Nuclear Engineering and Design xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Nuclear Engineering and Design

journal homepage: [Link]/locate/nucengdes

Validation of computer code ‘ATMIKA’ against RD-14M Small Break LOCA

experiments
Ashutosh Dixit , S.K. Yadav, N. Kumar, T.A. Khan, S. Hajela, M. Singhal
⁎

Nuclear Power Corporation of India Limited, Nabhikiya Urja Bhavan, Directorate of Reactor Safety & Analysis, Anushakti Nagar, Mumbai 400094, India

A R T I C L E I N F O A B S T R A C T

Keywords: The simulation of Small Break Loss-of-Coolant Accident (SBLOCA) experiments in the RD-14M integral test
SBLOCA facility is performed under the auspices of International Atomic Energy Agency (IAEA) as an International
RD-14M Collaborative Standard Problem (ICSP) with the objective to benchmark and validate the in-house developed
CANDU system thermal–hydraulic neutronic computer code ‘ATMIKA’, extensively used to analyze postulated events in
IAEA
Indian PHWRs. RD-14M is an 11 MW, full-elevation-scaled extensively instrumented thermal hydraulic
ICSP
ATMIKA
Canadian test facility, possessing most of the key components of a CANDU (CANada Deuterium Uranium) pri-
PHWR mary heat transport system (PHTS). The loop configuration is similar to figure-of-eight geometry of a typical
CANDU circuit and it is intended to reproduce the important geometric features of a reactor PHTS and the
appropriate operating conditions. ‘ATMIKA’ prediction and its comparison against SBLOCA experimental results
are compared in this paper. A specific SBLOCA experiment ‘B9006’ is selected for the Computer code ‘ATMIKA’
predictions. Test B9006 is a 7-mm inlet header break experiment with pressurized accumulator emergency
coolant injection (ECI) and represents most complete SBLOCA test conducted in RD-14M that includes all the
phases of the transient (blow-down, high-pressure ECI, secondary pressure ramp (crash cool), refill, low pressure
ECI, exponential pump ramp, and natural circulation). This simulation demonstrates that ‘ATMIKA’ is adequately
capable of predicting the break discharge, PHTS depressurization, channel flow rate, channel voiding, fuel
sheath temperatures and high pressure core injection flow through ECCS accumulator and initiation of low
pressure ECI for test B9006. ‘ATMIKA’ predicted results are compared with experimental results and it is seen
that predicted results for all phases of transient are in good agreement with experimental results.

1. Introduction parallel channel geometry).

Test facility RD-14M is an 11 MW, full-elevation-scaled thermal–-
NPCIL participated in an IAEA International Collaborative Standard hydraulic test facility possessing most of the key components of a
problem (ICSP) with the objective to benchmark and validate in-house CANDU primary heat transport system. RD-14M is extensively in-
developed system thermal-hydraulic neutronic computer code strumented facility with over 600 measurements such as temperatures,
‘ATMIKA’ against qualified data for Small Break Loss of Coolant pressures, flows, levels, and voids at various locations.
Accident (SBLOCA) scenarios generated on Canadian RD-14M Test This paper presents the modeling of Canadian test facility RD-14M
Facility (IAEA-TECDOC-1688, 2012). Computer code ‘ATMIKA’, is ex- with in-house developed computer code ‘ATMIKA’ for simulation of one
tensively used for safety analysis and licensing of Indian Pressurized of the selected test i.e. B9006. Analytically predicted transient by
Heavy Water Reactors (IPHWRs). computer code ‘ATMIKA’ has been compared with experimental results
In this IAEA-ICSP activity, two specific Small Break LOCA (SBLOCA) for the test B9006.
tests (B9006 and B9802) conducted in Canadian test facility RD-14M As seen through the results of PSA studies, performed for Indian
were selected and case B9006 has been presented in this paper as it PHWR which concludes that SBLOCA is the dominating event influen-
represents the most complete SBLOCA test with a 7-mm inlet header cing the core damage frequency. In view of that, benchmarking for most
break experiment which includes all the phases of the transient (blow- complete SBLOCA in PHWR becomes more relevant.
down, high-pressure ECI, secondary pressure ramp (crash cool), refill,
low pressure emergency coolant injection (ECI), exponential pump
ramp, natural circulation and bidirectional flow in the horizontal

⁎
Corresponding author.
E-mail address: adixit@[Link] (A. Dixit).

[Link]

0029-5493/ © 2017 Elsevier B.V. All rights reserved.

Received 23 September 2016; Received in revised form 21 July 2017; Accepted 31 July 2017

Please cite this article as: Dixit, A., Nuclear Engineering and Design (2017), [Link]
A. Dixit et al. Nuclear Engineering and Design xxx (xxxx) xxx–xxx

2. Brief description and assessment of the code with the new densities and energies are used by the momentum equa-
tion, and so on. In this manner, a time history of the fluid evolution is
Computer code ‘ATMIKA’ (IAEA-TECDOC-984, 1997; IAEA- obtained. This method is evolved to get maximum numerical stability
TECDOC-1149, 2000; IAEA-TECDOC-1688, 2012, and IAEA-TECDOC- with higher time step size.
1709, 2013) is a system thermal hydraulic neutronic computer code As a part of code development above models have been checked,
developed in Nuclear Power Corporation of India Limited (NPCIL) for verified and compared for their correctness with logical trends. The
the analysis of Loss-of-Coolant Accident (LOCA), main steam line break computer code ‘ATMIKA’ as an integration of all the above individual
and plant transient scenarios. The computer code ‘ATMIKA’ is based on models was assessed with respect to published results in open literature
Unequal Velocity Equal Temperature (UVET) model using three con- on Canadian experimental facilities. It is observed that overall predic-
servation equations with drift flux model. ATMIKA uses four equation tions of ATMIKA match reasonably well with the experimental results.
model where three conservation equations for homogeneous two phase Simulation of wide range of break sizes and locations on these facilities
and drift flux for phasic velocity difference. Phase separation and non- demonstrate the predicting capability of computer code ATMIKA.
thermal equilibrium in fuel channel is taken care by stratification Computer code ATMIKA is being extensively used in NPCIL to analyze
model. It is seen that with the modeling technique used in ATMIKA, it is loss of coolant accidents in Indian PHWRs.
equally capable of predicting all associated phenomena of SBLOCA and Computer code ATMIKA is in-house developed thermal hydraulic
can be compared with other six equation codes including RELAP5 with code which is well validated with experimental results as well as inter-
local minor deviation in the reactor system involving large source and code comparisons. Verification of ATMIKA code is done intensively in-
sinks. A staggered mesh arrangement is adopted where pressure, den- house within NPCIL. This code is used for licensing analysis of Indian
sity and enthalpy are defined at the node; flow is defined along the flow PHWRs and has been peer reviewed by the regulatory body, Atomic
path at the junction of two control volumes. Mass and energy con- Energy Regulatory Board (AERB) in India through various Committees/
servation equations are applied on lumped control volume and mo- Task Groups as a part of regulatory submission involving different
mentum equation is applied on flow paths. A semi-implicit scheme has scientific organizations and academic institutions.
been adopted for solving the set of differential equations.
Some of the important models that ATMIKA utilizes are wall heat 3. RD-14M facility description
transfer model for estimation of heat transfer from fuel to coolant and
coolant to steam generator/heat exchanger tube including various The RD-14M facility, shown schematically in Fig. 1, is a pressurized-
transition boundaries. Model uncertainties are suitably taken care by water loop with essential features similar to the primary heat transport
selecting reasonably conservative models based on engineering judg- loop of a typical CANDU reactor (RD-14M, 2010). The facility is de-
ment. Critical Heat Flux (CHF) is estimated based on Groeneveld’s signed so that reactor typical conditions, such as fluid mass flux, transit
Look-up table. AECL Look-up table is used for post dry out heat transfer time, pressure and enthalpy can be achieved in the primary-side for
coefficients (IAEA-TECDOC-1203, 2001). Critical discharge through the both forced and natural circulation. The design incorporates the basic
break is estimated using options from several available correlations e.g. “figure-of-eight” geometry of a CANDU reactor, with five horizontal
Burnell’s (for single-phase liquid), Moody, Homogeneous Equilibrium channels per pass and a 1:1 scaling of the vertical elevations throughout
Model, Frozen slip model. Friction Model is used to calculate pressure the loop. Each six-meter-long channel contains 7 electrically heated
drop across pipes. Among various options, model suggested by Fuel Element Simulators (FES), connected to end-fitting simulators. The
Chisholm based on Lockhart-Martinelli and Friedel correlation is used thermal characteristics of the FES are similar to CANDU fuel in terms of
to evaluate two-phase frictional pressure drop. There are options for power density, heat flux and heat capacity. Five reactor channel/feeder
various thermal hydraulic models related to two phase modeling in the geometries were selected, representing three middle channels, one top
code but during detailed validation exercise for special effects, appro- channel, and one bottom channel. Similar hydrostatic pressures main-
priate models are selected and kept same for remaining analysis. Based tained between RD-14M and a typical reactor by preserving the 1:1
on sensitivity studies, Lockhart-Martinelli model is dropped against vertical scaling.
Friedel/Chisholm. Pump Model is used for evaluating pump behavior The steam generators (SG) are scaled approximately 1:1 with typical
under single phase and two-phase conditions. Heat Conduction model is CANDU steam generators, in terms of tube diameter, mass flux, and
used for estimation of radial temperature profile in fuel rod and pipes. heat flux. Spiral-arm steam separators in the steam dome and flow re-
Computer code identifies the stratified flow regime in horizontal stricting orifices in the external downcomer of the SGs are used to
channel based on Taitel and Duckler model. The ability to calculate the produce reactor-typical recirculation each of the steam generators.
heat transfer from individual groups of pins in a fuel bundle subjected Primary fluid circulation is provided by two centrifugal pumps.
to stratified flow is incorporated into the code. Fuel temperature in case These deliver full reactor typical head (about 225 m) at flow rates si-
of stratified condition is estimated based on actual non-thermal equi- milar to a single reactor channel (about 24 kg/s). Primary circuit
librium vapor enthalpy conservatively. Thermodynamic properties of pressure is maintained by a loop pressurizer that contains an electrical
the coolant have been simulated in all range of sub-cooled, two phase heater.
and superheated conditions. A model has been incorporated for simu- The RD-14M facility is equipped with an Emergency Core Coolant
lation of indirectly heated Fuel Element Simulator (FES). In the ex- (ECC) system that provides cooling to the FES under postulated LOCA
periment, fuel bundles were electrically heated and (reactor) shutdown conditions. The ECC system injects emergency coolant into the primary
was simulated by tripping these electric heaters (which were tripped on heat transport system through any combination of the four headers. The
process signal of primary system pressure low). Therefore, no neu- ECC system is controlled by the primary loop pressure with the isolation
tronics were involved in the experiment. However, computer code A- valves automatically opening when header pressure drops below a
TMIKA is coupled with point/3D neutron kinetics code to predict predetermined pressure.
neutronic and thermal hydraulic behavior during postulated accident RD-14M is extensively instrumented facility to measure parameters
scenario. such as temperatures, pressures, flows, levels, and voids at various lo-
In the numeric, computer code ATMIKA, first solve the momentum cations.
equation assuming previous value of pressure which gives the flows
from one node to another. The updated flows are used by the mass and 4. Nodalization scheme
energy equations to update the mass and energy contents at each lo-
cation. The new mass and energy are given to the equation of state to The ATMIKA nodalization of the RD-14M test facility for tests
update the pressure distribution at each node. The new pressure, along B9006 consisting of primary, secondary and ECI system is shown in

2
A. Dixit et al. Nuclear Engineering and Design xxx (xxxx) xxx–xxx

Fig. 2. For the RD-14M test facility, nodalization has been done giving
due consideration to the instrumentation measurement locations. The
ATMIKA nodalization (Fig. 2) used in simulating tests B9006 consisted
of 225 thermal hydraulic nodes, 240 links and 181 heat transfer
structures. The volume, length, flow area and elevation change of each
pipe component are modeled, as closely as possible to the RD-14M test
facility. This ensured that the fluid transit time and hydrostatic pressure
changes around the loop were represented accurately in the simulation.
The primary side of RD-14M test facility is modeled with 195 nodes
which consist of all piping connecting the headers, heated sections, end-
fitting simulators, steam generators, pumps and pressurizer.
The heat transfer structures are used to model all solid components
in contact with the hot fluid. It also account for the heat transfer from
the primary fluid to the pipe walls and from pipe walls to the en-
vironment, or in the case of the steam generator tubes, to the secondary
side. Pipe radii (inner and outer) were used in defining the metal mass
and heat transfer area in contact with the primary fluid. Heat losses to
the environment are modeled based on steady state energy balance
through heat structure.
The steam generator secondary side consists of riser, drum and
downcomer volumes. The U-Tube heat slabs are connected to the four
volumes of the riser portion. Two volume downcomer connects the
drum to the secondary riser inlet. The feed is injected into the riser
portion and mixes with the saturated water from the downcomer. It
picks up heat from the U-tubes, converts into a two-phase mixture and
rises in the riser volume. At the exit of the riser, this two-phase mixture
enters the separator volume. The bottom volume of the SG drum is
connected to the downcomer. SG drum volume is simulated as a non-
homogenous control volume and steam lines were modeled as two
control volumes. The secondary side steam generator outlet pressures
were modeled using the pressure boundary conditions obtained from
boiler steam drum pressures. Time varying feed water flow rates and
temperature, provided based on experimental data are imposed as the
boundary conditions.
The nodalization of the ECI system includes the high pressure ECI
phase (high pressure ECI tank) and low pressure ECI phase (low pres-
sure pump) injection modes. Thirteen nodes were used for representing
the operation of both phases of the ECI system. Nominal test conditions
Fig. 1. RD-14M LOOP Schematic.
of Test are given in Table 1.

Fig. 2. Nodalization Diagram for Simulation of Test B9006.

3
A. Dixit et al. Nuclear Engineering and Design xxx (xxxx) xxx–xxx

Table 1 poised state of ECI accumulator. The ECC isolation valves were opened
Nominal test conditions. when headers pressure reached to 4.2 MPa and the low pressure
pumped ECI startedat1110 s. The test was terminated at 2284 s.
Outlet Header Pressure 10.0 MPa(g) @ Header 5
Nominal Input Power 4.0 MW per pass
Power in each Heated Section
Sections 5, 6, 7, 8, 9: 748.0, 748.0, 811.0, 944.0, 5.2. Discussion of results
757.0 kW
Sections 10, 11, 12, 13, 14: 749.0, 760.0, 822.0, 961.0, 760.0 kW In this test, the break is initiated at inlet-header HDR-8 (Fig. 2) at
Flow in each Heated Section (kg/s) 11.0 s, and the primary pumps were ramped down when PHT pressure
Sections 5,6,7,8, 9: 3.92, 4.00, 4.80, 5.03, 3.92
falls below 8.8 MPa which is imposed as boundary conditions with time
Sections 10, 11, 12, 13, 14: 3.80, 3.95, 5.22, 4.82, 3.88
Pump Flow 21.67 kg/s (full speed) to simulate the loss of class-IV power supply. The entire loop de-
Break Size 7.0 mm @ Header 8 pressurized relatively uniformly during this SBLOCA test, with differ-
Steam Drum Pressure 4.5 MPa(g) ential pressure driven mainly by the pumps running down. In this test,
Steam Separator Level 55%
following 7mm inlet header break at 11 s, the system pressure decreases
Feed water Temp. 180 °C
Pressurized Accumulator 4.2 MPa(g)
rapidly till it reaches saturation pressure of the fluid in outlet headers.
Low Pressure Pumped 1.5 MPa(g) System pressure falls to 8.8 MPa in HDR-7 at 11.3 s (ATMIKA predicts
11.6 s) and heaters trip as per logic and power reduces to decay level
(5%).
5. Results The initial depressurization to about 8 MPa in the outlet and
9.5 MPa in the inlet headers is the result of single-phase discharge from
Analysis for test B9006 is performed for 2284 s as per actual ex- the break until saturation is reached and two-phase discharge starts.
periment duration (RD-14M, 2007). Key parameters predicted by ‘A- The depressurization is considerably slower during the two-phase dis-
TMIKA’ as sequence of events is given in below Table 2 along with the charge and slows down even further when flashing occurs in the heated
experimental results for the purpose of comparison. channels both in experiments as well as in ATMIKA predictions (Fig. 3).
Emergency coolant injection begins when the selected header
pressure (HDR-8) drops below the pre-determined injection pressure.
5.1. Accident scenario
Since header 8 is the broken header, it has the fastest depressurization
rate among the four headers during the blowdown. Once sustained ECI
Test B9006 wasa7-mm inlet header break experiment with pres-
injection established, header pressure is determined by ECI pressure.
surized accumulator emergency coolant injection, performed in 1990
Pressure reduction in the primary system continues and pressure falls to
May. The break was represented by a fast-opening valve connected to
6.0 MPa, at 28.5 s (predicted at 27.2 s) at which ECI injection valves
an inlet header, and an orifice plate, scaled by the ratio of break area to
open as per logic.
loop volume to represent a feeder-sized break. Once the break valve
In this test, the ECI system was directed into all four headers (Figs. 4
was opened, single-phase liquid was discharged through the orifice,
and 5). This allows flow between the inlet header from one pass and
changing to two-phase flow when the inlet header pressure reached
outlet headers of the other pass, i.e. between HD5 and 6 and between
saturation. The break discharge flow was not measured directly. This is
HD7 and 8 (Fig. 5). This results in approximately 1/6 of the pump flow
the most complete SBLOCA test conducted in RD-14M in terms of in-
rate to bypass the heated sections, both in experiments as well as in
cluding all the phases of the transient (blowdown, exponential pump
ATMIKA predictions. ECC flow into each header starts when the par-
ramp, secondary pressure ramp (crash cool), high-pressure ECC, refill,
ticular header pressure drops below the ECC tank pressure of 4.2 MPa.
low pressure ECC, and natural circulation).The break was represented
Furthermore, when headers pressure reaches the ECC pressure of about
by a 7-mm orifice connected to the inlet header, and power was re-
4.2 MPa, the depressurization rate is slow due to a near-balance in
duced to decay power. This orifice valve was opened at inlet header 8 to
channel steaming and discharge from the break. Therefore, significant
simulate the break 11 s after data sampling started. About two seconds
net ECC inflow did not start until about 115 s in the experiment and in
after break initiation, the FES power was decreased to represent decay
ATMIKA predictions too.
power levels and the primary loop pump speeds were exponentially
The net flow of ECI into the system, based on the depletion of the
decreased to simulate the loss of class IV power. Secondary pressure
ECI tank, varied from 1.0 to 2.0 L/s in the initial stage. The high
ramp started when HDR-7 pressure reached to 6.0 MPa along with the
pressure accumulator tank ECI phase is terminated upon depletion of
the accumulator ECI tank at 1110 s (ATMIKA predicts 1063 s). As ex-
Table 2
pected, similar to experimental result, spike in header pressure
Sequence of events.
VALIDATION OF COMPUTER CODE ATMIKA AGAINST
Event/Action Experiments Prediction EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
( IAEA-ICSP : TEST - B9006 )
12.0
Data gathering started/calculation starts (s) 0.0 0.0
Surge tank isolated (s) 6.0 6.0
Legend
Blowdown valve (MV8) opens (s) 11.0 11.0 ATMIKA (BLIND)
Primary pump run down/ Power run down 11.3 11.6 RD-14M (EXP)
ATMIKA (OPEN)
(8.8 MPa @ HD7) (s) 8.0
PRESSURE (MPa)

Secondary pressure ramp ECI accumulator 28.5 27.2

poised
(6.0 MPa @ HD7) (s)
HP ECI valve open 78 74.3
4.0
(4.2 MPa) (s)
HP ECI start (s) 78 74.3
Channel reverse flow starts (HSx) (s) 196 (HS5, 10) 200(HS5, 10)
Loop 1 refills at (Plout)- > 0 (s) 440 400
HP ECI complete (10% of initial level) (s) 1110 1063 0.0
LP ECI start (s) 1110 1063 0 400 800 1200 1600 2000 2400
Loop 2 refills at (P2out)- > 0 (s) 1160 1013 Time (Sec)

End of test (s) 2284.3 2300

Fig. 3. Header 8 pressure.

4
A. Dixit et al. Nuclear Engineering and Design xxx (xxxx) xxx–xxx

VALIDATION OF COMPUTER CODE ATMIKA AGAINST VALIDATION OF COMPUTER CODE ATMIKA AGAINST
EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY. EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
( IAEA-ICSP : TEST - B9006 ) ( IAEA-ICSP : TEST - B9006 )
6 300

Legend
Legend RD-14M (EXP)
RD-14M (EXP) ATMIKA (OPEN)
ATMIKA (OPEN)
4

TEMPERATURE (DEG. C)
200
FLOW RATE (L/S)

100

-2 0
0 400 800 1200 1600 2000 2400
0 400 800 1200 1600 2000 2400
Time (Sec) Time (Sec)

Fig. 4. ECC to header 5 flow rate. Fig. 6. Boiler 1 inlet fluid temperature.

VALIDATION OF COMPUTER CODE ATMIKA AGAINST

VALIDATION OF COMPUTER CODE ATMIKA AGAINST
EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
( IAEA-ICSP : TEST - B9006 )
( IAEA-ICSP : TEST - B9006 )
300
6

Legend
Legend RD-14M (EXP)
RD-14M (EXP) ATMIKA (OPEN)
ATMIKA (OPEN)
4
TEMPERATURE (DEG. C)

200
FLOW RATE (L/S)

100

0
-2
0 400 800 1200 1600 2000 2400
0 400 800 1200 1600 2000 2400 Time (Sec)
Time (Sec)
Fig. 7. Boiler 1 outlet fluid temperature.
Fig. 5. ECC to header 8 flow rate.

VALIDATION OF COMPUTER CODE ATMIKA AGAINST

particularly in header-8 is observed due to isolation of high pressure EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
( IAEA-ICSP : TEST - B9006 )
ECI and initiation of low pressure ECI pump injection (Fig. 3). The final 400
loop pressure at the end of the high pressure ECI stage is around
Legend
1.4 MPa, both in experiment and predictions. In the low pressure RD-14M (EXP)
ATMIKA (OPEN)
pumped ECI phase, constant ECI flow in all headers is observed which 300
TEMPERATURE (DEG. C)

indicates that void from primary system is collapsed (Figs. 4 and 5). It is
observed that the predicted pressure and ECI flow transients are in good
agreement with experimental results (Figs. 3–5).
200
Boiler inlet and outlet temperature provide information about heat
transfer (removal/addition) capability of the boilers during the blow-
down transient. In initial period of transient, boiler 1 inlet temperature
decreases rapidly whereas in later period inlet temperature reduces 100

slowly in test as well as in ATMIKA predictions (Fig. 6). Boilers 1 outlet

temperatures are decreasing in accordance to falling boiler pressures
simulated by opening of main steam safety valves (Fig. 7). As soon as 0
cold ECI injection occurs in the system, boiler inlet temperature reduces 0 400 800 1200 1600 2000 2400
Time (Sec)
sharply in the predictions. After initial reduction (due to ECI), cooling
trend is similar in predictions as well as experiment. Boiler inlet and Fig. 8. FES temperature @ top pin, inlet HS13.
outlet temperature are predicted lower as compared to experimental
results in the initial period following cold water higher ECCS injection
header does not allow rise in sheath temperature. Sheath temperature
getting completely mixed in the simulation in all header injection mode
excursion has not been seen throughout the transient except some in-
with primary coolant at boiler inlet in UVET model.
termittent temperature rise at around 400 s. It is seen that in this
For heated section HS13, the inlet and outlet temperatures of the
channel (HS13), the sheath temperature is gradually increasing from
top FES show similar behaviours throughout the transient in the ex-
inlet to outlet FES along the length which clearly indicates that
periment and ATMIKA calculation (Figs. 8 and 9). Cold ECI flow in

5
A. Dixit et al. Nuclear Engineering and Design xxx (xxxx) xxx–xxx

VALIDATION OF COMPUTER CODE ATMIKA AGAINST VALIDATION OF COMPUTER CODE ATMIKA AGAINST
EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY. EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
( IAEA-ICSP : TEST - B9006 ) ( IAEA-ICSP : TEST - B9006 )
400 1.0

Legend
Legend
RD-14M (VF) (EXP)
RD-14M (EXP) 0.8 ATMIKA (OPEN)
ATMIKA (OPEN)

300
TEMPERATURE (DEG. C)

0.6

VOID FRACTION
0.4
200

0.2

0.0
100

-0.2

0 -0.4
0 400 800 1200 1600 2000 2400 0 400 800 1200 1600 2000 2400
Time (Sec) Time (Sec)

Fig. 9. FES temperature @ top pin, outlet HS13. Fig. 11. HS13 outlet void fraction.

continuous forward flow with no flow reversal in this channel. VALIDATION OF COMPUTER CODE ATMIKA AGAINST
EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
Channel void fractions are important as it indicate the onset of ( IAEA-ICSP : TEST - B9006 )
channel voiding and helps to explain the FES temperature behavior in 6.0

the heated sections. In this test, no voids are seen at the inlet of HS13 Legend
RD-14M (EXP)
except momentary void at around 200 s. Voids are seen at the outlet of ATMIKA (OPEN)
4.0
HS13 in the initial period of transient up to around 400 s. In code
predicted results, due to higher ECI flow in initial phase of transient,
FLOW RATE (L/SEC)

magnitude of voiding and duration of voiding are smaller as compared 2.0

to experimental results (Figs. 10 and 11).
In all the channels, flow starts reducing following the break together
with pump coast down due to depressurization of header and increase 0.0

in flow resistance in two phase conditions (Figs. 12–14). In heated

sections, HS8 and HS13, flow is maintained in its normal direction in
-2.0
experiment and ATMIKA predictions. Whereas in heated sections, HS5
and HS10, highest elevation channel, low flow/flow stagnation is ob-
served in initial period of transient subsequently reverse flow is main- -4.0
tained in experiment and ATMIKA predictions (Figs. 12–14). Inlet vo- 0 400 800 1200 1600 2000 2400

lumetric flow rate through HS-5 has captured less volumetric flow Time (Sec)

reversal as compared to experimental results in the initial period of the Fig. 12. HS5 inlet flow.
transient up to around 400 s. It is mainly due to relatively larger
voiding in the headers as seen in the experimental results as compared VALIDATION OF COMPUTER CODE ATMIKA AGAINST
to predictions in the initial phase of ECC mixing where non thermal EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
( IAEA-ICSP : TEST - B9006 )
equilibrium effects are dominating for short duration. However, after 8.0
400 s, predicted results are close to the experimental results. Legend
RD-14M (EXP)
In this test with several sensitivity analysis, it is observed that ATMIKA (OPEN)
6.0
usually, in broken pass flow reversal occurs in two channels (either in
FLOW RATE (L/SEC)

VALIDATION OF COMPUTER CODE ATMIKA AGAINST

EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY. 4.0
( IAEA-ICSP : TEST - B9006 )
1.0

Legend 2.0
RD-14M (VF) (EXP)
0.8
ATMIKA (OPEN)

0.6 0.0
VOID FRACTION

0.4
-2.0

0.2 0 400 800 1200 1600 2000 2400

Time (Sec)

0.0 Fig. 13. HS8 inlet flow.

-0.2
HS-12 or HS13 middle elevation channel and HS10 top channel)
whereas in unbroken pass, flow reversal occurs in only one high ele-
-0.4
vation channel (i.e. HS5). This is very well predicted by computer code
0 400 800 1200 1600 2000 2400
Time (Sec) giving confidence in the assessment of channel flow behavior under
failure of force circulation mode. This brings out the phenomenology of
Fig. 10. HS13 inlet void fraction.

6
A. Dixit et al. Nuclear Engineering and Design xxx (xxxx) xxx–xxx

VALIDATION OF COMPUTER CODE ATMIKA AGAINST

EXPERIMENTAL RESULTS OF RD-14M TEST FACILITY.
middle elevation channels in the broken pass under SBLOCA sce-
( IAEA-ICSP : TEST - B9006 ) nario. It is seen that during flow reversal, heat removal is effective
6.0
which is well predicted by computer code ATMIKA.

Legend
• ‘ATMIKA’ code is able to predict most of the important phenomenon
RD-14M (EXP) which are of great interest during the overall simulation of SBLOCA
4.0
ATMIKA (OPEN)
scenarios, mainly:
• Break discharge characteristics and critical flow
FLOW RATE (L/SEC)

2.0 • Consistent depressurization transient

• Two phase pressure drop
• Core flow distribution
0.0
• Natural Circulation
• Low flow and Flow reversal in horizontal parallel channels facility in
natural circulation.
-2.0
• Loop refill following ECC injection
• Timing of LP ECC start
-4.0 • The performed activity brings out the capability of a system thermal
0 400 800 1200 1600 2000 2400 hydraulics code with three conservation equations with drift flux
Time (Sec)
model as well as accounting non-thermal equilibrium effect in the
Fig. 14. HS10 inlet flow. core in particular estimation of fuel temperature in the stratified
condition in vapor region in the SBOCA experiment under natural
bidirectional flow in PHWRs in the later period of the transient fol- convective flow upon loss of forced circulation.
lowing MCP trip which gets further assisted following SBLOCA. • This work brings out the capability of a system thermal hydraulics
ATMIKA predictions for establishment of bidirectional flow in the core code with three conservation equations with drift flux model solved
under power supply failed condition are given below along with ex- in staggered mesh domain predicting all important phenomena ex-
perimental results: perienced in the SBLOCA experiment.
• This work also identifies the deviation in predicting few parameters
6. Conclusion and possible reason for the same and scope of further improvement
in the code.
The simulation of SBLOCA experiments in RD-14M integral test
facility is performed with the objective to benchmark and validate in- References
house developed system thermal–hydraulic neutronic computer code
‘RD-14M Small-Break LOCA Experiments for an IAEA International Collaborative
ATMIKA. With various phases of ICSP, following key conclusions are
Standard Problem’ 153–108210-440-001 Revision 0, October 2007.
drawn. ‘RD-14M Facility Description’, 2010. COG-00-034-R1 S & L WP-20104.
Bajaj, S.S., 1996. Computer Codes for Safety Analysis of Indian PHWR. IAEA, TECDOC-

• All main phenomena (e.g. break discharge, system depressurization, 984 on Advances in Heavy Water Reactor Technology proceeding of TCM held in
Mumbai, India.
temperature excursion, ECI injection and rewet in the heated sec- Dixit A., Yadav S.K., Kumar N., Khan T.A., Singhal M., Mohan N., Malhotra P.K., Ghadge
tions) are predicted well by the code. Deviations in quantitative S. G. and Chandra U., 2013. Assessment of Fuel Behaviour Under Large Break LOCA
terms are of minor in nature in particular dominated by local effects Condition For Indian Pressurized Heavy Water Reactor. IAEA-TECDOC-1709.
IAEA-TECDOC-1203, 2001. Thermo- hydraulic Relationship for Advanced Water Cooled
and do not affect the prediction of the overall system performance. Reactors.
• Prediction of well-established bidirectional flow in the horizontal IAEA-TECDOC-1688, 2012 .Comparison of Heavy Water Reactor Thermal Hydraulic Code
Predictions with Small Break LOCA Experimental Data. IAEA, Vienna.
parallel channel geometry of PHWR is distinguishably recognized in
Pramod P., Dubey A., Rammohan H.P., Bajaj S.S., 2000. Primary coolant circuit coast-
SBLOCA scenario in this international benchmark exercise. down and thermosyphon mode under failure of force circulation in Indian PHWR.
• Further, in horizontal parallel channel geometry, the flow reversal is IAEA-TECDOC-1149.
observed in high elevation channels in both the passes and in the