基于遗传神经网络方法的流动不稳定起始点研究

Study on Onset of Flow Instability by Genetic Neural Network

利用遗传神经网络(GNN)方法分析窄矩形通道内流动不稳定起始点(OFI),并检测其热流密度随各个系统参数的变化。检测结果显示,GNN的预测结果与实验值符合良好,误差在±10%范围内。进一步通过GNN模型预测各个系统参数对OFI的影响。结果显示:OFI点的热流密度随着系统压力、入口过冷度、质量流速的增加而增大;系统压力对OFI点热流密度的影响小于质量流速的影响,小于入口过冷度的影响。

The trend of OFI heat flux with the system parameters is studied by the method of genetic neural network. The test result shows that the results of GNN agree well with the results of experiments. The errors fall in the limits of ＋10%. By using the GNN model to predict the effect of parameters on OFI, we can find that the heat flux of OFI grows with the increasing of system pressure, the inlet subcooled temperature and the mass flow. The effect of system pressure on OFI is smaller than that of the mass flow and the inlet subcooled temperature.