Stochastic analysis of the fracture of solids with microcracks (Q1898022)

This paper is concerned with the stochastic formulation of the fracture of solids with microdefects. The analysis focuses on the development of microcracks in high-temperature compounds due to the application of an external constant load that ultimately leads to the failure of the material. For this purpose, the authors introduce the concepts of stochastic population processes and some results of the general theory of point processes. The development of the microcracks in a given material domain depends on certain probability laws with respect to the probabilistic function space. The collection of microcracks is regarded as a family of objects and is represented by its state. The material system is an abstract dynamical system in which the number of the microcracks is a random variable. The evolution with time of a given initial distribution of microcracks occuring in the solid is regarded as a stochastic population process. At first, the population growth process is treated as a continuous time parameter process. The development of the microcrack population is characterized by a Markov branching process for the microcrack field. The introduced definitions exclude any interaction effects between individual microcracks and assume that the development of the microcrack field is restricted to intergranular crack formations and is determined by the volume concentration of the microcracking. Then, the authors review some results of the theory of populations and branching processes. In accordance with the usual formulation of population mechanics, a continuous time parameter Markov process is specified by two functions. The first refers to the probability of an initiation of a microcrack, the second accounts for a transformation of a microcrack within the considered time interval. For a continuous time process, the transitions involving the microcrack population size as a whole occur in the form of sudden jumps, and thus the general theory of discontinuous Markov processes applies. The stochastic process is studied as a one-dimensional continuous time Markov process. In this case, a phenomenological model is expressed in form of the parabolic equation. In order to include a time homogeneous process in the more general non-homogeneous process, the authors discuss the probability measures and the \(\sigma\)-finite algebra of events observed in a given time interval. Governing equations for the development of microcracks are derived, and the results of the analysis are compared with experimental observations obtained from short-time creep tests on a silicon-nitride and tungsten-carbide compounds. Finally, in the frame of the microcrack random field concept, the consideration of excursion sets and the extreme values of strains in the elements of such sets are presented.