Intelligent Monitoring of Coal Mine Ventilation Safety Based on Improved Support Vector Machine

YE Yang

doi:10.16545/j.cnki.cmet.2025.04.016

Abstract：For the topology structure of ventilation networks in coal mines,ventilation related parameters were affected by branching and merging relationships,exhibiting complex nonlinear characteristics that make it difficult to accurately capture the relationships between ventilation related parameters,resulting in significant errors between monitoring results and reality.Therefore,research has been conducted on intelligent monitoring of coal mine ventilation safety based on improved support vector machines.A calculation model for the topology structure of coal mine underground ventilation network was constructed with ventilation related parameters air flow temperature as the core,and ventilation related parameter data for any node of the ventilation network topology structure considering bifurcation and convergence relationships were obtained.Using SVM mapping to output the relationship function between air temperature and air volume,in order to capture the complex relationship between air temperature and air volume,and iteratively search for the optimal values of kernel function parameters in a given search space,eliminating the interference of ventilation related parameter changes on the relationship function of air temperature and air volume,and improving the accuracy of monitoring.By using relaxation function to constrain the mapping output of wind temperature parameters in improved support vector machine,accurate coal mine ventilation status data can be obtained.In the test results,the error between the airflow density output by the monitoring method and the actual value is within 0.04 kg/m3,and the error in airflow volume of different dampers is always below 0.4 m3/s,all at a low level.