Application of an artificial neural network to reactor core analysis
- Seoul National Univ. (Korea, Republic of)
To analyze three-dimensional reactor core behaviors, the finite difference or the finite element method have generally been used. Nodal method is adopted as another tool for analyzing transient core characteristics. These methods, however, require much calculation time to solve very complicated iterations for better convergence. Especially when the transient states are to be predicted, none of these methods can meet the requirements within the time span in which the operator can react. To overcome these difficulties, a new analytic model based on the artificial neural networks (ANNs) is suggested. Because trained ANNs are capable of modeling the input/output relationships of a nonlinear system without complex analogy, they are able to map the power distributions and calculate the eigenvalue corresponding to the core conditions in a short time and utilize the previous results by updating the weights of inter-connection between input and output patterns. To confirm the accuracy and capability, daily load-follow operation in a pressurized water reactor (PWR) is simulated using the new analytic model.
- OSTI ID:
- 186687
- Report Number(s):
- CONF-950601--
- Journal Information:
- Transactions of the American Nuclear Society, Journal Name: Transactions of the American Nuclear Society Vol. 72; ISSN TANSAO; ISSN 0003-018X
- Country of Publication:
- United States
- Language:
- English
Similar Records
Development of Fast-Running Simulation Methodology Using Neural Networks for Load Follow Operation
Use of artificial neural networks for analysis of complex physical systems