【作者】 李伟波；

【导师】 金宁德；

【作者基本信息】 天津大学 ， 检测技术与自动化装置， 2005， 硕士

【摘要】 论文中首先采用符号时间序列分析表征了垂直上升管中油水两相流流型的变化。与先前分形维、混沌吸引子关联维及Kolmogorov熵得到的结果相比,时间不可逆转性T_fb及χ²统计量对水包油流型变化不敏感,而对过渡流型变化呈现不规则突变,显示出在区分水包油与过渡流型时,符号时间序列分析方法更具有独特特点,是辨识油水两相流流型的有用辅助诊断工具。相含率是油水两相流测量的一个重要参数,本文对国内外两相流相含率测量方法进行了总结,建立了油水两相流纵向多极电导式阵列传感器测量系统,实现了数据采集与分析,并利用数据采集卡采集传感器信号,利用LabVIEW编程同步采集三路信号,在实验室条件下对油水两相流测量系统做了初步试验,验证了该测量系统的适用性。将软测量技术应用于油水两相流的相含率预测中,分别从时域及频域提取特征量,并作为神经网络的输入。油水两相流波动信号的时域特征采用一些传统的统计量,频域特征借助语音信号处理中的线性预测方法展开,此外将纵向多极电导式传感器上下游测量数据计算的渡越时间作为神经网络的输入。应用人工神经网络技术取得了具有较高精度的相含率预测结果。更多还原

【Abstract】 The flow patterns of oil/water two-phase flow in vertical upward pipes werecharacterized by the analysis of symbolic time series based on the conductancefluctuating signals. The study showed that the symbolic sequence temporalirreversibility T_fb and chi-square χ² statistics had little changes with oil-in-water flowpattern variations for water cut （ K_w） ranging from 61% to 91% and irregular suddenchanges to transitional flow pattern variations of 51%. When distinguishing thetransitional flow pattern from oil-in-water, the symbolic time series analysis methodpresents more unique characteristic and is a useful assistant diagnostic tool for theidentification of oil/water two-phase flow patterns.The volume fraction is an important parameter of the two-phase flowmeasurement. The methods of the volume fraction measurement were summarizedand the vertical multi-electrode sensor measurement theory was analyzed in this paper.The measurement system of vertical multi-electrode sensor was established. The threeway signals of multi-electrode sensor are acquired by use of the dynamic signalacquisition device and LabVIEW. We did the exerperiments in the laboratory andachieved ideal results.The soft measurement method was applied to predict the volume fraction ofoil/water two-phase flow. The characteristics of time domain and frequency domainact as the inputs of artificial neural network. Several traditional statistics represent thetime domain characteristics. The frequency domain characteristics are obtained by thelinear prediction method. The transition time of the correlation electrodes also acts asthe input characteristic. The prediction results indicate that the soft measurementtechnique may be used to predict the volume fraction of oil/water two-phase flow.更多还原