The journal of Nuclear Science and Technology - Volume 10, Number 2, June 2020
Contents
Analysis of in-vessel accident progression in VVER1000 NPP during SBO accident
with external reactor vessel cooling method
Doan Manh Long, Nguyen Van Thai, Tran Chi Thanh…………………………………..
01
Numerical Investigation of Fluid – Structure Interaction (FSI) on Sodium Leakage
Accident at Prototype Fast Breeder Reactor Monju
Hoang Tan Hung, Bui Thi Hoa, Hoang Minh Giang…………………………………….. 15
Study on the preparation of selenium nanoparticles by gamma Co-60 method and
investigate the stability
Ngoc Duy Nguyen, Van Phu Dang, Anh Quoc Le, T. Kim Lan Nguyen,Quoc Hien
Nguyen, T. Thu Ngan Tran…………………………………………………………………...
26
Assessment of nitrogen nutrient sources in aquatic environment of Tuyen Lam sub-
catchment based on it’s stable isotopes ratio (δ15N-NO3) combined with
geochemical parameters
Nguyen Thi Huong Lan, Nguyen Minh Dao, Phan Quang Trung, Vo Thi Mong
32
44
Tham, Le Xuan Thang, Phan Son Hai, Nguyen Van Phuc………………………………
Radioactivity of some natural and artificial radionuclides in coastal seawater at
Ninh Thuan, Ba Ria - Vung Tau and Ca Mau provinces in 2018
Nguyen Dinh Tung, Le Nhu Sieu, Nguyen Van Phuc, Nguyen Van Phu, Tran Dinh
Khoa, Vuong Thi Thu Hang, Nguyen Thi Thanh Nga, Le Thi Minh Tuyen……………
Coupled reaction channels study of the 16O(d,6Li) reaction
Do Cong Cuong, Nguyen Hoang Phuc, Nguyen Tri Toan Phuc………………………... 52
Nuclear Science and Technology, Vol.10, No. 2 (2020), pp. 01-14
Analysis of in-vessel accident progression in VVER1000 NPP
during SBO accident with external reactor vessel cooling method
Doan Manh Long1, Nguyen Van Thai2, Tran Chi Thanh3
1) Nuclear Training Center, 140 Nguyen Tuan, Thanh Xuan, Hanoi
2) Department of Nuclear Engineering and Environmental Physics, School of Engineering Physics,
Hanoi University of Science and Technology, 01 Dai Co Viet, Hai Ba Trung, Hanoi
3) Viet Nam Atomic Energy Institute, 59 Ly Thuong Kiet, Hoan Kiem, Hanoi
Email(1): doanmanhlong@vinatom.gov.vn
(Received 20 February 2020, accepted 08 May 2020)
Abstract: In this study, the MELCOR v1.8.6 code was utilized to perform an analysis of the in-vessel
accident progression in VVER1000 reactor during the Station Black-Out (SBO) accident with and
without external reactor vessel cooling (ERVC) strategy. The analysis presented the predictions of the
main phenomena during the accident such as failure of fuel cladding, collapse of lower core support
plate, relocation of core debris to lower plenum and mass of debris components in lower plenum, and
provided comparisons between two cases in term of main parameters such as integrity time of reactor
and structure components of molten pool. These parameters are very important inputs for further
research on the application of external vessel cooling strategy for VVER1000 reactor.
Keywords: External Reactor Vessel Cooling, SBO, VVER1000, In-Vessel Melt Retention, Severe Accident.
such as RELAP5/SCDAP and MELCOR code,
in order to evaluate the thermal response of
lower head wall under heat load from molten
pool. The RELAP5/SCDAP was used to
provide bounding conditions including mass
of molten debris and its components,
temperature, and decay heat deposited in
molten debris in an extensive series of severe
accident calculations for AP1000 [4] and
APR1400 [5]. And the MELCOR code was
applied to analyze the in-vessel accident and
the capability of IVR strategy for a large scale
pressurized water reactor [12].
I. INTRODUCTION
The external reactor vessel cooling
strategy through submerging reactor vessel into
water has been well known as a novel severe
accident management strategy with the aim at
preventing lower head vessel from failure in
case of core melt accident, called the In-Vessel
melt Retention (IVR) strategy. This strategy
has been successfully adopted as severe
accident management strategy for low power
reactors such as VVER440 [1, 2] and AP600
[3]. Subsequently, the strategy has become a
crucial research issue in severe accident
management strategy for high power reactors
as AP1000 [4], APR1400 [5] and HPR1000 [6].
Recently, the preliminarily studies on
the feasibility of the IVR strategy for
VVER1000 reactor have been analyzed
through benchmark calculations under a very
conservative scenario of Large Break Loss of
Coolant Accident (LBLOCA) by using
different severe accident codes as MELCOR,
ASTEC, PROCOR and MAAP [13]. The
Right from the first, the severe accident
codes have greatly contributed to analysis of
IVR strategy. Since the analytical methods [7,
8] and numerical methods [9, 10, 11] require
initial bounding conditions which mostly
provided by severe accident analysis codes
ANALYSIS OF IN-
results of the benchmark calculations have
confirmed that the IVR strategy could prevent
VVER1000 lower head vessel from failure and
provided bounding conditions for stand-alone
calculations by using computational codes.
Although, the benchmark calculations also
showed that the large discrepancies were found
in the predicted mass components of debris bed
and heat flux distribution on the external
surface of lower head wall in different codes,
the results provided a bunch of necessary data
in order to specify a range of bounding
conditions for IVR analysis in case of
LBLOCA scenario for VVER1000 reactor.
The study provided predictions on the key
parameters for further studies on IVR
application for VVER1000 reactor, such as:
the timing of main events as collapse of
fuel cladding and lower core support plate;
relocation of core debris and mass of debris
components into lower plenum; and the
decay heat deposited in debris bed as well.
II. MELCOR INPUT
FOR VVER1000 REACTOR
The MELCOR v1.8.6 code [15] is a
recent version of MELCOR code with taking
full of advanced features which have been
capability of simulating a broad spectrum of
severe accident phenomena in light water
nuclear power plants such as thermal hydraulic
response of the reactor coolant system; core
heat-up, degradation and relocation; hydrogen
production, transport, and combustion; fission
product release and transport behavior. The
MELCOR v1.8.6 code has an improvement in
modeling formation of molten pool in lower
plenum and thermal-interaction of lower head
wall and molten pool. These features enable the
version to apply for studying the IVR strategy.
The IVR strategy adopted for
VVER1000 reactor was also analyzed through
the Station Black-Out (SBO) accident by using
ASTEC code and ANSYS Fluent [14] which
showed the external vessel cooling was
capability for the stabilization of the vessel and
the retention of the molten corium inside
VVER1000 vessel. However, the study did not
provide in details the mass of debris
components which relocated to lower plenum
and presented decay heat power in debris bed
clearly, which would not benefit for further
studying on heat load imposing to lower head
vessel by using numerical and analytical
methods. In addition, the results obtained from
one code are not enough for studying IVR
application for VVER1000 reactor during SBO
accident, there need additional number of
studies by other codes in order to contribute
additional data for specifying the range of
bounding conditions.
The VVER1000 reactor is a Russian
pressurized water reactor which produces
3000MW thermal power and 1000MW
electrical power. It comprises of four cooling
circulation loops. Each loop has a main coolant
pump, hot leg, cold leg and horizontal steam
generator (SG). The pressurizer (PRZ) is
connected to hot leg of the fourth loop and the
spray lines of pressurizer are connected to cold
leg of the first loop, and safety valves and
relief valves are mounted on the top of PRZ.
The reactor core consists of 163 hexahedral
fuel assemblies and 61 control rod clusters.
Each assembly comprises 312 fuel rods and 18
control rod tubes.
In the present study, an analysis of
in-vessel accident progression during SBO
accident combined with externally vessel
cooling for VVER1000 reactor was
performed by using MELCOR v1.8.6 code.
2
DOAN MANH LONG et al.
Fig. 1. A MELCOR nodalization scheme for thermal hydraulics volumes
The VVER1000 reactor was input into
relief tank were also modeled. Secondary
part of steam generator was modeled as a
control volume. The passive core cooling
system, comprising of four Hydro-
ACCumulators (HACCs), was also included.
MELCOR code through a nodalization step.
Fig.1 displays a nodes scheme for thermal
hydraulics volumes of reactor vessel and a
circulation loop including the pressurizer
systems. The internal volume of reactor
vessel consists of lower plenum, reactor
core, downcomer, upper plenum and upper
head which were assigned numerical
identifiers as CV100, CV101, CV102,
CV103 and CV104 respectively. In an
individual circulation loop, the hot leg was
divided into two control volumes (CV110
and CV111) and cold leg was divided into
three control volumes (CV130, CV131 and
CV132); heat exchanger tubes inside steam
generators were divided into five groups
which were split into hot part (HP) with red
boundary and cold part (CP) with blue
boundary; hot collector and cold collector
were modeled as two control volumes;
pressurizer systems including a surge line,
pressurizer, safety and relief valves; and
Fig. 2. A nodalization scheme for structures in
reactor core, lower plenum and downcomer
3
ANALYSIS OF IN-
Fig. 3. A node model for lower head wall
Fig. 2 demonstrates the nodalization
scheme of structures in reactor core, lower
plenum and downcomer which were divided
into six radically concentric rings and twelve
vertically levels. The lower head wall was
divided into nine segments and each segment
was sliced into six thickness-equaling layers, is
presented in Fig. 3.
(2)
Where is the effective stress (Pa) and
calculated as:
(3)
W
across the lower head;
and
are the
II. FAILURE MODEL OF LOWER HEAD
density and the depth of the debris in lower
plenum; g is gravitational acceleration; and
are the inner vessel radius and outer radius
of load-bearing vessel.
In this calculation, the Larson-Miller
creep rupture model in MELCOR code [15]
was used to evaluate failure of lower head
vessel. The model evaluates cumulative
damage in lower head wall following time in
response to mechanical loading at elevated
temperatures based on the Larson-Miller
parameter and a life-fraction rule.
The life-fraction rule gives the
cumulative damage, expressed as plastic strain,
, as:
(4)
The Larson-Miller creep-rupture failure
model gives the time to rupture, , in seconds, as:
Due to the division of , the plastic
strain expressed in Eq.4 is a dimensionless
value. And following the rule of the model,
the failure of lower head vessel was declared
(1)
Where T is the temperature of segment,
and
is the Larson-Miller parameter that is
when the value of
0.18 [15].
reaches to the value of
given as:
4
DOAN MANH LONG et al.
generator and batteries except batteries for
III. INITIAL CONDITIONS
AND SCENARIOS
BRU-A valves of SGs. The total SBO caused
failure of all active safety systems including
Emergency Feed Water systems. The
additional assumptions of the scenario were
made as follows:
Before initiation of the accident, the
VVER1000 reactor was assumed to be
operating at steady state condition with full
thermal power. The steady state was simulated
in MELCOR code and the key parameters are
compared to design values in Table I. The
steady state parameters obtained from
MELCOR simulation were mostly in
oscillating limitation of design values, except
the value of maximum coolant temperature at
reactor inlet was little higher. Although, there
were not oscillating limitations for mass flow
rate through reactor core and steam mass flow
rate at steam generator outlet, the values
obtained from MELCOR were close to these
design values. Generally, the steady state
conditions obtained from MELCOR are in
good agreement with the design values.
Pressure of PRZ was controlled by
relief valves and safety valves with
characteristics in Table II;
There was not taken into account main
coolant pumps seal leakages;
Pressure of SGs was controlled by
BRU-A valves to maintain pressure in SGs
below 6.7MPa;
Four Hydro-ACCumulators (ACCs)
were available;
The depressurization for reactor vessel
by opening safety valves on top of PRZ and
cavity flooding strategy were triggered when
temperature of steam in reactor core exceeded
650oC based on Accident Management
Measures (AMM) [14].
The total Station Black-Out accident
was initiated at zero second due to loss of
offsite and onsite power including diesel
Table I. Comparison of parameters at steady state condition
Parameters
MELCOR value
Design value
3000+210
15.7±0.3
559.15±2.0
593.15±3.5
17 611
Core power (MW)
3100
15.6
567
Primary pressure (MPa)
Maximum coolant temperature at reactor inlet (K)
Average coolant temperature at reactor outlet (K)
Mass flow rate through reactor core (kg/s)
Pressure in steam generator (MPa)
Steam mass flow rate at steam generator outlet (kg/s)
596
17650
6.29
420
6.28±0.2
437
Table II. The characteristics of the pressurizer relief and safety valves
Name
Characteristics
Design value
16.00
MELCOR value
16.00
Relief valves
Opening pressure (MPa)
Closing pressure (MPa)
Opening pressure (MPa)
Closing pressure (MPa)
Opening pressure (MPa)
Closing pressure (MPa)
15.70
15.70
Safety valves
Stage 1
18.11
18.11
16.67
16.67
Safety valves Stage
2, 3
18.60
18.60
17.07
17.07
5
ANALYSIS OF IN-
pressure was maintained at safety threshold
below 20 MPa (Fig. 4).
III. RESULTS AND DISCUSSION
A. The base case without ERVC
Since the decay heat kept on generating,
together with leakage through the opening of
relief valves, the water volume inside reactor
vessel constantly decreased. Fig. 6 displaying
volume of water in reactor core and lower
plenum points out the uncovery of reactor core
started at 2300 seconds and the reactor core
was totally uncovered at 4100 seconds, the
lower plenum even ran out of water before
reactor core at 3800 seconds.
The SBO accident was initiated at 0
second when the VVER1000 reactor was
operating at full power capacity. Because of
total loss of onsite and offsite power systems,
the active safety systems were entirely
neutralized. The passive safety system, which
consists of four hydro-accumulators (HACCs),
was available and only initiated when pressure
in primary loop decreased to below 5.8 MPa.
The main events of the accident are presented
in Table III.
The pressure of primary loop sharply
dropped because of a sudden reduction of
thermal power to decay heat power (Fig. 4). At
early phase, the decay heat was removed via
steam generators through dump valves.
Afterward, the steam generators nearly dried
out at 1000 seconds (Fig. 5), the pressure of
primary loop drastically increased (Fig. 4). The
primary pressure was controlled by relief and
safety valves with the opening and closing set
point listed in Table II. Therefore, the primary
Fig. 4. Pressure of primary loop
Table III. Main events of in-vessel accident progression
Main events
Time
(second)
Initiation of accident
0.0
Reactor tripped
1.6
Begin of core uncovery
Total core uncovery
2300
4100
9727
9910
9990
10000
10300
10846
Initiation of primary depressurization
Begin of water injection from HACCs
Stop of HACCs injection
Start of oxidation
Start of fuel cladding failure
Failure of lower core plate
Failure of lower head vessel (CREEP-RUPTURE)
21938s
6
DOAN MANH LONG et al.
The temperature of vapor in reactor core
oxidations. Only after the depressurization and
the water injection from HACCs brought a
beneficial condition for structural oxidations,
and the oxidations were predicted to happen at
10000 seconds. Fig. 8 presents mass of
hydrogen which was generated from oxidations
of steel and zircaloy, and total mass of
generated hydrogen was 372 kg.
over 650oC was predicted at 9727 seconds
which triggered the opening of relief and safety
valves on the top of PRZ to depressurize
primary loop. The depressurization created the
sharp drop of pressure in the primary loop at
9727 seconds (Fig. 4), and the pressure
dropped below 5.8 MPa at 9910 seconds which
initiated the water injection from four hydro-
accumulators. The water supply from HACCs
was stopped when their water volumes reduced
below 6 m3 in order to prevent noble gases
from being transported into reactor vessel. The
result showed the water volume in HACCs
dropped below 6 m3 at 9990 seconds (Fig. 7),
hence the water supply from HACCs only
lasted 80 seconds. The additional water from
four HACCs only helped the reactor core
recovering in a very short time (Fig. 6).
Afterward, the volume of water in reactor
vessel quickly decreased and the reactor core
was uncovery again at 12300 seconds (Fig. 6).
And due to the evaporation of additional water,
the pressure of primary loop slightly increased
(Fig. 4).
The occurrence of oxidations boosted
the reactor core heating. The temperature of
fuel claddings in ring 1 are displayed in Fig. 9
which shows the temperature of fuel claddings
from level 7 to 11 rapidly increased right after
the happening of oxidations, even at this time
the reactor core was submerged by additional
water. The failure of fuel claddings occurred
at 10300 seconds.
Fig. 6. Volume of water in reactor core
and lower plenum
The failure of fuel claddings marked the
collapse of reactor core and created debris
containing heat source. The core debris
relocated to the bottom of the reactor core area
and imposed thermal load on lower core
support plate. The simulation showed the
failure of lower core support plate occurred at
10846 seconds, only 846 seconds (14 minutes)
and 546 seconds (9 minutes) after the
depressurization and failure of fuel cladding
Fig. 5. Water volume of SGs
Before the initiation of depressurization,
the high pressure condition in primary loop did
not provide good condition for structural
7
ANALYSIS OF IN-
respectively. The failure of lower core plate
triggered massive relocation of core debris to
lower plenum. Fig. 10 presents evolution of
mass of core debris in lower plenum.
At 10864 seconds, the time of core
debris relocation to lower plenum, water
volume in lower plenum was about 4 m3
(Fig. 6). Under the massive amount of hot
debris, the remaining water in lower plenum
was quickly boiled off, and the area ran out
of water again at 11000 seconds.
Fig. 7. Volume water of HACCs
Fig. 10. Mass of debris components
in lower plenum
Fig. 8. Mass of generated hydrogen
Fig. 11. The average temperature of 9 segments
Because there were none of any cooling
measures from internal and external reactor
vessel, the decay heat generated in debris
directly imposed to lower head vessel. Fig. 11
demonstrating the average temperatures of 9
segments shows the average temperatures of
the segments slightly decreased after the
depressurization, they, however, started to
Fig. 9. Temperature of fuel claddings in ring 1
8
DOAN MANH LONG et al.
increase right after appearance of core debris in
lower plenum. Generally, the first five
segments were imposed thermal load more
than others which led to higher evolution of
their temperatures compared to others. Since
20000 seconds, temperatures of the first five
segments were over 1000oK which initiated the
thermal strain of the lower head vessel at the
segments. The evolutions of thermal strain
fraction at 9 segments are displayed in Fig. 12.
Among the first five segments, the 3rd segment
seemed to be imposed the highest heat hence
its temperature was the strongest increase. Due
to experience of the highest thermal load, the
evolution of elastic strain of the 3rd segment
was the sharpest increase and reached to the
value of 0.18 at 21938 seconds (6 hours) since
initiation of the accident, marked the failure of
lower head vessel.
Fig. 10 indicates the relocation of core
fuels started from 10864 seconds and stopped
at 15000 seconds, and together with other
components such as ZrO2, Zr and oxides of
steel was stable after 15000 seconds.
Meanwhile, the evolution of steel debris kept
on going due to the collapse of supporting
structure in lower plenum and stabilized after
20000 seconds. The results showed all core
fuels relocated to lower plenum and according
to time relocation of core fuels the decay heat
power deposited in debris bed with elimination
of decay power of volatile gases varied from
32 MW to 29 MW. At the time of lower head
failure, also end of simulation time, total mass
of debris components, predicted by MELCOR
v1.8.6, are listed in Table IV.
A code-to-code comparison for the main
events of in-vessel accident progress between
MELCOR code in this work and ASTEC in [14]
was presented in Table V. The comparison
shows the considerable differences of results
obtained in both codes. The prediction of
ASTEC on relocation of core melt to lower
plenum was at 9705 seconds which is 1141
seconds earlier than MELCOR prediction
which was at 10846 seconds. As a result, the
failure of lower head vessel predicted by
ASTEC was at 20886 seconds which is 1052
seconds earlier that MELCOR prediction
which was at 21938 seconds. However,
MELCOR results predicted a larger mass of
corium in lower plenum than ASTEC results
which were 131 tons and 73.3 tons respectively.
Fig. 12. Elastic strain of 9 segments
Table IV. Mass of debris in lower plenum
The differences in the time of main
events were caused by considerable differences
in calculation models in two codes. Besides the
reactor trip assumption of ASTEC also
contributed to differences in predictions when
it was assumed to happen at the same time with
initiation of SBO accident which is unrealistic.
Mass (kg)
80600
18150
8800
Debris
UO2
Zr
ZrO2
Steel
31000
2750
Oxide of steel
9
ANALYSIS OF IN-
In addition, the difference in model of core
debris relocation to lower plenum caused a
significant difference in prediction of mass of
UO2 relocating to lower plenum. In MELCOR
code, the relocation of core debris was
triggered when the lower core support plate
was failed by both mechanical and thermal
load, and when the lower core support plate in
a ring failed and lost capability of supporting,
its entire structure and all material resting on it
would totally collapse and relocate to lower
plenum. Meanwhile, ASTEC code [16] did not
predict the mechanical failure of the lower core
support plate but only its thermal failure, and
the core debris only relocated to lower plenum
by melting the plate and crossing through the
plate hole to lower plenum. In this case,
results of MELCOR simulation showed the
lower core support plate in five rings
completely collapsed, therefore all core fuel
relocated to lower plenum. In case of ASTEC,
the mass of core debris relocating to lower
plenum depended on the size of the hole in the
plate hence somehow the size was not large
enough for all core fuel relocating to lower
plenum in this scenario.
Table V. MELCOR and ASTEC main events
Time (second)
MELCOR ASTEC
Main events
Initiation of accident
0.0
1.6
0.0
0.0
Reactor tripped
Opening of steam-dump to atmosphere valves (BRU- A)
Begin of core uncovery
8.0
45.0
-
2300
4000
4100
8000
8000
9727
9910
9990
10846
21938
131
Heating up of the core
8767
-
Total core uncovery
Steam dry-out
3800.0
7200
10341
10514
17545
9705
20886
73.33
Closing of BRU-A valves
Initiation of primary depressurization
Begin of water injection from HACCs
Stop of HACCs injection
First material slump in lower plenum
Failure of lower head vessel (CREEP-RUPTURE)
Total mass of corium in lower head (tons)
of cold leg. Until now, there has not been
official design of ERVC for VVER1000
reactor. Therefore, the study proposed a
simplified scheme of ERVC strategy for
VVER1000 reactor which included a cooling
channel formed by reactor vessel and a frame
structure, and an unlimited water resource.
Fig. 13 demonstrated a nodalizational scheme
of the ERVC in MELCOR code.
B. Deployment of IVR strategy with ERVC
In this study, the external vessel cooling
strategy was adopted in order to cool down
the hot debris from external surface of the
VVER1000 lower head vessel by injecting
water into the cavity. The water injection was
initiated when the vapor temperature in
reactor core exceeded 650oC. The water level
in reactor cavity was maintained at the height
10
DOAN MANH LONG et al.
The SBO accident was re-simulated with
deployment of ERVC strategy. The results
showed the performance of ERVC strategy did
not affect the in-core accident progress. The
evolution of the accident remained the same as
the scenario without implementing ERVC
strategy. A comparison of main events between
two SBO scenarios with and without deploying
IVR strategy is presented in Table VI.
The initiation of ERVC strategy was
simultaneously happened at 9727 seconds with
the primary depressurization when vapor
temperature in reactor core exceeded 650oC.
Fig. 14 demonstrated water level in reactor
cavity and the water level reached the height of
cold leg at 12500 seconds, then it was
maintained at the level. Figs. 15-17 displayed
the temperature of 9 segments of VVER1000
lower head wall in two both scenarios. The
solid lines and the dash lines present the results
of the scenario without ERVC strategy and
with ERVC strategy respectively. The figures
indicate the decrease of temperature of lower
head walls in both cases happened right after
the initiation of depressurization at 9727
seconds. However, in case of implementing
ERVC (dash line) the temperature of lower
head at 9 segments decreased deeper.
Fig. 13. A nodalization for external reactor vessel
cooling strategy
Fig. 14. Water level in cooling channel and cavity
Table VI. Main events of in-vessel accident progression
Time (seconds)
Main events
With IVR
Without IVR
0.0
Initiation of accident
0.0
Reactor tripped
1.6
1.6
Begin of core uncovery
Total core uncovery
2300
4100
9727
9727
9910
9990
10000
10300
10846
27341
2300
4100
9727
9727
9910
9990
10000
10300
10846
21938
Initiation of primary depressurization
Start of cavity flooding
Begin of water injection from HACCs
Stop of HACCs injection
Start of oxidation
Start of fuel cladding failure
Failure of lower core plate
Failure of lower head vessel (CREEP-RUPTURE)
11
ANALYSIS OF IN-
However, Fig. 15 indicates the
temperature of segments 1-3 increased beyond
1000oK, among them the temperature of
segments 1 and 3 were beyond 1000oK at
24000 and 26000 seconds respectively which
caused the occurrence of thermal strain of the
segments. Fig. 18 displayed the evolution of
thermal strain fraction of all segments. The
results show the thermal strain of lower head at
segment 1 reached the value of 0.18 at 27341
seconds which marked the failure of
VVER1000 lower head reactor vessel.
Fig. 15. Temperature of segments 1-3
Fig. 15 and Fig. 16 indicate the
temperature at segments 1-6 started to increase
at 15000 seconds, and it strongly increased at
17000 seconds. Meanwhile, Fig. 17 shows the
temperature of lower head wall at segments 7-9
was efficiently cooled and kept at low
temperature below 600oC. Together with the
increase of temperature of segments 1-6, the
heat transfer between lower head vessel and
water drastically increased as well, which
caused the vibration of water level in cooling
channel (Fig. 14). Fig. 16 shows the temperature
of segments 4-6 was kept under 1000oK (727oC).
Fig. 17. Evolution of temperature of segments 7-9
Fig. 16. Evolution of temperature of segments 4-6
Fig. 18. Evolution of thermal strain of segments
12
DOAN MANH LONG et al.
occurrence of failure at low position when the
IV. CONCLUDING REMARKS
critical heat fluxes at the low positions were
low as seen in results of ULPU [7]. The
results of the study also showed the failure of
VVER1000 lower head vessel happened at
low positions, therefore, it is necessary to
optimize geometry of cooling channel for
VVER1000 reactor;
In this paper, the in-vessel accident
progress during the Station Black-Out (SBO)
accident combined with externally vessel
cooling was analyzed for VVER1000 reactor
by using MELCOR v.1.8.6 code. Some
conclusions were given as following:
Under SBO accident, only water
from four hydro-accumulators could not
prevent the collapse of reactor core and
relocation of core debris to lower plenum. The
collapse of reactor core marked by failure of
fuel cladding was occurred at 10300 seconds
(2.86 hours), the appearance of first core debris
in lower plenum was at 10846 seconds (3.01
hours), and all core fuels relocated to lower
plenum at 15000 seconds (4.2 hours). The
decay heat power deposited in debris bed
locating in lower plenum was estimated from
32 to 29 MW with taking the elimination of
decay power of volatile gases into
consideration. All the outcomes would be
significant information for further study on
thermal behavior of debris bed/molten pool in
lower plenum and thermal response of
VVER1000 lower head vessel as well;
In addition, the comparison on the
main phenomena during accident progress
between MELCOR and ASTEC suggested it is
necessary to take further investigation on in-
vessel accident progression in case of SBO
accident by other severe accident codes in
order to set up a range of key parameters for
the bounding configuration of molten pool in
VVER1000 lower plenum and specify decay
heat power deposited molten pool as well.
ABBREVIATIONS
AMM
Accident Management Measures
BRU-A
A kind of atmospheric dump
valves of steam generators
CC
Cold Collector
Cold Leg
CL
CP
Cold Part
In
the
base
case
without
CV
Control Volume
External Reactor Vessel Cooling
Hydro-ACCumulators
Hot Collector
implementing IVR strategy, the VVER1000
lower head vessel was failed at 21938 seconds
(6.09 hours) due to thermal creep-rupture.
Meanwhile, in the case of IVR strategy
deployment, the results indicated the strategy
could not prevent VVER1000 lower head
vessel from failure caused by thermal creep-
rupture and only prolonged the existence of
VVER1000 lower head vessel for 5437
seconds (1.5 hours) compared to the base case;
ERVC
HACCs
HC
HL
Hot Leg
HP
Hot Part
IVR
In-Vessel melt Retention
LBLOCA Large Break Lost of Coolant
Accident
LP
Lower Plenum
MSH
NPP
Main Steam Header
Nuclear Power Plant
The flat shape of VVER1000 lower
head vessel also raises a concern about the
13
ANALYSIS OF IN-
[8]. Esmaili, H., et al. An Assessment of Ex-
Vessel Steam Explosions in the AP600
Advanced Pressurized Water Reactor , Energy
Research, Inc., ERI/NRC 95-211, 1996.
PORV
PRZ
SBO
SG
Pilot-Operated Relief Valve
PRessuriZer
Station Black-Out
Steam Generator
Steam Line
[9]. Bui and Dinh. Modeling of heat transfer in
heated-generating liquid pools by an effective
SL
diffusivity-convectivity
approach ,
In:
UP
Upper Plenum
Proceedings of 2nd European Thermal-Sciences
ACKNOWLEDGEMENT
Conference, Rome, Italy, pp. 1365 1372, 1992.
[10].Tran, C.T., and Dinh, T.N. The effective
convectivity model for simulations of melt
pool heat transfer in a light water reactor
pressure vessel lower head . Part I&II.
Progress in Nuclear Energy 51, 849 871, 2009.
The authors are grateful for financial
funding of Ministry of Science and Technology
and Vietnam Atomic Energy Institute. This
work was performed under ministerial project
on
[11].Dombrovskii, L.A., et al.
Numerical
05/01/2018.
Simulation of the Stratified-Corium
REFERENCES
Temperature Field and Melting of the Reactor
Vessel for a Severe Accident in a Nuclear
Power Station . Thermal Engineering, Vol. 45,
No. 9, pp. 775-765, 1998.
[1]. Tuomisto, H., and Theofanous, T.G.
A
consistent approach to severe accident
management . Nuclear Engineering and
Design 148, 171-183, 1994.
[12]. Yue Jin et al. In-and ex-vessel coupled analysis of
IVR-ERVC phenomenon for large scale PWR .
Annals of Nuclear Energy 80 322-337, 2015.
[2]. Kym l inen, O., et al. In-vessel retention of
corium at the Loviisa plant . Nuclear
Engineering and Design, 169, 109-130, 1997.
[13].SANGIORGI Marco et al. In-Vessel Melt
Retention Analysis of a VVER1000 NPP .
EUR 27951; doi 10.2790/62596, 2016.
[3]. Theofanous, T.G., et al. In-vessel cool-ability
and retention of a core melt , DOE/ID-10460,
Volume 1, Revised October, 1996.
[14].Polina Tusheva et al. Analysis of severe
accidents in VVER1000 reactors using integral
code ASTEC . Proceeding of the 17th
[4]. Esmaili, H., and Khatib-Rahbar, M. Analysis of
in-vessel retention and ex-vessel fuel coolant
interaction for AP1000 . Energy Research, Inc.,
ERI/NRC 04-21, NUREG/CR-6849, 2004.
International
Conference
on
Nuclear
Engineering, Brussels, Belgium, 2009.
[5]. Rempe, J.L., et al. In Vessel Retention
strategy for higher power reactors . Final
Report, INEEL/EXT-04-02561, 2005.
[15].Sandia National Laboratories, MELCOR
computer code manuals , Ver.1.8.6, Rev.3 vol.
NUREG/CG 6119, 2005.
[6]. Ji Xing et al.
HPR1000: Advanced
[16].Belon, S., et al. Insight of core degradation
simulation in integral codes throughout
ASTEC/MELCOR crosswalk comparisons and
ASTEC sensitivities studies . The 8th
European Review Meeting on Severe Accident
Research, Warsaw, Poland, 2017.
pressurized water reactor with Active and
Passive safety Engineering 2 79-87, 2016.
[7]. Theofanous, T.G., et al. In-vessel cool-ability
and retention of a core melt , DOE/ID-10460,
Volume 1, Revised October, 1996.
14
Nuclear Science and Technology, Vol.10, No. 2 (2020), pp. 15-25
Numerical Investigation of Fluid – Structure Interaction (FSI) on
Sodium Leakage Accident at Prototype Fast Breeder Reactor Monju
Hoang Tan Hung, Bui Thi Hoa, Hoang Minh Giang
Institute for Nuclear Science and Technology, 179 Hoang Quoc Viet, Nghia Do, Cau Giay, Ha Noi
(Received 28 February 2020, accepted 21 April 2020)
Abstract: This study presents a numerical investigation of fluid-structure interaction about sodium
leakage accident at prototype fast breeder reactor Monju. The dynamic forces of sodium liquid flow
acting on the thermowell are the main cause of fatigue failure, which leads to the occurrence of
thermowell cracking due to intense stresses and high fatigue-cycle. Since the location and magnitude
of these stresses are unknown, an analysis of the vibrational characteristics and stresses caused by FSI
was performed by using ANSYS to prevent similar accidents in the future. The most dangerous case
was found for FSI analysis by comparisons between the natural frequency of thermowell and vortex
shedding frequency at several operating conditions. The results showed that a stress concentration
location is similar to the location of broken thermowell, and the amplitude of stress is large enough for
fatigue damage.
Keywords: Monju, Sodium leakage accident, FSI, ANSYS.
was to study the interaction between water and
circular cylinder structure under the
operational conditions of the Monju reactor.
I. INTRODUCTION
Monju is a fast breeder reactor (FBR)
that has been developed in Japan for a long –
term research, development, and utilization
program in nuclear energy. On 8th December
1995, the reactor suffered a serious accident
with sodium leaked through a temperature
sensor due to a break of thermowell inside a
secondary pipe resulting from intense vibration.
A study on sodium leakage at Monju was
performed in [1] to investigate the causes and
consequences of the accident. From the results,
it was concluded that the breakage of the
thermowell was caused by high cycle fatigue
of the well tube tip due to Vortex-Induced
Vibration (VIV) phenomenon.
On the other hand, to prevent the sodium
leakage accident, the improvement of
thermowell design was obtained by means of
an analytical study using the finite element
method (FEM) [3]. Various designs of
thermowell were taken into account. The
results showed that the design of thermowell
with a continuous low-gradient slope and no
neck could avoid the effects of VIV.
It can be seen that, after the sodium
leakage accident at the Monju reactor, many
studies were performed to find the cause of the
accident and also to prevent the accident from
happening again. This prevention could be
done by the thermowell design and its working
condition improvement. However, the number
of studies on thermowell stress concentration
prediction under the operational condition still
limited. Therefore, it is necessary to perform a
In the study of Ref. [2], it was found that
vortex-induced vibration depends on the
specific range of reduced velocity. Therefore, a
range of the reduced velocity was created using
the experimental facility to avoid that
phenomenon. The purpose of this experiment
©2020 Vietnam Atomic Energy Society and Vietnam Atomic Energy Institute
NUMERICAL INVESTIGATION OF FLUID – STRUCTURE INTERACTION (FSI)…
study on the prediction of the stress
concentration of thermowell. On the other hand,
the Monju accident happened due to the VIV
phenomenon as a consequence of FSI. One
study on the consequence of FSI using ANSYS
has done [4] to predict thermowell stress
concentration on the steam temperature sensor.
Base on this paper, our work focus on ANSYS
simulation to predict the stress concentration of
Monju’s thermowell.
generating capacity of 280 MW. Liquid
sodium is used in the primary and secondary
cooling systems. An accident occurred due to
the failure of a thermowell located in the
secondary cooling system in the course of
power
thermowell was broken apart by high cycle
fatigue failure. Figure shows the
buildup
test
programs.
The
1
configuration of thermowell before and after
the accident. The tip of the well-tube was
lost and the thermocouple was bent
approximately 45 degrees from the original
position. Thermowell is a type of step-shank
with its whole length of 185.5 mm and its
thickness of 3 mm as shown in Figure 1. It is
made of 304 stainless steel, which is 8000
kg/m3 in density, Young modulus, E, is in the
range of 193 – 200 GPa, and the Poisson
ratio equal to 0.29 [5].
In this study, following the study as
mentioned in [4], an analysis of FSI between
sodium liquid and Monju’s thermowell will be
performed by using ANSYS. Two approaches
of one-way FSI and two-way FSI are applied to
investigate the VIV phenomenon to find the
stress concentration location and the stress
amplitude acting on Monju’s thermowell.
II. EXPERIMENTAL FACILITY AND
CALCULATING METHODS
From the operation history of the
secondary loop of the Monju fast breeder
reactor [6] and thermodynamic and transport
properties of sodium liquid [7], the
characteristics of sodium liquid flow depend on
the history of the secondary heat transport
system operation is presented in Table I.
A. Overview of Monju facility
The Monju fast-breeder reactor is a
sodium-cooled reactor operating at a thermal
output of 714 MW and has an electrical
Fig. 1. Thermowell configuration before and after the accident [8]
Table I. Characteristics of sodium liquid flow in the secondary heat transport system
Temperature
(
Velocity
(m/s)
5.20
2.08
2.08
Dynamic viscosity
Density (kg/m3)
Reynold number
(10-4 Pa.s)
4.52
Case 1
Case 2
Case 3
Case 4
200
200
450
485
903.0352
903.0352
846.2525
838.1427
1.04x105
4.15x104
6.92x104
7.23x104
4.52
2.55
2.41
2.08
16
HOANG TAN HUNG et al.
phenomenon could lead to large amplitude
B. Calculation Method
The VIV phenomenon is the main cause
vibrations of structure, which in turn causes
fatigue damage. However, this situation occurs
in a range of frequencies, not a single value of
frequency as the normal resonance. The
following condition is given to avoid fatigue
damage [10]:
of damaging the thermowell of the Monju
reactor. Vortices were formed and detached
periodically from either side of cylindrical
bodies causes an unsteady oscillating flow. An
asymmetrical flow pattern develops around the
body because of the periodic shedding and
changes in the pressure distribution behind the
body. Vortex shedding occurs at a certain
frequency, which is called vortex shedding
frequency (fS). This frequency depends on flow
velocity U and the cylinder diameter D,
Strouhal number (St) and formulated as [9]:
fS < 0.4 fN
(2)
Because the Monju reactor was operated
under several conditions, it is necessary to find
the most dangerous case to perform FSI
analysis by using the criteria of equation (2).
Firstly, Modal analysis was applied to
determine the natural frequency (fN) and the
mode shape of the structure. Secondly, 2D
flow analysis in Fluent was conducted to
determine the vortex shedding frequency of
different flow conditions. A comparison
between the natural frequency of structure (fN)
and vortex shedding frequency (fS) could show
the most dangerous case to perform FSI
analysis. Finally, the FSI calculation was
performed to analyze the location and value of
stress concentration on the thermowell tube.
The value of stress concentration was
compared with the endurance limit of material
(SE). The diagram of the analysis process is
shown in Figure 2.
(1)
As a result of the periodic change of the
vortex shedding, the pressure distribution on
the cylinder outer surface may also change
periodically, thereby generating a periodic
force on the cylinder surface. Each cylindrical
body has its natural frequencies (fN). If the
vortex shedding frequency (fS) is close to one
of the natural frequencies (fN), resonance could
have happened. Then, as a result, the
cylindrical bodies start to vibrate with the same
frequency as the vortex shedding frequency,
called the lock-in phenomenon. This lock-in
Fig. 2. Analysis process
17
Tải về để xem bản đầy đủ
64 trang
14280
Free
Bạn đang xem 20 trang mẫu của tài liệu "The journal of Nuclear Science and Technology - Volume 10, Number 2, June 2020", để tải tài liệu gốc về máy hãy click vào nút Download ở trên
File đính kèm:
- the_journal_of_nuclear_science_and_technology_volume_10_numb.pdf
