Simultaneous Measuring Methods for Gas-Liquid Phase Flow Rates Using Wevelet Analysis

结合海上油田的开发，本研究项目的目的是开发一种测量气液混合物单相流量的新方法。空气和水的混合物通过一根管子向上流动，在管子里放一个楔子。当流体为单相流时，可以利用楔形板上游和下游的压差频率来测量流量。然而，在气液两相流动的条件下，频率和波形都发生了很大的变化。但压力波动的随机特征，如平均压差、波形概率密度分布的标准偏差和偏斜度以及傅立叶分量，可能会随着单个相流率的变化而特殊地改变，正如该方法在语音识别技术中预测的那样。因此，在本项目中，使用这些变量来测量两相流，并使用模式匹配方法和神经网络来识别单独的流量。利用在不同相流量条件下测得的压力波形，对随机特征变量进行了分析。并利用这些变化的强度来识别水的体积流量，得到的流量测量精度在3%以内。为了建立该方法，有必要进行进一步的研究，如对气相流速和S/氮比的适用性，然后继续

Relevant to the development of off-shore oil field, this research project have been started with the intention of developing a new method for measuring individual phase flow rates of gas-liquid mixtures. The air-water mixture flows upwards through a pipe, in which a wedge is put. The frequency of the differential pressure between upstream and downstream of the wedge is available to measure the flow rates when the flow is a single-phase flow. The frequency, however, is much changed under the conditions of gas-liquid flows as well as the waveform. But the stochastic features of the pressure fluctuations, such as the averaged differential pressure, standard deviation and skewness for the probability density distributions of waveforms, and Fourier components, may alter peculiarly with a change in individual phase flow rates, as predicted by the method in a voice recognition technique. Thus, the usage of these variables to measure the two-phase flows is investigated in this project with a pattern matching methods and neural network to identify the individual flow rates. Stochastic feature variables were analyzed using the pressure waveforms measured under the various conditions of phase flow rates. And, the volumetric flow rate of water are identified by using these variable intensity, the resultant flow rate are found to be measured within 3% accuracy. To establish the method, further investigation, such applicability for air-phase flow rates and S/N ratio are necessary, and then is continued