The main intake shaft in the mine serves as the main intake and return air shaft and also serves as a lifting shaft. The piston wind effect caused by the frequent operation of large and efficient hoisting cages in the shaft will disturb the normal flow of airflow in the transport lane connected to the shaft and affect the underground ventilation effect. Therefore, based on the SST k-ω turbulence model, this paper uses the dynamic mesh to simulate the air-fluid in the cage running in the shaft and verifies the simulation data through field measurement to study the influence of piston effect on the airflow field in the transport lane. The results show that the piston effect caused by cage operation in the shaft will disturb the flow of normal airflow in the transportation lane and affect the ventilation effect in the mine. The 1765.0 m transport lane is closest to the position of the shaft inlet, which is most significantly affected by the piston wind effect. The low-speed eddy current zone generated at the tail of the cage in the upwind operation has a large area and strong adsorption force, which causes the gas at the local position in the transportation lane to generate eddy current and reverse flow, hindering the flow of fresh air in the lane. The influence distance and strength of the piston effect on each measuring point in the transportation lane are limited. When the measuring point distance increases from 21.7 m to 71.7 m, the differential pressure of measuring points in 1414.0 m, 1584.0 m and 1765.0 m transport lanes decreases by 11.08 Pa, 9.62 Pa and 8.58 Pa, respectively.

The safety and reliability of a ventilation system relies on an accurate friction resistance coefficient (α), but obtaining α requires a great deal of tedious measurement work in order to determine the result, and many erroneous data are obtained. Therefore, it is vital that α be obtained quickly and accurately for the ventilation system design. In this study, a passive and active support indicator system was constructed for the prediction of α. An RF model, GSCV-RF model and BP model were constructed using the RF algorithm, GSCV algorithm and BP neural network, respectively, for α prediction. In the GSCV-RF and BP models, 160 samples complied with the prediction indicator system and were used to construct a prediction dataset and, this dataset was divided into a training set and a test set. The prediction results were based on the quantitative evaluation models of MAE, RMSE and R2. The results show that, among the three models, the GSCV-RF model’s prediction result for α was the best, the RF model performed well and the BP model performed worst. In the prediction for all the datasets obtained by GSCV-RF model, all the values of MAE and RMSE were less than 0.5, the values of R2 were more than 0.85 and the value of R2 of the passive and active support test sets were 0.8845 and 0.9294, respectively. This proved that the GSCV-RF model can offer a more accurate α and aid in the reasonable design and the safe operation of a ventilation system.