Abstract:

In this research, based on virtual ECT sensors and flow-electric coupling simulation method, four typical oil gas two-phase flow patterns and their corresponding measurement signals were dynamically simulated. Time-series signals of pressure and capacitance were analyzed by applying multi-resolution analysis (MRA) and time-frequency analysis methods based on wavelet transform. Continuous wavelet transform can display the energy distribution in the time-frequency two-dimensional domain and effectively locate flow fluctuations. Based on the wavelet MRA method, the original signals were decomposed into different frequency bands within the same time range. As results, both pressure and capacitance signal analyses showed the similar migration pattern of main frequency banded from middle to low and then to high frequency with flow pattern changes, which provided flow pattern identification with a potential criterion. This method was validated through experimental measurements in the intermittent flow, effectively reflecting the changes in slug frequency, flow fluctuation, and liquid slug morphology under different flow rates. The research method and results in this paper are expected to provide reliable means for flow pattern identification and process monitoring in practical industrial processes.

Key words: electrical capacitance tomography, two-phase flow, wavelet transform, multi-resolution analysis

摘要：

基于虚拟电容层析成像（electrical capacitance tomography，ECT）传感器和流电耦合仿真方法，本文动态模拟了4种典型的油气两相流流型及其相应的测量信号，并采用基于小波变换的多分辨率分析（multi-resolution analysis，MRA）和时频分析方法对压力和电容的时间序列信号进行分析。连续小波变换显示了时频二维平面中的能量分布，并在时间上有效定位了瞬时流量波动。而基于小波MRA方法对流型变化下的原始测量信号进行不同频带的分解后，压力和电容测量信号的主导频率均呈现出由中频到低频再到高频的迁移规律，为流型识别提供了有效的量化判据。在间歇流实验测量中应用信号时频分析方法，有效解析了不同流量下段塞频率、流量波动和液塞形态的变化。本文的研究方法和研究结果有望为实际过程中的流型识别和过程监测提供可靠手段。

关键词: 电容层析成像, 两相流, 小波变换, 多分辨率分析