【作者】 华陈权；

【导师】 王昌明；

【作者基本信息】 南京理工大学 ， 测试计量技术及仪器， 2010， 博士

【摘要】 随着海洋、沙漠及边远油气田的开发,由于开发成本、安装空间和优化管理的要求,迫切需要开发适应性强、低成本的在线湿气流量计替代昂贵笨重的计量分离器。把成熟的单相气体流量计用于湿气计量一直是国内外研究的热点。本文将槽式孔板和旋进漩涡流量计这两种成熟的气体流量计用于湿气计量技术的研究,主要研究成果如下：(1)通过对不同流型的湿气流经管道时诱发的管道振动特性的实验研究,提出了基于流体诱发振动的非接触式在线湿气流型识别新方法。通过安装于测试管道外壁的振动传感器测量不同流型下湿气流诱发的振动信号；采用小波包分析提取了表征流型变化的振动信号的能量特征向量；分别建立了基于径向基函数网络、概率神经网络、支持向量机的水平管湿气流型识别模型。这三种模型对实验样本都有较高的流型识别率。(2)对不同孔径比的槽式孔板的湿气流计量性能进行了研究,确定了“虚高”与L-M参数XLM、气体富劳德数、密度比和孔径比等无量纲参数显著相关,提出了两个槽式孔板湿气气相计量修正模型,计算精度优于其它修正模型,可用于湿气气相流量计量。(3)首次对旋进漩涡流量计的进动频率的湿气流计量性能进行了研究。当xLM>0.12时,由于进动频率失效,不能用于湿气气相流量计量；当XLM≤0.12时,进动频率呈现负偏差,并提出了基于进动频率的旋进漩涡流量计湿气气相计量模型,经该模型修正后的气相质量流量相对误差在95.4%的置信度下小于±5%,可用于XLM≤0.12的湿气气相流量计量。(4)首次对旋进漩涡流量计的差压的湿气流计量性能进行了研究,发现其与“文丘里”具有类似的“虚高”特性,并提出了基于差压的旋进漩涡流量计湿气气相计量模型。经该模型修正后的气相质量流量相对误差在96.4%的置信度下小于±5%,可用于湿气气相流量计量。(5)采用两个不同类型的流量传感器——槽式孔板的差压与旋进漩涡流量计的进动频率组合,进行了湿气两相计量技术的研究,并给出了两相流量迭代求解方法、步骤。实验结果表明：气相质量流量相对误差在89.2%的置信度下小于±6%,液相质量流量相对误差在100%的置信度下小于±20%,可用于XLM≤0.12的湿气两相流量计量。(6)采用旋进漩涡流量计的两个不同类型的输出信号——差压与进动频率组合,进行了湿气两相计量技术的研究,并给出了两相流量迭代求解方法、步骤。实验结果表明：气相质量流量相对误差在91.3%的置信度下小于±8%,液相质量流量相对误差在89.2%的置信度下小于±20%,可用于XLM≤0.12的湿气两相流量计量。这种方法只需要利用一个流量计就可实现湿气两相流量计量,从安装、使用、维护等方面都具有更高的性价比。更多还原

【Abstract】 With the advancing of exploitation of oil and natural gas field into offshore, desert and remote areas, it extremely desirable to have a wet gas metering system that is capable of providing cost effective, compact in size, and on-line measurement of wet gas with sufficient accuracy to replace a bulky and costly test separator, in terms of cost savings, production optimization, field monitoring and reservoir management. The method employing mature gas flowmeters to meter wet gas had been and still is a highly focused subject for many researchers around the world. In this dissertation, two mature gas flowmeter, namely a sloted orifice and a swirlmeter, are employed to meter wet gas. The main contributions can be enumerated as follows:First, a novel noninvasive approach, based on flow-induced vibration, to the online flow regime identification for wet gas flow in a horizontal pipeline is proposed. The flow-induced vibration signals were measured by a transducer installed on outside wall of pipe, and then the normalized energy features from different frequency bands in the vibration signals were extracted through wavelet package transform. The three classifies, namely the RBF neural network, the probabilistic neural network, and support vector machine classifiers, are respectively constructed to identify wet gas regimes. The results show that the three classifiers can identify flow regimes effectively.Second, researches into the wet gas flow performance of the horizontally installed slotted orifice meter were performed. After studying of the relationship between the over-reading and Lockhart-Martinelli parameter X_LM, gas Froude Number, the gas to liquid density ratio and beta ratio, two new metering correlations for low-pressure wet gas flow are proposed. The results show that the calculate accuracy of the correlations are higher than all of the previous ones, and therefore they can be used to meter the gas flow rate for wet gas flow.Third, the wet gas flow performance of a swirlmeter based on vortex precession frequency was uniquely investigated. When X_LM> 0.12, the flowmeter failed to meter the gas flow rate due to disappearance of vortex precession. WhenX_LM≤0.12, the entrained liquid in a gas stream tends to induce a negative bias in the gas flow rate reading of the swirlmeter, and therefore a metering correlation is proposed by using the surface-fitting technology. The results show the correlation improves the gas flow rate prediction errors of the wet gas flow to within±5%at a confidence level 95.4%, and therefore it can be used to meter the gas flow rate for wet gas flow at X_LM≤0.12. Fourth, the wet gas flow performance of a swirlmeter based on differential pressure was uniquely investigated. It shows that the differential pressure of swirlmeter has similar metering characteristics to that of Venturi tube, and and a metering correlation is proposed by using the surface-fitting technology. The results show the correlation improves the gas flow rate prediction errors of the wet gas flow to within±5%at a confidence level 96.4%, and therefore it can be used to meter the gas flow rate for wet gas flow.Fifth, a novel approach to meter individual flow rates for wet gas by using two dissimilar flow sensors, i.e. a sloted orifice and a swirlmeter, is proposed. The steps of iterative solution of obtaining individual flow rates are also given. The results show that the proposed approach predicts the gas mass flow rate within±6%at a confidence level 89.2%, and the gas mass flow rate within±20%at a confidence level 100%, and therefore It can be used to meter individual flow rates for wet gas flow at X_LM≤0.12.Sixth, a novel approach to meter individual flow rates for wet gas by using two dissimilar output signals of a swirlmeter, i.e. vortex precession frequency and differential pressure, is proposed. The steps of iterative solution of obtaining individual flow rates are also given. The results show that the proposed approach predicts the gas mass flow rate within±8%at a confidence level 91.3%, and the gas mass flow rate within±20%at a confidence level 89.2%, and therefore it can be used to meter individual flow rates for wet gas flow at X_LM≤0.12. In view of installation, maintenance and cost, the proposed approach is cost-effective due to using only a flow sensor.更多还原