カテゴリー Ⅰ 日本建築学会環境系論文集第 8 巻第 75 号,75-758,5 年 9 月 J. Environ. Eng., AIJ, Vol. 8 No. 75, 75-758, Sep., 5 アクリル製人体気道モデルを用いた気道内流れ場の PIV 計測と CFD 解析 FLOW FIELD MEASUREMENTS IN ACRYLIC RESPIRATORY TRACT MODEL BY PIV AND CFD 平瀬公太 *, 山下真登 *, 荒巻森一朗 ** ***, 伊藤一秀 Kota HIRASE, Masato YAMASHITA, Shin-ichiro ARAMAKI and Kazuhide ITO An investigation of air flow characteristics in the human respiratory tract will provide essential information to enhance understanding transportation of inhaled contaminants through respiration. In this study, we conducted in vitro experiments to investigate the flow pattern in the human respiratory airway model. The detail measurement by particle image velocimetry (PIV)technique is challenge in the field of respiratory infection in indoor environment. In this study, flow fields in trachea region were precisely measured by PIV under three inhalation conditions; 7.5 L/min, 5 L/min and L/min. Keywords : Respiratory Tract, Exposure analysis, PIV. VOCsPM ) Virtual Manikin4 Virtual Airway ) K. Inthavong and J. Tu,4) (CT DICOM)CAD CTD PIV () CT CFD CTD PIV 5,6) 7.5 L/min 5L/minL/minPIV Rek-(Abe Kondoh Nagano model) 7) RNG k- 8) SST k- 9,) * ** *** 九州大学大学院 総合理工学府 修士課程 九州大学大学院 総合理工学研究院 助教 工博 九州大学大学院 総合理工学研究院 准教授 工博 Master Course Student, IGSES, Kyushu iv. Assist. Prof., IGSES, Kyushu iv., Dr.Eng. Assoc. Prof., IGSES, Kyushu iv., Dr.Eng. 75
CFD 5,6). (Drug Delivery System; DDS) ) ) PIV 5,6) 5,6) ((7cm4)) 4 (STL) D() PIV ) (Trachea)PIV Fig. Fig. Respiratory Tract (Airway) Model () Respiratory Tract Model for in Vitro Experiment. PIV. (PIV) PIV m () ) CMOS (Photron FASTCAM APX) 54pixels(frame rate)5 fps(cw)w Gr een Laser (Beamtech Optronics, Diode- pumped solid state (DPSS) Green Laser W5 nm)6mm5mm mmpivdantech Dynami c Studio.Inter- rogation window pixelsoverlap 5%Adaptive (FFT) Cross- Corr elation Method PIV Fig. () Experimental Set Up Fig. Outline of PIV Measurement Table Numerical and Boundary Conditions Low Re Type k- model (AKN Model) Turbulence RNG k- model Model SST k- model Mesh 7.5 million mesh (structured, Tetra) Algorithm SIMPLE (steady state) Scheme Convection TermQUICK U in =.475 m/s (Nasal),.6 (Oral) (7.5L/min) U Inflow in =.95 m/s(nasal),.65 (Oral) (5L/min) U boundary in =.9 m/s(nasal),. (Ora;) (L/min) TI= % k in =/ (U in.), in =C /4 μ k / in l in, Outflow U out =Free slip, boundary k out =Free slip, out =Free slip Wall treatment Velocity: no slip, 75
( ). PIVPIV Fig. PIV J IS K 8 ) Cannon-Fenske Routine Viscometers =.747-6 m /s=.47 kg/m Refractive In dex (RI)( 7.9%8.%5) PIV 5 Reynolds() ( ) Q mixture mixture Q () air air Q mixture Q air (Kofloc model RK4). (Q air 7.5 L/min5 L/min L/min),4) Q mixture.8l/min (Re=5).6L/min (Re=).L/min (Re=) PIV 4. Rek-(Abe Kondoh Nagano ) 7) RNG k- 8) SST k- 9,) Rek-(Abe Kondoh Nagano ) k- RNG k- () 7.5 L/min () 5 L/min () L/min (a) Oral inhalation () 7.5 L/min () 5 L/min () L/min (b) Nasal inhalation Fig. Plots of two dimensional mean velocity vectors in central sagittal plane of oral inhalation (a) and nasal inhalation (b) (RNG) k-(c, C )k SST k- k-k- k- t SIMPLE QUICK No-slip PIV 7.5L/min5L/minL/min ((.88 m.9 m )U in.475 m/s) TI=% 5) Table 75
(a) Qair =7.5L/min (b) Qair =5L/min (a) Qair =7.5L/min. (b) Qair =5L/min (c) Qair =L/min Fig.4 Scalar Velocity Distributions by PIV and CFD under Oral Breathing Condition (c) Qair =L/min Fig.5 Scalar Velocity Distributions by PIV and CFD under Nasal Breathing Condition ࡋࡓ Ẽ㐨 ᗄఱᙧ ࢩ ศ 㸪ቃ ᮲ ୧ᅗ 㰯 ᮲ ᮲ Trachea(Ẽ ) ࡅ ᖹ ሗ) 㸬 ࢩ 㸪 ࢩ ࢨ ᆒ㢼 ศᕸ ィ ᮲ (Fig. 4) 㸪 㐣ࡋࡓ Ẽࡀဗ㢌㒊 ᛴ ゅ ὶ ሙゎᯒ ᗘ ๓ ド( ࢩ ౫Ꮡᛶ ウ) ᐇ ࡋࡓ ࢩ Ỵᐃࡋ ὀ)㸬ࡇ ヲ 5,6) ሗ ヲࡋ 㸬 ᯝ ࡋ 㸬 ᗘ ኚ Ẽ ὶධࡍ 㸬PIVィ CFDゎᯒ ࡋࡓẼ 㐨 ᗄఱᙧ 㸪ဗ㢌㒊 Ẽ ᥋ 㒊 ࢠ ࡀᏑ 5. PIV ィ ᯝ ゎᯒ ᯝ PIV Trachea(Ẽ )㒊ศ ὶ ሙ ᅾࡋ (Fig.ཧ )㸪ࡑ 㒊ศ 㞳ࡀ ࡌ 㸬ࡑ ᙳ㡪 ᐃ ᯝ 㢼 ࢡ PIVィ 㡿ᇦ 㒊㡿ᇦ 㒊 ᑐ ప ᇦࡀ ㄆࡉ 㸬PIV ࡋ Fig. ࡍ㸬Ẽ 㒊 㒊 ࡅ ᵝὶࡀᙧᡂࡉ ィ ぢ 㸪 ὶ(㰯 ᪉ )ࡀ ࡍ ὶ ሙ ࢫ 㢼 ศᕸ ࡁ ᕪ ㄆ 㸬 ᅇ ὶධὶ㔞ኚ ࡗ 㸬 㔞7.5L/min ࢣ ࢫ 5fps タᐃ᮲ Re ౫Ꮡᛶ ㄆࡉ ࡗࡓ㸬Reᆺk-İ ᯝ 7.5L/min㸪5L/min㸪L/min 㔞᮲ ḟඖ ᙳࡋࡓPIV ࢪ4ᯛ 㸪ᖹᆒ ฎ ࡗࡓ㸬ࡇ ᐇ (Abe Kondoh Nagano )㸪RNG k-i 㸪SST k-ȧ 㛫ࢫࢣ 8sec 㛫ᖹᆒ ᙜࡍ (7.5L/min ᮲ Ẽ㐨య ὶ ゎᯒ 㸪 ὶ 㛫 ὶ ሙゎᯒ ᯝ ὶධ Ẽ㔞 ฟࡋࡓ Ẽ 㛫.4sec)㸬 ᕪ㐪ࡀ ㄆࡉ 㸪 7.5L/min ࢣ ࢫ Reᆺk-İ ᯝࡀ ࡓ㸪PIVᐇ㦂 ᯝ CFD 㒊ศ ὶ ሙゎᯒ ᯝ 㧗㢼 㡿ᇦ ᑠࡉࡃ㸪SST k-ȧ ᯝࡀ㧗㢼 㡿ᇦ (ὶධ㢼 ḟඖࡋࡓࢫ 㢼 ศᕸ) Fig.4 Fig.5 ࡍ㸬 ࡁࡃホ౯ࡋࡓ㸬 Re ࡁࡃࡋࡓ5L/min㸪L/min ࢣ ࢫ 754
.5 RNG Low Re.5 SST.5.5.5 () S Line.5 PIV.5.5.5.5 () S Line.5.5.5.5.5 () S Line Fig.6 Profiles of normalized scalar velocity under Oral Breathing Condition (Left: 7.5L/min, Center: 5L/min, Right: L/min) ( U U in, U U x U y,).5 RNG.5.5 Low Re.5.5 () S Line.5 SST.5.5.5.5 () S Line.5.5.5.5.5 () S Line Fig.7 Profiles of normalized scalar velocity under Nasal Breathing Condition (Left: 7.5L/min, Center: 5L/min, Right: L/min) PIV Re PIV (Fig. 5)Fig.4 Reynold7.5L/min47 5L/min874L/min747Re PIV Rek-(Abe Kondoh Nagano )RNG k- SST k- 7.5L/minRNG k-rek- SST k- CFD Re Fig.6Fig.7 PIVCFD S S mmsmms Fig. 4Fig. 5 ( ) Fig. 6S 7.5L/min5L/minRNG k- PIV Rek-SST k- Rek-SST k- L/minSST k-piv SS Rek-PIV Fig. 77.5L/min SSRe k-piv 5L/minL/min Rek-L/min PIV7.5L/min 6. PIV 755
6 7, 8) PIV PIV 9 PIV Re k- (Abe Kondoh Nagano ) CFD 4) PIV PIV CFD PIV PIV PIV PIV 7. PIV CFD () Re k- RNG k- SST k- CFD 7.5L/min Re k- PIV () (5L/min L/min) PIV PIV CFD () Re Re k- (Abe- Kondoh-Nagano model) PIV PIV Nguyen Lu Phuong.5.5.5 () S Line (4) S Line (5) S Line Fig.8 Flow Pattern in Idealized Cylindrical model by PIV and CFD (Red circle: PIV, Black line: Low Re k-, ( U U in, U U x U y ) ( 6765) ) Objet Connex 5 (z )6m 6dpi(x)6dpi(y)6dpi(z) Objet VeroClear RGD 8 ) Sphericel P8 m (5m )..49 g/cc ) L/min 95% y+< 4) 9 Fig.8 7.5L/min PIV Re k- PIV Fig 7 ()() PIV CFD Fig 7 ()(5) Fig 7 () S S PIV CFD PIV CFD () CFD Benchmark test -) ), YAKUGAKU ZASSHI, Vol. 7, No., pp 49-46, 7 ) Sung-Jun Yoo Vol.8No. 79pp9-85. ) Xiangdong Li, Kiao Inthavong, Jiyuan Tu : Particle inhalation and deposition in a human nasal cavity from the external surrounding environment, Building and S S S 756
Environment, 47, pp 9, 4) Kiao Inthavong, Qin Jiang Gea, Xiang Dong Lia, Ji Yuan Tua : Detailed predictions of particle aspiration affected by respiratory inhalation and airflow, Atmospheric Environment, 6, pp7 7, 5) Nguyen Lu Phuong No.9pp-9 6) Nguyen Lu Phuong PIV CFD No.7pp-74 7) Abe, K., Kondoh, T., Nagano, Y.: A new turbulence model for predicting fluid flow and heat transfer in separating and reattaching flows II. Thermal field calculations. International Journal of Heat and Mass Transfer 8, pp467-48, 995 8) Yakhot, V. and Orszag, S. A. : Renormalization group analysis of turbulence. I. Basic theory. Journal of Scientific Computing, pp-5, 986 9) Wilcox, D.: Turbulence Modeling for CFD. DCW Industries, Inc., 554 Palm Drive, La Canada, California 9, 99 ) Menter, F. R., Kuntz, M., Langtry, R. B. : Ten Years of Industrial Experience with the SST Turbulence Model, in Turbulence, Heat and Mass Transfer 4, K. Hanjalic, Y. Nagano, M. Tummers, eds., Begell House Inc.,pp 65-6, ) Tu, J.Y., Inthavong, K. Ahmadi, G. : Computational Fluid and Particle Dynamics in the Human Respiratory System, Springer Publishing, Heidelberg, DE. ISBN: 978-94744875, ) JIS K 8, ) Brooks, G. A., Fahey, T. D., White, T. P. Exercise physiology: human bioenergetics and its applications. Mayfield Publishing Company. 996 4) J Tu, I Kiao and G Ahmadi, Computational Fluid and Particle Dynamics in the Human Respiratory System, Springer, ISBN 978-94-7-4487-5, 5) B. Ma and K R. Lutchen, CFD Simulation of Aerosol Deposition in an Anatomically Based Human Large Medium Airway Model, Annals of Biomedical Engineering, Vol. 7, No., pp.7 85, 9 6),,,, PIV, vol9no.pp.8-49. 7),,,,, PIV, 69 pp.6-68.8 8) CFD 77, 68, pp86-97,. 9) E. R. Weibel, Morphometry and lung models, Springer, ISBN 978--64-5-6, 967 ) Ito K, Inthavong K, Kurabuchi T, Ueda T, Endo T, Omori T, Ono H, Kato S, Sakai K, Suwa Y, Matsumoto H, Yoshino H, Zhang W, Tu J. CFD Benchmark Tests for Indoor Environmental Problems: Part Isothermal/non-isothermal flow in D and D room model, International Journal of Architectural Engineering Technology, Vol., No., pp-, 5 ) Ito K, Inthavong K, Kurabuchi T, Ueda T, Endo T, Omori T, Ono H, Kato S, Sakai K, Suwa Y, Matsumoto H, Yoshino H, Zhang W, Tu J. CFD Benchmark Tests for Indoor Environmental Problems: Part Cross-ventilation airflows and floor heating systems, International Journal of Architectural Engineering Technology, Vol., No., pp-49, 5 ) Ito K, Inthavong K, Kurabuchi T, Ueda T, Endo T, Omori T, Ono H, Kato S, Sakai K, Suwa Y, Matsumoto H, Yoshino H, Zhang W, Tu J. CFD Benchmark Tests for Indoor Environmental Problems: Part Numerical thermal manikins, International Journal of Architectural Engineering Technology, Vol., No., pp5-75, 5 ) Ito K, Inthavong K, Kurabuchi T, Ueda T, Endo T, Omori T, Ono H, Kato S, Sakai K, Suwa Y, Matsumoto H, Yoshino H, Zhang W, Tu J. CFD Benchmark Tests for Indoor Environmental Problems: Part 4 Air-conditioning airflows, Residential kitchen airflows and Fire-induced flow, International Journal of Architectural Engineering Technology, Vol., No., pp76-, 5 757
FLOW FIELD MEASUREMENTS IN ACRYLIC RESPIRATORY TRACT MODEL BY PIV AND CFD Kota HIRASE *, Masato YAMASHITA *, Shin-ichiro ARAMAKI ** and Kazuhide ITO *** * Master Course Student, IGSES, Kyushu iv. ** Assist. Prof., IGSES, Kyushu iv., Dr.Eng. *** Assoc. Prof., IGSES, Kyushu iv., Dr.Eng. A number of epidemiological studies have shown consistent associations between increases acute exposure to particulate air pollution and increases in human morbidity and mortality. When focusing on health problems related respiratory system, it is important to grasp the pollutant mass transported in respiration. In view point of contaminant transportation in respiratory tract, the essential factor is flow characteristic in it and deeper understanding and fundamental information of flow pattern and turbulence characteristic in respiratory tract will contribute to the understanding of respiratory exposure mechanism. In this study, we conducted an in vitro experiment and numerical prediction to investigate the flow distribution in the realistic geometry of a human airway. The in vitro experiment and numerical prediction models were reproduced from CT data of a real human airway. Detailed measurements through the particle image velocimetry technique (PIV) as well as the numerical simulation through Computational Fluid Dynamics (CFD) are challenging in the field of respiratory infection in an indoor environment. In this study, application results of PIV technique to measure flow pattern in realistic respiratory tract model by using dimensional printer and refraction control technique are reported. The PIV measurement results of flow patterns in realistic replica model of human respiratory tract; i.e. in vitro experiments, contribute to the validation process of numerical simulation and improvement of numerical respiratory tract model. By using STL data of numerical airway model, a respiratory tract model with three dimensional configuration was created by three-dimensional printer. Lamination layer (Z direction) is 6m and formative resolution is 6dpi(x) 6dpi(y) 6dpi(z). Respiratory tract model was created by transparent acrylic material. As described later, mesh resolution of CFD simulation is the order of. -.mm ( - m) and this CFD resolution is in the same range of that in D printer. PIV is a technique that measures the instantaneous velocity field within an illuminated plane of the fluid field using light scattered from tracer particles into the fluid. The study applied the analysis of single exposed image pairs, which is single exposure/single frame PIV, by means of cross-correlation. In our application, a mixture of water and sodium polytungstate was used as a working fluid to match the refractive index (RI) of the flow and acrylic material constituting the airway model. To duplicate real breathing conditions, the working fluid condition in this experiment should have the same Reynolds number (Re) in both the numerical simulations and the experimental measurements. The measured/visualization region (trachea) of the airway model was immersed in the working fluid (the same mixture of water and sodium polytungstate) and suspended in the rectangular Plexiglass box. CFD simulations were performed to calculate airflow profiles under three breathing conditions. Steady flow fields were analyzed using a low Reynolds (Re) number-type k- model, RNG type k- model and the SST k- model. The CFD predictions were reasonably consistent with the results of PIV experimental data, and good agreement was observed in the 7.5 L/min flow rate with a low Re number type k- model. For 5 L/min and L/min flow rates, the low Re type k- model showed relatively good agreement at peak location. PIV experimental data has two velocity peak positions, but CFD simulation data has one peak position for both 5 L/min and L/min flow rates. This is caused by the fact that CFD overestimates turbulence. A difference of the normalized velocities at the red line of the peak between PIV and CFD is 8% for the 5 L/min flow rate and % for the L/min flow rate. (5 年 月 9 日原稿受理,5 年 5 月 5 日採用決定 ) 758