Objective Fire zones in shallowly buried coal seams have long restricted safe and efficient coal mining due to their considerably hidden and hazardous nature. However, individual methods for fire source detection are susceptible to geological conditions and ambient noise, suffering from limited detection dimensions and the presence of technical blind spots. Therefore, they fail to accurately localize hidden fire sources in shallow coal seams.

Methods This study proposes a three-dimensional, collaborative multi-source detection approach that combines the self-potential (SP) method, the radon measurement (Rn) method, electrical resistivity tomography (ERT; also referred to as high-density resistivity method), and the transient electromagnetic method (TEM). Accordingly, the technical pathway is developed, consisting of fire zone delineation in the horizontal direction, fire source positioning in the vertical direction, and multi-source constraint. Specifically, fire zones in coal seams are preliminarily identified using the SP and Rn methods in the horizontal direction, with interference zones being ruled out through anomaly matching. For the hazardous zones delineated horizontally, vertical detection is conducted using the ERT and TEM methods to jointly determine the depths of fire sources. Finally, the detection results are verified through borehole thermometry.

Results and Conclusions  The proposed approach was applied to a shallowly buried coal seam in Inner Mongolia. The application results indicate that nine suspected fire zones were identified through SP-Rn joint detection after interference zones were ruled out. Then, three fire sources were localized through the vertical ERT-TEM joint detection. The borehole thermometry revealed the presence of high-temperature anomalies with temperatures ranging from 300 ℃ to 450 ℃ in the anomalous zones detected, highly consistent with the detection results. Compared to individual SP and Rn methods, the combination of both increased the detection accuracy by 42.76% and 70.03%, respectively in detecting anomalous zones in the horizontal direction, significantly reducing the misidentification rate of fire zones in the horizontal direction. Compared to individual ERT and TEM methods, the combination of both enhanced the accuracy by 86.07% and 87.49% for survey line 1 and 76.78% and 69.46% for survey line 2, respectively. Therefore, the three-dimensional, collaborative multi-source detection approach can significantly reduce the multiplicity of solutions of individual detection methods, effectively addressing the engineering challenges of difficult identification and low positioning accuracy in identifying fire sources in shallowly buried coal seams. The proposed approach provides a technical solution deserving widespread application for the rapid identification and precise control of fire zones in shallowly buried coal seams.