Refined Detection Technology for Discharge Tunnel in Karst Area
When constructing tunnels in karst areas,geological disasters such as collapse,mud and water inrush caused by karst often occur,which bring varying degrees of safety risks to tunnel construction.In order to prevent geological disasters and reduce their impact on tunnel construction,it is necessary to conduct accurate and advanced geological prediction of the tunnel.The drainage tunnel in karst areas has the characteristics of small cross-sec-tional size,large interference,high detection requirements,and complex geological condi-tions,which makes advance geological prediction more difficult and less accurate.Compre-hensive advance geological prediction techniques such as geological survey,seismic wave re-flection,convolutional neural network-based geological radar,transient electromagnetic method,and advance geological drilling can be used to accurately detect karst.This article u-ses the comprehensive advanced geological prediction technology to apply it in the XSDK59+017—XSDK58+887 section of the Qinghou Mountain drainage tunnel,predicting the mud filled karst caves in the XSDK58+928—XSDK58+918 section,improving the prediction ac-curacy and reducing construction risks.The results show that using the trained YOLOV8 model to identify karst caves in radar images in drainage tunnels in karst areas can achieve a confidence rate of over 80％,avoiding human subjectivity and improving data interpretation efficiency.Moreover,this comprehensive advanced geological prediction method can effec-tively predict the spatial distribution and filling materials of karst caves in front of the tun-nel,which provides a basis for design and construction,avoids the adverse effects of geolog-ical disasters on the project,and has certain reference significance for similar projects.
NETL
NSTL
万方数据
