Original Article
Correlation between failure mechanism and rupture lifetime of 2D-C/SiC under stress oxidation condition based on acoustic emission pattern recognition

https://doi.org/10.1016/j.jeurceramsoc.2020.06.070Get rights and content

Abstract

Creep tests of 2D-C/SiC in a wet oxidizing atmosphere were implemented for six samples. The loading process was monitored by acoustic emission (AE). Principal component analysis and a fuzzy clustering algorithm were used to perform pattern recognition of the AE data. All of the AE events were divided into four clusters and labelled as matrix cracking, interfacial damage, fiber breakage and fiber-bundle breakage respectively, according to their physical origin. It was found C/SiC has very scattered rupture lifetimes even under the same test conditions, and the evolution of AE events corresponding to fiber failure is quite different. With increasing rupture lifetime, the AE energy of fiber-bundle breakage is higher, while the number of these events is less. Thus, it is concluded that local oxidation and damage development is the controlling failure mechanism for short-lived specimens and uniform oxidation and damage development is the controlling failure mechanism for long-lived specimens.

Introduction

Continuous carbon fiber reinforced silicon carbide matrix composites (C/SiC) exhibit excellent mechanical properties at elevated temperature and have been developed for aeronautic and aerospace applications [1,2]. In service, C/SiC is subject to complex coupling effects between thermal, stress and chemical reactions, and its damage evolution and failure mechanisms are of wide interest [[3], [4], [5], [6]]. Generally, there is oxidation along with several kinds of microscopic fracture mechanisms involving stress oxidation of C/SiC in a wet oxidizing atmosphere, including matrix cracking, interface debonding, delamination and fiber breakage. Their evolution depends on the microscopic structure, especially defects in preparation, such as coating cracks, matrix cracks and porosities. Due to the random size and distribution of these defects, C/SiC usually shows a wide scatter in its performance, even under the same stress and environmental conditions.

In order to describe the failure process, researchers often rely on some physical information released from the internal changes in materials to describe the damage process, such as electrical resistance [[7], [8], [9]], mechanical parameters [[10], [11], [12]], thermal dissipation [13], acoustic emission (AE) [[14], [15], [16], [17], [18], [19], [20]] or their combinations [[21], [22], [23], [24]]. AE, which is a phenomenon whereby transient elastic waves are generated by strain energy release, provides a passive method of monitoring the changes within a material [25,26]. Since one micro-fracture event will generate one AE wave and various fracture mechanisms will generate AE signals with different characteristics, it is believed that AE is a powerful means to characterize the damage evolution of materials. Moreover, AE can be applied under conditions of elevated temperature using waveguides.

However, there are challenges in the application of AE in the study of the damage mechanisms of ceramic matrix composites (CMCs). One of the most important is that the complexity of the AE source itself, as well as the complex propagation and attenuation processes within the material, make it difficult to correlate AE events with microscopic fracture mechanisms. Researchers have tried to solve this problem with pattern recognition technology, mainly unsupervised cluster analysis [14,27,28]. The method has been used to analyze several kinds of CMCs and has been able to identify that AE signal grouping has a strong correlation with damage mechanisms [14,27].

In the present paper, a fuzzy C-means (FCM) algorithm was used instead of the more common K-means algorithm to improve the effectiveness of cluster analysis of AE data from creep tests of C/SiC in a wet oxidizing atmosphere. The damage mechanisms and their evolution were characterized by AE cluster analysis. Combined with the AE data and fracture surface observation, the dominant controlling failure mechanism for C/SiC with different rupture lifetimes under particular stress oxidation conditions is proposed.

Section snippets

Preparation of specimens

Plain woven carbon fiber preforms with a fiber volume fraction of 40 % were employed. A pyrolytic carbon (PyC) interface and SiC matrix were deposited via chemical vapor infiltration (CVI) processes. The precursors of PyC and SiC were Propene (C3H6) and Methyltrichlorosilane (MTS, CH3SiCl3), respectively [29]. Then the fabricated C/SiC was machined into dog-bone shaped samples, as shown in Fig. 1. Finally, a chemical vapor deposition (CVD) process was used to deposit a SiC coating.

Creep tests

Tensile creep

Pattern recognition techniques

AE signals generated by different microscopic fracture mechanisms have their special features. which can be described by a number of parameters (or descriptors) extracted from the AE waveforms, such as energy, average frequency, amplitude, rise time, duration, counts, etc. Pattern recognition is required to establish a mapping between AE signals and fracture mechanisms. Since no labelled data is available, an unsupervised methodology is used to perform the pattern cluster analysis. It involves

Creep test results

The creep curves of the six samples are shown in Fig. 2. The creep process includes two stages, namely the transient creep stage (t < 100 s) and the stationary, or steady, creep stage (t > 100 s). The creep strain increases rapidly and the creep rate decreases gradually in the first stage, corresponding to the loading process. The creep strain increases very slowly in the steady stage, with almost constant creep rate. The steady creep rate is defined to be the slope of the steady stage. The

Conclusions

Creep rupture lifetimes of C/SiC in a wet oxidizing atmosphere vary significantly. AE was used find a correlation between failure mechanism and rupture lifetime. AE data collected during creep tests were divided into four clusters by an FCM algorithm, corresponding to the physical damage mechanisms of matrix cracking, interfacial damage, fiber breakage and fiber bundle breakage respectively. For samples with different lifetimes, the evolution of AE events and the AE features of fiber-bundle

Declaration of Competing Interest

None.

Acknowledgements

This work has been financially supported by National Natural Science Foundation of China (51772244 and 11072195).

References (39)

Cited by (18)

  • Application of a remotely bonded fiber-optic Bragg grating sensor to acoustic emission testing for a carbon-carbon composite at a temperature of 1000 °C

    2022, Measurement: Journal of the International Measurement Confederation
    Citation Excerpt :

    Using this method, Momon et al. [30], Loutas et al. [26], Yong et al. [31], Hou et al. [32], and Li et al. [33] assigned damage mechanisms to AE signals generated in 2D and 3D woven composites. Recently, Lyu et al. [34] demonstrated that the fuzzy c-means clustering method cloud be used to examine the effect of material oxidation on the failure mechanisms and creep rupture lifetimes of ceramic matrix composites. However, it is still unclear if the fuzzy c-means is superior to the k-means due to difficulties in verifying cluster results [28].

  • Time-dependent creep fatigue damage evolution in C/SiC composite: Theory and analytical prediction

    2022, Ceramics International
    Citation Excerpt :

    C/SiC composites have great potential for application as high-temperature aerospace and aircraft thermo-structural constituents [10–14]. However, carbon fibers’ degradation under oxidizing conditions could cause strength reduction and constituent failure [15‒18]. For the C/SiC composites, multiple fatigue damage mechanisms (e.g., matrix cracks, interface debonding, and fibers fracture) occur with increasing applied cycles and degrade the mechanical performance (e.g., residual strength) of the composites [19‒22].

View all citing articles on Scopus
View full text