2.2. Void Fraction in Microchannels
2.2.1. Introduction
Several studies have been conducted over the years to understand various aspects of two-phase flow in microchannel for designing efficient and reliable thermal management systems for commercial and defense applications. Void fraction is one of the important factors in two-phase flow; it is used to calculate several other parameters of interest such as the heat transfer coefficient, pressure drop, two-phase flow density, two-phase flow viscosity, and average velocities of the respective phases.
Void fraction is defined in multiple ways. Volumetric void fraction is defined as the ratio of vapor volume to the total volume for a fixed channel length. In the area-averaged void fraction, fraction of cross-sectional area of channel occupied by the vapor phase is calculated. The chordal void fraction is evaluated by dividing the length of the vapor phase by the channel length. The local void fraction is referred to the ratio of time spent by vapor phase at a particular point to the total time. The void fraction in a channel varies both along the length of the channel and with time.
2.2.2. Different Methods of Void Fraction Measurement
Several methods have been developed for measuring the void fraction, some of which are discussed in this section. Pujara et al.
[48] summarized the void fraction measurement techniques as listed in
Table 2.6.
2.2.2.1. Experimental Methods
Singh et al.
[49] used image processing technique to study the heat transfer and different flow regime maps during the flow boiling of water in the microchannel of hydraulic diameter of 140µm. A 2-inch, 275±25-µm-thick, p-type, double-side polished silicon wafer was used to fabricate microchannels with dimensions of 173×119×20µm using CMOS technique. Surface roughness was less than 0.1µm, and sealing was achieved with a quartz plate. Deionized water was used as the working fluid. Chrome-gold thin-film stack was used as micro-heater material as it showed good linearity with respect to temperature along with high chemical and thermal stability.
Table 2.6
Different Void Fraction Measurement Techniques
| Quick closing valve | Measurement of volumetric void fraction | 1. Intrusive method 2. Requires finite time for closing of the valve 3. Requires considerable time for bringing the system back to the steady state between successive experiment |
| Conduction probe | Local time-averaged and chordal void fraction measurement | 1. Intrusive method 2. Limited to measure bubbly and slug flow regime |
| By pressure drop | Measurement of volumetric void fraction | 1. Nonintrusive method 2. Friction pressure drop and acceleration pressure drop neglected, manometer line filled by single phase |
| Radiation absorption and scattering method | Chordal void fraction measurement | 1. Nonintrusive method 2. Expensive, difficult to handle high energy radiation, presence of metal wall induces error |
| Laser beam method | Chordal void fraction measurement, flow pattern identification | 1. Nonintrusive method 2. Expensive |
| Photographic method | Chordal, cross section void fraction measurement | 1. Nonintrusive method 2. Subjective 3. Error due to operation performed on image |
| Impedance method | Volumetric void fraction, low cost, suitable for transition measurement | 1. Nonintrusive method |
| Capacitance method | Measurement of volumetric void fraction | 1. Nonintrusive 2. Relatively low cost 3. Not easy to calibrate the capacitance with void fraction due to signal dependency on the void fraction and flow patterns |
An image processing algorithm was developed to estimate the void fraction and evaluate the percentage of different flow regimes and heat transfer coefficient as the function of position, heat flux, and mass flow rate. In image processing, images were first recorded using a camera. The vapor region was identified next, and void fraction was estimated as the ratio of area of vapor to total area of the microchannel in the image. The images correspond to the plane normal to the flow direction (and not the cross-sectional view) of the microchannel. Steps for image processing techniques are:
1. First, background image was read and boundaries were whitened for edge detection. Microchannel was extracted by cropping. Whitening was required to know whether bubbles were fully enclosed by edges and its area was well defined.
2. Sample image was read, and microchannel was extracted by cropping.
3. Background was removed to reduce noise. It was done by subtracting background image from the sample image.
4. Sample image was then converted to intensity image or grayscale from RGB file type. Histogram equalization was done before converting to a binary image.
5. Noise was removed by median filter.
6. Edges were identified and vapor region with closed contours was filled.
7. Small objects with less than 40 pixels were left out for calculation of total area of vapor.
They observed that void fraction increased along the downstream of flow in the microchannel. This was observed due to the fact that the fraction of liquid in the microchannel decreased due to evaporation, which made a monotonic increase in the void fraction. Also, it was observed that the void fraction increased with lowering the volumetric flow rate. This was because a larger liquid fraction was evaporated to vapor at a given heat flux as the volumetric flow rate was reduced, thereby increasing the void fraction.
Gijsenbergh and Puers
[50] developed a capacitive fringing fields-based sensor to measure void fraction in silicon microchannels of 100µm (width) by 500µm (height) cross section. The test section consisted of 60 channels having dimensions of 500µm (depth)×100µm (width)×1000µm (long) and made of silicon die of 650-µm thickness. The change in permittivity due to change in fluid content was detected via the sensors connected to the top and the bottom electrodes. The top electrode was located at the back of each channel of silicon microchannels, and the bottom electrode was mounted on a glass substrate of each channel and insulated with a layer of Pyrex glass as shown in
Fig. 2.21. The capacitance network of the system is shown in
Fig. 2.22, where the total capacitance was given by,
total=2CSiliconwall+[1CSiliconceiling+1Cch]-1
The individual capacitance, CSiliconwall, CSiliconceiling, and Cch depend on the electrode area, the layer thickness, the absolute permeability abs and relative permeability rel. As the geometry of the sensor is fixed, all capacitances except relative permeability rel are known. The relative permeability rel, which varies due to change in fluid content in channel, is estimated for annular flow using numerical simulation. The relative permeability rel at two extremes, viz. channel completely filled with air and water, are evaluated and subsequently, the variation in rel with different proportions of air–water is correlated, which is found to be quasilinear in nature.
Figure 2.21 Schematic diagram of a single sensor
[50].
Figure 2.22 Capacitance network of a single sensor
[50].
The capacitance measured by the sensor is digitized using a 16-bit sigma-delta capacitance-to-digital converter (CDC). The digital signal is then converted into an analog voltage using a digital-to-analog converter (DAC). Their result shows a linear relationship with the measured voltage and void fraction.
Paranjape et al.
[51] developed an electrical impendence-based sensor to measure void fraction in microchannels with a square cross section of 780×780µm and length of 50.8mm. Deionized water and air were used in their two-phase experiments. The faces of two electrodes made of 304 stainless steel were flush mounted to the opposite sidewalls of...