- Email: [email protected]
Mixing and atomization characteristics in an internal-mixing twin-fluid atomizer
Fuel 97 (2012) 306–314
Contents lists available at SciVerse ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Mixing and atomization characteristics in an internal-mixing twin-fluid atomizer Zhouhang Li a, Yuxin Wu a,⇑, Chunrong Cai a, Hai Zhang a, Yingli Gong a, Keiji Takeno b, Kazuaki Hashiguchi b, Junfu Lu a a b
Key Laboratory for Thermal Science and Power Engineering of Ministry of Education, Department of Thermal Engineering, Tsinghua University, Beijing 100084, China Nagasaki R&D Center, Mitsubishi Heavy Industries, Ltd., Japan
a r t i c l e
i n f o
Article history: Received 17 June 2011 Received in revised form 14 November 2011 Accepted 6 March 2012 Available online 18 March 2012 Keywords: Twin-fluid atomization Internal mixing GLR SMD distribution
a b s t r a c t Twin-fluid atomizers have been successfully used in many industrial applications. This paper presents experimental studies on internal mixing and atomization in a water–air internal-mixing atomizer. Two-phase mixing process and flow patterns in the internal mixing chamber were visually studied through high speed CCD. Observation reveals that internal mixing was dominated by Gas to Liquid mass Ratio (GLR). As GLR increased, the flow patterns changed from slug flow to annular flow. The Oshinowo and Charles’ map can be used to predict the flow patterns for the designed atomizer. Droplet Sauter Mean Diameter (SMD) spatial distributions were measured with Phase Doppler Analyzer (PDA) at different operating conditions. Droplet SMD decreased with the increase of GLR at all operating pressure and locations. In the undeveloped region, a close relationship was observed between flow pattern transformation in internal mixing chamber and droplet SMD distribution, and there was an optimized pressure ranging from 0.2 MPa to 0.3 MPa for atomization since liquid films became thicker under a higher pressure. In the developed region, pressure promoted to generate finer atomization. Possibility Density Function (PDF) distribution of droplet size at different axial locations was analyzed to quantitatively represent the effect of droplet coalescence and breakup. As axial distance increased, PDF of both fine droplets and large droplets decreased. The particle size with the maximum PDF increased with the axial distance as well. The results imply that best atomization performance was acquired in the undeveloped region. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Twin-fluid atomizers have been widely used in many industrial applications such as oil burners and spray drying [1]. This kind of atomizer generates high relative velocity between gas and liquid by employing a relatively high gas velocity, which helps the formation of the smaller liquid sheets and ligaments. Then fine droplets come into formation due to sheet breakup. This process can be explained by the wavelengths that grow on the surface of the sheet, which are affected by surface tension, aerodynamic forces, and liquid viscosity [2]. Meanwhile, the introduction of gas phase may help dispersion of liquid and prevent from droplets coalescence. There are many types of twin-fluid atomizers, such as internal mixing [3,4], medium mixing (‘Y’ type) [4,5] and outside mixing atomizers [6]. In a boiler plant, the generally design of the oil/heavy oil burners are internal mixing or medium mixing atomizers. The characteristic of the medium mixing atomizers is that liquid and gas are mixed inside the atomizer within a limited space before they are injected out. The advantages of such kind of ‘Y’ type ⇑ Corresponding author. Tel./fax: +11 8610 62788523. E-mail address: [email protected] (Y. Wu). 0016-2361/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2012.03.006
atomizer are that it can be operated either keeping a constant steam-to-fuel flow rate ratio or a fixed fuel-to-steam pressure ratio. However, internal mixing process shall be strengthened further so that finer droplets can be acquired at a low GLR. Ferreira [4] has indicated that adding an internal mixing chamber may improve the performance of ‘Y’ type atomizer. In an internal mixing chamber, liquid and gas are mixed and interacted intensively before they are discharged out. Atomization in the internal mixing atomizer can be explained as follows [7]: liquid has to share flow area with gas, resulting in acceleration of liquid and formation of finer droplets. In addition, relative motion between two phases increases the flow instability at the interface, so that ligaments and sheets formation is accelerated. It is well recognized that atomizer geometry [4,8–10], fluid physical properties [11,12] and operating conditions [13,14] such as GLR and injection pressure have effects on the mixing process as well as atomization of internal mixing atomizers. Nguyen and Rhodes [9] found that length and diameter of the mixing chamber have little effect on atomization during a range of 20–45 mm and 1.7–2.3 mm, respectively. However, Kushari [10] showed that droplet size decreases with increase of mixing chamber length. Both Ferreira’s [4] and Kushari’s [10] results show that the ratio of the exit orifice area to the air injection area influences spray mean drop size in internal
307
Z. Li et al. / Fuel 97 (2012) 306–314
Nomenclature A di,l d0 FrTP g G GLR L _ m Q p r r⁄ SMD x
0:25 0:75 l rw factor, A ¼ l l qql rl w w mixing chamber’s inlet diameter (mm) exit orifice diameter (mm) 2 a þQ l Þ two-phase Froude number, FrTP ¼ ðQ 2 3 D gð p =4Þ gravitational constant (m/s2) gas mass flux (kg/m2 s1) Gas to Liquid mass Ratio liquid mass flux (kg/m2 s1) mass flow rate (kg/h) volumetric flow rate (m3/s) operating pressure (MPa) radial distance from spray axis (mm) non-dimensional radial distance, r/d0 Sauter Mean Diameter, D32 (lm) axial distance from the exit orifice (mm)
mixing twin-fluid atomizers, the same with effervescent atomizers [15]. Slant angle and specific shape of the liquid ports have little effect on atomization [4]. Broniarz-Press et al. [12] studied atomization of internal mixing atomizer using water and different oil emulsions, concluding that physical properties of liquid have significant influence on the spray characteristics. Flow patterns in internal-mixing twin-fluid atomizer were also investigated. Kufferath et al. [16] studied atomizer’s performance under different flow conditions of the liquid jet leaving the inlet port, including laminar and turbulent flows, and found that flow conditions have a strong influence on droplet mean size. Karnawat and Kushari [17,18] investigated different spray patterns in the developed jet regime and concluded correlations between spray patterns and operating conditions. The mixing process within the internal chamber is important for internal-mixing atomizer’s performance. Based on classical two-phase flow pattern maps, Lefebvre and Chin [19] investigated different flow patterns within the mixing chamber. Kim et al. [8] studied effects of GLR on flow patterns within the internal chamber and they reported breakup of water column was complicatedly changed with the increase of GLR. Ferreira et al. [20] found that flow patterns within the chamber change with internal mixing chamber geometry configuration. Although research work listed previously give useful information for internal mixing atomization design and research, the correlation between flow patterns in internal mixing chamber and droplet mean diameter has not been investigated in previous literatures. Operating pressure’s effect on internal mixing was seldom studied experimentally. Besides, most previous researchers (Whitlow and Lefebvre [21], Panchagnula and Sojka [22], and Jedelsky et al. [23]) reported droplet SMD at axial locations where are in the far-field region and spray is fully developed, while droplet SMD distribution in the near nozzle region was paid little attention. For oil burners, it’s important to acquire the axial distance where droplets have the smallest SMD in order to get an optimized condition. Moreover, spray with high velocity sometimes cools down excessively the reaction zone, leading to local flame extinction [24]. A reasonable atomization flow field is very important for oil burning. In present work, liquid–gas flow patterns in a water-inside–airoutside internal mixing twin-fluid atomizer were visually investigated through high speed CCD. The SMD spatial distribution and velocity profile in the near nozzle region and fully developed region were measured through Phase Doppler Analyzer (PDA). Effects of GLR and pressure on SMD spatial distribution were studied. The correlation between flow patterns in internal mixing chamber and droplet SMD was discussed. Droplet breakup and
x⁄
non-dimensional axial distance, x/d0
Greek letters k factor, k ¼ ðqg ql =qa qw Þ0:5 h i1=3 l q w factor, w ¼ rrwl l l qw w
q r l
l
density (kg/m3) surface tension (N/m) dynamic viscosity (N/s m2)
Subscripts a air l liquid g gas w water
coalescence effects were quantitatively expressed by SMD spatial distribution and possibility density distributions. 2. Experimental 2.1. internal-mixing twin-fluid atomizer and test rig system The geometry configuration of the internal-mixing twin-fluid atomizer is shown in Fig. 1. The atomizer is made by acrylic glass so that the internal mixing can be captured by high speed CCD. The overall length of the atomizer is 600 mm. The diameter of mixing chamber’s liquid inlet (di,l) is 4 mm. Liquid inlet is surrounded by four 1 mm air inlet holes with a slant angle of 45°. Atomizer’s exit orifice diameter, d0, is 3 mm. Fig. 2 shows the schematic drawing of the air–water atomization system. Both water and air were fed into the atomizer at an elevated operating pressure. A ball valve and a needle valve were installed on each route in order to control the flow rate of pressurized water and air. Metal tube rotameters with a maximum error of 1.5% were adopted to acquire the volume flow rate of water and air. 2.2. Measuring apparatus A high speed CCD, the PCO.dimax made by COOK Company, was used to capture the internal flow patterns and two-phase mixing process inside the chamber. The CCD has a high ISO value of 50,000. The minimum exposure time is 2 ls. Up to 1297 frames per second (fps) can be acquired with the size of 2016 2016 pixels for each frame. In this work, pictures and videos of the mixing processes were captured at an fps of 10,000 so that change of the two-phase flow can be analyzed in detail. A Phase Doppler Analyzer (PDA) system of Dantec Dynamics was used to measure droplet size and velocity distributions. The PDA was set in refractive scattering mode with the receiver positioned at 30° from the transmitter axis. The focal length for the
Fig. 1. Geometry configuration of the research atomizer.
308
Z. Li et al. / Fuel 97 (2012) 306–314
from 0.1 MPa to 0.5 MPa and GLR from 0.02 to 0.2. The detailed experimental cases are listed in Table 1. Under each case, droplet size and velocities were acquired at 88 spatial locations distributed at 8 different x, within a range of 10–400 mm. At each location, 2000 samples were acquired with PDA. 3. Results and discussion 3.1. Two-phase flow patterns and mixing process inside the mixing chamber
Fig. 2. Schematic of the experimental system. 1 – Pressurized water tank, 2 – adjusting valve, 3 – mass flowmeter, 4 – pressure meter, 5 – atomizer, 6 – observation tank, 7 – PDA laser lens, 8 – gas flow rotameter, 9 – water tank, and 10 – outlet.
transmitting lens was 800 mm and for the receiver was 600 mm. The maximum diameter which can be measured with this setup is 250 lm. Droplet distributions at different spatial locations were measured by adjusting a three dimensional auto control coordinate system with a size of 560 560 580 mm.
2.3. Experimental cases To conduct a complete analysis of the characteristic of the atomizer, experimental measurements under different parameters, such as operating pressure and GLR, are needed. In the designed atomizer, water and air were well mixed in the internal mixing chamber, and the pressure differential between water and air at the atomizer inlet was relatively small at all experimental cases. This is in accordance with the results of Jedelsky et al. [23]. In their studies on atomizers with an internal mixing chamber, pressure differential between gas and liquid was in a range of 3.3 to 10.6 kPa at different operation conditions and geometry configurations. The existence of the mixing chamber makes pressure differential between gas and liquid small enough to be neglected. Thus, operating pressure p represents the gauge pressure of the mixing chamber. Since pressure differential between water and air results in different mass flow rates of two phases at a given p, GLR is used to represent the effect of such pressure differential between liquid and gas routes. The experimental conditions covered the range of p
Table 1 Experimental cases of atomizer. Case No.
p (MPa)
GLR
_ l (kg/ m h)
Case No.
p (MPa)
GLR
_ l (kg/ m h)
1 2 3 4 5 6 7 8 9 10 11 12 13 14
0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.2 0.2 0.2 0.3 0.3
0.02 0.05 0.07 0.1 0.15 0.2 0.02 0.05 0.07 0.1 0.15 0.2 0.02 0.05
125 83 70 55 44 41 180 112 94 78 62 51 240 153
15 16 17 18 19 20 21 22 23 24 25 26 27 28
0.3 0.3 0.3 0.3 0.4 0.4 0.4 0.4 0.4 0.5 0.5 0.5 0.5 0.5
0.07 0.1 0.15 0.2 0.05 0.07 0.1 0.15 0.2 0.05 0.07 0.1 0.15 0.2
120 95 79 71 165 145 118 94 80 205 166 138 120 100
Flow patterns inside the internal mixing chamber have strong effects on spray characteristics and are discussed here. The flow patterns are usually correlated with empirical two-phase maps. For effervescent atomizers [25–28], the Baker’s map illustrated in Fig. 3a is often used to predict internal mixing flow patterns for a vertical fully developed flow. However, the Baker’s map was seldom used in the internal-mixing twin-fluid atomizer. And other maps are commonly used, such as the Oshinowo and Charles’ map [19] (illustrated in Fig. 3b). Both maps were used in present work to identify a better prediction of flow patterns for the internal-mixing atomizer. When p is 0.3 MPa and GLR is 0.02, the Baker’s map implies the flow pattern should be plug flow. As GLR increases, the flow pattern transforms to plug-annular flow, annular flow and further to dispersed flow. At the same p, the Oshinowo and Charles’ map implies the flow pattern is bubbly–slug when GLR is 0.02 and 0.05. As the GLR increases to 0.1 and 0.2, the flow transforms to bubbly-falling bubbly film flow and further to bubbly film flow, which can roughly be regarded as bubbly–annular flow and annular flow, respectively. Visualizations of flow patterns inside the internal mixing chamber at different cases are shown in Fig. 4. Fig. 4b shows flow pattern’s change with GLR at a p of 0.3 MPa. When GLR is 0.02, large water droplets were generated due to water slugs’ strong impinging on the wall. The observation indicates that the flow pattern was water slug flow. Increasing GLR to 0.05, less water slugs were generated in the central region of the chamber. The flow was very unsteady and central water slugs had strong oscillations. The majority of the water slipped downward as a thin film. The gas bubbles coalesced and flowed between the central water slugs and outer water film. This flow pattern was slug-annular. As GLR increased to 0.1, few water slugs were observed and the water film became thicker. Finally, the flow developed to steady annular flow when GLR is 0.2. Although both indications of two maps departs from the actual flow patterns, the Oshinowo and Charles’ map indicates better and is recommended to predict the flow regimes inside the internal-mixing chambers. Fig. 4 also shows the effect of p on gas–liquid mixing process. At lower GLR of 0.02, stronger impinging between the central liquid jet and the wall as well as stronger back-mixing inside the chamber were observed at a higher p due to a larger liquid mass flow rate. At a fixed GLR of 0.1, the flow pattern in the chamber was slug-annular flow given p of 0.1 MPa, and was much closer to annular flow given p of 0.3 MPa. When the pressure increased to 0.5 MPa, the flow pattern transformed to annular flow. These observations show that a steady annular flow is easier to be acquired at a higher pressure given a fixed GLR. However, it shall be emphasized that two-phase mixing process in the chamber is dominated by GLR. 3.2. SMD-GLR profiles at different p and axial locations Fig. 5 shows droplet SMD-GLR profiles at different p and x⁄, where x⁄ is the ratio of axial distance, x, to the diameter of exit orifice, d0. In all cases, SMD firstly decreased sharply with the increase
309
Z. Li et al. / Fuel 97 (2012) 306–314
Fig. 3. Empirical maps for vertical downward flow (p = 0.3 MPa). (a) Baker’s map and (b) Oshinowo and Charles’ map.
GLR= 0.02
0.05
0.10
0.20
0.10
0.20
(a) p=0.1MPa
GLR= 0.02
0.05
(b) p=0.3MPa
GLR= 0.05
0.10
(c) p=0.5MPa Fig. 4. Visualizations of two-phase flow in the internal mixing chamber at various GLR and p.
of GLR. When GLR continued increasing, SMD tended to decrease smoothly. Effect of operating pressure, p, on SMD is complex. At a low GLR of 0.02, when p increased from 0.1 MPa to 0.3 MPa, the droplet SMD decreased dramatically at all measuring locations. This can be explained by change of flow patterns discussed in Section 3. Given a fixed GLR of 0.02, the flow pattern in the chamber was slug flow at a low p of 0.1 MPa. Water’s breakup was weak due to a low water momentum. When p increased to 0.3 MPa, liquid momentum increased so that liquid slugs impacted on the
chamber wall more intensively. Momentum exchange between air and water was also strengthened, and more fine droplets came into formation. Both visualization and SMD measurement show that increase of p has an obvious positive effect on atomization at a low GLR of 0.02. When GLR is larger than 0.02, effects of p on SMD vary with x⁄. At the very near nozzle region of x⁄ = 3.3, as is shown in Fig. 5a, droplets SMD distribution was dominated by the primary atomization [23,29], and decreased monotonically with increase of
310
Z. Li et al. / Fuel 97 (2012) 306–314
Fig. 5. SMD change with p and GLR at various x⁄.
pressure. When x⁄ was in a range of 10–83 (Fig. 5b–e), the beneficial effect of an increase in p on promoting finer atomization was not so evident as the effect of an increase in GLR. This is not in accordance with the general rule that higher p improves atomization,
as is reported by Sovani et al. [29]. In past studies that showed higher p helps promote finer droplets, SMD were generally measured at larger x⁄, such as 100 (Jedelsky et al. [23]), 294 (Sataphthy et al. [30]) and 394–492 (Schmidt and Sojka [31]). In such fully
Z. Li et al. / Fuel 97 (2012) 306–314
311
Fig. 6. SMD change with p and GLR at various r⁄.
developed regions, both primary atomization and secondary atomization have finished. At a fixed GLR, injection velocity increases with pressure in this region, so that droplet breakup is stronger.
This trend was also repeated in Fig. 5f, although pressure effect is not as strong as GLR. When x⁄ is not large, SMD distribution depends on primary atomization and secondary atomization
312
Z. Li et al. / Fuel 97 (2012) 306–314
Fig. 7. Radial SMD and velocity distributions at different x⁄ when p is 0.1 MPa, GLR is 0.07. (a) SMD-r⁄ profile and (b) axial velocity-r⁄ profile.
Fig. 8. Possibility density distributions of droplet size at different x⁄. (a) GLR = 0.1, p = 0.3 MPa and (b) GLR = 0.1, p = 0.5 MPa.
effects. As pressure increases, liquid mass flow rate increases as well. At an unsteady flow pattern, more large droplets will be formed at higher pressure. In this case, flow patterns will have strong effects on SMD distribution. In order to explain this circumstance, the concept of undeveloped atomization region is introduced to represent the region where x⁄ is in 10–83. Visualization in Fig. 4 shows when GLR was 0.05, thin water films, which are very important for a fine droplets formation, were formed even
at a low p of 0.1 MPa. With an increase of p to 0.3 MPa, central water jet broke up extensively and contributed to formation of a thinner water film. As p increased to 0.5 MPa, breakup of water jet flow was so strong that the water film was destroyed, and thicker water films had more chance to appear. The change of water film was also demonstrated by SMD measurement. It shall be noted that the change tendencies of water film thickness with p were different at different GLR. This non-monotonic change of water film
Z. Li et al. / Fuel 97 (2012) 306–314
thickness can well explain SMD’s similar change tendency with p in the undeveloped atomization region. Fig. 4 also shows that at a constant p, increase of GLR could transform more water slugs into films and generate thinner films. This can explain why GLR has stronger promoting effect on atomization than p. According to Fig. 5b–e, an optimized pressure ranged from 0.2 MPa to 0.3 MPa were identified for atomization in the undeveloped region. When x⁄ increased to 133, which is in the developed region (100–492) discussed above, SMD’s variation with p followed the rule that higher p improves atomization. The effect of an increase in p was more pronounced at low p (0.1 MPa) than at high (>0.2 MPa). At this axial location, droplets coalescence became stronger due to the axial velocity’s decrease. 3.3. SMD-GLR profiles at different p and radial locations Droplet SMD-GLR profile at different p and r⁄ are depicted in Fig. 6. The results were acquired at a fixed x⁄ of 33, where nondimensional spray cone radius varied from 3.8 to 4.3 with p and GLR. When r⁄ is less than 3 (Fig. 6a–d), SMD decreased smoothly with the increase of GLR. However, this trend was not true when r⁄ was larger than non-dimensional spray cone radius, as is shown in Fig. 6e and f. The irregular SMD change with GLR in the outside jet region was observed at all the other x⁄. Jedelsky et al. [23] also observed this SMD irregular change with GLR in the outside jet region. p Also showed different effects on SMD at different r⁄. In the spray cone region (Fig. 6a–d), SMD’s change with p was non-monotonic, which has been explained previously. In the outside spray cone region, as shown in Fig. 6e and f, an increase in p had a remarkable effect on promoting finer atomization given a fixed GLR. The possible reason is that increasing of mass flow rate enlarges spreading angle at a larger p, which was observed in the experiments. SMD and velocity radial profiles at different x⁄ when GLR = 0.07 and p = 0.3 MPa, are shown in Fig. 7. SMD’s radial profile was axialsymmetric and in parabolic shape. At all axial distances, axial velocity profile shows a typical Gaussian distribution, which is the characteristic of a free turbulent jet. Comparing Fig. 7a and b, it is apparent that a larger axial velocity results in a smaller droplet SMD value. Since in high velocity region, stronger droplets breakup compensates droplets coalescence effect. Lund and Sojka [32] and Sutherland et al. [33] investigated the radial variation of SMD and got the similar observations. SMD radial distribution in Fig. 7a shows that larger SMD was observed at any r⁄ when x⁄ increased. Thus, the decrease of smaller droplets did not result from the radial convection of smaller droplets but from coalescence of smaller droplets. 3.4. droplet size Possibility Density Function (PDF) distribution In order to investigate droplet breakup and coalescence effects on atomization, PDF of droplet size at different axial locations given GLR of 0.1 and p of 0.3, 0.5 MPa, were analyzed, as shown in Fig. 8. At the very near nozzle region (x⁄ = 3.3), more fine droplets with diameters less than 10 lm were generated due to strong gas– liquid interactions. Due to unsteadiness of twin-fluid flow, large droplets of 200 lm were also observed. The maximum collected droplet size was 220 lm. As x⁄ increases, the droplet size range shrink, and the maximum droplet size decreases continually and become less than 200 lm when x⁄ is 133. The maximum possibility density also decreased with increase of x⁄. At both operating pressures, the droplet size where maximum possibility density appears moves slowly from (20, 30 lm) to (30, 40 lm) as x⁄ increased. The shape of the distribution became smoother at the developed region. This trend quantitatively represented coalescence of small
313
droplets and breakup of large droplets. As x⁄ increased, the possibility density of droplets with diameter less than 20 lm decreased sharply and possibility density of droplets with larger diameter increased. The decrease of smaller droplets means the coalescence of small droplets into larger droplets. Meanwhile, the possibility density of the very large droplets with diameter larger than 200 lm disappeared, which represents that large droplets break up as the jet develops. Fig. 8 also reveals the atomization mechanisms of this internalmixing twin-fluid atomizer in the outside nozzle region. The atomization process can be divided into two major steps. The primary atomization process includes gas–liquid interaction and momentum exchange, the formation of liquid films and pieces near the nozzle, which is due to the fierce interaction in the exit orifice and the gas expansion upon getting out of the exit orifice. In this very unsteady process, fine droplets are generated, as well as very large droplets. The secondary atomization process includes breakup of liquid films, liquid ligaments and larger droplets due to the slide velocity between liquid and gas phases. In this step, droplets will merge as well. When droplet coalescence is stronger than droplet breakup effect, droplet size will increase. Based on this knowledge and droplet size measurement shown in Figs. 5 and 8, the optimization of the primary and secondary atomization was acquired in the undeveloped atomization region. When x⁄ increased to 133, droplet diameter increased due to the stronger droplet collision and coalescence effects. 4. Conclusions In present work, an atomization test rig was built to investigate internal two-phase flow and atomization characteristics of an internal-mixing twin-fluid atomizer. High speed CCD was employed to visually investigate two-phase mixing and flow patterns in the mixing chamber. Droplet size and velocity profile outside the atomizer were acquired through PDA measurements under different GLR and operating pressures. Results show that internal twophase mixing has important influence on spray characteristics. GLR dominated the two-phase flow patterns in the mixing chamber. As GLR increases, the flow pattern transforms from liquid slug to slug-annular, further to steady annular flow. Annular flow generates thinner liquid films and finer droplets. GLR has a strong positive effect on atomization process at any operating pressure and position inside the jet flow. As GLR increases, droplet SMD decreases sharply at first, and then decreases slowly. In the undeveloped region, as p increased, a good correlation was observed between flow pattern transformation in internal mixing chamber and droplet SMD distribution. An optimum operating pressure appeared ranged from 0.2 MPa to 0.3 MPa for atomization, since thicker water films formed at a higher pressure. In the developed region, pressure was proved to have a positive effect on atomization. However, pressure played a minor effect on SMD distribution comparing with GLR. In radial direction, axial SMD got the minimum value and increased with radial distance. Larger SMD was observed in the lower velocity region, where droplet breakup was weak and droplets coalescence was strong. Droplet size PDF distribution quantitatively represented droplet breakup and coalescence effects. As x increased, possibility density of both fine droplets and large droplets decreased at a constant operating pressure. Acknowledgements This material is based upon work supported by National Nature Science Foundation of China (No. 51006062) and Mitsubishi Heavy Industries, Ltd.
314
Z. Li et al. / Fuel 97 (2012) 306–314
References [1] Lefebvre AH. Atomization and Sprays. New York: Hemisphere Publishing Corporation of Taylor & Francis; 1989. [2] Dombrowski N, Johns WR. The aerodynamic instability and disintegration of viscous liquid sheets. Chem. Eng. Sci. 1963;18(3):203. [3] Mullinger PJ, Chigier NA. The design and performance of internally-mixing multijet twin-fluid atomizers. J. Inst. Fuel. 1974;47:251–61. [4] Ferreira G, Garciia JA, Barreras F, Lozano A, Lincheta E. Design optimization of twin-fluid atomizers with an internal mixing chamber for heavy fuel oils. Fuel Process Technol. 2009;90(2):270–8. [5] Zhou YG, Zhang MC, Yu J, Zhu X, Peng J. Experimental investigation and model improvement on the atomization performance of single-hole Y-jet nozzle with high liquid flow rate. Powder Technol. 2010;199(3):248–55. [6] Liu XJ, Zhang WR, Park TJ. Modelling coal gasification in an entrained flow gasifier. Combust. Theor. Model. 2001;5(4):595–608. [7] Kushari A, Neumeier Y, Israeli O, Lubarsky E, Zinn BT. Internally mixed liquid injector for active control of atomization process. J. Propul. Power. 2001;4(17):878–82. [8] Kim S, Kondo S, Nishida K, Hiroyasu H. Effects of mixing chamber geometry and flow on spray characteristics from an internal mixing twin-fluid atomizer. Int. J. Fluid Mech. Res. 1997;24(1–3):76–87. [9] Nguyen DA, Rhodes MJ. Producing fine drops of water by twin-fluid atomisation. Powder Technol. 1998;99(3):285–92. [10] Kushari A. Effect of injector geometry on the performance of an internally mixed liquid atomizer. Fuel Process Technol. 2010;91(11):1650–4. [11] Tamaki N, Shimizu M, Hiroyasu H. Atomization of high-viscous liquid jet by internal mixing twin-fluid atomizer. Nihon Kikai Gakkai Ronbunshu. B Hen/ Trans. Jpn. Soc. Mech. Eng. 2005;71(712):3017–24. [12] Broniarz-Press L, Ochowiak M, Rozanski J, Woziwodzki S. The atomization of water-oil emulsions. Exp Therm Fluid Sci. 2009;33(6):955–62. [13] Ju SC, Li XW. Experimental study on internal mixing sonic flow air assist atomizer for heavy oils. In: International Gas Turbine and Aeroengine Congress and Exposition, June 11, 1990–June 14, 1990. Brussels, Belg: ASME; 1990. p. T6. [14] Lal S, Kushari A, Gupta M, Kapoor JC, Maji S. Experimental study of an air assisted mist generator. Exp Therm. Fluid Sci. 2010;34(8):1029–35. [15] Chin JS, Lefebvre AH. A design procedure for effervescent atomizers. J. Eng. Gas. Turb. Power. 1995;117(2):266–71. [16] Kufferath A, Wende B, Leuckel W. Influence of liquid flow conditions on spray characteristics of internal-mixing twin-fluid atomizers. Int. J. Heat Fluid Flow 1999;20(5):513–9.
[17] Karnawat J, Kushari A. Controlled atomization using a twin-fluid swirl atomizer. Exp. Fluids 2006;41(4):649–63. [18] Karnawat J, Kushari A. Spray evolution in a twin-fluid swirl atomizer. Atomization Sprays 2008;18(5):449–70. [19] Lefebvre AH, Chin JS. Flow patterns in internal-mixing, twin-fluid atomizers. Atomization Sprays 1993;3(4):463–75. [20] Ferreira G, Barreras F, Lozano A, Garcia JA, Lincheta E. Effect of the inner twophase flow on the performance of an industrial twin-fluid nozzle with an internal mixing chamber. Atomization Sprays. 2009;19(9):873–84. [21] Whitlow JD, Lefebvre AH. Effervescent atomizer operation and spray characteristics. Atomization Sprays 1993;3(2):137–56. [22] Panchagnula MV, Sojka PE. Spatial droplet velocity and size profiles in effervescent atomizer-produced sprays. Fuel 1999;78:729–41. [23] Jedelsky J, Jicha M, Slama J, Otahal J. Development of an effervescent atomizer for industrial burners. Energy Fuel 2009;23:6121–30. [24] Rasbash DJ, Rogowski ZW, Stark GWV. Mechanisms of extinction of liquid fires with water sprays. Combust. Flame 1960;4:223–34. [25] Sovani SD, Chou E, Sojka PE, Gore JP. High pressure effervescent atomization: effect of ambient pressure on spray angle. Fuel 2001;80:427–35. [26] Jedelsky J, Jicha M. Unsteadiness in effervescent sprays: a new evaluation method and the influence of operational conditions. Atomization Sprays 2008;18(1):49–83. [27] Huang X, Wang X, Liao G. Visualization of two phase flow inside an effervescent atomizer. J. Visualization 2008;11(4):299–308. [28] Liu M, Duan YF, Zhang TN. Evaluation of effervescent atomizer internal design on the spray unsteadiness using a phase/doppler particle analyzer. Exp. Therm. Fluid Sci 2010;34(6):657–65. [29] Sovani SD, Sojka PE, Lefebvre AH. Effervescent atomization. Prog. Energy Combust. 2001;27(4):483–521. [30] Sataphthy MR, Sovani SD, Sojka PE, Gore JP, Eckerle WA. The Effect of Ambient Density on the Performance of an Effervescent Atomizer Operating in the MPa Injection Pressure Range. Pittsburgh, PA: The Combustion Institute; 1998. p. 76–80. [31] Schmidt UT, Sojka PE. Air-assist pressure-swirl atomization. Atomization Sprays 1999;9(2):173–92. [32] Lund MT, Sojka PE. Effervescent atomization at low mass flow-rates. Part 2: the structure of the spray. In: Proceedings of the 5th annual conference on liquid atomization and spray systems, San Ramon, CA: 1992. p. 233–7. [33] Sutherland JJ, Sojka PE, Plesniak MW. Entrainment by ligament-controlled effervescent atomizer-produced sprays. Int. J. Multiphase Flow 1997;23(5): 865–84.
Contents lists available at SciVerse ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Mixing and atomization characteristics in an internal-mixing twin-fluid atomizer Zhouhang Li a, Yuxin Wu a,⇑, Chunrong Cai a, Hai Zhang a, Yingli Gong a, Keiji Takeno b, Kazuaki Hashiguchi b, Junfu Lu a a b
Key Laboratory for Thermal Science and Power Engineering of Ministry of Education, Department of Thermal Engineering, Tsinghua University, Beijing 100084, China Nagasaki R&D Center, Mitsubishi Heavy Industries, Ltd., Japan
a r t i c l e
i n f o
Article history: Received 17 June 2011 Received in revised form 14 November 2011 Accepted 6 March 2012 Available online 18 March 2012 Keywords: Twin-fluid atomization Internal mixing GLR SMD distribution
a b s t r a c t Twin-fluid atomizers have been successfully used in many industrial applications. This paper presents experimental studies on internal mixing and atomization in a water–air internal-mixing atomizer. Two-phase mixing process and flow patterns in the internal mixing chamber were visually studied through high speed CCD. Observation reveals that internal mixing was dominated by Gas to Liquid mass Ratio (GLR). As GLR increased, the flow patterns changed from slug flow to annular flow. The Oshinowo and Charles’ map can be used to predict the flow patterns for the designed atomizer. Droplet Sauter Mean Diameter (SMD) spatial distributions were measured with Phase Doppler Analyzer (PDA) at different operating conditions. Droplet SMD decreased with the increase of GLR at all operating pressure and locations. In the undeveloped region, a close relationship was observed between flow pattern transformation in internal mixing chamber and droplet SMD distribution, and there was an optimized pressure ranging from 0.2 MPa to 0.3 MPa for atomization since liquid films became thicker under a higher pressure. In the developed region, pressure promoted to generate finer atomization. Possibility Density Function (PDF) distribution of droplet size at different axial locations was analyzed to quantitatively represent the effect of droplet coalescence and breakup. As axial distance increased, PDF of both fine droplets and large droplets decreased. The particle size with the maximum PDF increased with the axial distance as well. The results imply that best atomization performance was acquired in the undeveloped region. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Twin-fluid atomizers have been widely used in many industrial applications such as oil burners and spray drying [1]. This kind of atomizer generates high relative velocity between gas and liquid by employing a relatively high gas velocity, which helps the formation of the smaller liquid sheets and ligaments. Then fine droplets come into formation due to sheet breakup. This process can be explained by the wavelengths that grow on the surface of the sheet, which are affected by surface tension, aerodynamic forces, and liquid viscosity [2]. Meanwhile, the introduction of gas phase may help dispersion of liquid and prevent from droplets coalescence. There are many types of twin-fluid atomizers, such as internal mixing [3,4], medium mixing (‘Y’ type) [4,5] and outside mixing atomizers [6]. In a boiler plant, the generally design of the oil/heavy oil burners are internal mixing or medium mixing atomizers. The characteristic of the medium mixing atomizers is that liquid and gas are mixed inside the atomizer within a limited space before they are injected out. The advantages of such kind of ‘Y’ type ⇑ Corresponding author. Tel./fax: +11 8610 62788523. E-mail address: [email protected] (Y. Wu). 0016-2361/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2012.03.006
atomizer are that it can be operated either keeping a constant steam-to-fuel flow rate ratio or a fixed fuel-to-steam pressure ratio. However, internal mixing process shall be strengthened further so that finer droplets can be acquired at a low GLR. Ferreira [4] has indicated that adding an internal mixing chamber may improve the performance of ‘Y’ type atomizer. In an internal mixing chamber, liquid and gas are mixed and interacted intensively before they are discharged out. Atomization in the internal mixing atomizer can be explained as follows [7]: liquid has to share flow area with gas, resulting in acceleration of liquid and formation of finer droplets. In addition, relative motion between two phases increases the flow instability at the interface, so that ligaments and sheets formation is accelerated. It is well recognized that atomizer geometry [4,8–10], fluid physical properties [11,12] and operating conditions [13,14] such as GLR and injection pressure have effects on the mixing process as well as atomization of internal mixing atomizers. Nguyen and Rhodes [9] found that length and diameter of the mixing chamber have little effect on atomization during a range of 20–45 mm and 1.7–2.3 mm, respectively. However, Kushari [10] showed that droplet size decreases with increase of mixing chamber length. Both Ferreira’s [4] and Kushari’s [10] results show that the ratio of the exit orifice area to the air injection area influences spray mean drop size in internal
307
Z. Li et al. / Fuel 97 (2012) 306–314
Nomenclature A di,l d0 FrTP g G GLR L _ m Q p r r⁄ SMD x
0:25 0:75 l rw factor, A ¼ l l qql rl w w mixing chamber’s inlet diameter (mm) exit orifice diameter (mm) 2 a þQ l Þ two-phase Froude number, FrTP ¼ ðQ 2 3 D gð p =4Þ gravitational constant (m/s2) gas mass flux (kg/m2 s1) Gas to Liquid mass Ratio liquid mass flux (kg/m2 s1) mass flow rate (kg/h) volumetric flow rate (m3/s) operating pressure (MPa) radial distance from spray axis (mm) non-dimensional radial distance, r/d0 Sauter Mean Diameter, D32 (lm) axial distance from the exit orifice (mm)
mixing twin-fluid atomizers, the same with effervescent atomizers [15]. Slant angle and specific shape of the liquid ports have little effect on atomization [4]. Broniarz-Press et al. [12] studied atomization of internal mixing atomizer using water and different oil emulsions, concluding that physical properties of liquid have significant influence on the spray characteristics. Flow patterns in internal-mixing twin-fluid atomizer were also investigated. Kufferath et al. [16] studied atomizer’s performance under different flow conditions of the liquid jet leaving the inlet port, including laminar and turbulent flows, and found that flow conditions have a strong influence on droplet mean size. Karnawat and Kushari [17,18] investigated different spray patterns in the developed jet regime and concluded correlations between spray patterns and operating conditions. The mixing process within the internal chamber is important for internal-mixing atomizer’s performance. Based on classical two-phase flow pattern maps, Lefebvre and Chin [19] investigated different flow patterns within the mixing chamber. Kim et al. [8] studied effects of GLR on flow patterns within the internal chamber and they reported breakup of water column was complicatedly changed with the increase of GLR. Ferreira et al. [20] found that flow patterns within the chamber change with internal mixing chamber geometry configuration. Although research work listed previously give useful information for internal mixing atomization design and research, the correlation between flow patterns in internal mixing chamber and droplet mean diameter has not been investigated in previous literatures. Operating pressure’s effect on internal mixing was seldom studied experimentally. Besides, most previous researchers (Whitlow and Lefebvre [21], Panchagnula and Sojka [22], and Jedelsky et al. [23]) reported droplet SMD at axial locations where are in the far-field region and spray is fully developed, while droplet SMD distribution in the near nozzle region was paid little attention. For oil burners, it’s important to acquire the axial distance where droplets have the smallest SMD in order to get an optimized condition. Moreover, spray with high velocity sometimes cools down excessively the reaction zone, leading to local flame extinction [24]. A reasonable atomization flow field is very important for oil burning. In present work, liquid–gas flow patterns in a water-inside–airoutside internal mixing twin-fluid atomizer were visually investigated through high speed CCD. The SMD spatial distribution and velocity profile in the near nozzle region and fully developed region were measured through Phase Doppler Analyzer (PDA). Effects of GLR and pressure on SMD spatial distribution were studied. The correlation between flow patterns in internal mixing chamber and droplet SMD was discussed. Droplet breakup and
x⁄
non-dimensional axial distance, x/d0
Greek letters k factor, k ¼ ðqg ql =qa qw Þ0:5 h i1=3 l q w factor, w ¼ rrwl l l qw w
q r l
l
density (kg/m3) surface tension (N/m) dynamic viscosity (N/s m2)
Subscripts a air l liquid g gas w water
coalescence effects were quantitatively expressed by SMD spatial distribution and possibility density distributions. 2. Experimental 2.1. internal-mixing twin-fluid atomizer and test rig system The geometry configuration of the internal-mixing twin-fluid atomizer is shown in Fig. 1. The atomizer is made by acrylic glass so that the internal mixing can be captured by high speed CCD. The overall length of the atomizer is 600 mm. The diameter of mixing chamber’s liquid inlet (di,l) is 4 mm. Liquid inlet is surrounded by four 1 mm air inlet holes with a slant angle of 45°. Atomizer’s exit orifice diameter, d0, is 3 mm. Fig. 2 shows the schematic drawing of the air–water atomization system. Both water and air were fed into the atomizer at an elevated operating pressure. A ball valve and a needle valve were installed on each route in order to control the flow rate of pressurized water and air. Metal tube rotameters with a maximum error of 1.5% were adopted to acquire the volume flow rate of water and air. 2.2. Measuring apparatus A high speed CCD, the PCO.dimax made by COOK Company, was used to capture the internal flow patterns and two-phase mixing process inside the chamber. The CCD has a high ISO value of 50,000. The minimum exposure time is 2 ls. Up to 1297 frames per second (fps) can be acquired with the size of 2016 2016 pixels for each frame. In this work, pictures and videos of the mixing processes were captured at an fps of 10,000 so that change of the two-phase flow can be analyzed in detail. A Phase Doppler Analyzer (PDA) system of Dantec Dynamics was used to measure droplet size and velocity distributions. The PDA was set in refractive scattering mode with the receiver positioned at 30° from the transmitter axis. The focal length for the
Fig. 1. Geometry configuration of the research atomizer.
308
Z. Li et al. / Fuel 97 (2012) 306–314
from 0.1 MPa to 0.5 MPa and GLR from 0.02 to 0.2. The detailed experimental cases are listed in Table 1. Under each case, droplet size and velocities were acquired at 88 spatial locations distributed at 8 different x, within a range of 10–400 mm. At each location, 2000 samples were acquired with PDA. 3. Results and discussion 3.1. Two-phase flow patterns and mixing process inside the mixing chamber
Fig. 2. Schematic of the experimental system. 1 – Pressurized water tank, 2 – adjusting valve, 3 – mass flowmeter, 4 – pressure meter, 5 – atomizer, 6 – observation tank, 7 – PDA laser lens, 8 – gas flow rotameter, 9 – water tank, and 10 – outlet.
transmitting lens was 800 mm and for the receiver was 600 mm. The maximum diameter which can be measured with this setup is 250 lm. Droplet distributions at different spatial locations were measured by adjusting a three dimensional auto control coordinate system with a size of 560 560 580 mm.
2.3. Experimental cases To conduct a complete analysis of the characteristic of the atomizer, experimental measurements under different parameters, such as operating pressure and GLR, are needed. In the designed atomizer, water and air were well mixed in the internal mixing chamber, and the pressure differential between water and air at the atomizer inlet was relatively small at all experimental cases. This is in accordance with the results of Jedelsky et al. [23]. In their studies on atomizers with an internal mixing chamber, pressure differential between gas and liquid was in a range of 3.3 to 10.6 kPa at different operation conditions and geometry configurations. The existence of the mixing chamber makes pressure differential between gas and liquid small enough to be neglected. Thus, operating pressure p represents the gauge pressure of the mixing chamber. Since pressure differential between water and air results in different mass flow rates of two phases at a given p, GLR is used to represent the effect of such pressure differential between liquid and gas routes. The experimental conditions covered the range of p
Table 1 Experimental cases of atomizer. Case No.
p (MPa)
GLR
_ l (kg/ m h)
Case No.
p (MPa)
GLR
_ l (kg/ m h)
1 2 3 4 5 6 7 8 9 10 11 12 13 14
0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.2 0.2 0.2 0.3 0.3
0.02 0.05 0.07 0.1 0.15 0.2 0.02 0.05 0.07 0.1 0.15 0.2 0.02 0.05
125 83 70 55 44 41 180 112 94 78 62 51 240 153
15 16 17 18 19 20 21 22 23 24 25 26 27 28
0.3 0.3 0.3 0.3 0.4 0.4 0.4 0.4 0.4 0.5 0.5 0.5 0.5 0.5
0.07 0.1 0.15 0.2 0.05 0.07 0.1 0.15 0.2 0.05 0.07 0.1 0.15 0.2
120 95 79 71 165 145 118 94 80 205 166 138 120 100
Flow patterns inside the internal mixing chamber have strong effects on spray characteristics and are discussed here. The flow patterns are usually correlated with empirical two-phase maps. For effervescent atomizers [25–28], the Baker’s map illustrated in Fig. 3a is often used to predict internal mixing flow patterns for a vertical fully developed flow. However, the Baker’s map was seldom used in the internal-mixing twin-fluid atomizer. And other maps are commonly used, such as the Oshinowo and Charles’ map [19] (illustrated in Fig. 3b). Both maps were used in present work to identify a better prediction of flow patterns for the internal-mixing atomizer. When p is 0.3 MPa and GLR is 0.02, the Baker’s map implies the flow pattern should be plug flow. As GLR increases, the flow pattern transforms to plug-annular flow, annular flow and further to dispersed flow. At the same p, the Oshinowo and Charles’ map implies the flow pattern is bubbly–slug when GLR is 0.02 and 0.05. As the GLR increases to 0.1 and 0.2, the flow transforms to bubbly-falling bubbly film flow and further to bubbly film flow, which can roughly be regarded as bubbly–annular flow and annular flow, respectively. Visualizations of flow patterns inside the internal mixing chamber at different cases are shown in Fig. 4. Fig. 4b shows flow pattern’s change with GLR at a p of 0.3 MPa. When GLR is 0.02, large water droplets were generated due to water slugs’ strong impinging on the wall. The observation indicates that the flow pattern was water slug flow. Increasing GLR to 0.05, less water slugs were generated in the central region of the chamber. The flow was very unsteady and central water slugs had strong oscillations. The majority of the water slipped downward as a thin film. The gas bubbles coalesced and flowed between the central water slugs and outer water film. This flow pattern was slug-annular. As GLR increased to 0.1, few water slugs were observed and the water film became thicker. Finally, the flow developed to steady annular flow when GLR is 0.2. Although both indications of two maps departs from the actual flow patterns, the Oshinowo and Charles’ map indicates better and is recommended to predict the flow regimes inside the internal-mixing chambers. Fig. 4 also shows the effect of p on gas–liquid mixing process. At lower GLR of 0.02, stronger impinging between the central liquid jet and the wall as well as stronger back-mixing inside the chamber were observed at a higher p due to a larger liquid mass flow rate. At a fixed GLR of 0.1, the flow pattern in the chamber was slug-annular flow given p of 0.1 MPa, and was much closer to annular flow given p of 0.3 MPa. When the pressure increased to 0.5 MPa, the flow pattern transformed to annular flow. These observations show that a steady annular flow is easier to be acquired at a higher pressure given a fixed GLR. However, it shall be emphasized that two-phase mixing process in the chamber is dominated by GLR. 3.2. SMD-GLR profiles at different p and axial locations Fig. 5 shows droplet SMD-GLR profiles at different p and x⁄, where x⁄ is the ratio of axial distance, x, to the diameter of exit orifice, d0. In all cases, SMD firstly decreased sharply with the increase
309
Z. Li et al. / Fuel 97 (2012) 306–314
Fig. 3. Empirical maps for vertical downward flow (p = 0.3 MPa). (a) Baker’s map and (b) Oshinowo and Charles’ map.
GLR= 0.02
0.05
0.10
0.20
0.10
0.20
(a) p=0.1MPa
GLR= 0.02
0.05
(b) p=0.3MPa
GLR= 0.05
0.10
(c) p=0.5MPa Fig. 4. Visualizations of two-phase flow in the internal mixing chamber at various GLR and p.
of GLR. When GLR continued increasing, SMD tended to decrease smoothly. Effect of operating pressure, p, on SMD is complex. At a low GLR of 0.02, when p increased from 0.1 MPa to 0.3 MPa, the droplet SMD decreased dramatically at all measuring locations. This can be explained by change of flow patterns discussed in Section 3. Given a fixed GLR of 0.02, the flow pattern in the chamber was slug flow at a low p of 0.1 MPa. Water’s breakup was weak due to a low water momentum. When p increased to 0.3 MPa, liquid momentum increased so that liquid slugs impacted on the
chamber wall more intensively. Momentum exchange between air and water was also strengthened, and more fine droplets came into formation. Both visualization and SMD measurement show that increase of p has an obvious positive effect on atomization at a low GLR of 0.02. When GLR is larger than 0.02, effects of p on SMD vary with x⁄. At the very near nozzle region of x⁄ = 3.3, as is shown in Fig. 5a, droplets SMD distribution was dominated by the primary atomization [23,29], and decreased monotonically with increase of
310
Z. Li et al. / Fuel 97 (2012) 306–314
Fig. 5. SMD change with p and GLR at various x⁄.
pressure. When x⁄ was in a range of 10–83 (Fig. 5b–e), the beneficial effect of an increase in p on promoting finer atomization was not so evident as the effect of an increase in GLR. This is not in accordance with the general rule that higher p improves atomization,
as is reported by Sovani et al. [29]. In past studies that showed higher p helps promote finer droplets, SMD were generally measured at larger x⁄, such as 100 (Jedelsky et al. [23]), 294 (Sataphthy et al. [30]) and 394–492 (Schmidt and Sojka [31]). In such fully
Z. Li et al. / Fuel 97 (2012) 306–314
311
Fig. 6. SMD change with p and GLR at various r⁄.
developed regions, both primary atomization and secondary atomization have finished. At a fixed GLR, injection velocity increases with pressure in this region, so that droplet breakup is stronger.
This trend was also repeated in Fig. 5f, although pressure effect is not as strong as GLR. When x⁄ is not large, SMD distribution depends on primary atomization and secondary atomization
312
Z. Li et al. / Fuel 97 (2012) 306–314
Fig. 7. Radial SMD and velocity distributions at different x⁄ when p is 0.1 MPa, GLR is 0.07. (a) SMD-r⁄ profile and (b) axial velocity-r⁄ profile.
Fig. 8. Possibility density distributions of droplet size at different x⁄. (a) GLR = 0.1, p = 0.3 MPa and (b) GLR = 0.1, p = 0.5 MPa.
effects. As pressure increases, liquid mass flow rate increases as well. At an unsteady flow pattern, more large droplets will be formed at higher pressure. In this case, flow patterns will have strong effects on SMD distribution. In order to explain this circumstance, the concept of undeveloped atomization region is introduced to represent the region where x⁄ is in 10–83. Visualization in Fig. 4 shows when GLR was 0.05, thin water films, which are very important for a fine droplets formation, were formed even
at a low p of 0.1 MPa. With an increase of p to 0.3 MPa, central water jet broke up extensively and contributed to formation of a thinner water film. As p increased to 0.5 MPa, breakup of water jet flow was so strong that the water film was destroyed, and thicker water films had more chance to appear. The change of water film was also demonstrated by SMD measurement. It shall be noted that the change tendencies of water film thickness with p were different at different GLR. This non-monotonic change of water film
Z. Li et al. / Fuel 97 (2012) 306–314
thickness can well explain SMD’s similar change tendency with p in the undeveloped atomization region. Fig. 4 also shows that at a constant p, increase of GLR could transform more water slugs into films and generate thinner films. This can explain why GLR has stronger promoting effect on atomization than p. According to Fig. 5b–e, an optimized pressure ranged from 0.2 MPa to 0.3 MPa were identified for atomization in the undeveloped region. When x⁄ increased to 133, which is in the developed region (100–492) discussed above, SMD’s variation with p followed the rule that higher p improves atomization. The effect of an increase in p was more pronounced at low p (0.1 MPa) than at high (>0.2 MPa). At this axial location, droplets coalescence became stronger due to the axial velocity’s decrease. 3.3. SMD-GLR profiles at different p and radial locations Droplet SMD-GLR profile at different p and r⁄ are depicted in Fig. 6. The results were acquired at a fixed x⁄ of 33, where nondimensional spray cone radius varied from 3.8 to 4.3 with p and GLR. When r⁄ is less than 3 (Fig. 6a–d), SMD decreased smoothly with the increase of GLR. However, this trend was not true when r⁄ was larger than non-dimensional spray cone radius, as is shown in Fig. 6e and f. The irregular SMD change with GLR in the outside jet region was observed at all the other x⁄. Jedelsky et al. [23] also observed this SMD irregular change with GLR in the outside jet region. p Also showed different effects on SMD at different r⁄. In the spray cone region (Fig. 6a–d), SMD’s change with p was non-monotonic, which has been explained previously. In the outside spray cone region, as shown in Fig. 6e and f, an increase in p had a remarkable effect on promoting finer atomization given a fixed GLR. The possible reason is that increasing of mass flow rate enlarges spreading angle at a larger p, which was observed in the experiments. SMD and velocity radial profiles at different x⁄ when GLR = 0.07 and p = 0.3 MPa, are shown in Fig. 7. SMD’s radial profile was axialsymmetric and in parabolic shape. At all axial distances, axial velocity profile shows a typical Gaussian distribution, which is the characteristic of a free turbulent jet. Comparing Fig. 7a and b, it is apparent that a larger axial velocity results in a smaller droplet SMD value. Since in high velocity region, stronger droplets breakup compensates droplets coalescence effect. Lund and Sojka [32] and Sutherland et al. [33] investigated the radial variation of SMD and got the similar observations. SMD radial distribution in Fig. 7a shows that larger SMD was observed at any r⁄ when x⁄ increased. Thus, the decrease of smaller droplets did not result from the radial convection of smaller droplets but from coalescence of smaller droplets. 3.4. droplet size Possibility Density Function (PDF) distribution In order to investigate droplet breakup and coalescence effects on atomization, PDF of droplet size at different axial locations given GLR of 0.1 and p of 0.3, 0.5 MPa, were analyzed, as shown in Fig. 8. At the very near nozzle region (x⁄ = 3.3), more fine droplets with diameters less than 10 lm were generated due to strong gas– liquid interactions. Due to unsteadiness of twin-fluid flow, large droplets of 200 lm were also observed. The maximum collected droplet size was 220 lm. As x⁄ increases, the droplet size range shrink, and the maximum droplet size decreases continually and become less than 200 lm when x⁄ is 133. The maximum possibility density also decreased with increase of x⁄. At both operating pressures, the droplet size where maximum possibility density appears moves slowly from (20, 30 lm) to (30, 40 lm) as x⁄ increased. The shape of the distribution became smoother at the developed region. This trend quantitatively represented coalescence of small
313
droplets and breakup of large droplets. As x⁄ increased, the possibility density of droplets with diameter less than 20 lm decreased sharply and possibility density of droplets with larger diameter increased. The decrease of smaller droplets means the coalescence of small droplets into larger droplets. Meanwhile, the possibility density of the very large droplets with diameter larger than 200 lm disappeared, which represents that large droplets break up as the jet develops. Fig. 8 also reveals the atomization mechanisms of this internalmixing twin-fluid atomizer in the outside nozzle region. The atomization process can be divided into two major steps. The primary atomization process includes gas–liquid interaction and momentum exchange, the formation of liquid films and pieces near the nozzle, which is due to the fierce interaction in the exit orifice and the gas expansion upon getting out of the exit orifice. In this very unsteady process, fine droplets are generated, as well as very large droplets. The secondary atomization process includes breakup of liquid films, liquid ligaments and larger droplets due to the slide velocity between liquid and gas phases. In this step, droplets will merge as well. When droplet coalescence is stronger than droplet breakup effect, droplet size will increase. Based on this knowledge and droplet size measurement shown in Figs. 5 and 8, the optimization of the primary and secondary atomization was acquired in the undeveloped atomization region. When x⁄ increased to 133, droplet diameter increased due to the stronger droplet collision and coalescence effects. 4. Conclusions In present work, an atomization test rig was built to investigate internal two-phase flow and atomization characteristics of an internal-mixing twin-fluid atomizer. High speed CCD was employed to visually investigate two-phase mixing and flow patterns in the mixing chamber. Droplet size and velocity profile outside the atomizer were acquired through PDA measurements under different GLR and operating pressures. Results show that internal twophase mixing has important influence on spray characteristics. GLR dominated the two-phase flow patterns in the mixing chamber. As GLR increases, the flow pattern transforms from liquid slug to slug-annular, further to steady annular flow. Annular flow generates thinner liquid films and finer droplets. GLR has a strong positive effect on atomization process at any operating pressure and position inside the jet flow. As GLR increases, droplet SMD decreases sharply at first, and then decreases slowly. In the undeveloped region, as p increased, a good correlation was observed between flow pattern transformation in internal mixing chamber and droplet SMD distribution. An optimum operating pressure appeared ranged from 0.2 MPa to 0.3 MPa for atomization, since thicker water films formed at a higher pressure. In the developed region, pressure was proved to have a positive effect on atomization. However, pressure played a minor effect on SMD distribution comparing with GLR. In radial direction, axial SMD got the minimum value and increased with radial distance. Larger SMD was observed in the lower velocity region, where droplet breakup was weak and droplets coalescence was strong. Droplet size PDF distribution quantitatively represented droplet breakup and coalescence effects. As x increased, possibility density of both fine droplets and large droplets decreased at a constant operating pressure. Acknowledgements This material is based upon work supported by National Nature Science Foundation of China (No. 51006062) and Mitsubishi Heavy Industries, Ltd.
314
Z. Li et al. / Fuel 97 (2012) 306–314
References [1] Lefebvre AH. Atomization and Sprays. New York: Hemisphere Publishing Corporation of Taylor & Francis; 1989. [2] Dombrowski N, Johns WR. The aerodynamic instability and disintegration of viscous liquid sheets. Chem. Eng. Sci. 1963;18(3):203. [3] Mullinger PJ, Chigier NA. The design and performance of internally-mixing multijet twin-fluid atomizers. J. Inst. Fuel. 1974;47:251–61. [4] Ferreira G, Garciia JA, Barreras F, Lozano A, Lincheta E. Design optimization of twin-fluid atomizers with an internal mixing chamber for heavy fuel oils. Fuel Process Technol. 2009;90(2):270–8. [5] Zhou YG, Zhang MC, Yu J, Zhu X, Peng J. Experimental investigation and model improvement on the atomization performance of single-hole Y-jet nozzle with high liquid flow rate. Powder Technol. 2010;199(3):248–55. [6] Liu XJ, Zhang WR, Park TJ. Modelling coal gasification in an entrained flow gasifier. Combust. Theor. Model. 2001;5(4):595–608. [7] Kushari A, Neumeier Y, Israeli O, Lubarsky E, Zinn BT. Internally mixed liquid injector for active control of atomization process. J. Propul. Power. 2001;4(17):878–82. [8] Kim S, Kondo S, Nishida K, Hiroyasu H. Effects of mixing chamber geometry and flow on spray characteristics from an internal mixing twin-fluid atomizer. Int. J. Fluid Mech. Res. 1997;24(1–3):76–87. [9] Nguyen DA, Rhodes MJ. Producing fine drops of water by twin-fluid atomisation. Powder Technol. 1998;99(3):285–92. [10] Kushari A. Effect of injector geometry on the performance of an internally mixed liquid atomizer. Fuel Process Technol. 2010;91(11):1650–4. [11] Tamaki N, Shimizu M, Hiroyasu H. Atomization of high-viscous liquid jet by internal mixing twin-fluid atomizer. Nihon Kikai Gakkai Ronbunshu. B Hen/ Trans. Jpn. Soc. Mech. Eng. 2005;71(712):3017–24. [12] Broniarz-Press L, Ochowiak M, Rozanski J, Woziwodzki S. The atomization of water-oil emulsions. Exp Therm Fluid Sci. 2009;33(6):955–62. [13] Ju SC, Li XW. Experimental study on internal mixing sonic flow air assist atomizer for heavy oils. In: International Gas Turbine and Aeroengine Congress and Exposition, June 11, 1990–June 14, 1990. Brussels, Belg: ASME; 1990. p. T6. [14] Lal S, Kushari A, Gupta M, Kapoor JC, Maji S. Experimental study of an air assisted mist generator. Exp Therm. Fluid Sci. 2010;34(8):1029–35. [15] Chin JS, Lefebvre AH. A design procedure for effervescent atomizers. J. Eng. Gas. Turb. Power. 1995;117(2):266–71. [16] Kufferath A, Wende B, Leuckel W. Influence of liquid flow conditions on spray characteristics of internal-mixing twin-fluid atomizers. Int. J. Heat Fluid Flow 1999;20(5):513–9.
[17] Karnawat J, Kushari A. Controlled atomization using a twin-fluid swirl atomizer. Exp. Fluids 2006;41(4):649–63. [18] Karnawat J, Kushari A. Spray evolution in a twin-fluid swirl atomizer. Atomization Sprays 2008;18(5):449–70. [19] Lefebvre AH, Chin JS. Flow patterns in internal-mixing, twin-fluid atomizers. Atomization Sprays 1993;3(4):463–75. [20] Ferreira G, Barreras F, Lozano A, Garcia JA, Lincheta E. Effect of the inner twophase flow on the performance of an industrial twin-fluid nozzle with an internal mixing chamber. Atomization Sprays. 2009;19(9):873–84. [21] Whitlow JD, Lefebvre AH. Effervescent atomizer operation and spray characteristics. Atomization Sprays 1993;3(2):137–56. [22] Panchagnula MV, Sojka PE. Spatial droplet velocity and size profiles in effervescent atomizer-produced sprays. Fuel 1999;78:729–41. [23] Jedelsky J, Jicha M, Slama J, Otahal J. Development of an effervescent atomizer for industrial burners. Energy Fuel 2009;23:6121–30. [24] Rasbash DJ, Rogowski ZW, Stark GWV. Mechanisms of extinction of liquid fires with water sprays. Combust. Flame 1960;4:223–34. [25] Sovani SD, Chou E, Sojka PE, Gore JP. High pressure effervescent atomization: effect of ambient pressure on spray angle. Fuel 2001;80:427–35. [26] Jedelsky J, Jicha M. Unsteadiness in effervescent sprays: a new evaluation method and the influence of operational conditions. Atomization Sprays 2008;18(1):49–83. [27] Huang X, Wang X, Liao G. Visualization of two phase flow inside an effervescent atomizer. J. Visualization 2008;11(4):299–308. [28] Liu M, Duan YF, Zhang TN. Evaluation of effervescent atomizer internal design on the spray unsteadiness using a phase/doppler particle analyzer. Exp. Therm. Fluid Sci 2010;34(6):657–65. [29] Sovani SD, Sojka PE, Lefebvre AH. Effervescent atomization. Prog. Energy Combust. 2001;27(4):483–521. [30] Sataphthy MR, Sovani SD, Sojka PE, Gore JP, Eckerle WA. The Effect of Ambient Density on the Performance of an Effervescent Atomizer Operating in the MPa Injection Pressure Range. Pittsburgh, PA: The Combustion Institute; 1998. p. 76–80. [31] Schmidt UT, Sojka PE. Air-assist pressure-swirl atomization. Atomization Sprays 1999;9(2):173–92. [32] Lund MT, Sojka PE. Effervescent atomization at low mass flow-rates. Part 2: the structure of the spray. In: Proceedings of the 5th annual conference on liquid atomization and spray systems, San Ramon, CA: 1992. p. 233–7. [33] Sutherland JJ, Sojka PE, Plesniak MW. Entrainment by ligament-controlled effervescent atomizer-produced sprays. Int. J. Multiphase Flow 1997;23(5): 865–84.