Erosion Evaluation of Gas-Turbine Grade CMC’s at Room and Elevated Temperatures

Abstract

The objective of this study is to predict ceramic matrix composites (CMCs) erosion behavior and Retained Strength (RS) under environmental conditions using an Integrated Computational Material Engineering (ICME) physics-based approach. The state-of-the-art erosion analysis using phenomenological algorithms and Finite Element Models (FEM) models follows a test duplication methodology and is not able to capture the physics of erosion. In this effort, two CMC systems are chosen for Erosion evaluation: (a) Oxide/Oxide N720/alumina; and (b) MI SiC/SiC. Experiments are conducted at room and elevated temperatures (RT/ ET). Erosion testing considers: (i) a high velocity oxygen fuel (HVOF) burner rig for ET, and (ii) a pressurized helium impact gun for RT. Erodent particles are chosen as alumina and garnet. Experimental observations show that the type of erodent materials affects CMC erosion degradation at ET. Alumina exhibits to be an effective erodent for maintaining a solid phase particle erosion, while Garnet, experiences some degree of melting. Erosion of the oxide/oxide composite is more severe for the same erodent, temperature, mass, and velocity conditions than the MI SiC/SiC composite for all conditions tested. In general, increasing erosion temperature results in increasing erosion rate for the same erodent mass/velocity condition. In conjunction with experiments, a computational Multi-Scale Progressive Failure Analysis (MS-PFA) is also used to predict erosion of the above-mentioned material systems at RT/ET. The MS-PFA augments FEM by a de-homogenized material modeling that includes micro-crack density, fiber/matrix, interphase, and degrades both fiber and matrix simultaneously during the erosion process. Erodent particles are modeled by Smooth Particle Hydrodynamic (SPH) elements. Erosion evolution in CMCs considering strain rate effect predicts a) spallation, b) mass-loss, and c) damages in fiber, matrix, and their interphase. ICME modeling is capable of predicting the erosion process and reproducing the test observation for the MI SiC/SiC at RT, where: a) erodent particles break up the layer of matrix covering fiber due to interlaminar shear (delamination); b) fiber is fractured because of brittle behavior; c) the process (erosion tunneling) continues till it gets to the next thick matrix layer that slows down the tunneling; and d) Erosion tunnel widens as exposed fiber layers are removed (eroded). Simulations are also performed for erosion of the oxide/oxide due to glass beads at RT and ET. Predictions show that erosion rate is lower at ET because voids in the CMC vanish and the glass beads are less effective at ET. Finally, prediction of retained strength of eroded CMC test specimens is predicted by MS-PFA.