For decades, monitoring and identification of the dynamic stress states of rock masses in reservoirs have been challenging tasks owing to the lack of effective observation and analysis methods. In this study, we used deep learning methods to identify whether acoustic emission (AE) signals of rock fractures are created by different loading conditions. We performed Brazilian split and uniaxial compressive experiments on six different engineering materials, and we obtained the AE waveforms. To take advantage of the powerful image processing capabilities of convolutional neural networks (CNNs), we transformed the AE waveforms into time–frequency images. We used five types of CNN to identify the time–frequency images of the AE signals created in the Brazilian split and uniaxial compressive experiments. As a result, we found that Xception model had the highest recognition accuracy. We analyzed the basis of the CNN models to recognize the AE signals using local interpretable model‐agnostic explanations and found that the Xception model mainly used the patterns of the low‐energy region of the time–frequency images to determine the loading modes of the rock fractures. Furthermore, the high‐energy time–frequency region had little effect on recognition. The findings of this work can aid in automatically monitoring the dynamic stress states of fracture areas in reservoir formations and ensure the safety of oil and gas resource exploitation.

The controlled presplitting blasting technique is widely used in mining engineering to improve the permeability and gas extraction efficiency of coal seams. One of the key factors is the appropriate arrangement of the blasting and drainage holes, which can help improve the gas drainage quantity. To optimize the drilling layout to enhance gas-drainage efficiency, a series of controlled presplitting blasting tests were conducted at the Pingdingshan No. 8 coal mine. Based on the analysis of variations in stress and longitudinal-wave velocity of the coal in different blasting ranges, the results show that the stress on the coal at a distance of 1 m from the blasting hole decreased significantly after blasting; thus, the coal exhibited negligible bearing capacity and the longitudinal-wave velocity decreased by 56%. However, the coal exhibited particular bearing capacity at a distance of 3 m away from the blasting hole, and the longitudinal-wave velocity decreased by 35%. The stress and longitudinal-wave velocity at a distance of 5 m from the blasting hole were unaffected by the blasting. The blasting integrity rate of coal k_v was defined to characterize the effect of blasting on the coal-seam fracture. The effective cracking and effective influence radii of blasting under these working conditions were predicted to be in the ranges 3.3–3.4 m and 7.2–7.3 m, respectively. According to the test results, the borehole layout was optimized in the field testes for gas drainage in coal roadway strips, and the amounts of pure gas extracted after blasting were thus increased by 1.54–2.24 times the amount before blasting.