Application Of New Information Processing Technology To The Flow Pattern Identification Of Gas-Liquid Two-Phase Flow

The two-phase flow system is a complex, nonlinear and dynamic system, which exists widely in processes such as power, chemical, petroleum and metallurgy industries. Research of hydrodynamics, mechanisms and accurate measurement of parameters in two-phase flow are very important for the development of modern industry. Flow pattern identification is one of most important research fields because the measurement accuracy of other flow parameters depends on the flow patterns.Based on theoretical analysis and a large amount of experimental data, nonlinear signal processing and data fusion techniques are applied to the parameter measurement of gas-liquid two-phase flow in this dissertation. Higher-order statistics, Hilbert-Huang Transform, multi-sensor data fusion and fuzzy set theory are used for the flow pattern identification of gas-liquid two-phase flow in horizontal pipeline. Meanwhile, a multi-parameter fuzzy data fusion model is established, and the disadvantage of sole identification method is overcome. So the proposed method has made great progress in flow pattern identification of two-phase flow field. The experimental results show that the presented methods are effective. The main works are listed as follows:1) Based on higher-order statistics technique, a new method for differential pressure fluctuation signal analysis of gas-liquid two-phase flow is proposed. The differential pressure fluctuation signals obtained from the differential pressure transducer are analyzed by using bispectral method and the average bispectral amplitude is regarded as an eigenvalue. The experimental studies prove that the extent of differential pressure fluctuation signals deviation from Gaussian distribution is different for different flow patterns. Combined with fuzzy pattern recognition theory, the fuzzy identification criterions of flow patterns are established. Theexperimental results prove the proposed method can suppress Gaussian noises of signal and is effective for flow pattern identification.2) Hilbert-Huang Transform(HHT) is applied to the nonlinear and non-stationary characteristics analysis of differential pressure fluctuation signal of gas-liquid two-phase flow in the pipeline of small - medium diameter. By the empirical mode decomposition(EMD) method, the local time-frequency characteristics are emphasized. The energy eigenvalues of differential pressure fluctuation signals are obtained to analyze the relationship between the energy distribution of signal and the flow patterns. Based on the extracted eigenvalues, a new method is presented to identify the flow patterns. The experimental results in pipelines of different diameters show that the energy eigenvalues reflect the flowing state of gas-liquid two-phase flow. The proposed method is simple, practical and free of the influence of the pipeline size. In order to illustrate the effectiveness of the adopted approach, the analysis results of differential pressure fluctuation signals based on HHT are compared with those based on traditional Wavelet Transform (WT).3) The application of data fusion technique for target identification has been studied. Some problems such as fusion level, architecture and fusion algorithm of data fusion are analyzed. Based on D-S evidence theory and fuzzy set theory, a multi-parameter fuzzy data fusion model is established to implement the fusion of the results of target identification at decision level. By fuzzifying the local decisions from separate sensors, the belief functions of the target are obtained. A modified D-S evidence theory is introduced as the fusion algorithm in the fuzzy fusion model, thus the effect of the conflicting evidences combination to the fusion result is avoided.4) Based on the proposed multi-parameter fuzzy data fusion model, a new flow pattern identification method of gas-liquid two-phase flow ispresented to improve the accuracy. The experiment is carried out in 40mm-diameter pipeline. By using two differential pressure transducers, two groups of differential pressure fluctuation sign...