KfK 4024März 1986
SEFLEXFuel Rod Simulator Effects in
Flooding ExperimentsPart 1:
Evaluation Report
P. Ihle, K. RustInstitut für Reaktorbauelemente
Projekt Nukleare Sicherheit
Kernforschungszentrum Karlsruhe
KERNFORSCHUNGS ZENTRUM KARLSRUHE
Institut für Reaktorbauelemente
Projekt Nukleare Sicherheit
KfK 4024
SEFLEX - Fuel Rod Simulator Effects in Flooding Experiments
Part 1: Evaluation Report
P. Ihle and K. Rust
Kernforschungszentrum Karlsruhe GmbH, Karlsruhe
Als Manuskript vervielfältigtFür diesen Bericht behalten wir uns alle Rechte vor
Kernforschungszentrum Karlsruhe GmbHPostfach 3640, 7500 Karlsruhe 1
ISSN 0303-4003
-1-
Abstract
This report is a summary of an experimental investigation which is apart of
the German LWR safety program. The aim of the SEFLEX program has been to
quantify the influence of the design and the physical properties of different
fuel rod simulators on heat transfer and quench front progression in un
blocked and blocked rod bundles during the reflood phase of a LOCA in a PWR.
Fuel rod simulators with Zircaloy claddings and a gas-filled gap between
claddings and pellets exhibit lower peak cladding temperatures and shorter
quench times than gapless heater rods with stainless steel claddings. Grid
spacers cause significant cooling enhancement downstream during the time span
at which maximum cladding temperatures occur. Ballooned Zircaloy claddings,
forming e.g. a 90 percent blockage, are quenched substantially earlier than
thickwall stainless steel blockage sleeves attached to the rods, and even
earlier than undeformed rod claddings. A comparison of test data with results
of the "Best Estimate" computer program COBRA-TF shows a good agreement with
unblocked bundle data including grid spacer effects.
This report is accompanied by a unblocked bundle data report (KfK 4025) and
a blocked bundle data report (KfK 4026). These three reports conclude the
SEFLEX program.
-IJ-
SEFLEX-Brennstab-Simulator-Effekte in Flutexperimenten
Teil 1: Auswertebericht
Kurzfassung
Dieser Bericht ist eine Zusammenfassung einer experimentellen Untersuchung,
die ein Teil des deutschen LWR Sicherheitsprogramms ist. Das Ziel des SEFLEX
Programms war die Quantifizierung des Einflusses von Aufbau und physikali
schen Stoffdaten von verschiedenen Brennstabsimulatoren auf den WärmeUbergang
und das Fortschreiten der Benetzungsfront in unblockierten und blockierten
StabbUndeIn während der Flutphase eines Kühlmittelverluststörfalles in einem
LWR. Brennstabsimulatoren mit Zircaloy-HUllrohren und einem gasgefUllten
Spalt zwischen Hüllrohren und Pellets fUhren zu niedereren Maximaltempera
turen der HUllrohre und zu kürzeren Wiederbenetzungszeiten als spaltlose
Heizstäbe mit EdelstahlhUllrohren. Abstandshalter verursachen eine bedeutende
Verbesserung der KUhlung in der Nachlaufströmung während der Zeitspanne, in
der die HUllrohre das Temperaturmaximum erreichen. Aufgeblähte Zircaloy-Hüll
rohre, die z.B. eine KÜhlkanalversperrung von 90 % darstellen, werden erheb
lich frUher benetzt als dickwandige, an den Stäben angebrachte Blockadehülsen
aus Edelstahl, und sogar früher als unverformte HUllrohre. Ein Vergleich der
Versuchsdaten mit Ergebnissen des "Best Estimate" Rechenprogramms COBRA-TF
zeigt eine gute Ubereinstimmung mit den Meßdaten der unblockierten BUndel
einschließlich der Abstandshaltereffekte.
Zu diesem Bericht gehören zwei getrennte Berichte, Meßdaten von Experimenten
mit unblockierten Bündeln (KfK 4025) und mit blockierten BUndeIn (KfK 4026).
Mit diesen drei Berichten ist das SEFLEX-Programm abgeschlossen.
-III-
TABLE OF CONTENTS
Listing of Figures
Listing of Tables
Preamble
IV
VIII
IX
1,
2.
Introduction
SEFLEX Reflood Program
1
5
3. Test Facility
3.1 Test Loop and Bundle Housing
3.2 FEBA Heater Rod and REBEKA Fuel Rod Simulator
3.3 Blockage Design
3.4 Ins t rumenta tion
3.5 Operational Procedure
4. Test Matrix
9
9
11
12
14
16
18
5.
6.
7.
8.
SEFLEX Reflood Test Results and Comparison with FEBA Data
5.1 Bundle Behavior
5.2 Grid Spacer Effects
5.3 Blockage Effects
Analytical Simulation of Reflood Experiments
6.1 COBRA-TF, A "Best-Estimate" Computer Program
6.2 Simulation of FEBA and SEFLEX Tests
6.3 Comparison of Test Data with COBRA-TF Calculations
Conclusions
References
20
20
27
38
42
42
44
46
50
51
-lV-
LISTlNG OF FIGURES
Figure
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
4-loop steam generator system and pressure vessel withinstallations of a pressurized water reaetor
Fuel rod eladding loading in a 2F-eold leg break LOCA
OECD Halden Reaetor Projeet: Comparison of nuelear fuel rodand SEMISCALE heater rod responses
OECD Halden Reaetor Projeet: Comparison of nuelear fuel rodand REBEKA fuel rod simulator responses
FEBA test loop used for SEFLEX tests
Photograph of FEBAjSEFLEX test rig
Cross-seetional view of a 5 x 5 rod bundle
Cross-seetional view of FEBAjSEFLEX test seetion
Original and modified upper bundle end and plenum ofFEBAjSEFLEX test seetion
Original and modified lower bundle end and plenum ofFEBAjSEFLEX test seetion
Cross seetion of a FEBA heater rod
Cross seetion of a REBEKA fuel rod simulator
Working drawing of the REBEKA fuel rod simulator modifiedfor SEFLEX tests
Axial power profile and loeation of grid spaeers of FEBAand REBEKA rod bundles in SEFLEX tests
Seetional view of the 90 pereent bloekage with bypassrealized for FEBA tests
Seetional view of the 90 pereent bloekage with bypassrealized for SEFLEX tests
Working drawing of ballooned REBEKA fuel rod simulatorswith instrumentation used for SEFLEX tests
Photograph of the SEFLEX 90 pereent bloekage after the tests
Sehematie diagram of SEFLEX instrumentation for unbloekedand bloeked rod bundle tests
Radial and axial positions of eladding, grid spaeer, fluid,and housing TCls for unbloeked rod bundle tests
56
57
58
59
60
61
62
63
64
65
66
67
69
71
72
73
75
77
78
79
Figure
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
-v-
Radial and axial positions of cladding, heater sheath, gridspacer, fluid, and housing Te's for blocked rod bundle tests
Cladding temperatures measured at four different axiallevels in FEBA and REBEKA rod bundles
Housing temperatures in FEBA and REBEKA rod bundles
Surface heat fluxes of FEBA and REBEKA rods
Cladding temperatures and surface heat fluxes of FEBA andREBEKA rods during the early portion of reflooding
Release of stored heat from FEBA and REBEKA rods
Heat transfer from FEBA and REBEKA rods (related to coolantsaturation temperature)
Cladding temperatures, surface heat fluxes, heat transfer,and heat release during quenching of FEBA and REBEKA rods(FEBA test No. 223, SEFLEX test No. 05)
Cladding temperatures, surface heat fluxes, heat transfer,and heat release during quenching of FEBA and REBEKA rods(FEBA test No. 223, SEFLEX test No. 07)
Cladding temperatures, surface heat fluxes, heat transfer,and heat release during quenching of FEBA and REBEKA rods(FEBA test No. 216, SEFLEX test No. 03)
Cladding temperatures, surface heat fluxes, heat transfer,and heat release during quenching of FEBA and REBEKA rods(FEBA test No. 218, SEFLEX test No. 06)
Cladding temperatures, surface heat fluxes, heat transfer,and heat release during quenching of FEBA and REBEKA rods(FEBA test No. 214, SEFLEX test No. 04)
Azimuthai cladding temperatures of a REBEKA rod measuredat the bundle midplane
Azimuthai cladding temperatures of a REBEKA rod duringquenching measured at the bundle midplane
Radial temperature profiles as function of time duringquenching of FEBA and REBEKA rods
Quench front progression and liquid inventory after275 seconds in FEBA and REBEKA rod bundles
Water carry over from FEBA and REBEKA rod bundles
Influence of the grid spacer at the bundle midplane on theaxial temperature profiles in FEBA and REBEKA rod bundles
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
Figure
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
-VI-
Influence of reflood conditions on cladding and grid spacerternperatures at the bundle rnidplane
Cladding ternperatures and surface heat fluxes at leadingedge and 12 rnrn downstrearn of bundle rnidplane grid spacer(test No. 03)
Cladding ternperatures and surface heat fluxes upstrearn anddownstrearn of bundle rnidplane grid spacer (test No. 03)
Fluid TC signals upstrearn and downstrearn of bundle rnidplanegrid spacer (test No. 03)
Cladding and grid spacer temperatures, and fluid TC signalon enlarged time scale at beginning of reflood (test No. 03)
Cladding and grid spacer temperatures, and fluid TC signal onenlarged time scale at quenching of grid spacer (test No. 03)
Cladding and grid spacer temperatures, and fluid TC signalabove the bundle rnidplane (test No. 03)
Cladding and grid spacer temperatures, and fluid TC signalabove the bundle rnidplane (test No. 05)
Cladding ternperatures and surface heat fluxes at leadingedge and 12 rnrn downstrearn of bundle rnidplane grid spacer(test No. 05)
Cladding ternperatures and surface heat fluxes upstrearn and17 rnrn downstrearn of the grid spacer above the bundle rnidplane(test No. 03)
Cladding ternperatures and surface heat fluxes upstrearn and17 rnrn downstrearn of the grid spacer above the bundle rnidplane(test No. 05)
Cladding ternperatures and surface heat fluxes upstrearn and17 rnrn downstrearn of the grid spacer above the bundle rnidplane (test No. 07)
Ternperatures rneasured at the rnidplane of a 90 percentblockage and in the blockage bypass of FEBA and REBEKA rodbundles
Ternperatures rneasured 10 rnrn downstrearn of a 90 percentblockage and in the blockage bypass of FEBA and REBEKArod bundles
Cornparison of rneasured and calculated heater sheath ternperatures and corresponding cladding ternperatures rneasured at thebund1e rnidplane in the blockage bypass of a REBEKA rod bundle
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
Figure
54
55
56
57
58
59
60
61
62
63
64
65
-VII -
Gladding and heater sheath temperatures measured upstreamand at the bundle midplane in the blocked rod cluster of aREBEKA rod bundle
Gladding and heater sheath temperatures measured downstreamof the bundle midplane in the blocked rod cluster of aREBEKA rod bundle
Radial noding scheme of the FEBA test section for GOBRA-TFcalculations
Axial noding scheme of the FEBA test section (fluid cells)for GOBRA-TF calculations
Initial axial temperature profiles of claddings and housing(SEFLEX test No. 03)
Flooding parameters (SEFLEX test No. 03)
Gomparison of measured and calculated center rod claddingtemperatures (SEFLEX test No. 03)
Gomparison of measured and calculated center rod claddingtemperatures downstream of the bundle midplane grid spacer(SEFLEX test No. 03)
Gomparison of measured and calculated grid spacer temperatures (SEFLEX test No. 03)
Comparison of measured and calculated housing temperatures(SEFLEX test No. 03)
Comparison of measured and calculated quench front progression(SEFLEX test No. 03)
Comparison of measured and calculated quench front progression(FEBA test No. 223, SEFLEX tests No. 03, 05, and 07)
113
114
115
116
117
118
119
120
121
122
123
124
-VII I-
LISTING OF TABLES
Tab1e
1
2
3
4
5
FEBA-program: Bund1e geometry of test series I through VIIIAxial arrangement of grid spacers and f10w b10ckages
SEFLEX-program: Bund1e geometry of test series 1 through 4Axial arrangement of grid spacers and flow blockages
Test matrix of the SEFLEX-program
Quench times of grid spacers in unb10cked rod bund1es asfunction of ref100d conditions
Grid spacer effects on cladding temperature and surfaceheat f1ux in unb10cked rod bund1es as function of ref100dconditions
6
7
19
36
37
-IX-
PREAMBLE
This report is an overall summary of an experimental investigation which is a
part of the German LWR safety program. Within the framework of this program
the Kernforschungszentrum Karlsruhe (KfK) started the Project Nuclear Safety
(PNS) in 1973 to investigate the fuel rod behavior of light water reactors
(LWR) under loss-of-coolant accident (LOCA) conditions. Subjects of special
importance were: The extent of core damage during a LOCA, the consequences of
fuel rod failure on core coolability and fission product release, and the
quantification of safety margins.
Two experimental programs of the PNS, performed in the Institut fUr Reaktor
bauelemente (IRB) of the KfK, contributed to: (1) Zircaloy deformation be
havior including interaction between fuel clad ballooning and thermal-hydrau
lics in a LOCA (REBEKA) and (2) Coolability of blocked rod bundles (FEBA).
Comparison and analysis of the results of both programs indicated, that the
two different types of rods used for simulation of nuclear fuel rods showed
different behavior which needed to be quantified. Furthermore, the question
arose how far the FEBA results concerning the coolability of severe
blockages were applicable to ballooned fuel rod clusters.
The experience obtained from both programs was used for defining a new in
vestigation in 1983:
"Fuel Rod Simulator Effects in Flooding Experiments (SEFLEX)
The publication of this report as weIl as two complementing data reports (KfK
4025, KfK 4026) marks the completion of this program.
Although many individuals have contributed to this program, we wish particu
larly to acknowledge the following:
Mr. H. Schneider
Mr. S. Barth
Modifications of the FEBA facility and of the REBEKA fuel
rod simulators, management of rod bundle and test section
assemblies, instrumentation and rig operations.
Data acquisition systems, instrumentation, data pro
cessing.
Mr. F. Erbacher
Mr. A. Fiege
-X-
Consulting and technical support.
General consulting and financial support by the PNS.
The main workshop VBW/HW of the KfK and the W. Bergmaier Co. at
D-7520 Bruchsal 5 mainly for construction and instrumentation of the
fuel rod simulators.
The authors gratefully acknowledge the support from the Nuclear Power Di
vision of the Electric Power Research Institute (EPRI), USA, especially the
efforts of Dr. W. B. Loewenstein, Dr. R. B. Duffey and Dr. A. Singh, for
providing the opportunity to simulate selected FEBA and SEFLEX tests by using
the COBRA-TF computer code. This analysis was sponsored by EPRI and carried
out in cooperation with EPRI staff of the Safety Technology Department at the
Palo Alto offices.
-1-
1. INTRODUCTION
The thermohydraulics in a nuclear reactor core during a loss-of-coolant acci
dent (LOCA) of a pressurized water reactor (PWR) depends mainly on the loca
tion and the size of the break in the primary coolant system. However, the
conditions of the plant at initiation of a LOCA as weIl as the design and the
operation of the emergency core cooling system influence time dependent core
cooling conditions as weIl.
During a large break in the cold leg, the water within the primary coolant
circuit rapidly depressurizes leading to a flow reversal in the core. The
flow direction from top to bot tom of the core prevails at least towards the
end of the blowdown phase, i.e. when the system pressure corresponds to the
pressure in the containment. The upper part of Fig. 1 shows a simplified
scheme of a 4-loop steam generator system of a PWR. The lower part of Fig. 1
shows the reactor pressure vessel and the installations.
During blowdown emergency core cooling systems (ECCS) are initiated following
the transient of the system pressure. However, it is assumed that the reactor
presssure vessel is empty at the end of the blowdown phase. The low pressure
emergency core cooling system already operating is assumed to need some time
to fill up the pressure vessel until the lower end of the core is beginning
to be submerged in the rising water column (refill phase). At that moment the
main flow direction through the core again is reversed to from bottom to top,
prevailing during the reflood phase.
The nuclear decay of the fission products heats up the pellets and the clad
dings of the fuel rods until the ECCS becomes effective. Some of the fuel
rods may reach temperatures which cause clad ballooning and burst. At be
ginning of the reflood phase the cladding temperatures are assumed to be
above the Leidenfrost temperature. As the liquid level reaches the bot tom end
of the core and starts to rise around the fuel rods, complex transient heat
transfer and two-phase flow processes occur. Ahead of the quench front the
cladding temperatures are affected by the rate of steam genera ted upstream
and the thermal-hydraulic behavior of entrained liquid droplets. The effect
of this precursory cooling prevailing until quenching is characterized by a
heat transfer coefficient decreasing with distance from the quench front. The
local cladding temperature starts dropping when the precursory cooling ex-
-2-
ceeds the heat genera ted in the rods. The reflood phase is terminated when
all rods are quenched over the whole length. Figure 2 shows schematically the
pressure difference across the cladding and a range of temperature transients
for different fuel rods in a 2F-cold leg break LOCA predicted by a conserva
ti ve evalua tion model.
The investigation presented contributes to answering the questions:
How fast are fuel rods cooled down realistically under given reflood condi
tions compared to swaged heater rods?
How and to what extend do coolant channel blockages (due to ballooned fuel
rod claddings) influence the effectiveness of the reflood core cooling com
pared to blockages simulated by stainless steel sleeves?
A number of out-of-pile experiments were conducted in order to genera te heat
transfer and fluid flow data needed for the safety analysis of nuclear reac
tors. The thermal-hydraulic phenomena in unblocked as weIl as blocked rod
bundle geometries were examined in reflood experiments such as FEBA [1],
FLECHT-SEASET [2], THETIS [3], CEGB blockage tests [4], SCTF [5], CCTF [6]
etc. The objectives of all these bundle tests have been to provide experimen
tal reflood heat transfer and two-phase flow da ta in simu1ated PWR geometries
for postulated LOCA conditions. The measured data were used to develop and
validate physical models for computer codes providing qualified analytical
tools for calculating realistic peak cladding temperatures and safety margins
for unblocked and blocked bundle configurations.
Most of the experiments performed so far to understand the quench front pro
gression and heat transfer in rod bundles were carried out using "solid-type"
electrically heated rods for simulation of nuclear fuel rods. Such rods are
characterized by a stainless steel cladding and a close thermal contact bet
ween cladding and the electric insulation filler material containing the
embedded heating element.
However, during in pile tests such as the OECD Halden Reactor Project, it was
observed that nuclear fuel rods, which are characterized by heat generating
fuel pellets stacked in a Zircaloy tube with a radial gap between pellets and
cladding, were quenched substantially earlier than electrically heated rods
with a close contact between filler material and stainless steel cladding [7,
8, 9]. In the same project, REBEKA fuel rod simulators with agas filled gap
-3-
between alumina pellets and Zircaloy claddings were observed to simulate
closely the actual fuel rod behavior during a LOCA [10]. Similar results were
obtained from NRU in pile tests [11].
Figure 3 shows a comparison of the transient cladding temperatures measured
during four different tests carried out in the Halden Boiling Water Reactor
(HBWR) under nearly identical test conditions. The temperatures were measured
at an axial level of about 600 mm from the bot tom end of the heated rod
length of the nuclear rods and the SEMISCALE heater rods. Although limited in
their validity, since three SEMISCALE heater rods of the seven rod bundle
failed, the comparison indicates that du ring the blowdown and heatup phases
the temperatures of the nuclear and the electrically heated rods essentially
overlap each other. However, quenching of the SEMISCALE rods was significant
ly delayed compared to the nuclear rods. This behavior was confirmed when the
SEMISCALE rod bundle was rebuilt and the test se ries was repeated.
Figure 4 shows a comparison of the transient cladding temperatures as mea
sured at the peak power level of the nuclear fuel rods and the electrically
heated REBEKA fuel rod simulators. Plot ted are the envelopes of all tempera
ture readings measured at the axial level mentioned. The REBEKA fuel rod
simulators duplicate the temperature and quenching behavior of nuclear rods,
as can be seen. The REBEKA fuel rod simulators have been developed for inves
tigation of the plastic deformation behavior of pressurized Zircaloy-4 clad
ded fuel rod simulators under LOCA conditions. Results of single rod, full
length 5 x 5 rod bundle and 7 x 7 rod bundle tests are summarized in Ref. 12.
Investigating the effects of cladding surface thermocouples and electrical
heater rod design on quench behavior using REBEKA fuel rod simulators and
FEBA heater rods, the different behavior of both the types of rods during
simulated reflood conditions, known qualitatively, had been confirmed [13].
Furthermore, the influence of thermal properties of different cladding mate
rials on the heat transfer and rewetting behavior was observed in experiments
using single rods or tubes of stainless steel or Zircaloy, respectively,
under falling film and bottom reflood conditions [14]. Similar bench-type
reflood experiments were carried out with a 4-rod bundle to study the quench
behavior of stainless steel and Zircaloy claddings [15, 16]. In the frame of
the "Indirect Action Research Programme" of the Commission of the European
-4-
Communities upon "Safety of Thermal Water Reactors" the effects of cladding
material and pin composition upon rewetting and quench phenomena were inves
tigated [17, 18, 191.
The results of the different experiments confirmed qualitatively that the
safety analysis based on reflood tests performed with conventional gapless
heater rods overpredicts the extent of core damage. Approaching the safety
margin quantitatively, computer code models for description of the thermal
response of different fuel rod simulators need to be improved and tested. The
validation of such models implemented in reflood codes needs experimental
da ta obtained from different experiments. For a strict comparison of the
effects of the design and the composition of different rods, such experiments
should be performed under identical conditions varying only the composition
of the fuel rod simulators.
Therefore the purpose of this investigation is to address the following open
questions concerning bottom reflooding:
- Reflood timing and resultant peak cladding temperatures in fuel rod bundles
under given reflood conditions in comparison to bundles of electrically
heated rods, mostly used for out-of-pile experiments.
- Coolability of flow blockages caused by ballooned fuel rod claddings in
comparison to blockages simulated by sleeves fixed on the outer surface of
conventional heater rods.
2. S E F LEX REFLOOD PROGRAM
-5-
The aim of the SEFLEX (Fuel Rod ~imulator !ffeets in Flooding Experiments)
program has been to quantify the influenee of the design and the physieal
material properties of different fuel rod simulators on queneh front progres
sion and heat transfer in unbloeked as weIl as bloeked rod bundles during the
reflood phase of a LOCA in a PWR.
Foreed feed bot tom injeetion reflood tests have been performed using bundles
of 5 x 5 REBEKA fuel rod simulators eharaeterized by Zirealoy eladdings,
alumina pellets and agas filled gap between eladding and pellets. For the
tests the FEBA test faeility has been used. These tests performed under va
rious reflood eonditions and the eomparison of the results with eorresponding
tests of the FEBA program represent the SEFLEX program.
The FEBA tests were performed with swaged, solid type heater rods without gap
between stainless steel eladding and filler material. Separate effeet tests
were earried out in eight test series with the objeetive to measure and to
evaluate thermal-hydraulie behavior of grid spaeers and of unbloeked versus
bloeked rod bundle geometries with and without bypass. Flow bloekages simula
ting ballooned fuel rod eladdings were aehieved with sleeves of stainless
steel attaehed to the rods. The initial and reflood eonditions varied were
repeated systematieally from series to series as elose as experimentally
possible to isolate the different geometrieal effeets. The bundle eonfi
gurations tested are listed in Table 1.
For the eonduetion of the subsequent SEFLEX tests the main reflood parameters
of the FEBA test program have been maintained; only the bundle of 5 x 5
eonventional heater rods had been replaeed by a bundle of 5 x 5 REBEKA rods
whieh more elosely represented the features that exist in the aetual fuel rod
design. The influenee of the eonduetivity of the gap between Zirealoy
eladding and pellet has been investigated replaeing the helium gas filling by
argon gas filling. Helium is the filling gas of nuelear fuel rods at be
ginning of life time. The heat eonduetivity of argon eorresponds to that of
the fission gas mixed with the helium after high fuel burnup. As for the
FEBA program unbloeked bundle tests served as base line tests. For bloeked
bundle tests a 90 pereent flow bloekage at 3 x 3 rods of the 5 x 5 rod bundle
was applied having identieal loeation and outer shape as that of the FEBA
-6-
Table I
FEBA-program: Bundle geometry of test series I through VIII.Axial arrangement of grid spacers and flow blockages.
Iml
rid Spacer
lockage
,!,!dl!>idplane
G/
/3I· . fl-l B
M
62%.c~ Eg'E..--' '"'"" '"
1L ,**=1=!=10 .w~ cI=t=1~ cI=t=1~ ~1=J=l, #I=J+
Blockage RatioTest Series 11
90% 90% ... 62%V VI
Series I: Baseline tests with undisturbed bundle geometry; seven grid spacers.
Series 11: Investigation of the effects of a grid spacer; without grid spacerat the bundle midplane.
Series 111: Investigation of the effects of a 90% flow blockage with bypass;blockage at the bundle midplane of 3 x 3 rods placed in the cornerof the 5 x 5 rod bundle; without grid spacer at the bundle midplane.
Series IV: Investigation of the effects of a 62% flow blockage with bypass;blockage at the bundle midplane of 3 x 3 rods placed in the cornerof the 5 x 5 rod bundle; without grid spacer at the bundle midplane.
Series V: Investigation of the effects of a 90% flow blockage with bypass combined with grid spacer effects; blockage immediately upstream of thebundle midplane at 3 x 3 rods placed in the corner of the 5 x 5 rodbundle; grid spacer at the bundle midplane.
Series VI: Investigation of the effects of 90% and 62% flow blockages with bypass combined grid spacer effects; 90% flow blockage immediately upstream of the bundle midplane; 62% flow blockage immediately downstream of the bundle midplane; both blockages at the same 3 x 3 rodsplaced in the corner of the 5 x 5 rod bundle; grid spacer at thebundle midplane.
Series VII: Investigation of the effects of a 62% flow blockage without bypass;blockage at the bundle midplane of all rods of the 5 x 5 rod bundle.
Series VIII: Investigation of the effects of a 90% flow blockage without bypass;blockage at the bundle midplane of all rods of the 5 x 5 rod bundle.
-7-
Tab1e 2
SEFLEX-program: Bund1e geometry of test series 1 through 4.Axial arrangement of grid spacers and f10w blockages.
VGrid Spacer
VBlockage
f..-. Bund!e__Midplane
Gas Filling Helium
Blockage RatioTest Series
Argon
2
Helium
90%
3
Argon
90%
4
effects of rod clad properties, conducgaps, and grid spacers.test series I and SEFLEX test series 2.
Series 1: Rods with helium-filled gapsalumina pellets; undisturbedspacers.Investigation of thetivity of gas filledComparison with FEBA
between Zircaloy claddingsbundle geonetry with seven
andgrid
Series 2:
Series 3:
Series 4:
Rods with argon-filled gaps between Zircaloy claddings andalumina pellets; undisturbed bundle geometry with seven gridspacers.Investigation of the effects of rod clad properties, conductivity of gas filled gaps, and grid spacers.Comparison with FEBA test series I and SEFLEX test series 1.
Rods with helium-filled gaps between Zircaloy claddings andalumina pellets; 90% flow blockage with bypass; blockage atthe bundle midplane of 3 x 3 rods placed in the corner of the5 x 5 rod bundle; without grid spacer at the bundle midplane.Investigation of the effects of rod clad properties, conductivity of gas filled gaps, grid spacers, and flow blockage.Comparison with FEBA test series 111 and SEFLEX test series 4.
Rods with argon-filled gaps between Zircaloy claddings andalumina pellets; 90% flow blockage with bypass; blockage atthe bundle midplane of 3 x 3 rods placed in the corner of the5 x 5 rod bundle; without grid spacer at the bundle midplane.Investigation of the effects of rod clad properties, conductivity of gas filled gaps, grid spacers, and flow blockage.Comparison with FEBA test series 111 and SEFLEX test series 3.
-8-
tests. However, the flow blockages were realized by artificially ballooned
Zircaloy claddings surrounding the pellet column.
The separate effect tests were carried out in four test series to measure and
to evaluate the influence of four major factors on the reflood heat transfer
and rod quenching:
- Rod clad properties
- conductivity of the gap between pellets and cladding
- grid spacers
- flow blockages.
The bundle configurations tested are listed in Table 2. The SEFLEX tests were'
conducted using REBEKA rod bundles in the FEBA test facility to minimize the
influence of the boundary conditions of different test rigs. The initial and
reflood conditions selected for the FEBA program were repeated as close as
experimentally possible for the comparison of the difference in the behavior
of the two rod designs on the basis of two-phase flow heat transfer phenomena
of SEFLEX test series 1 through 4 and FEBA test series land 111.
-9-
3. TEST FACILITY
The FEBA test facility was designed for aseparate effect test reflood pro
gram involving a constant flooding rate and a constant back pressure to allow
investigation of the influence of grid spacers and coolant channel blockages
independently of system effects. Since, the design of fuel rod simulators
represents an experimental parameter similar to that of design and location
of grid spacers or coolant channel blockages, the FEBA test facility as well
as the operational procedure and the measurement technique of the FEBA tests
have been maintained for the SEFLEX tests. Modifications necessary for re
placing the 5 x 5 FEBA rod bundle by bundles of 5 x 5 REBEKA rods are
described in Section 3.1.
3.1 Test Loop and Bundle Housing
Figure 5 shows schematically the FEBA test loop with its main components.
lt is a forced flow bot tom injection reflood facility with a back pressure
control system. Coolant water is atored in a tank (3). During operation,
coolant is pumped (4) through a throttle valve (7) and a turbine meter (8)
into the lower plenum region (10) of the test section (11). The coolant flow
may be directed either upwards through the test assembly, or through the
lower plenum (10) and water level regulation valve (9) back into the water
supp1y. When reflood is initiated, coolant water rises in the test assembly
and two-phase flow results when water reaches the hot zone of the fuel rod
simulators. Entrained water droplets are transported upwards by the steam
flow and may impinge on the steam water separator (13) placed above the test
assembly. The liquid separated from the steam then drains into a collecting
tank (17), where the water content is continuously measured. Steam passes
around the droplet deflector and is then flowing through a buffer tank (19)
and the back pressure control valve (10) to the atmosphere. A large external
steam supply is connected to the buffer to heat up the total system and the
buffer contents, and to maintain the system pressure.
For the performance of the FEBA test se ries [1], the heater rod instrumenta
tion, which was completely embedded in the rod claddings, did exit from the
lower end of the rod assembly as did the electric power connections for the
heater rods. However, the instrumentation of the sleeve blockages was led to
the top end of the housing such that the lead outs attached to the rod sur-
-10-
faces did not influence the two-phase mixture rising from the bot tom.
For the performance of the SEFLEX test series, the heater rod instrumentation
(15) and the electric power connections (14) for the heater rods were led out
from the upper plenum (12). Therefore, the upper plenum (12) and the steam
water separator (13) were modified as weIl as the lower plenum (10) where the
REBEKA fuel rod simulators were filled with helium or argon gas, respective
ly, (21).
Figure 6 shows a photography of the FEBA test rig with its main components
modified for the conduction of the SEFLEX tests.
Figure 7 shows a cross sectional view of the FEBA and the REBEKA rod bundle,
respectively, placed in a square stainless steel (Standard No. 1.4571, ASTM
410) housing having an inner edge length of 78.5 mm and a wall thickness of
6.5 mm. The reasons for the use of a thick-walled housing were:
- To simulate surrounding heat generating hot rods by having sufficient heat
storage in the wall prior to the individual tests (see Section 3.6).
- To facilitate assembling of the test rig.
- To allow easy penetration of the wall for instrumentation of the bundle
with fluid thermocouples (see Section 3.4).
The dimensions of the housing inner cross section had been so chosen that the
5 x 5 rod bundle array and an infinite bundle were to have the same subchannel
hydraulic diameter dH:
= 4A--C 13.47 mm
where A: flow area; C: wetted perimeter.
The outer diameter of the rods was 10.75 mm and the rod pitch 14.3 mm for
both, the FEBA heater rods as weIl as the REBEKA fuel rod simulators. Further
dimensions of the rods end the bundles, respectively, are described in Section
3.2. Original PWR grid spacers were attached to the rods by friction. They
were sliding in the bundle housing in axial direction when relative motion
between rod bund1e and housing occurred. The rods were bol ted to the top of
the test section. The lower ends of the rods were al10wed to hang free for
-11-
FEBA as weIl as SEFLEX tests. For replacing the FEBA rod bundle by the REBEKA
rod bundle the upper as weIl as the lower plenum were modified. The bundle
housing was identical for both the bundles. Figure 8 shows a cross sectional
view of the test section with the insulation at the housing outside for
reduction of the heat losses to the environment.
The modification of the upper plenum is shown in Fig. 9. Whilst the FEBA rods
were bol ted to the top grid plate, the REBEKA rods were bol ted to the top of
the upper plenum, and penetrated the top grid plate through square holes
which provided the same total cross section for coolant through flow as the
circular holes between the FEBA rods for the FEBA grid plate. The two-phase
flow leaving the rod bundles had to cross the REBEKA rods in radial direc
tion. The cross flow in that portion of the plenum has probably led to
slightly increased droplet evaporation compared with the FEBA flow condi
tions, i.e. without rods at that place. However, any effect of additional
evaporation was rather small because of the short flow path along and across
that - unheated - portion of the REBEKA rod bundle. After separation of the
water from the steam, the flow path of the water to the water collecting tank
was identical for both designs. The conditions for the steam flow, after
separation from the water, did not affect the flow conditions upstream, since
the pressure drop between bundle exit and buffer was very small.
Figure 10 shows the modification of the lower plenum. The FEBA rods penetrated
the bot tom of the plenum which was covered by a water film controlling the
temperature of the lower plenum including the O-ring sealings to the tempera
tu re of the feedwater during heat up of the bundle. The REBEKA rods were
hanging in a water-filled plenum. The water level was at the same elevation
for both designs, and the water temperature was controlled to that of the
feed water during the test. Therefore, no influence of the modification of
the plenum on the reflood conditions was observed.
3.2 REBEKA Fuel Rod Simulator and FEBA Heater Rod
Fuel rod simulators of PWR dimensions were used to simulate the nuclear fuel
rods. Figure 11 shows the cross section of a gapless FEBA heater rod which has
an outer diameter of 10.75 mm. A spiral wound heating element of NiCr 80 20
(ASTM B 344-60) is embedded in the electrical insulator (magnesium oxide),
and then encapsulated in the clad of NiCr 80 20 which has a wall thickness of
-12-
1.0 mm. In eontrast to a nuelear fuel rod with a Zirealoy eladding and agas
filled gap, this heater rod is a solid type widely used for thermal-hydraulie
tests. A elose thermal eontaet between eladding and filler material results
from swaging of the rods. More details including a working drawing are con
tained in Ref. [1].
Figure 12 shows the cross section of a REBEKA fuel rod simulator. This fuel
rod simulator eonsists of an eleetrieally heated rod of 6.02 mm outer diameter
placed in the center of annular alumina pellets simulating fuel pellets. As
for a nuelear rod, the pellets are encapsulated in the Zircaloy tube with a
wall thickness of 0.725 mm. By pressurization of the rod with filling gas the
gap between pellets and cladding is filled with helium or argon, respeetive
ly, to study the influenee of the gap eonductivity on the reflood behavior.
The thiekness of the Zirealoy eladding, the helium filling and the nominal
gap width of 0.05 mm of a REBEKA rod are identieal to a nuclear fuel rod at
the beginning of li fe time. Heater rod and alumina pellets represent about
110 pereent of the heat capaeity of fuel pellets. The heat eonductivity of
argon eorresponds roughly to that of the fission gas mixed with the helium
after high fuel burnup. Figure 13 shows a working drawing of a REBEKA fuel
rod simulator of nominal geometry modified for the SEFLEX tests.
The remaining eharaeteristics of both types of fuel rod simulators, FEBA rod
as well as REBEKA rod, were the same. Figure 14 represents an axial layout of
the fuel rod simulators. The eosine power profile of the rods with a heated
length of 3900 mm were approximated by seven steps of specifie power. The
axial power profile was flat with a peak-to-average ratio of 1.19. Seven grid
spacers without mixing vanes (height 38 mm) were installed a 545 mm axial
intervals throughout the bundles.
3.3 Bloekage Design
The influenee of the size and the shape of various coplanar bloekages on
loeal reflood heat transfer was already examined as part of the FEBA program.
For most of the geometries, improved eooling was found downstream of sueh
uniform bloekages eompared with base line tests without bloekages eondueted
under the same flooding eonditions. Only a 90 pereent bloekage with bypass
led to about the same peak eladding temperatures downstream of the bloekage
eompared with unbloeked bundle data [1]. The most signifieant differenee
-13-
between the temperatures so compared occurred after turnaround. Downstream
of the blockage the temperatures decreased more slowly than in the unblocked
portion of the bundles and a delayed quenching was observbed. However, the
FEBA blockage configuration using sleeves was a compromise between flow
channel constriction caused by ballooned claddings and the technical feasibi
lity of such a simulation having sufficient li fe time for repeated tests. To
examine and to isolate properly the blockage effects of FEBA and REBEKA rod
bundles with 90 percent flow blockage with bypass, identical outer dimensions
of the blockage geometries had to be selected.
The coplanar 90 percent blockage configuration with bypass used for the FEBA
tests is shown in Fig. 15. Hollow sleeves of stainless steel were used to
simulate ballooned claddings. The sleeves were attached to the rods. For the
simulation of the heat resistance between pellets and lifted cladding a gap
of 0.8 mm width filled with stagnant steam was provided between the outer
surface of the FEBA rod and the inner surface of the sleeve. In addition,
side plate devices were placed between the sleeves of the peripheral rods and
the housing walls for constriction of the coolant subchannels between the
3 x 3 rod cluster and the housing.
Figure 16 shows a sectional view of the rod bundle with coplanar 90 percent
flow blockage and bypass investigated in SEFLEX test series 3 and 4. The flow
blockage was placed symmetrically to the bundle midplane (axial level
2025 mm) gene rating a local coolant channel constriction of nine subchannels
of the 3 x 3 rod cluster. The balloons had an axial extension of 180 mm
including the conical ends. The length of the 90 percent flow channel con
striction amouted to 65 mm.
The outer shape and size of the blockages, i. e. the geometries and the
surfaces exposed to the coolant, were the same for both, the SEFLEX and FEBA
arrays. However, the heat capacities and the radial compositions underneath
the cooled surfaces were different from each other.
Figure 17 shows a working drawing of REBEKA fuel rod simulators with artifi
cally ballooned claddings as weIl as the instrumentation of the individual
simulator types with thermocouples in axial and circumferential directions.
The instrumention is described in detail in Section 3.4. To model a 90 per
cent flow blockage with bypass accordingly to the corresponding FEBA blocked
-14-
bundle configuration of test series 111, three types of artificially
ballooned Zircaloy claddings with different outer shape were produced. This
was necessary to avoid side wall blockages as used for the FEBA tests. The
cross sections of simulator type a and f, shown in Fig. 17, indicate the
geometries of regular ballooned rods and ballooned rods placed at the housing
wall of the 3 x 3 rod cluster, respectively. A third type of simulator was
used to constrict the coolant subchannel in the corner of the housing (see
cross section of rod No. 21 shown in Fig. 21). The required outer shape of
the ballooned Zircaloy claddings was produced in a furnace by heating up the
pressurized cylindrical tubes placed in correspondingly shaped molds. Subse
quently, the ballooned claddings were cooled down very slowly to avoid any
bursting or collapsing. During the reflood test series no deformation of the
ballons took place. The blockage array after performance of the test series
is shown in Fig. 18.
3.4 Instrumentation
Most part of the SEFLEX instrumentation consisted of thermocouples (Chromel
Alumel), since cladding (TS), heater sheath (TZ), grid spacer (TA), fluid
(TF) and housing (TK) temperatures were to be measured at various positions.
Figure 19 shows a schematic diagram of the axial levels of the thermocouples,
the pressure and the differential pressure measuring positions. This diagram
enables to relate the measuring positions to the blockage and the grid spacer
positions as weIl as to the different specific power zones.
The cladding temperatures were measured with 0.36 mm sheath outer diameter
thermocouples having insulated junctions. For test series 1 and 2 these ther
mocouples were embedded in grooves from the individual measurement position
up to the top end of the rods. The grooves were milled into the outer surface
of the Zircaloy claddings. The grooves were closed by peening over to avoid
any disturbance of the coolant flow. For test series 3 and 4 these thermo
couples were embedded in grooves of 20 mm length, which were milled into the
outer surface of the Zircaloy claddings. The short grooves were closed by
peening OVer as weIl. The remaining lead outs were attached to the outer
surface of the Zircaloy claddings by very small and thin straps of Zircaloy
which were spot welded to the claddings. For the instrumentation of the
ballooned portion of the claddings the same method was applied with the dif
ference that the thermocouple ti ps were not embedded in grooves but also were
-15-
attached to the rods by using straps (see Fig. 18). This external instrumen
tation was necessary with respect to the reduced wall thickness of the bal
loons. Aseparate effects experiment program [13] conducted in the LOFT Test
Support Facility (LTSF) at the ldaho National Engineering Laboratory (INEL)
with REBEKA fuel rod simulators to evaluate the effect of cladding external
thermocouples on the quench behavior indicated: "Cladding external thermo
couples have a negligible effect on the cooldown rate and quench behavior of
a REBEKA fuel rod simulator over the range of LOCA-type, high pressure ther
mal-hydraulic reflood condi tions examined."
As indicated in Figs. 12 and 21, some of the heater rods placed in the center
of the alumina pellets were instrumented for the conduction of SEFLEX test
series 3 and 4. The temperatures of the heater rod sheaths with an outer
diameter of 6.02 mm were measured with 0.25 mm sheath outer diameter thermo
couples having insulated junctions. These thermocouples were embedded in
grooves which were milled into the outer surface of the lnconel rod sheath.
The grooves were closed by peening over to keep the thermocouples at the
provided measuring positions and to maintain the geometry of the alumina
pellets. The leads were led out to the top end of the rod bundle close to the
insulated connections of the electrical rod power supply.
The grid spacer temperatures were measured with 0.5 mm outer sheath diameter
thermocouples having insulated junctions. The tips of these thermocouples
(see indication TA on Figs. 20 and 21) were placed each at about 2 mm from
the leading and the trailing edges, respectively, of the grid spacers. The
thermocouples were attached to the grid spacers by very small and thin straps
of stainless steel which were spot welded to the surface of the 0.4 mm thin
grid spacer sheetings. The leads were led from the subchannels surrounding
the central rod via trailing edge to the peripheral subchannels to avoid as
far as possible any disturbance of the coolant flow.
The fluid temperatures were measured with unshielded thermocouples of 0.25 mm
outer sheath diameter (see indication TF on Figs. 20 and 21). The junctions
protruded into the center of the individual bundle subchannels. The ability
of such fluid thermocouples for measuring steam temperature is demonstrated
in Ref. L
-16-
The housing temperatures were measured with 0.5 mm outer sheath diameter
thermocouples (see indication TK on Figs. 20 and 21) placed from the outside
close to the inner surface of the 6.5 mm thick housing wall.
Pressures and pressure differences were measured with pressure transducers.
In addition to the inlet and outlet pressure, the pressure differences were
measured along the entire bundle length, along both the lower and upper por
tion of the bundle as weIl as along a short section at the bundle midplane.
The flooding rate was measured with a turbo-flowmeter. The amount of water
carried over was measured continuously by apressure transducer at the water
collecting tank.
All data were recorded with a scan frequency of 10 cycles per second using
NEFF amplifiers, a PDP-ll mini-computer and disks for fast data recording.
3.5 Operational Procedure
The investigation of separate effects of core reflood during a PWR LOCA re
quires weIl defined system parameters for each test. The quality of the com
parison among the tests depends mainly on the repeatability of the individual
tests. Therefore, with respect to the real sequence of events during a LOCA,
the following modification of the heat up period during refill of a reactor
vessel had been made for the FEBA tests and maintained for the SEFLEX tests:
For about two hours prior to reflood, the fuel rod simulators were heated in
stagnant steam to the described initial cladding temperature, using a low rod
power. In the mean time the test housing was being heated up passively to the
desired initial temperature by radiation from the rods. This led to a wall
(6.5 mm thick) heat content of approximately the same as that of half a row
of heater rods including the heat input during a test (rod power). The aim of
choosing the "active wall" was to prevent premature quenching of the wall
relative to the bundle quench front progression. The hot steam film at the
surface of the wall ncts somewhat like a layer of insulation for the two
phase flow in the bundle sllbehannels. The "passive wall" design using a thin
wall of low heat capacity i. all alternative method which allows fast heat up
of the bundle and thC' hOlls1 ng. However, premature quenching may occur influ
encing the bundle heat transfer conditions. Furthermore, it complicates
instrumentation and assembling.
-17-
Reflood was initiated by closing the water exit and the steam inlet valve at
the lower bundle plenum and the drain valve of the water collecting tank (see
Fig. 5). The bundle power was stepped up to the controlled decay heat tran
sient, i.e. 120 percent ANS-Standard 40 seconds after shut down of a reactor
for most of the tests. About 30 seconds prior to reflood the data recording
system was started.
-18-
4. TEST MATRIX
The main test parameters varied are shown in Table 3:
- Bundle geometry
- Gap gas filling
- Flooding rate given as flooding velocity, i.e. the velocity of the rising
water level in the cold bundle
- System pressure.
For the comparison of the reflood behavior of the two rod bundles consisting
of either 5 x 5 FEBA or 5 x 5 REBEKA fuel rod simulators, the SEFLEX tests
were carried out for flooding velocities of 3.8 cm/s and 5.8 cm/s (in the
cold bundle) and system pressures of 2.1 and 4.1 bar. The test operational
procedures were also similar (see Section 3.5). The power input was stepped
up, when the rising water level reached the bot tom end of the heated bundle
length, to about 200 kW and decreased corresponding to the 120 percent ANS
decay heat transient. Flooding velocity, system pressure, and feedwater tem
perature were kept constant during each test. The internal gas pressure was
controlled to about 1 bar overpressure with respect to the system pressure.
010101010o (1ItHlli) 0 ••••••010lGI81G~W/W/Wß000Wffi
SEFLEX-ProgramTest Series 1 and 2
@)@)@)C9@)®(J)(J)C9(J)@@@@)@)@@@(J)@)@@@@).
SEFLEX-ProgramTest Series 3 and 490% BI oekageBa I I ooned cl add i ngs
Table 3
Test matrix cf the SEFLEX-program
•••••••••••••••I•••••
•••••0(00'#//W$ß/(%
FEBA- Prog ramTest Series I
••••••••••i1RBlRBlrl ••:' KJ] KJ] RBl ••ii"IjIjRBl••
FEBA- Prag ramTest Series 1I190% BI oekageSieeve blockages
<0Prog ram Test Test-No. Rod Design Cladding Gap Gas Flooding System Feedwater Reference Tests
Series Materia I F i I I i ng Velocity Pressure Tempe ra tu re FEBA-Test SEFLEX-Testcm/s ba r oe
SEFLEX 1 05 REBEKA Z i rca loy He [iurn 3.8 2.1 40 No. 223 No. 07SEFLEX 1 03 REBEKA Z i rca loy He I ium 3.8 4.1 40 No. 216SEFLEX 1 06 REBEKA Z i rca loy He I ium 5.8 2.1 40 No. 218SEFLEX 1 04 REBEKA Z i rca IOy He I i um 5.8 4.1 40 No. 214
SEFLEX 2 07 REBEKA Zircaloy Argon 3.8 2.1 40 No. 223 No. 05
FE8A 1 223 FE8A SS gapless 3.8 2.1 40 No. 05 and 07FEBA 1 216 FEBA SS gapless 3.8 4.1 40 Ne. 03FEBA 1 218 FEBA SS gapless 5.3 2.1 40 No. 06FEBA I 214 FEBA SS gapless 5.8 4.1 40 No. 04
SEFLEX 3 32 RE8EKA Zi rca loy He I ium 3.8 2.1 40 No. 241SEFLEX 3 35 REBEKA Z i rca loy He I j um 3.8 4.1 40 No. 239
SEFLEX 4 33 REBEKA Z i rca loy Argon 3.8 2.1 40 No. 241SEFLEX 4 34 REBEKA Zi rca IOy Argon 3.8 4.1 40 No. 239
FEBA 1I1 241 FE8A SS gapless 3.8 2.1 40 No. 32 and 33FEBA 111 239 FEBA SS gapless 3.8 4.1 40 No. 34 and 35
-w-
5. SEFLEX REFLOOD TEST RESULTS AND COMPARISON WITH FEBA DATA
The results of the SEFLEX test series 1 through 4 obtained with 5 x 5 REBEKA
rod bundles [20, 21] are summarized and compared with results of correspond
ing FEBA tests [1]. This comparison, comprised within the SEFLEX program,
deals with the overall bundle behavior, the grid spacer effects, the blockage
effects and quench phenomena discussing data measured and evaluated.
Performing the SEFLEX program most of the individual results have been pub
lished successively [22 through 361.
5.1 Bundle Behavior
At initiation of the reflood phase the cladding temperatures of the rods are
highly above the Leidenfrost temperature. As the liquid level reaches the
bot tom end of the bundle and starts to rise around the rods, complex heat
transfer and two-phase flow processes occur. Ahead of the quench front the
cladding temperatures are affected by the rate of steam generated upstream
and the thermal-hydraulic behavior of entrained liquid droplets. The effect
of this precursory cooling prevailing until quenching is characterized by a
heat transfer coefficient decreasing with axial distance from the quench
front. The local cladding temperature starts to decrease when the precursory
cooling exceeds the heat genera ted in the rods.
For a proper comparison, the test conditions overtaken from the FEBA tests
were systematically repeated for the individual tests of the SEFLEX series,
e.g. initial cladding temperatures, power input, flooding rate, system pres
sure, inlet water subcooling etc.
The reflooding behavior of the two bundles consisting of either 5 x 5 FEBA or
5 x 5 REBEKA rods is significantly different. Figure 22 shows cladding tempe
ratures versus time of three test runs performed with a flooding velocity of
3.8 cm/s and a system pressure of 2.1 bar in a FEBA rod bundle (square sym
bols), in a REBEKA rod bundle with helium-filled gaps (circular symbols), and
in a REBEKA rod bundle with argon-filled gaps (triangular symbols). The tem
perature transients were measured at four different axial levels, in each
case at about the half way between two grid spacer positions. Upstream of the
bundle midplane (Plot A: axial level 2225 mm), the influence of the different
-21-
rod design on the peak cladding temperature is not yet pronounced, but the
quench times of the REBEKA rods are almost 100 s, i.e. about 30 percent,
shorter than that of the FEBA rods. Downstream of the bundle midplane (Plot
B: axial level 1680 mm; Plot C: axial level 1135 mm; Plot D: axial level 590
mm), these differences become more pronounced towards the top end of the
bundles.
The reasons for the lower cladding temperatures and the faster quench front
progression of the REBEKA rod bundles are the lower heat capacity of the
Zircaloy claddings and the more pronounced thermal decoupling of the cladding
from the heat source, compared with the thick stainless steel claddings of
the swaged FEBA rods.
For a temperature of 500 °c the products of specific weight and thermal
capacity of Zircaloy-4 and NiCr 80 20 amount to [42J:
=
=
NiCr 80 20
Zircaloy-4
and the heat
NiCr 80 20
Zircaloy-4
pocp
pOcp
conductivities amout
A =
A =
4.17 Wos/(cm'OK)
2.15 Wos/(cm'OK)
to [42J:
0.21 W/(cmOK)
0.19 W/(cmOK)
Taking into account the following cladding dimensions
FEBA (NiCr 80 20) d = 10.75 mm0
d. = 8.65 mm1
REBEKA (Zircaloy-4) d = 10.75 mm0
d. = 9.30 mm1
the internal energies per unit length are given byFEBA cladding
REBEKA cladding
E =
E =
1.33 WOs/K
0.49 WOs/K
A comparison of the internal energies stored in FEBA and REBEKA rod claddings
leads to a ratio of about 2.7.
The assumptions of a heat transfer coefficient of 3.0 W/(cm'OK) between filler
material and FEBA rod cladding, a gap width of 0.05 mm between alumina pellets
and REBEKA rod cladding, and a filling gas temperature of 500°C in REBEKA rods
result in the following ratios of the thermal conductances at the interfaces:
-22-
FEBA/REBEKA (He-filling) 5
FEBA/REBEKA (Ar-filling) 40
REBEKA (He-filling)/REBEKA (Ar-filling) 8
This comparison does not include the heat transfer by radiation between the
ceramic and the cladding.
The comparison between a nuclear fuel rod and a SEMISCALE heater rod (see
Fig. 4), shows the same trend concerning shorter quench time for Zircaloy
cladded rods with gap. However, the SEMISCALE heater rod having a similar
design as a FEBA rod did not lead to significantly higher peak cladding
temperatures than the nuclear fuel rods. This finding only is consistent with
the temperature transients measured in the lower portions of the FEBA and the
REBEKA rod bundles, respectively (see Fig. 17, Plot A). It has to be men
tioned that the heated length of the rods used in the OECD Halden experiments
was only 1500 mm and the data shown were measured 600 mm from the bot tom end
of the heated length. For the FEBA and REBEKA rods in the SEFLEX tests the
heated length amounted to 3900 mm. Since in the upper bundle portions the
peak cladding temperatures are substantially lower for REBEKA rods than for
FEBA rods, it can be assumed that the heated length of the rods used in the
OECD Halden experiments was too short for promoting a similar behavior lead
ing to lower peak cladding temperatures for e.g. the fuel rods in the OECD
Halden experiment.
Analyzing the reasons for the lower peak cladding temperatures in the REBEKA
rod bundle the question arises whether the precursory cooling is really bet
ter or the about 10 percent lower amount of heat stored in the REBEKA rods
leads to lower peak cladding temperatures during precursory cooling.
The first indication for the differences between the transient heat transfer
coefficients prevailing in the individual rod bundles can be found qualita
tively reading the temperature transients of the bundle housing. Figure 23
shows the housing temperatures measured at the axial level 1680 mm du ring the
tests discussed above. The housing, which is identical for the total of the
tests, is cooled faster using the REBEKA rod bundles than using the FEBA rod
bundle. Therefore, there is an increased reflood heat transfer for rods with
Zircaloy claddings and helium gas filled gaps, e.g. SEFLEX test No. 03, com
pared with gapless heater rods with stainless steel claddings, e.g. FEBA test
No. 223. The heat transfer again increases for increased heat resistance
-23-
across the gap, e.g. for the argon gas filled REBEKA rods, e.g. SEFLEX test
No. 07.
The heat transfer analysis quantifies the cooling conditions for the dif
ferent rod bundles. Heat transfer coefficients, rod surface heat flux, tem
perature distribution in a cross section of a rod and stored heat per unit of
length of a rod have been calculated for various locations within the bundle.
Input data for the one-dimensional inverse heat conduction calculation using
the modified and supplemented HETRAP-computer code [41] are the measured
local cladding temperature, the corresponding specific rod power, the satura
tion temperature related to the system pressure, the temperature-dependent
material properties [42, 43] as weIl as the individual rod geometry. The FEBA
rod cross section was defined by 11 radial nodes and the REBEKA rods by 14
radial nodes to obtain the transient temperature distributions in the rod
cross sections.
The FEBA heater rod consists of four concentric rings of different material
regions. Contact resistances between material regions were specified by con
stant uniform input data.
The REBEKA fuel rod simulator required to describe eight concentric rings of
different materials. Again, contact resistances between two material regions
were defined by constant uniform input data. The heat conductance of the gas
filled gaps between heater rod and pellets as weIl as pellets and cladding
was assumed to consist of two components: (1) Heat transfer due to thermal
radiation and (2) Heat transfer due to conduction in the filling gas. Since
the conduction model does not calculate the effects of power history of the
rods, the gap width specified by input includes any changes from the as-built
conditions. Therefore, it was assumed that the gap between heater rod and
pellet was reduced from a nominal width of 0.04 mm to an effective width of
0.01 mm due to thermal expansion of the heater rod in radial direction.
The numerical procedure used in the conduction solution is based on a finite
difference resistance network approach.
The rod surface heat flux transients, evaluated for the tests discussed
above, for the axial level 1680 mm, confirm increased precursory cooling in
the REBEKA rod bundles as shown in Fig. 24. For the early portion of the
-24-
ref100d phase the heat f1ux at the surfaee of the gap1ess heater rods of the
FEBA tests is lower than that of the helium fi11ed REBEKA rods of the eorres
ponding SEFLEX tests. The highest heat f1ux density is eva1uated for REBEKA
rods with argon fi11ing. The individual transients are approaehing eaeh other
with inereasing ref100d time and reaeh about the same va1ues after approxi
mate1y 200 seeonds. However, at that time the e1adding temperatures are de
ereasing for all the eases shown, i.e. after the time of peak e1adding tempe
ratures. C1adding temperatures and eorresponding surfaee heat f1uxes are
eompared in Fig. 25 using an en1arged time sea1e for the first 140 seeonds of
the ref100d phase. The transients are the same as plotted in Fig. 22 (Plot B)
and Fig. 24.
The heat stored in a FEBA rod eompared with that stored in REBEKA rods with
either helium or argon filling is shown in Hg. 26 for axial level 1680 mm.
lt has to be mentioned that for the same initial e1adding temperatures at
beginning of the ref100d phase the amount of heat stored in a FEBA rod is
about 10 pereent 1arger than that of the REBEKA rods (see va1ues at t = 0
seeonds in Fig. 26). However, for identiea1 power input app1ied for the dif
ferent bund1es, a 1arger amount of heat remains in a FEBA rod than in the
REBEKA rods du ring the ear1y portion of the ref100d phase as indieated a1
ready by the surfaee heat f1ux transients, plot ted in Fig. 24. After the
turnaround points, i.e. when the heat removal exeeeds the heat input, again
the stored heat transients deerease faster for the REBEKA rods than for the
FEBA rods. Therefore, 1ater quenehing of the FEBA rod bund1e is main1y due to
the lower ref100d heat transfer to the eoo1ant.
As soon as the rod e1addings are quenehing the heat stored in the rods drops
sudden1y. The transients are simi1ar for the FEBA rod and the REBEKA rod with
helium gas fi11ed gap. Later in time after quenehing, the amount of stored
heat remaining in both rods is about the same. The behavior is different for
REBEKA rods with argon gas fi11ed gap: The deerease of the stored heat is
slower during quenehing and, 1ate after quenehing, the amount of stored heat
remaining in the rod is signifieant1y higher than for both types of rods
mentioned before. The inereased heat resistanee aeross the argon-fi11ed gap
is responsib1e for this phenomenon. lt eou1d be drawn a pre1iminary eon
e1usion, that the higher amount of stored heat remaining after quenehing in
rods with argon-fi11ed gaps is the reason for the faster queneh front pro
gression eompared with rods with he1ium-fi11ed gaps. However, this exp1ana-
-25-
tion is not sufficient, because the surface heat flux transients are diffe
rent for both cases. For most part of the reflood period the surface heat
flux of the REBEKA rod with argon-filling is substantially higher than that
of the REBEKA rod with helium-filled gap (see Fig. 24), and consequently, the
heat removal is increased for increased heat resistance across the gap (com
pare Fig. 26, circular and triangular symbols). The effect of cladding mate
rial is excluded in this comparison, since cladding thickness and material
are identical for both tests. Comparing the behavior of FEBA rod bundles with
that of REBEKA rod bundles the effect of radial heat resistance as well as
the effects of the physical properties and the dimensions of the rod cladd
ings have to be considered. The quantitative separation of these effects
seems not to be possible using the FEBA and SEFLEX results only. But, it can
be concluded qualitatively, that both, cladding and gap effects are respon
sible for increased reflood heat transfer of REBEKA rod bundles compared with
that of FEBA rod bundles. There is a limited number of experiments [17, 181
to examine these effects. However, for a quantitative analysis of the
question whether the cladding material or the gas-filled gap is the main
parameter of influence, additional experimental and analytical investigations
are needed.
Further informations about the heat transfer and quench behavior of the indi
vidual rods used for the FEBA and SEFLEX tests give the following diagrams:
The heat transfer coefficients evaluated for axial level 1680 from FEBA test
No. 223 and SEFLEX tests No. 05 and 07 are plot ted in Fig. 27. Emphasis is
placed on the conditions at the time of quenching of the individual rods at
the elevation indicated. Besides the fact that the quench times are differ
ent, the slopes of the transient heat transfer coefficients are different as
well. For the following representations the quench times of the individual
rod sections are set t = 0 obtaining a new time scale allowing a better
comparison of the different transients. A recording window of 40 seconds, i.
e. 20 seconds before through 20 seconds after the quench fronts passed the
axial level 1680 mm, is used for the da ta plotted versus an enlarged time
scale. Figure 28 shows cladding temperatures (Plot A), surface heat fluxes
(Plot B), heat transfer coefficients (Plot C) and stored heat quantities
(Plot D) for FEBA test No. 223 and SEFLEX test No. 05 with helium-filled gaps
of the REBEKA rods. Before quenching (t < 0) the temperature of the REBEKA
rod cladding decreases faster than that of the FEBA rod, and it drops sudden
ly down to saturation temperature at the moment of quenching. However, the
-26-
"queneh temperatures" of both the types of rods are almost the same (see Plot
A). The surface heat f1ux at the REBEKA rod is somewhat higher than that of
the FEBA rod prior to quenehing as diseussed a1ready. At the moment of quen
ching this trend is reversed: The peak surfaee heat f1ux at the FEBA rod then
is about twice as high as that of the REBEKA rod (see Plot B). A comparison
of the heat transfer eoefficients (Plot C, 10garithmie sca1e!) shows again
for the REBEKA rod a somewhat higher mean va1ue prior to quenehing, signifi
eant1y higher maximum at the moment of quenching and slower decrease after
quenching than that of the FEBA rod. Fina11y, Plot D of the Fig. 28 shows
that the removal of the heat stored in a REBEKA rod is increased prior to
quneching and slight1y de1ayed immediate1y after quenching compared with the
heat release of a FEBA rod.
The conditions for a FEBA rod a1ready shown are eompared with those of a
REBEKA rod with argon gas-fi11ed gap in Fig. 29. Due to the increased gap
heat resistanee for the REBEKA rod in this case, the effects discussed before
are somewhat more pronounced. Especia11y, the heat release is de1ayed signi
ficant1y after quenching of a REBEKA rod with argon gas-fi11ed gap (see Plot
D of Fig. 29).
For different ref100d conditions, e. g. different system pressures and f100d
ing ve10cities, the trends of the quenching behavior remain the same as shown
in Figs. 30, 31 and 32. The individual quantities on1y are different.
Ana1yzing the azimutha1 temperatures of a REBEKA rod c1adding for different
ref100d conditions some information is obtainab1e eoneerning the qua1ity of
the quench front progression. The transient c1adding temperatures measured at
bund1e midp1ane and at four circumferentia1 positions of a rod are plot ted in
Fig. 33. The data are plotted again in Fig. 34 versus an en1arged time sca1e
for the recording windows in which quenching of the individual rods oecurs.
The different times, at which the four considered positions of the rod cir
cumference are quenching, indicate that the quench front is not strict1y
circu1ar. This finding is eonsistent with optica1 observations. In some eases
dryout oceurs for a short period after the who1e circumference was quenched
a1ready as shown in Fig. 34, Plot B.
-27-
Radial temperature profiles in gap1ess FEBA rods and REBEKA rod with gaps
indicate the different mechanisms of heat release du ring the period of sur
face rewetting. Figure 35 shows ca1cu1ated radial temperature profiles as a
function of time for the axial level 1680 mm. For the sudden1y increasing
coo1ing conditions c10se to the quench front, the temperature of Zirca10y
c1addings drops fast (Plot Band Plot C) and the heat stored in the remaining
portion of the rod is being removed with a certain de1ay after the tempera
ture difference ac ross the gap is estab1ished. A sudden increase of the coo1
ing conditions or the arrival of the quench front at a FEBA rod surface has to
remove more heat stored in the c1ad and in the fi11er material which are in
c10se contact (Plot A). Therefore, the c1adding temperature does not drop as
fast as for a REBEKA rod or for a nuc1ear fue1 rod. This mechanism de1ays the
quench front progression, reduces the amount of heat re1eased from the rods
per unit of time, and lowers the effectiveness of precursory coo1ing.
Figure 36 shows that the quench front progression in the REBEKA rod bund1e
with he1ium-fi11ed gaps is faster than that in the FEBA rod bund1e. The plot
shows also an increased quench front velocity for the upper bund1e portion of
REBEKA rods with argon gas fi11ing compared with the he1ium-fi11ed gaps. This
is apparent1y due to the fact that the thermal conductivity of argon is
near1y one order of magnitude lower than for helium. In both cases the inter
na1 gas pressure amounted to about 3 bar. Furthermore, it can be seen that in
the REBEKA rod bund1es the quench front progression is inf1uenced by the grid
spacers, especia11y in the upper portion of the REBEKA rod bund1e with argon
gas filled gaps.
Figure 37 shows the water carry over co11ected downstream of the bund1e exit
versus ref100d time. A higher amount of water entrained by the steam is being
evaporated within the REBEKA rod bund1e subchanne1s removing more heat per
unit of time. Less water is carried over for identica1 injection rate. Less
entrainment cou1d not exp1ain the higher rate of heat removed from the un
wetted portion of the bund1e (see stored heat as function of ref100d time
plotted in Fig. 26).
5.2 Grid Spacer Effects
The grid spacer effects on loca1 coo1ing conditions inf1uence significant1y
size and shape of c1adding ba11oons. Experimental resu1ts make evident the
-28-
deerease of the eladding temperatures downstream of the individual grid spa
cers as weIl as the interaction between thermal-hydraulies and fuel elad
ballooning [12, 27, 30, 311.
The transient axial profile of the eladding temperature during refill and
reflood of a LOCA determines the amount of loeal eladding deformation along
the rods. In general, elad ballooning oeeurs first at the axial rod seetion
where a eritieal temperature level is reaehed. That loeation and the axial
extension of ballooning is mainly the result of the axial eladding tempera
ture profile between two grid spaeer positions. This temperature profile is
determined by the thermodynamie non-equilibrium in the two-phase flow and its
interaction with the grid spaeers. The presenee of a grid spaeer enhanees
substantially the heat transfer downstream.
However, this signifieant effeet deereases on the way to the next grid spaeer
in flow direetion and leads to the development of an axial temperature pro
file with a loeal maximum immediately upstream of the individual grid spa
cers. It has been found that Zirealoy eladdings - separated by even a rather
small gas filled gap from the internal heat source - are more sensitive to
the grid spacer effeets than swaged heater rods with stainless steel eladd
ings. A typical example is shown in Fig. 38. The axial temperature profiles
are reeorded 30, 90 and 150 seeonds after initiation of reflood. After 30
seeonds the eladding temperature at and downstream of the grid spacer drops
to about 30 Klower for REBEKA rods than for gapless FEBA rods. Upstream of
the grid spacer the cladding temperatures of both the bundles are still at
the same temperature level. After 90 and 150 seeonds, respectively, the over
all differenee between both the temperature profiles inereases. This is due
to the faster overall reflood transient during the SEFLEX test with REBEKA
rods. However, at and downstream of the grid spaeer the differenees are even
more pronouneed indieating again enhaneed grid spaeer effeets in SEFLEX eom
pared wi th FEBA.
Adequate fuel rod simulation is obtained using REBEKA rod bundles with helium
filled gaps. The results are still conservative concerning spent fuel rods
with fission gas mixed with the helium gas as shown by the results obtained
with REBEKA rods with argon-filling of the gaps. Therefore, the deseription
of the grid spacer effeets is coneentrated on results obtained from SEFLEX
test series 1 using undeformed REBEKA rods with helium-filled gaps.
-29-
From the sample of tests performed under different reflood conditions, Fig.
39 shows data measured at the bundle midplane. For the set of tests performed
with system pressures of 2.1 and 4.1 bar and with f100ding ve10cities of 3.8 and
5.8 cm/s, transient c1adding and grid spacer temperatures are plot ted. The
temperatures of the grid spacers are measured near the leading edge and near
the trai1ing edge, respective1y, as described in Section 3.4.
The c1adding temperatures are obtained from thermocoup1es embedded in the
c1adding of the center rod at the bund1e midp1ane, the position of the 1ead
ing edge of the grid spacer, and 12 mm downstream of the trai1ing edge of the
grid spacer with a height of 38 mm. At initiation of ref100d the temperatures
of the c1adding as we11 as of the grid spacers are c10se to 800 oe diverging
from each other rapid1y within the first 10 seconds of the ref100d transient.
The c1adding temperatures downstream of the grid spacer then are lower than
upstream by between 50 K up to 150 K depending on the ref100d conditions.
After the initial drop the temperatures of the grid spacer are stabi1ized
(quasi for a certain time span) at about 200 K below the cladding temperature
measured near the leading edge of the grid. The level of this temperature
transient results from the radiation heat transfer from the rods to the grid,
and from heat exchange with the dispersed flow, characterized by superheated
steam and water droplets being at saturation temperature, which passes the
grid spacer. The ratio grid quench time to c1adding quench time depends on
the reflood conditions. Tab1e 4 shows a 1isting of the quench times of the
grid spacers instrumented for the unblocked bund1e SEFLEX tests.
lt is important to mention that quenching is initiated at the trai1ing edge
of a grid spacer for all cases investigated. The velocity of the downwards
moving quench front between the trai1ing edge and the leading edge depends on
the ref100d conditions and the axial location of the grid spacer. The measu
rements do not confirm previous assumptions that the droplet impact on the
grid spacer would initate quenching at the 1eading edge.
For many of the cases premature quenching of the rod cladding is initiated
downstream of the grid spacers as we1l. For the bund1e midp1ane this is shown
in Fig. 39, Plot B, e and D. This leads to the conclusion that heat transfer
enhancement is significant in the ear1y portion of the dispersed f10w regime.
During that period the grid spacer is hot and dry. The cladding temperature
-~-
transients downstream and upstream of the grid spacer are unaltered when the
grid spacer wets. Therefore. the individual droplet breakup mechanisms at a
dry grid or the droplet deposition at a wet grid including re-entrainment at
the trailing edge seem to lead to comparable results concerning the overall
heat transfer enhancement downstream of the grids.
Heat transfer analysis from da ta measured confirms previous findings [1) that
cooling enhancement downstream of grid spacers shows a maximum for the early
portion of the transient mist flow regime. The heat flux at the rod surface
12 mm downstream of the trailing edge of the grid spacer is about 20 percent
higher than at the rod surface neighboring the leading edge of the grid
spacer. The enhancement disappears towards the end of the dispersed flow
regime as shown in Fig. 40. There is a second maximum for the cooling enhan
cement at the onset of transition film boiling eharacterized by inereased
water content in the two-phase flow. The inerease of the water eontent presu
mably leads to grid spaeer rewetting eoincidently in the case shown in Fig.
40. This eoincidenee is not observed in this test for the remaining grid
spaeers at different elevations in the bundle or for other tests performed
with different reflood eonditions. The eooling enhaneement from the grid
spaeer during the transition film boiling regime is of the same order of
magnitude as that during the dispersed flow regime. However. eoneerning elad
ballooning loeal eooling enhaneement is more important during the dispersed
flow regime.
The axial extension of the grid spacer effeet on the loeal eooling enhanee
ment depends on the reflood eonditions. Figure 41 shows surfaee heat fluxes
and eladding temperatures upstream and further downstream of the midplane
grid spaeer eomplementing the data of the test plot ted in Figs. 39 (Plot B)
and 40. At axial level 1925 mm, i. e. 62 mm downstream of the trailing edge
of the grid spaeer, eooling enhaneement ean be observed only for the early
portion of the dispersed flow regime compared with the eooling eonditions 100
mm upstream of the grid spaeer, i. e. axial level 2125 mm (Fig. 41, Plot A).
Plot Band eillustrate, that 162 mm as weIl as 262 mm downstream of the grid
spaeer the eooling eonditions are approximately the same as 100 mm upstream
of the grid spaeer during the early portion of the dispersed flow regime
inspite of the inereased distance from the queneh front. The transients com
pared in Plot D indicate that 362 mm downstream of the trailing edge there is
no more grid spacer effeet under the reflood eonditions mentioned.
-31-
It is evident from the da ta presented so far that a grid spaeer has signifi
eant effeets on the eooling eonditions downstream of it. However, for a quan
tifieation of the different heat transfer meehanisms superimposed, more in
formations are needed about droplet size and veloeity distributions as weIl
as about loeal flow turbulenee and steam temperature.
Figure 42 shows fluid TC signals measured upstream and downstream of the
5 x 5 rod bundle midplane eomplementing the data plot ted in Figs. 39 (Plot
B), 40 and 41. The fluid temperature signal measured 215 mm upstream of the
leading edge of the grid spaeer (Plot A) indieates signifieant steam super
heat. However, liquid droplets impinging upon the TC tip lead to repeated
quenehing of the probe preventing the measurement of the real steam tempera
ture. Far downstream of the grid (Plot D) steam superheat is elearly indi
eated lasting for the whole dispersed flow transient. Quenehing of the grid
spaeer at t = 140 seeonds does not affeet the steam temperature measured 362
mm downstream of the trailing edge of the grid. The transients measured at
the levels 1925 and 1825 mm (Plots Band C) give less information about the
steam temperature presumably due to inereased droplet impinging upon the
individual probe tips by the inereased number but smaller size of droplets in
that portion of the rod bundle.
In Figure 43 the temperature transients, almost presented in Figs. 39 (Plot B)
and 42 are plotted versus an enlarged time seale to elueidate in more detail
the effeets of droplets on the fluid TC signal as weIl as on the grid spaeer
temperatures. Flow pulsations of a wave period of about 4.5 seeonds influenee
the grid spaeer temperatures slightly and with some delay. The grid spaeer
temperature is mainly eontrolled by the radiation heat transfer from the rods
and the droplet eooling, beeause the vapor temperature seems to be elose to
the grid temperature. High fluid TC signal indieates high vapor superheat,
dry probe tip, and low eooling, presumably due to low vapor veloeity and low
water eontent in the two-phase flow. Low fluid TC signal indieates enhaneed
eooling espeeially at the grid and in the wake of it presumably due to in
ereased vapor veloeity, water entrainment, and droplet breakup leading to
quenehing of the fluid TC tip. The eladding temperature measured upstream of
the grid is nearly unaffeeted by the flow pulsation.
The signals of the fluid TC probes plaeed immediately downstream of the grid
spaeer (da ta not shown) indieate that the probe tips stay wet inspite of the
-32-
presence of superheated steam and the flow pulsations mentioned. Higher con
tent as weIl as different distribution of the water compared with the con
ditions upstream of the grid spacer can be assumed by the following reasons:
At a given elevation in the bundle the steam mass flux is decreasing during
half aperiod of oscillation of the dispersed flow transient. Therefore,
decreasing steam velocity may lead to fall back of apart of the water en
trained before by the steam of increased velocity. Locally increased overall
steam velocity within the subchannel constrictions along the individual grid
spacers hinders fall back across the grids leading to somewhat increased
water content immediately downstream of the grids compared with the mean
water content between two grid spacer elevations.
Figure 44 shows the temperature transients later in time plot ted again versus
an enlarged time scale. Quenching of the grid is initiated at the trailing
edge. The downwards moving quench front at the grid needs about 5 seconds to
reach the leading edge in this case.
For the next grid spacer, placed 545 mm above the bundle midplane, the con
ditions are similar to those at the midplane grid spacer as shown in Fig. 45.
However, the dispersed flow cooling is somewhat lower indicated by the longer
time span until rod quenching. The fluid contains less water than upstream of
the midplane grid spacer as indicated by the signal of the fluid TC placed
245 mm upstream of the leading edge of the next grid spacer. The fluid TC
probe tip is quenched only once for a short time during the flow oscillation
period and stays dry for the remaining - extended - dispersed flow period.
Inspite of the lower water content the grid spacer effect is increased com
pared with the midplane grid spacer situation. This is presumably due to the
increase of the steam velocity and the turbulence enhancement over the bundle
length given. The peak cladding temperatures are 120 Klower at the position
17 mm downstream of the trailing edge than 45 mm upstream of the leading edge
of the grid spacer. Under these reflood conditions this grid spacer is
quenching when the dispersed flow turns to transition film boiling as did the
midplane grid spacer.
This phenomenon has not been observed in the test run carried out with a
system pressure of 2.1 bar as shown in Fig. 46. The grid spacer is quenching
during the early portion of the dispersed flow regime, and the difference of
-33-
the cladding temperatures upstream/downstream of the grid spacer amounts to
220 K. The slope of the cladding temperature transient downstream of the grid
spacer is somewhat affected when the grid is quenching, indicating that a wet
grid spaeer improves the heat removal downstream of it. However, this is not
the only reason for the inerease of the grid spacer effeet (220 K eladding
temperature differenee in this test instead of 120 K for the 4.1 bar test).
For the ease of the low system pressure of 2.1 bar the steam velocity and
henee, the water entrainment and the turbulence are higher. The signal of the
fluid TC indieates that the probe tip plaeed 245 mm upstream of the grid
spacer stays wet for a long time span of the dispersed flow period. Dryout
of the probe tip and henee, measurement of steam superheat is possible during
the last period of the dispersed flow regime when the steam velocity is
lower. Besides the fact that the fluid flow from below is eooling the grid
spaeer, it is evident that the grid is eooled from above as weIl, sinee
quenehing is initiated at the trailing edge due to the eooling enhaneement,
and eventually aeeumulated water downstream of the grid.
The magnitude of the grid spaeer effect on the cladding temperatures in this
case - eompared with the 4.1 bar test - results from both, the effeets down
stream of the grid, i.e. turbulence enhancement again inereased by higher
steam velocity, and the effeet at the grid itself, i.e. additional steam
desuperheating after grid rewetting. The ratio of the effeets of the two
meehanisms is still uneertain. However, turbulenee enhaneement seems to be
dominant.
The transient cooling conditions during reflood tests make difficult the
distinction whether the change of the grid spacer condition from dry to wet
or the change of the arriving fluid influences the grid spacer effect.
From the transient 5 x 5 rod bundle tests boundary conditions are selected
eharacterized by rather stable dispersed flow conditions for the time span in
which a grid spacer is quenching. Furthermore, the bundle section upstream
and downstream of the grid spacer selected has to be instrumented sufficient
ly. For the midplane grid spacer the desired flow conditions are not given
during the test performed with p = 4.1 bar and v = 3.8 cm/s because of the
change of the fluid conditions at quenehing of the grid as shown in Fig. 40.
The test performed with p = 2.1 bar and v = 3.8 cm/s provides more stable
flow conditions durin~ the period in which the midplane grid is quenching.
-34-
Figure 47 shows cladding temperatures and surface heat fluxes versus time at
the leading edge and 12 mm downstream of the trailing edge of the midplane
grid spacer. Quenching of the grid is initiated at the trailing edge at
t = 65 seconds and terminated at the leading edge at t = 101 seconds. At
t = 65 seconds the heat transfer is increasing downstream as weil as upstream
by about the same amount as indicated by the heat flux transients. This is
presumably due to the change of the conditions of the fluid arriving at the
grid spacer elevation. At t = 101 seconds the surface heat flux is increasing
downstream of the grid spacer only.
The grid spacer effect may be increased at t = 101 seconds due to different
droplet effects induced by the wet grid as weil as due to increased droplet
volume flux. However, this situation does not last for a long time. The rod
surface heat flux increases then at the leading edge and decreases at the
trailing edge after the peak at t = 101 seconds.
Comparing the magnitude of the grid spacer effect during the early portion of
the dispersed flow regime with that evaluated for the period of grid rewet
ting, it becomes evident that the grid spacer effect is more important during
the early portion of reflood. There is a minor effect on the cooling enhance
ment later in time which is not dependent on whether the grid spacer is dry
or wet.
The quench front progressions in 5 x 5 REBEKA fuel rod bundles indicate more
pronounced effects at the grid spacers downstream of the midplane grid as
shown in Fig. 36.
In Fig. 48 rod surface heat flux transients are plotted which have been
evaluated from the cladding temperatures measured close to the grid spacer
above the midplane grid spacer. The cladding temperatures shown for reference
are taken from the test performed with p = 4.1 bar and v = 3.8 cm/s (see
Fig. 45). For a short time the heat flux at the rod surface 17 mm downstream
of the trailing edge of the grid spacer is up to 60 percent higher than that
at the rod surface 45 mm upstream of the leading edge of the grid spacer. The
enhancement disappears towards the end of the dispersed flow regime. There is
a second maximum for the cooling enhancement at the onset of transition film
boiling. The cooling enhancement from the grid during the transition film
boiling regime is of the same order of magnitude as that during the dispersed
-35-
flow regime. However, eoneerning elad ballooning loeal eooling enhaneement is
more important during the dispersed flow eooling period.
Under the reflood eonditions of the test mentioned (p = 4.1 bar, v = 3.8 em/s)
grid spaeer rewetting eoineides with the onset of transition film boiling
eharaeterized by inereased water eontent. The sudden rise of the water eon
tent presumably leads to grid spaeer rewetting in this test. For different
test eonditions, e.g. higher water injeetion rate and/or lower system pres
sure, the grid spaeers are quenehing during dispersed flow eonditions al
ready, as shown in Figs. 46 und 47.
Figure 49 shows heat flux transients of the test performed with p = 2.1 bar,
v = 3.8 em/s. The measurement positions are the same as those ehosen for
Fig. 48. Downstream of the grid spaeer the heat flux is up to 100 pereent
higher than upstream for a short time during the early portion of the dis
persed flow regime. After the maximum, the heat flux is deereasing more or
less steadily below the transient evaluated for the position upstream of the
grid spaeer.
Under identieal reflood eonditions exeept the gas filling of the rods, e.g.
argon instead of helium, the ratio of the heat flux downstream/upstream of
the grid spaeer remains approximately the same as for helium-filled rods.
However, the eladding temperatures deerease faster at all elevations, and the
eladdings are quenehing downstream of the grid spaeer earlier than upstream
as shown in Fig. 50. For the sample of tests performed with unbloeked bundles
the main grid spaeer effeets are summarized in Tables 4 and 5.
The influenee of fuel rod simulator geometry and physieal properties on over
all eooling eonditions and rod queneh behavior is being investigated separa
tely. Coneerning grid spaeer effeets the following ean be summarized: For
inereased heat resistanee between eladding and pellets the eladdings intent
to queneh downstream of the grid spaeers earlier than upstream of them. Al
though, the dispersed flow eooling enhaneement promoted by grid spaeers is
rather unaffeeted by the eonduetanee of agas filled gap. The resulting dif
ferenee of the eladding temperatures is e.g. about 220 K for most time of the
reflood transient in both tests with either helium- or argon-filled rods. The
eorresponding surfaee heat flux transients have similar slopes and ratios
(eompare Figs. 49 and 50).
-36-
Tab1e 4
Quench tirnes of grid spacers in unb10cked rod bund1esas function of ref100d conditions.
Grid Spacer TC Test No.Axial Level Axial Level Gap Gas Filling
rnrn rnm F100ding Velocity, crn/sSystem Pressure, barFeedwater Temperature, °c
Quench Time, s
05 03 06 04 07Helium Helium Helium Helium Argon
3.8 3.8 5.8 5.8 3.82.1 4.1 2.1 4.1 2.140 40 40 40 40
3660' no TC - - - - -3622' no TC - - - - -
3115' no TC - - - - -3077' 3075' 7 21 18 17 27
2570' 2568' 77 78 50 51 892532' 2534' 73 71 36 41 68
2025' 2023' 105 145 33 67 961987' 1989' 65 140 12 47 66
1480' 1478' 42 202 12 59 341442' 1444' 28 186 12 46 25
935' no TC - - - - -897' 899' 36 216 9 35 28
390' no TC - - - - -352' 354' 4 57 6 17 15
Leading edge of grid spacer,Trai1ing edge of grid spacer
,TC p1aced in subchanne1 surrounded by rods No. 13, 18, 17, and 12.,TC p1aced in subchanne1 surrounded by rods No. 13, 12, 7, and 8.
-37-
Table 5
Grid spacer effects on cladding temperature and surface heat fluxin unblocked rod bundles as function of reflood conditions.
Test No.Gap Gas FillingFlooding Velocity, cm/sSystem Pressure) barFeedwater Temperature, oe
05 03 06 04 07Helium Helium Helium Helium Argon
3.8 3.8 5.8 5.8 3.82.1 4.1 2.1 4.1 2.140 40 40 40 40
Naximum reduction of claddingtemperature (K) comparing axial 150 60 230 120 160level 2125 with 1975 mm. 1
Naximum heat flux enhancement (-)relating axial level 1975 rum to 1.6 1.2 2.1 1.3 1.5axial level 2025 mm. 1
Time (s) of maximum heat fluxenhancement after initiation of 5 20 12 20 23reflood. 1
-
Naximum reduction of claddingtemperature (K) comparing axial 220 120 230 180 220level 1525 with 1425 mm. 2
flaximum heat flux enhancement (-)relating axial level 1425 mm to 2.0 1.6 2.1 1.6 1.7axial level 1525 mm. 2
Time (s) of maximum heat fluxenhancement after initiation of 5 20 12 20 23reflood. 2
Grid spacer at bundle midplane(leading edge at axial level 2025 mm., trailing edge at axial level 1987 mm)
2 Grid spacer placed 545 mm downstream of bundle midplane(leading edge at axial level 1480 mm, trailing edge at axial level 1442 mm)
-38-
5.3 Blockage Effects
Ballooned fuel rod claddings may lead to flow blockages influencing local
reflood heat transfer conditions. A bloekage causes two opposite effects:
- Within and downstream of the blocked portion of the bundle, the coolant
mass flux is reduced, which can lead to reduced local eooling.
- Two-phase flow passing through a blockage ean lead to improved eooling due
to enhancements of turbulence and water droplet dispersion.
For the coplanar 90 pereent blockage formed by a 3 x 3 rod cluster with
Zircaloy claddings in a corner of the 5 x 5 REBEKA rod bundle, the eladdings
are highly lifted from the pellets and the heat capacity of the eladding
balloons is low. For the same outer shape of the blockage, investigated as
part of the FEBA program, the heat eapaeity of the stainless steel blockage
sleeves of 1 mm wall thiekness, attaehed at the FEBA rods, is significantly
higher [1] (compare Figs. 15, 16 and 17). Therefore, the two-phase flow pas
sing through the blockage may cool down the thin Zirealoy claddings faster
than the FEBA sleeves. Figure 51 shows temperature transients measured at the
bundle midplane, i.e. the midplane of the 90 percent blockage as well, for a
FEBA test (Plot A), for a SEFLEX test with helium-filled gaps (Plot B), and
for a SEFLEX test with argon-filled gaps (Plot C). The temperatures of the
sleeves and of the balloons, respeetively, are lower than the eladding tem
peratures of the rods placed in the bypass of the blocked portion of the
bundle. However, the balloons are quenching substantially earlier than the
sleeves. After about 20 seconds the Zircaloy balloons, and after about 400
seconds the FEBA sleeves are quenching. After quenching the portions of the
rods underneath the balloons and underneath the sleeves, respectively, remain
at a high temperature level indicating the magnitude of the heat resistance
between ballooned claddings and heater sheath inside the pellets as well as
between FEBA rod surface and sleeve (gap of 1.0 mm width filled with stagnant
steam). For the SEFLEX blockage the gap between the ballooned Zircaloy clad
dings and the pellets amounts to about 2.3 mm width filled with either helium
or argon. The cladding temperatures measured in the bypass of the blockage
are shown for reference.
For the same tests the cooling conditions downstream of the blockage are
indieated by the temperature transients shown in Fig. 52. In the FEBA test
-39-
the temperatures downstream of the blockage and in the bypass are roughly the
same during the early portion of reflood. The mass flux reduction in the
blockage is approximately compensated by the cooling enhancement effect of
the blockage. However, quenching is delayed compared with the bypass con
ditions. For the Zircaloy cladding the cooling enhancement, prevailing at
beginning of reflood mainly, is sufficient 'to quench the thin Zircaloy clad
dings rapidly. The quench front initiated within the SEFLEX blockage moves
fast through the whole blockage increasing the precursory cooling immediately
downstream of the blockage. Both effects, low heat capacity of the Zircaloy
claddings and high heat resistance between the heat source and the lifted
clad within the blockage are responsible for the quick cooling of the rod
claddings. Most part of the heat stored in the portions of the rods with
nominal gap width of 0.05 mm is being removed after quenching of the clad
dings as indicated by the heater sheath temperatures shown in Fig. 52, Plot
Band Plot C. Due to the low heat conductivity of argon the heater sheath
temperature remains at a higher level for the quasi steady state conditions
late after quenching (Plot C) compared with the temperature level measured in
the test with helium-filled gaps (Plot B). Again, the cladding temperatures
measured in the bypass of the blockage are shown for reference.
The instrumentation of some of the heater rods in the center of the alumina
pellet column of the REBEKA rods used in SEFLEX test se ries 3 and 4 provides
information about the real temperature distribution in the rod cross sec
tions. For the calculation of the temperature distribution concentric ar
rangement of the cladding, the alumina pellets and the heater rod has been
assumed. However, it is more probable that the arrangement is non-concentric
for a given measurement position. For analysis of data measured this fact has
to be taken into account.
Figure 53 shows a comparison of measured and calculated heater sheath tem
peratures and corresponding cladding temperatures measured at the bundle
midplane in the blockage bypass. In the individual plots da ta obtained from
different tests are plotted versus time. For each plot identical locations in
rod No. 4 of the 5 x 5 REBEKA rod bundle have been chosen. Across the gap of
0.05 mm nominal width between cladding and pellet significantly different
temperature differences are calculated depending on the filling of either
argon or helium gas. The temperature difference across the pellet is un
affected by the physical properties of the gases.
-~-
The temperature transients shown in Plot A of Fig. 53 indieate (1) large
temperature difference aeross the argon-filled gap and (2) rather good agree
ment between measured and calculated heater sheath temperatures. For the
helium-filled gap (Plot B) the temperature drop aeross the gap is small. The
comparison of the heater sheath temperatures measured and calculated seems to
indicate a bad agreement. However, there is no information about pellet excen
tricity at that location. Furthermore, there is an axial displacement between
heater rod and Zircaloy cladding. This is due to different thermal extensions
of the heater rod and the eladding during the tests. Therefore, with respect
to uncertainties of the geometrical conditions, the da ta comparison shows a
satisfactory matching.
Plot C shows the conditions for a argon-filled gap again. However, the system
pressure applied in the corresponding test is 2.1 bar instead of 4.1 bar
comparing Plot C with Plot A. For lower system pressure the reflood heat
transfer and hence, the heat flux aeross the gap are lower. Therefore, the
temperature difference between rod cladding and heater sheath is somewhat
smaller in Plot ethan in Plot A. The same trend is true comparing the data
of Plot B with those of Plot D.
In Fig. 54 the temperatures measured at various locations within the blocked
portion of the REBEKA rod bundle are shown. Upstream of the bloekage the
temperature difference between cladding and heater sheath is the same as in a
rod placed in the bypass as shown eomparing Plot A of Fig. 54 with Plot D of
Fig. 53. At the axial level 2075 mm, i.e. at the lower conical end of a rod
balloon, the temperature differenee between balloon surface and heater sheath
is substantially larger and the quench time for the cladding is shorter than
for the axial level upstream with nominal rod geometry (compare Plot B with
Plot A of Fig. 54). At the axial level 2025 mm, i.e. the blockage midplane,
the balloon is quenching rapidly as shown in Plot C of the Fig. 54. The
cooling conditions within the fully blocked subchannels seem to be rather
unstable as indicated by the peak of the balloon temperature after quenching
due to a short dryout period in a subchannel.
The upper conical end of a rod balloon, i.e. downstream of the blockage
midplane, is quenching earlier than the lower eonieal end as shown in Plot D
of Fig. 55 compared with Plot B of Fig. 54. Even downstream of the blockage
the quench times are substantially shorter for rod seetions of nominal geo-
-41-
metry compared with the conditions in the bypass (compare Plots E and F of
Fig. 55 with Plot B of Fig. 52). It has to be mentioned that the corner rod
of the blocked rod cluster, rod No. 13 (Plot F), is surrounded by one fully
blocked and three partly blocked subchannels whilst rod No. 17 (Plot E) is
surrounded by four fully blocked subchannels. Therefore, the cooling condi
tions for these two rods are different within as weIl as downstream of the
blockage. May be that water accumulation and enhanced turbulence downstream
of the blockage, as assumed previously [1], are responsible for increased heat
transfer at that location. The temperature transient measured at the heater
sheath of rod No. 17 at the axial level 1925 mm indicates rapid removal of
the heat stored in that section of the rod as shown in Fig. 55 Plot E. For
the sections of the rods underneath the clad balloons, most part of the heat
stored remains in the pellets and the heater rod as indicated by the tempera
ture transients plotted in Plots Band C of Fig. 54 and Plot D in Fig. 55.
Therefore, early quenching of the balloons is possible inspite of reduced
coolant flow through the blocked bundle subchannels. Poor heat transmission
from the pellets through a large gas-filled gap to the thin Zircaloy cladding
balloons having small heat capacity are responsible for rapid quenching of
the blocked bundle portion. In contrast to that behavior the FEBA blockage
array, characterized by substantial larger amount of stored heat and smaller
heat resistances in radial direction of the rods and the blockages, leads to
conservative results concerning the behavior of blocked fuel rod clusters.
Investigating the coolability of blocked fuel elements the simulation of a
90 percent blockage applied in the SEFLEX tests leads to rather realistic
results.
-42-
6. ANALYTICAL SIMULATION OF REFLOOD EXPERIMENTS
The reflood behavior of both the unblocked bundles consisting of either
5 x 5-FEBA and 5 x 5 REBEKA rods were calculated using the COBRA-TF (Coolant
!oiling in !od ~rrays - !wo !luid) computer code developed at the Pacific
Northwest Laboratory (PNL) as part of the cooperative USNRC, Electric Power
Research Institute (EPRI), and Westinghouse (~) FLECHT-SEASET program.
6.1 COBRA-TF, A "Best-Estimate" Computer Program
The COBRA-TF code was developed to predict the thermal-hydraulic response of
a LWR rod bundle during LOCA reflood. The computer code [37J provides a two
fluid, three-field representation of the two-phase flow. The two fluids are:
Water and its vapor. The three field are: Continuous vapor, continuous li
quid, and entrained liquid droplets. In addition, COBRA-TF allows for the
transport of a non-condensible gas mixture with the vapor field.
This two-fluid, three-field description of the two-phase flow results in a
set of nine conservation equations. Four continuity equations are required
for the vapor, continuous liquid, entrained liquid, and non-condensible gas
mixture. Three momentum eqllAtions are solved, allowing the liquid and en
trained liquid fields to flow with different velocities relative to the vapor
phase. Two energy equations are specified tor the vapor-gas mixture and the
combined liquid fields. The liquid and entrained liquid fields are assumed to
be in thermal equilibrium.
These conservation equations and the equations for heat transfer from and
within the solid structures in contact with the two fluids are solved using a
semi-implicit, finite-difference numerical technique on an Eulerian mesh. The
selection of either rectangular Cartesian or subchannel coordinates is pro
vided. This allows a fully three-dimensional treatment in geometries amenable
to description in a Cartesian coordinate system. The constitutive relations
include state-of-art physical models for the interfacial mass transfer, the
interfacial drag forces, the liquid and vapor wall drag, the wall and inter
facial heat transfer, the rate of liquid entrainment and de-entrainment, and
the thermodynamic properties of the fluid. A mixing length turbulence model
is included as an option.
-43-
A consistent set of heat transfer models was implemented. lt consists of five
components:
a) A conduction model specifies the conductor geometry (fuel rods, electri
cally heated rods, tubes, and walls) and material properties, and solves
the heat conduction equations.
b) A heat transfer package selects and evaluates the appropriate heat
transfer correlations.
c) A quench front model employs a fine mesh rezoning technique in which fine
mesh heat transfer cells with axial and radial conduction are superimposed
upon the coarse hydrodynamic mesh spacing. The quench front propagation is
calculated by applying a boiling heat transfer package to each node, so
that the resulting quench front velocity is a function of the axial con
duction, the boiling curve shape, the prequench heat transfer, and the
internal heat conduction within the structure.
d) A dynamic gap conduction model evaluates the fuel pellet to clad heat
conduction for a nuclear fuel rod.
e) A subchannel-based radiation model determines the rod to rod, rod to
vapor, and rod to droplet radiation heat transfer.
Physical models [38], [39] were implemented into the code to
describe, as realistic as possible,
a) the two-phase enhancement of convective heat transfer in the dispersed
flow,
b) the subchannel thermal radiation,
c) the effects of grid spacers, i.e. single-phase convective heat transfer
enhancement, droplet impact heat transfer, droplet breakup, droplet en
trainment snd de-entrainment, micro-droplet evaporation, snd grid spacer
rewetting,
d) the effects of blockages, i.e. flow redistribution, single- phase con
vective heat transfer enhancement, droplet impact heat transfer on the
blockage, droplet breakup due to the blockage.
-44-
6.2 Simulation of FEBA and SEFLEX tests
For the ca1cu1ation with COBRA-TF, the FEBA test section was mode1ed by using
two representative fluid channe1s, one center channel, and one periphera1
channe1. The 5 x 5 heater rods were simu1ated by two rods, one center rod and
one periphera1 rod. The test section housing was described by a wall with an
inside heat transfer surface and an insu1ated outer surface. Figure 56 shows
the radial noding scheme of the bund1e. Transverse connections were specified
between the coo1ant subchanne1s to comp1ete the multidimensional mesh for the
region taken into consideration. The f10w area between the channe1s was given
by the width and the vertica1 1ength increment for the mesh. Form drag 10ss
coefficients and wall friction factors were additional informations required
as input data for the transverse momentum equations. Figure 57 shows the
axial noding scheme of the FEBA test section. For the simulation of the
heated 1ength of 3900 mm, 18 vertica1 mesh ce11s were chosen. The vertica1
10cation of the faces of the individual ce11s are indicated correspondingly
to the input data for COBRA-TF (reference level or zero level at the coo1ant
in1et) as we11 as with respect to the experiments to be simu1ated (reference
level or zero level at the upper end of the housing). The 1atter informations
are written in the brackets. The numbers of these mesh ce11s depend on the
degree of detail required to reso1ve the fluid fie1d, the phenomena being
mode1ed, and practica1 restrictions such as computing time and computer sto
rage 1imitations. To capture the dominant physica1 phenomena and to coincide
with se1ected measurement 10cations, a variable node 1ength was provided. At
the mesh ce11 faces, the fluid ve10cities are computed. Loca1 pressure losses
in the vertica1 f10w due to grid spacers, orifice p1ates or other obstruct
ions in the f10w fie1d are mode1ed in the code as a velocity head 10ss. On
the other hand, the state variables, e.g., pressure, density, entha1py, and
phasic volume fractions are computed at the mesh ce11 center. This means the
three fie1d conservation equations for multidimensional f10w are solved using
a standard "staggared" differencing scheme for the convected quantities (do
nar ce11 differencing). Both fluid and rod temperatures are also ca1cu1ated
at the centers of the fluid continuity cel1. Wen the axial temperature diffe
rences between adjacent axial nodes exceed maximum surface temperature diffe
rences, an additional node row is inserted ha1fway between the two original
nodes. This splitting process continues unti1 the mesh is fine enough to
reso1ve the surface temperature adequate1y, when the temperature is near the
critica1 heat flux temperature. Converse1y, fine mesh nodes coa1esce when the
44
-45-
quench front has propagated downstream and the criterion based on minimum
temperature differences between adjacent nodes indicates coalescense of finer
nodes. The axial noding scheme was bounded by phantom bottom and top nodes
which contained the known boundary conditions. The data comparison was
carried out for the axial locations (2225, 1975, 1875, 1675, and 1125 mm
referred to the zero level of the experiments) marked by dots.
For the description of the conduction models, the characteristics of the
heater rods (referred to as solid cylinders) and housing (referred to as flat
plate) were specified. The modeling requirements included the following
features: Unequal mesh spacing, internal resistance due to gaps, radial heat
generating profiles, and temperature- and space-dependent material pro
perties.
A maximum of five different material tables can be used by the code data
input. The FEBA and REBEKA rods consist of concentric rings of material re
gions, as shown in Figs. 11 and 12, respectively. In each region, the number
of radial nodes, width, and power factor as weIl as the material type were
specified by the data input. Contact resistances were not calculated between
material regions but were modeled by including a region, one node wide, with
material properties which gave them the appropriate thermal resistances. The
FEBA and REBEKA rod geometries were defined by eight and nine radial nodes,
respectively. With respect to the maximum of five different material tables,
the material properties of the heater rod with 6.02 mm o.d., which is placed
in the center of the REBEKA rod, were described by those of a pseudomaterial.
The averaged thermophysical properties of that pseudomaterial took into
account the densities, heat capacities, and heat conductivities (including
heat resistances at the interfaces) of the filler material (magnesium oxide),
heating element (Inconel), electrical insulator material (boron nitride) and
sheath material (Inconel). This approach seemed to be adequate since calcula
tions [40] have shown that the dimensions of the alumina pellets and of the
encapsulated heater rod have in comparison with the gas filled gaps a minor
influence on the thermal behavior of the REBEKA fuel rod simulator during the
reflood phase.
The thick-walled housing, shown in Figure 56, was modeled by four radial nodes
to account for conduction and heat transfer from and to the inner surface.
The outer surface was assumed to be insulated.
-46-
All thermophysical material properties as function of temperature were taken
from Refs. 42 and 43. The initial axial temperature profiles of the center
rods, peripheral rods, and the housing as weIl as the flooding parameters,
i.e. flooding velocitYt system pressure, and feedwater temperature, were
specified corresponding to the individual test runs. The actual power profile
was slightly modified to fit the fluid cell boundaries with reference to the
actual power profile step changes. The total power of the rod bundle was
conserved in the input data.
6.3 Comparison of Test Data with COBRA-TF Calculations
A total of four computer runs were made to simulate forced flow bot tom re
flood tests of the FEBA and SEFLEX-program, respectively. The comparison of
code predictions against the experimental data for such main informations as
cladding temperatures and quench front velocities are described in detail in
Refs. 34 and 36. Therefore, the da ta comparisons presented in this report
document only the results for SEFLEX test No. 03 obtained on the latest
version of the COBRA-TF computer code available to EPRI in fall 1984.
This SEFLEX test is a run of test series 1 carried out with an unblocked
bundle geometry containing seven grid spacers. The 5 x 5 rod bundle consisted
of REBEKA fuel rod simulators with helium-filled gaps between the alumina
pellets and the Zircaloy claddings.
For the calculations measured initial and boundary conditions were used as
input data. Figure 58 shows the initial axial temperature profiles of the
center rods, the peripheral rods, and housing, obtained by averaging all
initial thermocouple readings of the instrumented axial levels. The tempera
ture profiles are roughly symmetric about the bundle midplane (axial level
2025 mm). Figure 59 shows the boundary conditions plotted as function of re
flood time. The flooding velocity (3.8 cm/s in the cold bundle), the system
pressure (4.1 bar), and the feedwater temperature (40 °C) were kept constant
during the test. For about two hours prior to reflood, the bundle and the
housing were heated in an essentially stagnant steam environment to the de
sired initial temperatures using a low bundle power. The power input was
stepped up, when the rising water level reached the bottom end of the heated
bundle length, to about 200 kW followed by a decay heat transient corres
ponding to 120 percent ANS standard 40 seconds after scram.
-47-
COBRA-TF was run from the start of reflooding, with the initial and boundary
conditions already described. The simulation continued until about 75 percent
of the heated bundle length was quenched, in view of the computing time.
Figure 60 shows the measured and calculated cladding temperatures for three
axial positions in the rod bundle, always approximately 300 mm downstream of
the trailing edge of a grid spacer. Plot A represents a fairly good agreement
between the measured and calculated data for the axial level 2225 mm, just
upstream of the bundle midplane, a position characterized by the maximum rod
power. The measured data (solid line) are taken from rod No. 18 (TC-position
18a1, see Fig. 20). The diamonds represent the results of COBRA-TF calcu
1ation for a simulated center rod. The time interval of the symbols is every
5 seconds (the print interval). Tt can be seen that for the early portion of
the dispersed flow, the code overpredicts slightly the measured data. Later
in time, the cladding temperature is underpredicted. The quench time is
reached slightly earlier than in the experiment; the quench temperature is
weIl predicted. Plot B represents a similar comparison for the axial level of
1680 mm, just downstream of the bundle midplane, again a region of maximum
rod power. The measured data are taken from rod No. 12 (TC-position 12b4,
see Fig. 20). The diagram indicates a good matching of the temperature
transient; and it is noticeable that the small divergences between the com
pa red data decrease with increasing distance from the bot tom end of the
heated bundle length. Plot C illustrates the comparison for the axial level
1135 mm, the beginning of the last quarter of heated bundle length. The peak
to-average rod power amounts to 1.06 at this position. The measured data are
taken again from rod No. 12 (TC-position 12b3, see Fig. 20). A comparison of
the measured data against COBRA-TF calcu1ation indicates an excellent agree
ment for temperature rise, turnaround time, quench temperature, and quench
time.
Figure 61 shows a comparison of the observed and computed cladding tempera
ture immediately downstream of the trai1ing edge of the grid spacer at the
bundle midplane, i.e. for axial levels at 1975 and 1875 mm, respectively. The
measurements were recorded from the center rod (rod No. 13; i.e. TC-position
13i3 at axial level 1975 mm and TC-position 13il at axial level 1875 mm,
respectively, see Fig. 20).
The grid spacer effects modeled in COBRA-TF predict weIl the trends observed
at these axial positions of the simulated SEFLEX test run.
-~-
The quench times of the grid spacers located just upstream of the bundle
midplane (No. 3), at the midplane (No. 4), and just downstream of it (No. 5)
show an excellent agreement as plotted on an enlarged time scale in Fig. 62,
Plot A through C. Only the characteristic temperature drops within the first
10 seconds of the reflood time are not very weIl predicted by the computer
code. Figure 63 shows a comparison of measured housing data against COBRA-TF
calculation for the axial positions 2225 mm (Plot A), 1680 mm (Plot B), and
1135 mm (Plot cl. These axial positions correspond to the data comparison for
the rod surface temperatures, plot ted in Fig. 60. The COBRA-TF simulations
show again a reasonable agreement with the reflood data, even if the quench
times are predicted slightly earlier « 20 %) than observed in the experi-
ment.
Figure 64 shows a comparison of the measured and calculated quench front pro
gression, which indicates an excellent matching of the data for the most
part of the heated bundle length. Only for the upper most portion of the rod
bundle, the differences increase and the quench times are slightly underpre
dicted.
Figure 65 shows a comparison of the measured and calculated quench front pro
gressions for four simulated test runs. All experiments were carried out
under identical initial and boundary conditions, as far as experimentally
possible; in particular, same axial temperature profiles at initiation of
reflood, same flooding velocities of the rising water levels in the cold
bundles (3.8 cm/s), and same feedwater temperatures (40 °C). However, the
reflood experiments were carried out with different system pressures of 4.1
and 2.1 bar, respectively, using rod bundles which have fuel rod simulators
of different design and in ca se of gapped rods with helium- or argon-filled
gaps. For the FEBA "solid-type" rod bundle (Plot A) flooded at a system
pressure of 4.1 bar, the calculated quench time becomes slightly shorter than
the experimental as the axial elevation increases. As discussed earlier, a
remarkably good agreement is obtained for the REBEKA rod bundle with helium
filled gaps (Plot B) under identical flooding conditions. For flooding ex
periments performed at a system pressure of 2.1 bar using REBERA rod bundles
with helium- (Plot C) and argon-filled gaps (Plot D), respectively, the com
puted quench times are slightly higher than the measured. The overprediction
of the quench times is presumed to be due to an underprediction of the liquid
content of the flow and hence, heat transfer at lower system pressure.
-49-
Finally should be mentioned that COBRA-TF does a remarkably good job of pre
dicting peak cladding temperatures and quench times of forced reflood tests,
though the code has undergone only a limited assessment since the completion
of its development. To define remaining deficiencies in the physical models
the COBRA-TF code should run against data from the SEFLEX experiments, since
a source version of the program was not available and hence no modifications
or enhancements were made to the code during the course of this study.
-50-
7. CONCLUSIONS
Fuel rod simulators with Zircaloy claddings and agas filled gap beetween clad
dings and pellets (REBEKA rods) exhibit lower peak cladding temperatures and
shorter quench times during reflood than gapless heater rods with stainless
steel claddings (FEBA rods).
Due to earlier quenching of the claddings the removal of the heat stored in
the pellets is accelerated for increasing gap heat resistance and nominal rod
geometry.
Grid spacers cause significant cooling enhancement downstream during the time
span at which maximum cladding temperatures occur. The effect is more pro
nounced for REBEKA rods than for FEBA rods.
Ballooned Zircaloy claddings, forming e.g. a coplanar 90 percent blockage,
are quenched substantially earlier than thickwall stainless steel blockage
sleeves, and even earlier than undeformed rod claddings.
The most recent version of COBRA-TF, "a best-estimate" computer code, deve
loped as part of the FLECHT-SEASET program, has been used to simulate se
lected test data. The comparison of measured and calculated data demonstrates
the capability of the code for reflood applications.
The results obtained comparing FEBA and SEFLEX data suggest a higher safety
margin than evaluated from gapless heater rods for the coolability of blocked
as weIl as unblocked PWR cores under LOCA conditions.
-51-
8. REFERENCES
[lJ P. Ihle and K. Rust:"FEBA - Flooding Experiments With B10cked Arrays."a) Evaluation Report, KfK 3657, March 1984,b) Data Report 1, Test Series I Through IV, KfK 3658, March 1984,c) Data Report 2, Test Series V Through VIII, KfK 3659, March 1984.
[2} M. J. Loftus et a1.:"PWR FLECHT-SEASET, 21-Rod Bundle F10w B10ckage Task,Data and Analysis Report."NUREG/CR-2444, EPRI NP-2014, WCAP-9992, Vo1. 1 and 2, September 1982.
[3] C. A. Cooper, K. G. Pearson and D. Jowitt:"The THETIS 80% Blocked Cluster Experiment,Part 3: Forced Reflood Experiments."AEEW - R 1765, September 1984.
[4] S. A. Fairbairn and P. D. G. Piggott:"F1ow and Heat Transfer in PWR Rod Bundles in the Presence of B10ckagedue to C1ad Ballooning."a) Experimental Data Report, Part 1, TPRD/B/0458/N84, May 1984,b) Experimental Data Report, Part 2, TPRD/B/0511/N84, November 1984,c) Experimental Data Report, Part 3, TPRD/B/0512/N84, November 1984.
[5] H. Adachi et al.:"SCTF Core I Reflooding Test Results."NUREG/CP-0041, Vol. 1, January 1983, p. 287.
[6] Y. Murao et a1.:"Findings in CCTF Core I Test."NUREG/CP-0041, Vo1. 1, January 1983, p. 275.
[7] C. Vitanza et a1.:"B1owdown/Reflood Tests With Nuc1ear Heated Rods,(IFA-511.2). "OECD Halden Reactor Project, HPR 248, May 1980.
[8] T. Johnsen and C. Vitanza:"Blowdown/Reflood Tests With SEMISCALE Heaters,(IFA-5ll.3, Data Collection)."OECD Halden Reactor Project, HWR 17, May 1981.
[9] T. Johnsen and C. Vitanza:"Results of LOCA Tests With SEMISCALE Heaters,(IFA-511.5). "OECD Halden Reactor Project, HPR 313, May 1984.
[10] C. Vitanza and T. Johnsen:"Results of Blowdown/Reflood Tests With REBEKA E1ectric Simulators,(IFA-511.4, Cyc1e 2)."OECD Halden Reactor Project, HWR 85, May 1983.
[11] C. L. Mohr et al.:"LOCA Simulation in National Research Universal Reactor Program. 1I
NUREG/CR-2528, PNL-4166, April 1983.
-52-
[12] F. J. Erbacher:"Interaction Between Fue1 Clad Ballooning and Thermal-Hydraulics in aLOCA. "KfK 3880, Vo1. 1, December 1984, pp. 299-310.
[13] R. C. Gottu1a:"Effects of C1adding Surface Thermocouples and Electrica1 Heater RodDesign on Quench Behavior."NUREG/CR-2691, EGG-2186, February 1984.
[14J B. D. G. Pigott and R. B. Duffey:"The Quenching of Irradiated Fue1 Pins."Nuc1ear Engineering and Design, Vo1. 32, 1975, pp. 182-190.
[15] V. K. Dhir and I. Catton:"Reflood Experiments With a 4-Rod Bundle."EPRI NP-1277, December 1979.
[16] V. K. Dhir, R. B. Duffey and I. Catton:"Quenching Studies on a Zircaloy Rod Bundle."Journal of Heat Transfer, Vo1. 103, No. 2, May 1981, pp. 293-299.
[17] N. E. Kaiser and O. Rathmann:"Study of Rewetting and Quench Phenomena by Single Pin Out-of-PileExperiments, With Special Emphasis on the Effect of the Fue1 PinComposition."Commission of the European Communities: Seminar on the Resu1ts ofthe Indirect Action Research Programme, Safety of Thermal WaterReactors (1979 - 1983),Brussels, Be1gium, October 1-3, 1984.
[18] T. C. de Boer and S. B. van der Molen:"Heat Transfer to a Dispersed Two-Phase F10w and Detailed QuenchFront Velocity Research."Commission of the European Communities: Seminar on the Results ofthe Indirect Action Research Programme, Safety of Thermal WaterReactors (1979 - 1983),Brusse1s, Belgium, October 1-3, 1984.
[19] M. K. Denham and D. Blackburn:"A Study of Rewetting Propagation Over Zirca10y Under Bottom FloodingConditions."Commission of the European Communities: Seminar on the Results ofthe Indirect Action Research Programme, Safety of Thermal WaterReactors (1979 - 1983),Brusse1s, Be1gium, October 1-3, 1984.
[20] P. Ihle and K. Rust:"SEFLEX - Fuel Rod Simulator Effects in F100ding Experiments,Part 2: Unb10cked Bundle Data."KfK 4025, March 1986.
[21] P. Ihle and K. Rust:"SEFLEX - Fue1 Rod Simulator Effects in Flooding Experiments,Part 3: B10cked Bund1e Data."KfK 4026, March 1986.
-53-
[22J P. Ihle, K. Rust und H. Schneider:"Brennstab-Simulator-Effekte in Flutexperimenten, (SEFLEX-Programm)."In: Projekt Nukleare Sicherheit, Jahresbericht 1983,KfK 3450, Juni 1984, S. 4200/97-111.
[23] P. Ihle, K. Rust und F. J. Erbacher:"Temperatur- und Wiederbenetzungsverhalten von Brennstäben beimKühlmittelverluststörfall: Einfluß des Spaltes zwischen Pelletund Hüllrohr ."Jahrestagung Kerntechnik '84, Frankfurt, 22.-24. Mai 1984, S. 65-68.
(24) P. Ihle:"Vergleich des thermischen Verhaltens verschiedener BrennstabSimulatoren unter DWR-Notkühlbedingungen."Commission of the European Communities: 7th Project Review Meetingin Area A on LOCA-ECCS (Loss-of-Coolant Accident - Emergency CoreCooling Systems),Ris~ National Laboratory, Denmark, June 13-15, 1984.
[25] P. Ihle, K. Rust and F. J. Erbacher:"Dispersed F10w Reflood Heat Transfer in Rod Bund1es of DifferentFue1 Rod Simulator Design."NUREG/CP-0060, December 1984, pp. 575-582.
[26] P. Ihle, K. Rust and F. J. Erbacher:"Quenching of Rod Bundles of Different Fuel Rod Simulator Design."NUREG/CP-0060, December 1984, pp. 171-178.
[27] P. Ihle, K. Rust and F. J. Erbacher:"Grid Spacer Effects in Reflooding Experiments Using Rod Bundlesof Different Fuel Rod Simulator Design."NUREG/CP-0060, December 1984, pp. 673-681.
[28J M. Nishida:"Fuel Rod Simulator Effects in Flooding Experiments, Single RodTests (Zry-Cladding)."KfK 3786 B, September 1984.
[29J F. J. Erbacher, P. Ihle, K. Rust and K. Wiehr:"Temperature and Quenching Behavior of Undeformed, Ballooned andBurst Fue1 Rods in a LOCA."KfK 3880, Vol. 1, December 1984, pp. 516-524.
[30] S. L. Lee and P. Ihle:"A Study of Mist Cooling Enhancement from Grid Spacers in LOCA Refloodof a PWR Combined Gross Heat Transfer and Loca1 Temperature and LDADroplet Sizing Analysis."NUREG/CP-0058, Vo1. 1, January 1985, pp. 286-306.
[31] P. Ihle and K. Rust:"Grid Spacer Effects on PWR Reflood Heat Transfer Measured in Bundlesof 5 x 5 Rods With Zircaloy Claddings and Pellets."Proceedings of 23rd ASME, AIChE, ANS National Heat Transfer Conference,Denver, CO, U.S.A., August 6-9, 1985.
-54-
[32] K. Rust, A. Singh, R. B. Duffey and P. Ihle:"Effects of Fuel Rod Simulator Geometry on Reflood Behavior Following aLOCA."Proceedings of 23rd ASME, AIChE, ANS National Heat Transfer Conference,Denver, CO, U.S.A., August 6-9, 1985.
[33] P. Ihle, K. Rust und H. Schneider:"Fue1 Rod Simulator Effects in F100ding Experiments (SEFLEX)."In: Projekt Nukleare Sicherheit, Jahresbericht 1984,KfK 3550, Juni 1984, S. 15-16.
[34] K. Rust:"Reflood Behavior of Rod Bundles Having Fuel Rod Simulators of DifferentDesign."EPRI NP-4l03-SR, July 1985.
[35] P. Ihle and K. Rust:"PWR Reflood Experiments Using Full-Length Bundles of Rods With ZircaloyCladdings and Alumina Pellets, (Results of the SEFLEX Program) ."Proceedings of Third International Topical Meeting on Reactor ThermalHydraulics, Vol. 2, Session 13, Paper 13.H,Newport, RI, U.S.A., October 15-18, 1985.
[36] A. Singh, K. Rust, R. B. Duffey and P. Ihle:"The Effects of Thermal Diffussion and Gap Conductance on QuenchVelocity During Bottom Reflooding of Rod Bundles."Proceedings of Third International Topical Meeting on Reactor ThermalHydraulics, Vol. 2, Session 13, Paper 13.G,Newport, RI, U.S.A., October 15-18, 1985.
[37] "COBRA/TRAC - A Thermal-Hydraulics Code for Transient Analysisof Nuclear Reactor Vesse1s and Primary Coolant Systems."NUREG/CR-3046, PNL-4385, March 1983.a) M. J. Thurgood et al.:
Vol. 1: "Equations and Constitutive Models."b) M. J. Thurgood and T. L. George:
Vol. 2: "COBRA/TRAC Nummerical Solution Methods."c) M. J. Thurgood et al.:
Val. 3: !lUsers' Manual. 1I
d) M. J. Thurgood et al.:Val. 4: "Developmental Assessment and Data Comparison. 1I
e) A. S. Koontz and J. M. Cuta:Val. 5: "programmers' Manual."
[38J J. M. Kelly and R. J. Kohrt:"COBRA-TF: Flow Blockage Heat Transfer Program."NUREG/CP-0048, Vol. 1, January 1984, pp. 209-232.
[39] J. M. Kelly, C. Y. Paik and L. E. Hochreiter:"The Analysis of the FLECHT-SEASET Flow Blockage Data With COBRA-TF."NUREG/CP-0057, Vol. 1, October 1984, p. 324.
[40] V. Casal, S. Malang and K. Rust:"Thermal and Mechanical Behaviour of PWR Fuel Rod Simulators for LOCAExperiments. 11
KfK 3331, May 1982.
-55-
[41] S. Malang:"HETRAP - A Heat Transfer Analysis Program."ORNL-TM-4555, September 1974.
[42] K. Rust, S. Malang und W. Götzmann:"PEW - Ein FORTRAN IV-Rechenprogramm zur Bereitstellung physikalischerEigenschaften von Werkstoffen für LWR-Brennstäbe und deren Simulatoren."KfK-Ext. 7/76-1, Dezember 1976.
[43] N. B. Vargaftik:"Tables on the Thermophysical Properties of Liquids and Gases."John Wiley & Sons, Inc., New York 1975.
-56-
Steam Generator
Hot Leg(1 of 4)
[ore
PressurizerHead Pump
IReactor Pressure Vessel
_________ Upper Plenum
'-
Cold Leg(1 0 f 4)
Figure 1. 4-loop steam generator system and pressure vessel
with installations of a pressurized water reactor.
-57-
160
high ratingFq.2.5
140120
normal ratingFq.l.2
10060 80time [s]
inlernal rod pressure :70 bar
pressure differenceacrass cladding
'-lJJJU ~LI
14----- reflood -----------
40
r' blowdown
1000
'U~900
Ö 8000..III
] 700
ö600f!!
::J
6' e SOO...a 0.=200 ~4oo0. -~ ISO r300~ "0~ 100 E2000. u
i 50 100-g!,III 0 0
0
Figure 2. Fuel rod cladding loading in a 2F-cold leg break LOCA.
-58-
1100 ~---------------------,
o
~::J-~Q)0.EQ)
IClC
"Cl"Cl
'"Ü
1000
900
800
700
600
500
400
300
200
100
-- Run 5246
----. Run 5263 }_.- Run 5264
-- Run 5265
Nuclear fuel rod
SEMISCALEheater rads
OL-__.L-__-l-__----l .l...-__--'
o 60 120 180Time (sec)
240 300
Figure 3. OECD Halden Reactor Projeet: Comparison of nuelear
fuel rod and SEMISCALE heater rod responses.
-59-
800 r-------------------,
300240
I--
I--
I--
-- IFA-541 Nuclear fuel rods700 _ k,:,:",::::q IFA-511,4 REBEKA tuel rod simulators
/".......,I ,,,,..... \
//""'''Qt+~ 4% higher power
I~I ~11V11 \1;\
/1 FilI/ t':;/:I/\J
'1 h:"lI Fil
300 I-- I r'l
200 ~\J' 111l;1"',100 L-_...1..--1 _----l1__.L-1 _--l..1_----l
o 60 120 180Time (sec)
~
U~ 600~:J
"§ 500Q)a.EQ)
I- 400Clc:"0"0coU
Figure 4. OECD Halden Reactor Project: Comparison of nuclear
fuel rod and REBEKA fuel rod simulator responses.
-60-
20 20
11
LEGEND
1 Water Supply
2 Steam Supply
3 Storage Tank
4 Water Pump
5 Filter
6 Heat Exchanger
7 Throttle Valve
8 Turbine Meter
9 Water Level Regu lation
Valve
2
•
- -67
10 Lower Plenum
11 Test Section
12 Upper Plenum
13 Water Separator
14 Power Supply
15 Rod Instrumentation Exits
16 Water Level Detector
17 Water Collecting Tank
18 Outlet Valve
19 BuHer
20 Pressure Regulator
21 Filling Gas Supply
Figure 5. FEBA test loop used for SEFLEX tests.
-61-
Legend
3 Storage Tank
4 Water Pump
9 Water Level
Regulation Valve
10 Lower Plenum
11 Test Section
12 Upper Plenum
14 Power Supply
15 Rod Instrumentation
Exits
17 Water Collecting Tank
18 Outlet Valve
19 BuHer
21 Filling Gas Supply
Figure 6. Photograph of FEBA/SEFLEX test rig.
--~
11-.,1',
~-~
I
stainless steel
Rod No.
-62-
6.5
~~~~~~~~
78.5
10.75
14.3
Figure 7. Cross-sectional view of a 5 x 5 rod bundle.
-63-
(/)......X Ol
QJ W C
"0 C (/) 0
C 0 :::J 0 :><:::::J ...... 0 :3 E
CD ro .c.""-...... 0
"0 C 0.... c ro :3 LJ•0 QJ<] 0 :><::Cl::: Ol E lf1 C>
C ...... • C>:::J ro C r- lf1
lf1 (/) '- "0 0
X :::J ...... :::J :::J ......0 (/) (/) '-
'- 0lf1 :r: C ..... c 1-"< .....
e0:jj0<llCIl
....CIl<ll....XW..Ju..W(I)
'-<C:lWu......0
E ~
E <ll':;
C>lf1 '"...j-
e0:jj0<llCIlICIlCIl0'-U
00
<ll'-:::J
.2lu..
ww OS,]
~
I, TC lead out
-
Grid Plate WithSquare Holes(Same total (rosssection tor(ooLant throughflow as for FEBAI
TC lead out
Steam OutletIto bufferl
75_______..1 _
Top End ofHeated Zone
Steam Outlet(to buffer)
Steam Supply duringPreheating Phase
I '1~~~~~~~~~~1t-~z~e~r~o~Level (referenee levelJJ - - for all axial bundle
positions)
0100A ...-..Temperature'1'~l Position Measuring
] . /Pressure. /;./ ~ \1>280=~ - Position Measuring
Housing
5 x 5 Rod Bundle
Upper Plenum
Separator
Water Outlet~(t0 Water t:::::::::.eolleding tank)
Grid Plate (36 holes of10 mm diameter foreoolant trough 110w)
Pressure Balance
FEBA SEFLEX
Figul"e 9. Ol"iginal and modified uppel" bundle end and plenum of FEBA/SEFLEX test section.
-65-
5 x 5 Rod Bundle Water
Pressure Balance It LevelHousing identical
---~
I I I II I I
LowerPlenum identical
I1 I11, I II!. ,
I i
111I IO-Ring Sealing I I 111
I I I 11
FEBA SEFLEX I I 11
11 11
I I 11
Rod InstrumentationREBEKA I IRods
Exits I I I IPower Supply 11 I1
11 11
WeightsI 1 11
11 [ I11
Electrical11 11Ground
11 I I
IConnection to Heliumor Argon Gas Supply
Figure 10. Original and modified lower bundle end and plenum
of FEBA/SEFLEX test section
Cladding (Ni Cr 80 20)
Insulator (MgOI
-66-
Filler Material (MgOI
Heater Element (Ni Cr 80 201
Thermocouple
-------- 8.65 -------I
-------10.75 --------1
Figure 11. Cross section of a FEBA heater rod.
-67-
Annular pellets (Alz~l
Magnesium oxide
Inconel heaterrod sheath
,--- Zircaloy cladding
r-- Heater (Inconel)
Thermocouples
Figure 12. Cross section of a REBEKA fuel rod simulator.
69-1" " !
, , ,--
H H
Pcllesiiule bei MontQQe 4000 lc.
7 r/ l I i I
oben
i I, 3900 beheitzle Länne ,! 432010.Iheitzte länne
40
i
, '080
B' j: )'22 "16 F
I I '~A A A A A A A - +- :- ~\ unlenVVyyyyyyy
1-/I, , . __ ._.-_. -_. _._.-_. __ ._-
bottom
3900 "0 298end F
F 432O2220 bis Mille beheitzte L
'581
~0 Zr4 tubex~ Legend<-
~"00
".c ~ 1 Heater Rod~
I0 m
'" 2 Soldering Tube3 Cold Welding (55 - Zr4)I .. . 4 SpringpSVY'ÜSSSjll ,I" Eff9+E&Fi'H;:S:~ 5 Connection6 Alumina Pellets7 Tul?e Plug8 Zircaloy-4 Tube
, 9 Flexible Copper Tape ,10 Connection11 Check Nut12 Stainless Steel Tube13 Tube14 Spiral Coil15 Screw Cap
I 16 Nut17 Top Cap
"-18 Screw ,,
, I // 19 Nut
I 20 Screw21 Sealing22 Sealing
j!111 23 Sealing
4i111 I,!.II ---.- ---....... Klrnlorschunguentlum1 I'
~, ,J I- .. ... _ - w =- Karlsruhe 0..111,,lll !±IJ .. H " . , ,_I:. - .~~
,, I I' • , - -- BE-SimulatorUnl'
llon 18. I.
~ (5e"e.)111111 Figura 13. Worki •.,..- -' hoI 1· 1
.f;::;-~. PN54260 R2.39-1-03a,lrl 1 "" " " - ,- "'" ~~. CrIt1z1do>"'"
...."'_ ...'w.
390
Axial Location oflirid Spacers
-71-
Axial Rod Power
Profile P/Paverage
o"",==R=e=f=e=re=n~c~e~L~e=v=el:::;:- It-__ Upper End75 + ll'l :f·T' of Housing
tO ::: ~ ::gj375 d t·'V
I
1375+---..!*-~:
.... 1-.
r- .....
935
1480
2675 +---"ffi!~
3275 +-----"'i
3675+-------";,;,;;:
......... 2570
3115
3660
3975+----"""'----'-" 114 "'---- +t__ Lower End
of Housing
Figure 14. Axial power profile and loeation of grid spaeers
of FEBA and REBEKA rod bundles in SEFLEX tests.
-72-
FEBA Rod
(ross Sectionat Midplane ot the Bundle
Local Blockage Ratio 90%Overall Blockage Ratio 31%
Sleeve .--Ir
'"
~l
u-I
14. 3 mm10.75mm
3900 mm
Bundle Data:
PitchRod DiameterHeated Length
I I
.L-----.11
~.Flow Area~ r
,Gap Filled with Stagnant Steam
Figure 15. Sectional view of the 90 percent blockage
with bypass realized for FEBA tests.
tt-+~ -1~---HH-1+-
~~l
-73-
Artificially balloonedZircaloy-cladding \ c-
(ross Seetionat Midplane of the Bundle
Local Blockage Ratio 90%Overall Blockage Ratio 31%
, I
Pellet column
~m
Bundle Data:
PitchRod DiameterHeated Length
14.3 mm10.75mm
3900 mm
Flow
Helium or Argon in the Gap of > Zmm Width
Figure 16. Sectional view of the 90 percent blockage
with bypass realized for SEFLEX tests.
" " I 75 , , ,
H H
Ii // . :
'I-E J j71
,, ,ob.n unten
t f ~I
rßJ
F F
LTE4
LTE 3LTE2
mulator Typ LTEI L TE 2 L TE 3 LTE4 Blockade Heizleiter TypLTEI
a 2430 2975 3520 4065, ,a 2430 2975 3520 4065 R2.39-2-23 ab 250 795 13&0 1885b 250 795 1340 1885 R2.39-2-23 b
TEl f 2330 2430 2530 2630- f 2330 2430 2530 2630 R2.39-2-24
oben 9 1830 1930 2030 2130TE2-T 9 1830 um ImID [ID1J R2.39-2-2l.
! J 1430 1530 1630 1730
0 k 305 405 505 605
-l0Fräserausll I 2130 2180 2230 2280 I
I (Nm 2180 2230 2280 R2.39-2- 23 I
o nichl fräsen, nur anreißen
, j um k 9 ,.'
" ~~.,..'/.<>
~"'.>.
"'+,
o~TEl ~TE4 '3 I
~ ~j~-o .~• ~E2 •". ~ATE
T X / "Hili Radius am Nutgrund ,HI·j 0,3 bis O,Sx Breite'1,11I!'! 1hlllJ --"- ---.....
~I(lflilofSthungszllIlrllfTl
" 1 ~ » '... _ ? '"' W'OV'RV
, 'ilj.J Klrlsruhe OmbH, H ±U H~ ..~ .. ~ , , ,
I, t j! ,~. : - "'_lOH, , - --.llUII - Blockad.n und TE'111 I Figura 17. Working drawini ~ AnordnungI 1~11
.,..- -- h,,,!l I P N 5 4260 R2.39-131
I " " " ,., ...... - .- ~,. [»,li ru.. EntIII""''''
...,..._.~" ..
-77-
Figure 18. Photograph of the SEFLEX 90 percent blockage after the tests.
-78-
-105
45L ,..J
100
200
JOO JO. 28531.!L * '00
'"590 590
9J2.. *1135 1135 1135
1225
1325
1~ *1425
1525
1625 1625 16251660
17251725
18251635
1835 18351875
1925 1925 1925 19252~ ** 2025
1975 20252025 2025,.2125 2125 2125
2225 2225 2225 221,5 2235
2325 23802]80
2425
25lQ... *2770 2770 2770
30]6 ]01831~ *
3315 3315 33 15
36.§1..
]725
3825 ]820]860
3925]932 - 3915
4025 4012 4091
e--.!~
0~
u~
~
;Z~ v>~ = .c -'< v> 0 ~ Cero .c -v> ~ a; 0 c;:
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ " 2 :r: Ce ~~ a. a. a. a. a. a. a. a. a. a. 0 > f-u >. >. >. >. >. >. >. >. >. >. N ro ~
~ ,ro f- f- f- f- f- f- f- f- f- f- ~
~-" ~ ~ =a. =
~~ f-
~V> .~ ~ v> ,~ ~ ~ ~ v> .~
v>~ ro ~ . ~
~ ~." ~ro ~ ro
~~ 0
Rod-TC ITSI~, 3 D- u: :r:CJ :r: o..~
* Grid spacer ins trumen ted
** Location of 90% blockage In serles 3 and 4
Figure 19. Schematic diagram of SEFLEX instrumentation
for unblocked and blocked rod bundle tests.
-79-
TC No.
,..-----,~~Rod No.
'--~'t--Rod Type
---ti"~- TA
~+-TK
TA
Rod TC AxialType No. Level
mm
a 1 22252 27703 33154 3860
b 1 452 5903 11354 1680
c 1 37252 38253 39254 4025
d 1 20252 20253 20254 2025
Rod TC AxialType No. Level
mm
f 1 21252 22253 23254 2425
g 1 16252 17253 18254 1925
h 1 19252 20253 21254 2225
i 1 18752 19253 19754 2025
Rod TC AxialType No. Level
mm
j 1 12252 13253 14254 1525
k 1 1002 2003 3004 400
x without TC's
Figure 20. Radial and axial positions of cladding, grid spacer, fluid,
and housing TC's for unblocked rod bundle tests.
-80-
TF
TF
Rod TC AxialType No. Level
mm
a 1 22252 27703 33154 3860
b 1 452 5903 11354 1680
f 1 21252 22253 23254 2425
Rod TC AxialType No. Level
mm
g 1 16252 17253 18254 1925
j 1 12252 13253 14254 1525
k 1 1002 2003 3004 400
Rod TC AxialType No. Level
mm
1 1 19252 19753 20254 2075
x without TC' s
Figure 21. Radial and axial positions of cladding, heater sheath, grid
spacer, fluid, and housing TC's for blocked rod bundle tests.
-B1-1000
BOO Plot D
l-' 600'"'":::>....er
~OO'"'"..>:
'".... 200
Level 590 "'"00 100 200 300 ~OO 500 600 700 BOO
TIHE 51000
600 Plot (
u 600•'"'":::>....er ~OO
'"'"..>:
'" 200....
0Level 1135"",
0 100 200 300 ~OO 500 600 700 600TIHE 5
1000 ---Plot B - - - - - -- 590mm
600
- - - - - -- 1135mmi-' 600'"'" - -- - - -- 1680 mm:::> Healed Lenglh....er
~OO 3900 mm'"'" - - -- 2225 mm.. - -->:UJ.... 200
Level 16BO "'"00 100 200 300 ~OO 500
1000TIHE S
---BOO Plot A 1'1'1"1'
FLoodlng rate 3.B ernteI-' 600 System pressure 2.1 bar'"'":::>.... 1!J FEBA lest. No . 223er
~OO FEBR rod bund Le'"UJ Gepless rode..>:UJ
~ SEFLEX tost No. OS.... 200RE6EKA rod bundloHOllum-f I ltod gops
Level 2225 "'"0 & SEFLEX tost No. 070 100 200 300 ~OO 500 AEBEKA rod bundle
TIHE 5 Argon-fl LLed gaps
Figure 22. Cladding temperatures measured at four different
axial levels in FE BA and REBEKA rod bundles.
200
-82-
1000
800
600.t--=::::::,jI::===:::::::::::J~::=--e!l--"''":::>....~ 1100"'..x"'....
o level 1680 nm
o 100 200 300TIHE. 5
~oo 500
Floodlng rete 3.8 cm/sSystem pressure 2.1 bar
1!J FE8A tes t No. 223FEBA rod bundleGepless rads
~ SEFLEX test No. 05REBEKA rod bundleHe LI um- f j LLed gaps
A SEFLEX lesl No. 01REBEKA rod bundLeArgon-filLed gaps
_. - -- 1680mmHealed lenglh -+++=+=lo
3900 mm
1"1'1"1'
Figure 23. Housing temperatures in FEBA and REBEKA rod bundles.
-83-
500~oo200 300TIME. 5
16
I~
12
N
••>: 10u,'"X 8=>-'"-0- 6a:
"':x:
~
2
Lave l 1660 lMl0
0 100
FLooding rate 3.8 cm/sSystem pressure 2. I bar
~ FEBA test No. 223fEBA rod bund LaGapLess rads
~ 5EFLEX lesl No. 05AEBEKR rod bundleHolium-fitLed gaps
.6 SEFLE'X test No. 07REBEKA rod bundleArgon-fi lLed gaps
_. - --1660mmHeated lenglh 404=1=1=10
3900 mm
1"1'1"1'
Figure 24. Surface heat fluxes af FE BA and REBEKA rads.
10001 12.5
0" 0.00 20 ~o 60 80 IOD 120 J~O
TIME. 5
Flooding rale 3.8 cm/s Axial Level 1680 mmSystem pressure 2. 1 ber- TC 12.b.~
FE BA test No. 223 5EFLEX test No. 05 SEFLEX test No. 07FEBR rod bund Le REBEKA rod bundLe REBEKA rod bundLeGepless rods Helium-fi LLed gaps Argon-fi lLed gaps~ CLadding temperature ~ Cladding temperature 6. Cladding lemperature+ Surfece hest flux X Surfece hest fLux Y Surfece haat (Lux
8001 10.0 -r ~'-/~
I -tt±tt-'680 mmru• Heatea le"9ln
• 3900 mm
"600 u 7.5u -• ""-..; '"'"::> x- ::>a: ~
'" u."'..
~OO - 5.0'" a:
"' "'- :z: , , , ...-.HH I
co...200j
, , ,2.5
Figure 25. Cladding temperatures and surface heat fluxes of FE BA and
REBEKA rods during the early portion of reflooding.
-85-
ijOOO
500ijOO0~L:!!e~V!e~l-..!..l16~8!!0~"",!"..- ~ ~!:;;;;;~~~~~~~
o 100 200 300TIHE. 5
'"u~ 3000
~
~ 2000x:ow~ 1000~
'"
FLoodlng rate 3.8 cm/sSyslem pressure 2. I bar
~ FEBA lesl No. 223FEBR rod bundleGapless rads
~ 5EFLEX lesl No. 05flEBEKR rod bundLeHelium-fi lled gaps
" 5EFLEX lesl No. 07REBEKR rod bundleRrgon-f ilLed gaps
- - - --1680mmHealed lenglh -4=#=1='1>
3900 mm -
Figure 26. Release of stored heat from FEBA and REBEKA rods.
10'
50 100 150 200 250TIME. 5
300 350 ~oo ~50 500
FLooding rate 3.8 cm/sSystem pressure 2.1 bar
~ FEBR test No. 223FEBA rod bund LeGap Less rads
~ 5EFLEX test No. 05AEBEKR rod bundLeHeLium-fj lLed gaps
A_ieL level 1680 mmTC 12.b.~
~ SEFLEX test No. 07AEBEKR rod bund~e
Argon-fi LLed gaps
Figure 27. Heat transfer from FEBA and REBEKA rods (related to coolant saturation temperature).
-87-
PLoL A800
Claddlng lemperetures
: 6oolFC:~~q!,=,=:l..t~__a: .~ '"~ LIDOa:a:
'"..~ 200~
I\,
0+----.------11------.-----:':"-20 -10 0 10 20
T1HE. 5
PLoL BN 200••>:u- 150,'"~ 100-'...~
:5 50>:
Ol~~~~~+_.....::~==fl::::==:t--20 -10 0 10 20
TlHE. 5
PLot C::: 10'><•N~ 10 1>:~,~ 10°C>
-'
10--~~~~~~----1---_c-----=-c-20 -10 0 10 20
T1HE. 5
PLo L 02000
>:u
~ I soor~-el..."=#==:::j:r_____~
a:~ 1000c
'"<C:= 500
'"0,L-----.-------l------=~~-20 -10 0 10 20
T1HE. 5
Surfece heet fLuX8S
Heet transfer coefficlenls
Storad heut quentltlss
Axlel level 1680 mm
Floodlno rete 3.8 ernteSystem pre99ure 2.1 bar
l!1 FE8R tost No. 223FE8A rod bund LuDuench time """"0 s
(!)SEFLEX tost No. OSAEBEKA rod bundLeHellum-fl lled gapsOuonch lImo 306 •
Figure 28. Cladding temperatures, surface heat fluxes, heat transfer,
and heat release during quenching of FE BA and REBEKA rods
(FEBA test No. 223, SEFLEX test No. 05).
-88-
P Lo l A800
Cleddlng lemperelures
w
'"~ ljOOco
'"w..~ 200~
"('1
o-l----~---+---;;_;_---;;:;_-20 -10 0 10 20
TlHE. 5
PLo l B;;; 200••'"u- 150"'"~ 100...JU.
~
~ 50
'"o~~~~~~-20 -10 0 10 20
TIHE. 5
PLol C~ 10 2
;;:•N: 10 I
'"u"~ 10°C>..J
10-'~::!:~~~~----;-:;----;:;--20 -10 0 10 20
TI HE. 5
PLo L 02000
'"~ lsoof===l!t:::::::=l!l:::::'=l!F==J~
co~ 1000ow
'"~ 500...o+-----.----+----:-!:---:;:;-20 -10 0 10 20
TI HE. 5
Surlace hesl (Luxes
Heul transfer coerrlclenls
Stored heat quentilles
Axial Level 1680 mm
FLoodlno rate 3.8 cm/sSystem pr&9sure 2.1 ber
1!I FE8R lesl Ho. 223FE8A rod bund luOuench LI me '&'&0 Si
~SEFLEX leel Ho. 07RE8EKA rod bundleArgon-filled gapsQuench time 270 s
Figure 29. Cladding temperatures, surface heat fluxes, heat transfer,
and heat release during quenching of FE BA and REBEKA rods
(FEBA test No. 223, SEFLEX test No. 07).
-89-
f' Lo l A Claddlng lempereluresBOO
wa:~ liDOa:a:w...~ 200...
20o 10!lNE. 5
-10o+------.--------f----r-----,-20
PLol B Surfete hest (luxesN 200••'"u::: ISO:x
~ 100-'lL...::5 SO
'"
-10 o 10!lNE. 5
20
=10'
'"•N: 10'>:~....;! 10°Cl...J
PLol C H88l transfer coerrlclenls
20o 10TlNE. 5
10-'F:--'----r----4-----.-----.--20 -10
PLo l 0 Slored hiat quentltl892000
Axle\ \eve\ 16BO nvn
FLoodlng reh 3.8 cm/,System pres9ufe 'I. I bar
I!IfEBA lesl Ho. 216fEBA rod bund\eQuench time 3DlI 9
1!15EfLEX leel Ho. 03REBEKA fod bundleHellum-fllted gapsQuench time 21.18 920o 10
TINE. 5-10
o-l----~--____j---__,_---___r_-20
>:u
~ 1500r--s-~;::::::~::::::J
cwa:~ 500
"'
...a:~ 1000
Figure 30. Cladding temperatures, 5urface heat fluxes, heat transfer,
and heat release during quenching of FE BA and REBEKA rads
(FEBA test No. 216, SEFLEX test No. 03).
-90-
PLot A Claddlng lemperelures800
...a:~ 'IODa:a:.....~ 200~
20o 10TIHE. S
-10o-l----~---~"----~---~-20
PLot B Surf8C8 heet (LOX86
N 200••'"u- 150"-
'"~ 100--'...~
~ 50:>:
20o 10TI HE. S
-10o~~=~~e-l-----'::::::::':!i!==B::=~-20
PLot C Hest transfer coefflclenls
- 10';;:•N=10 I
'"~"-:! 10°co--'
-10 o 10TI HE. S
20
P Lo t 0 Stored hest Quenlltles2000
owa:~ SOOon
-10 o 10TI HE. S
20
Axial level 1680 mm
Floodlng rate 5.6 cm/sSystem pressure 2.1 bel'"
Cl fE8A losl No. 216fEBA rod bundLeOuench time 319 9
~ SEfLEX losl No. 06REBEKA rod bund~o
Hellum-fl lled gepsOuench time 21.10 s
Figure 31. Cladding temperatures, surface heat fluxes, heat transfer,
and heat release during quenching of FEBA and REBEKA rods
(FEBA test No. 218, SEFLEX test No. 06).
-91-
PLot ABOO
~ 600
"'a:~ '100a:a:
"'"-~ 200....
0+----,..-----1----,-----:0-20 -10 0 10 20
TIHE. 5
PLot BN 200••'"u- ISO"'"~ 100-'......~ 50I
o~~~~~==-t_-=:::::::J~=fl=~-20 -10 0 10 20
TIHE. 5
PLot C:: 10'><•N: 10 1
'"~"~ 100C>-'
10·'I----~---''----+----_.:_---_=c
-20 -10 0 10 20TIHE. 5
PLot 02000
'"u": 1500r--e:f--~::;"''-1:!l-.......Ja:~ 1000
""'a:~ 500V>
0+----,..-----1----,-----r-20 -10 0 10 20
TlHE. 5
Cleddlng temperelur88
Surfece he8t (Luxes
Heet transfer coerrlclenls
Stored heet quenlilles
Allel level 1680 mm
Floodlng fate 5.8 cm/!System preS8ure ~. 1 bar
1!J FEBA tost No. 21~
FEBA rod bund teQuench time 217 9
~ SEFLEX tost No. o~
REBEKR rod bundLoHellum-fl Lled gapsQuench time 181 9
Figure 32. Cladding temperatures, surface heat fluxes, heat transfer,
and heat release during quenching of FE BA and REBEKA rads
(FEBA test No. 214, SEFLEX test No. 04).
-92-
1000
Plo l A FL~Od Ing ra Lo 3.8 ern/e800 Sy.lem pre89ure 2.1 bar
SEFLEX tos t No. 07~ RE8EKA rod bundLe..; 600 Argon-flLLad gapser:::>>-
'" ~OOer:
"'"-JE:
"'>- 200
00 100 200 300 ~OO
1000TIME. S
Plot B FLoodlng rate 3.8 ernte
800 Sy, lern p res su re 2. I bar
SEFLEX tost No. 05u RE8EKA rod bundLo0..; 600 Hellum-flLled gapser:::>>-
'" ~OOer:
"'"-JE:
"'>- 200
00 100 200 300 ~OO
1000TIME. S
Plot C Floodlng rate 3.8 cmJa
800 System pres8ure ~. I bar
SEFLEX tost No. 03u RE8EKA rod bundLo•..; 600 HeLlum-flLled Qepser:::>>-
'" ~OOer:
"'"-JE:
"'>- 200
00 100 200 300 ~OO
1000TIME. S
Plo t 0 FLoodlng rete 5.8 ernts
800 System pr8ssure ~. 1 bor
SEFLEX Lost No. O~u RE8EKA rod bundLo• 600 Hellum-flLled Qeps..;er:::>>-
'" ~OOer:"'"-JE:
"' 200 Rxlal Leve l 2025>- mm
1!I Rod No. I~. TC No. I
0Cl Rod No. I~. TC No. 2A Rod No. I~. TC No. 3
0 100 200 300 ~OO <) Rod No. I~. TC No. ~
TIME. S
Figure 33. Azimuthai cladding temperatures of a REBEKA
rod measured at the bundle midplane.
-93-
OU
.J 300CI::::>....~ 200'"..:I:
'".... 100
Plol A FLoodlng rate 3.8 cm/sSystem pres8ure 2.1 bar
5EFLEX tost No. 07REBEKA rod bundtoArgon-flLled gaps
FLoodlng rate 3.8 emlsSystem pres9ure 2. 1 bar
Plol B
gi"28:----2""T2":""9----2-,3-0----2~3,-1----2-3.-2----23....-3
TlHE. 5
500t?....~~::::::::~~~,, ___...
~oo
U•·300'"oe:::>....~ 200
'"..:I:
'"l- 100
5EFLEX tost No. 05REBEKA rod bundtoHellum-flLled gaps
FLoodlng fate 3.8 emteSystem pressure ~. I bar
5EFLEX t.st No. 03REBEKA rod bundtoHeLlum-fllled gaps
2~6
P lol C
o+----.------~---_,_---~---___,_2~1 2~2 2~3 2~~
500~~3;>"<.~~"":'?~T"l[..H':$E".......5
~oo
.U·300
'"CI::::>....~ 200'"..:I:
'".... 100
20~ 205202 203TlHE, 5
201o+----.------~---_,_---~---___,_200
500 i:--..
floodlng rate 5.8 ernteSystem pre99ure ij. 1 bar
5EFLEX tost No. O~
REBEKA rod bundtoHellum-fllLed gap8
Axial level 2025 mm
1!I Rod No. 1~. TC No. ICl Rod No. 1~. TC No. 2'" Rod No. I~. TC No. 3<I> Rod No. I~, TC No. 4155
P lol 0
154
\\
152 153T[HE. 5
151O+----__._---~----~---__._---___,-150
~oo
U
•·300'"oe:::>....~ 200'"..:I:
'"l- 100
Figure 34. Azimuthai cladding temperatures of a REBEKA rod
during quenching measured at the bundle midplane.
-94-
600
ij 5 6RRDIUS. HH
Level 16BO mmO+-'-'-'-'-T-=.:...r=--,.--,c--,.--~
023
Plot A
ftoodlng rate 3.8 cm/sSystem pressure 2.1 bor
FEBR test No. 223fEBA rod bundleGepl8s8 rode
Quench time ijijO 8
ij 5 6RRDIUS. HH
BOO
Level 16BO mmO+-'-=-T-=:...r"'---~-~--r---_,_
023
Plot B
Healed lenglh3900 rnm
- - - -- 16BOmen
Floodlng rate 3.8 ernteSystem preisure 2.1 bar
SEFLEX tsst No. 05REBEKA rod bundleHeLlum-fl lled gaps
BOO
Quench time 306 •Shlfled time scele
Plot C
l!]-IO.Os(') -7.5 s.. -5.0 s+ -2.5 sX 0.09 (eurfece rewettlno)<:> 2.5 s+ 5.0 sl': 7.5 sZ 10.0 s
o Level 16BO mm
o I 2 3
FLoodlng rate 3.8 ernlsSyetem prs88ure 2.1 bar
ij 5 6RRDIUS. HH
SEFLEX test No. 07REBE KR rod bund teRrgon-II lled oapa
Quench time 210 9
Anetys I SI
OnG-dlmenslonal Inverseheat conducllon problem
Flnlte-dlfference method
Expllclt time Integration
Tempereture-dependentmelerlel propertles
Figure 35. Radial temperature profiles as function of time
during quenching of FEBA and REBEKA rods.
-95-
F100ding Velocity (cold) 3.8 cml sSystem Pressure 2.1 bar
SIe_m_ndDroplets
Liquid
Helium Argon
FEBA REBEKA
5.5 rod bundles
600 800 1000Time (s)
400200
3500i:!==i======
4000 L..---,-+----,.--,..--.,...----ro
REBEKA FEBA0 \
)Argon Helium•
500
E 1000oSGi>GI 1500--'ni°ii...
2000
- ---------
2500
3000 Grid spacer
Figure 36. Quench front progression and liquid inventory after
275 seconds in FEBA and REBEKA rod bundles.
C>><
• 10
'""'>D
>-
'"'"a:u 5
'""'>-a:'"
100
-96-
200 300TI HE, S
YOO
Wale'Col'lCClll'IQ
lank
FLoodlng rate 3.8 erntsSystem pres9ure 2.1 bar
~ FEBR teet No. 223FEBA rod bundLeGepless rods
~ SEFLEX teet No. 05RBEKA rod bundleHellum-ri lled gaps
~ SEFLEX teet No. 07REBEKR rod bundleArgon-flLled gaps
Figure 37. Water carry oyer from FEBA and REBEKA rod bundles.
-97-
Time = ISO seconds
flO"
~900
800
a: a: a:700 w w w
u u ua: a: a:.. .. .. Healed lenglh'" '" '" 3900mm 2025mm600
2750 2500 2250 2000 1750 1500 1250RXIRL LEVEL HH
Time =90 seconds
900FLOW
(}...e 800,i! HH8-E 700
~
6001750 15002750 2500 2250 2000 1250
RXIRL LEVEL HH
A HBR Program.Test No. 223FEBA rOd bundle. ga~ss rOds
OSEfLEX ProgramTes l No. OSREBEKA,od bundle. helium·lilled gaps
Flooding Velocily 3.8 cm/tSyslem Pressure 2.1 bar
1750 1500 1250RXIRL LEVEL MH
2000
Time =30 seconds
22502500
- -fLOH
• -~ "" 7
,
800
700
900
6002750
Figure 38. Influence of the grid spacer at the bundle midplane on the axial
temperature profiles in FEBA and REBEKA rod bundles.
-98-1000
P La L A FLoodlng rate 3.6 cmJa600 System pres9ure 2. I bo'
SEFLEX lool No. 05u REDEKR rod bundLe0 600 HeLlum-fllled gepswa::::>I-0:
~OOa:w..>:w
200I-
00 50 100 150 200 250
IIHE. 51000
P La l B fLoodlng rate 3.8 cm/sSy, lem pre89ure ~;\ bor800SEFLEX loot No. 03
u RE8EKA rod bundle0 600 Hellum-filled gaps..;a::::>I-0:
~OOa:w..xWI- 200
00 50 100 150 200 250
IIHE. S1000
P La l C Flood Ing rate 5.8 cm/sSystem pressure 2.1 bor800SEFLEX test No. 06
u RE8EKR rod bundte0 600 Hellum-fllLed g8pS..;a::::>I-a:
~OOa:w..xw
200I-
00 50 100 150 200 250
TlHE. S1000
P La l 0 FLoodlng rate 5.8 cmtsSys tem pres9ure ~. I bor8005EFLEX test No. O~
u RE8EKA rod bundle0 600 HeLlum-fllLed g8pS..;a::::>I-a: ~OOa:w
~ Cloddlng ol 2025 mm..(upstreem of grld)x
w200 l!>Cloddlng ot 1975 mmI-
(downstreom of grld)A Grld spacer el 2023 mm
tt.8odlng edoe)0
200 250~ Grld specer el 1990 mm
0 50 100 150 (lrelling edgelIIME. S
Figure 39. Influence of reflood conditions on cladding and
grid spacer temperatures at the bundle midplane.
'"'"I
.a 1975 mm
Cl 2025 IM>
t t t I
~11i1I I I I II . . I II I I II I I I II I I I II I I ! I, ;, .l- I
r T , , I 11I I I I
II I I III
,I III I I I
lJ. Je LL Il.:<
Grid spacer
I
'~I~I~
Qu::nch j ng of I~grld specer ~
Leeding edge~
Tre j L j ng edge~,II
Oispersed (low regime
Rod surfece hest fLuxes
5IIIII Inverted annuLer! film bOiling..
aal I , , , ,
o 50 100 150 200 250 300TIME. S
o
200
;;;z•z:
u 600 ~ 15o '-. '""'er: •:::> ><~ =>a: ~er: ...
"'§: 1400 ::t 10
"' "'~ '"
B001 20V~--""'-"'''''''''''''''.4..~_~
Cladding temperetures
10001 25
FLooding rate 3.8 cm/sSystem pressure 4.1 bar
5EFLEX test No. 03REBEKR rod bundLeHelium-filLed gaps
Axial Levels:[!] FH Lead i ng edge of gr i d specer (2025 mmJ~ 12 mm downstreem cf trei ling edge of
grid specer (1975 mrnl
Figul"e 40. Cladding tempel"atul"es and sUl"face heat fluxes at leading edge and
12 mm downstl"eam of bundle midplane gl"id spacel" (test No. 03).
-100-
1725 mm
1925 mm
1625 mm
1825 mm
GridSpilcer2025 mm
2125 mm
Ol9lence (rom grld spacerel the bundle mldplene:
~ 100 mm upslroem(!)62 rrm downstreem& 162 mm dOWflslraem~ 262 nm downslreem+ 362 mm dowoslreem
! 1 I !
Floodlng rate 3.8 tm/eSystem pressure Ij.t bel'"
5EFLEX teel No. 03RE8EKR rod bundleHellum-fllled gapsUnblocked geomelry
rt, f~~II I . I
II I I II
iI I I I
-- I I I II I I
II
I I I 11I
I I J
I
I i I 1 11I I I
II I I I I--I I I
iI
I I I
~tl
250
250
250
250
200
200
200
200
100 150TlHE. S
100 150TI HE. 5
100 150TlHE. 5
100 150TlHE. 5
50
50
so
50
P lo l C
P lo l A
Levels 2125 end 1925 mm
Levels 2125 end 1725 mm
Plo l 0
Levels 2125 end 1825 mm
Levels 2125 end 1625 mm
Plol B
Cladding temperatures and surface heat fluxes upstream and
downstream of bundle midplane grid spacer (test No. 03).
1000 12.5
800 N 10.0•:I:
U U0 600 "- 7.5..; '"'" ,.;=>0- =>« ijOO -' 5.0'" "-w"- 0-:I: «w
200w
2.50- :L
0 0.00
1000 12.5
800 N 10.0•:I:
U U
• 600 "- 7.5..; '"'" ,.;=>0- =>« ijOO -' 5.0'" "-w"- 0-:I: «w
200w
2.50- :L
0 0.00
1000 12.5
800 N 10.0•:I:
U U
• 600 "- 7.5..; '"'" ,.;=>0- =>« ijOO -' 5.0'" "-w"- 0-:I: '""' 200 "'0- :L 2.5
0 0.00
1000 12.5
800 N 10.0•:I:
U U
• 600 "- 7.5..; '"'" ,.;=>0- =>« ijOO -' 5.0'" "-"'"- 0-:I: «"' 200 "' 2.50- :L
0 0.00
Figure 41.
-101-
Fluid TC signals upstream and downstream of
bundle midplane grid spacer (test No. 03).
P Lo l 0
1925 mm
182S mm
2225 mm
1625 mm
GridSpacer
Olstance fram grld 'paceret lhe bundle mldplene:
~ 210 mm upslreem(9 62 mm downstreamA 162 mm downslream.,. 362 mm downslream
I I I I
floodlng rate 3.8 cmJsSystem preesure 4.1 bar
SEFLEX test No. 03RE8EKR rod bundleHellum-filled gapsUnbLocked geometry
'* cl ~ r<',I I
I ,
I I ,
I , ,i I 1
I i I,I ,
Ii I,I , II ,I I,
I ~ : 11, , I
I I,
I I ,
:
,
:i, ,
I,,,
~ i l+ltL
250
250
250
P Lo l A
P Lo l B
PLol C
200
200
200
200 250
I I
100 ISOTIME. 5
\00 ISOTIME. 5
100 ISOTIME. 5
u• 600,.ja:::l....a: ~OOa:"'..>:
"' 200....I
Level 1825 lMl0
0 50
1000
800
~u0 600,.ja:::l....a: ~ooa:"'..
I>:
"'.... 200
0Level 1925 lMl
0 50
1000
800
u• 600,.ja:::l....a: ~OOa:"'..>:
"' 200....
Level 22~0 lMl0
0 50
Figure 42.
1000
800
u0 600,.ja:::l....a: ~OOa:
"'..>:
"' 200....
LeveL 1625 lMl0
0 50 100 ISOTIME. 5
1000
800
~
o'"
6 1975 nm~ 1990 "'"Cl 2023 "'"
;5t 22~0 "'"
1!I 2125 "'"
t i t t
rtn:1rt1I I I II . I II I II I I II I I II I ! II
II I ,I I , I T fI
I I I
II I III I I
I I I I
~J.~~J.200
BOO
1000
w 6000
..;a::::>~
1 I U 1 1111 /In Grid spaceraoa:"'~ liDO
"'~
o I , , , , ,o 5 10 15 20 25 30
TI ME. S
FLooding rate 3.8 cm/sSystem pressure 4. 1 bar
SEFLEX tesl No. 03REBEKA rod bundleHeLium-fi LLed gaps
All iaL LeveLs:~ CLadding. 2125 mm6. CLadd ing. 1975 rrmC) Grid spacer. Leeding edge. 2023 mmo Grid specer. trai Ling edge. 1990 mm~ FLuid TC signal. 22110 mm
Figure 43. Cladding and grid spacer temperatures, and fluid TC signal on
enlarged time scale at beginning of reflood (test No. 03)-.
o'"I
A 1975 rrrn(> 1990 rrm
Cl 2023"",
X 22110 rrm
Cl 2125"",
t t t t
~r.c 1 rr~II I I I
I II I
I I I II I I I II I I I II I I ! II
I II II : I I 11
I I I I
II I I I!I I I III I I I
lJ.J.JU I..;c
Gridspacer
1501~0 1~5
TIME. S135130125
I
I
~
A ...... ~=
o 0 '\
u,L ! . Id! ! l~~ ·r
o120
200
800
1000
u 600o
"''"=>~
a:'""'~ llOO"'~
F~ooding rate 3.8 cm/sSystem pressure ll.l bar
SEFLEX test No. 03REBE KR rod bundLeHelium-fi Lled gaps
AxiaL Levels:[!] CLadding. 2125 rrm.6.CLadding. 1975 l'Ml~Grid spacer. leading edge. 2023 mm~Grid spacer. traiLing edge. 1990 mm~ FLuid TC signal. 22110 mm
Figure 44. Cladding and grid spacer temperatures, and fluid TC signal on
enlarged time scale at quenching of grid spacer (test No. 03).
1000
11
.0. 1l.l25 lMl
Cl 1~78 rrtll
~ 1~~5 rrtll
~ 1525 rrtll
l': 1625 rrtll
rr '"" irr 'iI I I II . I II I II I I II I I II I 1 II
1I II : , I 1 11
I I iII I I III
,I III I I I
L1J.J. lJ. I\.;c
200
800
u 6000..;a:::::> I 11 I '11'IAM I tfh. ~ Grid spacer~
a:a::w§: \tODw~
oI, "'"o 50 100 150 200 250 300 350 ~OO
TI HE. S
t t t t
FLooding ~ate 3.8 cm/sSystem pressure 4. I bar
SEFLEX test No. 03REBEKA rod bundLeHelium-fi LLed gaps
FlxiaL Levels:~CL.dd;ng. 1525 rrtll
b. CLadding. 142511Yn~Grid spacer. Leadin\jl edge. 14.78 mm~ Grid Spacer. trai ling edge. ll.l.l.I.S mmX Fluid TC signaL. 1625 rrrn
Figure 45. Cladding and grid spacer temperatures, and fluid TC signal
above the bundle midplane (test No. 03).
o'"
Jl: 1625 "'"
l!l 1525 "'"
.. 1~25 "'"Cl 1~7B "'"~ I~~5 "'"
I,, l : I; 11
mItI;" II I II I II !
11
II 'I I~III
I1 I II . . i II I I II I I I II I I I I: ! ! ! I
11 I I I
f1~
200
BOO
1000
u 600
•.~
a::::> 11"\ \ 1rI11~ T'I'I"r'[JN. "'A,. ~ Grid spacer~
a:a:w§: !.lODw~
t 1 t t~oo
o I , , , , " ,o 50 100 150 200 250 300 350
TI ME. S
Flooding rate 3.8 cm/sSystem pressure 2.1 bar
SEFLEX test No. 05AEBEKR rod bundLeHeLium-fiLLed gaps
Axial Levels:[!]CL.adding. 1525 mm6. CLadd ing. lLj25 mmc)Grid spacer. Leeding edge. lIn8 mmo Grid spacer. trai Ling edge. 11.14.5 mm~ FLuid TC signal. 1625 ITVTI
Figure 46. Cladding and grid spacer temperatures, and fluid TC signal
above the bundle midplane (test No. 05).
oCl
I
Cl 2025 rrm
.6 J975 rrm
rt i 1rt1I I I I II . . I II I I II I I I II 1 I I II I I ! II
I 11I II I . ' I I 11
I I I I
II I I III
,I II II I I
JJ.Jl Il.;<
5200
10001 25
ICLäOding tempera~ures
eJ I ----20
Nxx>: Quenching of
u 600 ~ 15 grid specer0 ,
'""' Lead; ng edge'~a:=> x
TraiLing edge~~ ~,,~~ =>\/.1 Grid spacer'" -'a: ~
"'.. ijOO ~ 10>:
"' "'~ rRod surface heat fLuxes
o o I , , , , , ,o 50 100 150 200 250 300
TI ME- S
t t t I
Flooding rate 3.8 cm/sSystem pressure 2.1 bar
SEFLEX test No. 05AEBEKR rod bundLeHeLium-f j LLed gaps
Axial Levels:I:l Rl lead j ng edge of gr id spacer (2025 mmJ612 mm downstream of trai Ling edge of
grid spacer (1975 mml
Figure 47. Cladding temperatures and surface heat fluxes at leading edge and
12 mm downstream of bundle midplane grid spacer (test No. 05).
CLadding lemperatures
o""
6. 11125 lI'In
Cl 1525 "'"
,.......:-::.-':<"",.,<I I I I I
I I I I [I . , I II I I II I I I II I I I II I' I I
,~I , ' ,
+t I "r '-I :
. ,
: I 11. ,[I I I I
11-Ir I I III
,I III I I I
JJ. Je l.L I..:c
Grid spacer
j'
Ouenching of
::~:I~:e:::e ~TreiLlng edg~
Rod 5urface heat fLuxes
5200
'"••>::u 600 ~ 15D ,
XWce •:> x~ :>'" -'ce ~
w~ lI00 ~ 10"'w w~ '"
800
10001 25
o olit , , , , , , , ,o 50 100 150 200 250 300 350 ~OO
TIME. 5
t i t t
FLooding rate 3.8 cm/sSystem preS$ure 4.1 bar
5EFLEX lesl No, 03AEBEKR rod bundleHelium-fiLLed gaps
~xieL Levels:Cl 1.15 mm upstream cf grid spacer (1525 mm).6.17 mm downstreem cf grid spacer (1425 mm)
Figure 48. Cladding temperatures and surface heat fluxes upstream and 17 mm
downstream of the grid spacer above the bundle midplane (test No. 03).
I~
iilI
6 It125 nrn
Cl \525"",
, , ,
I i I I I
; I I : 'I ; i I II I [ I II I I I I: ! ~ ! I
Grid spacer lJ,-H,...;-,J,-;...rr...M-+l
,
Cladding temperatures25
200
\000
800, 20
~N
~••600
r I Ouench ,ng o(u ~ 15• , ~ grtd sp~cer
w '" ~Lead;ng edgaa: ,.;:::>~ :::> ~ Trelling edgea: -'a: ..w"- ~OO ~ 10rw w~ :>:
Rod surface hest fLuxes
ot i t t
o I , , , , , , , ,o 50 100 150 200 250 300 350 ~OO
TI ME. S
Flooding rate 3.8 cm/sSystem pressure 2.1 bar
SEFLEX test No. 05AEBEKR rod bundLeH8Lium-filLed gaps
Rxiel Levels:(!] 45 mm upstream cf gri d specer (1525 rrml~ 17 mm downslreem cf grid spacer 11Y.25 mmJ
Figure 49. Cladding temperatures and surface heat fluxes upstream and 17 mm
downstream of the grid spacer above the bundle midplane (test No. 05).
ocoI
.:.. 11.125 lI1fI
Cl 152S ....
Grid spacer
200i 5 , ,Rod $urface heat fLuxes .. '1i!iI III III "T"'T T-I
t 4 t tIOJ 0'1' , , , , , , , ,
0 50 100 150 200 250 300 350 400TIME. 5
1000
125
ICladding temperatures
1 .~~800
120
;;;•
~•'"u 600 ~ IS
~ Ouench ,ng of• ......; '"a: X
~ grld specer=> l Leed; ng edge~ =>
'" -'a: ...~ TralLlng edge"'"- 400 ;: 10
'""' "'~ :I:
FLooding rate 3.8 cm/sSystem pressure 2.1 bar
5EFLEX test No. 07AEBEKR rod bundLeF1rgon-f i LLed gaps
Axiel Levels:r:l45 mfTl upstreem of grid spacer (1525 mml617 mm downslreem cf grid spacer 11425 mml
Figure 50. Cladding temperatures and surface heat fluxes upstream and 17 mm
downstream of the grid spacer above the bundle midplane (test No. 07).
-110-
1000
1825 mm
2125 mm
2075 mm
2225 mrn
1680 mm
1725 film
1615 mm
1925 mm
1975 mm
+-l+HI-H+++,- 1015 mm
(\9@)(\9(\9(\9(\9@)(\9(\9(\9000(\9(\9000(\9(\9000(\9(\9
PLol C
SEFLEX test No. 32REBEKA rod bundleHellum-fltled gapsBalLooned cLaddlngs
SEFLEX test No. 33REBEKA rod bundleArgon-(Illed gapsBatlooned claddlngs
RXlal level 2025
REBEKA rod bundle~ Bypass.rod ctaddlng(!) Blocktlge. rod c tadd IngX BLockage, healer sheolh
P Lo l B
100 200TIHE. 5
100 200TIHE. S
FLoodlng rote 3.8 cm/sSystem pressure 2. I bar
fEBR rod bund tel!J Bypass, rod cladding* Blockege. sloeve+ Blockege. rod claddlng
underneaLh sLeeve
u0
600Wa::::>~
a:~OOa:..,..
>:..,~ 200
00
1000
800
u0
600Wa::::>~
a: 400a:..,..>:..,~ 200
00
u0 600..,a::::>~
a: ~OOa:..,.. P La l R>:..,~ 200
FEBA test No. 2~1
FE BA rod bundLe
0Sleav9 bLockages
0 100 200 300 ~oo 500TIHE. 5
1000
BOO
Figure 51. Temperatures measured at the midplane of a 90 percent blockage
anJ in the blockage bypass of FE BA and REBEKA rod bundles.
-111 -
1000
1825 mm
1680 mm
1725 mrn
1625 mm
1925 mm
2125 mm
2025 mm
2075 mm
2225 mm
1975 mm
®@)®@)®®@)®@)@)CDCDCD@)@)CDCDCD@)@)CDCDCD®@)
500400•300
SEFLEX tes t No. 33AEBEKA rod bundleRrgon-fllLed gapsBellooned claddings
SEFLEX test No. 32AEBEKA rod bundLeHelium-fi lled gapsBeLlooned claddings
PLot C
AlCial lovel 1925
AEBfKA rod bundlel!lBypess.rod ctaddlng~8lockage. rod cleddingXBLockaQe. healer sheeth
PLot B
200TI ME. S
•200
TIME. S
I I
100 200TIME. S
100
100
PLa t A
o
FEBR test No. 241FEBA rod bund LeSle8ve blockages
o •
200
600
200
600
uo
"'a:=>>-~ LIDO"'"-
""'>-
00
10004
800
u0
600"'a:=>>-oe 400a:w"-
""' 200>-
00
Flooding rate 3.8 cm/sSystem pressure 2.1 ber
FE BA rod bund La~ Bypass. rod cLadding~ BLockage. rod cLedding
"'a:=>>-~ liDO
"'"-
""'>-
uo
Figure 52. Temperatures measured 10 mm downstream of a 90 percent blockage
and in the blockage bypass of FE BA and REBEKA rod bundles.
-112-
1000
600
200
u•""a::::>....:l: ijOO
""..x""....
PLo L A Floodlng rate 3.0 ernteSyst.em praS8ure ~.l bar
SEFLEX toot No. 3ijftEBEKR rod bundtoRrgon-flttod gop.
100 1SOTI ME. 5
soO+----~---~---~o
1000
200
600
Floodlng rate 3.8 ernlsSystem pres9ure ~.I bar
SEFLEX toot No. 35ftEBEKR rod bundtoHotlum-flttod gop.
PLo L B
uo
""a::::>....:l: ijOO
""..x""....
100 150TIME. 5
50O+----,..----.-----~o
1000
200
600
800
uo
""a::::>....:l: ijOO
""..x""....
PLo L C Floodlng rate 3.8 ernteSystem pressure 2.1 bar
SEFLEX to.t No. 33ftEBEKR rod bundtoArgon-fl LLed gaps
200100 150TIME. 5
50O.J.----,..----.-----~---~o
1000
Axial Level 2025 mmRod No. ij. pleced In bypass
~He8sur8d cLeddlno temperature~He8sured sheelh temperelure~ CeLculeled shseth lempereture
FLoodlng rate 3.8 cm/sSystem pres9ure 2.1 bar
SEFLEX toot No. 32REBEKA rod bundleHeLlum-fllled gaps
PLo L 0
200100 150TIME. 5
50Ol-l----~----~---~---~o
600
200
uo
""er:::>....:l: ijOO""..x""....
Figure 53. Comparison of measured and calculated heater sheath temperatures
and corresponding cladding temperatures measured at the bundle
midplane in the blockage bypass of a REBEKA rod bundle.
-113-
1000
PLa L CBOO 1625 mm
1680 mmu0 600..; 1725 mm'"=>>-a: ijOO'"w 1625 mm..>:w
200>-
1925 mrn0
LeveL 2025 nm. rod No. 17
0 50 100 150 200 1975 mm
1000TIHE. 5
2025 mm
PLa L B 2075 mm
800 2125 mm
u0 600w 2225 mm'"=>>-a: ijOO'"w
t t t t0->:w>- 200
0Level 2075 nm. rod No. 17
0 50 100 150 200
1000T1HE. 5
No. 17
PLai A No. 18BOO
u0 600..;'"=>>-a: ijOO floodlng rate 3.8 ernts'"w System pressure 2.1 bor..>:w>- 200 SEFLEX te9t No. 32
RE8EKR rod bundleLevel 2225 nm. rod No. 1B Hotlum-rlltod OOP9
0 8ellooned cleddlngs0 50 100 150 200
TIHE. 5l!l Rod cloddlno
(!) Heeter sheeth
Figure 54. Cladding and heater sheath temperatures measured upstream
and at the bundle midplane in the blacked rod cluster of a
rod cluster of a REBEKA rod bundle.
-114-
Figure 55. Cladding and heater sheath temperatures measured downstream
of the bundle midplane in the blocked rod cluster of a
REBEKA rod bundle.
Rod diameter
Rod pitch
Housing width
Housing thickness
10.75 mm
14.3 mm
78.5 mm
6.5mm
ITJ Rod 1
0 ROd2
Gw.n
-115-
w
Figure 56. Radial noding scheme of the FEBA test section
for COBRA-TF calculations.
-116-
Elevation (mm)
COBRA-TF (FEBA - test section)
3900 (75) Outlet
3600 (375)
3300 (675)
3000 (975)
2700 (1275)
2500 (1475)
2350 (1625)2250 (1725)2150 (1825)2050 (1925)1950 (2025) Midplane1850 (2125)
1650 (2325)
1400 (2575)
1150 (2825)
900 (3075)
600 (3375)
300 (3675)
0 (3975) InletI II J
I
•
•••
•
Phantom cell
Phantom cell 1----,
Grid spacer 3585 (390)
Grid spacer 1405 (2570)
Grid spacer 2495 (1480)
Grid spacer 3040 (935)
Grid spacer 1950 (2025)
Grid spacer 860 (3115)
Grid spacer 315 (3660)
• Comparison: measurement - calculation
Figure 57. Axial noding scheme of thp. FEBA test section
(fluid cells) for COBRA-TF calculations.
-117-
o
500
1000
1500
""~
2000Center rods
w,. PerlpheraL rodsw-' Houslng-'erxer
2500
3000
3500
1000500 750TEHPERATURE, 'e
2504000+--------;::=-------;:....!....---,------.--
o
FLoodlng rele 3.8 cm/sSystem pressure 4. 1 bar
SEFLEX test ND. 03REBEKA rod bundleHellum-fl lled gaps
Figure 58. Initial axial temperature profiles of claddings
and housing (SEFLEX test No. 03).
-118-
10 10 100
'''Ia:V> a:"- 8 <D 8 u 80 200>: 0u W
. '".: '"W a:6 => 6
>:60 ci 150~ V> "'a: V> ~ "'a: "' '"a: a: '"'" .. "' ..
z~ ~
~~O "' 100>: a:
0 "' '" -'
'" ~ 0 0
'" V> "' Z-' >- "' =>... 2 V> 2 ... 20 <D 50
0 0 0 0-100 0 100 200 300 ~OO
TIHE. 5
FLoodlng rale 3.8 cm/sSystem pressure ~.l bar
SEFLEX test No. 03RE8EKR rod bundleHeLlum-flLled gaps
~ FLoodlng rate, tin cold bund leiA System pressure~ Feedwaler lemperelure(!J Bund Le power
Figure 59. Flooding parameters (SEFLEX test No. 03).
-119-
1000Plot [
800 ~~
~~~~~
600~~
u ~~" ~~wa:: ~:::>.... 400a:a::w"-'" 200'"....
Level 1135 nm0
0 50 100 150 200 250 300 350
1000TIHE,S
~
~~~800~~
Plot B~~
600~~
P ~~",' ~~a:::::>.... 400a:a::
'""-'"'" 200....
LeveL 1680 nm0
0 50 100 \50 200 250 300
1000TIHE,S
~~
800 ~
~~ Plot A~~~~
.... 600 ~~ - - - - - -- 1135mm
'"~~
a:: ~:::> - -- - - -- 1680 mm.... 400 Healed Lenglha:a:: 3900 mmw"-'" - -- - - -- 2225 mmw 200....
Level 2225 nm0
0 SO 100 ISO 200 250TIHE, S
---FLoad Ing r8le 3.8 cm/s HHSystem pressure 4.\ bar
SEfLEX la.1 No. 03 --- ExperimentRE8EKR rod bundle ~ ~ ~ CD8RR- TfHellum-fillad gap.
Figure 60. Comparison of measured and calculated center rod
cladding temperatures (SEFLEX test No. 03).
-120-
1000
O~800
u6000
..;a:=>~
~OOa:a:
"'"-
'" 200"'~
Level 1815 lMl0
0 50
1000
;; 600
"'a:=>:I IWOa:"'"-
'"uJ 200~
100 150TIHE. S
PLo L B
200
PLo L A
250
; ; • ; ==_ _ 1875 mm-- 1975 mm
0~L~e.::ve::.:t~I.::9.::15~lMl:~ ,,~ ~, -~, ~,
o 50 100 150 200 250TI HE. S
FLoodlng rate 3.B cm/sSystem pres$ure ~.1 bar
t tt t
SEfLEX lest No. 03REBEKA rod bundLeHeLium-f Illed gaps
----- Experimento 0 0 C08RR- TF
Figure 61. Comparison of measured and calculated center rod cladding
temperatures downstream of the bundle midplane grid spacer
(SEFLEX test No. 03).
-121-
1000
Grid 9pecer No.l!) LeedIng edge(!) Tral Li ng edge
5lIij78 mmJIIYij5 mm)
PLa L B
Grld spacer No.[!] Lead I ng edge(!) Tralli ng edge
200175
150
150125
125
ij(2023 mm)(1990 mm)
100
10075
75TlHE. 5
TI HE. 5
50
5025
250
0
1000
800
uQ
600W'"=>>-<r ijOO'"w..xw>- 200
00
1000
PLa t A
Midplane -. -' - 4 +11=1=++
Grid 5paCer 5
ijOO
Grld $p6Cer No. 3
200 I!J Lead Ing edge 12568 mmJ(!) Tralling edge 12535 mm)
00 25 50 75 100
TIHE. 5
uo
w
'"=>><r
'"w..xw>-
Flooding rale 3.8 cm/sSystem pressure 4.1 bar
SEFLEX lest No. 03REBEKA rod bundleHelium-filled gaps
tttt----- E~perimenl
(> (> (> C08RA- TF
Figure 62. Comparison of measured and calculated grid spacer temperatures
(SEFLEX test No. 03).
-122-
1000
PLa l CBOO
u 600<:>~<:>0,.... <:><:><:>0:
:::> ~~~.... 400a: <:>0:...."-
'" <:>.... 200.... <:>~
Level 1135 nm0
0 50 100 \50 200 250 300 350 400
1000TIHE, S
PLa l BBOO
u 600 ~<:><:>~~ <:>~'>0: ~'>~:::> ~<:>.... 400 <:>a:0:...."- <:>'".... 200 ~~....
0Level 16BO nm
0 50 100 150 200 250 300 350
\000TI HE, 5
PLa l ABOO
u 600 <:><:>~ 1'JSmm., - - - - . --.... <:>~~0:
<:>~~:::>.... 400 1680 mma:~ - -- - - --
0: Heale<J Leng"".... 3900mm"-'" ~.... 200 - - - - - -- 2225 mm.... ~
0Lovel 2225 nm
0 50 100 150 200 250TIHE, S
Flooding rele 3.6 cm/sSystem pres9ure 4. I bar
SEFLEX to. t No. 03REBEKR rod bundloHellum-filled gaps
----- Experiment~ ~ ~ COBRR- TF
1'1'1"1'
Figure 63. Camparison of measured and calculated housing temperatures
(SEFLEX test No. 03).
-123-
Grid 9p8cer
ijOO200 300TIME. S
100
0
~0 ")
<!> /0
./0
./0
II
I"
50
100
~oooo
3000
2500
3500
150
'"'"...;
200w>w-'
-''"><'"
Floodlng rete 3.B cm/aSystem pressure 4. I bar
SEFLEX test No. 03AEBEKR rod bundleHelium-fl lled gaps
----- ExperIment<!> <!> <!> COBRA- TF
Figure 64. Comparison of measured and calculated quench front progression
(SEFLEX test No. 03).
500
3000
1000
100 200 300 ~OO 500TIME, S
~ooo'+----~---~--~-~-~o
-124-
0 Plot B
500
1000
1500
"'",-' 2000w>..,..J
-' 2500a:xa:
3000
3500
100 200 300 ~OO 500TI ME, S
Plot A
)
~ /~
~/
~~~/
~7 Grld ,pacer
-I/
11~ooo
o
3500
o
'"a:
1500
'"'",üJ 2000>..,..J
Ci! 2500
Floodlng rale 3.8 ernteSystem pressure 4.1 bar
FEBA lesl No. OSFEOR rod bundLeG8pless rods
----- Experiment~ ~ ~ COBRA- TF
floodlng fate 3.8 cm/sSystem pressure 4.0 bar
SEFLEX lesl No. 03AEBEKA rod bundleHelium-filled gaps
--- Experiment~ ~ ~ COBRA- TF
o
500
1000
1500
"",üJ 2000>..,-'
Ci! 2500
"a:3000
3500
Plot ()
)Z
k~~
;::~~
fl~
o
500
1000
1500
"",üJ 2000>w-'
Ci! 2500
3000
3500
) Plot 0
7/ ~-l ~~~~
~~
j~
-l~ooo
o 100 200 300 ~OO 500TIME, S
~ooo
o 100 200 300 ~oo 500TIME, S
Floodlng fale 3.8 em/sSystem pres9ure 2.1 bar
SEFLEX lesl No. OSAEBEKA rod bundleHellum-filled gaps
-- Experiment~ ~ ~ COBRA- TF
Floodlng raLe 3.8 ernteSystem pressure 2. t ber
SEFLEX lesl No. 07AEBEKA rod bundleArgon-lllled gaps
-- Experiment~ ~ ~ COBRA- TF
Figure 65. Compari50n of measured and calculated quench front progression
(FEBA test No. 223, SEFLEX tests No. 03, 05, and 07).