基于经验模态分解技术,采用能量评估和阈值概率统计手段,提出了一种油气两相流流型状态监测的新方法。油气两相流差压信号是一典型的非平稳多组分信号,经验模态分解首先将差压信号分解成9阶本征模函数,按照频率范围可以划分为3个子带:高频带(30~50Hz),中频带(5~30Hz)和低频带(0~5Hz).在不同流型下,中频带的能量变化很显著,跟踪捕捉中频带能量变化可以监测两相流流型的跃迁.首先确定不同流型下对应的归一化能量阈值,阈值概率统计技术通过移动时间窗扫描中频带子信号的方式来监测流型状态变化.油气两相流的实验结果表明该方法是有效的,为两相流流型状态监测提供了新途径。
Based on Empirical Mode Decomposition (EMD), a new method was presented to monitor flow pattern of oil-gas two-phase flow by using energy evaluation and threshold probability statistical technique. Because the differential pressure signal of oil-gas two-phase flow was a typical non-stationary and multicomponent signal, EMD was first applied to break up it into Intrinsic Mode Functions (IMFs) which can be divided into three frequency bands such as high-band (30-50 Hz), mid-band (5-30Hz) and low-band (0-5 Hz). The energy variation in mid-band was remarkable under different flow patterns, and the transition of flow pattern can be monitored by catching the energy variation in mid-band. Normalized energy thresholds of different flow patterns in mid-band were obtained, and the threshold probability statistical technique monitored flow pattern by scanning components in mid-band with a moving time-window. The experimental results of oil-gas two-phase flow show th at the proposed method is effective for monitoring the transition of flow pattern of two-phase flow.