Nuclear Power Plant Burst Parameters Prediction During a Loss-of-Coolant Accident Using an Artificial Neural Network
Several researchers have concentrated on analyzing the nature of fuel claddings through performing burst experiments on computed loss-of-coolant accident scenarios and creating practical and conceptual computer programs. In comparison to experimental observation, the established burst criteria (a) a...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Published: |
Springer Science and Business Media Deutschland GmbH
2022
|
| Online Access: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85120849564&doi=10.1007%2f978-981-16-6636-0_31&partnerID=40&md5=4ed1fdf5b8a0397cc285d0a646c8ed36 http://eprints.utp.edu.my/28915/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Several researchers have concentrated on analyzing the nature of fuel claddings through performing burst experiments on computed loss-of-coolant accident scenarios and creating practical and conceptual computer programs. In comparison to experimental observation, the established burst criteria (a) assumes that the cladding tube deforms in a symmetrical manner (b) infers the characteristics of Zircaloy-4 cladding for mixed-phase of α + β step (c) ignores azimuthal temperature variations. To resolve all of the shortcomings of the burst criteria, this paper proposed an artificial neural network to forecast the burst parameters. In this research, a feedforward backpropagation algorithm with the logsig activation function is used to build this neural network model. A neural network architecture of 2-15-15-15-3, which is a model of three hidden layers containing fifteen neurons in each layer is designed. The mean deviation of burst temperature, burst stress, and burst strain gained from the burst criteria is 1.15, 3.82, and 39.41, respectively, while these parameters are predicted by the proposed neural network includes mean deviations of 0.43, 1.57, and 3.85, respectively. The proposed neural network has been discovered to be more efficient than existing models. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. |
|---|