DOI: 10.1615/ICHMT.1992.ExpSystComputSimEnergEngin
ISBN Print: 1-56700-031-2
Neural Network Model for On-Line Thermal Margin Estimation of a Nuclear Power Plant
Краткое описание
A new approach for on-line thermal margin monitoring of a PWR core is proposed in this paper, where a neural network model is introduced to predict the DNBR values for the given reactor operating conditions. The neural network is learned by the Back Propagation algorithm with the optimized random training data, prepared by Latin Hypercube Sampling technique on the wide range of input parameters. The trained network is tested on the steady state operating region as well as for the transient situations where DNB is of me primary concern. The test results show that the higher level of accuracy in predicting the DNBR can be achieved by the neural network which is trained with an appropriate data size.