Change search
ReferencesLink to record
Permanent link

Direct link
Test and Analysis of Homogeneity Regarding Failure Intensity of Components in Nuclear Power Plants
Uppsala University, Disciplinary Domain of Science and Technology, Mathematics and Computer Science, Department of Mathematics.
2009 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

In the Probabilistic Safety Assessment (PSA) all potential failure events of a system, for example a nuclear power plant, are identified in order to evaluate the total probability for a severe accident. The input data to the PSA, the reliability parameters, are derived by a two-stage Bayesian method which comprises an assumption of homogeneity among the components of a population. If the components are assumed to be inhomogeneous each individual component is assessed its own failure rate. Contrary, if the components are assumed to be homogeneous the data is pooled before a common reliability parameter is derived. However, the motives for making these assumptions are sparsely documented and the purpose of this study is to design a statistical method for testing the homogeneity of sparse data. The chosen test method, a chi-square goodness-of-fit test with consideration taken to operation (or standby) time, is implemented and applied on failure event data for the Nordic utilities. From the tests it can be concluded that the failure intensity for continuously operating components for most populations can be considered homogeneous with regard to failure rate, since in 6 % of the test the null hypothesis cannot be rejected at 0.05 significance level. Test results indicate that populations of standby components are to a larger extent inhomogeneous, which might be explained by differences in the data set due to unequal number of demands. Also larger populations, i.e. components of all plants, must be considered as more inhomogeneous. If the populations are to be regrouped in order to increase the homogeneity a statistical test could be a controlling tool. Besides statistical tests, there is a need to study the consequences in the PSA before a conclusion can be made on what approach is to be preferred.

Place, publisher, year, edition, pages
2009. , 39 p.
UPTEC ES, ISSN 1650-8300 ; 09002
National Category
Engineering and Technology
URN: urn:nbn:se:uu:diva-162564OAI: diva2:460916
Educational program
Master Programme in Energy Systems Engineering
Available from: 2011-12-01 Created: 2011-12-01 Last updated: 2011-12-01Bibliographically approved

Open Access in DiVA

fulltext(1568 kB)335 downloads
File information
File name FULLTEXT01.pdfFile size 1568 kBChecksum SHA-512
Type fulltextMimetype application/pdf

By organisation
Department of Mathematics
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
Total: 335 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 262 hits
ReferencesLink to record
Permanent link

Direct link