S-70/PO |
Neural Networks application to CRDM and thermo-hydraulic data validation |
|
M. Sepielli 1), M. Palomba 2), A. Ratto 3), R. Rosa 2), M. Bernabucci 3) |
1) ENEA, Roma, Italy
2) ENEA, FPN-FISON
3) ENEA, Guests |
|
Abstract:
The work is carried out on ENEA RC-1, a Triga Mark II 1 MW - demineralized water / natural convection cooled - research reactor, located at ENEA Casaccia Centre. Activity is aimed at intelligent processing of the data obtained by reactor measurements through soft-computing models based on neural networks (NN). A first application is made for CRDM rod position validation and a second one is devoted to the fuel temperature prediction. Both use the NN-based algorithms working on real data coming from sensors. The bias between real and calculated values is used to train the NNs for improving the performance in the further experiments.
|
|