Journal Name:
- International Journal of Science and Engineering Investigations
Key Words:
Author Name | University of Author |
---|---|
Abstract (2. Language):
In a reactor accident like a loss of coolant
accident (LOCA), one or some signals can not be monitored by
control panel for some reasons such as interruptions and so on.
Therefore a fast alternative method could guarantee the safe
and reliable exploration of nuclear power planets. In this study,
an artificial neural network (ANN) with Elman recurrent
structure is used to predict five thermal hydraulic signals in a
LOCA after the upper plenum break. In the prediction
procedure, a few previous samples are fed to the ANN and the
output value of the next time step is estimated by the network
output. The Elman recurrent network is trained with data
obtained from the benchmark simulation of a LOCA in VVER.
The results reveal that the predicted values follow the real
trends well and ANNs can be used as a fast alternative
prediction tool in LOCA.
Bookmark/Search this post with
- 7
1-4