Article ID Journal Published Year Pages File Type
10629640 Journal of the European Ceramic Society 2014 7 Pages PDF
Abstract
We have performed a statistical evaluation of more than 10 000 experimental strength values of test pieces from a serial production of alumina products where in each production series 12 samples were broken. The Weibull parameters were obtained by three different methods: linear regression, maximum-likelihood and moments' method. The methods give different values of parameters even for this large number of data, but all the correlation coefficients are high and similar. Evaluating Weibull parameters for each testing group gives expected systematically different values of particular Weibull modulus as compared to the whole population of collected data. However, when the order of experimental data is mixed, the average Weibull modulus is lower than that corresponding to time ordered data. This difference can be served as an indication for variation of products' quality from series to series.
Related Topics
Physical Sciences and Engineering Materials Science Ceramics and Composites
Authors
, , ,