應用梯度照明與機器學習減噪演算法之二維微粒紋痕測速儀於微流道內之聲射流之研究

簡易檢索 / 詳目顯示

回結果列表

研究生：	Abdul Qadir Abdul Qadir
論文名稱：	應用梯度照明與機器學習減噪演算法之二維微粒紋痕測速儀於微流道內之聲射流之研究 Study of Acoustic Streaming in Microchannel using 2-D Particle Streak Velocimetry with Graded Illumination and Machine-Learning Based Noise Reduction Scheme
指導教授：	田維欣 Wei-Hsin Tien
口試委員:	田維欣 Wei-Hsin Tien 劉孟昆 Meng-Kun Liu 林怡均 Yi-Jiun Lin
學位類別：	碩士 Master
系所名稱：	工程學院 - 機械工程系 Department of Mechanical Engineering
論文出版年：	2023
畢業學年度：	112
語文別：	英文
論文頁數：	126
中文關鍵詞：	微粒紋痕影像測速儀 (PSV) 、聲射流、聲射阻抗、流場可視化
外文關鍵詞：	Particle Streak Velocimetry, Acoustic Streaming, Acoustic impedance, Flow Visualization
相關次數：	點閱：34 下載：0
分享至:	分享至facebook 分享至twitter

查詢本校圖書館目錄查詢臺灣博碩士論文知識加值系統勘誤回報

微粒紋痕影像測速儀 (Particle Streak Velocimetry, PSV) 為受廣泛使用的微粒影像測速儀 (Particle Image Velocimetry, PIV) 和微粒循跡測速儀 (Particle Tracking Velocimetry, PTV) 技術之延伸，常應用於流場分析。在 PSV 技術中，微粒的軌跡是透過攝影機以長曝光時間下記錄影像所獲得的。在實際流場可視化應用中，經常受到硬體的限制而需要以更長的曝光時間成像，導致拍攝移動微粒的影像拉伸為微粒紋痕影像。在Qureshi & Tien (2022) 研發的PSV技術可以提供實驗流體力學有價值的觀察結果，但其性能表現受到影像雜訊和流動方向模糊的困擾而限制了其準確性和適用性。首先發展了一種基於深度學習特別為 PSV 影像所設計的去雜訊方法，藉由結合卷積神經網路 (Convolutional Neural Networks, CNNs) 與特製的之影像處理技術，我們的方法能有效地處理各種類型的 PSV 影像雜訊。對比分析證實其性能優越，增加流體速度量測的準確性且提高計算效率。其次，實驗設置採用了一個梯度變化照明的發光二極體光源來解決實驗量測時單幀影像含多條紋痕的實驗流場方向。這種新的實驗技術將光源照明強度從全亮度的100% 線性變化至50% 。紋痕解析的演算法亦根據此改善方案進行調整，以便以最小平方擬合法決定流動的方向。為了評估驗證此法的概念可行性，將此技術應用於恆定平行流。平行流進一步分為恆定照明和梯度照明，也就是變化照明光強度，並進行測試比較。結果顯示所提出的方法改善了PSV的精確度與有效性，讓結果改善了96%。這些改善隨後被應用於研究微流道內以三角形微結構誘發的聲射流 (Acoustic Streaming) 現象，其中振盪透過電壓20V、頻率12kHz的壓電轉換器傳遞至流體。透過PSV 技術研究微結構周圍所產生的穩定聲射渦旋，分別採用三種不同頂角 (α=20°, 35°, 70°) 以及傾斜角 (β=30°, 45°, 60°) 的非對稱三角形微結構進行速度場分析。實驗結果顯示聲射渦旋的流動模式會隨著非對稱三角形微結構的傾斜角度變化，且與過往的研究比較顯示了合理的一致性。這些結果顯示，本研究所提出的方法可以在無需昂貴的高速成像硬體之下解析速度場，因此適用於各種應用。

Particle streak velocimetry (PSV) is an extension of widely used methods, particle image velocimetry (PIV), and particle tracking velocimetry (PTV), to resolve flow fields experimentally. The particle's trajectory is obtained by using a long exposure time of the camera in PSV. Practical applications often encounter hardware limitations, necessitating longer exposure times that cause captured images of moving particles to elongate into particle streak images. The currently developed PSV technique by Qureshi & Tien (2022) provides valuable insights for experimental fluid mechanics; however, its performance is plagued by image noise and direction ambiguity in the flow, limiting its accuracy and applicability. Two major modifications are proposed in this study to denoise the images and resolve the direction ambiguity problem. Firstly, a deep-learning-based denoising scheme specifically designed for PSV images is developed. By combining Conventional Neuro Network (CNNs) and custom image processing, our approach effectively handles various types of PSV image noise. Comparative analysis demonstrates its superior performance, enhancing the accuracy of fluid velocity measurements and computational efficiency. Secondly, in this study, a graded illumination LED light source is used for the experimental setup to determine the experimental flow field direction for a single image frame containing multiple streaks. The new experimental approach involves varying the illumination intensity of each exposure from 100% to 50% linearly of the full intensity of the light source. The streak-resolving algorithm is modified accordingly so the flow direction can be found through the least-square fitting process. To assess the validity of the concept, we applied this technique to a constant parallel flow. The Parallel flows were tested and compared with constant illumination and graded illumination, in which the intensity is varied. The results show that the proposed approach enhances the accuracy and reliability of PSV, by 95% in the tested flow cases. The modifications were then applied to the study of acoustic streaming phenomenon induced by triangular microstructures in the microchannels, in which the oscillation is transmitted to the fluid through the piezoelectric transducer at 20V and 12kHz. The asymmetrical triangular structure of three different tip angles (α=20°, 35°, 70°) and inclined angles (β=30°, 45°, 60°) is used for velocity field analysis. The results show that flow pattern of the acoustic streaming vortex changes with the inclined angle of the asymmetric triangular microstructure, and the results were compared to the previous study and showed reasonable agreements. These results suggest that the proposed approach can be used to resolve velocity fields without expensive hardware for high-speed imaging and is thus suitable for diverse applications.

Table of Content
摘要	ii
Abstract	iii
Acknowledgment	vii
Nomenclature	x
Greek Symbols	xii
Chapter 1 Introduction	1
1	MOTIVATION	1
1.1 Non-Intrusive Flow Measurement Techniques	1
1.2 Psv Direction Ambiguity	2
2	LITERATURE REVIEW	3
        1.2.1 Acoustic Streaming Technique	3
2.2 Psv Optical System	9
2.3 Psv Particle Information	14
2.4 Psv Various Flow Speed And Applications	17
2.5 Image Noise Reduction	25
3	SUMMERY	28
4	OBJECTIVES	29
8        ORGANIZATION OF THESIS	30
Chapter 2 Methods And Experimental Principles	31
1 EXPERIMENTAL SETUP	31
1.1 Optical Setup	31
1.2 Microcofluidic Devises	33
1.3 Acoustic Appliances	41
1.4 Parallel Flow Experimental Measurement System Setup	44
1.5 Acoustic Streaming Experimental Measurement Setup	46
2 DATA PROCESSING	48
2.1  Psv Image Preprocessing	48
2.1.2  Deep Learning Process For Psv Image	50
2.1.3  Deep Learning Model Training And Validation Process	54
2.1.4  Deep Learning Model Testing Phase	56
2.1.5  Real Scenarios Application	58
2.2    The Principle Of The Streak-Resolving Algorithm	60
2.2.1  Streak Resolved By Least Square Curve-Fitting	62
2.2.2  Velocity Measurement	63
2.2.3  Graded Illumination And The Determination Of The Direction Of Streak Velocity Vector	65
2.2.4  Algorithm Key Features	65
3 GEOMETRICAL DESIGN OF THE ACOUSTIC STREAMING MICROCHANNELS	69
Chapter 3 Results & Discussion	71
1 SECTION A: THE MACHINE-LEARNING BASED DENOISING SCHEME  	71
1.1 Parallel Flow Psv Image Denoising	71
1.2 Resolving Direction Ambiguity	76
3 SECTION B: APPLICATION TO ACOUSTIC STREAMING FLOW IN MICROCHANNEL 	80
3.1 Case 1 (?=20°, ?=30°) Acoustic Streaming Observation Results	81
3.2 Case 2 (?=35°, ?=30°) Acoustic Streaming Observation Results	88
3.3 Case 3 (?=35°, ?=45°) Acoustic Streaming Observation Results	97
3.4 Case 4 (?=35°, ?=60°) Acoustic Streaming Observation Results	105
3.5 Case 5 (?=70°, ?=30°) Acoustic Streaming Observation Results	111
4 DISCUSSION	117
Chapter 4 Conclusions And Future Work	122
References	124
                                

Adatrao, S., & Sciacchitano, A. (2019). Elimination of unsteady background reflections in PIV images by anisotropic diffusion. Measurement Science and Technology, 30(3), 035204.
Bergthorson, J. M., Goodwin, D. G., & Dimotakis, P. E. (2005). Particle streak velocimetry and CH laser-induced fluorescence diagnostics in strained, premixed, methane–air flames. Proceedings of the Combustion Institute, 30(1), 1637-1644.
Dimotakis, P. E., Debussy, F. D., & Koochesfahani, M. M. (1981). Particle streak velocity field measurements in a two‐dimensional mixing layer. The Physics of Fluids, 24(6), 995-999.
Doinikov, A. A., Gerlt, M. S., Pavlic, A., & Dual, J. (2020). Acoustic streaming is produced by sharp-edge structures in microfluidic devices. Microfluidics and Nanofluidics, 24(5), 32.
Dong, X., Wang, X., Zhou, W., Wang, F., Tang, X., & Cai, X. (2023). 3D particle streak velocimetry by defocused imaging. Particuology, 72, 1-9.
Estevadeordal, J., & Goss, L. (2005). PIV with LED: particle shadow velocimetry (PSV) technique. 43rd AIAA aerospace sciences meeting and exhibit,
Fan, L., Vena, P., Savard, B., & Fond, B. (2022). Experimental and numerical investigation on the accuracy of phosphor particle streak velocimetry. Experiments in Fluids, 63(10), 165.
Fan, L., Vena, P., Savard, B., Xuan, G., & Fond, B. (2021). High-resolution velocimetry technique based on the decaying streaks of phosphor particles. Optics Letters, 46(3), 641-644.
Fan, Y., Guo, C., Han, Y., Qiao, W., Xu, P., & Kuai, Y. (2023). Deep-learning-based image preprocessing for particle image velocimetry. Applied Ocean Research, 130, 103406.
Huang, P.-H., Xie, Y., Ahmed, D., Rufo, J., Nama, N., Chen, Y., Chan, C. Y., & Huang, T. J. (2013). An acoustofluidic micromixer based on oscillating sidewall sharp-edges. Lab on a Chip, 13(19), 3847-3852.
Jansen, M. (2012). Noise reduction by wavelet thresholding (Vol. 161). Springer Science & Business Media.
JunáHuang, T. (2014). A reliable and programmable acoustofluidic pump powered by oscillating sharp-edge structures. Lab on a Chip, 14(22), 4319-4323.
Kessler, M., & Leith, D. (1991). Flow measurement and efficiency modeling of cyclones for particle collection. Aerosol science and technology, 15(1), 8-18.
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. nature, 521(7553), 436-444.
Malavasi, S., Corretto, R., & Fossati, F. (2011). PSV and marker detection techniques for investigating the wake of an oscillating cylinder in a large wind tunnel. Flow Measurement and Instrumentation, 22(5), 428-437.
Mendez, M., Raiola, M., Masullo, A., Discetti, S., Ianiro, A., Theunissen, R., & Buchlin, J.-M. (2017). POD-based background removal for particle image velocimetry. Experimental Thermal and Fluid Science, 80, 181-192.
Müller, D., Müller, B., & Renz, U. (2001). Three-dimensional particle-streak tracking (PST) velocity measurements of a heat exchanger inlet flow A new method to measure all three air-flow velocity components in a plane is applied to a steady-state three-dimensional flow: A new method to measure all three air-flow velocity components in a plane is applied to a steady-state three-dimensional flow. Experiments in Fluids, 30(6), 645-656.
Nabavi, M., Siddiqui, M. K., & Dargahi, J. (2007). Simultaneous measurement of acoustic and streaming velocities using synchronized PIV technique. Measurement Science and Technology, 18(7), 1811.
Ovchinnikov, M., Zhou, J., & Yalamanchili, S. (2014). Acoustic streaming of a sharp edge. The Journal of the Acoustical Society of America, 136(1), 22-29.
Qureshi, M. H., & Tien, W.-H. (2022). Novel streak-resolving algorithm for particle streak velocimetry. Flow Measurement and Instrumentation, 87, 102208.
Qureshi, M. H., Tien, W.-H., & Lin, Y.-J. P. (2020). Performance comparison of particle tracking velocimetry (PTV) and particle image velocimetry (PIV) with long-exposure particle streaks. Measurement Science and Technology, 32(2), 024008.
Rosenstiel, M., & Grigat, R.-R. (2010). Segmentation and classification of streaks in a large-scale particle streak tracking system. Flow Measurement and Instrumentation, 21(1), 1-7.
Scholzen, F., & Moser, A. (1996). Three-dimensional particle streak velocimetry for room Airflow with automatic stereo-photogrammetric image processing. Proceedings of 5 th International Conference on Air Distribution in Rooms, ROOMVENT,
Sheremet, O., Sheremet, K., Sadovoi, O., & Sokhina, Y. (2018). Convolutional neural networks for image denoising in infocommunication systems. 2018 International Scientific-Practical Conference Problems of Infocommunications. Science and Technology (PIC S&T),
Sun, Y., Zhang, Y., Zhao, L., & Wang, X. (2004). An Algorithm of Stereoscopic Particle Image Velocimetry for Full-Scale Room Airflow Studies. ASHRAE transactions, 110(1).
Völker, C., Mämpel, S., & Kornadt, O. (2014). Measuring the human body's microclimate using a thermal manikin. Indoor air, 24(6), 567-579.
Wang, H., Wang, G., & Li, X. (2018). High-performance color sequence particle streak velocimetry for 3D airflow measurement. Applied Optics, 57(6), 1518-1523.
Wiklund, M., Green, R., & Ohlin, M. (2012). Acoustofluidics 14: Applications of acoustic streaming in microfluidic devices. Lab on a Chip, 12(14), 2438-2451.
Zhang, C., Guo, X., Royon, L., & Brunet, P. (2020). Acoustic streaming generated by sharp edges: The coupled influences of liquid viscosity and acoustic frequency. Micromachines, 11(6), 607.

簡易檢索 / 詳目顯示

相關論文