人工神经网络在圆管临界热流密度数据处理中的研究

利用人工神经网络理论对均匀加热垂直上升圆管内的临界热流密度(CHF)进行了预测。分别采用进口条件、出口条件以及局部条件假设，利用收集到的6941个CHF实验数据中的一半作为神经网络训练的样本，采用训练成功的网络预测CHF值可得到比常规方法更好的效果，其均方差分别为6．6％、10．39％和21．39％。

Study on Tube Critical Heat Flux Data Treatment With Artificial Neural Networks

Prediction of the Critical Heat Flux(CHF) are analyzed by Artificial Neural Networks(ANN) to a CHF database for upward flow of water in uniformly heated vertical round tubes. The analysis is performed with three viewpoints hypothesis, i.e. for fixed inlet condition, fixed exit condition and local condition. Half of 6 941 from CHF database data is trained through ANN, the trained ANN predicts the total CHF data better than any other conventional correlations, showing RMS error of 6.6%, 10.39% and 21.39%,respectively.