CIESC Journal

• 多相流 • 上一篇    下一篇

基于Hilbert-Huang变换和支持向量机的油水两相流流型识别

马龙博 张宏建 周洪亮 贺庆   

  1. 浙江大学控制科学与工程学系,工业控制技术国家重点实验室自动化仪表所
  • 出版日期:2007-03-05 发布日期:2007-03-05

Flow regime identification of oil-water two-phase flow based on HHT and SVM

  

  • Online:2007-03-05 Published:2007-03-05

摘要: 针对油水两相流的测量难题,利用文丘里管对水平管内油水两相流流型进行了研究。基于差压波动信号,提出了Hilbert-Huang变换与支持向量机相结合的流型识别方法。首先计算差压波动信号的均方根,并对其进行归一化处理后作为表征流型的特征向量之一;然后对差压信号进行Hilbert-Huang变换,利用经验模态分解后的多分辨率特征,提取第一层和第二层的能量比作为表征流型的另外两个特征向量;最后利用训练好的支持向量机进行流型识别,对分层流型及分散流型取得了较好的流型识别效果。如果将流型识别与推导得到的文丘里油水两相流流量测量模型相结合,可以较好地实现油水两相流的流量测量。

Abstract: To overcome the difficulties in measuring the oil-water two-phase flow, the oil-water two-phase flow regime was studied with a horizontal Venturi tube and a model of flow measurement was established.A flow regime identification method combining Hilbert-Huang transform (HHT) with support vector machine (SVM) was developed, based on the differential pressure signals obtained from the Venturi tube.With this method, the root mean square of the differential pressure signal was calculated and used as one of the characteristic vectors after it was normalized.Meanwhile, the differential pressure signal was handled with Hilbert-Huang transform.The energy ratios of the first and the second layers were obtained as the other two characteristic vectors closely associated with the flow regime,by using the multi-resolution feature of the empirical mode analysis.Then,the flow regime was identified with SVM, and better results were obtained.The oil-water flow measurement could be better made if flow regime identification was combined with the model of the oil-water two-phase flow measurement using a Venturi tube.