Artificial neural network introductions

The artificial intelligence has attracted broad attention and has been widely applied in the field of nuclear engineering. This section introduce the artificial neural network (ANN), including the concept, the mathematical-physical theory, and the applications in the thermal-hydraulic problems. The back propagation network (BPN), wavelet neural network, and genetic neural network are all included. The training process of neural network has been introduced as well. The existing correlations were derived within a certain experiment range, which can only be developed to the specific conditions. Therefore new methods should been employed to expand the application. Neural network works as an effective way to describe the complicated phenomenon and obtain better results due to its abilities of associative memory, nonlinear mapping, and knowledge processing. The ANN has been used to predict leak rate, pressure drop, onset of nucleate boiling, critical heat flux, onset of nucleation boiling, boiling curve, and heat-transfer coefficient. Reasonable accuracy has been obtained for different neural networks, proving the ability of various neural network in the thermal-hydraulic problems for nuclear engineering.