基于压力信号的分析,本文对模拟海洋平台输油管道的集输-立管系统中严重段塞流进行了研究,划分了流型,并对系统中各种流型下压力信号的特征参数进行了提取。通过对这些特征参数的分析,获得了流型的判别规则,并利用人工智能最小二乘支持向量机技术对流型进行了在线识别。通过在线实验表明,使用本文的特征参数提取方法和智能识别技术,能够很好地判断管线中严重段塞流的发生,管中流型的平均识别率达到了96.4%。 Based on the analysis of pressure signals, this paper studies the severe slugging flow in the pipeline-riser system of the offshore pipeline platform, divides the flow pattern, and analyzes the characteristic parameters of the pressure signals under various flow patterns in the system extract. Through the analysis of these characteristic parameters, the discriminant rules of the flow pattern are obtained, and the flow pattern is identified online by the artificial intelligence least square support vector machine technology. The experiments show that using the method of feature parameter extraction and intelligent recognition can judge the serious slug flow seriously. The average recognition rate of flow pattern in the pipe reaches 96.4%.