- Email: [email protected]
Development of Methods of Gas Flow Computation in Short Diffusers
Available online at www.sciencedirect.com
ScienceDirect Procedia Engineering 113 (2015) 259 – 263
International Conference on Oil and Gas Engineering, OGE-2015
Development of methods of gas flow computation in short diffusers Filkin N.Yu.a *, Yusha V.L.a , Litunov S.N.a a
Omsk State Technical University, 11, Mira Pr., Omsk 644050, Russian Federation
Abstract The paper suggests numerical and experimental computation methods of gas flow in short diffusers. The results of numerical investigations, carried out in ANSYS CFX for thirteen variants of turbulence, as well as experimental results are presented. For turbulence model k-ω, good repeatability of experimental data is obtained. The possibility of numerical computation methodology implementation while developing critical infrastructure systems with short diffusers is proved. © 2015 byby Elsevier Ltd.Ltd. This is an open access article under the CC BY-NC-ND license 2015The TheAuthors. Authors.Published Published Elsevier (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Omsk State Technical University. Peer-review under responsibility of the Omsk State Technical University
Keywords: methods; short diffuser; velocity profile; turbulence model; verification.
1. Introduction It is known that a diffuser is aimed to decrease gas dynamic looses thanks to the smooth transition from the smaller section of the flowing part to the bigger one. However, this idea works well until the angle of diffuser expansion is not more than 16-18 degrees. The more the angle of expansion, the higher is the pressure gradient in the diffuser near-wall part, which leads to the flow avulsion from diffuser walls; the more the angle of expansion is, the more the avulsion boundary moves towards the diffuser inlet. As a result, in a short diffuser with the big expansion angle (more than 90 degrees) a bigger part of the flow at the outlet moves mainly in the central part, which is the projection of the diffuser inlet [1]. Short diffusers are widely used in filters and heat exchangers of the critical infrastructure systems, where the requirement of providing uniform flow at the inlet diffuser section and reduction of overall dimensions are relevant [2]. There are various technical solutions and recommendations on the diffuser work improvement [3, 4], but the influence of these structural modifications on the functional efficiency of critical infrastructure systems elements is
* Corresponding author. Tel.: +7-923-674-2093. E-mail address: [email protected]
1877-7058 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Omsk State Technical University
doi:10.1016/j.proeng.2015.07.332
260
N. Yu. Filkin et al. / Procedia Engineering 113 (2015) 259 – 263
understudied. The solving of this task requires carrying out gas dynamic research with the aim of obtaining recommendations on designing flow part of critical infrastructure systems elements with a short diffuser. Nowadays, various programs are widely used for carrying out gas dynamic researches, one of them being ANSYS CFX. They provide solving a wide range of tasks, though, in every particular case, there is a question about the reliability of the obtained results, which depends on the characteristic of the object [5-8]. At the same time, there is some data [9], proving the significant difference in quantitative results of diffuser calculations and a good qualitative results of the gas stream flow characteristics. That is why the task of development methodology of gas flow computation in short diffusers is up-to-date and is the object of this paper. 2. Study subject The subject of the study is velocity profile of the air stream in the flow part, containing a short diffuser. The flow moves along the straight pipeline, and then to the flow part of this or that device or unit (heat exchanger, filter and so on), consisting of a short diffuser and a body (fig. 1). The first alternative is that the diffuser flow part cannot have any additional elements; the second alternative is that some additional equally spaced cone guides can be placed there. Velocity profile is considered in the pipeline at some distance from the diffuser inlet (section а-а), as well as after the diffuser (section b-b).
Fig. 1. Model for carrying out numerical research: а-а – section, where velocity profile before the diffuser was measured; b-b – section, where velocity profile after the diffuser was visualized.
3. Methods 3.1. Experimental methods With the purpose of computational model verification experimental test bench with the analogous to the computational model geometry and layout was designed and made (fig. 2). Methodology of the experimental research consists of two parts: measuring velocity profile before the diffuser and visualizing velocity profile after the diffuser.
N. Yu. Filkin et al. / Procedia Engineering 113 (2015) 259 – 263
Fig. 2. Experimental test bench for the verification of the computational model: а-а – section, where velocity profile before the diffuser was measured ; b-b – section, where velocity profile after the diffuser was visualized; 1 – static pressure tube; 2 – total pressure tubes; 3 – short diffuser; 4 – concentric guides; 5 – case with fabric.
During the first part, air flow was supplied into the pipeline, the pressure being atmospheric, the temperature being 25 °С and the volume flow rate being 65 m3/ hour. With the digital differential pressure gauge, one end of which was connected to the static pressure tube, while another one was alternately connected to each of total pressure 19 tubes, pressure drop in corresponding part of a pipeline section was measured. Then, according to the measured pressure drop, flow speed in the corresponding flow point was calculated [10]. Velocity profile was developed on the basis of calculation results for the total pressure central tube, located on the pipeline axis, as well as on the basis of average values, obtained for each of the three diameters, where the measurements were carried out by the rest total pressure tubes. During the second part, the pipeline area with tubes was replaced by the analogous one without tubes, but with a device of contaminant supply (fine breeze). After the diffuser the case with air-permeable white fabric which entrapped coal particles on its surface was placed. Supposing, that the path of particles coincides with the air trajectory, according to the profile of the entrapped particles on the fabric surface it is possible to tell the character of the air at the diffuser outlet. Visualization was made for the diffuser with equally spaced cone guides and without them. Flow parameters are the following: volume flow rate is 65 m3/hour, temperature is 25 °С, pressure is atmospheric. 3.2. Computational method Numerical research was carried out in ANSYS CFX package on the basis of the designed and developed 3D model of the study subject, shown in fig. 1. Computation included the following stages: geometrical model import; dividing the model into the final elements net; specification of initial and boundary conditions; calculation; analyzing the results. Input boundary condition – average flow velocity Vm, calculated on the basis of the experimental data. Output boundary condition is pressure, equal to atmospheric. Working object is air with temperature 25° С; the flow is isometric. The research was carried out for thirteen turbulence types: SST, none (Laminar), k-ε, BSL Reynolds Stress, SSG Reynolds Stress, Zero Equation, RNG k-ε, k-ω, Eddy Viscosity, BSL, LRR Reynolds Stress, QI Reynolds Stress, Omega Reynolds Stress. For each turbulence type before the diffuser in points, corresponding to the total pressure tap-off points at the test bench, flow velocity was measured, at the diffuser output velocity profile was developed for two variants of the diffuser (with guides and without them).
261
262
N. Yu. Filkin et al. / Procedia Engineering 113 (2015) 259 – 263
4. Results and discussion Verification of the computational method was carried out on the basis of the comparison of the velocity profiles, measured in the pipeline before the diffuser, and obtained visually just after the diffuser. The research itself was carried out in two variants simultaneously: with cone guides, equally placed in the diffuser flow part, and without them. When comparing velocity profiles (fig. 3) it was found out, that most close to the experimental ones, computational values were obtained for turbulences SST, BSL Reynolds Stress, k-ω, BSL и Omega Reynolds Stress, from which the least расхождение (no more than 1.6%) corresponds to k-ω turbulence. Comparison of the visual flow stream profile at the diffuser outlet (fig. 4) obtained for turbulence k-ω and in the experiment has shown likewise flow movement profile.
Fig. 3. Veocity profile, obtained after numerical and experimental computation.
The results given in the paper prove the verification of calculation methodology and its applicability for numerical research and short diffusers design. 5. Conclusion Summing up, the methodology of gas flow computation in short diffusers is developed. Comparison analysis of the numerical and experimental results obtained has proved the applicability of numerical computation methodology with using turbulence model k-ω, the discrepancy with the experiment being no more than 1.6%. This let us underline the possibility of its implementation while designing short diffusers, including those of the critical infrastructure systems.
N. Yu. Filkin et al. / Procedia Engineering 113 (2015) 259 – 263
Fig. 4. Visualization of the flowstream at the diffuse output, obtained after numerical (up) and experimental computation (doen) for short diffuser with с guides (right) and without them (left).
References [1] Spravochnik po gidravlicheskim soprotivleniyam, pod red. М.О. Shteinberg, М.: Mashinostroenie, 1992, 672 p. (in Russian). [2] V. А. Anan’ev., L.N. Balueva, V.P. Murashko, Systemy ventilyatsii i konditsionirovaniya. Teoriya i praktika, Moscow, Tehnosphera: Evroclimat, 2008, 503 p. (in Russian). [3] B. Sahin, A. J. Ward-Smith, The use of perforated plates to control the flow emerging from a wide-angle diffuser, with application to electrostatic precipitator design. International Journal of Heat and Fluid Flow. 8 (1987) 124-131. [4] A. I. Reshmin, S. Kh. Teplovodskii, V. V. Trifonov. Short round diffuser with a high area ratio and a permeable partition, Fluid Dynamics, 5 (2012) , pp. 583–589. [5] J. Feng, F.-K. Benra, H. J. Dohmen, Qualitative comparison between numerical and experimental results of unsteady flow in a radial diffuser pump, Journal of Visualization. Volume 10, Issue 4, 349-357. [6] J. Feng, F.-K. Benra, H. J. Dohmen, Unsteady flow visualization at part-load conditions of a radial diffuser pump: by PIV and CFD, Journal of Visualization. Volume 12, Issue 1, 65-72. [7] Jianjun Feng, F.-K. Benra, H. J. Dohmen, Application of Different Turbulence Models in Unsteady Flow Simulations of a Radial Diffuser Pump, Forschung im Ingenieurwesen. Volume 74, Issue 3, 123-133. [8] R.A. Van den Braembussche, J. Prinsier, A. Di Sante, Experimental and numerical investigation of the flow in rotating diverging channels, Journal of Thermal Science. Volume 19, Issue 2, 115-119. [9] Yu.F. Kashkin, A.E. Konovalov, S.Yu. Krasheninnikov, D.A. Lyubimov, D.E. Pudovikov, V.A. Stepanov, Experimental and numerical investigation of separated flows in subsonic diffusers, Fluid Dynamics. Volume 44, Issue 4, 555-565. [10] S.V. Biryukov, A.I. Cheredov, Metody i sredstva izmereniy: uchebnoe posobie, Omsk, OmSTU, 2001. (in Russian).
263
ScienceDirect Procedia Engineering 113 (2015) 259 – 263
International Conference on Oil and Gas Engineering, OGE-2015
Development of methods of gas flow computation in short diffusers Filkin N.Yu.a *, Yusha V.L.a , Litunov S.N.a a
Omsk State Technical University, 11, Mira Pr., Omsk 644050, Russian Federation
Abstract The paper suggests numerical and experimental computation methods of gas flow in short diffusers. The results of numerical investigations, carried out in ANSYS CFX for thirteen variants of turbulence, as well as experimental results are presented. For turbulence model k-ω, good repeatability of experimental data is obtained. The possibility of numerical computation methodology implementation while developing critical infrastructure systems with short diffusers is proved. © 2015 byby Elsevier Ltd.Ltd. This is an open access article under the CC BY-NC-ND license 2015The TheAuthors. Authors.Published Published Elsevier (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Omsk State Technical University. Peer-review under responsibility of the Omsk State Technical University
Keywords: methods; short diffuser; velocity profile; turbulence model; verification.
1. Introduction It is known that a diffuser is aimed to decrease gas dynamic looses thanks to the smooth transition from the smaller section of the flowing part to the bigger one. However, this idea works well until the angle of diffuser expansion is not more than 16-18 degrees. The more the angle of expansion, the higher is the pressure gradient in the diffuser near-wall part, which leads to the flow avulsion from diffuser walls; the more the angle of expansion is, the more the avulsion boundary moves towards the diffuser inlet. As a result, in a short diffuser with the big expansion angle (more than 90 degrees) a bigger part of the flow at the outlet moves mainly in the central part, which is the projection of the diffuser inlet [1]. Short diffusers are widely used in filters and heat exchangers of the critical infrastructure systems, where the requirement of providing uniform flow at the inlet diffuser section and reduction of overall dimensions are relevant [2]. There are various technical solutions and recommendations on the diffuser work improvement [3, 4], but the influence of these structural modifications on the functional efficiency of critical infrastructure systems elements is
* Corresponding author. Tel.: +7-923-674-2093. E-mail address: [email protected]
1877-7058 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Omsk State Technical University
doi:10.1016/j.proeng.2015.07.332
260
N. Yu. Filkin et al. / Procedia Engineering 113 (2015) 259 – 263
understudied. The solving of this task requires carrying out gas dynamic research with the aim of obtaining recommendations on designing flow part of critical infrastructure systems elements with a short diffuser. Nowadays, various programs are widely used for carrying out gas dynamic researches, one of them being ANSYS CFX. They provide solving a wide range of tasks, though, in every particular case, there is a question about the reliability of the obtained results, which depends on the characteristic of the object [5-8]. At the same time, there is some data [9], proving the significant difference in quantitative results of diffuser calculations and a good qualitative results of the gas stream flow characteristics. That is why the task of development methodology of gas flow computation in short diffusers is up-to-date and is the object of this paper. 2. Study subject The subject of the study is velocity profile of the air stream in the flow part, containing a short diffuser. The flow moves along the straight pipeline, and then to the flow part of this or that device or unit (heat exchanger, filter and so on), consisting of a short diffuser and a body (fig. 1). The first alternative is that the diffuser flow part cannot have any additional elements; the second alternative is that some additional equally spaced cone guides can be placed there. Velocity profile is considered in the pipeline at some distance from the diffuser inlet (section а-а), as well as after the diffuser (section b-b).
Fig. 1. Model for carrying out numerical research: а-а – section, where velocity profile before the diffuser was measured; b-b – section, where velocity profile after the diffuser was visualized.
3. Methods 3.1. Experimental methods With the purpose of computational model verification experimental test bench with the analogous to the computational model geometry and layout was designed and made (fig. 2). Methodology of the experimental research consists of two parts: measuring velocity profile before the diffuser and visualizing velocity profile after the diffuser.
N. Yu. Filkin et al. / Procedia Engineering 113 (2015) 259 – 263
Fig. 2. Experimental test bench for the verification of the computational model: а-а – section, where velocity profile before the diffuser was measured ; b-b – section, where velocity profile after the diffuser was visualized; 1 – static pressure tube; 2 – total pressure tubes; 3 – short diffuser; 4 – concentric guides; 5 – case with fabric.
During the first part, air flow was supplied into the pipeline, the pressure being atmospheric, the temperature being 25 °С and the volume flow rate being 65 m3/ hour. With the digital differential pressure gauge, one end of which was connected to the static pressure tube, while another one was alternately connected to each of total pressure 19 tubes, pressure drop in corresponding part of a pipeline section was measured. Then, according to the measured pressure drop, flow speed in the corresponding flow point was calculated [10]. Velocity profile was developed on the basis of calculation results for the total pressure central tube, located on the pipeline axis, as well as on the basis of average values, obtained for each of the three diameters, where the measurements were carried out by the rest total pressure tubes. During the second part, the pipeline area with tubes was replaced by the analogous one without tubes, but with a device of contaminant supply (fine breeze). After the diffuser the case with air-permeable white fabric which entrapped coal particles on its surface was placed. Supposing, that the path of particles coincides with the air trajectory, according to the profile of the entrapped particles on the fabric surface it is possible to tell the character of the air at the diffuser outlet. Visualization was made for the diffuser with equally spaced cone guides and without them. Flow parameters are the following: volume flow rate is 65 m3/hour, temperature is 25 °С, pressure is atmospheric. 3.2. Computational method Numerical research was carried out in ANSYS CFX package on the basis of the designed and developed 3D model of the study subject, shown in fig. 1. Computation included the following stages: geometrical model import; dividing the model into the final elements net; specification of initial and boundary conditions; calculation; analyzing the results. Input boundary condition – average flow velocity Vm, calculated on the basis of the experimental data. Output boundary condition is pressure, equal to atmospheric. Working object is air with temperature 25° С; the flow is isometric. The research was carried out for thirteen turbulence types: SST, none (Laminar), k-ε, BSL Reynolds Stress, SSG Reynolds Stress, Zero Equation, RNG k-ε, k-ω, Eddy Viscosity, BSL, LRR Reynolds Stress, QI Reynolds Stress, Omega Reynolds Stress. For each turbulence type before the diffuser in points, corresponding to the total pressure tap-off points at the test bench, flow velocity was measured, at the diffuser output velocity profile was developed for two variants of the diffuser (with guides and without them).
261
262
N. Yu. Filkin et al. / Procedia Engineering 113 (2015) 259 – 263
4. Results and discussion Verification of the computational method was carried out on the basis of the comparison of the velocity profiles, measured in the pipeline before the diffuser, and obtained visually just after the diffuser. The research itself was carried out in two variants simultaneously: with cone guides, equally placed in the diffuser flow part, and without them. When comparing velocity profiles (fig. 3) it was found out, that most close to the experimental ones, computational values were obtained for turbulences SST, BSL Reynolds Stress, k-ω, BSL и Omega Reynolds Stress, from which the least расхождение (no more than 1.6%) corresponds to k-ω turbulence. Comparison of the visual flow stream profile at the diffuser outlet (fig. 4) obtained for turbulence k-ω and in the experiment has shown likewise flow movement profile.
Fig. 3. Veocity profile, obtained after numerical and experimental computation.
The results given in the paper prove the verification of calculation methodology and its applicability for numerical research and short diffusers design. 5. Conclusion Summing up, the methodology of gas flow computation in short diffusers is developed. Comparison analysis of the numerical and experimental results obtained has proved the applicability of numerical computation methodology with using turbulence model k-ω, the discrepancy with the experiment being no more than 1.6%. This let us underline the possibility of its implementation while designing short diffusers, including those of the critical infrastructure systems.
N. Yu. Filkin et al. / Procedia Engineering 113 (2015) 259 – 263
Fig. 4. Visualization of the flowstream at the diffuse output, obtained after numerical (up) and experimental computation (doen) for short diffuser with с guides (right) and without them (left).
References [1] Spravochnik po gidravlicheskim soprotivleniyam, pod red. М.О. Shteinberg, М.: Mashinostroenie, 1992, 672 p. (in Russian). [2] V. А. Anan’ev., L.N. Balueva, V.P. Murashko, Systemy ventilyatsii i konditsionirovaniya. Teoriya i praktika, Moscow, Tehnosphera: Evroclimat, 2008, 503 p. (in Russian). [3] B. Sahin, A. J. Ward-Smith, The use of perforated plates to control the flow emerging from a wide-angle diffuser, with application to electrostatic precipitator design. International Journal of Heat and Fluid Flow. 8 (1987) 124-131. [4] A. I. Reshmin, S. Kh. Teplovodskii, V. V. Trifonov. Short round diffuser with a high area ratio and a permeable partition, Fluid Dynamics, 5 (2012) , pp. 583–589. [5] J. Feng, F.-K. Benra, H. J. Dohmen, Qualitative comparison between numerical and experimental results of unsteady flow in a radial diffuser pump, Journal of Visualization. Volume 10, Issue 4, 349-357. [6] J. Feng, F.-K. Benra, H. J. Dohmen, Unsteady flow visualization at part-load conditions of a radial diffuser pump: by PIV and CFD, Journal of Visualization. Volume 12, Issue 1, 65-72. [7] Jianjun Feng, F.-K. Benra, H. J. Dohmen, Application of Different Turbulence Models in Unsteady Flow Simulations of a Radial Diffuser Pump, Forschung im Ingenieurwesen. Volume 74, Issue 3, 123-133. [8] R.A. Van den Braembussche, J. Prinsier, A. Di Sante, Experimental and numerical investigation of the flow in rotating diverging channels, Journal of Thermal Science. Volume 19, Issue 2, 115-119. [9] Yu.F. Kashkin, A.E. Konovalov, S.Yu. Krasheninnikov, D.A. Lyubimov, D.E. Pudovikov, V.A. Stepanov, Experimental and numerical investigation of separated flows in subsonic diffusers, Fluid Dynamics. Volume 44, Issue 4, 555-565. [10] S.V. Biryukov, A.I. Cheredov, Metody i sredstva izmereniy: uchebnoe posobie, Omsk, OmSTU, 2001. (in Russian).
263