See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/281746540

Numerical Simulation of a Natural Circulation Loop

Conference Paper · October 2011

CITATIONS READS

0 317

3 authors:

Gabriel Verissimo Maria d de Lourdes Moreira

Federal University of Rio de Janeiro Nuclear Engineering Institute
13 PUBLICATIONS   28 CITATIONS    21 PUBLICATIONS   80 CITATIONS   

SEE PROFILE SEE PROFILE

J. L. H. Faccini
Nuclear Engineering Institute
32 PUBLICATIONS   122 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

CFD Euler-Euler modeling of sugarcane bagasse gasification in fluidized beds View project

Estudos Computacionais e Experimentais de Termo-hidráulica de Reatores Nucleares e Transporte de Radionuclídeos View project

All content following this page was uploaded by Gabriel Verissimo on 15 September 2015.

The user has requested enhancement of the downloaded file.

2011 International Nuclear Atlantic Conference - INAC 2011
Belo Horizonte,MG, Brazil, October 24-28, 2011
ASSOCIAÇÃO BRASILEIRA DE ENERGIA NUCLEAR - ABEN
ISBN: 978-85-99141-04-5

NUMERICAL SIMULATION OF A NATURAL CIRCULATION LOOP

Gabriel L. Verissimo1, Maria de Lourdes Moreira1 and José Luiz H. Faccini1

1
Instituto de Engenharia Nuclear – Comissão Nacional de Energia Nuclear,
Rua Hélio de Almeida, 75 – Cidade Universitária - Rio de Janeiro – BRAZIL
Zip Code 21941-906 - Postal Box 68550
gabrielverissimo@poli.ufrj.br; malu@ien.gov.br; faccini@ien.gov.br

ABSTRACT

This work presents a numerical simulation of a natural circulation loop using computational fluid dynamics
(CFD). The simulated loop is an experimental model in a reduced scale of 1:10 of a passive heat removal system
typical of advanced PWR reactors. The loop is composed of a heating vessel containing 52 electric heaters, a
vertical shell-tube heat exchanger and a column of expansion. The working fluid is distilled water. Initially it
was created a tridimensional geometric model of the loop components. After that, it was generated a
tridimensional mesh of finite elements in order to calculate the variables of the problem. The boundaries of the
numerical simulation were the power of the electric resistances and the cooling flow in the secondary side of the
heat exchanger. The initial conditions were the temperature, the pressure and the fluid velocity at the time just
before the power has been switched on. The results of this simulation were compared with the experimental
data, in terms of the evolution of the temperatures in different locations of the loop, and of the average natural
circulation flow as a function of time for a given power.

1. INTRODUCTION

The natural circulation is a phenomenon that has been largely used in many different
applications like solar heater, extraction of geothermic energy, coffee machines, boilers and
others. On the nuclear engineering field this phenomenon can be used in a residual heat
removal system of nuclear reactors. This kind of system has the objective to maintain the
temperature of the nuclear reactor core under control in case of forced cooling system fault or
in case of safe shutdown [1,2]. It presents competitive costs, less than forced cooling systems
and furthermore it is intrinsically safe since it works based on a natural physical phenomenon
with no active pumps, diesel generators or valves and no dependency on human intervention
[3,4].

Nowadays the natural circulation has been very much studied, experimentally and
numerically [5,6], for the better understanding of this phenomenon in order to increase the
safety of the future nuclear reactors. The use of computational codes for the study of this
phenomenon has grown in the last years as a consequence of the great and fast increase in the
velocity of computer processors. The numerical simulation is becoming a cheaper and faster
way to study the fluid dynamics in this field of engineering than it was some years ago.

The nuclear reactor design more common in the world is the Pressurized Water Reactor
(PWR), and for this reason many studies about this kind of reactor have been developed. The
two Brazilian nuclear power plants (Angra I and Angra II) use this type of reactor. This kind
of reactor is based on the capacity of the water remaining liquid when submitted to a very
high pressure [6,7]. The AP1000® pressurized water reactor works on the simple concept that,
in the event of a design-basis accident (such as a coolant pipe break), the plant is designed to
achieve and maintain safe shutdown condition without any operator action and without the
need for ac power or pumps. Instead of relying on active components such as diesel
generators and pumps, the AP1000 relies on the natural forces of gravity, natural circulation
and compressed gases to keep the core and containment from overheating. However, many
active components are included in the AP1000, but are designated as non safety-related.
Multiple levels of defense for accident mitigation are provided, resulting in extremely low
core-damage probabilities while minimizing occurrences of containment flooding,
pressurization and heat-up. The AP600 is a model of relatively small, 600 MWe nuclear
power plant designed by Westinghouse Electric Company. The AP600 has passive safety
features characteristic of the Generation III reactor concept. The design has been scaled up
and improved with the AP1000. Certification testing and analysis of the AP600 and AP1000
reactor designs for Westinghouse were conducted at the APEX facility at Oregon State
University. The one-quarter scale reduced pressure integral system certified the passively safe
systems that cool the reactor core using gravity and natural circulation.

This work presents a numerical study of the natural circulation in an experimental model in
reduced scale of a passive heat removal system similar to the AP600.

2. NATURAL CIRCULATION LOOP

The natural circulation loop studied in this work is formed by three main components: a heat
vessel, a heat exchanger and a column of expansion as is shown in the figure below.

Figure 1. The natural circulation loop.

The heat vessel has 52 electric heaters responsible for heating the water and inducing the
natural circulation phenomenon. As the natural circulation starts the water begins to move

INAC 2011, Belo Horizonte, MG, Brazil.

getting out of the heat vessel entering in a hot pipe leg and going to the heat exchanger. The
heat exchanger is a shell-tube type located 0.944m center to center above the heat vessel. In
the heat exchanger there are 7 vertical stainless steel straight tubes with 0.6 m length and
0.0314 m of internal diameter. Water at 25 OC and atmospheric pressure flows around the 7
tubes cooling the heated water that flows inside them.

The cooled water enters in a cold pipe leg which is connected to the expansion column
responsible to absorb the loop water volume changes. There is a safety valve installed in the
top of the column to protect it against over-pressure.

A data acquisition and control system has instrumentation in order to acquire flow, power and
temperature data. All the signal outputs from the instrumentation are received, processed and
recorded in this system, which includes one industrial computer containing A/D interface
boards and LabView software package, signal amplifiers/conditioners, switching panels,
A/D converters and computer displays. The power for the electric heaters is controlled from a
rack, which includes one AC/DC transformer, current shunt, power multipliers and other
electronic devices.

The loop has 12 thermocouples for monitoring the temperature of water in different locations
along the loop. An electromagnetic flow meter measures the natural circulation flow, and a
rotameter measures the flow entering the heat exchanger shell side. The heat exchanger shell
side flow rate is controlled by a needle valve during the experiments. More details about the
loop and the acquisition of data can be found in [9].

3. NUMERICAL SIMULATION

Numerical simulations provide a helpful tool to reduce the usually large experimental
expenses for nuclear reactor design. Due to the increase in computational power, direct
numerical simulations using mathematical models based on first principles become feasible
and prove extremely useful for understanding and predicting fundamental processes and
phenomena.

In this work it was used a commercial software for the numerical simulation of the loop
called CFX 12.1 from Ansys. To execute the simulation it was used parallel processing in a
machine with 12 processers.

3.1. Geometry and Mesh Generation

The geometry of the loop was created based on the mechanical drawings of the loop with a
tool of the software Ansys CFX. This geometry was used as basis for generate a mesh with
another tool of this software that uses the finite volume method. With the default options
given for the software it wasn’t possible to create a mesh. Because the complex geometry of a
few parts of the loop, specially the regions of the active length, it was necessary to improve
the quality of the mesh by reducing the element size. Another refinement was necessary at the
regions with curvature, like the regions near the electric heaters and near the tubes of the heat
exchanger. After that was necessary to increase the relevance of the mesh generated in the
software. After those improvements, a non structured mesh with 394,214 nodes and

INAC 2011, Belo Horizonte, MG, Brazil.

1,685,335 elements was created to be used in the numerical simulation. A picture of the mesh
generated is shown below.

Figure 2. The mesh of the loop.

3.2. Boundaries and Initial Conditions

After defining the mesh, the boundaries and initial conditions for the problem were modeled.
As initial conditions it was prepared a simulation of the loop with the heat exchanger working
and the electrical resistances turned off in the steady state.

The boundaries for this first simulation were the water entering the heat exchanger with a
velocity of 0.0044 kg/s and a temperature of 298 K. All the walls are defined with the
condition of no slip except the upper water level of the expansion column that is defined as
free slip, because it is a surface where the water is in contact with air. The results of this
simulation were used as initial condition for the natural circulation flow simulation.

The boundaries of the second simulation were the same as for the conditions for water
entering the heat exchanger and the walls conditions. The energy heating the water in the
region of the active length chosen was 800 W. The model of turbulence used was the κ-ε
model [10]. Initially it was considered that the loop was adiabatic, but the results show the
necessity of estimating a value for the heat loss for the environment. So after the simulation
with the assumption of adiabatic wall, a second simulation taking into account the loss of heat
for the environment was made.

3.3. Governing Equations

The governing equations are based in three physics principles called principle of Mass
Conservation, principle of Momentum Conservation and principle of Energy Conservation.
These equations are showed bellow.

INAC 2011, Belo Horizonte, MG, Brazil.

Equation of Continuity:

(1)

Equation of Momentum Conservation:

(2)

It is important to note that the last term of equation 2, is a special term added to take in
account the effects of buoyancy in the equation at the momentum equation.

Equation of Energy Conservation:

(3)

The equations of ideal gas must be added to the governing equations, they are given by:

Equation of State for density:

(4)

Equation of State for density:

(5)

Where: is the density, is the vector of velocity, is stress tensor due to the strain rate,
is a source term related to external forces, is the source term accounting the
effects of buoyancy, h is the enthalpy of the fluid, is the thermal conductivity, is a
source term related to energy.

4. RESULTS AND DISCUSSIONS

After the end of the simulations it was possible to visualize the behavior of the flow inside
the circuit. It was numerically obtained the temperatures of determined regions of the loop, as
the volumetric flow in the hot leg, and they were compared with the experimental values
given by the thermocouples and the rotameter. Some plots were made to compare the
behavior of the temperature and the volumetric flow against time and to observe the
differences between the two simulations and the experiment. The graphics are showed bellow
as a brief discussion about the results.

INAC 2011, Belo Horizonte, MG, Brazil.

 Figure 3. Entrance Hot Leg Fluid Temperature.

Figure 4. Hot leg fluid temperature.

INAC 2011, Belo Horizonte, MG, Brazil.

 Figure 5. Heat Exchanger Tube Fluid Temperature.

Figure 6. Entrance Cold Leg Fluid Temperature.

INAC 2011, Belo Horizonte, MG, Brazil.

 Figure7. Entrance Downcommer Fluid Temperature.

Figure 8. Lower Plenum Fluid Temperature.

INAC 2011, Belo Horizonte, MG, Brazil.

The comparison between the numerical simulation data and the experimental data shows that
the adiabatic simulation shows a tendency to stabilize with a greater temperature than it is
expected, compared to the values of the experiment. When these losses are considered it is
possible to observe that the temperatures of the numerical simulation are lower than the
temperatures of the experiment.

Figure 9. Volumetric Flow.

This graphic shows the volumetric flow on the cold leg given for the adiabatic simulation and
the simulation with losses plotted against the volumetric flow measured during the
experiment. It is possible to note a good agreement between the curves, with the volumetric
flow given for the simulation with loss lower than that for the adiabatic simulation.

5. CONCLUSIONS

After the comparison between the experimental data and the data given for the simulation, it
is possible to drawn some conclusions about the flow inside the loop. The first conclusion is
that the simulation of the loop considering the control volume adiabatic, presents a valid
option from the point of view of security. Since the fluid’s temperature of the simulation will
be always higher than the real temperature.

The second conclusion is that to obtain numerical results more precisely there is a need in
taking into account the losses of the loop to the environment. So, it is necessary to measure
the temperature of the environment during the experiment and to estimate the heat transfer

INAC 2011, Belo Horizonte, MG, Brazil.

 coefficient between the loop and the environment. This coefficient can be calculated using the
experimental and the numerical data. We note from the graphics that the coefficient
considered for the simulation with losses was higher than the real one, because of that the
losses of this simulation were higher than the losses observed in the experiment.

ACKNOWLEDGMENTS

To CNEN and CNPq for the financial support.

To Prof. Su Jian (COPPE/UFRJ) for the help with the numerical code.

REFERENCES

1. J. A. Perrota, “Proposta de um Núcleo de Reator PWR Avançado com Características

Adequadas para o Conceito de Segurança Passiva”, Doctor Thesis IPEN, São Paulo,
Brazil (1999).
2. L. A. Macedo, “Controle de Sistemas Passivos de Resfriamento de Emergência de
reatores Nucleares por Meio de Linhas de Desvio”, Master Thesis IPEN, São Paulo,
Brazil (2001).
3. Y. Zvirin, “A Review of Natural Circulation Loops in PWR and Other Systems”, Nuclear
Engineering and Design, 67 (2) pp. 203–225 (1982).
4. H. Bielinski, J. Mikielewicz, “Natural Circulation in Single and Two Phase
Thermosyphon Loop with Conventional Tubes and Minichannels”, Heat Transfer –
Mathematical Modelling, Numerical Methods and Information Technology, 19, 475-496
(2011)
5. D. A. de Andrade, G. S., P. E. Umbehaun, “Simulação e Análise do Fenômeno de
Circulação Natural Monofásica e Bifásica no Circuito Experimental Instalado NE
Engenharia Química POLI-USP, com o Código RELAP5”, INAC 05, Santos (2005)
6. C. B. Vieira, “Simulação Computacional da Convecção Natural em Cavidades Contendo
um Fluído com Geração Interna de Calor”, Master Thesis UFRJ, Rio de Janeiro, Brazil
(2010)
7. R. O. Pessoa, “Metodologia para Estudos de Circulação Natural em Circuitos Fechados”,
Master Thesis, IEN, Rio de Janeiro, Brazil (2009)
8. T. L. Fernandez Filho, “Análise de Eventuais Acidentes em Circuito Experimental de
Água, Utilizando o Código RELAP4”, Master Thesis IPEN, São Paulo, Brazil
9. “Natural Circulation Analysis”,
visl.technion.ac.il/NI/site/curriculum/docs/library_19a.doc (2000)
10. K. S. Chung, D. H. Thompson, “Study of turbulent Natural-Circulation Flow and Low-
Prandtl-Number Forced-Convection Flow”, National Heat Transfer Conference ,
Argonne, Illinois, United of States (1980)

INAC 2011, Belo Horizonte, MG, Brazil.

View publication stats