- Email: [email protected]
Enhancement of louver dust collector efficiency using modified dust container
Powder Technology 325 (2018) 69–77
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Enhancement of louver dust collector efficiency using modified dust container Jung-Bo Sim a, Un-Hak Yeo a, Gwang-Hun Jung a, Su-Beom Park a, Gwi-Nam Bae b, Se-Jin Yook a,⁎ a b
School of Mechanical Engineering, Hanyang University, Seoul 04763, Republic of Korea Korea Institute of Science and Technology, Seoul 02792, Republic of Korea
a r t i c l e
i n f o
Article history: Received 28 July 2017 Received in revised form 23 September 2017 Accepted 29 October 2017 Available online 03 November 2017 Keywords: Inertial separation Louver dust collector Dust container Collection efficiency
a b s t r a c t External air injected into a gas turbine contains many solid particles, which can reduce the performance and life of the turbine. In this study, a louver dust collector, which is a type of inertial dust collector, was used to remove solid particles from air, and the shape of the dust container, which is part of the louver dust collector, was modified to improve dust collection efficiency. As existing dust containers have a structure for isolating particles, the inflow of the air into the dust containers is limited and effective removal of particles is difficult. In this study, slits were drilled in the dust container, baffle plates were used, and raised spots were added to improve particle collection efficiency through improved air flow in the dust container and increased inertia effect of the particles. The trajectory of the particles and the collection efficiency for each dust container shape were predicted using numerical analysis and the numerical analysis results were verified using a wind tunnel test. Results indicate that for an air flow rate of 3 m3/min, the collection efficiencies of the louver dust collectors with the one-slit model dust container and two-slit model dust container improved by 40.1% and 43.5%, respectively compared with that of the louver dust collector with the existing dust container. Furthermore, for an air flow rate of 6 m3/min, the collection efficiencies of the louver dust collectors with the one-slit model dust container and two-slit model dust container improved by 32.9% and 37.6%, respectively, compared with that of the louver dust collector with the existing dust container. Therefore, it is expected that the particle collection efficiency of the existing louver dust collector can be effectively increased by utilizing the shape of the dust container proposed in this study. © 2017 Elsevier B.V. All rights reserved.
1. Introduction A gas turbine is a heat engine that drives a turbine using hightemperature and high-pressure combustion gas, and for its operation, external air is usually injected into the gas turbine. However, as external air contains many solid particles, its continuous inflow may cause corrosion of the compressor blades and damage to the engine, and possibly may lower the performance of the gas turbine [1–3]. Various studies have been conducted to prevent performance degradation due to various solid particles contained in the external air introduced into the gas turbine. The particle collection efficiency was evaluated through the fixed valve tray column for the removal of fly ash particles entering the gas turbine [4]. Coating treatment was performed to improve the erosion resistance and prevent gas turbine compressor blades from being eroded by solid particles [5,6]. A computational fluid dynamics (CFD) analysis was performed to identify the impact adhesion properties of particles and guidelines for filtration systems were prepared to prevent blade sediments and compressor performance degradation [7]. Several types of inertial separators were studied to protect gas ⁎ Corresponding author. E-mail address: [email protected] (S.-J. Yook).
https://doi.org/10.1016/j.powtec.2017.10.047 0032-5910/© 2017 Elsevier B.V. All rights reserved.
turbines from dust and sand [8]. The use conditions of an electrostatic precipitator for separating solid particles from combustion gas were investigated [9]. In this study, an inertial dust collector was considered to be installed in the secondary flow path of a gas turbine engine, to remove various solid particles floating in the external air entering the gas turbine. Inertial dust collectors collect dust using the inertia of the particles and can have a high dust collection efficiency under the condition that the gas moves at a high speed as in a gas turbine. Among them, the louver dust collector can effectively separate and remove solid particles using the louver blade in the flowing gas [10]. A louver refers to long plates, regardless of the material, continuously placed at a regular interval. There have been many studies to improve ambient air flow and to separate and remove particles in air using louvers. Dust particles in the combustion gas were removed through the circulating granular bed filter (CGBF) using louver plates [11]. The size distribution of dust particles was measured by the centrifugal separation method using a louver media filter, and particle collection efficiency was analyzed [12]. A fixed louver was installed in the burner duct of a pulverized coal power plant to separate solid particles from gas [13]. A louversublouver system was developed to remove fine dust particles and fugitive dust [14,15]. A device using a louver was installed on the
70
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
upper wall of a factory to improve indoor air ventilation efficiency [16]. The natural ventilation efficiency of the wind tower system was improved using louvers [17]. The angle of the louver was a factor that significantly influenced the dust collection efficiency of the CGBF [18]. Analytical, numerical, and experimental investigations of louvered classifiers demonstrated the influences of louver shape, louver angle, number of louvers, inlet velocity, powder feed-rate, and blowdown on fractional collection efficiency [19–21]. These parameters were also examined in an extensive literature on louvered classifiers in both rectilinear and cylindrical geometries. A partial review of that literature is provided in [22] together with a correlation of fractional collection efficiency, showing the influences of gas and particle dynamics and array geometry. Particularly relevant to the present study is the work of [23,24]. In that study, a louvered classifier with a dust container, analogous to the “original model” herein, was applied to dust removal in the secondary flow path of a gas turbine engine. The present study uses and improves upon the results of [24]. As mentioned earlier, many studies have been conducted to improve particle collection efficiency by changing the properties of louver blades; however, only few studies have attempted to improve the collection efficiency by changing the shape of the dust container. The purpose of this study is to improve the particle collection efficiency of the louver dust collector by changing the shape of the dust container. The louver dust collector design considered in this study was based on that of the previous literature [23,24]. Slits were drilled and additional baffle plates were installed in the dust container to increase the amount of particles collected in the dust container through improved air flow inside the dust container and maximized inertia effect of the particles. The air flow through each dust container shape was predicted using CFD simulation, and the particle collection efficiency of each dust container shape was predicted through the analysis of the particle trajectory. The particle collection efficiency of the louver dust collector was measured through a wind tunnel test and compared with the predicted collection efficiency from simulation. Furthermore, changes in the collection efficiency due to the application of grease were evaluated and analyzed for each dust container model. Results indicate that changing the dust container shape could improve particle collection efficiency and suppress particle re-scattering. Therefore, the collection efficiencies of existing louver dust collectors can be significantly
improved with low investment cost by replacing only their dust container shapes with those proposed in this study. 2. Numerical Fig. 1a shows the two-dimensional (2D) model of the louver dust collector used for numerical analysis. It consists of continuous louver blades and one dust container downstream. The louver was constructed using five continuously arranged straight louver blades with louver blade lengths of 46 mm and gaps of 15 mm. In this study, to compare the collection performances of different dust container shapes, the collection efficiency of the louver dust collector was predicted by changing the dust container shape while the louver blade shape remained unchanged. Fig. 1b shows the grid system used for 2D simulation of flow field and particle trajectory. Triangular grids were created using the Gambit software. After performing grid independence test, the louver dust collector comprised approximately 70,000 triangular grid cells and the inside of the dust container comprised about 10,000 triangular grid cells. Fig. 2a shows the existing dust container structure. It is a closed structure without gap except the entrance where dust is introduced through the louver blades (original model). The dust container of Fig. 2b has a slit in the left part of the original model (one-slit model). The width of the slit was 10 mm. A raised spot was added beneath the slit to prevent dust collected in the dust container from being rescattered. The one-slit model is identical to the original model of Fig. 2a except for the slit and the raised spot. Fig. 2c shows a dust container with 10-mm-width slits in the left and right parts of the original model dust container and a 30-mm baffle plate at the center of the dust container (two-slit model). Raised spots were also added beneath the left and right slits and the baffle plate to prevent particles collected in the dust container and the baffle plate from being rescattered. ANSYS FLUENT Release 16.1, a commercial CFD software, was used for numerical analysis. The flow inside the louver dust collector was assumed to be two-dimensional, steady, incompressible, and turbulent. Turbulent flow analysis was conducted using the standard k-ω turbulence model based on the Reynolds-averaged Navier–Stokes (RANS) model [23–24]. The boundary conditions for flow analysis
Fig. 1. Shape of louver dust collector: (a) Cross-sectional view; (b) Grid system for 2-dimensional simulation.
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
71
(Ncollected) was counted and the collection efficiency (η) of the louver dust collector was calculated using the equation: η¼
N collected : N injected
ð1Þ
For each tested particle size, Ninjected was set at 1000. 3. Experimental
Fig. 2. Dust container shapes: (a) Original model; (b) One-slit model; (c) Two-slit model.
were defined as follows: the velocity inlet condition was set at the entrance of the louver dust collector, the pressure outlet condition was set at the exit of the louver dust collector, and the louver blades as well as the dust container were set as walls. The SIMPLE algorithm was selected for pressure-velocity coupling, while the second-order upwind discretization scheme was selected to discretize the momentum [25]. The convergence criteria were set to 10−6 for continuity, momentum, and energy equations. The particle trajectory was analyzed using the discrete phase models (DPM) embedded in FLUENT [25]. Only the walls of the dust container were set as the ‘trap’ type boundary conditions, while the other walls, including louver blades, were set as the ‘reflect’ type boundary conditions [25]. In other words, the particles were assumed to be permanently trapped once they hit the dust container wall, while they were presumed to be bounced off once they hit the walls other than the dust container wall. The particle density was set to 2.65 g/cm3 considering the Arizona Test Dust that was used in the experiment. The particle was assumed to be spherical, and the drag acting on the particle was calculated using the slip correction factor according to the particle size. In addition, the Brownian motion of particles caused by the collision between the particles and air molecules was considered. In order to account for turbulent dispersion of particles, the stochastic random walk model provided in the FLUENT software was employed as described in [26]. After the injection of Ninjected particles from an equidistant position in the horizontal and vertical directions of the cross-sectional area of the flow path at the entrance of the louver dust collector, the number of particles collected in the dust container
Fig. 3a is a schematic diagram of the experimental setup for measuring the collection efficiency of the louver dust collector in a wind tunnel. A louver dust collector with the same size used in the numerical analysis model was installed in a wind tunnel with a cross sectional area of 0.01 m2 (0.1 m × 0.1 m). A photo of the louver dust collector, for example, with the one-slit model dust container, is shown in Fig. 3b. Air was injected using a turbofan. A HEPA filter was installed at the inlet of the wind tunnel to ensure that only particles used in the experiment enter the wind tunnel. The Arizona Test Dust (ISO 12103-1, A4 type) was dispersed into the air entering the wind tunnel through the HEPA filter, using a solid aerosol generator (SAG 410, TOPAS, Dresden, Saxony, Germany). The particles were injected in the opposite direction to the flow and there was sufficient distance for the development of turbulent flow so that the particles could be distributed evenly across the cross section of the wind tunnel. A honeycomb was used to obtain a uniform flow velocity distribution. The particle number concentrations by particle size were measured before and after the louver dust collector using an optical particle counter (OPC, Model 1.109, GRIMM, Ainring, Bayern, Germany). In addition, the inflow velocity of the air at the entrance of the louver dust collector was measured using a velocity meter (Model 9535, TSI, Shoreview, MN, USA), and the pressure drop through the louver dust collector was measured using a differential pressure transmitter (CP 210, KIMO, Montpon, Dordogne, France). Air velocities and particle number concentrations were measured at five points both in the horizontal and vertical directions (that is, a total of 25 points on the cross sectional area of the flow path at the entrance of the louver dust collector), and the measured values showed uniformities of ±5.5% and ±18.2% from the average values, respectively. The collection efficiency in the experimental method was obtained by η ¼ 1−
C downstream C upstream
ð2Þ
where Cupstream and Cdownstream are the particle number concentrations measured before and after the louver dust collector, respectively. 4. Results and discussion Through numerical analysis, the pressure distribution of the ambient air for each dust container model was examined at an air flow rate of 3 m 3/min. Fig. 4a shows the pressure distribution in the louver dust collector that used the original model dust container. Because there was almost no flow from inside the dust container to the outside, a relatively constant pressure was observed inside and around the dust container. On the other hand, in the case of the one-slit model of Fig. 4b, high pressure was observed inside the dust container, which spread around the dust container through the slit on the left side. This indicates that a large amount of air entered the dust container and then escaped through the slit. Similarly, for the two-slit model of Fig. 4c, a relatively high pressure was maintained inside the dust container, which spread around the dust container through the left and right slits. This indicates that a large amount of air was introduced to the dust container and then escaped through the left and right slits.
72
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
Fig. 3. Schematic diagram of the experimental setup for measuring the collection efficiency of the louver dust collector using a wind tunnel.
Table 1 presents a comparison of the numerical analysis results with wind tunnel test results in terms of the differential pressure before and after the louver dust collector according to the dust container model. As illustrated in Fig. 3a, the inlet pressure was measured before the louver blades and outlet pressure was measured after the dust container. The air flow rates were set at 3 m3/min and 6 m3/min. Results of the comparison indicate that the differential pressures were in good agreement with the maximum error of 3.4%. As the air flow rate increased twice, the differential pressure became roughly quadrupled. In terms of the differential pressure of each dust container model, the louver dust collector with the one-slit model showed the highest differential pressure of all the models. The original model and the two-slit model showed similar differential pressures, while a smaller differential pressure was observed in the louver dust collector with the two-slit model. Fig. 5 represents the trajectory of 5-μm particles in the louver dust collector when the air flow rate was 3 m3/min; the figure shows the
patterns in which particles were collected and released through each dust container model. It should be noted that Fig. 5 shows the trajectories of only 20 particles, while 1000 particles were tracked to numerically estimate the collection efficiency. Fig. 5a shows the trajectory of the particles passing through the louver dust collector with the original model dust container. It was observed that the injected particles moved upward toward the dust container. However, since there was almost no air flow through the dust container, the behavior of the particles to be introduced into the dust container was restricted. Therefore, it was confirmed that the particles floating in air could not be sufficiently introduced into the dust container and most particles were either collected near the entrance of the dust container or released from the louver dust collector without being collected. Fig. 5b shows the trajectory of the particles passing through the louver dust collector in the one-slit model dust container. Unlike in the case of Fig. 5a, many of the particles that passed over the louver blades were
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
73
Fig. 4. Pressure distribution around the dust container: (a) Original model; (b) One-slit model; (c) Two-slit model.
introduced into the dust container, and most of them could not escape the dust container due to inertia and were collected in the dust container as the flow was suddenly forced from the right to the left. Fig. 5c shows the trajectory of the particles passing through the louver dust collector with the two-slit model dust container. Since there were slits on the left and right sides of the dust container, it was observed that some of the particles that entered the dust container escaped through the slits whereas the others were collected in the dust container. Owing to the baffle plate installed at the center of the dust container,
Table 1 Differential pressures by dust container model: comparison of numerical analysis results with experimental results. Air flowrate (m3/min)
Dust container model
Differential pressure (Pa) Numerical analysis
Experimental measurement
3
Original model One-slit model Two-slit model Original model One-slit model Two-slit model
378 560 352 1695 2087 1612
365 ± 32 551 ± 34 340 ± 28 1642 ± 124 2064 ± 135 1575 ± 115
6
particles could not directly escape through the right slit and moved downward in the dust container. Then, the flow of the particles was suddenly forced toward the left and right slits, and particles were collected near the bottom of the dust container due to inertia. Fig. 6a shows the collection efficiencies of the louver dust collectors with the original model and the one-slit model dust containers. The numerical analysis results and the wind tunnel test results were in good agreement. When the air flow rate was 3 m3/min, the cut-off size (particle size corresponding to collection efficiency of 50%) was 8.41 μm for the original model and 5.04 μm for the one-slit model, confirming that the one-slit model had a 3.37 μm lower cut-off size. Fig. 6b compares the collection efficiencies of the louver dust collectors of the original model and two-slit model dust containers. When the air flow rate was 3 m3/min, the cut-off size of the two-slit model was 4.75 μm and it was 3.66 μm lower than that of the original model. Fig. 7a compares the collection efficiencies of the louver dust collectors of the original model and one-slit model dust containers when the air flow rate was 6 m3/min. The cut-off size was 5.35 μm for the original model and 3.59 μm for the one-slit model, confirming that the one-slit model had a 1.76 μm lower cut-off size. Fig. 7b compares the collection efficiencies of the louver dust collectors of the original model and two-slit model. The cut-off size of the two-slit model was 3.34 μm, which was 2.01 μm lower than the original model. As seen from Figs. 6 and 7, the one-slit model and two-slit model showed similar cut-off sizes and
74
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
Fig. 5. Trajectory of particles (5 μm): (a) Original model; (b) One-slit model; (c) Two-slit model.
both showed lower cut-off sizes than that of the original model. Therefore, it was confirmed that changing the dust container shape alone can improve the collection efficiency of the louver dust collector. The aforementioned experimental results of Figs. 6 and 7 were obtained from dust containers in which grease was applied to the walls to prevent the particles collected in the dust containers from being re-scattered. However, when a large number of particles are collected in the dust containers, the grease application effect could disappear and particle re-scattering problem could occur. Therefore, the collection efficiencies of the louver dust collector with and without grease application were measured and compared. Fig. 8 shows the results when the air flow rate into the louver dust collector was 3 m3/min, while Fig. 9 shows the results when the air flow rate was 6 m3/min. Figs. 8a and 9a show the collection efficiencies of the louver dust collector with the original model dust container, Figs. 8b and 9b show the collection efficiencies of the louver dust collector with the one-slit model, and Figs. 8c and 9c show the collection efficiencies of the louver dust collector with the two-slit model. In all cases, the collection efficiencies without grease application were lower than the collection efficiencies with grease application. In particular, when the particle size was larger than the cut-off size, the collection efficiency significantly decreased due to re-scattering. Furthermore, it was found that the collection efficiency decreased more due to re-scattering when the air flow rate was 6 m3/min than when the air flow rate was 3 m3/min. For the dust container shapes without grease application, a comparison of the degree of re-scattering revealed that the original model dust container showed the highest decrease in collection efficiency while the modified dust containers showed less re-scattering degrees, and hence higher collection efficiencies than the original model. In particular, the two-slit model dust container showed a significant reduction in particle
re-scattering under the same conditions compared with the one-slit model dust container, and its collection efficiency was relatively high even without grease application. Therefore, it was confirmed that improved dust container designs can enhance the collection efficiency of the louver dust collector, as well as suppress re-scattering of particles. However, by considering the particle re-scattering problem without grease application, it is needed to improve the collection efficiency further. The reason that the louver collector with the one-slit model or the two-slit model dust container showed improved collection efficiency compared with the original model dust container is that the flow that did not occur in the original model dust container was activated by drilling slits in the container and the particles were more likely to be collected in the dust container due to inertia when the flow was forced to change its direction to escape through the slits. Another reason is that the re-scattering of the particles was suppressed due to the presence of the baffle plate or raised spots. The differential pressure of the one-slit model dust container was approximately 1.2–1.5 times the differential pressure of the original model. This is because the pressure increased while air introduced into the dust container was forced to change its flow direction dramatically to escape through the slit located at a position opposite the inflow direction. Therefore, the collection efficiency of the louver dust collector was improved while the differential pressure increased. Meanwhile, the differential pressure of the two-slit model dust container was about 0.93–0.95 times the differential pressure of the original model. In other words, the two-slit model dust container showed the least differential pressure, the highest collection efficiency, and the least particle re-scattering of all the tested dust container models. This is because the pressure increased due to the left slit located at a position opposite the inflow direction, as well as the baffle plate
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
75
Fig. 7. Comparison of particle collection efficiency by dust container model (6 m3/min): (a) Original model and One-slit model; (b) Original model and Two-slit model. Fig. 6. Comparison of particle collection efficiency by dust container model (3 m3/min): (a) Original model and One-slit model; (b) Original model and Two-slit model.
located in the center, which was offset by the presence of the right slit located in the same direction as the inflow direction. It is expected that not only the one-slit model but also the two-slit model dust container can be applied to the louver dust collector, in which the airflow is caused by an external means such as a pump or a blower, to collect particles more effectively. The two-slit model dust container is also anticipated to be used in the louver dust collector for both securing the airflow rate and improving the collection efficiency when the ambient air is introduced into the louver dust collector in the absence of a pump or a blower, e.g., when the airflow is induced to enter into the louver dust collector by the running of a subway train as employed in recent studies [27,28], in which a louver dust collector was installed at the bottom of a subway train to remove coarse particles generated or re-suspended in the subway tunnel.
5. Conclusions This study was conducted in an effort to enhance the collection efficiency of a louver dust collector, of which design was based on that of the previous literature [23,24]. To improve the collection efficiency of the louver dust collector, the shape of the dust container, which is a component of the louver dust collector, was varied. The existing dust
container shape is not effective for dust collection because it simply isolates the particles introduced through the louver blades for collection. Therefore, in this study, slits were drilled in the existing dust container to induce flow into the dust container and the flow direction was sharply changed in the dust container to maximize the collection of the particles through inertia. The proposed one-slit model dust container had a slit located at a position opposite the direction of flow entering the container while the proposed two-slit model dust container had a slit located at a position opposite the direction of flow entering the container, another slit located in the same direction as the inflow direction, and a baffle plate for sharply changing the flow direction. In addition, raised spots were added around the slits to suppress the re-scattering of particles. For these three dust container models, the collection efficiencies predicted by numerical analysis and those obtained from measurements in a wind tunnel test were compared. The numerical analysis results and the wind tunnel test results were in good agreement for dust containers with grease application. It was confirmed that changing the dust container shape could improve the particle collection efficiency and suppress particle re-scattering. Therefore, it is expected that the collection efficiencies of the existing louver dust collectors will be significantly improved with low investment cost by replacing only their dust container shapes with those proposed in this study. Future studies need to be performed to develop more efficient dust container designs by considering particle re-scattering problem without grease application.
76
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
Fig. 8. Particle collection efficiencies with and without grease application (3 m3/min): (a) Original model; (b) One-slit model; (c) Two-slit model.
Fig. 9. Particle collection efficiencies with and without grease application (6 m3/min): (a) Original model; (b) One-slit model; (c) Two-slit model.
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
Acknowledgment The work was supported by research grants for the Railway Technology Research Project from the Ministry of Land, Infrastructure and Transport, Republic of Korea (17RTRP-B082486-04). References [1] P.F. Batcho, J.C. Moller, C. Padova, M.G. Dunn, Interpretation of gas turbine response due to dust ingestion, ASME, transactions, J. Eng. Gas Turb. Power. 109 (3) (1987) 344–352. [2] M.G. Dunn, C. Padova, J.E. Moller, R.M. Adams, Performance deterioration of a turbofan and a turbojet engine upon exposure to a dust environment, J. Eng. Gas Turb. Power. 109 (3) (1987) 336–343. [3] W. Tabakoff, W. Hosny, A. Hamed, Effect of solid particles on turbine performance, J. Eng. Power 98 (1) (1976) 47–52. [4] Q. Wang, X. Chen, X. Gong, Theoretical and experimental investigation on the characteristics of fly-ash scrubbing in a fixed valve tray column, AICHE J. 59 (6) (2013) 2168–2178. [5] S.A. Muboyadzhyan, Erosion-resistant coatings for gas turbine compressor blades, Russ. Metall. 2009 (3) (2009) 183–196. [6] A.V. Ryzhenkov, G.V. Kachalin, A.F. Mednikov, A.B. Tkhabisimov, The investigation of construction materials and protective coatings wear resistance to solid particle erosion, Mod. Appl. Sci. 9 (4) (2014) 85–95. [7] A. Suman, M. Morini, R. Kurz, N. Aldi, K. Brun, M. Pinelli, P.R. Spina, Estimation of the particle deposition on a subsonic axial compressor blade, J. Eng. Gas Turb. Power. 139 (1) (2017) 1–12. [8] M.R. Caskey, Dust and sand protection for marine gas turbines, J. Eng. Power 104 (2) (1982) 260–267. [9] J. Brandt, H. Heer, G.J. Clessuras, Criteria for electrostatic collection of unfavorable dusts, J. Eng. Power 97 (3) (1975) 461–464. [10] J.L. Smith Jr, M.J. Goglia, The mechanism of separation in the louver-type dust separator, J. Air Pollut. Control Assoc. 5 (1) (1955) 51–62. [11] J.C. Bai, S.Y. Wu, A.S. Lee, C.Y. Chu, Filtration of dust in a circulating granular bed filter with conical louver plates (CGBF-CLPs), J. Hazard. Mater. 142 (1) (2007) 324–331. [12] L.J. Czerwonka, J.M. Carey, Application of Particle-Size Analysis Data for the Determination of Air/Cleaner Performance. In ASME Gas Turbine Conference and Products Show 1965: American Society of Mechanical Engineers, 1965 (V001T01A006).
77
[13] G. Živković, S. Nemoda, P. Stefanović, P. Radovanović, Numerical simulation of the influence of stationary louver and coal particle size on distribution of pulverized coal to the feed ducts of a power plant burner, Therm. Sci. 13 (4) (2009) 79–90. [14] J. Smid, S.S. Hsiau, S.A. Tsai, C.C. Tzeng, Y.P. Chyou, Study on gravity flow of granules in beds supported by louver–sublouver system, Adv. Powder Technol. 20 (2) (2009) 127–138. [15] N. Arashi, T. Komuro, Y. Hishinuma, O. Kanda, H. Kuroda, Particle movements on louver supporting moving granular bed, Powder Technol. 43 (1985) 175–179. [16] J.H. Kang, S.J. Lee, Improvement of natural ventilation in a large factory building using a louver ventilator, Build. Environ. 43 (12) (2008) 2132–2141. [17] S. Liu, C.M. Mak, J. Niu, Numerical evaluation of louver configuration and ventilation strategies for the windcatcher system, Build. Environ. 46 (8) (2011) 1600–1616. [18] J.C. Bai, S.Y. Wu, A.S. Lee, Dust collection efficiency analysis in a two-dimensional circulating granular bed filter, J. Air Waste Manage. Assoc. 56 (5) (2006) 684–694. [19] H. Yoshida, K. Fukui, H. Morizaki, Particle classification of fly ash using a modified louver-type separator and reduction of unburned carbon amount, KONA Powder and Part. J. 18 (2000) 221–229. [20] M. Higashino, T. Maeda, K. Fukui, H. Yoshida, Improvement of particle separation in louver-type separator by use of numerical simulation, J. Soc. Powder Technol. Japan 52 (5) (2015) 252–259. [21] H. Yoshida, K. Fukui, K. Kanagawa, M. Okamoto, Y.U.N. Chi-Mun, K. Iinoya, Particle classification of a louver-type separator, J. Soc. Powder Technol. Japan 34 (9) (1997) 684–689. [22] T.R. Blake, Assessment of fractional collection efficiency in louvered inertial particle classifiers, Powder Technol. 311 (2017) 432–438. [23] G.O. Musgrove, M.D. Barringer, K.A. Thole, E. Grover, J. Barker, Computational design of a louver particle separator for gas turbine engines, ASME Turbo Expo 2009: Power for Land, Sea, and Air 2009, pp. 1313–1323. [24] G.O. Musgrove, K.A. Thole, E. Grover, J. Barker, Performance measurements of a unique louver particle separator for gas turbine engines, J. Eng. Gas Turb. Power. 135 012001 (2013) 1–10. [25] FLUENT 12.0 User's Guide, FLUENT Inc. April, 2009. [26] A. Dehbi, A. CFD, Model for particle dispersion in turbulent boundary layer flows, Nucl. Eng. Des. 238 (2008) 707–715. [27] K.R. Lee, W.G. Kim, S.H. Woo, J.B. Kim, G.N. Bae, H.K. Park, H. Yoon, S.J. Yook, Investigation of airflow and particle behavior around a subway train running in the underground tunnel, Aerosol Sci. Technol. 50 (2016) 669–678. [28] J.B. Sim, S.H. Woo, W.G. Kim, S.J. Yook, J.B. Kim, G.N. Bae, H.H. Yoon, Performance estimation of a louver dust collector attached to the bottom of a subway train running in a tunnel, Aerosol Air Qual. Res. 17 (2017) 1954–1962.
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Enhancement of louver dust collector efficiency using modified dust container Jung-Bo Sim a, Un-Hak Yeo a, Gwang-Hun Jung a, Su-Beom Park a, Gwi-Nam Bae b, Se-Jin Yook a,⁎ a b
School of Mechanical Engineering, Hanyang University, Seoul 04763, Republic of Korea Korea Institute of Science and Technology, Seoul 02792, Republic of Korea
a r t i c l e
i n f o
Article history: Received 28 July 2017 Received in revised form 23 September 2017 Accepted 29 October 2017 Available online 03 November 2017 Keywords: Inertial separation Louver dust collector Dust container Collection efficiency
a b s t r a c t External air injected into a gas turbine contains many solid particles, which can reduce the performance and life of the turbine. In this study, a louver dust collector, which is a type of inertial dust collector, was used to remove solid particles from air, and the shape of the dust container, which is part of the louver dust collector, was modified to improve dust collection efficiency. As existing dust containers have a structure for isolating particles, the inflow of the air into the dust containers is limited and effective removal of particles is difficult. In this study, slits were drilled in the dust container, baffle plates were used, and raised spots were added to improve particle collection efficiency through improved air flow in the dust container and increased inertia effect of the particles. The trajectory of the particles and the collection efficiency for each dust container shape were predicted using numerical analysis and the numerical analysis results were verified using a wind tunnel test. Results indicate that for an air flow rate of 3 m3/min, the collection efficiencies of the louver dust collectors with the one-slit model dust container and two-slit model dust container improved by 40.1% and 43.5%, respectively compared with that of the louver dust collector with the existing dust container. Furthermore, for an air flow rate of 6 m3/min, the collection efficiencies of the louver dust collectors with the one-slit model dust container and two-slit model dust container improved by 32.9% and 37.6%, respectively, compared with that of the louver dust collector with the existing dust container. Therefore, it is expected that the particle collection efficiency of the existing louver dust collector can be effectively increased by utilizing the shape of the dust container proposed in this study. © 2017 Elsevier B.V. All rights reserved.
1. Introduction A gas turbine is a heat engine that drives a turbine using hightemperature and high-pressure combustion gas, and for its operation, external air is usually injected into the gas turbine. However, as external air contains many solid particles, its continuous inflow may cause corrosion of the compressor blades and damage to the engine, and possibly may lower the performance of the gas turbine [1–3]. Various studies have been conducted to prevent performance degradation due to various solid particles contained in the external air introduced into the gas turbine. The particle collection efficiency was evaluated through the fixed valve tray column for the removal of fly ash particles entering the gas turbine [4]. Coating treatment was performed to improve the erosion resistance and prevent gas turbine compressor blades from being eroded by solid particles [5,6]. A computational fluid dynamics (CFD) analysis was performed to identify the impact adhesion properties of particles and guidelines for filtration systems were prepared to prevent blade sediments and compressor performance degradation [7]. Several types of inertial separators were studied to protect gas ⁎ Corresponding author. E-mail address: [email protected] (S.-J. Yook).
https://doi.org/10.1016/j.powtec.2017.10.047 0032-5910/© 2017 Elsevier B.V. All rights reserved.
turbines from dust and sand [8]. The use conditions of an electrostatic precipitator for separating solid particles from combustion gas were investigated [9]. In this study, an inertial dust collector was considered to be installed in the secondary flow path of a gas turbine engine, to remove various solid particles floating in the external air entering the gas turbine. Inertial dust collectors collect dust using the inertia of the particles and can have a high dust collection efficiency under the condition that the gas moves at a high speed as in a gas turbine. Among them, the louver dust collector can effectively separate and remove solid particles using the louver blade in the flowing gas [10]. A louver refers to long plates, regardless of the material, continuously placed at a regular interval. There have been many studies to improve ambient air flow and to separate and remove particles in air using louvers. Dust particles in the combustion gas were removed through the circulating granular bed filter (CGBF) using louver plates [11]. The size distribution of dust particles was measured by the centrifugal separation method using a louver media filter, and particle collection efficiency was analyzed [12]. A fixed louver was installed in the burner duct of a pulverized coal power plant to separate solid particles from gas [13]. A louversublouver system was developed to remove fine dust particles and fugitive dust [14,15]. A device using a louver was installed on the
70
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
upper wall of a factory to improve indoor air ventilation efficiency [16]. The natural ventilation efficiency of the wind tower system was improved using louvers [17]. The angle of the louver was a factor that significantly influenced the dust collection efficiency of the CGBF [18]. Analytical, numerical, and experimental investigations of louvered classifiers demonstrated the influences of louver shape, louver angle, number of louvers, inlet velocity, powder feed-rate, and blowdown on fractional collection efficiency [19–21]. These parameters were also examined in an extensive literature on louvered classifiers in both rectilinear and cylindrical geometries. A partial review of that literature is provided in [22] together with a correlation of fractional collection efficiency, showing the influences of gas and particle dynamics and array geometry. Particularly relevant to the present study is the work of [23,24]. In that study, a louvered classifier with a dust container, analogous to the “original model” herein, was applied to dust removal in the secondary flow path of a gas turbine engine. The present study uses and improves upon the results of [24]. As mentioned earlier, many studies have been conducted to improve particle collection efficiency by changing the properties of louver blades; however, only few studies have attempted to improve the collection efficiency by changing the shape of the dust container. The purpose of this study is to improve the particle collection efficiency of the louver dust collector by changing the shape of the dust container. The louver dust collector design considered in this study was based on that of the previous literature [23,24]. Slits were drilled and additional baffle plates were installed in the dust container to increase the amount of particles collected in the dust container through improved air flow inside the dust container and maximized inertia effect of the particles. The air flow through each dust container shape was predicted using CFD simulation, and the particle collection efficiency of each dust container shape was predicted through the analysis of the particle trajectory. The particle collection efficiency of the louver dust collector was measured through a wind tunnel test and compared with the predicted collection efficiency from simulation. Furthermore, changes in the collection efficiency due to the application of grease were evaluated and analyzed for each dust container model. Results indicate that changing the dust container shape could improve particle collection efficiency and suppress particle re-scattering. Therefore, the collection efficiencies of existing louver dust collectors can be significantly
improved with low investment cost by replacing only their dust container shapes with those proposed in this study. 2. Numerical Fig. 1a shows the two-dimensional (2D) model of the louver dust collector used for numerical analysis. It consists of continuous louver blades and one dust container downstream. The louver was constructed using five continuously arranged straight louver blades with louver blade lengths of 46 mm and gaps of 15 mm. In this study, to compare the collection performances of different dust container shapes, the collection efficiency of the louver dust collector was predicted by changing the dust container shape while the louver blade shape remained unchanged. Fig. 1b shows the grid system used for 2D simulation of flow field and particle trajectory. Triangular grids were created using the Gambit software. After performing grid independence test, the louver dust collector comprised approximately 70,000 triangular grid cells and the inside of the dust container comprised about 10,000 triangular grid cells. Fig. 2a shows the existing dust container structure. It is a closed structure without gap except the entrance where dust is introduced through the louver blades (original model). The dust container of Fig. 2b has a slit in the left part of the original model (one-slit model). The width of the slit was 10 mm. A raised spot was added beneath the slit to prevent dust collected in the dust container from being rescattered. The one-slit model is identical to the original model of Fig. 2a except for the slit and the raised spot. Fig. 2c shows a dust container with 10-mm-width slits in the left and right parts of the original model dust container and a 30-mm baffle plate at the center of the dust container (two-slit model). Raised spots were also added beneath the left and right slits and the baffle plate to prevent particles collected in the dust container and the baffle plate from being rescattered. ANSYS FLUENT Release 16.1, a commercial CFD software, was used for numerical analysis. The flow inside the louver dust collector was assumed to be two-dimensional, steady, incompressible, and turbulent. Turbulent flow analysis was conducted using the standard k-ω turbulence model based on the Reynolds-averaged Navier–Stokes (RANS) model [23–24]. The boundary conditions for flow analysis
Fig. 1. Shape of louver dust collector: (a) Cross-sectional view; (b) Grid system for 2-dimensional simulation.
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
71
(Ncollected) was counted and the collection efficiency (η) of the louver dust collector was calculated using the equation: η¼
N collected : N injected
ð1Þ
For each tested particle size, Ninjected was set at 1000. 3. Experimental
Fig. 2. Dust container shapes: (a) Original model; (b) One-slit model; (c) Two-slit model.
were defined as follows: the velocity inlet condition was set at the entrance of the louver dust collector, the pressure outlet condition was set at the exit of the louver dust collector, and the louver blades as well as the dust container were set as walls. The SIMPLE algorithm was selected for pressure-velocity coupling, while the second-order upwind discretization scheme was selected to discretize the momentum [25]. The convergence criteria were set to 10−6 for continuity, momentum, and energy equations. The particle trajectory was analyzed using the discrete phase models (DPM) embedded in FLUENT [25]. Only the walls of the dust container were set as the ‘trap’ type boundary conditions, while the other walls, including louver blades, were set as the ‘reflect’ type boundary conditions [25]. In other words, the particles were assumed to be permanently trapped once they hit the dust container wall, while they were presumed to be bounced off once they hit the walls other than the dust container wall. The particle density was set to 2.65 g/cm3 considering the Arizona Test Dust that was used in the experiment. The particle was assumed to be spherical, and the drag acting on the particle was calculated using the slip correction factor according to the particle size. In addition, the Brownian motion of particles caused by the collision between the particles and air molecules was considered. In order to account for turbulent dispersion of particles, the stochastic random walk model provided in the FLUENT software was employed as described in [26]. After the injection of Ninjected particles from an equidistant position in the horizontal and vertical directions of the cross-sectional area of the flow path at the entrance of the louver dust collector, the number of particles collected in the dust container
Fig. 3a is a schematic diagram of the experimental setup for measuring the collection efficiency of the louver dust collector in a wind tunnel. A louver dust collector with the same size used in the numerical analysis model was installed in a wind tunnel with a cross sectional area of 0.01 m2 (0.1 m × 0.1 m). A photo of the louver dust collector, for example, with the one-slit model dust container, is shown in Fig. 3b. Air was injected using a turbofan. A HEPA filter was installed at the inlet of the wind tunnel to ensure that only particles used in the experiment enter the wind tunnel. The Arizona Test Dust (ISO 12103-1, A4 type) was dispersed into the air entering the wind tunnel through the HEPA filter, using a solid aerosol generator (SAG 410, TOPAS, Dresden, Saxony, Germany). The particles were injected in the opposite direction to the flow and there was sufficient distance for the development of turbulent flow so that the particles could be distributed evenly across the cross section of the wind tunnel. A honeycomb was used to obtain a uniform flow velocity distribution. The particle number concentrations by particle size were measured before and after the louver dust collector using an optical particle counter (OPC, Model 1.109, GRIMM, Ainring, Bayern, Germany). In addition, the inflow velocity of the air at the entrance of the louver dust collector was measured using a velocity meter (Model 9535, TSI, Shoreview, MN, USA), and the pressure drop through the louver dust collector was measured using a differential pressure transmitter (CP 210, KIMO, Montpon, Dordogne, France). Air velocities and particle number concentrations were measured at five points both in the horizontal and vertical directions (that is, a total of 25 points on the cross sectional area of the flow path at the entrance of the louver dust collector), and the measured values showed uniformities of ±5.5% and ±18.2% from the average values, respectively. The collection efficiency in the experimental method was obtained by η ¼ 1−
C downstream C upstream
ð2Þ
where Cupstream and Cdownstream are the particle number concentrations measured before and after the louver dust collector, respectively. 4. Results and discussion Through numerical analysis, the pressure distribution of the ambient air for each dust container model was examined at an air flow rate of 3 m 3/min. Fig. 4a shows the pressure distribution in the louver dust collector that used the original model dust container. Because there was almost no flow from inside the dust container to the outside, a relatively constant pressure was observed inside and around the dust container. On the other hand, in the case of the one-slit model of Fig. 4b, high pressure was observed inside the dust container, which spread around the dust container through the slit on the left side. This indicates that a large amount of air entered the dust container and then escaped through the slit. Similarly, for the two-slit model of Fig. 4c, a relatively high pressure was maintained inside the dust container, which spread around the dust container through the left and right slits. This indicates that a large amount of air was introduced to the dust container and then escaped through the left and right slits.
72
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
Fig. 3. Schematic diagram of the experimental setup for measuring the collection efficiency of the louver dust collector using a wind tunnel.
Table 1 presents a comparison of the numerical analysis results with wind tunnel test results in terms of the differential pressure before and after the louver dust collector according to the dust container model. As illustrated in Fig. 3a, the inlet pressure was measured before the louver blades and outlet pressure was measured after the dust container. The air flow rates were set at 3 m3/min and 6 m3/min. Results of the comparison indicate that the differential pressures were in good agreement with the maximum error of 3.4%. As the air flow rate increased twice, the differential pressure became roughly quadrupled. In terms of the differential pressure of each dust container model, the louver dust collector with the one-slit model showed the highest differential pressure of all the models. The original model and the two-slit model showed similar differential pressures, while a smaller differential pressure was observed in the louver dust collector with the two-slit model. Fig. 5 represents the trajectory of 5-μm particles in the louver dust collector when the air flow rate was 3 m3/min; the figure shows the
patterns in which particles were collected and released through each dust container model. It should be noted that Fig. 5 shows the trajectories of only 20 particles, while 1000 particles were tracked to numerically estimate the collection efficiency. Fig. 5a shows the trajectory of the particles passing through the louver dust collector with the original model dust container. It was observed that the injected particles moved upward toward the dust container. However, since there was almost no air flow through the dust container, the behavior of the particles to be introduced into the dust container was restricted. Therefore, it was confirmed that the particles floating in air could not be sufficiently introduced into the dust container and most particles were either collected near the entrance of the dust container or released from the louver dust collector without being collected. Fig. 5b shows the trajectory of the particles passing through the louver dust collector in the one-slit model dust container. Unlike in the case of Fig. 5a, many of the particles that passed over the louver blades were
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
73
Fig. 4. Pressure distribution around the dust container: (a) Original model; (b) One-slit model; (c) Two-slit model.
introduced into the dust container, and most of them could not escape the dust container due to inertia and were collected in the dust container as the flow was suddenly forced from the right to the left. Fig. 5c shows the trajectory of the particles passing through the louver dust collector with the two-slit model dust container. Since there were slits on the left and right sides of the dust container, it was observed that some of the particles that entered the dust container escaped through the slits whereas the others were collected in the dust container. Owing to the baffle plate installed at the center of the dust container,
Table 1 Differential pressures by dust container model: comparison of numerical analysis results with experimental results. Air flowrate (m3/min)
Dust container model
Differential pressure (Pa) Numerical analysis
Experimental measurement
3
Original model One-slit model Two-slit model Original model One-slit model Two-slit model
378 560 352 1695 2087 1612
365 ± 32 551 ± 34 340 ± 28 1642 ± 124 2064 ± 135 1575 ± 115
6
particles could not directly escape through the right slit and moved downward in the dust container. Then, the flow of the particles was suddenly forced toward the left and right slits, and particles were collected near the bottom of the dust container due to inertia. Fig. 6a shows the collection efficiencies of the louver dust collectors with the original model and the one-slit model dust containers. The numerical analysis results and the wind tunnel test results were in good agreement. When the air flow rate was 3 m3/min, the cut-off size (particle size corresponding to collection efficiency of 50%) was 8.41 μm for the original model and 5.04 μm for the one-slit model, confirming that the one-slit model had a 3.37 μm lower cut-off size. Fig. 6b compares the collection efficiencies of the louver dust collectors of the original model and two-slit model dust containers. When the air flow rate was 3 m3/min, the cut-off size of the two-slit model was 4.75 μm and it was 3.66 μm lower than that of the original model. Fig. 7a compares the collection efficiencies of the louver dust collectors of the original model and one-slit model dust containers when the air flow rate was 6 m3/min. The cut-off size was 5.35 μm for the original model and 3.59 μm for the one-slit model, confirming that the one-slit model had a 1.76 μm lower cut-off size. Fig. 7b compares the collection efficiencies of the louver dust collectors of the original model and two-slit model. The cut-off size of the two-slit model was 3.34 μm, which was 2.01 μm lower than the original model. As seen from Figs. 6 and 7, the one-slit model and two-slit model showed similar cut-off sizes and
74
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
Fig. 5. Trajectory of particles (5 μm): (a) Original model; (b) One-slit model; (c) Two-slit model.
both showed lower cut-off sizes than that of the original model. Therefore, it was confirmed that changing the dust container shape alone can improve the collection efficiency of the louver dust collector. The aforementioned experimental results of Figs. 6 and 7 were obtained from dust containers in which grease was applied to the walls to prevent the particles collected in the dust containers from being re-scattered. However, when a large number of particles are collected in the dust containers, the grease application effect could disappear and particle re-scattering problem could occur. Therefore, the collection efficiencies of the louver dust collector with and without grease application were measured and compared. Fig. 8 shows the results when the air flow rate into the louver dust collector was 3 m3/min, while Fig. 9 shows the results when the air flow rate was 6 m3/min. Figs. 8a and 9a show the collection efficiencies of the louver dust collector with the original model dust container, Figs. 8b and 9b show the collection efficiencies of the louver dust collector with the one-slit model, and Figs. 8c and 9c show the collection efficiencies of the louver dust collector with the two-slit model. In all cases, the collection efficiencies without grease application were lower than the collection efficiencies with grease application. In particular, when the particle size was larger than the cut-off size, the collection efficiency significantly decreased due to re-scattering. Furthermore, it was found that the collection efficiency decreased more due to re-scattering when the air flow rate was 6 m3/min than when the air flow rate was 3 m3/min. For the dust container shapes without grease application, a comparison of the degree of re-scattering revealed that the original model dust container showed the highest decrease in collection efficiency while the modified dust containers showed less re-scattering degrees, and hence higher collection efficiencies than the original model. In particular, the two-slit model dust container showed a significant reduction in particle
re-scattering under the same conditions compared with the one-slit model dust container, and its collection efficiency was relatively high even without grease application. Therefore, it was confirmed that improved dust container designs can enhance the collection efficiency of the louver dust collector, as well as suppress re-scattering of particles. However, by considering the particle re-scattering problem without grease application, it is needed to improve the collection efficiency further. The reason that the louver collector with the one-slit model or the two-slit model dust container showed improved collection efficiency compared with the original model dust container is that the flow that did not occur in the original model dust container was activated by drilling slits in the container and the particles were more likely to be collected in the dust container due to inertia when the flow was forced to change its direction to escape through the slits. Another reason is that the re-scattering of the particles was suppressed due to the presence of the baffle plate or raised spots. The differential pressure of the one-slit model dust container was approximately 1.2–1.5 times the differential pressure of the original model. This is because the pressure increased while air introduced into the dust container was forced to change its flow direction dramatically to escape through the slit located at a position opposite the inflow direction. Therefore, the collection efficiency of the louver dust collector was improved while the differential pressure increased. Meanwhile, the differential pressure of the two-slit model dust container was about 0.93–0.95 times the differential pressure of the original model. In other words, the two-slit model dust container showed the least differential pressure, the highest collection efficiency, and the least particle re-scattering of all the tested dust container models. This is because the pressure increased due to the left slit located at a position opposite the inflow direction, as well as the baffle plate
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
75
Fig. 7. Comparison of particle collection efficiency by dust container model (6 m3/min): (a) Original model and One-slit model; (b) Original model and Two-slit model. Fig. 6. Comparison of particle collection efficiency by dust container model (3 m3/min): (a) Original model and One-slit model; (b) Original model and Two-slit model.
located in the center, which was offset by the presence of the right slit located in the same direction as the inflow direction. It is expected that not only the one-slit model but also the two-slit model dust container can be applied to the louver dust collector, in which the airflow is caused by an external means such as a pump or a blower, to collect particles more effectively. The two-slit model dust container is also anticipated to be used in the louver dust collector for both securing the airflow rate and improving the collection efficiency when the ambient air is introduced into the louver dust collector in the absence of a pump or a blower, e.g., when the airflow is induced to enter into the louver dust collector by the running of a subway train as employed in recent studies [27,28], in which a louver dust collector was installed at the bottom of a subway train to remove coarse particles generated or re-suspended in the subway tunnel.
5. Conclusions This study was conducted in an effort to enhance the collection efficiency of a louver dust collector, of which design was based on that of the previous literature [23,24]. To improve the collection efficiency of the louver dust collector, the shape of the dust container, which is a component of the louver dust collector, was varied. The existing dust
container shape is not effective for dust collection because it simply isolates the particles introduced through the louver blades for collection. Therefore, in this study, slits were drilled in the existing dust container to induce flow into the dust container and the flow direction was sharply changed in the dust container to maximize the collection of the particles through inertia. The proposed one-slit model dust container had a slit located at a position opposite the direction of flow entering the container while the proposed two-slit model dust container had a slit located at a position opposite the direction of flow entering the container, another slit located in the same direction as the inflow direction, and a baffle plate for sharply changing the flow direction. In addition, raised spots were added around the slits to suppress the re-scattering of particles. For these three dust container models, the collection efficiencies predicted by numerical analysis and those obtained from measurements in a wind tunnel test were compared. The numerical analysis results and the wind tunnel test results were in good agreement for dust containers with grease application. It was confirmed that changing the dust container shape could improve the particle collection efficiency and suppress particle re-scattering. Therefore, it is expected that the collection efficiencies of the existing louver dust collectors will be significantly improved with low investment cost by replacing only their dust container shapes with those proposed in this study. Future studies need to be performed to develop more efficient dust container designs by considering particle re-scattering problem without grease application.
76
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
Fig. 8. Particle collection efficiencies with and without grease application (3 m3/min): (a) Original model; (b) One-slit model; (c) Two-slit model.
Fig. 9. Particle collection efficiencies with and without grease application (6 m3/min): (a) Original model; (b) One-slit model; (c) Two-slit model.
J.-B. Sim et al. / Powder Technology 325 (2018) 69–77
Acknowledgment The work was supported by research grants for the Railway Technology Research Project from the Ministry of Land, Infrastructure and Transport, Republic of Korea (17RTRP-B082486-04). References [1] P.F. Batcho, J.C. Moller, C. Padova, M.G. Dunn, Interpretation of gas turbine response due to dust ingestion, ASME, transactions, J. Eng. Gas Turb. Power. 109 (3) (1987) 344–352. [2] M.G. Dunn, C. Padova, J.E. Moller, R.M. Adams, Performance deterioration of a turbofan and a turbojet engine upon exposure to a dust environment, J. Eng. Gas Turb. Power. 109 (3) (1987) 336–343. [3] W. Tabakoff, W. Hosny, A. Hamed, Effect of solid particles on turbine performance, J. Eng. Power 98 (1) (1976) 47–52. [4] Q. Wang, X. Chen, X. Gong, Theoretical and experimental investigation on the characteristics of fly-ash scrubbing in a fixed valve tray column, AICHE J. 59 (6) (2013) 2168–2178. [5] S.A. Muboyadzhyan, Erosion-resistant coatings for gas turbine compressor blades, Russ. Metall. 2009 (3) (2009) 183–196. [6] A.V. Ryzhenkov, G.V. Kachalin, A.F. Mednikov, A.B. Tkhabisimov, The investigation of construction materials and protective coatings wear resistance to solid particle erosion, Mod. Appl. Sci. 9 (4) (2014) 85–95. [7] A. Suman, M. Morini, R. Kurz, N. Aldi, K. Brun, M. Pinelli, P.R. Spina, Estimation of the particle deposition on a subsonic axial compressor blade, J. Eng. Gas Turb. Power. 139 (1) (2017) 1–12. [8] M.R. Caskey, Dust and sand protection for marine gas turbines, J. Eng. Power 104 (2) (1982) 260–267. [9] J. Brandt, H. Heer, G.J. Clessuras, Criteria for electrostatic collection of unfavorable dusts, J. Eng. Power 97 (3) (1975) 461–464. [10] J.L. Smith Jr, M.J. Goglia, The mechanism of separation in the louver-type dust separator, J. Air Pollut. Control Assoc. 5 (1) (1955) 51–62. [11] J.C. Bai, S.Y. Wu, A.S. Lee, C.Y. Chu, Filtration of dust in a circulating granular bed filter with conical louver plates (CGBF-CLPs), J. Hazard. Mater. 142 (1) (2007) 324–331. [12] L.J. Czerwonka, J.M. Carey, Application of Particle-Size Analysis Data for the Determination of Air/Cleaner Performance. In ASME Gas Turbine Conference and Products Show 1965: American Society of Mechanical Engineers, 1965 (V001T01A006).
77
[13] G. Živković, S. Nemoda, P. Stefanović, P. Radovanović, Numerical simulation of the influence of stationary louver and coal particle size on distribution of pulverized coal to the feed ducts of a power plant burner, Therm. Sci. 13 (4) (2009) 79–90. [14] J. Smid, S.S. Hsiau, S.A. Tsai, C.C. Tzeng, Y.P. Chyou, Study on gravity flow of granules in beds supported by louver–sublouver system, Adv. Powder Technol. 20 (2) (2009) 127–138. [15] N. Arashi, T. Komuro, Y. Hishinuma, O. Kanda, H. Kuroda, Particle movements on louver supporting moving granular bed, Powder Technol. 43 (1985) 175–179. [16] J.H. Kang, S.J. Lee, Improvement of natural ventilation in a large factory building using a louver ventilator, Build. Environ. 43 (12) (2008) 2132–2141. [17] S. Liu, C.M. Mak, J. Niu, Numerical evaluation of louver configuration and ventilation strategies for the windcatcher system, Build. Environ. 46 (8) (2011) 1600–1616. [18] J.C. Bai, S.Y. Wu, A.S. Lee, Dust collection efficiency analysis in a two-dimensional circulating granular bed filter, J. Air Waste Manage. Assoc. 56 (5) (2006) 684–694. [19] H. Yoshida, K. Fukui, H. Morizaki, Particle classification of fly ash using a modified louver-type separator and reduction of unburned carbon amount, KONA Powder and Part. J. 18 (2000) 221–229. [20] M. Higashino, T. Maeda, K. Fukui, H. Yoshida, Improvement of particle separation in louver-type separator by use of numerical simulation, J. Soc. Powder Technol. Japan 52 (5) (2015) 252–259. [21] H. Yoshida, K. Fukui, K. Kanagawa, M. Okamoto, Y.U.N. Chi-Mun, K. Iinoya, Particle classification of a louver-type separator, J. Soc. Powder Technol. Japan 34 (9) (1997) 684–689. [22] T.R. Blake, Assessment of fractional collection efficiency in louvered inertial particle classifiers, Powder Technol. 311 (2017) 432–438. [23] G.O. Musgrove, M.D. Barringer, K.A. Thole, E. Grover, J. Barker, Computational design of a louver particle separator for gas turbine engines, ASME Turbo Expo 2009: Power for Land, Sea, and Air 2009, pp. 1313–1323. [24] G.O. Musgrove, K.A. Thole, E. Grover, J. Barker, Performance measurements of a unique louver particle separator for gas turbine engines, J. Eng. Gas Turb. Power. 135 012001 (2013) 1–10. [25] FLUENT 12.0 User's Guide, FLUENT Inc. April, 2009. [26] A. Dehbi, A. CFD, Model for particle dispersion in turbulent boundary layer flows, Nucl. Eng. Des. 238 (2008) 707–715. [27] K.R. Lee, W.G. Kim, S.H. Woo, J.B. Kim, G.N. Bae, H.K. Park, H. Yoon, S.J. Yook, Investigation of airflow and particle behavior around a subway train running in the underground tunnel, Aerosol Sci. Technol. 50 (2016) 669–678. [28] J.B. Sim, S.H. Woo, W.G. Kim, S.J. Yook, J.B. Kim, G.N. Bae, H.H. Yoon, Performance estimation of a louver dust collector attached to the bottom of a subway train running in a tunnel, Aerosol Air Qual. Res. 17 (2017) 1954–1962.