Random forest-based integration of TSP and TEM for water inrush prediction in tunnels

Currently, advanced geological prediction in railway tunnels primarily relies on methods such as TSP (Tunnel Seismic Prediction), TEM (Transient Electromagnetic Method), and horizontal drilling for water inrush forecasting. Each method has limitations, requiring manual data analysis and cross-validation. This study pioneers the fusion of TSP and TEM data by extracting characteristic parameters and employing machine learning to establish an automated prediction model for water inrush ahead of tunnel faces, significantly enhancing forecast accuracy and construction safety.