Pump failure data analysis: VTT's contribution to the EuReDatA Benchmark Exercise on data analysis

Kaisa Simola, Tapio Huovinen, Esko Lehtinen, Urho Pulkkinen, Antti Lyytikäinen

Research output: Book/ReportReport


The report describes the contribution of Technical Research Centre of Finland (VTT) to the EuReDatA Benchmark Exercise-project on data analysis. In the project, a common data set taken from the Component Event Data Bank was analysed by seven participating organizations. The aim of the study was to compare statistical methods to estimate reliability parameters from raw data. The data set consisted of operating experiences of pumps in condensate and feedwater system, and in auxiliary feedwater system in 16 European power plants. During the Benchmark Exercise, a more homogeneous subset of the original data set, including only condensate and feedwater system pumps, was selected for further analyses in order to obtain comparable results. In VTTs approach, preliminary analyses are conducted in order to identify typical failed parts and failure trends. Simple performance indicators, which can easily be obtained from the data without statistical methods are introduced. Prior to the statistical analyses, the homogeneity and classification of the data are discussed. Statistical ana1yses are performed for failure and restoration times of pumps in order to obtain parameters describing failure and restoration rates. Both classical and Bayesian methods were applied in parameter estimation. The effect of various renewal and censoring principles on results are also discussed.
Original languageEnglish
Place of PublicationEspoo
PublisherVTT Technical Research Centre of Finland
Number of pages118
ISBN (Print)951-38-4029-8
Publication statusPublished - 1991
MoE publication typeNot Eligible

Publication series

SeriesValtion teknillinen tutkimuskeskus. Tiedotteita


  • failure rate
  • data analysis
  • pumps
  • power plants
  • trends


Dive into the research topics of 'Pump failure data analysis: VTT's contribution to the EuReDatA Benchmark Exercise on data analysis'. Together they form a unique fingerprint.

Cite this