ESPAC-Based Linear Array Microtremor Tomography Method and Its Application in Goaf Detection

Collapse columns and goafs present significant risks to coal mine safety. Accurate detection of their geometrical configurations is essential for ensuring safe operations and guiding mining design. Traditional Spatial Autocorrelation (SPAC) detection techniques, however, face notable limitations, including narrow applicability, vulnerability to noise interference, and reduced resolution caused by one-dimensional model splicing. Using the Talahao Coal Mine as a case study, a linear array configuration is implemented built upon background noise imaging and the Extended Spatial Autocorrelation (ESPAC) method to collect microtremor data. A specialized noise suppression algorithm is developed to eliminate various sources of ambient and instrumental noise, thereby enhancing the signal-to-noise ratio (SNR) of the microtremor recordings. By means of human–computer interaction, dispersion curves between station pairs are extracted, fitted, and assessed for quality. To support quality control, the theoretical waveform of the spatial autocorrelation function is incorporated. A special point constraint method is introduced to guarantee the dependability of the dispersion data and enable the successful extraction of low-frequency surface waves. Based on this processed data, tomographic inversion is conducted to derive a two-dimensional structural model beneath the survey line. The imaging process accounts for topographic variations, resulting in seamless model stitching and improved resolution. Potential goaf locations are identified through the spatial distribution of shear wave low-velocity zones. This information serves as a foundation for delineating hazardous regions and developing mitigation strategies, thereby contributing to the reduction of structural instability risks during mining operations.