- Email: [email protected]
Liquid film dynamics of two-phase annular flow in square and tight lattice subchannels
Nuclear Engineering and Design 300 (2016) 467–474
Contents lists available at ScienceDirect
Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes
Liquid film dynamics of two-phase annular flow in square and tight lattice subchannels Daisuke Ito a,∗ , Petros Papadopoulos b , Horst-Michael Prasser b a b
Research Reactor Institute, Kyoto University, Osaka, Japan Department of Mechanical and Process Engineering, ETH Zürich, Zürich, Switzerland
h i g h l i g h t s • • • •
A pair of liquid film sensors was applied to flow measurements in a double subchannel geometry. Film thickness distributions in square and tight lattice subchannels were measured. The characteristics of the water film flow were compared between each subchannel. Correlation feature of film thickness fluctuations in a subchannel gap were studied.
a r t i c l e
i n f o
Article history: Received 29 June 2015 Received in revised form 17 February 2016 Accepted 20 February 2016 Available online 5 March 2016 Classification: K. Thermal hydraulics
a b s t r a c t The flow behavior of a liquid film on the wall in square-lattice and tight-lattice double subchannels was studied by using a pair of flexible liquid film sensors based on the electrical conductance method. The liquid film sensor has a network of electrodes on the surface and the electrical conductance between transmitter and receiver electrodes is detected by using a high-speed wire-mesh measurement system. The sensors were fixed on rod simulators with 20 mm diameter and the liquid film thickness distribution was estimated from the measured conductance data array. With this method, it is possible to observe the flow structure of the liquid film with a high sampling speed up to 10 kHz and a spatial resolution of 2 mm. In the present study, the spatial-temporal distributions of liquid film thickness in the annular flow in a square-lattice and a tight-lattice pair of adjacent subchannels were measured by using two liquid film sensors installed on the surface of opposing rods. The traveling of disturbance waves was observed in the time-series of two-dimensional film thickness distributions. The interaction of the liquid films on the opposing rods was investigated by correlating the signals of both sensors. As a result, the effect of the channel geometry on the averaged film thickness profiles and the correlation characteristics was analyzed for the first time. © 2016 Elsevier B.V. All rights reserved.
1. Introduction A stable energy source has to be ensured for a sustainable energy supply in the future. Nuclear energy has been considered as an option to cover the base load of the electricity supply. However, the management and effective utilization of plutonium is very important in terms of sustainability and issues of non-proliferation. High conversion boiling water reactors (HCBWR) are potential candidates for such an application as well as fast breeder reactors (Iwamura et al., 2006). The HCBWR can achieve a high conversion
∗ Corresponding author. Tel.: +81 72 451 2373. E-mail address: [email protected] (D. Ito). http://dx.doi.org/10.1016/j.nucengdes.2016.02.034 0029-5493/© 2016 Elsevier B.V. All rights reserved.
ratio by reducing the volume of water that is acting as coolant and moderator in the same time. For this purpose, a tight-lattice rod bundle with a narrow gap between the fuel rods is adopted. The main advantage of the HCBWR over Gen IV reactor concepts consists in the direct use of the well-established technology of light water reactors for the improvement of fuel sustainability. In order to improve and assess the safety of the HCBWR, experimental data is needed to characterize the two-phase flow in the narrow subchannels of tight-lattice rod bundles. A number of studies on thermal-hydraulics in the tight-lattice rod bundle have been conducted by the Japan Atomic Energy Agency (JAEA). Kureta (2007) measured void fraction distributions in a rod bundle using neutron radiography. Tamai et al. (2004, 2006) studied the heat transfer characteristics in a 37-rod bundle. Furthermore, several
468
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
numerical studies were also carried out to evaluate the two-phase flow phenomena in the tight-lattice configuration (Yoshida et al., 2008; Zhang et al., 2008). For the validation of high fidelity simulations using an interfacial tracking technique, an experimental database on the spatio-temporal characteristics of the flow structure is needed. Of particular interests are the bubble behavior at low void fraction conditions and the dynamics of the liquid film in the annular flow regime. Two-phase flow experiments with high-resolution instrumentation are not only interesting for tight-lattice geometries, but can contribute to a better understanding and modeling of traditional square-lattice rod bundles of the existing light water reactors. Recently, several measurement tools have been developed for such experiments in the rod bundle. There is considerable progress in the field of imaging by X-rays, gamma radiation and neutrons. The common main advantage is the non-intrusiveness of the measurement. Neutron radiography is an excellent tool for the visualization of a gas-liquid two-phase flow in metallic pipes (Mishima and Hibiki, 1998) because of the better contrast of low mass number nuclides (H2 ) close to channel walls and fuel rod simulators than photons, which are much more strongly attenuated by the heavier elements in the metal wall. Kureta (2007) carried out neutron radiography experiments for a 3D visualization of the two-phase flow structure in a tight-lattice rod bundle. Kickhofel et al. (2011) and Zboray and Prasser (2013) have applied the technique of cold neutron imaging in a 3D tomography setup to measure the liquid film thickness profile along the walls of a tight-lattice subchannel model as well as in a square-lattice. However, cold neutron tomography requires large infrastructures, which is in case of the work of them the spallation neutron source facility SINQ of the Paul Scherrer Institute in Switzerland. Furthermore, the temporal resolution of the radiography is low. In the experiments of Zboray and Prasser (2013), a complete 3D scan took several hours and the obtained phase distributions are therefore time averaged, which requires keeping flow conditions constant over a long time. Individual interfacial structure, like wavy films or bubbles, cannot be resolved. In order to study dynamic structures forming in the liquid film at the wall in the annular flow regime, Damsohn and Prasser (2009a, 2009b) developed a liquid film sensor which is based on a high speed electrical conductance measurement. They applied the sensor to time resolved measurements of the film thickness distribution in the square-lattice rod bundle (Damsohn and Prasser, 2010). Conductance measurements were used before in numerous other studies to characterize the liquid film thickness, however, the liquid film sensor has the advantage to acquire instantaneous two-dimensional distributions of the film thickness on a matrix of typically 16 × 64 individual measuring locations with a time resolution of 10 kHz. The lateral resolution is 2 mm and the film thickness range is about 0.8 mm. The resolution of the film thickness measurement is about 20 m (Damsohn and Prasser, 2010). The aim of this work is to clarify the difference of flow characteristics between a square-lattice and a triangular tight-lattice subchannel and to get a better understanding of the liquid film behavior in such subchannel geometry, which is important to improve the safety of the reactor core. Therefore, in the present study, a pair of the liquid film sensors is applied to measurements in an annular flow in the subchannel geometry of both the square-lattice and the tight-lattice configuration. The test channel consisted of a segment comprising a pair of adjacent subchannels (double subchannel geometry). Then, the liquid film dynamics are investigated by using the time-resolved film thickness distributions measured by the film sensor. Finally, the film thickness fluctuations on the opposing walls in the subchannels are correlated to study the dependence on the subchannel geometries and the inlet flow conditions.
2. Experiments 2.1. Experimental setup The scheme of the experimental facility is given in Fig. 1. This setup was used for the experiments of Damsohn and Prasser (2010) (CALVIN test facility). The working fluids in the present study were air and water at room temperature. The compressor was used to send a high flow rate of air to the test channel. Water was injected upstream of the test section. The injection port is forming a slit spreading the water to the surface of the inner walls of the flow duct forming the subchannel geometry. The mixture at the exit of the vertical duct was sent to a separator. From there, both water and air were circulated back to the test section via storage tanks and a heat exchanger for keeping the temperature constant. Two test channels with different subchannel geometries were used in the experiments (Fig. 2). The flow duct follows the walls and symmetry boundaries of a pair of adjacent subchannels of square fuel rod lattice (Fig. 2(a)) and a triangular tight-lattice (Fig. 2(b)) Both test sections were made of acrylic glass. They have the same length of 2.5 m. The square-lattice channel is confined by 6 adjacent rods. Two out of these 6 rods are in contact with the fluid of the simulated pair of subchannels over an angle of 180◦ . These rods form the subchannel gap. The outer 4 rods are only quarter rods covering 90◦ each. The pair of triangular tight-lattice subchannels is surrounded by 4 rods, two of them, forming the subchannels gap, cover an angle of 120◦ , the other two of 60◦ each. The rod simulators have a diameter of 20 mm. This means that the dimensions of the simulated subchannels are scaled up by a factor of 2–3 compared to realistic reactor cores. This makes it easier to reach higher Reynolds numbers in experiments at ambient temperature. It also better responds to the fact that the surface tension of water at room temperature and with it the capillary length scale is higher compared to reactor conditions. In the square-lattice subchannels, the rod pitch is 26.5 mm and the free cross-sectional area is 776 mm2 , the width of the subchannel gap is 6.5 mm. On the other hand, the tight-lattice channel has a gap width of 2 mm and the pitch is 22 mm. Its flow cross-section has an area of 34.9 mm2 . The hydraulic equivalent diameters of the square- and tight-lattice channels are 18.9 mm and 1.97 mm, respectively. A pair of film thickness sensors was installed on those two opposite fuel rod models, which form the subchannel gap, as shown in Fig. 2. The outer radius of these
Fig. 1. Experimental test facility CALVIN (Damsohn and Prasser, 2010).
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
469
a Sensor2 6.5 mm Sensor1
b Fig. 3. Photo of the liquid film sensor electrodes (T: Transmitter electrode, R: Receiver electrode, G: Ground electrode).
Sensor2 2 mm Sensor1
Fig. 2. Scheme of the cross-sections of square-lattice and triangular tight-lattice double subchannels.
curved sensors is 10 mm. The lower edge of the sensitive area of both sensors was located at 1.7 m downstream of the water inlet. In the experiments, the superficial gas velocity was varied from 20 to 70 m/s and the superficial liquid velocity was set to a value between 0.2 and 0.5 m/s. At these flow conditions, an annular flow is established in test section, and the ranges of the dimensionless parameters ReL , ReG and WeL are presented in Table 1. These parameters are calculated by the following equations: ReL =
JL Dh . L
(1)
ReG =
JG Dh . G
(2)
WeL =
L JL2 Dh L
.
(3)
where Dh , , and are the hydraulic equivalent diameter, the kinematic viscosity and the surface tension, respectively. 2.2. Measurement method The pair of the liquid film sensors applied to the annular flow measurement in square-lattice and tight-lattice subchannels is inserted from the side into slots of the acrylic glass plates forming the test channel. The sensing surface is composed of a flexible printed circuit board (PCB) wrapped around a semicircular
Table 1 Re and We for the current flow conditions. geometry ReL ReG WeL
Square-lattice 3800–9500 25,000–86,000 10.4–65
Tight-lattice 400–1000 2600–9000 1.1–6.8
cylindrical carrier. It carries transmitter, receiver and ground electrodes, as shown in Fig. 3. The flexible surface is bent with a radius of curvature of 10 mm, which corresponds to the rod diameter of 20 mm. The axial length of the sensitive area is 128 mm. The transmitter and receiver electrodes have a diameter of 0.5 mm, the ground electrodes of 0.9 mm. The thickness of the conductive layer forming the electrode pads is 35 m. This creates a certain surface roughness between electrodes and insulating base material of the PCB board. However, the intrusive effect on the flow is not considered significant because the resulting unevenness is smaller than the characteristic wavelength of the flow under the current experimental conditions. The lateral distance between transmitters as well as between receiver electrodes is 2 mm, which is the spatial step width of the two-dimensional matrix of sensitive points, thus the lateral resolution of the sensors. The electrical current measured between transmitter and receiver reflects the local instantaneous film thickness at the individual measuring position within the matrix in a non-linear way (Damsohn and Prasser, 2009a). Since the measured conductance is also proportional to the conductivity of the liquid phase, the sensor signals are related to the saturation values obtained in the situation of a test channel which is completely filled with water. A calibration for the film thickness was performed to obtain a polynomial fit, which is used to convert the electrical signals into local instantaneous liquid film thickness information. Exact water film layer was formed on the sensing part by using an acrylic flat plate wall in a water vessel, as shown in Fig. 4. The film thickness was varied at a range from 100 to 1000 m by using a supporting frame which is placed outside of the sensing part. Because the sensor surface is curved, the sensor was rotated on the flat plate to get the signal at all circumferential sensing position. Then, the electrical conductance corresponding to each film thickness was obtained for all measurement points, individually. The calibration results of both sensors are shown in Fig. 5. The dimensionless current is a ratio of the measured electrical conductance to the saturated conductance value which is reached asymptotically for a large film thickness. Here, averaged values for all sensitive points of each sensor are plotted as the circular and triangular points, while the actual calibration is performed for each sensitive point of the matrix individually. It is seen that the sensor characteristics are the almost same for two sensors. The liquid film thickness was calculated from the measured conductance by polynomial curve fitting of the calibration points. The
470
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
tomography with cold neutrons has been presented by Zboray et al. (2011). Their results show a good agreement with the time averaged liquid film thickness. 3. Results and discussion 3.1. Spatio-temporal distributions of the liquid film thickness
Supporting frame
The instantaneous distributions of liquid film thickness on the rods in the square-lattice and the triangular tight-lattice configurations are represented in an isometric view in Figs. 6 and 7, respectively. The liquid film behavior on both opposing rod surfaces is visualized by splitting the subchannels graphically in two half shells. It is seen that the ratio between the thick (blue) and thin (white) film regions changes by the flow condition. In Fig. 6(a), thick liquid film regions, so-called roll waves, flow upward. The wave velocity can be roughly estimated from the location of the wave at each frame. In addition, the waves on both walls were wellcorrelated as shown in the figure. The similar wave patterns were observed on the both walls, as shown Fig. 6(b) and (c). These correlations were also found in the tight-lattice geometry, as shown in Fig. 7. In Fig. 7(a), the small ripple waves flow upward on both walls at the same time. Although the liquid film behavior changes depending on the flow condition as with the square-lattice geometry, the ratio of the thick region in tight-lattice configuration seems smaller than the square-lattice one, especially, at high liquid velocity and low gas velocity condition. This might be caused by the thin liquid film formation due to the higher gas flow rate in the narrow gap. The fluctuations of the instantaneous local film thickness in the center region of the subchannels are shown in Fig. 8. The passage of disturbance waves, as represented in Figs 6(a) and 7(a), is reflected by a short peak of the film thickness. For identical flow condition a significant difference of the wave passage frequency was not observed between both geometries, as shown in Fig. 8(a) and (b). Obviously, the channel geometry affects the mechanism of wave formation.
Curved sensor surface Calibration point
Fig. 4. Schematic diagram of sensor calibration method.
Film thickness [μm]
1000
800
Sensor1 Sensor2 Fitted curve(Sensor1)
600
400
200
0
200
400
600
800
1000
Dimensionless current [-] Fig. 5. Calibration results for the liquid film sensors.
coefficients of the polynomial curve with fourth order were estimated, as follows: a1 X 4 + a2 X 3 + a3 X 2 + a4 X + a5 . ı= (X − b1 )(X − b2 )
(4)
The coefficients (a1 –a5 ) were determined by fitting the calibration curves for each measurement point. b1 and b2 are the poles of the calibration function, which help to better reflect the asymptotic character of the curve at a large film thickness. Here, X is defined as a dimensionless sensor signal calculated as follows, X=
wall . L
(5)
where wall is the measured local instantaneous conductance at a given sensitive point and L is the asymptotic conductance at signal saturation conditions recorded at the same point. The fitted curve obtained by the above procedure is also shown in Fig. 5. A wire-mesh signal acquisition system, WMS200 (Teletronic Rossendorf GmbH), was used for the high speed electrical conductance measurements with the liquid film sensors. In the given sensor configuration, the sampling rate was set to the maximum of 5000 Hz. Two sensors occupied 32 transmitter lines and 128 receiver inputs, since each sensor has a matrix size of 16 × 64 sensitive points. The same sensor was applied to the liquid film measurement by Damsohn and Prasser (2010) in the square-lattice double subchannel, where an accuracy of about 20 m was found. In addition, a comparison between conductance sensor and a
3.2. Statistical properties of the film thickness Time-averaged film thickness profiles are compared in Fig. 9, which displays only profiles from one of both opposite rods, since they are perfectly identical on both sides. The horizontal axis denotes the angle in polar coordinates and the origin of the coordinate system refers to the center of the double subchannel, i.e. the narrowest point of the subchannel gap. The figure presents the dependence on the superficial gas velocity at constant liquid flow rate. In all analyzed cases, the film thickness is larger in the narrow subchannel gap and smaller on the walls facing the center of the subchannels. This might be an effect of the liquid flow pattern on the rod surface and ripple wave dominated flow was observed in the present flow conditions. Therefore, most ripple waves are found in the subchannel gap region and water tends to accumulate in the center. Compared to the square-lattice, the triangular tight-lattice results show thinner films in the subchannel gap and more liquid in the side edge of the wall. Note that the edge effect is too much pronounced due to the unrealistic non-slip boundary conditions, which result from the cutting out of the double subchannel geometry from the real lattice. This edge effect is less pronounced in the square-lattice geometry, since the triangular tight-lattice has a narrower edge region and the water tends more to be collected in the corners. It should be mentioned that subchannel configurations representing at least one central subchannel without unrealistic boundary conditions were studied under the cold neutron beam for
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
471
Fig. 6. Instantaneous distributions of liquid film thickness on the opposing walls in the square-lattice double subchannel configuration at (a) JL = 0.2 m/s, JG = 60 m/s, (b) JL = 0.4 m/s, JG = 40 m/s and (c) JL = 0.5 m/s, JG = 20 m/s. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
both square and triangular geometries (Zboray et al., 2011; Zboray and Prasser, 2013). The location which is least affected by the unrealistic boundary conditions is the region of the subchannel gap. Therefore, the time-averaged film thickness at the position of 0◦ in the subchannel gap is plotted against the volumetric liquid flux ratio in Fig. 10. Here, the volumetric liquid flux ratio was defined as follow, 1−ˇ =
JL . JG + JL
(6)
The averaged film thickness in the square-lattice channel increases significantly as the volumetric liquid flux ratio increases. On the other hand, this tendency is nearly absent for the triangular tight-lattice. Furthermore, the influence of the superficial liquid velocity on the averaged film thickness is weaker in the tightlattice case, however, a relatively large sensitivity was found for the square-lattice channel. This can be explained by stronger acceleration of the liquid film limiting the film growth in the narrow gap of
the tight-lattice channel, which is a result of more intensive shear forces impeded by the fast flowing gas core.
3.3. Correlation of the film thickness fluctuations between two opposing walls As noted in Section 3.1, the wave structures in the opposing walls are correlated, whereas the degree of correlation depends on the flow conditions. This gave the motivation for a correlation analysis. A cross-variance of the film thicknesses fluctuations in the center of the subchannel gap was calculated by the following equation:
n C1,2 =
i=0
n i=0
ı1,i − ı¯ 1
ı1,i − ı¯ 1
ı2,i − ı¯ 2
2 n i=0
ı2,i − ı¯ 2
2
(7)
472
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
Fig. 7. Instantaneous distributions of liquid film thickness on the opposing walls in the triangular tight-lattice double subchannel configuration at (a) JL = 0.2 m/s, JG = 60 m/s, (b) JL = 0.4 m/s, JG = 40 m/s and (c) JL = 0.6 m/s, JG = 20 m/s.
where ı is the local instantaneous film thickness at an angle of 0◦ , while ı¯ is the corresponding time-averaged thickness. The index i is the number of the time sample and n is a total number of the frames. Fig. 11 shows the obtained cross-variance. On the horizontal axis the volumetric liquid flux ratio is given. The vertical axis is the calculated cross-variance. The correlation becomes weaker with a decreasing volumetric liquid flux ratio. This means that the flows on each wall become more independent
at higher gas flow. More similar wave patterns appear on both walls, when the gas flow is low. Fig. 12 shows the effect of the superficial liquid velocity on the cross-variance. Although the dependence on the superficial liquid velocity was high in the square-lattice channel, the cross-variance for the triangular tight-lattice is higher and shows a little variation with changing superficial liquid velocity. This might be caused by the stronger interaction between both films due to the narrower gap.
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
a
473
δ [μm]
600
0.4
400
0
0.1
0.2
0.3
Time [s]
b δ [μm]
600 400 200 0
0.1
0.2
Cross-variance[-]
200
0.3
0.2
0.1
0.3
Time [s] Fig. 8. Comparison of fluctuations of liquid film thickness in the subchannel gap of (a) the square-lattice and (b) the triangular tight-lattice (JL = 0.2 m/s, JG = 60 m/s, z = 0 mm).
0.02
0.03
1-β [-]
Cross variance [-]
Film thickness [μm]
500 400 300 200
0 -90
0.01
Fig. 11. Comparison of cross-variance at the narrowest place of the subchannel gap.
600
100
0
Square-Lattice Tight-Lattice JL=0.2m/s JL=0.2m/s JL=0.3m/s JL=0.3m/s JL=0.4m/s JL=0.4m/s JL=0.5m/s JL=0.5m/s
Square-Lattice JG=20m/s JG=40m/s JG=60m/s -60
-30
Tight-Lattice JG=20m/s JG=40m/s JG=60m/s 0
30
60
Tight-Lattice JL=0.2m/s JL=0.4m/s
0.4 0.3 0.2 0.1
0
Square-Lattice JL=0.2m/s JL=0.4m/s 20
40
60
80
JG [m/s] 90
Fig. 12. The effect of superficial liquid velocity on the cross-variance at the narrowest place of the subchannel gap.
Polar position [deg.] Fig. 9. Time-averaged film thickness profiles at JL = 0.4 m/s.
Mean film thickness [μm]
500
400
300
200
100
0
Square-Lattice Tight-Lattice JL=0.2m/s JL=0.2m/s JL=0.3m/s JL=0.3m/s JL=0.4m/s JL=0.4m/s JL=0.5m/s JL=0.5m/s 0.01
0.02
0.03
1-β [-] Fig. 10. Comparison of the time-averaged film thickness at the narrowest place of subchannel gap.
4. Conclusions and outlooks To clarify the difference of the liquid film dynamics between a square-lattice and a triangular tight-lattice fuel rod bundle, annular two-phase flow measurements were performed. The test channel consisted of a segment comprising a pair of adjacent subchannels (double subchannel geometry). A pair of liquid film sensors measuring electrical conductance has been utilized to visualize the film flow on opposing rod surfaces. Time resolved distributions of the film thickness in the subchannels were obtained. From the instantaneous film thickness distributions, the characteristics of the film flow were extracted and the liquid film structure on the opposing walls in the double subchannel geometry was measured simultaneously. While the wave behaviors were visualized well in the instantaneous distributions, large difference was not observed between square- and tight-lattice channels. Furthermore, film thickness fluctuations on both opposing walls in the subchannel gap were found to be correlated to varying degrees depending on inlet conditions. Dependencies on the geometry and on liquid and gas flow rates were studied by the calculation of a crossvariance. Such correlation characteristics of the film flow between the rod surfaces were first investigated in this study. As a result, specifics of the flow behavior in a tight-lattice bundle were highlighted. As an outlook it can be suggested to use the data acquired by the present technique for the clarification of the mixing process
474
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
and of droplet deposition and entrainment in the subchannels, as well. In addition, the effect of functional spacers can be investigated by using the liquid film sensors in the future. This was done for the square-lattice already (Damsohn and Prasser, 2010). The neutron tomography studies on the spacer effect have also carried out for square-lattice (Zboray et al., 2011) and tight-lattice geometry (Zboray and Prasser, 2013). Liquid film dynamics with the spacer in the tight-lattice channel will be studied by using the present technique. Such knowledge will be very meaningful for the validation of multidimensional analysis using the advanced CFD code. References Damsohn, M., Prasser, H.-M., 2009a. High-speed liquid film sensor for two-phase flows with high spatial resolution based on electrical conductance. Flow Meas. Instrum. 20, 1–14. Damsohn, M., Prasser, H.-M., 2009b. High-speed liquid film sensor with high spatial resolution. Meas. Sci. Technol. 20, 114001. Damsohn, M., Prasser, H.-M., 2010. Experimental studies of the effect of functional spacers to annular flow in subchannels of a BWR fuel element. Nucl. Eng. Des. 240, 3126–3144. Iwamura, T., Uchikawa, S., Okubo, T., Kugo, T., Akie, H., Nakano, Y., Nakatsuka, T., 2006. Concept of innovative water reactor for flexible fuel cycle (FLWR). Nucl. Eng. Des. 236, 1599–1605.
Kickhofel, J., Zboray, R., Damsohn, Kaestner, M.A., Lehmann, E.H., Prasser, H.-M., 2011. Cold neutron tomography of annular coolant flow in a double subchannel model of a boiling water reactor. Nucl. Instrum. Methods Phys. Res. Sec. A: Accel. Spectrometers Detectors Assoc. Equip. 651 (1), 297–304. Kureta, M., 2007. Experimental study of three-dimensional void fraction distribution in heated tight-lattice rod bundles using three-dimensional neutron tomography. J. Power Energy Syst. 1 (3), 225–238. Mishima, K., Hibiki, T., 1998. Development of high-frame-rate neutron radiography and quantitative measurement method for multiphase flow research. Nucl. Eng. Des. 184, 183–201. Tamai, H., Kureta, M., Yoshida, H., Akimoto, H., 2004. Pressure drop characteristics in tight-lattice rod bundles for reduced-moderation water reactors. JSME Int J., Ser. B: Fluids Therm. Eng. 47 (2), 293–298. Tamai, H., Kureta, M., Ohnuki, A., Sato, T., Akimoto, H., 2006. Pressure drop experiments using tight-lattice 37-rod bundles. J. Nucl. Sci. Technol. 43 (6), 1–8. Yoshida, H., Nagayoshi, T., Takase, K., Akimoto, H., 2008. Development of design technology on thermal-hydraulic performance in tight-lattice rod bundles: III—Numerical evaluation of fluid mixing phenomena using advanced interfacetracking method. J. Power Energy Syst. 2 (1), 250–261. Zboray, R., Kickhofel, J., Damsohn, M., Prasser, H.-M., 2011. Cold-neutron tomography of annular flow and functional spacer performance in a model of a boiling water reactor fuel rod bundle. Nucl. Eng. Des. 241 (8), 3201–3215. Zboray, R., Prasser, H.-M., 2013. Neutron imaging of annular flows in a tight lattice fuel bundle model. Nucl. Eng. Des. 262, 589–599. Zhang, W., Yoshida, H., Ose, Y., Ohnuki, A., Akimoto, H., Hotta, A., Fujimura, K., 2008. Numerical investigation of cross flow phenomena in a tight-lattice rod bundle using advanced interface tracking method. J. Power Energy Syst. 2 (2), 456–466.
Contents lists available at ScienceDirect
Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes
Liquid film dynamics of two-phase annular flow in square and tight lattice subchannels Daisuke Ito a,∗ , Petros Papadopoulos b , Horst-Michael Prasser b a b
Research Reactor Institute, Kyoto University, Osaka, Japan Department of Mechanical and Process Engineering, ETH Zürich, Zürich, Switzerland
h i g h l i g h t s • • • •
A pair of liquid film sensors was applied to flow measurements in a double subchannel geometry. Film thickness distributions in square and tight lattice subchannels were measured. The characteristics of the water film flow were compared between each subchannel. Correlation feature of film thickness fluctuations in a subchannel gap were studied.
a r t i c l e
i n f o
Article history: Received 29 June 2015 Received in revised form 17 February 2016 Accepted 20 February 2016 Available online 5 March 2016 Classification: K. Thermal hydraulics
a b s t r a c t The flow behavior of a liquid film on the wall in square-lattice and tight-lattice double subchannels was studied by using a pair of flexible liquid film sensors based on the electrical conductance method. The liquid film sensor has a network of electrodes on the surface and the electrical conductance between transmitter and receiver electrodes is detected by using a high-speed wire-mesh measurement system. The sensors were fixed on rod simulators with 20 mm diameter and the liquid film thickness distribution was estimated from the measured conductance data array. With this method, it is possible to observe the flow structure of the liquid film with a high sampling speed up to 10 kHz and a spatial resolution of 2 mm. In the present study, the spatial-temporal distributions of liquid film thickness in the annular flow in a square-lattice and a tight-lattice pair of adjacent subchannels were measured by using two liquid film sensors installed on the surface of opposing rods. The traveling of disturbance waves was observed in the time-series of two-dimensional film thickness distributions. The interaction of the liquid films on the opposing rods was investigated by correlating the signals of both sensors. As a result, the effect of the channel geometry on the averaged film thickness profiles and the correlation characteristics was analyzed for the first time. © 2016 Elsevier B.V. All rights reserved.
1. Introduction A stable energy source has to be ensured for a sustainable energy supply in the future. Nuclear energy has been considered as an option to cover the base load of the electricity supply. However, the management and effective utilization of plutonium is very important in terms of sustainability and issues of non-proliferation. High conversion boiling water reactors (HCBWR) are potential candidates for such an application as well as fast breeder reactors (Iwamura et al., 2006). The HCBWR can achieve a high conversion
∗ Corresponding author. Tel.: +81 72 451 2373. E-mail address: [email protected] (D. Ito). http://dx.doi.org/10.1016/j.nucengdes.2016.02.034 0029-5493/© 2016 Elsevier B.V. All rights reserved.
ratio by reducing the volume of water that is acting as coolant and moderator in the same time. For this purpose, a tight-lattice rod bundle with a narrow gap between the fuel rods is adopted. The main advantage of the HCBWR over Gen IV reactor concepts consists in the direct use of the well-established technology of light water reactors for the improvement of fuel sustainability. In order to improve and assess the safety of the HCBWR, experimental data is needed to characterize the two-phase flow in the narrow subchannels of tight-lattice rod bundles. A number of studies on thermal-hydraulics in the tight-lattice rod bundle have been conducted by the Japan Atomic Energy Agency (JAEA). Kureta (2007) measured void fraction distributions in a rod bundle using neutron radiography. Tamai et al. (2004, 2006) studied the heat transfer characteristics in a 37-rod bundle. Furthermore, several
468
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
numerical studies were also carried out to evaluate the two-phase flow phenomena in the tight-lattice configuration (Yoshida et al., 2008; Zhang et al., 2008). For the validation of high fidelity simulations using an interfacial tracking technique, an experimental database on the spatio-temporal characteristics of the flow structure is needed. Of particular interests are the bubble behavior at low void fraction conditions and the dynamics of the liquid film in the annular flow regime. Two-phase flow experiments with high-resolution instrumentation are not only interesting for tight-lattice geometries, but can contribute to a better understanding and modeling of traditional square-lattice rod bundles of the existing light water reactors. Recently, several measurement tools have been developed for such experiments in the rod bundle. There is considerable progress in the field of imaging by X-rays, gamma radiation and neutrons. The common main advantage is the non-intrusiveness of the measurement. Neutron radiography is an excellent tool for the visualization of a gas-liquid two-phase flow in metallic pipes (Mishima and Hibiki, 1998) because of the better contrast of low mass number nuclides (H2 ) close to channel walls and fuel rod simulators than photons, which are much more strongly attenuated by the heavier elements in the metal wall. Kureta (2007) carried out neutron radiography experiments for a 3D visualization of the two-phase flow structure in a tight-lattice rod bundle. Kickhofel et al. (2011) and Zboray and Prasser (2013) have applied the technique of cold neutron imaging in a 3D tomography setup to measure the liquid film thickness profile along the walls of a tight-lattice subchannel model as well as in a square-lattice. However, cold neutron tomography requires large infrastructures, which is in case of the work of them the spallation neutron source facility SINQ of the Paul Scherrer Institute in Switzerland. Furthermore, the temporal resolution of the radiography is low. In the experiments of Zboray and Prasser (2013), a complete 3D scan took several hours and the obtained phase distributions are therefore time averaged, which requires keeping flow conditions constant over a long time. Individual interfacial structure, like wavy films or bubbles, cannot be resolved. In order to study dynamic structures forming in the liquid film at the wall in the annular flow regime, Damsohn and Prasser (2009a, 2009b) developed a liquid film sensor which is based on a high speed electrical conductance measurement. They applied the sensor to time resolved measurements of the film thickness distribution in the square-lattice rod bundle (Damsohn and Prasser, 2010). Conductance measurements were used before in numerous other studies to characterize the liquid film thickness, however, the liquid film sensor has the advantage to acquire instantaneous two-dimensional distributions of the film thickness on a matrix of typically 16 × 64 individual measuring locations with a time resolution of 10 kHz. The lateral resolution is 2 mm and the film thickness range is about 0.8 mm. The resolution of the film thickness measurement is about 20 m (Damsohn and Prasser, 2010). The aim of this work is to clarify the difference of flow characteristics between a square-lattice and a triangular tight-lattice subchannel and to get a better understanding of the liquid film behavior in such subchannel geometry, which is important to improve the safety of the reactor core. Therefore, in the present study, a pair of the liquid film sensors is applied to measurements in an annular flow in the subchannel geometry of both the square-lattice and the tight-lattice configuration. The test channel consisted of a segment comprising a pair of adjacent subchannels (double subchannel geometry). Then, the liquid film dynamics are investigated by using the time-resolved film thickness distributions measured by the film sensor. Finally, the film thickness fluctuations on the opposing walls in the subchannels are correlated to study the dependence on the subchannel geometries and the inlet flow conditions.
2. Experiments 2.1. Experimental setup The scheme of the experimental facility is given in Fig. 1. This setup was used for the experiments of Damsohn and Prasser (2010) (CALVIN test facility). The working fluids in the present study were air and water at room temperature. The compressor was used to send a high flow rate of air to the test channel. Water was injected upstream of the test section. The injection port is forming a slit spreading the water to the surface of the inner walls of the flow duct forming the subchannel geometry. The mixture at the exit of the vertical duct was sent to a separator. From there, both water and air were circulated back to the test section via storage tanks and a heat exchanger for keeping the temperature constant. Two test channels with different subchannel geometries were used in the experiments (Fig. 2). The flow duct follows the walls and symmetry boundaries of a pair of adjacent subchannels of square fuel rod lattice (Fig. 2(a)) and a triangular tight-lattice (Fig. 2(b)) Both test sections were made of acrylic glass. They have the same length of 2.5 m. The square-lattice channel is confined by 6 adjacent rods. Two out of these 6 rods are in contact with the fluid of the simulated pair of subchannels over an angle of 180◦ . These rods form the subchannel gap. The outer 4 rods are only quarter rods covering 90◦ each. The pair of triangular tight-lattice subchannels is surrounded by 4 rods, two of them, forming the subchannels gap, cover an angle of 120◦ , the other two of 60◦ each. The rod simulators have a diameter of 20 mm. This means that the dimensions of the simulated subchannels are scaled up by a factor of 2–3 compared to realistic reactor cores. This makes it easier to reach higher Reynolds numbers in experiments at ambient temperature. It also better responds to the fact that the surface tension of water at room temperature and with it the capillary length scale is higher compared to reactor conditions. In the square-lattice subchannels, the rod pitch is 26.5 mm and the free cross-sectional area is 776 mm2 , the width of the subchannel gap is 6.5 mm. On the other hand, the tight-lattice channel has a gap width of 2 mm and the pitch is 22 mm. Its flow cross-section has an area of 34.9 mm2 . The hydraulic equivalent diameters of the square- and tight-lattice channels are 18.9 mm and 1.97 mm, respectively. A pair of film thickness sensors was installed on those two opposite fuel rod models, which form the subchannel gap, as shown in Fig. 2. The outer radius of these
Fig. 1. Experimental test facility CALVIN (Damsohn and Prasser, 2010).
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
469
a Sensor2 6.5 mm Sensor1
b Fig. 3. Photo of the liquid film sensor electrodes (T: Transmitter electrode, R: Receiver electrode, G: Ground electrode).
Sensor2 2 mm Sensor1
Fig. 2. Scheme of the cross-sections of square-lattice and triangular tight-lattice double subchannels.
curved sensors is 10 mm. The lower edge of the sensitive area of both sensors was located at 1.7 m downstream of the water inlet. In the experiments, the superficial gas velocity was varied from 20 to 70 m/s and the superficial liquid velocity was set to a value between 0.2 and 0.5 m/s. At these flow conditions, an annular flow is established in test section, and the ranges of the dimensionless parameters ReL , ReG and WeL are presented in Table 1. These parameters are calculated by the following equations: ReL =
JL Dh . L
(1)
ReG =
JG Dh . G
(2)
WeL =
L JL2 Dh L
.
(3)
where Dh , , and are the hydraulic equivalent diameter, the kinematic viscosity and the surface tension, respectively. 2.2. Measurement method The pair of the liquid film sensors applied to the annular flow measurement in square-lattice and tight-lattice subchannels is inserted from the side into slots of the acrylic glass plates forming the test channel. The sensing surface is composed of a flexible printed circuit board (PCB) wrapped around a semicircular
Table 1 Re and We for the current flow conditions. geometry ReL ReG WeL
Square-lattice 3800–9500 25,000–86,000 10.4–65
Tight-lattice 400–1000 2600–9000 1.1–6.8
cylindrical carrier. It carries transmitter, receiver and ground electrodes, as shown in Fig. 3. The flexible surface is bent with a radius of curvature of 10 mm, which corresponds to the rod diameter of 20 mm. The axial length of the sensitive area is 128 mm. The transmitter and receiver electrodes have a diameter of 0.5 mm, the ground electrodes of 0.9 mm. The thickness of the conductive layer forming the electrode pads is 35 m. This creates a certain surface roughness between electrodes and insulating base material of the PCB board. However, the intrusive effect on the flow is not considered significant because the resulting unevenness is smaller than the characteristic wavelength of the flow under the current experimental conditions. The lateral distance between transmitters as well as between receiver electrodes is 2 mm, which is the spatial step width of the two-dimensional matrix of sensitive points, thus the lateral resolution of the sensors. The electrical current measured between transmitter and receiver reflects the local instantaneous film thickness at the individual measuring position within the matrix in a non-linear way (Damsohn and Prasser, 2009a). Since the measured conductance is also proportional to the conductivity of the liquid phase, the sensor signals are related to the saturation values obtained in the situation of a test channel which is completely filled with water. A calibration for the film thickness was performed to obtain a polynomial fit, which is used to convert the electrical signals into local instantaneous liquid film thickness information. Exact water film layer was formed on the sensing part by using an acrylic flat plate wall in a water vessel, as shown in Fig. 4. The film thickness was varied at a range from 100 to 1000 m by using a supporting frame which is placed outside of the sensing part. Because the sensor surface is curved, the sensor was rotated on the flat plate to get the signal at all circumferential sensing position. Then, the electrical conductance corresponding to each film thickness was obtained for all measurement points, individually. The calibration results of both sensors are shown in Fig. 5. The dimensionless current is a ratio of the measured electrical conductance to the saturated conductance value which is reached asymptotically for a large film thickness. Here, averaged values for all sensitive points of each sensor are plotted as the circular and triangular points, while the actual calibration is performed for each sensitive point of the matrix individually. It is seen that the sensor characteristics are the almost same for two sensors. The liquid film thickness was calculated from the measured conductance by polynomial curve fitting of the calibration points. The
470
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
tomography with cold neutrons has been presented by Zboray et al. (2011). Their results show a good agreement with the time averaged liquid film thickness. 3. Results and discussion 3.1. Spatio-temporal distributions of the liquid film thickness
Supporting frame
The instantaneous distributions of liquid film thickness on the rods in the square-lattice and the triangular tight-lattice configurations are represented in an isometric view in Figs. 6 and 7, respectively. The liquid film behavior on both opposing rod surfaces is visualized by splitting the subchannels graphically in two half shells. It is seen that the ratio between the thick (blue) and thin (white) film regions changes by the flow condition. In Fig. 6(a), thick liquid film regions, so-called roll waves, flow upward. The wave velocity can be roughly estimated from the location of the wave at each frame. In addition, the waves on both walls were wellcorrelated as shown in the figure. The similar wave patterns were observed on the both walls, as shown Fig. 6(b) and (c). These correlations were also found in the tight-lattice geometry, as shown in Fig. 7. In Fig. 7(a), the small ripple waves flow upward on both walls at the same time. Although the liquid film behavior changes depending on the flow condition as with the square-lattice geometry, the ratio of the thick region in tight-lattice configuration seems smaller than the square-lattice one, especially, at high liquid velocity and low gas velocity condition. This might be caused by the thin liquid film formation due to the higher gas flow rate in the narrow gap. The fluctuations of the instantaneous local film thickness in the center region of the subchannels are shown in Fig. 8. The passage of disturbance waves, as represented in Figs 6(a) and 7(a), is reflected by a short peak of the film thickness. For identical flow condition a significant difference of the wave passage frequency was not observed between both geometries, as shown in Fig. 8(a) and (b). Obviously, the channel geometry affects the mechanism of wave formation.
Curved sensor surface Calibration point
Fig. 4. Schematic diagram of sensor calibration method.
Film thickness [μm]
1000
800
Sensor1 Sensor2 Fitted curve(Sensor1)
600
400
200
0
200
400
600
800
1000
Dimensionless current [-] Fig. 5. Calibration results for the liquid film sensors.
coefficients of the polynomial curve with fourth order were estimated, as follows: a1 X 4 + a2 X 3 + a3 X 2 + a4 X + a5 . ı= (X − b1 )(X − b2 )
(4)
The coefficients (a1 –a5 ) were determined by fitting the calibration curves for each measurement point. b1 and b2 are the poles of the calibration function, which help to better reflect the asymptotic character of the curve at a large film thickness. Here, X is defined as a dimensionless sensor signal calculated as follows, X=
wall . L
(5)
where wall is the measured local instantaneous conductance at a given sensitive point and L is the asymptotic conductance at signal saturation conditions recorded at the same point. The fitted curve obtained by the above procedure is also shown in Fig. 5. A wire-mesh signal acquisition system, WMS200 (Teletronic Rossendorf GmbH), was used for the high speed electrical conductance measurements with the liquid film sensors. In the given sensor configuration, the sampling rate was set to the maximum of 5000 Hz. Two sensors occupied 32 transmitter lines and 128 receiver inputs, since each sensor has a matrix size of 16 × 64 sensitive points. The same sensor was applied to the liquid film measurement by Damsohn and Prasser (2010) in the square-lattice double subchannel, where an accuracy of about 20 m was found. In addition, a comparison between conductance sensor and a
3.2. Statistical properties of the film thickness Time-averaged film thickness profiles are compared in Fig. 9, which displays only profiles from one of both opposite rods, since they are perfectly identical on both sides. The horizontal axis denotes the angle in polar coordinates and the origin of the coordinate system refers to the center of the double subchannel, i.e. the narrowest point of the subchannel gap. The figure presents the dependence on the superficial gas velocity at constant liquid flow rate. In all analyzed cases, the film thickness is larger in the narrow subchannel gap and smaller on the walls facing the center of the subchannels. This might be an effect of the liquid flow pattern on the rod surface and ripple wave dominated flow was observed in the present flow conditions. Therefore, most ripple waves are found in the subchannel gap region and water tends to accumulate in the center. Compared to the square-lattice, the triangular tight-lattice results show thinner films in the subchannel gap and more liquid in the side edge of the wall. Note that the edge effect is too much pronounced due to the unrealistic non-slip boundary conditions, which result from the cutting out of the double subchannel geometry from the real lattice. This edge effect is less pronounced in the square-lattice geometry, since the triangular tight-lattice has a narrower edge region and the water tends more to be collected in the corners. It should be mentioned that subchannel configurations representing at least one central subchannel without unrealistic boundary conditions were studied under the cold neutron beam for
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
471
Fig. 6. Instantaneous distributions of liquid film thickness on the opposing walls in the square-lattice double subchannel configuration at (a) JL = 0.2 m/s, JG = 60 m/s, (b) JL = 0.4 m/s, JG = 40 m/s and (c) JL = 0.5 m/s, JG = 20 m/s. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
both square and triangular geometries (Zboray et al., 2011; Zboray and Prasser, 2013). The location which is least affected by the unrealistic boundary conditions is the region of the subchannel gap. Therefore, the time-averaged film thickness at the position of 0◦ in the subchannel gap is plotted against the volumetric liquid flux ratio in Fig. 10. Here, the volumetric liquid flux ratio was defined as follow, 1−ˇ =
JL . JG + JL
(6)
The averaged film thickness in the square-lattice channel increases significantly as the volumetric liquid flux ratio increases. On the other hand, this tendency is nearly absent for the triangular tight-lattice. Furthermore, the influence of the superficial liquid velocity on the averaged film thickness is weaker in the tightlattice case, however, a relatively large sensitivity was found for the square-lattice channel. This can be explained by stronger acceleration of the liquid film limiting the film growth in the narrow gap of
the tight-lattice channel, which is a result of more intensive shear forces impeded by the fast flowing gas core.
3.3. Correlation of the film thickness fluctuations between two opposing walls As noted in Section 3.1, the wave structures in the opposing walls are correlated, whereas the degree of correlation depends on the flow conditions. This gave the motivation for a correlation analysis. A cross-variance of the film thicknesses fluctuations in the center of the subchannel gap was calculated by the following equation:
n C1,2 =
i=0
n i=0
ı1,i − ı¯ 1
ı1,i − ı¯ 1
ı2,i − ı¯ 2
2 n i=0
ı2,i − ı¯ 2
2
(7)
472
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
Fig. 7. Instantaneous distributions of liquid film thickness on the opposing walls in the triangular tight-lattice double subchannel configuration at (a) JL = 0.2 m/s, JG = 60 m/s, (b) JL = 0.4 m/s, JG = 40 m/s and (c) JL = 0.6 m/s, JG = 20 m/s.
where ı is the local instantaneous film thickness at an angle of 0◦ , while ı¯ is the corresponding time-averaged thickness. The index i is the number of the time sample and n is a total number of the frames. Fig. 11 shows the obtained cross-variance. On the horizontal axis the volumetric liquid flux ratio is given. The vertical axis is the calculated cross-variance. The correlation becomes weaker with a decreasing volumetric liquid flux ratio. This means that the flows on each wall become more independent
at higher gas flow. More similar wave patterns appear on both walls, when the gas flow is low. Fig. 12 shows the effect of the superficial liquid velocity on the cross-variance. Although the dependence on the superficial liquid velocity was high in the square-lattice channel, the cross-variance for the triangular tight-lattice is higher and shows a little variation with changing superficial liquid velocity. This might be caused by the stronger interaction between both films due to the narrower gap.
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
a
473
δ [μm]
600
0.4
400
0
0.1
0.2
0.3
Time [s]
b δ [μm]
600 400 200 0
0.1
0.2
Cross-variance[-]
200
0.3
0.2
0.1
0.3
Time [s] Fig. 8. Comparison of fluctuations of liquid film thickness in the subchannel gap of (a) the square-lattice and (b) the triangular tight-lattice (JL = 0.2 m/s, JG = 60 m/s, z = 0 mm).
0.02
0.03
1-β [-]
Cross variance [-]
Film thickness [μm]
500 400 300 200
0 -90
0.01
Fig. 11. Comparison of cross-variance at the narrowest place of the subchannel gap.
600
100
0
Square-Lattice Tight-Lattice JL=0.2m/s JL=0.2m/s JL=0.3m/s JL=0.3m/s JL=0.4m/s JL=0.4m/s JL=0.5m/s JL=0.5m/s
Square-Lattice JG=20m/s JG=40m/s JG=60m/s -60
-30
Tight-Lattice JG=20m/s JG=40m/s JG=60m/s 0
30
60
Tight-Lattice JL=0.2m/s JL=0.4m/s
0.4 0.3 0.2 0.1
0
Square-Lattice JL=0.2m/s JL=0.4m/s 20
40
60
80
JG [m/s] 90
Fig. 12. The effect of superficial liquid velocity on the cross-variance at the narrowest place of the subchannel gap.
Polar position [deg.] Fig. 9. Time-averaged film thickness profiles at JL = 0.4 m/s.
Mean film thickness [μm]
500
400
300
200
100
0
Square-Lattice Tight-Lattice JL=0.2m/s JL=0.2m/s JL=0.3m/s JL=0.3m/s JL=0.4m/s JL=0.4m/s JL=0.5m/s JL=0.5m/s 0.01
0.02
0.03
1-β [-] Fig. 10. Comparison of the time-averaged film thickness at the narrowest place of subchannel gap.
4. Conclusions and outlooks To clarify the difference of the liquid film dynamics between a square-lattice and a triangular tight-lattice fuel rod bundle, annular two-phase flow measurements were performed. The test channel consisted of a segment comprising a pair of adjacent subchannels (double subchannel geometry). A pair of liquid film sensors measuring electrical conductance has been utilized to visualize the film flow on opposing rod surfaces. Time resolved distributions of the film thickness in the subchannels were obtained. From the instantaneous film thickness distributions, the characteristics of the film flow were extracted and the liquid film structure on the opposing walls in the double subchannel geometry was measured simultaneously. While the wave behaviors were visualized well in the instantaneous distributions, large difference was not observed between square- and tight-lattice channels. Furthermore, film thickness fluctuations on both opposing walls in the subchannel gap were found to be correlated to varying degrees depending on inlet conditions. Dependencies on the geometry and on liquid and gas flow rates were studied by the calculation of a crossvariance. Such correlation characteristics of the film flow between the rod surfaces were first investigated in this study. As a result, specifics of the flow behavior in a tight-lattice bundle were highlighted. As an outlook it can be suggested to use the data acquired by the present technique for the clarification of the mixing process
474
D. Ito et al. / Nuclear Engineering and Design 300 (2016) 467–474
and of droplet deposition and entrainment in the subchannels, as well. In addition, the effect of functional spacers can be investigated by using the liquid film sensors in the future. This was done for the square-lattice already (Damsohn and Prasser, 2010). The neutron tomography studies on the spacer effect have also carried out for square-lattice (Zboray et al., 2011) and tight-lattice geometry (Zboray and Prasser, 2013). Liquid film dynamics with the spacer in the tight-lattice channel will be studied by using the present technique. Such knowledge will be very meaningful for the validation of multidimensional analysis using the advanced CFD code. References Damsohn, M., Prasser, H.-M., 2009a. High-speed liquid film sensor for two-phase flows with high spatial resolution based on electrical conductance. Flow Meas. Instrum. 20, 1–14. Damsohn, M., Prasser, H.-M., 2009b. High-speed liquid film sensor with high spatial resolution. Meas. Sci. Technol. 20, 114001. Damsohn, M., Prasser, H.-M., 2010. Experimental studies of the effect of functional spacers to annular flow in subchannels of a BWR fuel element. Nucl. Eng. Des. 240, 3126–3144. Iwamura, T., Uchikawa, S., Okubo, T., Kugo, T., Akie, H., Nakano, Y., Nakatsuka, T., 2006. Concept of innovative water reactor for flexible fuel cycle (FLWR). Nucl. Eng. Des. 236, 1599–1605.
Kickhofel, J., Zboray, R., Damsohn, Kaestner, M.A., Lehmann, E.H., Prasser, H.-M., 2011. Cold neutron tomography of annular coolant flow in a double subchannel model of a boiling water reactor. Nucl. Instrum. Methods Phys. Res. Sec. A: Accel. Spectrometers Detectors Assoc. Equip. 651 (1), 297–304. Kureta, M., 2007. Experimental study of three-dimensional void fraction distribution in heated tight-lattice rod bundles using three-dimensional neutron tomography. J. Power Energy Syst. 1 (3), 225–238. Mishima, K., Hibiki, T., 1998. Development of high-frame-rate neutron radiography and quantitative measurement method for multiphase flow research. Nucl. Eng. Des. 184, 183–201. Tamai, H., Kureta, M., Yoshida, H., Akimoto, H., 2004. Pressure drop characteristics in tight-lattice rod bundles for reduced-moderation water reactors. JSME Int J., Ser. B: Fluids Therm. Eng. 47 (2), 293–298. Tamai, H., Kureta, M., Ohnuki, A., Sato, T., Akimoto, H., 2006. Pressure drop experiments using tight-lattice 37-rod bundles. J. Nucl. Sci. Technol. 43 (6), 1–8. Yoshida, H., Nagayoshi, T., Takase, K., Akimoto, H., 2008. Development of design technology on thermal-hydraulic performance in tight-lattice rod bundles: III—Numerical evaluation of fluid mixing phenomena using advanced interfacetracking method. J. Power Energy Syst. 2 (1), 250–261. Zboray, R., Kickhofel, J., Damsohn, M., Prasser, H.-M., 2011. Cold-neutron tomography of annular flow and functional spacer performance in a model of a boiling water reactor fuel rod bundle. Nucl. Eng. Des. 241 (8), 3201–3215. Zboray, R., Prasser, H.-M., 2013. Neutron imaging of annular flows in a tight lattice fuel bundle model. Nucl. Eng. Des. 262, 589–599. Zhang, W., Yoshida, H., Ose, Y., Ohnuki, A., Akimoto, H., Hotta, A., Fujimura, K., 2008. Numerical investigation of cross flow phenomena in a tight-lattice rod bundle using advanced interface tracking method. J. Power Energy Syst. 2 (2), 456–466.