Detection of cracking behaviors in granite with open precut cracks by acoustic emission frequency spectrum analysis

Preexisting cracks are widely distributed in brittle rocks, and these cracks strongly influence the rock failure mechanisms. The cracking process is accompanied by rapid energy release in the form of elastic waves, known as acoustic emission (AE), enabling detection of the cracking behaviors and providing an early warning of sudden failure. Granites with three types of open precut cracks were tested in uniaxial and triaxial compression tests with AE monitoring; the cracking process was recorded by a high-speed camera during the uniaxial compression tests. The peak values of the AE count rate and normalized peak amplitude corresponded well to the peak stress points, while the normalized peak amplitude was more sensitive and effective in revealing the cracking behaviors. The peak frequency can be divided into three bands. Five types of AE waveform signals were identified, according to the value of the normalized peak amplitude and the range of the peak frequency band: low normalized peak amplitude with low peak frequency (LL), high normalized peak amplitude with low peak frequency (LH), low normalized peak amplitude with moderate peak frequency (ML), low normalized peak amplitude with high peak frequency (HL), and high normalized peak amplitude with high peak frequency (HH). A comparison of the high-speed images, b values, and AE frequency spectrum characteristics showed that the LH and HH signals corresponded to macrocracks, while the LL and HL signals corresponded to microcracks. Additionally, LH and ML signals always appeared just before failure, meaning that they can be used as precursor information.