Reconstruction of Water Level of Steam Generator Based on Elman Neural Network

Falsehood or forfeiture occurs frequently on the water level indication meters of the steam generator,which grievously affects the judgment of the operator on the nuclear-power plant.Elman Neural Network is a typical dynamic neural network.It can reflect the characteristics of dynamic system directly,especially when dealing with the complicated non-linear object.A method that rebuilding the water level of steam generator with Elman Neural Network in the circumstance of the crevasse of primary stream pipeline appears is described in the paper.In comparison with the emulation results,it shows that Elman Neural Network can reconstruct the water level of steam generator exactly.The outcome can meet the practical needs and give guidance on the safe motion of the marine nuclear-power plant.