . a y . . ----;--------;------.----d------- 10 C > 50.00 ------ 0.00 0 0 1 2 3 4 5 6 7 t (second) 1 Fig. 6.30 Water and gas discharge rate in 10.7 MPa test with 4.0 cm nozzle i f l 145 NUREG/CR-6510 50 200 . . . . .
- : Water test, Dj=4.0cm n .160 ----
- - - - - - - i - - - - -l- - - - - - - - - l- - - - - - - - - h - -40
- - -m- - - - m ,, . . y x. u v . .,f,, . . . , ,, . . . O a - - - - - - - - . - - - - - - - - 30 w - - - - - - - - ~ a 120 ---- - - ' . . . e p . . . 3 o . . o c . . . . C m a 10 N E EO '*** ' ''J'*************'''.********~ A E
- g g ,
a is . % : : : : --------s -- --- -------- e--------, ---- - - - 10 40 -- ----- - 1 0 ] 0 6 7 8 9 10 4 5 t (second) Fig. 6.31 Water and gas discharge rate in 14.2 MPa test with 4.0 cm nozzle i l NUREG/CR-6510. 146 f. t 200 . . . . 40
- : : Water test, Dje5.0cm 2
f_150 ----b. --i------k----------k----------- 30 2 w C. . u v 2 , o ~ Q . a 8, 100 -- ---- -h - ---?--------;----------;----------- 20 a$ m a . . . . m g . . . . = . . . g 2 . : g j So 4 .....'.... . . . . . . } . . . . . . . . . ' . . . . . .10. 0. . . i . . . 0 0 0 2 4 6 8 10 t (second) Fig. 6.32 Water and gas discharge rate in 6.9 MPa test with 5.0 cm nozzle 147 NUREG/CR-6510 r. l ] l 1 1 'l 250 . 75 Water test, Dj=5.0 cm. n 200 - ------:----- u,- - - ? - - - - - - - - : - - - - - - - -: - - - - - - - :- - - - - - 60 G A : i.s :. : : : . u .x 6 . ~ c . . . . . g- ............j............................ 45 y e., g 150 ............... _ 3 o . . *
- O g .
5
- S>. .
= 4 . . . . :. . . . . . . . ; . . . . . . . . .: . . . . . . . . . :. . . . . . . . .. 30 g 8E 300 ..........:..... E g- . . . M 2 . C ,C" . . . . --------~.----- ---.--------.- -- --.--------- ----- --- *3 . 15 50 . 0 0 0 1 2 3 4 5 6 t (second) Fig. 6.33 Water and gas discharge rate in 14.2 MPa test with 5.0 cm nozzle NUREG/CR-6510 148 l l = 180
- : Watek test, Dj=4.0cm 160 i-
-l
- g. .
cc . g 140 ..... . . . . . . . . h. .. o . . . A . 120 *- ff ^ 100 1 6 7 8 9 10 t (second) Fig. 6.34 Cavity pressure transient in 6.9 MPa test water test with 4.0 cm nozzle 149 NUREG/CR-6510 r-r 1 1 l l 1 l 240.00 . . . . Water test, Dj=4.0 cm 220.00 - i -- i - po=10,qo upa . . . . j 200.00 :- 180.00 .- . ~.
- n. . . .
4 . . . b 160.00 l - i. . m . . 1 A : : : : 140.00 - - 120.00 - ;- ;- 100.00 - --k-k-----k-- ) 80.00 0 2 4 6 8 10 t (second) Fig. 6.35 Cavity pressure transient in 10.7 MPa test water test with 4.0 cm nozzle 150 NUREG/CR-6510 i 300 . . . . .
- : : Water test, Dj 4.0cm 250 -
i- -- W- : : t-- s ./ n 200 ........,. 49 . . , . . A . . . . . f . q A , . . . . ' -+ - - 150 '.- '. ) ) 'r ; 100 %. .: :. . 50 4 5 6 7 8 9 10 t (second) l Fig. 6.36 Cavity pressure transient in 14.2 MPa test water test with 4.0 cm nozzle l l 1 151 NUREG/CR-6510 .200 . .
- : ; Water test, Dje5.0cm 180 -'
~ . . . { . . 160 ..........s..... n . W . . . . A v ' 140
- x. .
120 - - ,..s~~ 100 - 80 2 4 6 8 10 0 t (second) Fig. 6.37 Cavity pressure transient in 6.9 MPa test water test with 5.0 cm nozzle 152 NUREG/CR-6510 400 . . . . . Water test, Dj=5.0 cm. 350 -.- . 300 -.- , m . k . . . . e4 . . . . b250 .- . .- . O . . . . . C . . . . . 200 . 150 .- ^ 100 0 1 2 3 4 5 6 I t (second) Fig. 6.38 Cavity pressure transient in 14.2 MPa test water test with 5.0 cm nozzle 153 NUREG/CR-6510 n l l I
- ; Water test, Dj=4.0cm 80 -
\ -l - -l - - q 60 -< 40' .- .- -. . 20 -- ; - - 0 7 8 9 10 6 t (second) Fig. 6.39 Cavity gas velocity transient in 6.9 MPa test water test with 4.0 cm nozzle l NUREG/CR-6510 154 140.00 . . . . . . Water test, Dj=4.0 cm 120.00 .' .'- : PO=10.10 MPa - 100.00 *;-:- ;-;--:~--;- 2 80.00 . . .3 . 60.00 .~- - - 40.00 -:;- :- . 20.00 - - -- - 0.00 0 1 2 3 4 5 6 7 t (second) Fig. 6.40 Cavity gas velocity transient in 10.7 MPa test water test with 4.0 cm nozzle 155 NUREG/CR-6510
- c. .
1. 1-l I 200 . . .
- : : Water test, Dj=4.0cm 150 g .
$100 - :+:i- -50 - - . 1 0 .; 6 7 8 9 10 4 -5 t (second) Fig. 6.41. Cavity gas velocity transient in 14.2 MPa test water test with 4.0 cm nozzle i 1 156 NUREG/CR-6510 - ; i 140 . . , .: : : Water test lDj=5.0cm 120 q 100 ....... . . . . . . . ... . ~a 80 ,- .- -y m . . . > 60 ~ ~ 40 :- 20 .- O l 1 0 1 2 3 4 5 6 7 8 ; t (second) : 1 i Fig. 6.42 Cavity gas velocity transient in 6.9 MPa test water test with 5.0 cm nozzle 1 \ i l 157 NUREG/CR-6510 . i 300 . . -
- Water test, Dj=5.0 cm.
250 .- 200 .- . . . A v 1 50 .- . . . p . 100 .- l z . . . . 1 50 .- i 0 0 1 2 3 4 5 6 t (second) ) Fig. 6.43 Cavity gas velocity transient in 14.2 MPa test water test with 5.0 cm nozzle NUREG/CR-6510 - 158 l ,,,, j . 3/8" Isokinetic sampling probe . SE-3 - Dmax=5884m - O' o Dvm=1734m y . ' D32=122 km D 4E-3 - e o - f /- - 's - _ Experimental . data = i s I l @ 3E 3 - ' 's - Lt -- 's g M, ~ 's, Upper limit Log normal fitting s 2E 3 - ' N/ - 3 i s ' 's _ i 1 E-3 - ~ I i I ~~,~~. % OE0 ~!' ' '' ''' ' 0 100 200 300 400 500 600 Droplet Diameter ( m) Fig. 6.44 Water droplet size distribution in cavity (6.9 MPa test,4.0 cm nozzle) l l 159 NUREG/CR-6510 i # . 3/8" isokinetic sampling probe "" * ~ S E -- Experimental data a u : Dvm=142pm D ' 's D32=100 g m D 4E 3 - l ' s _, s g . , s , s, .g ' s m 3E-3 - LE l O Upper limit Log-normal fitting . .- - t s E m 2E 3 ', s '/ - o 'J l 's . 's ' - 1E-3 - i ~,, -l ___ OE0 0 100 200 300 400 500 600 Droplet Diameter ( m) Fig. 6,45 Water droplet size distribution in cavity (10.7 MPa test,4.0 cm nozzle) NUREG/CR4510 160 l ,,y , k . 3/8' isokinetic sampling probe Dmax=450pm SE-3 - i 's - - a o r Dvm=128pm u . Experimental data - ?> # _ 3: 8 % - D32=96 Hm u 4E 3 - i ( s 8 - .; _s . = , s M 3E-3 - 7 , E. 5 ,' S Upper limit Log-normal fitting . b 2E-3 ,' 's, - a 's . ,' 's ' 1 E-3 _ 1 , l r OE0 - - 0 100 200 300 400 500 600 Droplet Diameter ( m) Fig. 6.46 . Water droplet size distribution in cavity (14.2 MPa test,4.0 cm nozzle) 1 161 NUREG/CR-6510 . - m.. I 8 E-3 . N i . . 3/8" isokinetic sampling probe 7E-3 - i ' \ Dmax=4004m O / \ ' = Experimental data Dvm=1044m s 6E-3 ~ r s I > s D32=82 Mm - v ' SE 3 - 'r-8 - l s 8 4E-3 - l \ tt . s @ 3E-3 - 's - 2 . > '. - $ 2E 3 - 1 s Upper limit Log-normal fitting - 'sf s 1 E-3 - I ' ' * ~., J ' ' ' ' ' ' ' ' ' ' " ' ' ' ' ' ' ' ' ' OE0 O 100 200 300 400 500 Droplet Diameter ( m) Fig. 6.47 Water droplet size distribution in cavity (6.9 MPa test,5.0 cm nozzle) NUREG/CR-6510 162 e,f . 1 E-2 . . , . . . , . . 3/8' isokinetic sampling probe , s Dmax=2424m O 8E-3 - 7, s u 's Dvm=96 4 m > r 3 V s D32: 77 km . i s c 6E 3 - e s ~ Upper limit Log-normal fitting _ o " t i o '( s G - 1 ' ,f' e 4E-3 - i . 's s : Experimental data _ E , U / 'g 2 E-3 - o 's ~ l I / y OE0 ' - - ' - - - ' - - ' 50 100 150 200 250 Droplet Diameter ( m) Fig. 6.48 Water droplet size distribution in cavity (14.2 MPa test, 5.0 cm nozzle) 163 NUREG/CR-6510 0.20 -A- Dj=3.5 cm --x- Dj=4.0 cm -*- Dj=5.0 cm 0.18 - - - 0.16 - n . . g v E 0.14 - - - -
- e . . .
0.12 *- w 0.10 - - - 0.08 6 8 10 12 14 16 P0 (MPa) Fig. 6.49 Effects of vessel pressure and nozzle size on D. in cavity for water tests NUREG/CR-6510 164 1 i 6.00 .
- =
.I 5.00 - m . . . . # 4.00 v -i- -- -- -} -- -- -- b - - Z n . . 3.00 -*- ' U x . . .? A - w ' ' Ea 2.00 -- - - - - - - 1.00 - A Dj=3.5 cm x Dj=4.0 cm o Dj=5.0 cm 0.00 6 8 10 12 14 16 P0 (MPa) Fig. 6.50 Effects of vessel pressure and nozzle size on water carryover 165 NUREG/CR-6510 1 I
- 7.
SUMMARY
AND CONCLUSIONS The Purdue University 1:10 scale DCH separate effect experiments were carried out under wide range of pressure conditions with water and molten woods metal to simulate the core melt in the prototypic accident. A 1/10 scale Zion reactor model was designed and constructed to investigate the blowdown liquid dispersion and transport. Various instruments were employed to capture the parameters that are important for understanding the corium dispersion phenomenon in the accident. Among them, two unique instruments, i.e., the isokinetic droplet sampling probe and the woods metal film thickness probe, were developed to measure the dispersed droplet size and the liquid metal film thickness in the test cavity, respectively. The tests were grouped in two categories: the 1.4 MPa low pressure experiments and the 6.9 MPa to 14.2 MPa high pressure experiments. For the 1.4 MPa low pressure tests,3.5 cm liquid discharge nozzle was used. The gas supply came from a separate 10.16 cm pipe line that provided sufficient gas velocity in the cavity to preserve the prototypic entrainment phenomenon at a vessel pressure of 6.9 MPa. Three kinds of tests were included in the low pressure category. These were the standard air-water tests, the standard air-woods metal tests, and the parametric tests. In the standard tests,7 liters of water or woods metal were used to simulate the core melt (about 40% of the total core mass). According to the measurements and flow visualization, detailed information regarding the dispersion phenomena were obtained. These included (1) liquid discharge and liquid jet spread-out upon the impingement on the cavity f!oor, (2) liquid film motion and transport, (3) liquid entrainment, entrained droplet size distribution and mass flux, (4) gas velocity and velocity profiles, (5) liquid trapping in the subcompartment, and (6) liquid carry over into the containment / exhaust chamber. In the case of water tests, it was found that about 15% of the liquid was ejected from the cavity before gas blowdown without dispersion. In the cavity, 43% of the discharged water was entrained into droplets with a volume median diameter of 406 m. After the tests, 2.7% of water was recovered in the upper containment and the rest deposited in the subcompartment. No liquid was observed in the test cavity. For the woods metal tests, and due to the low liquid jet velocity (11m/s), the liquid film flow could not reach the cavity exit before gas blowdown. All the discharged woods metal was thus subjected to the entrainment process. About 11% ofliquid metal was entrained into droplets in the test cavity with a volume median diameter of 1.1 mm. After the tests, 1.6% of woods metal was recovered in the upper contamment and 8% remained in the cavity in the form of thin crust over the cavity walls. In the parametric studies, the effects of gas discharge nozzle size (or cavity gas velocity), water inventory, and subcompartment trapping condition on the liquid dispersion process were obtained. The increase of gas velocity in the test cavity significantly increased the liquid entrainment, but had negligible influence on the liquid film motion in the cavity. Larger liquid inventory tended to enhance the entrainment rate in the test cavity due to the increase in the Reynolds number of the liquid film flow. However, the liquid carryover to the containment remained constant (~2.7%) regardless of the changes of discharged liquid volume in the test range. The subcompartment trapping experiments demonstrated that NUREG/CR-6510 166
the trajectory dependent phenomenon was predominant in determining the effectiveness ofliquid entrapment. In the high pressure DCH tests, the water tests were conducted at vessel pressures of 6.9 MPa and 14.2 MPa, and the woods metal tests were performed at 10.7 MPa and 14.2 MPa vessel pressures. For woods metal tests at 14.2 MPa vessel Pressure,38% of the discharged liquid was entrained into droplets in the cavity with a volume median diameter of 857 mm. About 3.4% of the total 56 kg woods metal was dispersed into the upper containment chamber. Compared to the previous 1.4 MPa experiments as summarized in Table 7.1, the entrainment under high vessel pressure was higher and the resultant droplet size was smaller. In the case of the 14.2 MPa water tests, it was found that about 5.3% of the total water was dispersed into the upper containment. In the cavity,26% of the water was entrained into droplets with a volume median diameter of 151 m. As the vessel pressure dropped to 6.9 MPa, about 36% of the discharged water was dispersed into droplets with a volume median diameter of 184 m, and the dispersed liquid fraction in the containment was only 2.4%. A comparison of the data under different vessel pressures in Table 7.1 indicated that the liquid transport process was much faster as the vessel pressure increased. Consequently, the entrainment time became shorter, which tended to reduce the fraction ofliquid dispersion in the test cavity. However, as a result of rise in vessel pressure, the gas velocity increased, which in turn enhanced the entrainment process. These two competing factors determined the degree of the entire liquid dispersion transient. In general, as the vessel pressure increased in the high pressure tests, the results showed obvious increase in the dispersed liquid carryover to the upper containment. Table. 7.1 Comparison of the major results under different test vessel pressures Test type vesselpressure Liquid dispersion Droplet volume Liquid carry-over' fraction in cavity median diameter at (%) cavity exit ( m) (%) 1.4
- 43 406 2.7 Water tests 6.9 36 184 2.4 14.2 26 151 5.3 Woods metal 1.4* 11 1200 1.6 test 14.2 38 875 3.4
*: The gas flow in the cavity is compensated to the velocity caused by 6.9 MPa vessel pressure.
t: % of total mass discharged from vessel. The experimental results also indicated that corium dispersion mechanisms identified in the scaling study were adequate. According to a preliminary estimate based on the scaling study, the entrainment rate correlation proposed in the scaling study was accurate for both air-water and air-woods metal tests (Table 7.2). However, the mean droplet size correlation seemed to under-predict the 1.4 MPa test results by a factor of 2, though it 167 NUREG/CR-6510
matched the results from the high pressure tests as shown in Tab. 7.2. There were two possible explanations for this discrepancy: the entrance effect in the test cavity due to the less deposition of the entrained large droplets (38], and the manner in triggering of the auxiliary gas discharge slightly prior to the completion of the liquid discharge in the 1.4 MPa water tests resulting in liquid jet disintegration [39]. The entrance effect is important for a short channel with large size, where the measured droplet size should be larger due to the less deposition of large droplets based on a trajectory analysis [38]. However, this explanation seemed not convincing for the other two high pressure tests, which were run with the same test cavity. Careful comparisons showed that the only difference was the test actuating mechanism: one was controlled by solenoid valve and the others were triggered with rupture discs. For the case of solenoid valve control, the closing of the solenoid valve was automatic after liquid discharge within certain time period that was overlapped with the gas discharge time. Although this time interval was very short (less than 0.06 s), the gas flow blew the liquid jet apart and generated some large droplets. This explains why the high pressure test results agreed with the correlation. Table 7.2 Comparison between experimental data and preliminary estimation (7 liters of liquid, with 3.5 cm nozzle) Air-water woods metal 1.4 MPa 6.9 MPa 1.4 MPa 10.4 MPa D. ( m) Estimation 150 167 640 620 Experiment 406 184 1120 973 2 3.72 e(g/cm s) Estimation 1.97 1.90 - Experiment 1.75 1.84 4.05 - Finally, it should be emphasized that the droplet entrapment in the subcompartment greatly affected the droplet carryover to the upper containment, regardless of the entrainment fraction in the test cavity. In the high pressure and large nozzle tests, in spite of the fluctuation of the entrainment fraction in the test cavity, more carryover was obtained as the gas discharge rate increased. This phenomenon implied that droplet re-entrainment and transport in the subcompartment were the key factors in determining the total carryover to the upper containment. More theoretical work is therefore needed on the analysis of droplet transport from the subcompartment to the upper containment building. NUREG/CR-6510 168
REFERENCES
- 1. Commonwealth Edison Company (CECO), " Zion probabilistic safety study", Chicago, IL,(1981).
- 2. Power Authority of the State of New York (PASNY) and Consolidated Edison Company (Con.Ed.) of New York, Inc., " Indian Point probabilistic safety study",
(1982).
- 3. W. W. Tarbell, et al, "Results from the DCH-1 experiments", Sandia National Laboratory Report SAND 86-2483, NUREG/CR-4817 (1987).
- 4. W. W. Tarbell, et al, "DCH-2: Results from the second experiments performed in the SURTSEY direct heating test facility", Sandia National Laboratory Report SAND 87-0976, NUREG/CR-4917 (1988).
- 5. B. W. Spencer, D. Kilsdonk and J. J. Sienicki, "Corium/ water dispersal phenomena in ex-vessel cavity interactions", Proc. Intl. Mtg. on LWR Severe Accident Evaluation 2, Cambridge, MA, 15.5-1 (1983).
- 6. N. K. Tutu and T. Ginsberg, " Melt dispersal experiments with the 1/42 scale Watts Bar cavity model", Brookhaven National Laboratory, Technical Report No. A-3024 (1990).
- 7. M.M. Pilch, W.W. Tarbell, and J.E. Brockmann, "The influence of selected contamment structure on debris dispersal and transport following high pressure melt ejection from the reactor vessel", NUREG/CR-4914, SNL (1988).
- 8. M.D. Allen, M. Pilch, R.O. Griffith, D.C. Williams, and R.T. Nichols, "The third integral effects test (IET-3) in the SURTSEY test facility", Sandia National Laboratory, SAND 92-0166 (1992).
- 9. M.D. Allen, M.M. Pilch, T.K. Blanchat, and R.O. Griffith, "Experimnets to investigate Direct Containment Heating phenomena with scaled models of the Zion nuclear power plant in the SURTSEY test facility", NUREG/CR-6044, SAND 93-1049, SNL (1993).
- 10. S. Levy, " Debris dispersal from reactor cavity during low temperature simulant tests of direct containment hHeating, Part I - Tests with constant gas flow rate, Part II -
Tests with blowdown gas conditions", ANS Proc.1991 National Heat Transfer Conf., Minneapolis, MN,323-341 (1991). I1. T. Ginsberge and N. K. Tutu, " Progress in Understanding of Direct Containment Heating Phenomena in Pressurized Light Water Reactors", Invited Paper, Third International Topical Mtg. on Nuclear Power Plant Thermal Hydraulics and Operations, Seoul, South Korea (1988). 169 NUREG/CR-6510
- 12. B. W. Spencer, J. J. Sienecki, and L. M. McUmber, " Hydrodynamics and heat transfer aspects of corium/ water interactions", Electric Power Research Institute, EPRT-NP-5127 (1987).
- 13. J.L. Binder, L.M. McUmber, and B.W. Spencer, " Direct containmen: seating experiments in Zion nuclear power plant geonietry using prototypic materials",1993 ANS Winter Annual Meeting, San Francisco, CA, Nov.,1993.
- 14. R. V. Macbeth and R. Trenberth, " Experimental Modeling of Core Debris Dispersion from the Vault Under a PWR Pressure Vessel: Part 1. Preliminary Experimental Results", AEEW-R1888 (1987).
- 15. N. Zuber, et al, "An integrated structure and scaling methodology for severe accident technical issue resolution", NUREG/CR-5809, EGG-2659 (1991).
- 16. M. Ishii, " Scaling study of corium dispersion in DCH, Appendix G.of An integrated structure and scaling methodology for severe accident technical resolution",
NUREG/CR-5809, EGG-2659 (1991).
- 17. N.T. Obot and M. Ishii, "Two-phase flow regime transition criteria in post dryout region based on flow visualization experiments", Int.1 Heat Mass Transfer 31,2559-2570 (1989).
- 18. J.G. Denten and M. Ishii, " Flow visualization study of post critical heat flux region for inverted bubbly, slug and annular flow regime", NUREG/CR-5171, ANL-88-27 (1988).
- 19. I. Kataoka, M. Ishii, and K. Mishima, " Generation and size distribution of droplet in annular two-phase flow", J. Fluid Eng. 105,230-238 (1983).
- 20. G. DeJarlais, M. Ishii, and J. Linehan, " Hydrodynamic stability of inverted annular flow in adiabatic simulation",2 ofHeat Transfer 108,84-92 (1986).
- 21. L. Bolle and J.C. Moureau, " Spray cooling of hot surfaces", Multiphase Science and Technolog s 1 Edited by G.F. Hewitt, J.M. Delhaye and N. Zuber, 1-98 (1982).
22.R. Henry, "On evaluation of fission product release rates during debris dispersal", Proc. ANS International Topical Meeting - Probability, Reliability and Safety Assessment 1, April 2-7,375-383 (1989).
- 23. G.B. Wallis and J.E. Dobson, "The onset of slugging in horizontal stratified air-water flow", Int. 2 Multiphase Flow 1,173-183 (1973).
- 24. K. Mishima and M. Ishii, " Theoretical prediction of onset of horizontal slug flow",1 Fluids Engineering 102,441-445 (l980).
NUREG/CR-6510 170 l
- 25. Y. Taitel and A.E. Dukler, "A model for predicting flow regime transition in horizontal and near horizontal gas-liquid flow", AIChE J. 22,47-55 (1976).
- 26. M. Ishii and M.A. Grolmes, " Inception criteria for droplet entrainment in two-phase concurrent film flow", AIChE J. 21,308-318 (1975).
- 27. I. Kataoka and M. Ishii, " Mechanism and Correlation of Droplet Entrainment and Deposition in Annular Two-phase Flow", NUREG/CR-2885, ANL-82-44 (1982).
- 28. M. Ishii, S.T. Revankar, G.J. Zhang, Q. Wu and P. O'Brien, " Air-water simulation of phenomena of corium dispersion in direct containment heating", NUREG/CR-6367, PU NE-93/1 (1993).
- 29. L.E. Gill, G.F. Hewitt, J.W. Hitchon, and P.M.C. Lacey, " Sampling ptobe studies of the gas core in annular two-phase flow", Chem. Engr. Sci.18,525-535 (1963).
- 30. T. Allen, Paricle Size Measurement (3rd edition), p. 54. Chapman and Hall, New York, (1981).
- 31. G.J. Zhang and M. Ishii, " Isokinetic sampling probe and image processing system for droplet size measurement in two-phase flow", Int. J. Heat Mass Transfer 38, 2019-2027(1995).
- 32. S.V. Paras and A.J. Karabelas, " Droplet entrainment and deposition in horizontal annular flow", Int. J. Multiphaseflow 17,455-468, (1991).
- 33. K.V.L. Rao, " Liquid nitrogen cooled sampling probe for the measurement of spray drop size distribution in moving liquid-air sprays", presented at the 1st Int. Conf on Liquid Atomization and Spray System, Tokyo, Japan (l 978).
- 34. R.A. Mugele and H.D. Evans, " Droplet size distribution in sprays", Ind. Engr. Chem.
43,1317-1324 (1951). 35.J.E. Koskie, I. Mudawar, and W.G. Tiederman, " Parallel-wire probes for the measurement ofliquid film thickness in two-phase flow", Int. J. Multiphase Flow 15, 512-530 (1989). 36.M.W.E. Coney, "The theory and application of conductance probes for the measurement ofliquid film thickness in two-phase flow", J. ofPhysics E: Scientific Instruments 6,903-910 (1973).
- 37. Q. Wu and M. Ishii, " Investigation of liquid film flow in reactor cavity for DCH separate effect experiments",1Oth Proc. ofNuclear Thermal Hydraulics, ANS Winter Meeting, Nov. 13-17,1994, Washington, DC. p. 47-55.
- 38. G. Zhang, " Study of droplet entrainment and dispersion in two-phase flow", Ph. D.
Thesis, Dept. of Nuclear Engineering, Purdue Univ.,1994. 171 NUREG/CR-6510
- 39. Q. Wu, " Transient two-phase flow and application to severe nuclear reactor accident",
Ph. D Thesis, Dept. of Nuclear Engineering, Purdue Univ.,1995. NUREG/CR-6510 172
NRCFORM 346 U.S. NUCLEAR REIULATORY COMMiss4ON 1. REPORT NUMBER 048) (Ameigned by NRC, Add Vol., supp., Rev., E BIBLIOGRAPHIC DATA SHEET *"''"'"""""*"*'"*"d NUREG/CR-6510, Vol.1
- 2. TITLE AND SUBTITLE PU NE-96/2 Corium Dispersion in Direct Containment Heating
- 3. DA1E REPORT PUBUSHED Separate Effect Experiments With Water and Woods "*" l "^8 Metal Simulating Core Melt for Zion Reactor Conditions September 1999
- 4. FIN OR GRAN' NUheER L1990
- 5. AUTHOR (S) 6. TYPE OF REPORT M. Ishii, Q. Wu, S.T. Revankar, S. Kim, G. Zhang, Purdue University R.Y. Lee, C.G. Tinkler, U.S. Nuclear Regulatory Commission Technical
- 7. PERIOD COVERED (inciusue osass) 3/92-10/95
- 8. PERFORMING ORGANIZATION . NAME AND ADDRESS rrNRc. prove Ovam oske or Asam u s. Nucasar Aspuisenry t , and meene ad*ess; # conencear provafe nemo and meeng addoss.)
Schoolof Engineering Purdue University West Lafayette, IN 47907
- 9. SPONSORING ORGANIZATION NAME AND ADoRESS (r ARC, type 'seme es above' aconesesar, prove NRC Dvem once or Asam & S. NucJoerRepuissary commisam and meene ed*ess }
Division of Systems Analysis and Regulatory Effectiveness Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-0001
- 10. SUPPLEnENTARY NOTES R.Y. Lee. NRC Project Manager
- 11. ABSTRACT 000 words or Asso The research at Purdue University addresses corium dispersion during the Direct Containment Heating scenario in a severe nuclear reactor accident. The degree of corium dispersion has not only the strongest parametric effects on the containment pressurization, but also has the highest uncertainty in predicting it. In view of this, a separate effect testing program on the corium dispersion mechanisms in the reactor cavity and the subcompartment trapping mechanisms was initiated at the Purdue University.
Th3 four major objectives of this study are: (1) to perform a detailed study using a step-by-step integral scaing method, and to evaluate existing models and correlations for dropiet entrainment, particle size distribution and particle trapping, (2) to design and construct a 1/10 scale Zion reactor model, and to perform carefuly scaled experiments using air-water and air-woods metal to simulate the prototypic steam and core melt, (3) to develop reliable mechanistic models for the corium dispersion and transport in the accident scenario, which can be used to predict the Equid and gas blowdown, entrainment droplet size, liquid carryover to the containment, and the subcompartment trapping, and (4) to use the models to perform stand alone calculations for the prototypic conditions. In this report (volume 1), efforts are focused on the first two objectives, whereas the modeing study is documented in volume 2.
- 12. KEY WORDS/DESCRIPTORS (Ust weres erparem ret we esamt rmerchers ei Jaceang am repert) 11 AVALABdU1Y STATDsEW dir co tain nt heating ""
y scaling (** #*) severe accident unclassrfied separate effect experiments (rn Asparo unclassiP.d
- 15. NUMBER OF PAGES
- 16. PRICE NRc FORM 336 040)
on recycled paper Federal Recycling Program
1 LD I I AA MP 7 DS 6-RE AE DF G O ND AN N TA S E LG R AA I E P CT .. ES PO SP N O I S1 S0 I 0 0 0 M0 3 M-5 O5 E S E C50 SS SU E TY2 N E AR I T SA TO C UV STD BR I DA E L N, LP A I R T I UOGT CO I FF N EG F OT Y U RN I L - RH A N AS E EA P LW C _ U _ N ll ll}}