Aerosol and Air Quality Research, 14: 1675–1684, 2014 Copyright © Taiwan Association for Aerosol Research ISSN: 1680-8584 print / 2071-1409 online doi: 10.4209/aaqr.2013.09.0303 Study and Application of a Cyclone for Removing Amine Droplets from Recycled Hydrogen in a Hydrogenation Unit Liang Ma1, Jingping Wu1, Yanhong Zhang1, Qisong Shen2, Jianping Li2, Hualin Wang1* 1 State Environmental Protection Key Laboratory of Environmental Risk Assessment and Control on Chemical Process, East China University of Science and Technology, 130 Meilong Road, Shanghai, 200237, China 2 Shanghai Huachang Environmental Protection Co. Ltd., 188 Maoting Road, Shanghai, 201611, China

ABSTRACT

Recycled hydrogen is fed into a hydrogenation unit using a compressor with oil gas, amine droplets, and other hazardous gases at a super-high operating pressure after having been passed through a desulfurization system. This study developed a gas-liquid cyclone for removing amine droplets in recycled hydrogen based on computational fluid dynamics (CFD), an experimental study and industrial application. The results showed that the highest separation efficiency of the gas-liquid cyclone was 94.8% (experiment) and 97% (CFD) under 20°C and atmospheric pressure. Sampling and analysis showed that the highest separation efficiency for industrial application was 99.9% under 50°C and 10.7 MPa. The device purified the recycled hydrogen, and reduced both raw material losses and desulfurizer consumption. The energy consumption of the cycled hydrogen compressor and the pollution discharge of the hydrogenation unit were also reduced. Keywords: Amine droplet; Desulfurization; Gas-liquid cyclone separation; Hydrogenation unit. INTRODUCTION

The hydrogenation unit, an agent of technology in the petroleum refining industry, is one of the main technologies used for the deep processing of heavy-fraction oil. In the hydrogenation reaction, the heavy oil undergoes hydrogenation, cracking, and isomerization by catalyst at a high pressure and temperature (> 10 MPa, 400°C). The heavy oil is then changed into gasoline, kerosene, diesel oil, liquefied petroleum gas, dry gas, etc.

Amine droplet (i.e., N-methyl diethanolamine) is commonly produced once the desulfurized recycled hydrogen in the hydrogenation unit passes through the desulfurizing tower. Before entering the cycled hydrogen compressor, the recycled hydrogen undergoes gas-liquid separation, which is generally conducted with a gravity settler. However, the separation accuracy and separation efficiency of gravity settlers are very low. Thus, the recycled hydrogen still contains a large quantity of amine droplet after settlement separation. This phenomenon is not only increases desulfurizer consumption and raw material loss but also causes environmental pollution. It will also affect * Corresponding author.

Tel.: +86-021-64251894 E-mail address: samwhl@163.com, maliang3678@163.com

the long-term operation of the compressor and other key equipment. Therefore, new approaches and equipment that can effectively separate amine droplet from the recycled hydrogen at high pressure are urgently needed.

Compared with bulky and expensive gravity settlers, the cyclone separator is simple, compact, light, and inexpensive (Movafaghian et al., 2000; Wang et al., 2010), and it has higher separation accuracy and separation efficiency. In recent years, the industry has shown keen interest in the development and application of cyclone separators. Earni et al. (2003) used slug detection as a tool for the predictive control of gas-liquid cylindrical cyclone (GLCC) compact separators. Tsai et al. (2004, 2007) used an axial flow cyclone to remove nanoparticles at low pressures or vacuum conditions. Lin et al. (2013) designed a high-efficiency wet electrocyclone for removing fine and nanosized particles. Zhao et al. (2004, 2005) studied the different inlet configurations of a cylinder-on-cone cyclone.

Wang et al. (2003) designed a novel GLCC capable of separating liquid from a wet gas stream. Molina et al. (2008) carried out a detailed experimental investigation on the novel GLCC for low and high pressure conditions. Su (2006) and Mao (2006) used a three-dimensional particle dynamics analyzer to study the flow in a square-shaped cyclone separator that was designed for large circulating fluidized bed application. Similarly, Su et al. (2011) and Diao et al. (2009) studied the numerical simulation of the effect of inlet and exhaust configuration on a square-shaped cyclone. Chen et al. (2007), Hsu et al. (2005), and Tsai et

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1676

al. (2008, 2011, 2012) applied a modified axial flow cyclone to assess the exposure of nano-sized and respirable particles at different workplaces. Zhao et al. (2010, 2012) predicted the cylinder-on-cone cyclone separation efficiency and pressure drop by developing two different models. The cylinder-on-cone cyclone was also used in Chemical Looping Combustion (CLC) process (He et al., 2007) and Flidised Bed Combustion (FBC) boiler process (Gu et al., 2013).

Many studies on cyclone separators and their applications have been carried out. However, no study has explored the actual industrial application of a cyclone separator in high-pressure (> 10 MPa) and sulfur-hydrogen conditions as well as its long-term operation in industrial settings. In the present study, a gas-liquid cyclone was developed for removing amine droplet by simulation and experiment. The cyclone separator was set in the top of the desulfurization tower for the first time. METHODS Simulation

A cylinder-on-cone cyclone with diameter of 75 mm was designed in this study. The cyclone geometry was showed in Fig. 1. The detailed dimensions of the cyclone are listed in Table 1.

The multiphase flow in a cyclone is quite complicated. In this work, separation efficiency and velocity fields of the cyclone with different inlet velocity were numerically simulated using FLUENT 6.3.26 computational fluid dynamics (CFD) software. Some works of simulation of cyclone using the Reynolds stress model (RSM) have been reported to be suitable for modeling the flow in the cyclone in recent years (Safikhani et al., 2010; Elsayed et al., 2011; Qiu et al., 2012). Therefore, the Reynolds stress model (RSM) of the cyclone internal turbulent flow simulation is used in this paper. Gambit, which is the main preprocessor of Fluent, is used to create geometry, meshing and specifying boundary types of cyclone. Boundary types are considered as inlet mass flow rate, overflow and underflow outlet pressure. The amine droplet concentration of feed aerosol

Fig. 1. Cyclone structure (a) and the angular positions (b).

Table 1. Dimensions (mm) of the cyclone.

d a b di dc S H1 H2 H3 In

75 45 30 35 38 50 150 225 150 100

was 5 g/m3 and its density was 1032 kg/m3. Amine droplet with different sizes varying between 1 and 10 µm were injected from the feed inlet boundary zone along the surface. The droplet particle size distribution was set with an R-R distribution: dmin=1 µm, dmax=10 µm, davg=7 µm. Experiment

A cylinder-on-cone cyclone was designed based on the high throughput and low pressure drop, as shown in Fig. 1. It was made of SUS304 with a volute inlet. The experimental apparatus was illustrated in Fig. 2. Air stream was fed into the cylinder-on-cone cyclone through a Roots blower (TY0-1000, Shanghai Yangye Liquid Equipment Technology Co., Ltd. China ) with an airflow capacity of 1000 m3/h. Fine droplets were produced by the nozzle (SUE25B, Spraying Systems Co. USA) under high-pressure air from the air compressor (FB-420/7, Shanghai Jiebao Compressor Manufacturing Co., Ltd. China) and high-speed liquid from the metering pump (JW-JM-60/0.5, Shanghai Baoheng Pump Co., Ltd. China ). Fine droplets were ejected into the blending tank (Ø580 × 5300, Shanghai Huachang Environmental Protection Co. Ltd., China), and mixed with the air stream generated by the Roots blower. The pressure drop through the cyclone was determined by measuring the pressure difference between the inlet and outlet with a U-tube manometer. The inlet gas flow was measured by a digital flow meter (AMI KS300, Kimo Instruments, France) and controlled by throttle and bypass valves. The experimental test was conducted under temperature 20°C, pressure 0.101 MPa, air density 1.205 kg/m3 and amine liquid density 1032 kg/m3. The amine liquid composed of 30% N-methyl diethanolamine (MDEA) and 70% water.

According to product specification of the spray nozzle (SUE25B), the wanted droplets were generated by controlling the pressure and flow of both air compressor and metering pump. Fig. 3 showed the particle size distribution of droplets when air compressor pressure was 0.63 MPa, air flow was 411 L/min, metering pump pressure was 0.3 MPa and liquid flow was 52 L/h.

The sample was detected by an image particle size analyzer. Light Emitting Diode (LED) lights and Charge Coupled Device (CCD) cameras were mounted on the same platform located at the focal plane of the CCD cameras. The LED light source moved in sync with the traverse gauge to ensure the illumination of the area being measured. The step length and the frequency of the movement of the transverse gauge were fed into the computer’s operating interface. The particle size and concentration distribution of the droplets were measured and then displayed on the operating interface of the computer.

We adopted three kind of standard particles (1 µm Si-N-1000, 5 µm PS-1005, and 10 µm PS-1010, Suzhou Nanomicro Technology Co. Ltd., China) to calibrate the image particle size analyzer, and the calibration was repeated

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1677

1. Air filter, 2. Roots blower, 3. Metering pump, 4. Air compressor, 5. Nozzle, 6. Blending tank, 7. Throttle valve,

8. Bypass valve, 9. U-tube manometer, 10. Cyclone, 11. Sealed tube, 12. Amine tank, 13. Computer, 14. Stepping motor controller, 15. Stepping motor, 16. CCD camera, 17. Transverse gauge, 18. Telecentric lens,

19. LED lights, 20. Cable socket, 21. LED light controller, F-Flow meter.

Fig. 2. Schematic diagram of the experimental setup:

Fig. 3. Particle size distribution of droplets

6 times for each kind of standard particle. The density of the polystyrene standard particle was 1.08 g/cm3. The results showed that the measurement error was within ± 7%.

Separation efficiency E, the main performance index of a cyclone, is defined as follows: E = (1 –Co/Ci) × 100 (1) where Co and Ci are the concentrations of fine droplets in the overflow and inlet, respectively. RESULTS AND DISCUSSION Distribution of Three-dimensional Velocities

The position of vortex finder tip was defined as Z = 0. The four-dimensional velocity distributions in the position Z = –50, –150, –250, and –350 of the cyclone were studied separately. The negative positions Z = –50, Z = –150, Z = –250 and Z = –350 were corresponding to the cylindrical part, the interface between cylindrical and conical parts, the middle of conical part and the bottom of conical part, respectively. These negative positions have significant representativeness and are helpful to analyze the entire velocity distribution in cyclone separator. The simulation results were described as follows.

A 60% to 120% operation flexibility of equipment is desired in the industrial production process according to the Hydrogenation Unified Regulation by Li et al. (2009). Therefore, the changes of velocity, pressure drop and separation efficiency under different flow rates are important for industrial application. Axial Velocity

The axial velocity profiles in different cross-sections were shown in Fig. 4. The axial velocity in each cross-section basically showed a symmetric distribution, changed gradually from negative to positive in the direction from the wall to the center. Both upward velocity and downward velocity were increased with the inlet velocity (Vin). The fluid along the wall moved to the underflow outlet and the downward velocity increased with its radial distance from the cyclone axis. The fluid around the axis moved to the overflow outlet and the upward velocity increased with its radial distance from the cyclone wall.

The difference of the Z = –50 section from the other

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1678

Fig. 4. Axial velocity distributions of cyclone.

three sections was attributed to the two regions with positive axial velocity found in the axial velocity distribution chart. That was, a smaller region with positive axial velocity existed in the r = –0.035 region. The Z = –50 and r = –0.035 regions were just below the inlet of the cyclone separator. When the inlet velocity was increased, the upward axial velocity was also improved.

Axial velocity didn’t show better symmetry (Chu et al., 2011; Shukla et al., 2012) because of the single-inlet cyclone separator. Particularly, the downward axial velocity at one side of the rotating flow field passing through the cone section (Z = –250, Z = –350) was obviously higher than that in the other side. Radial velocity

The radial velocity profiles in different cross-sections were shown in Fig. 5. The radial velocity in the cross-section was characterized by the diagonal symmetry. It increased gradually from wall to axis and decreased sharply in the proximity of central axis. The difference of the Z = –50 section from the three other sections was attributed to the two velocities in the positive value region and two velocities

in the negative value region, as shown in the radial velocity distribution chart. The overflow pipe region inserted into the cyclone separator was in the region r = –0.02 to 0.02. The corresponding value was the radial velocity of the rotating flow field in the overflow pipe. In the other regions, the radial velocity was measured at the cylindrical section of the cyclone separator outside the overflow pipe. Tangential Velocity

Fig. 6 gave out the tangential velocity profiles in different cross-sections of cyclones. The tangential velocity increased in the same direction with the reduction of radius from wall to axis. In the area near to the wall, the tangential velocity reached its maximum and decreased with the reduction of radius until it reached to zero in the central axis. A portion of the overflow pipe was inserted into the cyclone separator in the Z = –50 section. Thus, four peak values were observed. The largest tangential velocity in the overflow pipe was obviously higher than that in the other regions.

Tangential velocity is the most important one among the three dimensional velocities, which is one of the key indexes

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1679

r (m) r (m)

Fig. 5. Radial velocity distributions of cyclone.

of the separation factor by determining the value of centrifugal force, because the tangential velocity is greater in value than the other two velocities. And what is more important is that, the centrifugal force deriving from the tangential velocity is the prerequisite for the separation of the two or more phases in cyclones. Separation efficiency

The grade efficiency curves of cyclone were shown in Fig. 7. The variation of inlet velocity has no significant effect on the trend of the grade efficiency curve. And the grade efficiency was 100% when the particle size bigger than 10 µm. The sharpness of grade efficiency curve of Vin = 10.29 m/s was slightly higher than other. And the grade efficiency showed a slight decreasing trend when the inlet velocity exceed 10.29 m/s. The cut size changed from 1.5 µm to 3 µm at different inlet velocity. That was, the inlet velocity has significant effects on the cut size.

The corresponding flow-efficiency curves were obtained from the FLUNT simulation and experimental tests (Fig. 8). The continuous phase in both simulation and experiment was compound gas, and the dispersed phase was the aqueous phase droplets. The temperatures were normal.

Droplet concentration was 0.5 g/m3, and the average particle size was 7 µm. Work flow in single cyclone within 100 m3/h (Vin = 20.58 m/s) was considered. Both in simulation and experiment, the separating efficiency reached a maximum value and slowly decreased with increasing flow. In the experiment, the separating efficiency reached the maximum value of 94.8% when the flow is 25 m3/h (Vin = 5.14 m/s). The separating efficiency slowly decreased when the flow was higher than 25 m3/h (Vin = 5.14 m/s). The measured separating efficiency was higher than 80% when the flow was in the range of 15 m3/h (Vin = 3.09 m/s) to 45 m3/h (Vin = 9.26 m/s). Therefore, the cyclone separator met the requirements of 60% to 120% of operational flexibility in industrial application.

Fig. 8 showed that the separating efficiency in the simulation was always higher than the experiment, in which the best separation efficiency of the cyclone in experimental test and simulation were 94.8% and 97%, respectively. This result was attributed to two reasons. First, the boundary conditions of the cyclone separator were set to fully capture conditions in the simulation status. That means all particles reached the side wall were considered to be separated from the main flow. During the experiment, the droplets separated

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1680

Fig. 6. Tangential velocity distributions of cyclone.

Fig. 7. Grade efficiency curves of cyclone.

to the side wall were possibly stripped from the side wall back to the rotating flow field. Second, in the simulation, the droplets were considered as indivisible global particles. However, large droplets were possibly broken into irregular

Fig. 8. Relationship of E and Q.

small droplets in the centrifugal flow field at experiment. These irregular small droplets were more difficult to separate than spherical particles.

The amine droplet concentration (Ci) changes with

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1681

different factors such as raw oil, hydrogen pressure, and reaction temperature effect. Therefore, an issue worth exploring is whether separation efficiency also changes with different droplet concentrations in a gas-liquid cyclone. Fig. 9 showed that there was not obvious change in the separation efficiency as the droplet concentration increased under the same inlet velocity. INDUSTRIAL APPLICATION Technological Process

The desulfurization of cycled hydrogen in the hydrogenation unit is shown in Fig. 10. This desulfurization process is commonly used in international oil refineries.

However, no cyclone separator (6) is set in the top of the desulfurization tower for the typical process. In this project, the cyclone separator was first set in the top of the

Fig. 9. Relationship of Ci and E under different Vin.

desulfurization tower. The gas-liquid cyclone was applied in the 200 million ton/year hydrogen cracking desulfurization system of Sinopec Guangzhou Refining & Chemical Company since April 2012.

In heat exchanger 1, lean amine solution is heated by hot water to 45°C and then transferred to buffer tank 2. Based on the sulfur content of the recycled hydrogen, pump 3 is used to extract the corresponding lean amine solution, which is used for spraying in the top of desulfurization tower 4. Recycled hydrogen from the cold high-pressure segregator passes through buffer tank 5 in front of the tower and then enters the middle and lower parts of desulfurization tower 4. Lean amine solution flows downward from the top of the desulfurization tower. Recycled hydrogen flows upward from the bottom of the desulfurization tower. Thus, recycled hydrogen containing sulfur is contacted with amine solution to remove sulfur in the recycled hydrogen. The amine solution always flows in the form of droplets with the recycled hydrogen to the gas phase outlet at the top of the desulfurization tower because of the effect of the tray, filler, temperature, and foreign solid particles in the desulfurization tower. Gas-liquid cyclone 6 that was developed in this research is used to recover amine droplets at the outlet. After being purified, the recycled hydrogen enters into gravity settler 7 and then into the cycle hydrogen compressor 8 to increase the pressure and mix with new hydrogen. The recycled hydrogen then enters the hydrogenation catalyst case for recycling. The field application was conducted under temperature 50°C, pressure 10.7 MPa, recycled hydrogen density 21.45 kg/m3, amine liquid density 989 kg/m3.

The gas-liquid cyclone device parameters are shown in Table 2.

High-pressure gas sampling for cycled hydrogen was conducted in the outlet of the desulfurizing tower. The samples collected with a sealed sampling bag were analyzed by a gas-phase chromatogram. An HP–6890 gas

1. Heat exchanger, 2. Amine buffer tank, 3. Pump, 4. Desulfurizing tower, 5. Hydrogen buffer tank,

6. Gas-liquid cyclone, 7. Hydrogen buffer tank, 8. Cycle hydrogen compressor.

Fig. 10. Technological process of the hydrocracking unit.

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1682

Table 2. Equipment parameters.

No. Title Parameter 1 Fluid stuff (V%) N2, O2, CO, CO2, H2, C1~5, C5

+, H2O, H2S 2 Operation pressure 10.7 MPa 3 Operation temperature 50C 4 Flow rate 257000 Nm3/h 5 Diameter 2000 mm 6 Total height 2300 mm 7 Shell material Q345R (R-HIC) 8 Cyclone material SUS304 9 Cyclone number 30

10 Total weight 2 t

chromatograph was employed for hydrocarbon analysis. A GC–9A gas chromatograph was utilized for water analysis, whereas a GC–14B gas chromatograph was used for H2S analysis.

Industrial Application Analysis

One sample place for recycled hydrogen piping is shown in Fig. 10. The results of the analysis are shown in Table 3.

The amine solution is at least composed of more than 50% water. Amine droplets sampled with the gas can be detected through H2O. Sampling data show that almost no H2O is detected in the gas. This observation confirms that all amine droplets are separated by the cyclone separator and are recovered in the desulfurization tower, which informed that the separation efficiency of cyclone for field application is up to 99.9%. The possible reason for this finding is that all amine droplets in the industrial production are greater than 10 µm in particle diameter.

The gravity settler behind desulfurization tower also shows that all amine droplets are separated by the cyclone separator. We continuously monitor the liquid level of the gravity settler (Fig. 10, Number 7) at actual production. Almost no increase can be found in the liquid level for

Table 3. Gas composition.

item 2013.4.1 2013.3.31 2013.3.30 2013.3.29N2 5.32 4.64 5.07 6.58 O2 0.16 0.10 0.11 0.15

CO2 < 0.01 < 0.01 < 0.01 < 0.01 CO < 0.01 < 0.01 < 0.01 < 0.01 H2 65.5 66.36 66.8 64.45

CH4 25.89 25.42 24.42 25.57 C2H6 1.04 1.15 1.11 1.21 C2H4 < 0.01 < 0.01 < 0.01 < 0.01 C3H8 0.98 1.09 1.11 1.05 C3H6 < 0.01 < 0.01 < 0.01 < 0.01

i-C4H10 0.25 0.29 0.34 0.29 n-C4H10 0.33 0.43 0.45 0.39 1-C4H8 < 0.01 < 0.01 < 0.01 < 0.01 2-C4H8 < 0.01 < 0.01 < 0.01 < 0.01

C5+ 0.18 0.34 0.33 0.30

H2O < 0.01 < 0.01 < 0.01 < 0.01 H2S 0.35 0.17 0.28 0.20

twelve months. This result indicates that amine droplets were not collected after long-term settlement and that amine droplets in the gas were fully separated by the cyclone separator at the top of the desulfurization tower.

Both analysis data and the liquid level of the gravity settler show that the cyclone separator can effectively separate amine droplets and can completely replace the setting tank. Therefore, a cyclone separator rather than a gravity settler can be set in the top of the desulfurization tower in future process designs. This modification will save investment not only the gravity settler itself but also on civil engineering, meters, valves, and pipes for the gravity settler.

CONCLUSIONS

Simulation and experimental study of the cyclone with different flow quantity showed that the change of the inlet velocity has significant effect on the value of three-dimensional velocities, separation efficiency, pressure drop, and cut size. The highest separation efficiency of the gas-liquid cyclone was 94.8% (experiment) and 97% (CFD) under 20°C and atmospheric pressure.

There was an efficient work area of the cyclone separator where the separation efficiency exceeds 80% when inlet flow quantity ranges from 15 m3/h (Vin = 3.09 m/s) to 45 m3/h (Vin = 9.26 m/s). The cyclone separator achieved the desired industrial operation flexibility of 60% to 120%.

Both analysis data and the liquid level of the gravity settler showed that the highest separation efficiency for industrial application was 99.9% under 50°C and 10.7 MPa. The cyclone separator which set in the top of the desulfurization tower could effectively separate amine droplets and could completely replace the setting tank. This design will save investment not only the gravity settler itself but also on civil engineering, meters, valves, and pipes for the gravity settler.

SYMBOLS USED a (mm) cyclone inlet height b (mm) cyclone inlet width d (mm) cyclone cylindrical body diameter di (mm) cyclone vortex finder diameter dc (mm) cyclone conical liquid outlet diameter s (mm) extrusive vortex finder height

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1683

H1 (mm) cyclone cylindrical body height H2 (mm) cyclone conical body height H3 (mm) cyclone underflow pipe height In (mm) cyclone inlet length G’(d) ( - ) grade efficiency E ( - ) separation efficiency Vin ((m/s) inlet air velocity Co (kg/m3) droplet concentration of overflow air Ci (kg/m3) droplet concentration of inlet air Q (m3/h) feeding flow rate Z (mm) cyclone cross section height REFERENCES Chen, S.C. and Tsai, C.J. (2007). An Axial Flow Cyclone

to Remove Nanoparticles at Low Pressure Conditions. J. Nanopart. Res. 9: 71–83.

Chu, K.W., Wang, B., Xu, D.L., Chen, Y.X. and Yu, A.B. (2011). CFD-DEM Simulation of the Gas-solid Flow in a Cyclone Separator. Chem. Eng. Sci. 66: 834–847.

Diao, Y.F., Li, X., Gu, P.D. and Li, H. (2009). Numerical Simulation of Gas-solid Flow in the Axisymmetric Inlets Square Separator. J. Chem. Eng. 26: 879–883.

Earni, S., Wang, S.B., Mohan, R.S., Shoham, O. and Marrelli, J.D. (2003). Slug Detection as a Tool for Predictive Control of GLCC© Compact Separators. J. Energy Res. Technol. 125: 145–153.

Elsayed, K. and Lacor, C. (2011). The Effect of Cyclone Inlet Dimensions on the Flow Pattern and Performance. Appl. Math. Modell. 35: 1952–1968.

Gu, H.Y., Zhang, Y.D., Huang, Y., Hewitt, N. and Roskilly, A.P. (2013). Waste Biomass from Production Process Co-firing with Coal in a Steam Boiler to Reduce Fossil Fuel Consumption: A Case Study. J. Energy Chem. 22: 413–419.

He, F., Wang, H. and Dai, Y.N. (2007). Application of Fe2O3/Al2O3 Composite Particles as Oxygen Carrier of Chemical Looping Combustion. J. Nat. Gas Chem. 16: 155–161.

Hsu, Y.D., Chein, H.M. and Tsai, C.J. (2005). Axial Flow Cyclone for Segregation and Collection of Ultrafine Particles: Theoretical and Experimental Study. Environ. Sci. Technol. 39: 1299–1308

Li, L.Q., Zheng, Q. and Chen, J. (2009). 200 Million ton/year Hydrogen Cracking Unified Regulation of Sinopec Guangzhou Refining & Chemical Company, Luoyang Petrochemical Engineering Corporation, Luoyang, China.

Lin, G.Y., Cuc, L.T., Lu, W., Tsai, C.J., Chein, H.M. and Chang, F.T. (2013). High-efficiency Wet Electrocyclone for Removing Fine and Nanosized Particles. Sep. Purif. Technol. 114: 99–107.

Molina, R., Wang, S.B., Gomez, L.E., Mohan, R.S., Shoham, O. and Kouba, G. (2008). Wet Gas Separation in Gas-Liquid Cylindrical Cyclone Separator. J. Energy Res. Technol. 130: 1–13.

Movafaghian, S., Jaua-Marturet, J.A., Mohan, R.S., Shoham, O. and Kouba, G.E. (2000). The Effects of Geometry, Fluid Properties and Pressure on the Hydrodynamics of Gas-liquid Cylindrical Cyclone Separators. Int. J.

Multiphase Flow 26: 999–1018. Qiu, Y.F., Deng, B.Q. and Kim, C.N. (2012). Numerical

Study of the Flow Field and Separation Efficiency of a Divergent Cyclone. Powder Technol. 217: 231–237.

Safikhani, H., Akhavan-Behabadi, A.B., Shams, M. and Rahimyan, M.H. (2010). Numerical Simulation of Flow Field in Three Types of Standard Cyclone Separators. Adv. Powder Technol. 21: 435–442.

Shukla, S.K., Shukla, P. and Ghosh, P. (2013). The Effect of Modeling of Velocity Fluctuations on Prediction of Collection Efficiency of Cyclone Separators. Appl. Math. Modell. 37: 5774–5789.

Su, Y.X. (2006). The Turbulent Characteristics of the Gas-solid Suspension in a Square Cyclone Separator. Chem. Eng. Sci. 61: 1395–1400.

Su, Y.X. and Mao, Y.R. (2006). Experimental Study on the Gas-solid Suspension Flow in A Square Cyclone Separator. Chem. Eng. J. 121: 51–58.

Su, Y.X., Zhao, B.T. and Zheng, A.Q. (2011). Simulation of Turbulent Flow in Square Cyclone Separator with Different Gas Exhaust. Ind. Eng. Chem. Res. 50: 12162–12169.

Su, Y.X., Zheng, A.Q. and Zhao, B.T. (2011). Numerical Simulation of Effect of Inlet Configuration on Square Cyclone Separator Performance. Powder Technol. 210: 293–303.

Tsai, C.J., Chen, D.R., Chein, H.M., Chen, S.C., Roth, J.L., Hsu, Y.D., Li, W.L. and Biswas, P. (2004). Theoretical and Experimental Study of an Axial Flow Cyclone for Fine Particle Removal in Vacuum Conditions. J. Aerosol Sci. 35: 1105–1118.

Tsai, C.J., Chen, S.C., Przekop, R. and Moskal, A. (2007). Study of an Axial Flow Cyclone to Remove Nanoparticles in Vacuum. Environ. Sci. Technol. 41: 1689–1695.

Tsai, C.J., Chang, C.S., Chen, S.C., Chen, P., Shih, T.S., Pui, D.Y.H., Karasev, V.V., Onischuk, A.A. and Li, S.N. (2008). Laboratory and Field Tests of a Novel Three-Stage Personal Dust Sampler for Sampling Three Dust Fractions Simultaneously. Aerosol Sci. Technol. 42: 86–95.

Tsai, C.J., Huang, C.Y., Chen, S.C., Ho, C.E., Huang, C.H., Chen, C.W., Chang, C.P., Tsai, S.J. and Ellenbecker, M.J. (2011). Exposure Assessment of Nano-sized and Respirable Particles at Different Workplaces. J. Nanopart. Res. 13: 4161–4172.

Tsai, C.J., Liu, C.N., Hung, S.M., Chen, S.C., Uang, S.N., Cheng, Y.S. and Zhou, Y. (2012). Novel Active Personal Nanoparticle Sampler for the Exposure Assessment of Nanoparticles in Workplaces. Environ. Sci. Technol. 46: 4546–4552.

Wang, S.B., Gomez, L.E., Mohan, R.S., Shoham, O. and Kouba, G.E. (2003). Gas-Liquid Cylindrical Cyclone (GLCC©) Compact Separators for Wet Gas Applications. J. Energy Res. Technol. 125: 43–50.

Wang, S.B., Gomez, L., Mohan, R., Shoham, O., Kouba, G. and Marrelli, J. (2010). The State-of-the-Art of Gas-Liquid Cylindrical Cyclone Control Technology: From Laboratory to Field. J. Energy Res. Technol. 132: 1–9.

Zhao, B.T. (2004). A Theoretical Approach to Pressure

Ma et al., Aerosol and Air Quality Research, 14: 1675–1684, 2014 1684

Drop across Cyclone Separators. Chem. Eng. Technol. 27: 1105–1108.

Zhao, B.T., Shen, H.G. and Kang, Y.M. (2004). Development of a Symmetrical Spiral Inlet to Improve Cyclone Separator Performance. Powder Technol. 145: 47–50.

Zhao, B.T. (2005). Experimental Investigation of Flow Patterns in Cyclones with Conventional and Symmetrical Inlet Geometries. Chem. Eng. Technol. 28: 969–972.

Zhao, B.T. and Su, Y.X. (2010). Artificial Neural Network-

based Modeling of Pressure Drop Coefficient for Cyclone Separators. Chem. Eng. Res. Des. 88: 606–613.

Zhao, B.T. (2012). Prediction of Gas-particle Separation Efficiency for Cyclones: A Time-of-flight Model. Sep. Purif. Technol. 85: 171–177.

Received for review, October 8, 2013 Accepted, January 24, 2014