A quantitative acoustic emission study on fracture processes in ceramics based on wavelet packet decomposition

We base a quantitative acoustic emission (AE) study on fracture processes in alumina ceramics on wavelet packet decomposition and AE source location. According to the frequency characteristics, as well as energy and ringdown counts of AE, the fracture process is divided into four stages: crack closure, nucleation, development, and critical failure. Each of the AE signals is decomposed by a 2-level wavelet package decomposition into four different (from-low-to-high) frequency bands (AA₂, AD₂, DA ₂, and DD₂). The energy eigenvalues P₀, P ₁, P₂, and P₃ corresponding to these four frequency bands are calculated. By analyzing changes in P₀ and P ₃ in the four stages, we determine the inverse relationship between AE frequency and the crack source size during ceramic fracture. AE signals with regard to crack nucleation can be expressed when P₀ is less than 5 and P₃ more than 60; whereas AE signals with regard to dangerous crack propagation can be expressed when more than 92% of P₀ is greater than 4, and more than 95% of P₃ is less than 45. Geiger location algorithm is used to locate AE sources and cracks in the sample. The results of this location algorithm are consistent with the positions of fractures in the sample when observed under a scanning electronic microscope; thus the locations of fractures located with Geiger's method can reflect the fracture process. The stage division by location results is in a good agreement with the division based on AE frequency characteristics. We find that both wavelet package decomposition and Geiger's AE source locations are suitable for the identification of the evolutionary process of cracks in alumina ceramics.