Statistical Analysis of Data for Timber Strengths

Publikation: Bog/antologi/afhandling/rapportRapportForskning

542 Downloads (Pure)

Abstract

Statistical analyses are performed for material strength parameters from approximately 6700 specimens of structural timber. Non-parametric statistical analyses and fits to the following distributions types have been investigated: Normal, Lognormal, 2 parameter Weibull and 3-parameter Weibull. The statistical fits have generally been made using all data (100%) and the lower tail (30%) of the data. The Maximum Likelihood Method and the Least Square Technique have been used to estimate the statistical parameters in the selected distributions. 8 different databases are analysed. The results show that 2-parameter Weibull (and Normal) distributions give the best fits to the data available, especially if tail fits are used whereas the LogNormal distribution generally gives poor fit and larger coefficients of variation, especially if tail fits are used.
OriginalsprogEngelsk
UdgivelsesstedAalborg
ForlagDept. of Building Technology and Structural Engineering
Antal sider61
StatusUdgivet - 2001
NavnStructural Reliability Theory
Nummer206
Vol/bindR0132
ISSN1395-7953

Bibliografisk note

PDF for print: 68 pp.

Emneord

  • Material Strength Parameters
  • Structural Timber
  • Maximum Likelihood Method
  • Least Square Technique
  • Bending Strength
  • Tension Strength
  • Compression Strength

Citationsformater