Opertional Modes Two Phase Flow
Opertional Modes Two Phase Flow
Opertional Modes Two Phase Flow
www.elsevier.com/locate/ijhmt
Abstract
The system behaviour of an open single tube, steam heated natural circulation evaporator has been investigated at heating steam pres-
sures up to 0.44 MPa and different subcoolings of the process medium water. In the studied range of parameters, diverse operational
states occurred which have been recognized for the first time as geysering coupled with manometer oscillation as well as first, second
and third orders of density wave oscillation of type I. Due to the strong nonlinear behaviour of the system, up to three states appear
in certain ranges of parameter constellations. By defined setting of initial conditions and distinct perturbations of the system, the indi-
vidual states can be set.
2006 Elsevier Ltd. All rights reserved.
Keywords: Natural circulation evaporator; Multiple modes; Geysering; Density wave oscillation
0017-9310/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2005.10.046
A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314 2305
Nomenclature
Greek symbols
k wave length of enthalpy oscillation, m
# temperature, C
inert gas, Jiang et al. [18] could maintain a stable single- evaporator. Both modes feature an oscillating void fraction
phase mode over a certain range of subcooling during at the outlet of the heat exchanger. Due to the nearly zero
start-up. Hence, low quality DWO was suppressed. Besides steam quality/low void fraction, gravitational pressure
the mentioned parameters of system pressure and subcool- drop governs the system in case of type I, while frictional
ing, the concrete behaviour during start-up strongly pressure drop dominates type II. Yadigaroglu and Bergles
depends on heat flux, geometry of the apparatus and the [23] reported as the only ones on higher mode density wave
physical properties of the process medium. oscillations which appeared at high subcoolings and low
In regard to the operational states mentioned above, heat fluxes. They define the order of the oscillation as
geysering/flashing has been investigated systematically for follows:
the first time by Griffith [19] in closed vertical liquid filled
n ¼ zbb =k. ð1Þ
lines. In the following, different types of geysering have
been recognized in forced and natural circulation evapora- The value zbb represents the length of the single-phase re-
tors such as geysering coupled with manometer oscillations gion and k the wavelength of the enthalpy oscillations in
[20], geysering induced by condensation [15,16,21] or gey- the single-phase region. The enthalpy oscillations result
sering similar to that occurring in closed vertical liquid from the oscillating residence time of the individual liquid
filled lines [17]. Flashing has also been investigated by elements in the single-phase part of the heat exchanger.
[14,16]. All works show the typical sequence: (1) incuba- After passing the boiling boundary, they are transferred
tion, (2) vapour generation and expulsion, (3) and refilling into void oscillations. For k > zbb Yadigaroglu et al. call
described by [12]. the mode ‘‘fundamental’’ or ‘‘zero order’’; they observe a
The phenomenon of density wave oscillation is well maximum order of six.
known in literature. A comprehensive description of this Due to the nonlinear behaviour of natural circulation
mode is given by [12]. Fukuda and Kobori [22] distinguish evaporators, multiple operational states can appear at
between oscillations with nearly zero steam quality (type I) one parameter constellation. This has been shown theoret-
and relatively high steam quality (type II) at the exit of the ically by Knaani et al. [24], who explain this bifurcation
2306 A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314
708
gle tube natural circulation evaporator the coexistence of (4)
(1)
973
H
DWO of different orders at transition points. These are LIR
not stable and change arbitrarily. Chen et al. [9] studied
LIR
experimentally the effect of power changing procedure on
the thermohydraulic behaviour of a double channel natural TIR
circulation loop. They found a hysteresis effect at the onset
of two-phase flow in which, depending on the power chang-
(5) TIRC
ing procedure, single-phase flow or an oscillating opera-
tional state arises. Steam
(3)
An evaluation of literature shows that hysteresis effects
on the field of two-phase flow and boiling heat transfer
2270
include flooding transition [25] or the onset of nucleate To expansion tank
2215
boiling [26]. However, experimental works, studying multi- PIRC
66
66
190
2.1. Apparatus a unique static pressure exists in all tanks at the liquid
surfaces.
The test facility consists of a U-tube connected to a This apparatus constellation satisfies the condition for
supply tank (1) and steam separator (4) at its open ends natural circulation mode if the liquid charging level H in
(see Fig. 1). One side of the U-tube is steam heated (3). the supply tank is greater than or equal to zero. Here,
To realise both, proper adjustable temperatures of the pro- the coordinate H starts at the height of the overfall in the
cess medium and a low pressure drop in the feeding line of steam separator and grows with decreasing liquid level in
the heat exchanger, the supply tank and the steam separa- the supply tank, Fig. 1. The pressure drop coefficient of
tor are not connected to each other directly. Hence, liquid the feeding line can be set by a ball valve (2). The steam
is conveyed from the expansion tank by a radial pump heated tube in the evaporator with a heated length of 2 m
through a pipe in pipe cooler into the supply tank (1). and a wall thickness of 2 mm features an electro polished
The coolant flow rate of the heat exchanger is controlled; inner wall. The heating steam pressure is controlled. Before
therefore, the liquid enters the supply tank with a defined entering the shell side of the evaporator, a steam dryer sep-
temperature. One part of the liquid leaves this tank by an arates the liquid from the vapour phase of the heating
adjustable overfall back into the expansion tank in order steam. After condensation, the liquid leaves the evaporator
to maintain a constant liquid level in the chamber. The by a ball float trap and a condensate cooler to avoid evap-
remaining part enters the U-tube. In the steam separator oration as a result of pressure release. In all tanks, the
(4), the liquid phase returns to the expansion tank by an evaporator and its feeding line feature inspection glasses
overfall situated in the wall. The vapour leaves the separa- for observation of two-phase flow pattern and the shape
tor by a second tube for the expansion tank, from where it of liquid surfaces respectively. The tubes and tanks consist
disappears into the environment. The head spaces of all of stainless steel. Insulation of the apparatus with glass
three tanks are connected to each other by tubes. Hence, fibre minimises heat removal.
A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314 2307
2.2. Data acquisition limited temporal changes of the height of liquid level in the
feeding tank and the pressure drop coefficient in the feeding
Platinum thermometers (PT 100) in 4-wire technique in line. This is achieved by either, closing valve (5) or by clos-
casing pipes (diameter 2 mm, wall thickness 0.2 mm) serve ing valve (2) in Fig. 1, respectively.
for temperature measurements at both the inlet and the The periodic time of transient operational modes is cal-
outlet of the evaporator (TIR in Fig. 1). The sensors have culated by applying fast Fourier transformation on the vol-
been calibrated together with the transmitters. The volume ume flux/time curve delivered by the magnetic inductive
flux in the U-tube (FIR) is determined by a magnetic induc- flowmeter in the feeding line. Due to the strong nonlinear
tive flowmeter (Endress and Hauser, Promag 33 H) and the behaviour of the system this method is only used to obtain
height of liquid levels in the supply tank and steam separa- the fundamental frequency.
tor (LIR) by differential pressure transmitter (Vega, Vega-
dif 34). A pressure transmitter by Haenni (ED 510) detects 3. Results
the pressure of heating steam (PIRC). The flow rate of the
heating steam condensate is determined by a simple For convenience, the results are partly interpreted by
method of temperature and time measurement the liquid using dimensionless numbers. This is not a complete set
requires filling a defined volume. A PLC (programmable of numbers, but adequate to present the results more
logic controller) by Texas Instruments (Typ 545) controls clearly. The time t and the periodic time T are normalized
heating steam pressure, feeding temperature of the U-tube by LH/vc, where LH represents the lengthpofffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
the heatedffi tube
and flow rate of the liquid entering the supply tank. For and vc the characteristic velocity vc ¼ gðH þ d=4Þ. The
data acquisition, a board by National Instruments (PCI- latter value is characteristic for the velocity, the liquid
MIO-16E-4) and LabView is used in combination with a would flow from the steam separator to the feeding tank.
personal computer. The parameter H denotes the height of liquid in the feeding
tank and d the diameter of the overfall in the steam separa-
2.3. Experimental procedure/data analysis tor. The term d/4 takes in consideration, that the maximum
liquid level in the steam separator is always increased due
The experiments start by heating up the system and set- to the operational behaviour of the overfall. The velocity
ting the following parameters: temperature of liquid in the at the inlet of the evaporator vI is divided by the character-
feeding line #F, heating steam pressure p, liquid charging istic velocity vc to obtain the normalized mass flux m_ .
level H in the supply tank, pressure drop coefficient fI of Here, identical densities of the origin and referred mass flux
feeding line and the flow rate through the cooler. All exper- are supposed. The subcooling at inlet (I) and exit (E) of the
iments have been carried out with water as the working heat exchanger are referred to (#b #F), the temperature
fluid at ambient pressure. The parameters and the investi- difference between boiling temperature at the exit of the
gated range are listed in Table 1. The pressure drop coeffi- heated tube and the temperature of the feeding liquid. If
cient is obtained experimentally and referred to the specific superheating is not taken into consideration, this repre-
kinetic energy of the liquid at the inlet velocity of the sents the maximum appearing temperature difference
heated channel. within the process medium. In the following, the ratio
After setting the parameters, the temporal progressions NBo/NSub is used as a non-dimensional parameter. The
of flow rate and temperatures are observed. If no temporal boiling number NBo describes the ration of added heat flow
change of the behaviour appears within eight minutes and Q_ to the process medium and the heat flow required for
if #F shows at most a maximum deviation of ±0.5 K from evaporating the process medium. The value Q_ is calculated
the scheduled value, data are recorded every 0.1 s for from the condensate mass flow and corrected by the exper-
8 min. During each measurement, the experimenter imentally obtained rate of heat loss. The mass flow in the
observes the two-phase flow pattern in the steam separator denominator is formed with the characteristic velocity vc.
leaving the evaporator. The repeatability of measurements In comparison to the boiling number, the subcooling num-
is investigated by repetitions at selected parameter constel- ber NSub contains in the numerator the heat flow required
lations carried out at different days. To investigate the sta- to shift the process medium from subcooled to saturated
bility of every operational mode, the system is disturbed by state.
Chiang et al. [27] apply a similar but dimensional value
which they call modified heat flux. It contains the quotient
Table 1
heat flux to subcooling of the process medium in the feed-
Parameter settings
ing line of the evaporator. They plot the time averaged
Heating steam pressure, p 0.12–0.44 MPa
velocity at the inlet of the evaporator versus the modified
Temperature feeding liquid, #F 61; 71; 81; 86 C
Liquid charging level, H 25; 50; 75; 100 mm heat flux for an oscillating operational state in a natural
Pressure drop coefficient inlet heat exchanger, fI 0.4–72 circulation evaporator. Here, a unique curve appears, inde-
System pressure Ambient pressure pendent from the inlet subcooling.
Volume flux cooler 4, 8 · 103 m3/s Fig. 2 shows the periodic time of the investigated oper-
Working fluid Deionised water
ational states versus heating steam pressure for a tempera-
2308 A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314
2 z* 0.5
m*
. 1 0
*
z
0 m*
.
-0.5
-1 -1
0 5 10 15 20
τ
1
0.8 ΔϑI Sub
Δϑ*I Sub , ΔϑE*Sub
0.6
0.4
0.2 ΔϑE Sub
0
-0.2
0 5 10 15 20
τ
Fig. 3. Mass flux m_ , liquid level in the steam separator z*, subcooling at the inlet D#I Sub , and exit D#E Sub of the heated tube versus time s in dimensionless
form for geysering (#F = 71 C, p = 0.24 MPa, H = 75 mm, fI = 0.4).
A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314 2309
where line can exceed one to the twofold of the heated tube. Dur-
X AS ing vapour generation and expulsion, the pressure in the
L¼ Li . liquid decreases due to change of geodetic height, density
i
Ai
and acceleration of the fluid. The superheated liquid in
Here, Li and Ai represent the lengths and the cross sectional the feeding line contributes to an additional generation of
areas of all liquid filled tubes/tanks of the U-tube. The vapour volume. This leads to an increased intensity of
comparison of the analytical solution with measured values the geyser in comparison to pure geysering. Due to flash-
shows a maximum deviation of ±3.6%. ing, the maximum mass fluxes, 20–60 times higher than
In the heated case, the manometer oscillation can be rec- the time averaged value, can be observed.
ognized easily during refilling and incubation. At the begin- The geometry of the experimental set-up influences
ning of the refilling period, liquid flows from the feeding strongly its operational behaviour. Therefore, the appear-
tank towards the steam separator while the liquid level in ance of geysering with manometer oscillation has been
the steam separator is rising. A partial change from kinetic rarely mentioned in literature. In comparison to plants
to potential energy takes place. The remaining part of described in literature [15,17,22,28], this facility features
energy leaves the U-tube system with the liquid at the over- extreme low pressure drop coefficients of the feeding line
fall in the steam separator and/or is converted to internal (fI = 0.4) and an extremely high ratio of cross sectional area
energy of the fluid by friction. At the beginning of incuba- steam separator to heated tube (AS/AH = 81). This enables
tion, the liquid velocity in the U-tube is equal to zero and the appearance of weakly damped manometer oscillations
the liquid level in the steam separator is higher than in the with time periods greater than that of geysering.
feeding tank. Therefore, an acceleration of the liquid from
steam separator to feeding tank and a partial change from 3.2. Density wave oscillations type I
potential energy to kinetic energy occurs. The onset of the
geyser stops the conversion. Hence, this can be described as Fig. 4 depicts typical curves of mass flux and subcooling
a forced oscillation – the periodically appearing geyser at the exit of the heated channel versus time for DWO type
feeds kinetic energy to the system to maintain the damped I, first, second and third order.
manometer oscillation. All curves are characterized by a periodic, sinusoidal
In the investigated parameter range, the periodic time of behaviour with almost constant amplitudes, similar to
the isothermal manometer oscillations TU is greater than velocity measurements by Saha et al. [29]. In comparison
that of the geyser T. Therefore, no complete period of to geysering, this mode shows a less violent behaviour with
the manometer oscillation arises (Fig. 3). The periodic time much smaller amplitudes of mass fluxes and subcoolings.
T decreases with rising heating steam pressure. This results The flow pattern at the outlet of the heated tube exhibit
from the increased heat flow and leads to shorter incuba- bubbly and partially slug flow. The mechanism of density
tion periods. Geysering also exists without manometer wave oscillations is sufficiently described in literature,
oscillations [17]. In the case of pure geysering, the liquid see [12] for a review, and therefore not mentioned here.
remains almost static in the heated channel during incuba- The classification of the observed operational states to
tion. Hence, the heat transfer is dominated by conduction DWO type I of different orders is based upon two charac-
in this phase. With manometer oscillation, transient flow teristics:
and therefore convective heat transfer arises in the evapo-
rator for time periods: (1) Thermodynamic state of the fluid at the outlet of the
evaporator.
(1) T > TU with alternating flow direction, observed by (2) Relation between time period of oscillation and tran-
[20] and for sit time of an enthalpy perturbation in the liquid filled
(2) T < TU with uniform negative flow direction, pre- part of the heated channel.
sented in this work.
Measurements of temperature at the outlet of the evap-
During incubation, liquid in riser (T > TU) and feeding orator reveal for first and second order DWO, a time aver-
line (T > TU, T < TU) is heated up in addition to liquid aged subcooled fluid (Fig. 4). Hence, these operational
in the heated channel. The residence time distribution of states are characterized by a low steam quality in the evap-
liquid in the heated channel leads to altered radial but espe- orator. Furthermore, the time averaged mass flux of all
cially axial temperature profiles in the U-tube in compari- orders increases with rising difference between heating
son to the static case. This influences time and intensity steam temperature and boiling temperature (#HS #b)
of the succeeding geyser: As result of the backflow during respectively with growing NBo/NSub (Fig. 5).
incubation, the liquid is heated up from outlet to inlet of The non-dimensional numbers vary in the range of one,
the heated tube. The negative values of the subcooling at which confirms that the characteristic values are of relevant
the entry of the heated tube indicate that superheated order of magnitude. The data points show a nearly linear
liquid (with respect to ambient pressure) is stored in the behaviour with NBo/NSub. The extrapolation of NBo+/
feeding line. The volume of preheated liquid in the feeding NSub ! 0 leads to the physically correct result m_ ! 0.
2310 A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314
3 st 1
1 order DWO
2 ΔϑE Sub 0.5
1 0
m*
Δϑ*E Sub
.
.
0 m* -0.5
-1 -1
0 5 10 15 20
τ
3 nd 1
2 order DWO
2 0.5
ΔϑE Sub
ΔϑE* Sub
m*
.
1 0
0 .
m* -0.5
-1 -1
0 5 10 15 20
τ
3 rd 1
3 order DWO
2 0.5
ΔϑE Sub
ΔϑE* Sub
1 0
m*
0 .* -0.5
m
-1 -1
0 5 10 15 20
τ
Fig. 4. Mass flux m_ , and subcooling at the exit D#E Sub of the heated section versus time s in dimensionless form for density wave oscillation first order
(#F = 71 C, p = 0.40 MPa, H = 75 mm, fI = 0.4), second order (#F = 71 C, p = 0.24 MPa, H = 75 mm, fI = 0.4) and third order (#F = 71 C,
p = 0.24 MPa, H = 75 mm, fI = 0.4).
61 75 . 61 75
300 71 25
0.4 71 25
71 50 71 50
.
200 71 75 71 75
71 100 0.2 71 100
81 75 81 75
100 86 75 86 75
0 0
0 20 40 60 80 0 0.2 0.4 0.6 0.8
(ϑHS-ϑb) / K NBo / NSub
Fig. 5. Time averaged mass flux versus difference between heating steam temperature and boiling temperature (#HS #b) and NBo/NSub for first to third
order DWO.
The course of mass flux versus NBo/NSub and (#HS #b) number of periodic times a fluid element needs to pass
indicates that phase change induced gravitational pressure the liquid filled length zbb of the heated channel. To receive
drop governs the flow behaviour in the evaporator. zbb, thermodynamic equilibrium of the fluid and a constant
Fukuda and Kobori [22] recognize this as a characteristic heat flux along the heated channel are assumed. The appli-
of DWO type I. cation of energy equation leads to
The fulfilment of the second characteristic can be shown
by applying the relation by Yadigaroglu and Bergles [23] to vI AH qF cp ð#b #F Þ
zbb ¼ LH ; ð3Þ
the experimental data (Eq. (1)). The value n describes the Q_
A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314 2311
where AH describes the cross sectional area, LH the length 15 ϑF /°C H/mm
of the heated tube, qF the density, TF the temperature of 61
71
75
25
the feeding liquid, Q_ the heat flow and #b the boiling tem- Geysering
71
71
50
75
10
perature. The wave length of the enthalpy oscillations is 71
81
100
75
T*
obtained from n = 1 (calculation)
n = 2 (calculation)
n = 3 (calculation)
k ¼ vI T ; ð4Þ 5 st
1 order DWO
+/- 0.2 3 1
3
2 0.5
ΔϑE Sub
Δϑ*E Sub
+/- 0.13
m*
n
.
2 1 0
.*
+/- 0.06 0 m -0.5
1
-1 -1
0 5 10 15 20
0 τ
0 0.2 0.4 0.6 0.8
NBo / NSub Fig. 8. Mass flux m_ , and subcooling D#E Sub at the exit of the heated
section versus time s for transition of DWO between first and second order
Fig. 6. Order n versus NBo/NSub for fI = 0.4. (#F = 71 C, p = 0.32 MPa, H = 75 mm, fI = 0.4).
2312 A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314
0.4
0.3 nd
2 order DWO
mFC
0.2
*
NBo/NSub = 0.16
0.1
p = 0.26 MPa
0 -1 0 1 2
10 10 T* 10 10
0.4
0.3 2nd order DWO
mFC
0.2
*
st
1 order DWO
NBo/NSub = 0.17
0.1
p = 0.28 MPa
0 -1
10 100 T* 101 102
0.4
0.3 nd
2 order DWO
mFC
0.2
*
st
1 order DWO
NBo/NSub = 0.19
0.1
p = 0.30 MPa
0 -1
10 100 T
* 101 102
0.4
0.3
mFC
0.2
*
st
2nd order DWO 1 order DWO
NBo/NSub = 0.20
0.1
p = 0.32 MPa
0 -1
10 100 T
* 101 102
0.4
1st order DWO
0.3
mFC
*
0.2
NBo/NSub = 0.22
0.1
p = 0.34 MPa
0
10-1 100 T* 101 102
Fig. 9. Spectra of FFT analysis from the mass flux (#F = 71 C, H = 75 mm, fI = 0.4).
pressure drop coefficient fI in the feeding line or (b) short- from geysering to second order DWO due to short-time
time rise in liquid charging level H in the feeding tank. For throttling results from the change of kinetic to internal
this, two ball valves (see (2) and (5) in Fig. 1) are available. energy. The influence of the increased value of fI can be
The first one serves for throttling, the second one to recognized especially during vapour generation and expul-
heighten the liquid level in the feeding tank. sion. In this phase, the highest flow velocities and kinetic
Starting with geysering at a parameter constellation energies of the fluid appear. Throttling interrupts the cycle
where all three operational states can occur, a temporary of periodic conversion from kinetic to potential energy by
increase in the pressure drop coefficient leads to second change of kinetic energy to internal energy. The lack of
and third order DWO (Fig. 11). In this diagram the ampli- potential energy at the beginning of incubation prevents
tude of the mass flux is plotted versus the pressure drop backflow in the heat exchanger. This leads to higher resi-
coefficient fI. The transition of modes takes place at a crit- dence times of liquid elements in the heated tube. Further-
ical value of fI. By putting the pressure drop coefficient more, almost no preheated liquid is stored in the feeding
back to the initial value, the current mode remains. The line. Hence, vapour generation starts earlier but less inten-
period of perturbation can range from a few seconds to sive. The kinetic energy during vapour generation is not
several minutes to set the new operational state. sufficient to fan geysering. The mode then changes to sec-
The rise in liquid charging level H in the feeding tank ond order DWO.
acts conversely to throttling. Depending on the height, a In the case of DWO, an increase in the pressure drop
transition to a mode of higher periodic time can be coefficient leads to reduced time averaged mass fluxes in
observed. A sequence of changes of modes after short-time the system, and therefore to a displacement of the enthalpy
perturbations is displayed in Fig. 12. Here, the mass flux in nodes in the liquid filled part of the heated tube towards
the U-tube and the height of liquid level H in the feeding the entry. During a sufficiently high decrease in mass flux,
tank are plotted versus time. Beginning with third order a new node is formed at the boiling boundary, which
DWO as the initial state, a temporal limited rise in H pro- results in a change from second to third order DWO.
vokes the transition to second order DWO and a further The rise in liquid level in the feeding tank acts conversely
increase to geysering. The initial state is restored by to throttling, whereby enthalpy nodes move towards the
short-time throttling. boiling boundary (transition third to second order DWO)
The change in operational states by perturbation can be respectively potential energy is fed into the system (transi-
explained by the impact of the specific perturbation in com- tion second order DWO – geysering). Starting with third
bination with the mechanism of the mode. The transition order DWO, a change to second order DWO or geysering
takes place depending on the perturbation of the liquid
charging level in the feeding tank.
4 With the knowledge of the effect of temporary perturba-
Geysering
tions on the system, it is possible to carry out defined
3 changes between the operational states, presented in
Fig. 12.
mAmp
.*
2
2nd order DWO 4. Conclusion
1
3rd order DWO The operational behaviour of an open single tube, steam
steady mode
0
heated, natural circulation evaporator has been investi-
0 1 2
10 10 10 gated. At subcoolings of 14–39 K and heating steam pres-
ζI
sures of 0.12–0.44 MPa, different operational modes
Fig. 11. Amplitude of mass flow m_ Amp versus pressure drop coefficient fI appear. With rising heating steam pressure and decreasing
(p = 0.25 MPa, #F = 71 C, H = 75 mm). subcooling, geysering coupled with manometer oscilla-
Fig. 12. Defined change between second, third order DWO, and geysering by short-time perturbation of the system (p = 0.25 MPa, #F = 71 C,
H = 75 mm, fI = 0.4).
2314 A. Baars, A. Delgado / International Journal of Heat and Mass Transfer 49 (2006) 2304–2314
tions, third order, second order and first order of density [9] W.L. Chen, S.B. Wang, S.S. Twu, C.R. Chung, C. Pan, Hysteresis
wave oscillations type I arise. All modes show a periodic effect in a double channel natural circulation loop, Int. J. Multiphase
Flow 27 (2001) 171–187.
behaviour and have been rarely described in literature up [10] J.M. Kim, S.Y. Lee, Experimental observation of flow instability in a
to now. The first operational state shows the typical semi-closed two-phase natural circulation loop, Nucl. Eng. Des. 196
sequence of geysering: (a) incubation, (b) vapour genera- (2000) 359–367.
tion/expulsion and (c) refilling. Due to the superposed [11] J.A. Bouré, A.E. Bergles, L.S. Tong, Review of two-phase flow
manometer oscillations, backflow occurs in the heated tube instability, Nucl. Eng. Des. 25 (1973) 165–192.
[12] G. Yadigaroglu, Two-phase flow instabilities and propagation
during incubation. Hence, saturated liquid is stored in the phenomena, in: J.M. Delhaye, M. Giot, M.L. Riethmuller (Eds.),
feeding line. During vapour generation, the stored liquid Thermohydraulics of Two-phase Systems for Industrial Design and
intensifies the geyser by flashing which leads to consider- Nuclear Engineering, McGraw-Hill, New York, 1981, pp. 353–403.
able vibrations of the plant. The remaining modes show a [13] G. Yun, G.H. Su, J.Q. Wang, W.X. Tian, S.Z. Qiu, D.N. Jia, J.W.
strong correlation between periodic time of the oscillation Zhang, Two-phase instability analysis in natural circulation loops of
China advanced research reactor, Ann. Nucl. Energy 32 (2005) 379–
and travelling time of enthalpy perturbations in the liquid 397.
phase of the heated tube. Furthermore, low vapour quality [14] A. Manera, T.H.J.J. van der Hagen, Stability of natural-circulation-
could be observed at the exit of the heat exchanger, which cooled boiling water reactors during startup: experimental results,
are characteristics of type I density wave oscillations. Nucl. Technol. 143 (2003) 77–88.
By calculating the order of the DWO (ratio between [15] M.H. Subki, M. Aritomi, N. Watanabe, H. Kikura, T. Iwamura,
Transport mechanism of thermohydraulic instability in natural
length of single-phase region and wavelength of the enthalpy circulation boiling water reactors during startup, J. Nucl. Sci.
oscillations in the single-phase region) from experimental Technol. 40 (2003) 918–931.
data, the oscillations can be divided into first, second and [16] S.Y. Jiang, Y.J. Zhang, X.X. Wu, J.H. Bo, H.J. Jia, Flow excursion
third order. The periodic time of DWO behaves proportion- phenomenon and its mechanism in natural circulation, Nucl. Eng.
ally to subcooling number and reciprocally proportional to Des. 202 (2000) 17–26.
[17] I.K. Kyung, S.Y. Lee, Experimental observations on flow character-
boiling number and order. istics in an open two-phase natural circulation loop, Nucl. Eng. Des.
Due to the strong nonlinear behaviour of natural circu- 159 (1994) 163–176.
lation evaporators, up to three modes (geysering coupled [18] S.Y. Jiang, M.S. Yao, J.H. Bo, S.R. Wu, Experimental simulation
with manometer oscillations, second and third order study on start-up of the 5 MW nuclear heating reactor, Nucl. Eng.
DWO) appear at one parameter constellation. By adopting Des. 158 (1995) 111–123.
[19] P. Griffith, Geysering in liquid-filled lines, ASME Paper 62-HT-39,
a defined setting of the boundary conditions or short-time 1962.
perturbation of the system, the different modes can be set. [20] B.A. Hands, The flow stability of a liquid-nitrogen thermosyphon
with 8 mm bore riser, AIChE Symp. Ser. No. 189, vol. 75, 1979,
pp. 177–184.
References [21] M. Aritomi, J.H. Chiang, T. Nakahashi, M. Wataru, M. Mori,
Fundamental study on thermo-hydraulics during start-up in natural
[1] K. Hibi, H. Ono, T. Kanagawa, Integrated modular water reactor circulation boiling water reactors (I), Nucl. Sci. Technol. 29 (1992)
(IMR) design, Nucl. Eng. Des. 230 (2004) 153–266. 631–641.
[2] A.K. Nayak, P.K. Vijayan, D. Saha, V. Venkat Raj, M. Aritomi, [22] K. Fukuda, T. Kobori, Two-phase flow instability in parallel
Study on the stability behaviour of a natural circulation pressure tube channels, in: Sixth International Heat Transfer Conference, 1978,
type boiling water reactor, Nucl. Eng. Des. 215 (2002) 127–137. FB-17, pp. 369–374.
[3] J.H. Hills, W.E. Jones, A.K. Ibrahim, The behaviour of a pilot-scale [23] G. Yadigaroglu, A.E. Bergles, Fundamental and higher-mode den-
horizontal thermosyphon reboiler, Trans. IChemE, Part A 75 (1997) sity-wave oscillations in two-phase flow, J. Heat Transfer Trans.
652–656. ASME 94 (1972) 189–195.
[4] A.W. Sloley, Properly design thermosyphon reboilers, Chem. Eng. [24] A. Knaani, Y. Zvirin, Bifurcation phenomena in two-phase natural
Progr. 3 (1997) 52–64. circulation, Int. J. Multiphase Flow 19 (1993) 1129–1151.
[5] A. Baars, F. Werner, A. Delgado, Investigations on the maximum [25] S.G. Bankoff, S.C. Lee, Flooding and hysteresis effects in nearly
inner wall temperature in a steam heated tube with natural circulation horizontal countercurrent stratified steam-water flow, Int. J. Heat
for the use in biotechnology, in: B.J. Grochal, J. Mikielewicz, B. Mass Transfer 30 (1987) 581–588.
Sunden (Eds.), Proceedings of the Third Baltic Heat Transfer [26] H. Bräuer, F. Mayinger, Onset of nucleate boiling and hysteresis
Conference, IFFM Publishers, Gdansk, 1999, pp. 509–514. effects under forced convection and pool boiling, in: Proceedings of
[6] N. Abe, S. Yokobori, H. Nagasaka, S. Tsunoyama, Two-phase flow the Engineering Foundation Conference on Pool and External
natural circulation characteristics inside BWR vessels, Nucl. Eng. Boiling, Santa Barbara, 1992, pp. 15–36.
Des. 146 (1994) 253–265. [27] J.H. Chiang, M. Aritomi, M. Mori, Fundamental study on thermo-
[7] D. Brück, Einfluss mechanischer Belastungen auf dispergierte, emp- hydraulics during start-up in natural circulation boiling water
findliche Inhaltsstoffe von Flüssigkeiten in Leitungssystemen und reactors (II), J. Nucl. Sci. Technol. 30 (1993) 203–211.
Behälterströmungen, PhD thesis, Technische Universität München, [28] V.K. Chexal, A.E. Bergles, Two-phase instabilities in a low pressure
1997. natural circulation loop, AIChE Symp. Ser. 131 (1973) 37–54.
[8] C. Schuster, A. Ellinger, J. Knorr, Analysis of flow instabilities at the [29] P. Saha, M. Ishii, N. Zuber, An experimental investigation of the
natural circulation loop DANTON with regard to non-linear effects, thermally induced flow oscillations in two-phase systems, J. Heat
Heat Mass Transfer 36 (2000) 557–565. Transfer, Trans. ASME 98 (1976) 616–622.