Prediction of Strength for Inhomogeneous - Defective Glass Elements Based on the Sequential Partitioning of the Data and Weibull Statistical Distribution

  • Alexander Shabetia G.S. Pisarenko Institute for Problems of Strength
  • Yurii Rodichev G.S. Pisarenko Institute for Problems of Strength
  • Frederic Veer Delft University of Technology
  • Elena Soroka G.S. Pisarenko Institute for Problems of Strength

Abstract

An analytical approach based on the on the sequential partitioning of the data and Weibull Statistical Distribution for inhomogeneous - defective materials is proposed. It allows assessing the guaranteed strength of glass structures for the low probability of fracture with a higher degree of reliability. Parameters of equations for the piecewise linear approximation for Weibull statistical distribution have been defined on the example of processing of bending tests results for float glass. The advisability of using this approach to structural elements of different size is proved. It was shown that excluding the minimum values from the sample does not lead to the uni-modal distribution. A group of values, forming the lower branch of the distribution, appears again. Statistical analysis of the distribution curves made it possible to identify groups of defects, the technological removal of which would ensure an increase in the guaranteed level of strength. The results are the basis for solving optimization problems when you need to get a guaranteed level of strength for a given probability of fracture with minimal costs for glass element manufacture and treatment.

How to Cite
SHABETIA, Alexander et al. Prediction of Strength for Inhomogeneous - Defective Glass Elements Based on the Sequential Partitioning of the Data and Weibull Statistical Distribution. Challenging Glass Conference Proceedings, [S.l.], v. 6, p. 749-758, may 2018. ISSN 2589-8019. Available at: <https://journals.open.tudelft.nl/index.php/cgc/article/view/2195>. Date accessed: 18 nov. 2018. doi: https://doi.org/10.7480/cgc.6.2195.
Published
2018-05-06