A study of the bootstrap method for estimating the accuracy of artificial neural networks in predicting nuclear transient processes

Author

Zio, Enrico

Author_Institution

Dept. of Nucl. Eng., Politecnico di Milano, Italy

Volume

53

Issue

3

fYear

2006

fDate

6/1/2006 12:00:00 AM

Firstpage

1460

Lastpage

1478

Abstract

The quantification of the uncertainty associated to the results provided by artificial neural networks is essential for their confident and reliable use in practice. This is particularly true for control and safety applications in critical technologies such as those of the nuclear industry. In this paper, the results of a study concerning the use of the bootstrap method for quantifying the uncertainties in the output of supervised neural networks are reported. A thorough parametric analysis is performed with reference to a literature problem. A case study is then provided, concerning the prediction of the feedwater flow rate in a Boiling Water Reactor.

Keywords

fission reactor cooling; fission reactor monitoring; fission reactor safety; neural nets; nuclear engineering computing; artificial neural networks; boiling water reactor; bootstrap method; confidence intervals; control application; critical technologies; feedwater flow rate; nuclear industry; nuclear transient processes; parametric analysis; prediction intervals; safety applications; Artificial neural networks; Bayesian methods; Biological neural networks; Inductors; Industrial control; Intelligent networks; Monitoring; Power system modeling; Safety; Uncertainty; Boiling Water Reactor; bootstrap; confidence intervals; neural network uncertainty; prediction intervals;

fLanguage

English

Journal_Title

Nuclear Science, IEEE Transactions on

Publisher

ieee

ISSN

0018-9499

Type

jour

DOI

10.1109/TNS.2006.871662

Filename

1645061