It is widely recognized that brittle material strengths are statistical
in nature. Specimen-to-specimen strength differences arise
from variations in the flaws and the microstructure. Hence, it is necessary
to use sample sets of ten or more specimens for even the
simplest of goals such as determining an average strength for material
ranking, material development, or materials specification purposes.
Random and systematic testing errors can generate faulty
data that obscure or spoil even these simple tasks. Such errors can
also contaminate data for other tests such as fracture toughness experiments
that use precracked beams in bending. If the goal of the
testing is to estimate the intrinsic strength scatter of a material, any
superimposed testing variability will render estimates incorrect.
The use of statistics to account for variability from an assigned
cause, i.e., the material’s intrinsic flaw and microstructural variations,
is different from the normal usage of statistics in experimental
work. This creates a requirement that the experimental error of
ceramic strength test data must be negligible compared to the intrinsic
material variability.
It is widely recognized that brittle material strengths are statisticalin nature. Specimen-to-specimen strength differences arisefrom variations in the flaws and the microstructure. Hence, it is necessaryto use sample sets of ten or more specimens for even thesimplest of goals such as determining an average strength for materialranking, material development, or materials specification purposes.Random and systematic testing errors can generate faultydata that obscure or spoil even these simple tasks. Such errors canalso contaminate data for other tests such as fracture toughness experimentsthat use precracked beams in bending. If the goal of thetesting is to estimate the intrinsic strength scatter of a material, anysuperimposed testing variability will render estimates incorrect.The use of statistics to account for variability from an assignedcause, i.e., the material’s intrinsic flaw and microstructural variations,is different from the normal usage of statistics in experimentalwork. This creates a requirement that the experimental error ofceramic strength test data must be negligible compared to the intrinsicmaterial variability.
การแปล กรุณารอสักครู่..