A modified Weibull distribution is evaluated for characterizing the statistical strength of ceramic fibers and whiskers with varying diameters from filament to filament. Many commercial ceramic fibers and whiskers have a significant range of diameters. A single-modal Weibull distribution is found inadequate to describe the statistical strength of these fibers and whiskers because of the effect of fiber diameter variation on strength. Procedures for extracting distribution parameters for the modified Weibull distribution from experimental data are presented. Comparison of the modified Weibull distribution with the single-modal Weibull distribution is made for the strength data from Nicalon fibers, Nextel (Al2O3) fibers, hydridopolysilzazlane (HPZ) Si-N-C-O fibers, Al2O3 whiskers, Si3N4 whiskers, and SiC whiskers. Due to its ability to account for the diameter effect on strength, the modified Weibull distribution can yield a more accurate β value than the single-modal Weibull distribution. The Modified Weibull distribution is shown to fit experimental data well and is recommended for characterizing the strength of ceramic fibers and whiskers, the diameters of which vary from filament to filament.

Plane strain compression is a versatile laboratory testing method for simulating industrial hot working operations such as plate and strip rolling. The deformation can be closely controlled to the required conditions of temperature and strain rate, and high strains can be achieved without instability. However, for accurate determination of flow stress, care must be taken with experimental procedure and the interpretation of the measured force-displacement data. This paper reports the results of work carried out in three laboratories on samples of the same cast of Al-1% Mn alloy. In particular, the effects of spread and friction are analyzed as a function of initial specimen geometry. When consistent procedures are used it is shown that excellent reproducibility of flow stress data is obtained between laboratories. In deriving constitutive relationships, the importance of considering the effects of lubricant films, of deformational heating, and of heat transfer are clearly demonstrated.

Steel welded T-joints with 102 mm thick base plates were fatigue tested in three-point bending. A reversing direct current potential drop (DCPD) system was used to detect the initiation of multiple surface cracks along the transverse weld toes of these joints and to monitor the size and shape of fatigue cracks that developed from these cracks. Potential drop readings were obtained by fixed probes straddling the weld toes and normalized with respect to potential drop readings from a remote reference probe. The normalized potential drop readings from each probe were related to the local crack depth by two-dimensional (2-D) calibrations derived by boundary element analyses and the foil analogue method. The fixed-probe arrangement was able to detect 0.4 mm to 1.0 mm deep surface cracks along the transverse weld toes of the T-joints. The 2-D calibrations were able to predict the crack depth at the deepest points of these cracks to within ±10%.

The numerical solution to the frequency equation for the transverse vibration of a simple beam with symmetric overhang is found. The numerical results converge to the analytical solutions for the two limiting cases of a beam with no overhang and a beam with no span and agree with the case in which the supports are at the nodal points of a freely vibrating beam. An approximation to the solution of the frequency equation for beams with small overhang is presented and compared to the numerical solution. This simple yet accurate approximation is most useful to determine a beam's flexural stiffness, EI, or modulus of elasticity, E, by freely vibrating a simply supported beam.

An experimental methodology was developed for identification of mechanical behavior laws, in which the evolution of the stresses and strains in the surface of the samples is determined by X-ray diffraction and electric resistance strain gages, respectively, as a function of the loading in four-point bending tests. The determined stress-strain curve is then characteristic of the near-surface layers of the samples. The method was validated for different steels by comparing the results of samples with homogeneous properties on the cross section with the results of tension tests. The developed methodology was applied to the study of steel samples submitted to surface treatments, presenting gradients of chemical composition, microstructure, and mechanical properties from the surface to the bulk material.