针对电容测量灵敏度分布易受被测两相流介质分布的影响，对由电容传感器阵列采集的反映两相流介质分布的信息进行模糊处理，经Ｋｏｈｏｎｅｎ 网络的自组织学习，提取输入模式样本的分类特征，通过ＢＰ网络的监督学习，可实现两相流流型的有效判别和分类 Sensitivity distribution of capacitance measurement is sensitive to the distribution of the measured two-phase flow media. The information collected by the capacitive sensor array reflecting the distribution of two-phase flow media is obfuscated. The classification of input mode samples is extracted by self-organizing learning of Kohonen network Features, through the BP network of supervised learning, two-phase flow patterns can be effectively identified and classified