池式沸腾和流动沸腾的人工神经网络研究

Study on Pool Boiling and Flow Boiling with Artificial Neural Networks

成功构建了2个分别用于预测同心圆管开式热虹吸器内自然循环临界热流密度(CHF)和池式核态沸腾换热系数的人工神经网络。其预测均方误差分别为16.43%和19.57%。用训练成功的人工神经网络分析了2种沸腾换热的影响因素,分析结果表明:热虹吸器内同心内管的出现使CHF增加,热虹吸器内的CHF随内管外径的增加先增加后减小。池式核态沸腾表面传热系数随压力的增加先呈线性增加,当压力接近临界压力时,增加速度增大。

In this paper,two artificial neural networks (ANNs) are trained successfully to predict the CHF of thermosyphon and heat transfer coefficient of pool nucleate boiling respectively. The root mean square of predicated value are 16.43% and 19.57%,respectively. The analysis results indicate that CHF would be improved by inserting an inner tube in the thermosyphon. CHF increases initially as inner tube diameter increases and then decreases with the further increase of inner tube diameter. The heat transfer coefficient of pool nucleate boiling increases linearly as pressure increases,and when the pressure is close to the critical pressure,the increasing rate increases.