Numerical study and analysis of thermal parameters of subcooled flow boiling and presentation of prediction models based on artificial neural network algorithm

In the present study, using axisymmetric numerical simulation based on the Euler-Euler method, subcooled flow boiling of pure water in a pipe was investigated and the local and average heat transfer coefficient, local and average vapor volume fraction, and local and average wall temperature under different boundary conditions have been investigated. According to the results obtained from the numerical simulations, the wall temperature increases with increasing pressure. Also, the vapor volume fraction has decreased with increasing pressure. The effect of heat flux on wall temperature and vapor volume fraction is greater than all other boundary conditions. Although numerical approaches give a complete insight into the flow pattern and thermal characteristics, the simulation of complex multiphase flows requires high computational resources and is very time-consuming. In conclusion, we present a deep learning approach based on artificial neural networks to predict the mentioned parameters in pure water. The models presented in the present study accurately predict the output parameters using the hyperparameter tuning method. The results of the prediction models show that these models are able to accurately predict the objective functions with an average absolute error of less than 2.5% and a coefficient of determination greater than 0.9.