Complex Network‐Based Characterization of Acoustic Emission Precursors in Coal–Rock Fracture Under Uniaxial Compression Load

Dynamic hazards in coal and rock have emerged as a critical impediment to the safety and efficiency of deep mining operations. The analysis of elastic wave evolution—manifested as acoustic emission (AE) or microseismicity—during the failure process of coal and rock under load is a pivotal technique for both laboratory and in situ investigations of such hazards. In this study, uniaxial compression experiments incorporating AE monitoring were performed on coal specimens under various constant stress rates. The AE data spanning the entire fracture process were collected. Subsequently, leveraging complex network theory, we analyzed the correlations among AE characteristic parameters to construct a network. The spectral clustering method was employed to classify the network nodes. This analysis yielded three distinct classes of AE signals, the characteristics of which align with the successive stages of coal fracture. The results indicate that the critical node proportion (CNP) of high‐intensity fracture signals (Class III) is a crucial metric. A CNP value surpassing 30% acts as a robust precursor to plastic instability, whereas a value exceeding 50% signals the onset of final macroscopic failure. This quantitative criterion provides a core technical foundation for the real‐time and high‐precision early warning of dynamic hazards in coal and rock.