International Journal of Heat and Mass Transfer 54 (2011) 2618–2629

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ijhmt

A generalized flow equation for single phase natural circulation loops obeying
multiple friction laws
B.T. Swapnalee, P.K. Vijayan ⇑
Reactor Engineering Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India

a r t i c l e i n f o a b s t r a c t

Article history: The heat transport capability of natural circulation loops is directly proportional to the flow rate it can
Received 16 June 2010 generate. Therefore, reliable prediction of flow rate is essential for design and performance evaluation
Received in revised form 31 December 2010 of natural circulation loops. The reported generalized flow equation applicable for single-phase natural
Accepted 13 January 2011
circulation is only valid for cases where the entire loop follows a single friction law. Such a situation
Available online 24 February 2011
arises when the natural circulation loop is either fully laminar or fully turbulent. It is possible that a nat-
ural circulation loop can be partly laminar and partly in transition or turbulent flow. In such cases, a sin-
Keywords:
gle friction law is not applicable throughout the loop. In the present study, a generalized flow equation is
Single-phase natural circulation
Steady state flow equation
proposed for cases where a single friction law is not applicable for the entire loop. The proposed equation
Stability is tested with experimental data generated in a uniform diameter rectangular loop and is found to be in
Friction laws for laminar good agreement. Subsequently the equation is tested with data reported in the literature.
Transition and turbulent flow Stability analysis reported in literature for single-phase loops are either for laminar or turbulent flows.
In practice, in natural circulation loops, all the flow regimes like laminar, transition and turbulent are
observed. Therefore, it is required to develop a stability map, which is valid for all the three regions. In
the present paper, such a stability map is presented.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction
Natural circulation is the result of thermally generated buoy-
ancy pressure differential and hence such systems enhance reli-
ability and safe operational potential compared to pumped
systems. Due to this, natural circulation loops are extensively used
in several industries. In nuclear industry, some of the innovative
reactors MASLWR [1], CAREM [2], LFR [3], ABV [4], SSTAR and
STAR-LM [5] are designed to remove core heat by single-phase nat-
ural circulation, as these designs eliminate the need for recircula-
tion pumps. Natural circulation loops are susceptible to
instability. Operation with an unstable natural circulation loop is
undesirable as it can lead to problems in control and occurrence
of Critical Heat Flux (CHF) during low flow periods.
The primary function of a natural circulation loop is to transport
heat from a source to a sink. Since the heat transport capability of
natural circulation loops depend on the flow rate generated, reli-
able prediction of flow rate is essential. Dimensionless equations
are desirable to compare the performance of different natural cir-
culation loops. The reported dimensionless equation [6] for steady
state flow (in both uniform and non-uniform diameter loops) is
⇑ Corresponding author. Tel.: +91 22 25595157; fax: +91 22 25595151/+91 22
25519613.
E-mail address: vijayanp@barc.gov.in (P.K. Vijayan).
applicable only when the loop is obeying a single friction law
throughout. In many instances, the loop may follow more than
one friction law. The simplest example, where a part of loop fol-
lows one friction law and another part follows a different friction
law is a natural circulation loop with the cold leg in the laminar
and hot leg in the transition or turbulent flow regime. It can also
be caused by geometrical variations within the loop even if the
flow regime is the same. For example, nuclear fuel bundles follow
a different friction law compared to tubes even though both are in
the same flow regime. Similar situation can arise for a loop com-
prising of a combination of circular and non-circular (like annulus,
rectangular, trapezoidal, etc.) conduits in laminar flow regime.
Thus, even though the flow regime is same, the friction laws re-
quired can be more than one. Therefore, a generalized flow equa-
tion is essential for natural circulation loop obeying multiple
friction laws.
Besides the above, better accuracy in flow rate estimation can
be obtained even in case of single friction law if the viscosity var-
iation in the cold and hot legs of the loop are accounted as it is the
property which changes significantly with temperature. Large vis-
cosity variations occur in liquid metal systems due to the large
temperature difference (between hot and cold legs of the loop)
caused by the very low specific heat of liquid metals. It is usual
to have a temperature difference of the order of 200 °C or more
in liquid metal natural circulation systems using lead or

0017-9310/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2011.01.023
 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629 2619

Nomenclature

A flow area (m2) DTr reference temperature difference (°C) ðAQlhCHp Þ

b exponent in Eq. (2)
U overall heat transfer coefficient (W/m2 K)
C constant in Eq. (1)
W mass flow rate (kg/s)
Cp specific heat (J/kg K)
z elevation (m)
D hydraulic diameter (m)
Dz center line elevation difference between cooler and hea-
f Darcy friction factor
ter (m)
g gravitational acceleration (m/s2)
D3 q2 bg DT
Grm modified Grashof number ð 0l2 r Þ Greek symbols
H loop height (m) b thermal expansion coefficient (1/K)
k thermal conductivity (W/m K) l dynamic viscosity (Pa s)
l dimensionless length (L/Lt) q density (kg/m3)
L length (m) q0 reference density (kg/m3)
L1 laminar region length
L2 transition region length ¼ L02 þ L002 Subscripts
L3 turbulent region length 1 laminar region
Nt total number of pipe segments 2 transition region
NG geometric number defined by Eq. (16) 3 turbulent region
Num modified Nusselt number (UiLt/k) c cooler
Pr Prandtl number (Cpl/k) cl cold leg
P perimeter (m) h heater
p constant in Eq. (2) hl hot leg
q heat flux (W/m2) l laminar
Q total heat rate (W) r reference value
r constant in Eq. (1) s secondary side
Re Reynolds number (WD/Al) ss steady state
Stm modified Stanton number (4Num/RessPr) t total
s co-ordinate around the loop, m tr transition
t time (s) tu turbulent
T temperature (°C)

lead–bismuth eutectic as coolant [3,7,8]. In addition, a theoretical a heat sink supplied with specified coolant flow rate and inlet tem-
formulation which can account for the large property variation perature, a generalized equation for steady state flow (applicable
near the thermodynamic critical point better than the earlier gen- to both uniform and non-uniform diameter loops) has already been
eralized equation [6] is required. In the present paper therefore, derived in [6] and can be expressed as:
the generalized flow equation has been extended to cases where  r
the loop follows multiple friction laws. The proposed equation Grm
Re ¼ C ð1Þ
has been tested with experimental data generated in a uniform NG
diameter loop and is found to be in good agreement. Subsequently,
The above correlation is applicable if a friction law of the form
it has been tested with data reported in literature. Results of these
studies are reported in this paper. p
f ¼ ð2Þ
Stability analyses reported in the literature for single-phase Reb
loops are either for the laminar or for the turbulent flow regime
is valid throughout the loop with the same values of p and b. The
only. These analyses have generated stability maps for laminar
constants C and r are related to p and b as follows:
and turbulent flows using corresponding friction laws. The stability
maps are found to be a strong function of the friction laws em-  1
2 3b
ployed. Also, the stability map for laminar flow reported earlier C¼ ð3Þ
p
shows laminar condition only at low Grm values. Similarly, stability
map for turbulent flow shows that the flow is actually laminar in
1
the region of low Grm values. Hence, there is a need to generate a r¼ ð4Þ
single stability map valid for laminar, transition and turbulent re- 3b
gions for a single-phase loop. An attempt is made in this paper to The constants C and r take the values of 0.1768 and 0.5 for a fully
generate a single stability map valid for all three-flow regimes. laminar loop (p = 64 and b = 1) and 1.956 and 0.364 for a fully tur-
bulent loop (p = 0.316 and b = 0.25). The generalized equation given
above was previously tested by various authors Vijayan [6], Misale
2. Steady state flow in single-phase natural circulation loops and Garibaldi [9], Ambrosini et al. [10], Bodkha [11] and Vijayan and
Austregesilo [12]. From these reports, it has been observed that
Steady state flow prevails in a natural circulation loop when the most of the steady state experimental data are in close agreement
driving buoyancy pressure differential is balanced by the retarding with the proposed theoretical steady state correlation in the fully
frictional pressure losses. Based on this, an equation for the flow laminar and fully turbulent regimes. However, in the transition re-
can be derived. For an incompressible fluid with negligible viscous gion the agreement is not as good. In addition, the above mentioned
dissipation and axial conduction effects flowing in a natural circu- correlation is applicable only when the loop is either fully laminar
lation loop with a heat source supplied with uniform heat flux and or fully turbulent.
2620 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629

The simplest natural circulation loop conceivable is with a point Table 1

heat source and a point heat sink, connected by adiabatic vertical Flow regimes in natural circulation loop.

pipes (as considered by Welander [13]). Such a loop has essentially Sr. No. Nature of flow
two parts, i.e. a hot leg and a cold leg (see Fig. 1). As shown in Table Cold leg Hot leg
1, several flow regimes are possible even in such a simple natural
1 Laminar Laminar
circulation loop. As seen from the table, the assumption of a single 2 Turbulent Turbulent
friction law is valid for the first three cases only. 3 Transition Transition
For practical loops with finite lengths of heater and cooler, the 4 Laminar Turbulent
number of regions and the possible combination of flow regimes 5 Laminar Transition
6 Transition Turbulent
become much larger than that in Table 1. Even for a uniform diam-
eter loop, the hot leg can be in transition/turbulent flow with the
cold leg in laminar/transition flow regime. Further, consider the
case of a non-uniform diameter loop in which it is possible that
some pipe sections are in turbulent flow and some in laminar
and still others in transition flow. Therefore, the assumption that
the same friction law is applicable throughout the loop is not valid
always even in the simplest of natural circulation loops. Besides,
due to geometric variations in the loop such as presence of rod
bundle and noncircular conduits influences the friction law. Thus
to cater to these situations, it is necessary to develop a generalized
equation valid for the case of multiple friction laws which is done
in the following section.

3. Derivation of the generalized equation for a loop obeying

multiple friction laws

For this development, consider the uniform diameter rectangu-

lar closed loop shown in Fig. 2. Assuming one-dimensional incom-
pressible flow with negligible viscous dissipation and axial
conduction effects, the mass conservation equation for a closed
loop can be written in terms of the mass flow rate W as
@W
¼0 ð5Þ Fig. 2. Rectangular loop considered for analysis.
@s
This means that the mass flow rate is independent of the location in
I XNt  
the loop, and is a function of time only. The integral momentum Lt dW fi Li W2
equation can be written as  g q0 b T dz þ ¼0 ð6Þ
A dt i¼1
D 2 q 0 A2

In writing the above equation, it has been assumed that the local
pressure losses are negligible1 and Boussinesq approximation is va-
lid to account for the density variation. The energy equation can be
written for the loop as
8 qP
>
for heater
>
h

@T W @T < ACpq0
þ ¼ UPðTT Þ ð7Þ
@t q0 A @s > ACpq0
s
for cooler
>
:
0 for pipes
The steady state equations are obtained by dropping the temporal
derivatives. At steady state the momentum equation reduces to
I XNt  
fi Li W 2ss
g q0 b T dz ¼ ð8Þ
i¼1
D 2q0 A2

The energy equation at steady state is

8 qP
>
>
h

W ss dT < ACpq0
¼ UPðTT s Þ ð9Þ
q0 A ds >
>
:
ACpq0

0
The energy equation for the heater can be integrated with the
boundary condition that at s = 0, T = Tcl to obtain

Fig. 1. Schematic of a loop with point source and sink.
1
The assumption of negligible local pressure losses is justified for uniform
diameter loops. In a uniform diameter loop the contribution of local pressure losses is
negligible under laminar and transition condition where the friction factor is
reasonably high. Under turbulent conditions, the contribution of local pressure losses
is comparatively higher but is still less than 10% of the total pressure losses.
 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629 2621

qh P Lt Lt
ðT h Þss ¼ s þ ðT cl Þss ð10Þ NG ¼ ½l1 þ l2 þ l3  ¼ ð20Þ
W ss Cp D D
Similarly, the equation for the steady state temperature in the cool- since l1 + l2 + l3 = 1. This is same as that reported in Vijayan [6].
er can be obtained as (c) Length of transition region is negligible compared to the
UPðshl sÞ
length of laminar and turbulent regions (two-region model)
ðT c Þss ¼ T s þ ½ðT hl Þss  T s e W ss Cp ð11Þ " #
Lt p1 Reb33
where the boundary condition, at s = shl, T = Thl has been used. Using NG ¼ l1 b þ l3 ð21Þ
D p3 Re11
these equations, the integral in Eq. (8) can be calculated as
I (d) In many instances it can so happen that the same friction
Q hH
T dz ¼ ððT hl Þss  ðT cl Þss ÞH ¼ ð12Þ correlation is applicable, i.e. p1 = p3 and b1 = b3 = b
W ss Cp
"   #
Assuming the loop lengths in laminar, transition and turbulent flow Lt l b
NG ¼ l1 1 þ l3 ð22Þ
to be L1, L2 and L3 (see Fig. 2) respectively, we can rewrite the steady D l3
state momentum equation as
    If the viscosity variation is neglected, then this correlation becomes
Q hH f1 L1 f2 L2 f3 L3 W 2ss same as the one that was given in Eq. (20) above. The above equa-
g q0 b ¼ þ þ ð13Þ
W ss Cp D D D 2q0 A2 tion has been derived for a rectangular loop with horizontal heater
and horizontal cooler. For other orientations of the heater and cool-
Such that Lt = L1 + L2 + L3 where Lt is the total circulation length of
er, the same result can be used by taking center line elevation dif-
the loop. Alternatively, one can consider three regions in which
ference between the cooler and heater instead of the loop height.
the friction laws followed are different but has a general form of
When the loop is fully laminar, i.e. p1 = p3 = 64 and b1 = b3 = 1, Eq.
fi ¼ Repibi . In all the previous investigations, the fluid viscosity is as-
i (15) reduces to
sumed to be a constant throughout the loop. In the present investi-
gation, the viscosity is assumed to be constant in each region with 2 312
l1, l2 and l3 being the viscosities in region L1, L2 and L3 respec- 6 Gr 7
tively. Hence, the Reynolds number in each region is different due Re3 ¼ 0:17686
4   m 7
5 ð23Þ
Lt l1 b
to the difference in the viscosity value. With this, the steady state D
l1 l þ l3
3
momentum equation can be expressed as
  X ! When the loop is fully turbulent, i.e. p1 = p3 = 0.316 and
Q hH 3
p i Li W 2ss b1 = b3 = b = 0.25, Eq. (15) reduces to
g q0 b ¼ b 2
W ss Cp i¼1 Rei i 2Dq0 A 0 12:75
1
2 3
B Gr m C
6 p L1 p L2 p L3 7 W 2ss Re3 ¼ 1:956B
@  l b C
A ð24Þ
¼ 4 1 b þ  2 b þ  3 b 5 ð14Þ Lt
1
DW ss
2 3
DW ss DW ss 2Dq0 A2 D
l1 l1 þ l3
3
Al1 Al2 Al3

Taking region 3 as reference, the above equation can be expressed

in dimensionless form as 4. Validation of the steady state equation
 r
Grm The steady state equation derived above has been validated
Re3 ¼ C ð15Þ
NG with steady state natural circulation data generated in a uniform
diameter rectangular loop and with literature data.
where
" # 4.1. Experimental facility
Lt p1 Reb33 p2 Reb33
NG ¼ l1 þ l2 þ l3 ð16Þ
D p3 Reb11 p3 Reb22
A rectangular natural circulation test facility as shown in Fig. 3a
has been constructed and installed at BARC with the length scale of
  1
2 3b3 the different components as shown in Fig. 3b. It is a uniform diam-
C¼ ð17Þ
p3 eter rectangular loop of inside diameter 13.88 mm made of stain-
less steel (SS-347). The loop pressure is maintained with the help
1 of a pressurizer. The loop fluid fills the pressurizer up to a certain
r¼ ð18Þ level. Beyond this level, a cover gas (nitrogen) is filled. The loop
3  b3
had two heaters and two coolers. The heater was made of ni-
chrome wire wound uniformly on the outside surface. The cooler
3.1. Special cases was tube-in-tube type with secondary coolant flowing through
the annulus. The outside diameter of the inner tube is 21.34 mm
(a) Same friction law is applicable for all the three regions such and the inside diameter of the outer tube is 73.7 mm. Any of the
that p1 = p2 = p3 = p and b1 = b2 = b3 = b heater (either horizontal or vertical) and cooler (either horizontal
"    b #
Lt Re3
b
Re3 or vertical) combination for the experiments can be used. The un-
NG ¼ l1 þ l2 þ l3 used cooler secondary side was drained. The entire loop was insu-
D Re1 Re2
"   # lated with ceramic mat. The loop was extensively instrumented
 b
Lt l b l with K-type thermocouples to measure loop fluid temperature,
¼ l1 1 þ l2 2 þ l3 ð19Þ heater surface temperature as well as primary and secondary side
D l3 l3
coolant inlet and outlet temperatures. The loop had 44 thermocou-
(b) Same friction law is applicable and viscosity variation is neg- ples in all, 24 of which were used to measure the outside wall tem-
ligible so that l1 = l2 = l3 and Re1 = Re2 = Re3 perature in the heater sections. Sixteen thermocouples (two at
2622 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629
Fig. 3a. Photograph of the facility.
each of the eight locations shown in Fig. 3b) are used to measure
the fluid temperature in the primary side of the loop. In addition,
two thermocouples were used to measure the inlet and outlet tem-
peratures of the secondary side coolant. Since the secondary side
coolant flow was forced, this was adequate to measure the average
temperature. The heater power was measured using a Wattmeter.
The overall heat transfer coefficient of the cooler is estimated from
the measured power and fluid temperatures at inlet and outlet of
primary and secondary side. The secondary side flow rate was
measured with the help of a turbine flow meter. The accuracy of
temperature readings were within ±1 °C. Similarly, the accuracy
of the primary and secondary flow as well as power measurements
was within ±0.5% of the reading. The output of all instruments was
recorded on a PC based datalogger with a scanning speed of
100 ms.
4.2. Experimental data generated
Experiments were carried out with water at 30-bar pressure.
The secondary side coolant was chilled water. Steady state obser-
vations were taken for the following orientations of the heater
and the cooler:
(1) Horizontal Heater and Horizontal Cooler (HHHC).
(2) Horizontal Heater and Vertical Cooler (HHVC).
(3) Vertical Heater and Horizontal Cooler (VHHC).
(4) Vertical Heater and Vertical Cooler (VHVC).
For each orientation, the steady state data were generated for
different heater powers. The range of parameters for the
Fig. 3b. Schematic of the facility (all dimensions in mm).
experimental data for each orientation is given in Table 2. From
the measured temperature difference across the heater (DTh) and
the power Qh, steady state natural circulation mass flow rate Wss
is obtained as follows: W ss ¼ ðDT QÞh Cp where the specific heat Cph
h ss h
was calculated based on the average temperature (Th) of the heater
section (taken as the mean of the inlet and outlet temperatures).
Knowing the mass flow rate and the loop geometry, the Reynolds
numbers and Grashof numbers were calculated using the fluid
properties at the loop average temperature. The center line eleva-
tion difference between the heater and cooler used for calculation
of modified Grashof number is given in Table 2. Detailed experi-
mental data generated is given in Appendix A.
5. Friction factor correlation considered for laminar, transition
and turbulent region
Natural circulation loops are started up essentially from stag-
nant conditions. Hence, these loops experience laminar, transition
and turbulent flow conditions with progressive increase in power.
For example immediately after the flow initiation at low powers
laminar flow is expected to prevail throughout the loop for which
friction law is given by f = 64/Re. As the power increases the flow
increases and the loop enters the transition regime first at the hea-
ter outlet, followed by the hot leg and finally the cold leg making
the entire loop flow turbulent for which case the friction law is gi-
ven by f = 0.316/Re0.25.
The transition from laminar to turbulent flow is reported to oc-
cur at Re = 2000 for straight pipes. However, in a closed loop with
horizontal heater, cooler and several elbows, the transition is ex-
pected to occur at much lower Re values. For example, Creveling
et al. [14] reports that the transition to turbulence occurs at a value
of Re = 1500. On the other hand, Hallinan and Viskanta [15] reports
that the transition occurs at a Re value of 340. In numerical simu-
lations, generally the intersecting point of laminar and turbulent
 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629 2623

Table 2
Range of parameters for the test data.

Sr. No. Orientation Center line elevation difference Range of parameters for the data
between cooler and heater, m
Power, W Re Grm Secondary flow at 10 °C, l pm
Min Max Min Max Min Max Min Max
1 HHHC 4.10 225.4 3507.1 4.82E+2 1.02E+4 2.11E+10 1.55E+13 7.95 8.2
2 HHVC 2.75 504.4 3522.5 1.29E+3 1.01E+4 9.06E+10 2.27E+13 7.95 8.6
3 VHHC 2.70 244.9 3067.2 5.81E+2 5.91E+3 1.72E+10 5.87E+12 7.6 8.4
4 VHVC 1.35 249.5 2996.0 5.01E+2 6.58E+3 1.54E+10 6.35E+12 6.8 8.6

friction correlation is taken as the transition point [16,17]. The

transition zone is generally considered to be between 2000 <
Re 6 4000 for pipe flows [17]. In natural circulation loops due to
the presence of secondary flows, the transition could occur earlier
than that in straight pipe. Since transition occurs earlier in natural
circulation loops the transition to turbulence is also expected to
occur earlier in natural circulation loops.
Generally well-established friction law does not exist for the
transition zone. Most of the reported correlations are a mathemat-
ical extrapolation from laminar to turbulent friction law. In the
present paper, therefore it was decided to utilize the steady state
data on natural circulation flow in the transition zone to obtain a
friction factor correlation. As shown in Fig. 4, a flow correlation
was fitted to the transition region test data. From this correlation,
a power law of the form f ¼ Repb was obtained for the transition re-
gion using Eqs. (3) and (4). The friction law so obtained for the
transition zone is given by f ¼ 1:2063
Re0:416
. The same equation has been
compared with a few correlations reported for transition zone as
well as turbulent zone for forced flow in Fig. 5. As shown in Fig. 5. Comparison of transition region friction factor correlation with reported
Fig. 5, many of the reported transition correlations are close to correlations. (See above-mentioned references for further information.)
the proposed friction law.
In the present work the change from laminar to transition re-
gion and the transition to turbulent region is assumed based on
the criterion

f = max[fl, ftr] for change from laminar to transition.

f = max[ftr, ftu] for change from transition to turbulent.

These criterion actually causes the change from laminar to tran-
sition region to occur at Re = 898 and change from transition to tur-
bulent region to occur at Re = 3196. These are considered
appropriate for numerical solution where avoidance of numerical
instability while switching friction law is the primary concern.

Fig. 6. Friction coefficient for laminar, transition and turbulent regimes.

Fig. 6 shows a plot of f vs. Re for laminar, transition and turbulent

flow regimes with the above transition criteria. As expected, there
is no discontinuity in the plots of these equations. Similarly, the
same friction laws can be used in Eq. (15) to obtain the steady state
flow equation for fully laminar, fully transition and fully turbulent
loop as below.
 0:5
Gr m
Re ¼ 0:1768 fully laminar loop ð25aÞ
NG
 0:387
Grm
Re ¼ 1:216 fully transition loop ð25bÞ
NG
Fig. 4. Transition region correlation obtained from experimental data.
2624 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629
Fig. 7. Dimensionless flow for laminar, transition and turbulent regimes.
 0:364
Gr m
Re ¼ 1:956 fully turbulent loop ð25cÞ
NG
Plots of these three equations are given in Fig. 7. In this figure, the
three equations mentioned above have been plotted neglecting
property variation so that NG = Lt/D. As expected smooth transitions
has been observed while switching flow regimes in all the three
cases. In addition, Eqs. (16) and (21) has been used for calculating
NG where the three and two region models have been used respec-
tively with variation of viscosity with temperature and plotted in
the same figure. Between the plots of Eqs. (16) and (21), practically
no difference is found indicating that a two-region model is ade-
quate. A marginal difference is observed with plots of Eq. (25) with
maximum deviation of 7%.
6. Comparison with experimental data
Having finalized the friction laws for the laminar, transition and
turbulent zone we are in a position to compare the prediction of
Eq. (25) (where NG has been estimated by using Eq. (21)) with
the test data generated in the loop shown in Figs. 3a and 3b. As
mentioned above for a uniform diameter loop, it is possible that
the hot leg can be in transition/turbulent flow with the cold leg
in laminar/transition flow regime. For using the proposed equation
(as given in Eq. (25)), Reynolds number at hot leg and cold leg has
been calculated (assuming point heat source and point heat sink
and the corresponding hot and cold leg lengths are given in Table
3) from the steady state data generated in the experiment.
Depending upon the value of Reynolds number prevailing in hot
and cold legs different friction laws (i.e. 0 < Re 6 898 for laminar,
898 6 Re 6 3196 for transition and Re P 3196 for turbulent) has
been used in Eq. (25) and the resulting plot is given in Fig. 8 with
the estimated experimental error bar2.
The steady state equation derived above has been validated
with the steady state natural circulation data generated for various
orientations of the heater and cooler in the present loop. It has
been observed from Fig. 8 that the calculated Re by using the pro-
posed correlation is showing good agreement with the experimen-
2
Using the accuracy in measurement of power and temperature as given in section
4.1, the maximum error in the estimation of mass flow rate by heat balance is
estimated as ±5% , ±4%, ±3% and ±4% for HHHC, HHVC, VHHC and VHVC respectively.
Table 3
Hot and cold leg lengths for the differentiations of the test
loop.
Orientation L1/Lt L3/Lt
HHHC 0.5 0.5
HHVC 0.3 0.7
VHHC 0.7 0.3
VHVC 0.5 0.5
tal data. The entire data generated in the present experiments are
compared with the present correlation in Fig. 9.
7. Comparison with literature data
Single-phase natural circulation experiments have been re-
ported by a large number of researchers [9,11,22–27]. However,
in many cases, [9,23] the data on hot leg and cold leg temperatures
are not reported. For the validation of the steady state flow equa-
tion presented in this report usable data could be found only in
Refs. [11,22,24–27]. The entire data from literature including the
present experimental data where NG has been calculated using
Eq. (21) is compared in Fig. 10. From Fig. 10, it has been observed
that the data falling in the transition region is very close to the pro-
posed correlation. As shown in Fig. 7, the maximum deviation be-
tween Eq. (25) (single region model with constant properties) and
the two or three region models are only 7%. Thus, Eq. (25) is ade-
quate for most practical situations. The advantage of Eq. (25) is that
NG can be calculated as Lt /D and all the reported data on single-
phase natural circulation can be utilized for its validation. Fig. 11
compares the data available in the literature [28–36] (where NG
has been calculated using Eq. (20)) along with all the data plotted
in Fig. 10.
It is observed from Fig. 11 that by using Eq. (25) the agreement
with the proposed correlation is reasonable (approximately within
±54%). Considering the variation of viscosity with temperature the
prediction accuracy can be significantly improved (within ±12%) as
shown in Fig. 10.
8. Stability analysis
Eq. (15) enables us to calculate the steady state flow in a Natural
Circulation Loop (NCL). However, it does not tell us whether that
particular steady state is stable. Thus, stability analysis is necessary
to ascertain whether a particular steady state is stable. In general,
stability of a NCL is analyzed either by linear stability method or by
non-linear stability method. Stability analysis using non-linear
method involves tedious computations and also is very expensive.
Therefore, linear stability analysis provides the best route to gener-
ate a stability map. In the linear stability method, the governing
equations are perturbed over the steady state and the perturbed
governing equations are linearized to obtain a characteristic equa-
tion for the stability parameter. If this equation has any root with
positive real part then the system is considered to be unstable in
accordance with the Nyquist criterion. In most of the published lit-
erature, the results of the linear stability analysis have been given
in dimensionless form. Usually the stability map in dimensionless
form is plotted as a function of modified Grashof number, Stanton
number, and Lt/D for a specified geometry of the loop. But for cer-
tain conditions like, when the Grashof number is very high, it be-
comes difficult to ascertain whether the system remains in
single-phase condition. For this reason, a computer code LISA (LIn-
ear Stability Analysis) has been developed to carry out the steady
state and linear stability analysis of a single phase natural circula-
tion loop in dimensional form. The geometry and co-ordinate
 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629 2625

Fig. 8. Comparison of test data with proposed equation.

Fig. 9. Steady state natural circulation flow in a uniform diameter loop. Fig. 10. Steady state natural circulation flow in uniform diameter loops.
2626 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629

Fig. 11. Steady state natural circulation flow in uniform diameter loop (literature data).

system considered for the stability analysis is shown in Fig. 12. In The small perturbations T0 and W0 can be expressed as
the linear stability method, the steady state flow and temperature
are perturbed by an infinitesimal amount as below: T 0 ¼ TðsÞeert and W 0 ¼ W eert ð31Þ
0 0 So that
T ¼ T ss þ T and W ¼ W ss þ W ð26Þ
where T0 and W0 are small perturbations over the steady state val- @T 0
¼ TðsÞeert r ð32Þ
ues. Substituting these in Eq. (6), the perturbed momentum equa- @t
tion after linearizing can be written as
I @T 0 @TðsÞ rt
dW
0
g q bA ð1bÞ
ð2  bÞplb W ss W0 ¼ ee ð33Þ
¼ 0 0
T dz þ ð27Þ @s @s
dt Lt 1þb 1b
2D q0 A
where e is a small quantity and r is the stability parameter, which is
Similarly, the perturbed energy equations for the different segments a complex number. Substituting Eqs. (31)–(33), the momentum
of the loop can be written as equation (i.e. Eq. (27)) can be written as
 
@T 0 W ss @T 0 W 0 qh P I
þ þ ¼ 0 heater 0 < s 6 sh ð28Þ g q0 bA ð2  bÞplb W 1b
ss W
@t q0 A @s q0 A W ss Cp FðrÞ ¼ rW  TðsÞdz þ ¼0 ð34Þ
Lt 2D1þb q0 A1b
@T 0 W ss @T 0
þ ¼ 0 pipes sh < s 6 shl and sc < s 6 st ð29Þ Similarly substituting T0 and W0 in the energy equations and solving
@t q0 A @s the resulting equation, the following equation can be obtained for
" # the heater
@T 0 W ss @T 0 UP 0 W= " q Asr #
þ þ T  ðT ss  T s Þ ¼ 0 cooler shl < s 6 sc qh PW 0 q0 Asr
@t q0 A @s q0 ACp W ss TðsÞ ¼ e W ss W
 1 þ T cl e ss ð35Þ
q0 AW ss Cpr
ð30Þ

Fig. 12. Loop geometry and co-ordinates for HHHC orientation. Fig. 13. Nyquist plot.
 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629 2627

where the boundary condition that at s = 0, TðsÞ ¼ T cl , has been valid for shl < s 6 sc, where (Thl)ss is given by
used. Similarly, one can obtain the following equation for hot leg " #
q0 Arðsh sÞ qh Psh 1
TðsÞ ¼ T h e W ss for sh < s 6 shl ð36Þ ðT hl Þss ¼ T s þ ð39Þ
W ss Cp 1  eUPðs hl sc Þ
W ss Cp
where the boundary condition is that at s = sh, TðsÞ ¼ T h has been
used to eliminate the integration constant. Similarly, the equation Using these, the integral in the momentum equation, i.e. Eq. (34)
for cold leg is obtained as can be evaluated as
I 

q0 Arðsc sÞ W ss q0 Arðsc st Þ q0 Arðsh shl Þ

TðsÞ ¼ T c e W ss for sc < s 6 st ð37Þ TðsÞdz ¼ ðT h  T c Þ þ T c e W ss  T h e W ss ð40Þ

q0 Ar
where the boundary condition that at s = sc, TðsÞ ¼ T c has been used. where
Equation for the cooler is obtained as  q Ars 
q0 Arsh Wqh P 0 h

"  # T h ¼ T cl e W ss þ e W ss  1 ð41Þ
q0 Aðshl sÞ q0 ArCpW ss
WUP½ðT hl Þss  T s  UPðsWhl sÞ rþq UP
TðsÞ ¼ e ss  e W ss 0 AC p
and
q0 AC p r
 
q0 Aðshl sÞ rþ UP
q0 AC p
þ T hl e W ss ð38Þ

Fig. 14. Stability map.

2628 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629

2   3
q0 A rþ UP ðshl sc Þ 9. Conclusions
WUP½ðT hl Þss  T s  6 UPðshl sc Þ q0 ACp
7
Tc ¼ 4e W ss Cp e W ss
5
q0 ArW ss Cp A generalized equation for the steady state flow in single-phase
h i natural circulation loops was derived based on 1-D theory by
q0 A rsc þ UP ðshl sc Þ
q0 ACp assuming the loop to be partly in laminar and partly in transition
þ T cl e W ss or turbulent flow. The derived dimensionless flow equation is
2 n o n o3 applicable for any loop obeying multiple friction laws. The general-
q0 A rsc þ UP ðshl sc Þ q0 A rðsh sc Þþq UP ðs s Þ
Wqh P 6 q0 ACp 0ACp hl c
7 ized flow equation was tested with the experimental data
þ 4e W ss e W ss
5 generated in a uniform diameter rectangular loop for all four orien-
q0 ArCpW ss
tations such as, Horizontal Heater and Horizontal Cooler (HHHC),
ð42Þ Horizontal Heater and Vertical Cooler (HHVC), Vertical Heater
and Horizontal Cooler (VHHC), Vertical Heater and Vertical Cooler
(VHVC). From the above, it is observed that the same generalized
XþY correlation is applicable for all the four orientations of heater
T cl ¼ n o ð43Þ
and cooler tested. Data reported in the literature is also showing
q0 A rst þ UP ðshl sc Þ
q0 ACp
good agreement with the generalized equation developed.
1e W ss
The stability code LISA based on linear stability analysis meth-
odology has been validated with previously reported stability
where map for both laminar and turbulent flows. Subsequently a stability
map valid for laminar, transition and turbulent regions has been
2 n o n o3
q0 A rðsc st Þþq UP ðs s Þ q0 A rðshl st Þþq UP ðs s Þ
obtained using the LISA code for the same loop. The results showed
ACp hl c ACp hl c
WUP½ðT hl Þss  T s  6 0 0
7 that there is no stable zone in the subcritical region for Stm < 7 for
X¼ 4e W ss e W ss
5
q0 ArCpW ss turbulent flow for the loop considered in this study. Also by consid-
ering the various effects such as heat loss, wall effects, local pres-
ð44Þ sure losses and multidimensional effects the prediction can be
improved.
2 n o n o3
q0 A rst þ UP ðshl sc Þ q0 A rðsh st Þþq UP ðs s Þ
ACp hl c
Wqh P 6 q0 ACp 0
7
Y¼ 4e W ss e W ss
5 ð45Þ Appendix A
q0 ArCpW ss
Steady state experimental data generated for the various orien-
for the HHHC orientation. The roots of the characteristic equation tations of the heater and cooler (see Tables A.1–A.4).
(i.e. Eq. (34)) were found with the help of Regula-Falsi method. A
Table A.1
computer code LISA was written to identify the roots. Nyquist plots Steady state data for HHHC orientation.
were also made as shown in Fig. 13 for a few cases, which confirmed
the code predictions. Power, Tmean Thi Tho Wss Grm Grm(D/Lt) Re
W (°C) (°C) (°C) (kg/s)
The stability map (i.e. the locus of neutrally stable points) pre-
dicted for laminar flow reported in [37] is reproduced in Fig. 14(a) 225.4 29.6 24.4 34.8 0.005175 2.11E+10 1.82E+07 4.82E+02
759.0 46.1 38.3 53.9 0.011665 2.81E+11 2.59E+08 1.53E+03
and (b) using LISA code. As can be seen, LISA code is found to ade- 1251.0 59.0 49.2 68.8 0.015245 1.05E+12 9.04E+08 2.61E+03
quately predict the stability threshold. However, for the upper 1716.6 68.3 56.9 79.6 0.018168 2.10E+12 1.97E+09 3.86E+03
threshold in turbulent flow, the LISA code predicts always unstable 2233.6 77.7 64.9 90.5 0.020918 4.72E+12 4.41E+09 6.38E+03
operation right up to the critical point. Ambrosini et al. [38] also 2540.8 82.1 68.5 95.6 0.022374 6.66E+12 6.23E+09 6.52E+03
2734.2 85.9 71.8 99.9 0.023202 8.20E+12 7.68E+09 7.20E+03
predicted using Churchill friction law and friction law suggested
3096.5 89.6 74.6 104.5 0.024673 1.06E+13 9.89E+09 8.54E+03
by Vijayan et al. that the flow is fully unstable in turbulent zone. 3253.4 93.1 77.7 108.5 0.025189 1.22E+13 1.14E+10 9.12E+03
However, in the laminar flow stability map (Fig. 14a), the flow is 3507.1 96.1 80.3 111.9 0.026320 1.55E+13 1.44E+10 1.02E+04
not really laminar beyond a Grm value of 6.9E+9 (for Re = 898).
Likewise, in the turbulent flow stability map (Fig. 14b) the flow
is not turbulent below a Grm value of 1.8E+11 (for Re = 3196).
Table A.2
Therefore, these stability maps do not represent the true stability
Steady state data for HHVC orientation.
map of a single-phase rectangular natural circulation loop. In the
present work a stability map is generated which is valid for lami- Power, Tmean Thi Tho Wss Grm Grm(D/Lt) Re
W (°C) (°C) (°C) (kg/s)
nar, transition and turbulent flow regimes (see Fig. 14c). In the
present analysis, the upper stable zone of the stability map has 504.4 42.9 36.1 49.7 0.008856 9.06E+10 8.42E+07 1.29E+03
749.5 52.5 45.6 61.3 0.011433 2.63E+11 2.48E+08 2.01E+03
not been found in the subcritical region for Stm < 7 for turbulent
999.8 60.1 51.2 69.0 0.013442 5.34E+11 5.03E+08 2.64E+03
flow. Also for the present test loop a stability map is generated in 1263.4 70.4 60.3 80.4 0.015061 1.22E+12 1.14E+09 3.41E+03
dimensional form (Fig. 14d). It has been observed from the 1497.8 75.3 64.6 86.0 0.016717 1.91E+12 1.82E+09 4.06E+03
Fig. 14d that a few stable experimental data fall in the unstable re- 1775.5 82.9 71.3 94.5 0.018256 3.13E+12 2.99E+09 4.84E+03
gion. Similar observations are reported by Garibaldi [39]. This can 2000.3 87.8 75.4 100.1 0.019300 4.17E+12 3.98E+09 5.43E+03
2265.4 94.8 81.7 107.9 0.020574 6.33E+12 6.04E+09 6.27E+03
be attributed to the simplifying assumptions made in the linear
2521.8 99.6 85.8 113.3 0.021832 8.66E+12 8.27E+09 7.00E+03
stability analysis. For example, heat loss, wall effects, local pressure 2790.3 105.4 90.9 119.8 0.022906 1.22E+13 1.16E+10 7.84E+03
losses and multidimensional effects have not been considered. In 3035.6 109.5 94.5 124.5 0.023977 1.47E+13 1.40E+10 8.53E+03
this context, it may also be noted that the analysis by Misale 3051.6 110.6 95.6 125.6 0.024136 1.53E+13 1.46E+10 8.68E+03
et al. [40] showed that the wall effects play a significant role on 3275.9 114.8 99.2 130.3 0.024919 1.80E+13 1.72E+10 9.29E+03
3522.5 118.9 103.0 134.8 0.026085 2.27E+13 2.17E+10 1.01E+04
stability of single-phase natural circulation.
 B.T. Swapnalee, P.K. Vijayan / International Journal of Heat and Mass Transfer 54 (2011) 2618–2629 2629

Table A.3 [11] K. Bodkha, Study on stabilization of single-phase natural circulation system
Steady state data for VHHC orientation. using mechanical gadgets, M. Tech. Thesis, Homi Bhabha National Institute,
Mumbai, India, 2009.
Power, Tmean Thi Tho Wss Grm Grm(D/Lt) Re [12] P.K. Vijayan, H. Austregesilo, Scaling laws for single-phase natural circulation
W (°C) (°C) (°C) (kg/s) loops, Nucl. Eng. Des. 152 (1994) 331–347.
[13] P. Welander, On the oscillatory instability of a differentially heated fluid loop,
244.9 32.3 26.3 38.2 0.004937 1.72E+10 1.40E+07 5.81E+02
J. Fluid Mech. 29 (1967) 17–30.
265.1 32.5 24.4 40.6 0.003915 1.75E+10 1.32E+07 4.51E+02 [14] H.F. Creveling, J.F. De Paz, J.Y. Baladi, R.J. Schoenhals, Stability characteristic of
501.2 38.5 29.5 47.5 0.006670 5.84E+10 5.89E+07 8.84E+02 a single-phase free convection loop, J. Fluid Mech. 67 (1975) 65–84.
746.5 47.4 38.7 56.1 0.010279 1.82E+11 1.63E+08 1.61E+03 [15] K.P. Hallinan, R. Viskanta, Heat transfer from a rod bundle under natural
1004.4 50.9 39.9 61.9 0.010964 2.84E+11 2.49E+08 1.81E+03 circulation conditions, NUREG/CR-4556.
1232.3 60.4 49.6 71.1 0.013683 6.81E+11 6.06E+08 2.65E+03 [16] P.K. Vijayan, H. Austregesilo, V. Teschendorff, Simulation of the unstable
1546.9 63.2 49.7 76.6 0.013791 9.33E+11 7.93E+08 2.75E+03 behavior of single phase natural circulation with repetitive flow reversals in a
1749.3 59.9 49.8 69.9 0.020732 9.67E+11 8.95E+08 4.01E+03 rectangular loop using computer code ATHLET, Nucl. Eng. Des. 155 (1995)
2042.4 71.2 56.2 86.1 0.016318 1.87E+12 1.70E+09 3.65E+03 623–641.
2266.7 75.8 59.8 91.7 0.016944 2.73E+12 2.48E+09 4.06E+03 [17] W. Ambrosini, J.C. Ferreri, Stability analysis of single-phase thermosyphon
2564.9 78.9 63.2 94.7 0.019426 3.57E+12 3.24E+09 4.89E+03 loops by finite difference methods, Nucl. Eng. Des. 201 (2000) 11–23.
2797.9 83.8 66.3 101.3 0.019072 4.85E+12 4.39E+09 5.06E+03 [18] J.G. Knudsen, D.L. Katz, Fluid Dynamics and Heat Transfer, International
Student Edition., McGraw-Hill Book Company, New York, 1958. pp. 146–212.
3067.2 86.3 69.3 103.2 0.021530 5.87E+12 5.33E+09 5.91E+03
[19] D.W. Green, RobertH Perry, Perry’s Chemical Engineer’s Handbook’’, eighth ed.,
McGraw-Hill, 2007. pp. 6.1–6.56.
[20] International Atomic Energy Agency, Thermo hydraulic relationships for
Table A.4 advanced water-cooled reactors, IAEA-TECDOC-1203, April 2001, pp. 118–119.
Steady state data for VHVC orientation. [21] M. Thirumaleshwar, Fundamentals of heat and mass transfer, Pearson Educ.
(2006) 454–455.
Power, Tmean Thi Tho Wss Grm Grm(D/Lt) Re [22] P.K. Vijayan, M.H. Bade, D. Saha, R.K. Sinha, V. Venkat Raj, A generalized
W (°C) (°C) (°C) (kg/s) correlation for the steady state flow in single state natural circulation loops,
BARC/2000/E/021.
249.5 35.9 27.9 44.0 0.003704 1.54E+10 1.27E+07 5.01E+02 [23] H.H. Bau, K.E. Torrance, On the stability and flow reversal of an asymmetrically
250.6 29.9 21.8 37.9 0.00372 8.25E+09 6.70E+06 4.28E+02 heated open convection loop, J. Fluid Mech. 106 (1981) 417–433.
509.9 44.1 32.6 55.5 0.005325 4.99E+10 5.02E+07 8.05E+02 [24] D.C. Hamilton, F.E. Lynch, L.D. Palmer, The nature of flow of ordinary fluids in a
735.4 50.7 40.3 61.0 0.008511 1.28E+11 1.16E+08 1.45E+03 thermal convection harp, ORNL-1624, March 16, 1954.
1000.3 55.4 44.3 66.5 0.010766 2.20E+11 1.98E+08 1.99E+03 [25] B.J. Huang, R. Zelaya, Heat transfer behavior of a rectangular Thermosyphon
1253.6 65.9 53.5 78.4 0.012042 5.08E+11 4.61E+08 2.58E+03 loop, J. Heat Transfer 110 (1988) 487–493.
1556.9 74.7 60.9 88.4 0.0135 9.70E+11 8.75E+08 3.27E+03 [26] P. Dubey, G.S.S.P. Rao, D.S. Pilhawal, P.K. Vijayan, D. Saha, Analysis of
1752.3 80.7 66.1 95.3 0.014321 1.36E+12 1.23E+09 3.75E+03 experimental data on two-phase natural circulation from the flow pattern
2039.7 85.9 70.2 101.7 0.015437 2.02E+12 1.87E+09 4.29E+03 transition instability studies facility at Apsara reactor, Bhabha Atomic
2342.01 96.4 79.2 113.6 0.016171 3.40E+12 3.16E+09 5.03E+03 Research Centre, Mumbai, India, 2004. BARC/2004/E/031.
[27] C.D. Alstad, H.S. Isbin, N.R. Amundson, Transient behavior of single phase
2498.7 96.6 79.2 114.0 0.017075 3.74E+12 3.48E+09 5.37E+03
natural circulation loop systems, AIChE J. 1 (4) (1955) 417–425.
2756.3 103.5 85.3 121.6 0.018023 4.90E+12 4.56E+09 6.04E+03
[28] P.K. Vijayan, A.K. Nayak, D. Saha, M.R. Gartia, Effect of loop diameter on the
2996.0 107.0 88.2 125.8 0.018870 6.35E+12 5.91E+09 6.58E+03 steady state and stability behavior of single phase and two phase natural
circulation loops, Science and technology of nuclear installations, vol. 2008,
article ID: 672704, 17 pages, doi:10.1155/2008/672704.
[29] T. Nishihara, Oscillatory instability of a single-phase natural circulation loop,
References in: proceedings of the 8th International Topical Meeting on Nuclear Reactor
Thermal-Hydraulics (NURETH’97), Kyoto, Japan, September–October 1997, pp.
[1] S.M. Modro, J. Fisher, K. Weaver, P. Babka, J. Reyes, J. Groome, G. Wilson, 839–847.
Generation-IV Multi-Application Small Light Water Reactor (MASLWR), in: [30] H.H. Bau, K.E. Torrance, Transient and steady state behavior of an open
Proceedings of ICONE-10, Tenth International Conference on Nuclear Energy, symmetrically heated free convection loop, Int. J. Heat Mass Transfer 24 (4)
April 14–18, 2002, Arlington, Virginia, USA (1981) 597–609.
[2] R. Mazzi, CAREM: an innovative integrated PWR, in: 18th International [31] M. Misale, P. Garibaldi, J.C. Passos, G.G. de Bitencourt, Exp. Thermal Fluid Sci.
Conference on Structural Mechanics in Reactor Technology (SMiRT-18), 31 (8) (2007) 1111–1120.
Beijing, China, August 7–12, 2005, SMiRT18-S01-2. [32] S.K. Mousavian, M. Misale, F. D’ Auria, M.A. Salehi, Transient and stability
[3] C.F. Smith, W.G. Halsey, N.W. Brown, J.J. Sienicki, A. Moisseytsev, D.C. Wade, analysis in single-phase natural circulation, Ann. Nucl. Energy 31 (10) (2004)
SSTAR: The US-lead-cooled fast reactor (LFR), J. Nucl. Mater. 376 (2008) 255– 1177–1198.
259. [33] C.J. Ho, S.P. Chiou, C.S. Hu, Heat transfer characteristic of a rectangular natural
[4] V.I. Kostin, O.B. Samoilov, V.N. Vavilkin, Yu.K. Panov, A.V. Kurachenkov, M.A. circulation loop containing water near its density extreme, Int. J. Heat Mass
Bolshukhin, V.I. Alekeev, I.V. Shmelev, Yu.D. Baranaev, A.A. Pekanov, Small Transfer 40 (15) (1997) 3553–3558.
floating nuclear power plants with ABV reactors for electric power generation, [34] J.P. Holman, J.H. Boggs, Heat transfer to Freon 12 near the critical state in a
heat production and seawater desalination, in: Fifteenth Annual Conference of natural circulation loop, J. Heat Transfer 82 (1960) 221–226.
the Indian Nuclear Society, INSAC-2004, November 15–17, Mumbai, India. [35] M. Misale, L. Tagaliafico, G. Tanda, Experiments in a free convection
[5] International Atomic Energy Agency, Design Features to Achieve Defence in rectangular loop, in: Proceedings of the 4th International Symposium on
Depth in Small and Medium Sized Reactors, IAEA Nuclear Energy Series, No. transport Phenomena in Heat and Mass Transfer, Sydney, Australia, July 1991,
NP-T-2.2, Annex IX, Vienna, 2009, pp. 228-245. pp. 203–211.
[6] P.K. Vijayan, Experimental observations on the general trends of the steady [36] W. Ambrosini, N. Forgione, F. Oriolo, E. Semeraro, M. Tarantino, Experimental
state and stability behavior of single phase natural circulation loops, Nucl. Eng. study on natural circulation and air-injection enhanced circulation with
Des. 215 (2002) 139–152. different fluids, J. Eng. Gas Turbine Power 131 (2009). 032902-1–032902-7.
[7] James J. Sienicki, Bruce W. Spencer, Power optimization in the STAR-LM [37] P.K. Vijayan, M. Sharma, D. Saha, Steady state and stability characteristics of
modular convection reactor system, in: Proceedings of ICONE 10, Tenth single phase natural circulation in a rectangular loop with different heater and
International Conference on Nuclear Engineering, Arlington, Virginia, USA, cooler orientation, Exp. Thermal Fluid Sci. 31 (2007) 925–945.
April 14–18, 2002. [38] W. Ambrosini, N. Forgione, J.C. Ferreri, M. Bucci, The effect of wall friction in
[8] Dulera, A. Basak, R.K. Sinha, Options for Design of 600MWth Indian High single-phase natural circulation stability at the transition between laminar
Temperature Reactor for Hydrogen Production INSAC-2005, BARC, Mumbai, and turbulent flow, Ann. Nucl. Energy 31 (2004) 1833–1865.
2005. [39] P. Garibaldi, Single phase natural circulation loops: effects of geometry and
[9] M. Misale, P. Garibaldi, Dynamic behavior of a rectangular single phase natural heat sink temperature on dynamic behavior and stability, Ph.D. Thesis.
circulation loop: influence of loop inclination, in: Proceedings of 20th National [40] M. Misale, P. Ruffino, M. Frogheri. The influence of the wall thermal capacity
and 9th International ISHMT-ASME Heat and Mass Transfer Conference, 4–6 and axial conduction over a single-phase natural circulation loop: 2D
January, 2010, Mumbai, India. numerical study Eurotherm seminar no. 63, Single and two-phase natural
[10] W. Ambrosini, N. Forgione, J.C. Ferreri, M. Bucci, The effect of wall friction in circulation, 1999, pp. 177–186
single phase natural circulation stability at the transition between laminar and
turbulent flow, Ann. Nucl. Energy 31 (2004) 1833–1865.