Applied statistics in glass technology

JSGT 1957 V41 T185-T219

In glass technology it is often difficult to achieve precision in small-scale experimental tests, because of the characteristic variability of certain attributes of the material, such as strength and dimensions. By employing statistical analysis to measure the experimental error the risk of drawing false conclusions can be covered. In addition to the experimental field, applied statistics can also be used in such things as the study of glass-working processes, and the engineering design of glass vacuum envelopes. Examples of applied statistics in glass technology include the following: (a) the study of dimensional variation, (b) the measurement of high values of electrical insulation resistance of glass–metal seals, (c) thermal shock tests, (d) photoelastic stress measurement, (e) the study of variability in glass engineering design, (f) the determination of risk of failure in glass engineering design, (g) the analysis of gauging machines. Statistical techniques which are illustrated include: (a) calculation of mean and variance, (b) prediction (of proportions within a range), (c) the logarithmic transformation, (d) significance tests, (e) analysis of variance, (f) sampling of attributes, (g) Monte Carlo experiments, (h) standard error of a function of variables, (i) methods of graphical analysis.

S. J. Morrison