Dynamic safety assessment technology of mine ventilation system based on 3D digital twinning

The traditional evaluation method of mine ventilation system is difficult to reflect the dynamic working conditions in real time and lacks quantitative analysis means. This paper proposes a dynamic safety evaluation technology of mine ventilation system based on 3D digital twinning. By integrating BIM (Building Information Model) and GIS(Geographic Information System) technologies to build a high-precision 3D ventilation network model, and combining laser point cloud and multi-source sensor data, the virtual and real mapping and visual expression of roadway, equipment and personnel can be realized. Kalman filter and space-time graph convolution network (ST-GCN) are introduced to establish multi-source data fusion and anomaly correction mechanism to improve dynamic perception. Further construct a multi-index dynamic evaluation model covering ventilation efficiency, disaster risk and equipment reliability, and realize early warning and intelligent regulation of risks such as gas accumulation and smoke spread. The experimental results show that the simulation error of air volume of the system is less than 5%, which can give an early warning 90s before the gas concentration exceeds the limit, and significantly reduce the harmful gas concentration in the working face through intelligent regulation, which verifies its effectiveness in improving mine ventilation safety and emergency response capability.