Time-Dependent Reliability Analysis of Monolithic Ceramic Components Using the CARES/LIFE Integrated Design Program

Abstract

The computer program CARES/LIFE calculates the time-dependent reliability of monolithic ceramic components subjected to thermomechanical loading. This program is an extension of the CARES (Ceramics Analysis and Reliability Evaluation of Structures) computer program. The CARES/LIFE code consists of separately executable modules. These models function to calculate material parameters from requisite data of naturally flawed specimens; create a neutral database from either MSC/NASTRAN, ANSYS, or ABAQUS finite element analysis results files; and compute component reliability. The phenomenon of subcritical crack growth (SCG) is modeled with the power law relation where the crack velocity is a function of the equivalent stress intensity factor. Another model for SCG is based on a phenomenological criterion (Paris law), traditionally used for metals. The two-parameter Weibull cumulative distribution function is used to characterize the variation in component strength. The effects of multiaxial stresses are modeled by using either the principle of independent action, Weibull's normal stress averaging method, or Batdorf's theory. Batdorf's model combines linear elastic fracture mechanics with extreme value statistics. It requires a user-selected flaw geometry and fracture criterion to describe volume or surface strength defects. Material/environmental inert strength and fatigue parameters are obtained from rupture strength data of naturally flawed specimens. For inert strength fracture data, Weibull parameter estimation can be performed for unimodal or concurrent surface and volume flaw populations by using least-squares analysis or the maximum likelihood method. For static, dynamic, or cyclic fatigue fracture data, fatigue parameters are obtained using least-squares analysis or the median value technique. A method for obtaining the requisiste material inert strength and fatigue parameters from the fatigue data is included. Kolmogorov-Smirnov and Anderson-Darling goodness-of-fit test statistics are available for data analyzed with these procedures. For fast fracture (inert) data, 90% confidence intervals on the Weibull parameters and the unbiased value of the shape parameter for complete samples are also provided. Example problems demonstrating various features of the program are included.