微通道气液两相流流型识别系统设计

气液两相流广泛存在于能源、化工、微电子、航空航天等诸多重要领域。其中流型是表征气液两相流整体流动形态和相互作用的重要参数，准确识别气液两相流流型对于理解流动机制、优化工业设备设计并确保其安全运行至关重要。本文基于深度学习理论，对气液两相流流动进行了流型识别研究。

本文创建了气液两相流流动视频数据集。包括20帧视频序列的平行微通道和叉排凸台阵列微通道的数据集、50帧和100帧视频序列的叉排凸台阵列微通道数据集。其中平行微通道气液两相流的流型分别为泡状流、环状流和塞状流；叉排凸台阵列微通道气液两相流的流型分别为泡状流、环状流和搅拌流。
针对20帧视频序列数据集，本文提出了基于ResNet-GRU网络 的微通道气液两相流流型识别模型。通过深度残差网络对图像帧提取空间图像特征，然后将图像特征送入门控循环单元中对时间序列进行学习训练。训练结果显示，平行微通道流型识别的整体准确率为95.3%，叉排凸台阵列微通道流型识别的整体准确率为96.1%。

针对20帧、50帧和100帧视频序列数据集，本文提出了基于Timesformer框架的微通道气液两相流流型识别模型。该模型选择时空自注意力机制来对视频序列进行建模和特征提取。训练结果显示，20帧、50帧和100帧数据集中叉排凸台阵列微通道的准确率分别为98.4%、100%和100%。
本文开发了一种用于识别气液两相流流型的系统。整体采用C++编写，使用Qt设计软件界面，借助OPenCV、Libtorch等库实现相关功能模块。该系统可以选择训练好的模型，对单个或者多个原始的两相流流动视频完成流型的准确识别。

[1] 方立军，胡月龙，武生. 气液两相流流型识别理论的研究进展. 锅炉制造，2012:33-36.
[2] 周会龙，孙斌，陈飞. 气液两相流型智能识别理论及方法. 科学出版社，2007.
[3] LIJ, PETERSON G. Boiling nucleation and two-phase flow patterns in forced liquid flow in microchannels. International Journal of Heat and Mass Transfer, 2005, 48(23-24): 4797-4810.
[4] CUBAUD T, HO CM. Transport of bubbles in square microchannels. Physics of Fluids, 2004, 16(12): 4575-4585.
[5] CHUNG P Y, KAWAJI M. The effect of channel diameter on adiabatic two-phase flow characteristics in microchannels. International Journal of Multiphase Flow, 2004, 30(7-8): 735-76.
[6] CHINNOV E, KABOV O. Two-phase flow regimes in a short flat channel. Technical Physics Letters, 2008, 34: 699-701.
[7] HAVERKAMP V, HESSEL V, LOWE H, et al. Hydrodynamics and mixer-induced bubble formation in micro bubble columns with single and multiple-channels. Chemical Engineering & Technology: Industrial Chemistry-Plant Equipment-Process Engineering-Biotechnology, 2006, 29(9): 1015-1026.
[8] YUE J, LUO L, GONTHIER Y, et al. An experimental investigation of gas-liquid two-phase flow in single microchannel contactors. Chemical Engineering Science, 2008, 63(16): 4189-4202.
[9] KUZNETSOV V V, SHAMIRZAEV A S, KOZULIN T, et al. Correlation of the flow pattern and flow boiling heat transfer in microchannels. Heat Transfer Engineering, 2013, 34(2-3): 235-245.
[10] 季常伟，曹复昕，孙立成，等. 竖直窄矩形通道气液两相流流型识别研究. 原子能科学技术，2012, 46:1055-1060.
[11] 全卫国，庞雪纯，朱赓宏. 基于卷积神经网络的气液两相流流型识别方法. 系统仿真学报，2021, 33: 883-891.
[12] 周云龙，刘博，刘袖，等. T型微通道内两相流流型及相分离特性. 化学反应工程与工艺，2012, 28:300-305.
[13] 房贤在，李秋英，陈杰，等. 管内气液两相流流型研究现状与发展. 东北电力大学学报，2022, 42:1-7.
[14] 刘起超，周云龙，陈聪. 基于CEEMDAN和概率神经网络的起伏振动气液两相流型识别. 仪器仪表学报，2021, 42:84-93.
[15] MATSUI G. Automatic identification of flow regimes in vertical two-phase flow using differential pressure fluctuations. Nuclear Engineering and Design, 1986, 95: 221-231.
[16] ELPERIN T, KLOCHKO M. Flow regime identification in a two-phase flow using wavelet transform. Experiments in Fluids: Experimental Methods and Their Applications to Fluid Flow, 2002, 32: 674-682.
[17] 王晓萍. 气固流化床压力脉动信号的Hilbert-Huang变换与流型识别. 高校化学工程学报，2005: 474-479.
[18] 周云龙，主强，杨志行，等. 基于子波能量特征的气液两相流流型辨识方法. 化工学报，2007:1948-1954.
[19] DING H, HUANG Z, SONG Z, et al. Hilbert-Huang transform based signal analysis for the characterization of gas-liquid two-phase flow. Flow Measurement and Instrumentation, 2007, 18(1): 37-46.
[20] BIN S, ERPENG W, YANG D, et al. Time-frequency signal processing for gas-liquid two-phase flow through a horizontal venturi based on adaptive optimal-kernel theory. Chinese Journal of Chemical Engineering, 2011, 19(2): 243-252.
[21] SANTOSO B, INDARTO D, THOMAS S. The identification of gas-liquid co-current two-phase flow pattern in a horizontal pipe using the power spectral density and the artificial neural network (ANN). Modern Applied Science, 2012, 6(9): 56-67.
[22] LIU H, ZHU Y, PEI S, et al. Flow regime identification for air valves failure evaluation in water pipelines using pressure data. Water Research, 2019, 165:115002.
[23] KHAN U, PAO W, SALLIH N, et al. Flow Regime Identification in Gas-Liquid Two-Phase Flow in Horizontal Pipe by Deep Learning. Journal of Advanced Research in Applied Sciences and Engineering Technology, 2022, 27(1): 86-91.
[24] AKJ, GD, PK D. Flow regime identification of two-phase liquid-liquid upflow through vertical pipe. Chemical Engineering Science, 2006, 61: 1500-1515.
[25] WANG L, HUANG Z, WANG B, et al. Flow pattern identification of gas-liquid two-phase flow based on capacitively coupled contactless conductivity detection. IEEE Transactions on Instrumentation and Measurement, 2012, 61(5): 1466-1475.
[26] LI Z, FAN C L. A novel method to identify the flow pattern of oil-water two-phase flow. Journal of Petroleum Exploration and Production Technology, 2020, 10: 3723-3732.
[27] 乔守旭，钟文义，谭思超，等. 基于PCA-GA-SVM的竖直下降两相流流型预测. 核动力工程，2022, 43; 85-93.
[28] TANC, DONG F, WUM. Identification of gas/liquid two-phase flow regime through ERT-based measurement and feature extraction. Flow Measurement and Instrumentation, 2007, 18(5-6): 255-261.
[29] DE KERPELK, DE SCHAMPHELEIRE S, DE KEULENAER T, et al. Two-phase flow regime assignment based on wavelet features of a capacitance signal. International Journal of Heat and Fluid Flow, 2015, 56: 317-323.
[30] MORSHIED M, KHIAN M S, RAHMAN M A, et al. Flow regime, slug frequency and wavelet analysis of air/Newtonian and air/non-Newtonian two-phase flow. Applied Sciences, 2020, 10(9): 3272.
[31] LI X, LI L, MA W, et al. Two-phase flow patterns identification in porous media using feature extraction and SVM. International Journal of Multiphase Flow, 2022, 156: 104222.
[32] LI X, LI L, ZHAO H, et al. Identification of two-phase flow pattern in porous media based on signal feature extraction. Flow Measurement and Instrumentation, 2022, 83: 102123.
[33] ZHANG Y, AZMAN A N, XU K W, et al. Two-phase flow regime identification based on the liquid-phase velocity information and machine learning. Experiments in Fluids, 2020, 61: 1-16.
[34] LIN Z, LIU X, LAO L, et al. Prediction of two-phase flow patterns in upward inclined pipes via deep learning. Energy, 2020, 210:118541.
[35] LI X, LI L, WANG W, et al. Machine learning techniques applied to identify the two-phase flow pattern in porous media based on signal analysis. Applied Sciences, 2022, 12(17): 8575.
[36] OUYANG L, JIN N, REN W. A new deep neural network framework with multivariate time series for two-phase flow pattern identification. Expert Systems with Applications, 2022, 205:117704.
[37] KHAN U, PAO W, SALLIH N, et al. Identification of horizontal gas-liquid two-phase flow regime using deep learning. CFD Letters, 2022, 14(10): 68-78.
[38] 施丽莲. 基于数字图像识别技术的气液两相流参数检测的研究. 浙江大学，2004.
[39] SUNDE C, AVDIC S, PAZSIT I. Classification of two-phase flow regimes via image analysis and a neuro-wavelet approach. Progress in Nuclear Energy, 2005, 46(3-4): 348-358.
[40] 周云龙，李季洪伟，袁俊文. 一种识别气液两相流流型的新方法. 热能动力工程，2009, 24: 68-72+143.
[41] 周云龙，季洪伟，陈飞. 基于图像纹理特征和粒子群优化神经网络的气液两相流流型识别. 吉林大学学报（工学版）, 2009, 39: 673-678.
[42] 常佃康，薛婷，张旖婷. 基于高速摄影传感器的气液两相流型分层模糊识别. 传感器与微系统，2016, 35: 58-60.
[43] RAFALKO G, MOSDORF R, GORSKIG. Two-phase flow pattern identification in minichannels using image correlation analysis. International Communications in Heat and Mass Transfer, 2020, 113:104508.
[44] 梁法春，陈婧，冉云麒，等. 基于迁移学习的水平管气液两相流型智能识别. 实验室研究与探索，2021, 40: 1-5.
[45] URBINA-SALAS I, VAZQUEZ-RAMiREZ E, GARCiA-SANCHEZ E, et al. Application of convolutional neural networks for the classification of two-phase flow patterns. 2021 IEEE International Autumn Meeting on Power, Electronics and Computing (ROPEC): Vol.5. IEEE, 2021: 1-6.
[46] KADISH S, SCHMID D, SON J, et al. Computer Vision-Based Classification of Flow Regime and Vapor Quality in Vertical Two-Phase Flow. Sensors, 2022, 22(3): 996.
[47] BRANTSON E T, ABDULKADIR M, AKWENST PH, et al. Gas-liquid vertical pipe flow patterns convolutional neural network classification using experimental advanced wire mesh sensor images. Journal of Natural Gas Science and Engineering, 2022, 99: 104406.
[48] NIE F, WANG H, SONG Q, et al. Image identification for two-phase flow patterns based on CNN algorithms. International Journal of Multiphase Flow, 2022, 152: 104067.
[49] XU H, TANG T, ZHANG B, et al. Identification of two-phase flow regime in the energy industry based on modified convolutional neural network. Progress in Nuclear Energy, 2022, 147: 104191.
[50] 张立峰，王智，吴思橙. 基于卷积神经网络与门控循环单元的气液两相流流型识别方法. 计量学报，2022, 43: 1306-1312.
[51] 李爽，李玉星，王冬旭. 基于小波变换与神经网络的上领管流型识别方法. 油气储运，2020, 39: 912-918.
[52] LIANG F, ZHENG H, YU H, et al. Gas-liquid two-phase flow pattern identification by ultrasonic echoes reflected from the inner wall of a pipe. Measurement Science and Technology, 2016, 27(3): 035304.
[53] SUN B, TAN C, SHI X, et al. Gas-liquid two-phase flow pattern identification by differential pressure and ultrasonic echoes. 2021 IEEE International Instrumentation and Measurement Technology Conference (I2MTC). 2021: 1-6.
[54] AMEEL B, DE KERPEL K, CANIERE H, et al. Classification of two phase flows using linear discriminant analysis and expectation maximization clustering of video footage. International Journal of Multiphase Flow, 2012, 40: 106-112.
[55] 解爽晨，庄晓如，岳思君，等. HFE-7100平行微通道流动沸腾实验. 物理学报，2023: 1-24.
[56] 常亮，邓小明，周明全，等. 图像理解中的卷积神经网络. 自动化学报，2016, 42: 1300-1312.
[57] 季彦冬，娜宗波，雷航. 卷积神经网络研究综述. 计算机应用，2016, 36: 2508-2515+2565.
[58] 将昂波，王维维. ReLU激活函数优化研究. 传感器与微系统，2018, 37: 50-52.
[59] 张荣，季伟平，莫同. 深度学习研究综述. 信息与控制，2018, 47: 385-397+410.
[60] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016: 770-778.
[61] 杨丽，臭雨茜，主俊丽，等. 循环神经网络研究综述. 计算机应用，2018, 38: 1-6+26.
[62] 姚程文，杨苹，刘泽健. 基于CNN-GRU混合神经网络的负荷预测方法. 电网技术，2020, 44: 3416-3424.
[63] CHO K, VAN MERRIENBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation. 2014.
[64] HINTON G E, SRIVASTAVA N, KRIZHEVSKY A, et al. Improving neural networks by preventing co-adaptation of feature detectors. 2012.
[65] 蒋兆军，成孝刚，彭雅琴，等. 基于深度学习的无人机识别算法研究. 电子技术应用，2017, 43: 84-87.
[66] PASZKE A, GROSS S, MASSA F, et al. Pytorch: An imperative style, high-performance deep learning library. Advances in Neural Information Processing Systems, 2019, 32.
[67] BERTASIUS G, WANG H, TORRESANI L. Is space-time attention all you need for video understanding? ICML: Vol. 2. 2021: 4.
[68] SU J, AHMED M, LU Y, et al. Roformer: Enhanced transformer with rotary position embedding. Neurocomputing, 2024, 568: 127063.
[69] ZHANG H, HAO Y, NGO C W. Token shift transformer for video classification. Proceedings of the 29th ACM International Conference on Multimedia. 2021: 917-925.
[70] ARNAB A, DEHGHANI M, HEIGOLD G, et al. Vivit: A video vision transformer. Proceedings of the IEEE/CVF International Conference on Computer Vision. 2021: 6836-6846.
[71] BLANCHETTE J, SUMMERFIELD M. C++ GUI programming with Qt 4. Prentice Hall Professional, 2006.
[72] 王师，蒋林，刘镇变，等. Qt建模仿真用户界面设计. 实验室研究与探索，2016, 35: 70-74.
[73] CULJAK I, ABRAM D, PRIBANIC T, et al. A brief introduction to OpenCV. Proceedings of the 35th International Convention MIPRO. IEEE, 2012: 1725-1730.
[74] SEVER M, OGUT S. A Performance Study Depending on Execution Times of Various Frameworks in Machine Learning Inference. 2021 15th Turkish National Software Engineering Symposium (UYMS). IEEE, 2021: 1-5.