信息处理技术在气固两相流检测中的应用
|
摘要
气固两相流在多相流体系中普遍存在，例如，自然界的大漠扬沙，化工、冶金、能源和粮食等领域的燃料配送，散装物料(如沙子、谷物、塑料粒)的气力输送和粉尘处理等，硫化床、旋风除尘、沉降室及过滤过程中也会涉及到气固两相流。由于气固两相流动规律比单相流动的流动特性更复杂，且各相间的界面效应等原因，致使对两相流动过程参数(如流态、浓度、流量和粒径等)的检测难度很大。然而，要认清气固两相流体系的复杂现象，以及对工业过程进行预测、设计和控制，就必须解决气固两相流的检测问题。目前，许多发达国家如英国、美国、德国、日本、挪威、澳大利亚等均对此进行了大量的研究工作。我国在八五、九五科研规划以及863高科技计划中也都给予了高度重视，相应的两相流测试技术也有了很大的进步，但其发展水平还未很好满足工业应用的要求。因此，气固两相流检测技术尚属一个急待发展的研究领域。
本文在课题研究小组近年来取得的研究成果基础上，进一步将信息处理技术引用到气固两相流参数检测中。文中主要应用神经网络和数据融合这两种方法分别进行气固两相流流型的辨识及固相颗粒流速的检测，并进行了仿真研究。主要工作内容包括：
1．总结和整理了课题小组前期研究工作所取得的成果。
2．阅读国内外有关参考文献，在分析研究神经网络的功能及特点的基础上，选取学习向量量化神经网络作为气固两相流流型辨识的手段，并进行了仿真研究。
3．应用数据融合技术及互相关测速技术进行固相颗粒截面平均速度的检测，包括单一传感器数据的互相关分析，以及对所有传感器的数据进行融合处理，再对其进行相关分析。
Gas-solid two-phase flow is existed widely in multiphase flow system, for example, desert sand in nature, fuel transportation in field of chemical industry, metallurgy, energy sources and foodstuff, and so on, as well as pneumatic conveying and dust treatment of materiel in bulk (i.e. sands, corn and plastic granula etc.). In addition, sulfureted bed, whirlwind dust-exhaustion, sedimentation room and filtration process are involved with gas-solid two-phase flow. It is very difficult to measure some process parameters during two-phase flowing, such as flow status, concentration, flux, particle size and so forth, because the flowing rules of gas-solid two-phase flow is more complex than that of single-phase flow, together with interphase effect between every phase. However, to know well complex phenomena in gas-solid two-phase flow system, to predict, design and control industrial process, it is necessary to solve the difficulties in measurement of gas-solid two-phase flow. At present, many developed countries
such as England, America, Germany, Japan, Norway and Australia are all devoted to correlative research. Our country, China, is also pay much attention to it in the eighth five-year scientific research programming, the ninth five-year programming and '863' high-tech plan. And the measure technology of two-phase flow has achieved great development, but its level still doesn't satisfy the requirements of industrial application. So the measure technology of gas-solid two-phase flow yet belongs to the research field urgently waiting for developing
Based on the research achievements obtained by our task team recently, this dissertation further employed information process technology into parameter measurement of gas-solid two-phase flow, where the neural network technology was used to recognize the flow regime and the data fusion technology to measure the flow velocity of solid-phase particles, and the simulation research was worked. The main work is described as follow:
1. Summarizing and arranging the achievements obtained by the task team before.
2. Collecting and reading the correlative reference documents all over the world; based on analyzing the performances and characteristics of neural network, employing the learning vector quantification neural network to recognize the flow regime of gas-solid two-phase flow, and carrying out simulation research.
3. Combining the data fusion technology with the cross-correlation technology for measuring velocity to measure the mean velocity of solid-phase particles in section, including the cross-correlation analysis of data from every sensor and that of data from all sensors after processed by fusion technology.
本文在课题研究小组近年来取得的研究成果基础上，进一步将信息处理技术引用到气固两相流参数检测中。文中主要应用神经网络和数据融合这两种方法分别进行气固两相流流型的辨识及固相颗粒流速的检测，并进行了仿真研究。主要工作内容包括：
1．总结和整理了课题小组前期研究工作所取得的成果。
2．阅读国内外有关参考文献，在分析研究神经网络的功能及特点的基础上，选取学习向量量化神经网络作为气固两相流流型辨识的手段，并进行了仿真研究。
3．应用数据融合技术及互相关测速技术进行固相颗粒截面平均速度的检测，包括单一传感器数据的互相关分析，以及对所有传感器的数据进行融合处理，再对其进行相关分析。
Gas-solid two-phase flow is existed widely in multiphase flow system, for example, desert sand in nature, fuel transportation in field of chemical industry, metallurgy, energy sources and foodstuff, and so on, as well as pneumatic conveying and dust treatment of materiel in bulk (i.e. sands, corn and plastic granula etc.). In addition, sulfureted bed, whirlwind dust-exhaustion, sedimentation room and filtration process are involved with gas-solid two-phase flow. It is very difficult to measure some process parameters during two-phase flowing, such as flow status, concentration, flux, particle size and so forth, because the flowing rules of gas-solid two-phase flow is more complex than that of single-phase flow, together with interphase effect between every phase. However, to know well complex phenomena in gas-solid two-phase flow system, to predict, design and control industrial process, it is necessary to solve the difficulties in measurement of gas-solid two-phase flow. At present, many developed countries
such as England, America, Germany, Japan, Norway and Australia are all devoted to correlative research. Our country, China, is also pay much attention to it in the eighth five-year scientific research programming, the ninth five-year programming and '863' high-tech plan. And the measure technology of two-phase flow has achieved great development, but its level still doesn't satisfy the requirements of industrial application. So the measure technology of gas-solid two-phase flow yet belongs to the research field urgently waiting for developing
Based on the research achievements obtained by our task team recently, this dissertation further employed information process technology into parameter measurement of gas-solid two-phase flow, where the neural network technology was used to recognize the flow regime and the data fusion technology to measure the flow velocity of solid-phase particles, and the simulation research was worked. The main work is described as follow:
1. Summarizing and arranging the achievements obtained by the task team before.
2. Collecting and reading the correlative reference documents all over the world; based on analyzing the performances and characteristics of neural network, employing the learning vector quantification neural network to recognize the flow regime of gas-solid two-phase flow, and carrying out simulation research.
3. Combining the data fusion technology with the cross-correlation technology for measuring velocity to measure the mean velocity of solid-phase particles in section, including the cross-correlation analysis of data from every sensor and that of data from all sensors after processed by fusion technology.
引文
[1] 连桂森．多相流动基础．杭州：浙江大学出版社，1989，6：1-25
[2] 李海青．多相流测试技术现状及趋势．多相流检测技术进展．北京：石油工业出版社，1996，8：33-42
[3] 李海青．两相流参数检测及应用．杭州：浙江大学出版社，1991，11：1-51
[4] 李海青．智能型检测仪表及控制装置．北京：化学工业出版社，1998，9：43-73
[5] 郑之初．海洋高技术与多相流检测技术．多相流检测技术进展．北京：石油工业出版社，1996，8：58-67
[6] Ozanyan K. B, Carey S.J, Hindle F. All-optoelectronic solutions for process tomography. Lasers and Electro-Optics Europe, Conference Digest, 2000: 125-134
[7] 阎春生，曾楠，赖淑蓉等．光学层析成像技术的研究动态．激光杂志，2001，22(5)：5-7
[8] Turchin I.v, Dolin L.S, Kamensky V.A. Reconstruction of in vivo biological characteristics from OCT images by theoretical modeling of light scattering in biotissue. Lasers and Electro-Optics, 2002(1):477-478
[9] Hoerauf H, Wirbelauer C. Slit-lamp-adapted optical coherence tomography of the anterior segment. Clinical investigation, 2000, 238(8):8-18
[10] Bernini R, Pierri R, Zeni L. An iterative method for optical reconstruction of graded index profiles in planar dielectric waveguides. Light wave Technology,2000, 18(5):729-736
[11] L.Zeni R, Bernini, R.Pierri. Optical Tomography for Dielectric Profiling in Processing Electronic Materials. 1st World Congress on Industrial Process Tomography, Buxton, Greater Manchester, April
14-17, 1999
[12] Green R.G. Sensor system for lightly loaded pneumatic conveyors. Power Technology, 1998,95:79-92
[13] N.Ramli, R.G.Green, R.Abdul Rahim et al. Fibre Optic Lens Modelling for Optical Tomography 1st World Congress on Industrial Process Tomography, BuxtonGreater Manchester, April, 1999
[14] S.Ibrahim, R.G.Green, K.Dutton Ruzairi.Abdul Rahim. Application of Optical Tomography in Industrial Process Control System 2000 IEEE, Ⅰ-493-498
[15] L.McMackin, R.J.Hugo, K.P.Bishop et al. High Speed Optical Tomography System for Quantitative Measurement and Visualization of Dynamic Features in a Round Jet Experiments in Fluids 26(1999)
[16] Carey SH, McCann H. et al. Chemical Species tomography by near infra-red absorption. Chemical Engineering Journal,2000,77:111-118,
[17] http://www.tomography.umist.ac.uk/optictom.shtml, Researches on Optical Tomography in UMIST,
[18] Francis Hindle, Hugh McCann, Krikor B.Ozanyan. First Demonstration of Optical Fluorescence Auto-projection Tomography. Chemical Engineering Journal 77(2000) 127-135
[19] Chang-Fu, Wu. Experiment evaluation of a radial beam geometry for mapping air pollutions using optical remote sensing and computed tomography. Atmospheric Environment, 1999,33:4709-4716
[20] S.Ibrahim, R.G.Green, K.Dutton,etc. Optical Fibres for Process Tomography: A Design Study. 1st World Congress on Industrial Process Tomography, 1999:511-516
[21] 杨莉松，王桂英，徐至展．光学层析成像机理分析．激光与红外，1997，27(4)：239-241
[22] 岳建华，孔庆林，王泽章．激光粉尘浓度测量技术的研究与应用．内蒙古电力技术，1997(1)：6-8
[23] 庄天戈．CT原理与算法．上海：上海交通大学出版社，1992：2-6
[24] H．H巴雷特，W．斯温德尔著；庄天戈，周颂凯译．放射成象——图象形成、检测和处理的理论．北京：科学出版社，1988，6：379-486
[25] 邹旭华．激光传感过程层析成像技术研究：[硕士论文]．广州：广东工业大学图书馆，2003，5：38-46
[26] 李扬，汪仁煌，郑莹娜等．应用透射式光层析术反演颗粒流的二维浓度分布．中国电机工程学报，2004，24(3)：193-199
[27] 邹旭华，郑莹娜，李扬．激光传感过程层析成像技术探讨．现代制造工程，2002，12：38-42
[28] 李扬，汪仁煌，郑莹娜．应用线性神经网络映射光学过程层析成像的逆问题．中国图象图形学报，2003，8(7)：738-743
[29] 郑莹娜，李扬，陈健等．基于激光阵列传感的层析成像系统研究．仪器仪表学报，2004，25(1)：61-64
[30] Yang Li, Renhuang Wang, Yingna Zheng. Analysis of sensing field to array sensors used for optical tomography. In: Zhang Guangjun, Zhao Huijie, Wang Zhongyu eds. Fifth International Symposium on Instrumentation and Control Technology. Beijing, China. October 2003. Proc. of SPIE, vol.5253, 2003:414-419
[31] Yang Li, Renhuang Wang, Yingna Zheng. Image reconstruction of 2D concentration distribution of particles flow with extinction optical tomography. In: Zhang Guangjun, Zhao Huijie, Wang Zhongyu eds. Fifth International Symposium on Instrumentation and Control Technology. Beijing, China. October 2003. Proc. of SPIE, vol. 5253: 407-413
[32] 李扬，汪仁煌，郑莹娜等．光CT的扇束投影重排方法及投影几何分析．电子学报，2004，(4)
[33] 邹旭华，郑莹娜，李扬等．基于正交激光器阵列的图像重建技术研究．仪表技术与传感器，2003，7：39-42
[34] 邹旭华．激光传感过程层析成像技术研究：[硕士论文]．广州：广东工业大学图书馆，2003，5：47-58
[35] 陈之航，曹柏林，赵在三．气液双相流动和传热．北京：机械工业出版
社，1983，9：1-80
[36] 赵昀．神经网络在多相流参数检测中的应用：[博士论文]．杭州：浙江大学图书馆，2001，5：86-101
[37] 孙涛．基于数据融合技术的两相流流型辨识与流量测量方法研究：[博士论文]．杭州：浙江大学图书馆，2002，1：1-46
[38] Embrechts M J, et al. The application of neural networks to two-phase flow regime identification. In: Proceeding of the American Power Conference. 1996:860-864
[39] Monji H, Matsui G. Flow pattern identification of gas-liquid two-phase flow using a neural network. In: The Third International Conference on Multiphase Flow, ICMF'98, Lyon, France, June 8～12, 1998
[40] Mi Y, Ishii M, Tsoukalas L H. Vertical two-phase flow identification using advanced instrumentation and neural networks. Nuclear Engineering and Design, 1998, 184:409-420
[41] Annunziato A, Pizzuti S. Fuzzy fusion between fuidodynamic and neural models for monitoring multiphase flows. International Journal of Approximate Reasoning, 1999, 22:53-71
[42] 彭黎辉，张宝芬，姚丹亚．基于模糊神经元网络的两相流流型辨识方法．模式识别与人工智能，1997，10(4)：332-337
[43] 夏靖波，杨晓铁，王师．模糊神经网络两相流流型辨识算法的研究．控制与决策，1999，14(s)：531-535
[44] 夏靖波，夏军利，白鸥．BP神经网络在电容成像流型识别中的应用．沈阳工业大学学报，1999，21(3)：262-264
[45] 吴浩江，周芳德．应用RBF神经网络智能识别油气水多相流流型．石油学报，2000，21(3)：57-60
[46] 白博峰，郭烈锦，陈学俊．基于反传神经网络和压差波动识别气液两相流流型．化工学报，2000，51(6)：848-851
[47] 白博峰，郭烈锦，陈学俊．气液两相流流型在线智能识别．中国电机工程学报，2001，21(7)：46-50
[48] 白博峰，郭烈锦，赵亮．汽(气)液两相流流型在线识别的研究进展．力学进展，2001，31(3)：437-446
[49] 赵振宇，徐用懋．模糊理论和神经网络的基础与应用．北京：清华大学出版社，1996，6：80-115
[50] Martin T. Hagan, Howard B.Demuth, Mark H.Beale. Neural Network Design. Beijing: China Machine Press, 2002,9:8-425
[51] 飞思科技产品研发中心．MATLAB6.5辅助神经网络分析与设计．北京：电子工业出版社，2003，1：165-192
[52] 许东，吴铮．基于MATLAB6.x的系统分析与设计——神经网络．西安：西安电子科技大学出版社，2002，9：166-237
[53] D．Hanselman，B．Littlefield著；张航，黄攀译．精通MATLAB6．北京：清华大学出版社，2002，6：23-309
[54] 李海涛，邓樱．MATLAB6.1基础及应用技巧．北京：国防工业出版社，2002，3：235-261
[55] 徐苓安．相关流量测量技术．天津：天津大学出版社，1988：1-150
[56] 邓湘．基于层析成像技术的智能化两相流测量系统研究：[博士论文]．天津：天津大学图书馆，2001，7：1-95
[57] 卢文祥，杜润生．机械工程测试·信息·信号分析．武汉：华中理工大学出版社，1999，8：6-264
[58] 胡丹丹，肖书明，王燕清等．基于多传感器的数据融合技术．东北电力学院学报，2004，24(1)：62-67
[59] 秦志强．数据融合技术及其应用．网络信息技术，2003，22(5)：25-28
[2] 李海青．多相流测试技术现状及趋势．多相流检测技术进展．北京：石油工业出版社，1996，8：33-42
[3] 李海青．两相流参数检测及应用．杭州：浙江大学出版社，1991，11：1-51
[4] 李海青．智能型检测仪表及控制装置．北京：化学工业出版社，1998，9：43-73
[5] 郑之初．海洋高技术与多相流检测技术．多相流检测技术进展．北京：石油工业出版社，1996，8：58-67
[6] Ozanyan K. B, Carey S.J, Hindle F. All-optoelectronic solutions for process tomography. Lasers and Electro-Optics Europe, Conference Digest, 2000: 125-134
[7] 阎春生，曾楠，赖淑蓉等．光学层析成像技术的研究动态．激光杂志，2001，22(5)：5-7
[8] Turchin I.v, Dolin L.S, Kamensky V.A. Reconstruction of in vivo biological characteristics from OCT images by theoretical modeling of light scattering in biotissue. Lasers and Electro-Optics, 2002(1):477-478
[9] Hoerauf H, Wirbelauer C. Slit-lamp-adapted optical coherence tomography of the anterior segment. Clinical investigation, 2000, 238(8):8-18
[10] Bernini R, Pierri R, Zeni L. An iterative method for optical reconstruction of graded index profiles in planar dielectric waveguides. Light wave Technology,2000, 18(5):729-736
[11] L.Zeni R, Bernini, R.Pierri. Optical Tomography for Dielectric Profiling in Processing Electronic Materials. 1st World Congress on Industrial Process Tomography, Buxton, Greater Manchester, April
14-17, 1999
[12] Green R.G. Sensor system for lightly loaded pneumatic conveyors. Power Technology, 1998,95:79-92
[13] N.Ramli, R.G.Green, R.Abdul Rahim et al. Fibre Optic Lens Modelling for Optical Tomography 1st World Congress on Industrial Process Tomography, BuxtonGreater Manchester, April, 1999
[14] S.Ibrahim, R.G.Green, K.Dutton Ruzairi.Abdul Rahim. Application of Optical Tomography in Industrial Process Control System 2000 IEEE, Ⅰ-493-498
[15] L.McMackin, R.J.Hugo, K.P.Bishop et al. High Speed Optical Tomography System for Quantitative Measurement and Visualization of Dynamic Features in a Round Jet Experiments in Fluids 26(1999)
[16] Carey SH, McCann H. et al. Chemical Species tomography by near infra-red absorption. Chemical Engineering Journal,2000,77:111-118,
[17] http://www.tomography.umist.ac.uk/optictom.shtml, Researches on Optical Tomography in UMIST,
[18] Francis Hindle, Hugh McCann, Krikor B.Ozanyan. First Demonstration of Optical Fluorescence Auto-projection Tomography. Chemical Engineering Journal 77(2000) 127-135
[19] Chang-Fu, Wu. Experiment evaluation of a radial beam geometry for mapping air pollutions using optical remote sensing and computed tomography. Atmospheric Environment, 1999,33:4709-4716
[20] S.Ibrahim, R.G.Green, K.Dutton,etc. Optical Fibres for Process Tomography: A Design Study. 1st World Congress on Industrial Process Tomography, 1999:511-516
[21] 杨莉松，王桂英，徐至展．光学层析成像机理分析．激光与红外，1997，27(4)：239-241
[22] 岳建华，孔庆林，王泽章．激光粉尘浓度测量技术的研究与应用．内蒙古电力技术，1997(1)：6-8
[23] 庄天戈．CT原理与算法．上海：上海交通大学出版社，1992：2-6
[24] H．H巴雷特，W．斯温德尔著；庄天戈，周颂凯译．放射成象——图象形成、检测和处理的理论．北京：科学出版社，1988，6：379-486
[25] 邹旭华．激光传感过程层析成像技术研究：[硕士论文]．广州：广东工业大学图书馆，2003，5：38-46
[26] 李扬，汪仁煌，郑莹娜等．应用透射式光层析术反演颗粒流的二维浓度分布．中国电机工程学报，2004，24(3)：193-199
[27] 邹旭华，郑莹娜，李扬．激光传感过程层析成像技术探讨．现代制造工程，2002，12：38-42
[28] 李扬，汪仁煌，郑莹娜．应用线性神经网络映射光学过程层析成像的逆问题．中国图象图形学报，2003，8(7)：738-743
[29] 郑莹娜，李扬，陈健等．基于激光阵列传感的层析成像系统研究．仪器仪表学报，2004，25(1)：61-64
[30] Yang Li, Renhuang Wang, Yingna Zheng. Analysis of sensing field to array sensors used for optical tomography. In: Zhang Guangjun, Zhao Huijie, Wang Zhongyu eds. Fifth International Symposium on Instrumentation and Control Technology. Beijing, China. October 2003. Proc. of SPIE, vol.5253, 2003:414-419
[31] Yang Li, Renhuang Wang, Yingna Zheng. Image reconstruction of 2D concentration distribution of particles flow with extinction optical tomography. In: Zhang Guangjun, Zhao Huijie, Wang Zhongyu eds. Fifth International Symposium on Instrumentation and Control Technology. Beijing, China. October 2003. Proc. of SPIE, vol. 5253: 407-413
[32] 李扬，汪仁煌，郑莹娜等．光CT的扇束投影重排方法及投影几何分析．电子学报，2004，(4)
[33] 邹旭华，郑莹娜，李扬等．基于正交激光器阵列的图像重建技术研究．仪表技术与传感器，2003，7：39-42
[34] 邹旭华．激光传感过程层析成像技术研究：[硕士论文]．广州：广东工业大学图书馆，2003，5：47-58
[35] 陈之航，曹柏林，赵在三．气液双相流动和传热．北京：机械工业出版
社，1983，9：1-80
[36] 赵昀．神经网络在多相流参数检测中的应用：[博士论文]．杭州：浙江大学图书馆，2001，5：86-101
[37] 孙涛．基于数据融合技术的两相流流型辨识与流量测量方法研究：[博士论文]．杭州：浙江大学图书馆，2002，1：1-46
[38] Embrechts M J, et al. The application of neural networks to two-phase flow regime identification. In: Proceeding of the American Power Conference. 1996:860-864
[39] Monji H, Matsui G. Flow pattern identification of gas-liquid two-phase flow using a neural network. In: The Third International Conference on Multiphase Flow, ICMF'98, Lyon, France, June 8～12, 1998
[40] Mi Y, Ishii M, Tsoukalas L H. Vertical two-phase flow identification using advanced instrumentation and neural networks. Nuclear Engineering and Design, 1998, 184:409-420
[41] Annunziato A, Pizzuti S. Fuzzy fusion between fuidodynamic and neural models for monitoring multiphase flows. International Journal of Approximate Reasoning, 1999, 22:53-71
[42] 彭黎辉，张宝芬，姚丹亚．基于模糊神经元网络的两相流流型辨识方法．模式识别与人工智能，1997，10(4)：332-337
[43] 夏靖波，杨晓铁，王师．模糊神经网络两相流流型辨识算法的研究．控制与决策，1999，14(s)：531-535
[44] 夏靖波，夏军利，白鸥．BP神经网络在电容成像流型识别中的应用．沈阳工业大学学报，1999，21(3)：262-264
[45] 吴浩江，周芳德．应用RBF神经网络智能识别油气水多相流流型．石油学报，2000，21(3)：57-60
[46] 白博峰，郭烈锦，陈学俊．基于反传神经网络和压差波动识别气液两相流流型．化工学报，2000，51(6)：848-851
[47] 白博峰，郭烈锦，陈学俊．气液两相流流型在线智能识别．中国电机工程学报，2001，21(7)：46-50
[48] 白博峰，郭烈锦，赵亮．汽(气)液两相流流型在线识别的研究进展．力学进展，2001，31(3)：437-446
[49] 赵振宇，徐用懋．模糊理论和神经网络的基础与应用．北京：清华大学出版社，1996，6：80-115
[50] Martin T. Hagan, Howard B.Demuth, Mark H.Beale. Neural Network Design. Beijing: China Machine Press, 2002,9:8-425
[51] 飞思科技产品研发中心．MATLAB6.5辅助神经网络分析与设计．北京：电子工业出版社，2003，1：165-192
[52] 许东，吴铮．基于MATLAB6.x的系统分析与设计——神经网络．西安：西安电子科技大学出版社，2002，9：166-237
[53] D．Hanselman，B．Littlefield著；张航，黄攀译．精通MATLAB6．北京：清华大学出版社，2002，6：23-309
[54] 李海涛，邓樱．MATLAB6.1基础及应用技巧．北京：国防工业出版社，2002，3：235-261
[55] 徐苓安．相关流量测量技术．天津：天津大学出版社，1988：1-150
[56] 邓湘．基于层析成像技术的智能化两相流测量系统研究：[博士论文]．天津：天津大学图书馆，2001，7：1-95
[57] 卢文祥，杜润生．机械工程测试·信息·信号分析．武汉：华中理工大学出版社，1999，8：6-264
[58] 胡丹丹，肖书明，王燕清等．基于多传感器的数据融合技术．东北电力学院学报，2004，24(1)：62-67
[59] 秦志强．数据融合技术及其应用．网络信息技术，2003，22(5)：25-28