基于航空发动机轴承腔润滑中所存在的气液两相流问题, 采用基于神经网络的理论方法建立预测模型, 以便进行轴承腔内气液两相流流型的识别。研究以管道气液两相流为原型, 采用 3种典型的神经网络对流型进行模式识别,通过考察 3种网络的辨识率, 发现BP网络的识别方法具有较高的准确性。 Based on the problem of gas-liquid two-phase flow in aero-engine bearing cavity lubrication, a prediction model based on neural network is established to identify the flow pattern of gas-liquid two-phase flow in the bearing cavity. In this paper, the pipeline gas-liquid two-phase flow is taken as the prototype, and three kinds of typical neural networks are used to pattern recognition. By examining the recognition rates of the three networks, it is found that the BP neural network recognition method has high accuracy.