中文摘要：

本文基于光电二极管阵列传感器与Fisher判别分析（FDA）,提出了一种新的小管道气液两相流流型辨识方法。该方法采用片状激光照射透明管道,利用光电二极管阵列传感器接收透射激光分布信号。通过提取信号的特征向量,采用FDA训练得到流型分类器,从而实现两相流流型辨识。本实验针对四种典型流型,在四种不同内径的小管道内进行实验研究。实验结果表明提出的基于光电二极管阵列传感器与FDA判别分析的流型辨识方法是可行的,四种流型辨识精度均高于90%。

英文摘要：

Based on photodiode array sensor and Fisher Discriminate Analysis（FDA）,a new flow pattern identification method for gas-liquid two-phase flow in small channel is proposed.In this method,a planar laser sheet is emitted and passes through the transparent small channel,and the exit laser signal is obtained by the photodiode array sensor.The feature vector of the signal is extracted and then the flow pattern is identified by flow pattern classifiers,which are developed by FDA.In this work,four typical flow patterns are investigated and the experiments are carried out in four small channels with different inner diameters.The experimental results show that the proposed method based on photodiode array sensor and FDA is effective,and the identification accuracy is higher than90%.