Prediction of heat transfer coefficients for steam condensation in the presence of air based on ANN method

Abstract

Artificial neural network (ANN) methods have been gradually used in the field of nuclear reactor thermal-hydraulics as new methods to improve accuracy or fast prediction. This study establishes a back propagation (BP) neural network model based on the ANN methods to predict the steam condensation heat transfer coefficient outside a heat tube in the presence of air. The main factors affecting condensing heat transfer, such as pressure, air mass fraction, subcooling, and tube diameter, are used as input quantities, and the condensation heat transfer coefficient is used as output quantity. A complete set of neural networks for predicting the heat transfer coefficient for steam condensation in the presence of air is established based on the relevant experimental data collected over the world. The results predicted by the ANN model are compared with the experimental data and those of traditional correlation methods. The data from 2276 experiments are distributed within a ±10% error band at the 95% confidence level. This means that the prediction accuracy of the ANN model is higher than that of the traditional experimental correlation. Therefore, the neural network model developed in this study can be used for the prediction of the heat transfer coefficient for steam condensation in the presence of air.

Keywords

• Artificial neural network
• Back propagation
• Steam condensation
• Non-condensable gas
• Passive containment cooling system