Research on multi-physics field collaborative detection methods for concealed fire zones: Based on variable-temperature magnetic-dielectric-resistive characteristics

Accurate detection of hidden coal fire areas is crucial for early warning and disaster mitigation, yet single-physics methods often have limitations in scope and applicability. This study proposes a multi-physics cooperative detection approach based on temperature variations, integrating magnetic, dielectric, and resistivity measurements, and verifies it on a large-scale physical simulation platform. The results show that as the fire temperature increases (500–700 °C), the magnetic susceptibility of coal and rocks significantly enhances, raising the magnetic anomaly intensity from 1100nT to 1600nT, effectively delineating fire boundaries. Variable-temperature dielectric measurements reveal a three-stage evolution pattern, with burned-out zones exhibiting reduced permittivity, causing polarity inversion of ground-penetrating radar (GPR) reflections, which can be effectively identified in the 20–24 ns time window. High-density resistivity surveys indicate a distinct transition from high-resistivity anomalies (~10^5Ω·m) at ambient conditions to low-resistivity anomalies (50 - 200 Ω·m) at elevated temperatures, with inversion results consistent with forward modeling. The integration of magnetic, dielectric, and resistivity methods demonstrates strong complementarity in boundary delineation, void detection, and spatial inversion, ultimately achieving precise localization of fire centers. This study establishes a cooperative multi-physics detection framework for hidden coal fires, providing a new technical approach for integrated detection, disaster early warning, and fire control design, with potential applicability in geothermal monitoring and hydrocarbon leakage detection.