Probabilistic Life Evaluation of Ceramic Matrix Composite Combustor Liner

Probabilistic progressive failure analysis (PFA) is used to predict the life of a Sylramic/MI ceramic matrix composite (CMC) combustor liner in a JTAGG-III engine under both rig and engine conditions. PFA simulations of the combustor liner were made to predict fracture paths and life cycles based on microcrack damage evolution and the contribution of failure modes to total failure. Combustor liner life was also assessed probabilistically by simulating the effects of material and fabrication uncertainties on damage characteristics. The simulation results showed that: 1) creep is strongly dependent on uncertainties in design variables and 2) failure risk can be most effectively reduced by controlling a CMC’s fiber content, fiber tensile and compressive strengths, fiber modulus, and matrix compression strength. Introduction A Sylramic/MI CMC combustor liner was computationally simulated using GENOA [1,2] software to perform: 1) progressive failure analysis (PFA) and 2) probabilistic PFA analysis of CMC structure. GENOA’s progressive failure analysis computer code builds from the constituent material properties at the micromechanics level to establish properties at the macro-structural level. The GENOA PFA simulation of CMC degradation under loading [3] takes into account the percent contribution of all possible failure mechanisms to the occurrence of total failure. The capability of the GENOA software for analysis of CMC structure was enhanced with the specific capability to predict the creep life of a CMC combustor liner under both the engine and rig environmental conditions. Damage initiation, growth, accumulation, and propagation were simulated to global failure. GENOA’s probabilistic composite structural analysis code was used to predict the uncertainty of the CMC combustor liner’s creep life. The constituent material strengths and moduli were specified as random variables and used to predict cumulative distribution curves for creep life. The sensitivities to significant random variables were also predicted. Extensive efforts to identify the various modes of damage in composite materials have been undertaken in recent years. The primary finding of most of these investigations was that a macroscopic fracture usually resulted from coalescing of microscopic damage induced by various failure modes. Additionally, it was generally found that analyses based on classical fracture mechanics did not adequately support prediction of fracture initiation and growth in composite structures [4,5]. Predicting damage propagation in CMC structure is influenced by both local factors, such as constituent material properties at the location of damage, and global factors, such as structural geometry and boundary conditions. The interaction of these factors, complicated by the numerous possibilities of material combinations, composite geometry, fiber orientations, and loading conditions, renders the assessment of composite damage progression very complex.